This repository has been archived on 2023-10-09. You can view files and clone it, but cannot push or open issues or pull requests.
Files
blender-archive/source/blender/editors/space_image/image_ops.c

4008 lines
108 KiB
C
Raw Normal View History

/*
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
2010-02-12 13:34:04 +00:00
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
*
* The Original Code is Copyright (C) 2001-2002 by NaN Holding BV.
* All rights reserved.
*/
/** \file
* \ingroup spimage
2011-02-27 20:29:51 +00:00
*/
#include <errno.h>
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#include <fcntl.h>
#include <stddef.h>
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include <stdlib.h>
#include <string.h>
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#ifndef WIN32
# include <unistd.h>
#else
# include <io.h>
#endif
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "MEM_guardedalloc.h"
#include "BLI_blenlib.h"
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
#include "BLI_fileops.h"
#include "BLI_fileops_types.h"
#include "BLI_ghash.h"
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
#include "BLI_linklist.h"
#include "BLI_math.h"
#include "BLI_string.h"
#include "BLI_utildefines.h"
#include "BLT_translation.h"
#include "DNA_camera_types.h"
#include "DNA_node_types.h"
#include "DNA_object_types.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "DNA_packedFile_types.h"
#include "DNA_scene_types.h"
#include "DNA_screen_types.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "BKE_colortools.h"
#include "BKE_context.h"
#include "BKE_global.h"
#include "BKE_icons.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "BKE_image.h"
#include "BKE_image_save.h"
#include "BKE_layer.h"
#include "BKE_lib_id.h"
#include "BKE_main.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "BKE_packedFile.h"
#include "BKE_paint.h"
#include "BKE_report.h"
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#include "BKE_scene.h"
#include "BKE_screen.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "DEG_depsgraph.h"
#include "GPU_draw.h"
#include "GPU_immediate.h"
#include "GPU_state.h"
Color Management, Stage 2: Switch color pipeline to use OpenColorIO Replace old color pipeline which was supporting linear/sRGB color spaces only with OpenColorIO-based pipeline. This introduces two configurable color spaces: - Input color space for images and movie clips. This space is used to convert images/movies from color space in which file is saved to Blender's linear space (for float images, byte images are not internally converted, only input space is stored for such images and used later). This setting could be found in image/clip data block settings. - Display color space which defines space in which particular display is working. This settings could be found in scene's Color Management panel. When render result is being displayed on the screen, apart from converting image to display space, some additional conversions could happen. This conversions are: - View, which defines tone curve applying before display transformation. These are different ways to view the image on the same display device. For example it could be used to emulate film view on sRGB display. - Exposure affects on image exposure before tone map is applied. - Gamma is post-display gamma correction, could be used to match particular display gamma. - RGB curves are user-defined curves which are applying before display transformation, could be used for different purposes. All this settings by default are only applying on render result and does not affect on other images. If some particular image needs to be affected by this transformation, "View as Render" setting of image data block should be set to truth. Movie clips are always affected by all display transformations. This commit also introduces configurable color space in which sequencer is working. This setting could be found in scene's Color Management panel and it should be used if such stuff as grading needs to be done in color space different from sRGB (i.e. when Film view on sRGB display is use, using VD16 space as sequencer's internal space would make grading working in space which is close to the space using for display). Some technical notes: - Image buffer's float buffer is now always in linear space, even if it was created from 16bit byte images. - Space of byte buffer is stored in image buffer's rect_colorspace property. - Profile of image buffer was removed since it's not longer meaningful. - OpenGL and GLSL is supposed to always work in sRGB space. It is possible to support other spaces, but it's quite large project which isn't so much important. - Legacy Color Management option disabled is emulated by using None display. It could have some regressions, but there's no clear way to avoid them. - If OpenColorIO is disabled on build time, it should make blender behaving in the same way as previous release with color management enabled. More details could be found at this page (more details would be added soon): http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.64/Color_Management -- Thanks to Xavier Thomas, Lukas Toene for initial work on OpenColorIO integration and to Brecht van Lommel for some further development and code/ usecase review!
2012-09-15 10:05:07 +00:00
#include "IMB_colormanagement.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "IMB_imbuf.h"
#include "IMB_imbuf_types.h"
Image cache rewrite to using generic movie cache Summary: Behaves very much the same as cache for Movie Clip datablock: - Image now have `MovieCache *cache` field which replaced legacy `ListBase ibufs`. This allows image datablock to easily keep of image buffers which are owned by itself. This field isn't saved to the file and getting restored on undo steps. However, cache limit is global for movies, sequences and image datablocks now. So overall cached image buffers size will not go above cache limit size in user preferences. - Image buffers which are marked as BITMAPDIRTY will never be freed from the cache. - Added utility function to iterate over image buffers saved in movie cache. - Movie cache cleanup check callback now have ImBuf argument which can be used in a condition of cleanup. - Added some utility functions which replaces legacy ibufs iterations with image cache iteration which happens from inside a lock. - Fixed `image_mem_size()` which was only counting one of the buffers if both float and byte buffer present. Additional notes: - `BKE_image_get_first_ibuf()` is rather stupid, but direct access to ibufs->first was also the same stupid idea. Would consider avoid this function is another project. - There are some places which doesn't look threadsafe, but they already were not so much threadsafe anyway before. So think not a big deal with solving this later. Finally solves infinite memory usage by image sequences! :) Reviewers: brecht, campbellbarton Reviewed By: brecht CC: sebastian_k Differential Revision: http://developer.blender.org/D95
2013-12-13 16:22:08 +06:00
#include "IMB_moviecache.h"
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#include "intern/openexr/openexr_multi.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "RE_pipeline.h"
#include "RNA_access.h"
#include "RNA_define.h"
#include "RNA_enum_types.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "ED_image.h"
#include "ED_mask.h"
#include "ED_paint.h"
#include "ED_render.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "ED_screen.h"
#include "ED_space_api.h"
#include "ED_util.h"
#include "ED_util_imbuf.h"
#include "ED_uvedit.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "UI_interface.h"
#include "UI_resources.h"
#include "UI_view2d.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "WM_api.h"
#include "WM_types.h"
#include "PIL_time.h"
#include "RE_engine.h"
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
#include "image_intern.h"
/* -------------------------------------------------------------------- */
/** \name View Navigation Utilities
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
static void sima_zoom_set(
SpaceImage *sima, ARegion *region, float zoom, const float location[2], const bool zoom_to_pos)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
float oldzoom = sima->zoom;
int width, height;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
sima->zoom = zoom;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (sima->zoom < 0.1f || sima->zoom > 4.0f) {
/* check zoom limits */
ED_space_image_get_size(sima, &width, &height);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
width *= sima->zoom;
height *= sima->zoom;
if ((width < 4) && (height < 4) && sima->zoom < oldzoom) {
sima->zoom = oldzoom;
}
else if (BLI_rcti_size_x(&region->winrct) <= sima->zoom) {
sima->zoom = oldzoom;
}
else if (BLI_rcti_size_y(&region->winrct) <= sima->zoom) {
sima->zoom = oldzoom;
}
}
if (zoom_to_pos && location) {
float aspx, aspy, w, h;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
ED_space_image_get_size(sima, &width, &height);
ED_space_image_get_aspect(sima, &aspx, &aspy);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
w = width * aspx;
h = height * aspy;
sima->xof += ((location[0] - 0.5f) * w - sima->xof) * (sima->zoom - oldzoom) / sima->zoom;
sima->yof += ((location[1] - 0.5f) * h - sima->yof) * (sima->zoom - oldzoom) / sima->zoom;
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static void sima_zoom_set_factor(SpaceImage *sima,
ARegion *region,
float zoomfac,
const float location[2],
const bool zoom_to_pos)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
sima_zoom_set(sima, region, sima->zoom * zoomfac, location, zoom_to_pos);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/**
* Fits the view to the bounds exactly, caller should add margin if needed.
*/
static void sima_zoom_set_from_bounds(SpaceImage *sima, ARegion *region, const rctf *bounds)
{
int image_size[2];
float aspx, aspy;
ED_space_image_get_size(sima, &image_size[0], &image_size[1]);
ED_space_image_get_aspect(sima, &aspx, &aspy);
image_size[0] = image_size[0] * aspx;
image_size[1] = image_size[1] * aspy;
/* adjust offset and zoom */
sima->xof = roundf((BLI_rctf_cent_x(bounds) - 0.5f) * image_size[0]);
sima->yof = roundf((BLI_rctf_cent_y(bounds) - 0.5f) * image_size[1]);
float size_xy[2], size;
size_xy[0] = BLI_rcti_size_x(&region->winrct) / (BLI_rctf_size_x(bounds) * image_size[0]);
size_xy[1] = BLI_rcti_size_y(&region->winrct) / (BLI_rctf_size_y(bounds) * image_size[1]);
size = min_ff(size_xy[0], size_xy[1]);
CLAMP_MAX(size, 100.0f);
sima_zoom_set(sima, region, size, NULL, false);
}
static Image *image_from_context(const bContext *C)
{
/* Edit image is set by templates used throughout the interface, so image
* operations work outside the image editor. */
Image *ima = CTX_data_pointer_get_type(C, "edit_image", &RNA_Image).data;
if (ima) {
return ima;
}
/* Image editor. */
SpaceImage *sima = CTX_wm_space_image(C);
return (sima) ? sima->image : NULL;
}
static ImageUser *image_user_from_context(const bContext *C)
{
/* Edit image user is set by templates used throughout the interface, so
* image operations work outside the image editor. */
ImageUser *iuser = CTX_data_pointer_get_type(C, "edit_image_user", &RNA_ImageUser).data;
if (iuser) {
return iuser;
}
/* Image editor. */
SpaceImage *sima = CTX_wm_space_image(C);
return (sima) ? &sima->iuser : NULL;
}
2019-10-02 01:16:36 +10:00
static bool image_from_context_has_data_poll(bContext *C)
{
Image *ima = image_from_context(C);
ImageUser *iuser = image_user_from_context(C);
if (ima == NULL) {
return false;
}
void *lock;
ImBuf *ibuf = BKE_image_acquire_ibuf(ima, iuser, &lock);
const bool has_buffer = (ibuf && (ibuf->rect || ibuf->rect_float));
BKE_image_release_ibuf(ima, ibuf, lock);
return has_buffer;
}
2019-10-02 01:16:36 +10:00
/**
* Use this when the image buffer is accessed without the image user.
*/
static bool image_from_context_has_data_poll_no_image_user(bContext *C)
2019-10-02 01:16:36 +10:00
{
Image *ima = image_from_context(C);
return BKE_image_has_ibuf(ima, NULL);
}
static bool image_not_packed_poll(bContext *C)
{
/* Do not run 'replace' on packed images, it does not give user expected results at all. */
Image *ima = image_from_context(C);
return (ima && BLI_listbase_is_empty(&ima->packedfiles));
}
2018-07-02 11:47:00 +02:00
bool space_image_main_region_poll(bContext *C)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
/* XXX ARegion *region = CTX_wm_region(C); */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (sima) {
return true; /* XXX (region && region->type->regionid == RGN_TYPE_WINDOW); */
}
return false;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/* For IMAGE_OT_curves_point_set to avoid sampling when in uv smooth mode or editmode */
2018-07-02 11:47:00 +02:00
static bool space_image_main_area_not_uv_brush_poll(bContext *C)
{
SpaceImage *sima = CTX_wm_space_image(C);
Scene *scene = CTX_data_scene(C);
ToolSettings *toolsettings = scene->toolsettings;
if (sima && !toolsettings->uvsculpt && (CTX_data_edit_object(C) == NULL)) {
return 1;
}
return 0;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name View Pan Operator
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
typedef struct ViewPanData {
float x, y;
float xof, yof;
int launch_event;
bool own_cursor;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
} ViewPanData;
static void image_view_pan_init(bContext *C, wmOperator *op, const wmEvent *event)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
wmWindow *win = CTX_wm_window(C);
SpaceImage *sima = CTX_wm_space_image(C);
ViewPanData *vpd;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
op->customdata = vpd = MEM_callocN(sizeof(ViewPanData), "ImageViewPanData");
/* Grab will be set when running from gizmo. */
vpd->own_cursor = (win->grabcursor == 0);
if (vpd->own_cursor) {
WM_cursor_modal_set(win, WM_CURSOR_NSEW_SCROLL);
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
vpd->x = event->x;
vpd->y = event->y;
vpd->xof = sima->xof;
vpd->yof = sima->yof;
vpd->launch_event = WM_userdef_event_type_from_keymap_type(event->type);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
WM_event_add_modal_handler(C, op);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static void image_view_pan_exit(bContext *C, wmOperator *op, bool cancel)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
ViewPanData *vpd = op->customdata;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (cancel) {
sima->xof = vpd->xof;
sima->yof = vpd->yof;
ED_region_tag_redraw(CTX_wm_region(C));
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (vpd->own_cursor) {
WM_cursor_modal_restore(CTX_wm_window(C));
}
MEM_freeN(op->customdata);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_view_pan_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
float offset[2];
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
RNA_float_get_array(op->ptr, "offset", offset);
sima->xof += offset[0];
sima->yof += offset[1];
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
ED_region_tag_redraw(CTX_wm_region(C));
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_view_pan_invoke(bContext *C, wmOperator *op, const wmEvent *event)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
if (event->type == MOUSEPAN) {
SpaceImage *sima = CTX_wm_space_image(C);
float offset[2];
offset[0] = (event->prevx - event->x) / sima->zoom;
offset[1] = (event->prevy - event->y) / sima->zoom;
RNA_float_set_array(op->ptr, "offset", offset);
image_view_pan_exec(C, op);
return OPERATOR_FINISHED;
}
image_view_pan_init(C, op, event);
return OPERATOR_RUNNING_MODAL;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_view_pan_modal(bContext *C, wmOperator *op, const wmEvent *event)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
ViewPanData *vpd = op->customdata;
float offset[2];
switch (event->type) {
case MOUSEMOVE:
sima->xof = vpd->xof;
sima->yof = vpd->yof;
offset[0] = (vpd->x - event->x) / sima->zoom;
offset[1] = (vpd->y - event->y) / sima->zoom;
RNA_float_set_array(op->ptr, "offset", offset);
image_view_pan_exec(C, op);
break;
default:
if (event->type == vpd->launch_event && event->val == KM_RELEASE) {
image_view_pan_exit(C, op, false);
return OPERATOR_FINISHED;
}
break;
}
return OPERATOR_RUNNING_MODAL;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static void image_view_pan_cancel(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
image_view_pan_exit(C, op, true);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
void IMAGE_OT_view_pan(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Pan View";
ot->idname = "IMAGE_OT_view_pan";
ot->description = "Pan the view";
/* api callbacks */
ot->exec = image_view_pan_exec;
ot->invoke = image_view_pan_invoke;
ot->modal = image_view_pan_modal;
ot->cancel = image_view_pan_cancel;
ot->poll = space_image_main_region_poll;
/* flags */
ot->flag = OPTYPE_BLOCKING | OPTYPE_GRAB_CURSOR_XY | OPTYPE_LOCK_BYPASS;
/* properties */
RNA_def_float_vector(ot->srna,
"offset",
2,
NULL,
-FLT_MAX,
FLT_MAX,
"Offset",
"Offset in floating point units, 1.0 is the width and height of the image",
-FLT_MAX,
FLT_MAX);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name View Zoom Operator
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
typedef struct ViewZoomData {
float origx, origy;
float zoom;
int launch_event;
float location[2];
/* needed for continuous zoom */
wmTimer *timer;
double timer_lastdraw;
bool own_cursor;
/* */
SpaceImage *sima;
ARegion *region;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
} ViewZoomData;
static void image_view_zoom_init(bContext *C, wmOperator *op, const wmEvent *event)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
wmWindow *win = CTX_wm_window(C);
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
ViewZoomData *vpd;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
op->customdata = vpd = MEM_callocN(sizeof(ViewZoomData), "ImageViewZoomData");
/* Grab will be set when running from gizmo. */
vpd->own_cursor = (win->grabcursor == 0);
if (vpd->own_cursor) {
WM_cursor_modal_set(win, WM_CURSOR_NSEW_SCROLL);
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
vpd->origx = event->x;
vpd->origy = event->y;
vpd->zoom = sima->zoom;
vpd->launch_event = WM_userdef_event_type_from_keymap_type(event->type);
UI_view2d_region_to_view(
&region->v2d, event->mval[0], event->mval[1], &vpd->location[0], &vpd->location[1]);
if (U.viewzoom == USER_ZOOM_CONT) {
/* needs a timer to continue redrawing */
vpd->timer = WM_event_add_timer(CTX_wm_manager(C), CTX_wm_window(C), TIMER, 0.01f);
vpd->timer_lastdraw = PIL_check_seconds_timer();
}
vpd->sima = sima;
vpd->region = region;
WM_event_add_modal_handler(C, op);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static void image_view_zoom_exit(bContext *C, wmOperator *op, bool cancel)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
ViewZoomData *vpd = op->customdata;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (cancel) {
sima->zoom = vpd->zoom;
ED_region_tag_redraw(CTX_wm_region(C));
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (vpd->timer) {
WM_event_remove_timer(CTX_wm_manager(C), vpd->timer->win, vpd->timer);
}
if (vpd->own_cursor) {
WM_cursor_modal_restore(CTX_wm_window(C));
}
MEM_freeN(op->customdata);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_view_zoom_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
sima_zoom_set_factor(sima, region, RNA_float_get(op->ptr, "factor"), NULL, false);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
ED_region_tag_redraw(region);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
enum {
VIEW_PASS = 0,
VIEW_APPLY,
VIEW_CONFIRM,
};
static int image_view_zoom_invoke(bContext *C, wmOperator *op, const wmEvent *event)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
if (event->type == MOUSEZOOM || event->type == MOUSEPAN) {
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
float delta, factor, location[2];
UI_view2d_region_to_view(
&region->v2d, event->mval[0], event->mval[1], &location[0], &location[1]);
delta = event->prevx - event->x + event->prevy - event->y;
if (U.uiflag & USER_ZOOM_INVERT) {
delta *= -1;
}
factor = 1.0f + delta / 300.0f;
RNA_float_set(op->ptr, "factor", factor);
const bool use_cursor_init = RNA_boolean_get(op->ptr, "use_cursor_init");
sima_zoom_set(sima,
region,
sima->zoom * factor,
location,
(use_cursor_init && (U.uiflag & USER_ZOOM_TO_MOUSEPOS)));
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
}
image_view_zoom_init(C, op, event);
return OPERATOR_RUNNING_MODAL;
}
static void image_zoom_apply(ViewZoomData *vpd,
wmOperator *op,
const int x,
const int y,
const short viewzoom,
const short zoom_invert,
const bool zoom_to_pos)
{
float factor;
float delta;
if (viewzoom != USER_ZOOM_SCALE) {
if (U.uiflag & USER_ZOOM_HORIZ) {
delta = (float)(x - vpd->origx);
}
else {
delta = (float)(y - vpd->origy);
}
}
else {
delta = x - vpd->origx + y - vpd->origy;
}
delta /= U.pixelsize;
if (zoom_invert) {
delta = -delta;
}
if (viewzoom == USER_ZOOM_CONT) {
double time = PIL_check_seconds_timer();
float time_step = (float)(time - vpd->timer_lastdraw);
float zfac;
/* oldstyle zoom */
zfac = 1.0f + ((delta / 20.0f) * time_step);
vpd->timer_lastdraw = time;
/* this is the final zoom, but instead make it into a factor */
factor = (vpd->sima->zoom * zfac) / vpd->zoom;
}
else {
factor = 1.0f + delta / 300.0f;
}
RNA_float_set(op->ptr, "factor", factor);
sima_zoom_set(vpd->sima, vpd->region, vpd->zoom * factor, vpd->location, zoom_to_pos);
ED_region_tag_redraw(vpd->region);
}
static int image_view_zoom_modal(bContext *C, wmOperator *op, const wmEvent *event)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
ViewZoomData *vpd = op->customdata;
short event_code = VIEW_PASS;
/* execute the events */
if (event->type == TIMER && event->customdata == vpd->timer) {
/* continuous zoom */
event_code = VIEW_APPLY;
}
else if (event->type == MOUSEMOVE) {
event_code = VIEW_APPLY;
}
else if (event->type == vpd->launch_event && event->val == KM_RELEASE) {
event_code = VIEW_CONFIRM;
}
if (event_code == VIEW_APPLY) {
const bool use_cursor_init = RNA_boolean_get(op->ptr, "use_cursor_init");
image_zoom_apply(vpd,
op,
event->x,
event->y,
U.viewzoom,
(U.uiflag & USER_ZOOM_INVERT) != 0,
(use_cursor_init && (U.uiflag & USER_ZOOM_TO_MOUSEPOS)));
}
else if (event_code == VIEW_CONFIRM) {
image_view_zoom_exit(C, op, false);
return OPERATOR_FINISHED;
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
return OPERATOR_RUNNING_MODAL;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static void image_view_zoom_cancel(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
image_view_zoom_exit(C, op, true);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
void IMAGE_OT_view_zoom(wmOperatorType *ot)
{
PropertyRNA *prop;
/* identifiers */
ot->name = "Zoom View";
ot->idname = "IMAGE_OT_view_zoom";
ot->description = "Zoom in/out the image";
/* api callbacks */
ot->exec = image_view_zoom_exec;
ot->invoke = image_view_zoom_invoke;
ot->modal = image_view_zoom_modal;
ot->cancel = image_view_zoom_cancel;
ot->poll = space_image_main_region_poll;
/* flags */
ot->flag = OPTYPE_BLOCKING | OPTYPE_GRAB_CURSOR_XY | OPTYPE_LOCK_BYPASS;
/* properties */
prop = RNA_def_float(ot->srna,
"factor",
0.0f,
-FLT_MAX,
FLT_MAX,
"Factor",
"Zoom factor, values higher than 1.0 zoom in, lower values zoom out",
-FLT_MAX,
FLT_MAX);
RNA_def_property_flag(prop, PROP_HIDDEN);
WM_operator_properties_use_cursor_init(ot);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
#ifdef WITH_INPUT_NDOF
/** \} */
/* -------------------------------------------------------------------- */
/** \name NDOF Operator
* \{ */
/* Combined pan/zoom from a 3D mouse device.
* Z zooms, XY pans
* "view" (not "paper") control -- user moves the viewpoint, not the image being viewed
* that explains the negative signs in the code below
*/
static int image_view_ndof_invoke(bContext *C, wmOperator *UNUSED(op), const wmEvent *event)
{
if (event->type != NDOF_MOTION) {
return OPERATOR_CANCELLED;
}
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
float pan_vec[3];
const wmNDOFMotionData *ndof = event->customdata;
const float speed = NDOF_PIXELS_PER_SECOND;
WM_event_ndof_pan_get(ndof, pan_vec, true);
mul_v2_fl(pan_vec, (speed * ndof->dt) / sima->zoom);
pan_vec[2] *= -ndof->dt;
sima_zoom_set_factor(sima, region, 1.0f + pan_vec[2], NULL, false);
sima->xof += pan_vec[0];
sima->yof += pan_vec[1];
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
}
void IMAGE_OT_view_ndof(wmOperatorType *ot)
{
/* identifiers */
ot->name = "NDOF Pan/Zoom";
ot->idname = "IMAGE_OT_view_ndof";
ot->description = "Use a 3D mouse device to pan/zoom the view";
/* api callbacks */
ot->invoke = image_view_ndof_invoke;
ot->poll = space_image_main_region_poll;
Option to lock the interface while rendering Added function called WM_set_locked_interface which does two things: - Prevents event queue from being handled, so no operators (see below) or values are even possible to run or change. This prevents any kind of "destructive" action performed from user while rendering. - Locks interface refresh for regions which does have lock set to truth in their template. Currently it's just a 3D viewport, but in the future more regions could be considered unsafe, or we could want to lock different parts of interface when doing different jobs. This is needed because 3D viewport could be using or changing the same data as renderer currently uses, leading to threading conflict. Notifiers are still allowed to handle, so render progress is seen on the screen, but would need to doublecheck on this, in terms some notifiers could be changing the data. For now interface locking happens for render job only in case "Lock Interface" checkbox is enabled. Other tools like backing would also benefit of this option. It is possible to mark operator as safe to be used in locked interface mode by adding OPTYPE_ALLOW_LOCKED bit to operator template flags. This bit is completely handled by wm_evem_system, not with operator run routines, so it's still possible to run operators from drivers and handlers. Currently allowed image editor navigation and zooming. Reviewers: brecht, campbellbarton Reviewed By: campbellbarton Differential Revision: https://developer.blender.org/D142
2014-01-29 16:07:14 +06:00
/* flags */
ot->flag = OPTYPE_LOCK_BYPASS;
}
/** \} */
#endif /* WITH_INPUT_NDOF */
/* -------------------------------------------------------------------- */
/** \name View All Operator
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
/* Updates the fields of the View2D member of the SpaceImage struct.
* Default behavior is to reset the position of the image and set the zoom to 1
* If the image will not fit within the window rectangle, the zoom is adjusted */
static int image_view_all_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima;
ARegion *region;
float aspx, aspy, zoomx, zoomy, w, h;
int width, height;
const bool fit_view = RNA_boolean_get(op->ptr, "fit_view");
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
/* retrieve state */
sima = CTX_wm_space_image(C);
region = CTX_wm_region(C);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
ED_space_image_get_size(sima, &width, &height);
ED_space_image_get_aspect(sima, &aspx, &aspy);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
w = width * aspx;
h = height * aspy;
float xof = 0.0f, yof = 0.0f;
if ((sima->image == NULL) || (sima->image->source == IMA_SRC_TILED)) {
/* Extend the shown area to cover all UDIM tiles. */
int x_tiles, y_tiles;
if (sima->image == NULL) {
x_tiles = sima->tile_grid_shape[0];
y_tiles = sima->tile_grid_shape[1];
}
else {
x_tiles = y_tiles = 1;
LISTBASE_FOREACH (ImageTile *, tile, &sima->image->tiles) {
int tile_x = (tile->tile_number - 1001) % 10;
int tile_y = (tile->tile_number - 1001) / 10;
x_tiles = max_ii(x_tiles, tile_x + 1);
y_tiles = max_ii(y_tiles, tile_y + 1);
}
}
xof = 0.5f * (x_tiles - 1.0f) * w;
yof = 0.5f * (y_tiles - 1.0f) * h;
w *= x_tiles;
h *= y_tiles;
}
/* check if the image will fit in the image with (zoom == 1) */
width = BLI_rcti_size_x(&region->winrct) + 1;
height = BLI_rcti_size_y(&region->winrct) + 1;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (fit_view) {
const int margin = 5; /* margin from border */
zoomx = (float)width / (w + 2 * margin);
zoomy = (float)height / (h + 2 * margin);
sima_zoom_set(sima, region, min_ff(zoomx, zoomy), NULL, false);
}
else {
if ((w >= width || h >= height) && (width > 0 && height > 0)) {
zoomx = (float)width / w;
zoomy = (float)height / h;
/* find the zoom value that will fit the image in the image space */
sima_zoom_set(sima, region, 1.0f / power_of_2(1.0f / min_ff(zoomx, zoomy)), NULL, false);
}
else {
sima_zoom_set(sima, region, 1.0f, NULL, false);
}
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
sima->xof = xof;
sima->yof = yof;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
void IMAGE_OT_view_all(wmOperatorType *ot)
{
PropertyRNA *prop;
/* identifiers */
ot->name = "Frame All";
ot->idname = "IMAGE_OT_view_all";
ot->description = "View the entire image";
/* api callbacks */
ot->exec = image_view_all_exec;
ot->poll = space_image_main_region_poll;
/* flags */
ot->flag = OPTYPE_LOCK_BYPASS;
/* properties */
prop = RNA_def_boolean(ot->srna, "fit_view", 0, "Fit View", "Fit frame to the viewport");
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Center View To Cursor Operator
* \{ */
static int view_center_cursor_exec(bContext *C, wmOperator *UNUSED(op))
{
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
ED_image_view_center_to_point(sima, sima->cursor[0], sima->cursor[1]);
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
}
void IMAGE_OT_view_center_cursor(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Center View to Cursor";
ot->description = "Center the view so that the cursor is in the middle of the view";
ot->idname = "IMAGE_OT_view_center_cursor";
/* api callbacks */
ot->exec = view_center_cursor_exec;
ot->poll = ED_space_image_cursor_poll;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Frame Selected Operator
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
static int image_view_selected_exec(bContext *C, wmOperator *UNUSED(op))
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima;
ARegion *region;
Scene *scene;
ViewLayer *view_layer;
Object *obedit;
/* retrieve state */
sima = CTX_wm_space_image(C);
region = CTX_wm_region(C);
scene = CTX_data_scene(C);
view_layer = CTX_data_view_layer(C);
obedit = CTX_data_edit_object(C);
/* get bounds */
float min[2], max[2];
if (ED_space_image_show_uvedit(sima, obedit)) {
uint objects_len = 0;
Object **objects = BKE_view_layer_array_from_objects_in_edit_mode_unique_data_with_uvs(
view_layer, ((View3D *)NULL), &objects_len);
bool success = ED_uvedit_minmax_multi(scene, objects, objects_len, min, max);
MEM_freeN(objects);
if (!success) {
return OPERATOR_CANCELLED;
}
}
else if (ED_space_image_check_show_maskedit(sima, view_layer)) {
if (!ED_mask_selected_minmax(C, min, max)) {
return OPERATOR_CANCELLED;
}
}
rctf bounds = {
.xmin = min[0],
.ymin = min[1],
.xmax = max[0],
.ymax = max[1],
};
/* add some margin */
BLI_rctf_scale(&bounds, 1.4f);
sima_zoom_set_from_bounds(sima, region, &bounds);
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
2018-07-02 11:47:00 +02:00
static bool image_view_selected_poll(bContext *C)
{
return (space_image_main_region_poll(C) && (ED_operator_uvedit(C) || ED_operator_mask(C)));
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
void IMAGE_OT_view_selected(wmOperatorType *ot)
{
/* identifiers */
ot->name = "View Center";
ot->idname = "IMAGE_OT_view_selected";
ot->description = "View all selected UVs";
/* api callbacks */
ot->exec = image_view_selected_exec;
ot->poll = image_view_selected_poll;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name View Zoom In/Out Operator
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
static int image_view_zoom_in_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
float location[2];
RNA_float_get_array(op->ptr, "location", location);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
sima_zoom_set_factor(
sima, region, powf(2.0f, 1.0f / 3.0f), location, U.uiflag & USER_ZOOM_TO_MOUSEPOS);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_view_zoom_in_invoke(bContext *C, wmOperator *op, const wmEvent *event)
{
ARegion *region = CTX_wm_region(C);
float location[2];
UI_view2d_region_to_view(
&region->v2d, event->mval[0], event->mval[1], &location[0], &location[1]);
RNA_float_set_array(op->ptr, "location", location);
return image_view_zoom_in_exec(C, op);
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
void IMAGE_OT_view_zoom_in(wmOperatorType *ot)
{
PropertyRNA *prop;
/* identifiers */
ot->name = "Zoom In";
ot->idname = "IMAGE_OT_view_zoom_in";
ot->description = "Zoom in the image (centered around 2D cursor)";
/* api callbacks */
ot->invoke = image_view_zoom_in_invoke;
ot->exec = image_view_zoom_in_exec;
ot->poll = space_image_main_region_poll;
/* flags */
ot->flag = OPTYPE_LOCK_BYPASS;
Option to lock the interface while rendering Added function called WM_set_locked_interface which does two things: - Prevents event queue from being handled, so no operators (see below) or values are even possible to run or change. This prevents any kind of "destructive" action performed from user while rendering. - Locks interface refresh for regions which does have lock set to truth in their template. Currently it's just a 3D viewport, but in the future more regions could be considered unsafe, or we could want to lock different parts of interface when doing different jobs. This is needed because 3D viewport could be using or changing the same data as renderer currently uses, leading to threading conflict. Notifiers are still allowed to handle, so render progress is seen on the screen, but would need to doublecheck on this, in terms some notifiers could be changing the data. For now interface locking happens for render job only in case "Lock Interface" checkbox is enabled. Other tools like backing would also benefit of this option. It is possible to mark operator as safe to be used in locked interface mode by adding OPTYPE_ALLOW_LOCKED bit to operator template flags. This bit is completely handled by wm_evem_system, not with operator run routines, so it's still possible to run operators from drivers and handlers. Currently allowed image editor navigation and zooming. Reviewers: brecht, campbellbarton Reviewed By: campbellbarton Differential Revision: https://developer.blender.org/D142
2014-01-29 16:07:14 +06:00
/* properties */
prop = RNA_def_float_vector(ot->srna,
"location",
2,
NULL,
-FLT_MAX,
FLT_MAX,
"Location",
"Cursor location in screen coordinates",
-10.0f,
10.0f);
RNA_def_property_flag(prop, PROP_HIDDEN);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_view_zoom_out_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
float location[2];
RNA_float_get_array(op->ptr, "location", location);
sima_zoom_set_factor(
sima, region, powf(0.5f, 1.0f / 3.0f), location, U.uiflag & USER_ZOOM_TO_MOUSEPOS);
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_view_zoom_out_invoke(bContext *C, wmOperator *op, const wmEvent *event)
{
ARegion *region = CTX_wm_region(C);
float location[2];
UI_view2d_region_to_view(
&region->v2d, event->mval[0], event->mval[1], &location[0], &location[1]);
RNA_float_set_array(op->ptr, "location", location);
return image_view_zoom_out_exec(C, op);
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
void IMAGE_OT_view_zoom_out(wmOperatorType *ot)
{
PropertyRNA *prop;
/* identifiers */
ot->name = "Zoom Out";
ot->idname = "IMAGE_OT_view_zoom_out";
ot->description = "Zoom out the image (centered around 2D cursor)";
/* api callbacks */
ot->invoke = image_view_zoom_out_invoke;
ot->exec = image_view_zoom_out_exec;
ot->poll = space_image_main_region_poll;
/* flags */
ot->flag = OPTYPE_LOCK_BYPASS;
Option to lock the interface while rendering Added function called WM_set_locked_interface which does two things: - Prevents event queue from being handled, so no operators (see below) or values are even possible to run or change. This prevents any kind of "destructive" action performed from user while rendering. - Locks interface refresh for regions which does have lock set to truth in their template. Currently it's just a 3D viewport, but in the future more regions could be considered unsafe, or we could want to lock different parts of interface when doing different jobs. This is needed because 3D viewport could be using or changing the same data as renderer currently uses, leading to threading conflict. Notifiers are still allowed to handle, so render progress is seen on the screen, but would need to doublecheck on this, in terms some notifiers could be changing the data. For now interface locking happens for render job only in case "Lock Interface" checkbox is enabled. Other tools like backing would also benefit of this option. It is possible to mark operator as safe to be used in locked interface mode by adding OPTYPE_ALLOW_LOCKED bit to operator template flags. This bit is completely handled by wm_evem_system, not with operator run routines, so it's still possible to run operators from drivers and handlers. Currently allowed image editor navigation and zooming. Reviewers: brecht, campbellbarton Reviewed By: campbellbarton Differential Revision: https://developer.blender.org/D142
2014-01-29 16:07:14 +06:00
/* properties */
prop = RNA_def_float_vector(ot->srna,
"location",
2,
NULL,
-FLT_MAX,
FLT_MAX,
"Location",
"Cursor location in screen coordinates",
-10.0f,
10.0f);
RNA_def_property_flag(prop, PROP_HIDDEN);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name View Zoom Ratio Operator
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
static int image_view_zoom_ratio_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
sima_zoom_set(sima, region, RNA_float_get(op->ptr, "ratio"), NULL, false);
/* ensure pixel exact locations for draw */
sima->xof = (int)sima->xof;
sima->yof = (int)sima->yof;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
void IMAGE_OT_view_zoom_ratio(wmOperatorType *ot)
{
/* identifiers */
ot->name = "View Zoom Ratio";
ot->idname = "IMAGE_OT_view_zoom_ratio";
ot->description = "Set zoom ratio of the view";
/* api callbacks */
ot->exec = image_view_zoom_ratio_exec;
ot->poll = space_image_main_region_poll;
Threaded object update and EvaluationContext Summary: Made objects update happening from multiple threads. It is a task-based scheduling system which uses current dependency graph for spawning new tasks. This means threading happens on object level, but the system is flexible enough for higher granularity. Technical details: - Uses task scheduler which was recently committed to trunk (that one which Brecht ported from Cycles). - Added two utility functions to dependency graph: * DAG_threaded_update_begin, which is called to initialize threaded objects update. It will also schedule root DAG node to the queue, hence starting evaluation process. Initialization will calculate how much parents are to be evaluation before current DAG node can be scheduled. This value is used by task threads for faster detecting which nodes might be scheduled. * DAG_threaded_update_handle_node_updated which is called from task thread function when node was fully handled. This function decreases num_pending_parents of node children and schedules children with zero valency. As it might have become clear, task thread receives DAG nodes and decides which callback to call for it. Currently only BKE_object_handle_update is called for object nodes. In the future it'll call node->callback() from Ali's new DAG. - This required adding some workarounds to the render pipeline. Mainly to stop using get_object_dm() from modifiers' apply callback. Such a call was only a workaround for dependency graph glitch when rendering scene with, say, boolean modifiers before displaying this scene. Such change moves workaround from one place to another, so overall hackentropy remains the same. - Added paradigm of EvaluaitonContext. Currently it's more like just a more reliable replacement for G.is_rendering which fails in some circumstances. Future idea of this context is to also store all the local data needed for objects evaluation such as local time, Copy-on-Write data and so. There're two types of EvaluationContext: * Context used for viewport updated and owned by Main. In the future this context might be easily moved to Window or Screen to allo per-window/per-screen local time. * Context used by render engines to evaluate objects for render purposes. Render engine is an owner of this context. This context is passed to all object update routines. Reviewers: brecht, campbellbarton Reviewed By: brecht CC: lukastoenne Differential Revision: https://developer.blender.org/D94
2013-12-26 17:24:42 +06:00
/* flags */
ot->flag = OPTYPE_LOCK_BYPASS;
Option to lock the interface while rendering Added function called WM_set_locked_interface which does two things: - Prevents event queue from being handled, so no operators (see below) or values are even possible to run or change. This prevents any kind of "destructive" action performed from user while rendering. - Locks interface refresh for regions which does have lock set to truth in their template. Currently it's just a 3D viewport, but in the future more regions could be considered unsafe, or we could want to lock different parts of interface when doing different jobs. This is needed because 3D viewport could be using or changing the same data as renderer currently uses, leading to threading conflict. Notifiers are still allowed to handle, so render progress is seen on the screen, but would need to doublecheck on this, in terms some notifiers could be changing the data. For now interface locking happens for render job only in case "Lock Interface" checkbox is enabled. Other tools like backing would also benefit of this option. It is possible to mark operator as safe to be used in locked interface mode by adding OPTYPE_ALLOW_LOCKED bit to operator template flags. This bit is completely handled by wm_evem_system, not with operator run routines, so it's still possible to run operators from drivers and handlers. Currently allowed image editor navigation and zooming. Reviewers: brecht, campbellbarton Reviewed By: campbellbarton Differential Revision: https://developer.blender.org/D142
2014-01-29 16:07:14 +06:00
/* properties */
RNA_def_float(ot->srna,
"ratio",
0.0f,
-FLT_MAX,
FLT_MAX,
"Ratio",
"Zoom ratio, 1.0 is 1:1, higher is zoomed in, lower is zoomed out",
-FLT_MAX,
FLT_MAX);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name View Border-Zoom Operator
* \{ */
static int image_view_zoom_border_exec(bContext *C, wmOperator *op)
{
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
rctf bounds;
const bool zoom_in = !RNA_boolean_get(op->ptr, "zoom_out");
WM_operator_properties_border_to_rctf(op, &bounds);
UI_view2d_region_to_view_rctf(&region->v2d, &bounds, &bounds);
const struct {
float xof;
float yof;
float zoom;
} sima_view_prev = {
.xof = sima->xof,
.yof = sima->yof,
.zoom = sima->zoom,
};
sima_zoom_set_from_bounds(sima, region, &bounds);
/* zoom out */
if (!zoom_in) {
sima->xof = sima_view_prev.xof + (sima->xof - sima_view_prev.xof);
sima->yof = sima_view_prev.yof + (sima->yof - sima_view_prev.yof);
sima->zoom = sima_view_prev.zoom * (sima_view_prev.zoom / sima->zoom);
}
ED_region_tag_redraw(region);
return OPERATOR_FINISHED;
}
void IMAGE_OT_view_zoom_border(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Zoom to Border";
ot->description = "Zoom in the view to the nearest item contained in the border";
ot->idname = "IMAGE_OT_view_zoom_border";
/* api callbacks */
ot->invoke = WM_gesture_box_invoke;
ot->exec = image_view_zoom_border_exec;
ot->modal = WM_gesture_box_modal;
ot->cancel = WM_gesture_box_cancel;
ot->poll = space_image_main_region_poll;
/* rna */
WM_operator_properties_gesture_box_zoom(ot);
}
/* load/replace/save callbacks */
static void image_filesel(bContext *C, wmOperator *op, const char *path)
{
RNA_string_set(op->ptr, "filepath", path);
WM_event_add_fileselect(C, op);
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Open Image Operator
* \{ */
typedef struct ImageOpenData {
PropertyPointerRNA pprop;
ImageUser *iuser;
ImageFormatData im_format;
} ImageOpenData;
static void image_open_init(bContext *C, wmOperator *op)
{
ImageOpenData *iod;
op->customdata = iod = MEM_callocN(sizeof(ImageOpenData), __func__);
iod->iuser = CTX_data_pointer_get_type(C, "image_user", &RNA_ImageUser).data;
UI_context_active_but_prop_get_templateID(C, &iod->pprop.ptr, &iod->pprop.prop);
}
static void image_open_cancel(bContext *UNUSED(C), wmOperator *op)
{
MEM_freeN(op->customdata);
op->customdata = NULL;
}
static Image *image_open_single(Main *bmain,
wmOperator *op,
ImageFrameRange *range,
const char *relbase,
bool is_relative_path,
bool use_multiview)
{
bool exists = false;
Image *ima = NULL;
errno = 0;
ima = BKE_image_load_exists_ex(bmain, range->filepath, &exists);
if (!ima) {
if (op->customdata) {
MEM_freeN(op->customdata);
}
BKE_reportf(op->reports,
RPT_ERROR,
"Cannot read '%s': %s",
range->filepath,
errno ? strerror(errno) : TIP_("unsupported image format"));
return NULL;
}
if (!exists) {
/* only image path after save, never ibuf */
if (is_relative_path) {
BLI_path_rel(ima->filepath, relbase);
}
/* handle multiview images */
if (use_multiview) {
ImageOpenData *iod = op->customdata;
ImageFormatData *imf = &iod->im_format;
ima->flag |= IMA_USE_VIEWS;
ima->views_format = imf->views_format;
*ima->stereo3d_format = imf->stereo3d_format;
}
else {
ima->flag &= ~IMA_USE_VIEWS;
BKE_image_free_views(ima);
}
if ((range->length > 1) && (ima->source == IMA_SRC_FILE)) {
if (range->udim_tiles.first && range->offset == 1001) {
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
ima->source = IMA_SRC_TILED;
LISTBASE_FOREACH (LinkData *, node, &range->udim_tiles) {
BKE_image_add_tile(ima, POINTER_AS_INT(node->data), NULL);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}
}
else {
ima->source = IMA_SRC_SEQUENCE;
}
}
}
return ima;
}
static int image_open_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
Main *bmain = CTX_data_main(C);
ScrArea *area = CTX_wm_area(C);
Scene *scene = CTX_data_scene(C);
Object *obedit = CTX_data_edit_object(C);
ImageUser *iuser = NULL;
ImageOpenData *iod = op->customdata;
Image *ima = NULL;
int frame_seq_len = 0;
int frame_ofs = 1;
const bool is_relative_path = RNA_boolean_get(op->ptr, "relative_path");
const bool use_multiview = RNA_boolean_get(op->ptr, "use_multiview");
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
const bool use_udim = RNA_boolean_get(op->ptr, "use_udim_detecting");
if (!op->customdata) {
image_open_init(C, op);
}
ListBase ranges = ED_image_filesel_detect_sequences(bmain, op, use_udim);
LISTBASE_FOREACH (ImageFrameRange *, range, &ranges) {
Image *ima_range = image_open_single(
bmain, op, range, BKE_main_blendfile_path(bmain), is_relative_path, use_multiview);
/* take the first image */
if ((ima == NULL) && ima_range) {
ima = ima_range;
frame_seq_len = range->length;
frame_ofs = range->offset;
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}
BLI_freelistN(&range->udim_tiles);
}
BLI_freelistN(&ranges);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
if (ima == NULL) {
return OPERATOR_CANCELLED;
}
/* hook into UI */
iod = op->customdata;
if (iod->pprop.prop) {
/* when creating new ID blocks, use is already 1, but RNA
* pointer use also increases user, so this compensates it */
id_us_min(&ima->id);
PointerRNA imaptr;
RNA_id_pointer_create(&ima->id, &imaptr);
RNA_property_pointer_set(&iod->pprop.ptr, iod->pprop.prop, imaptr, NULL);
RNA_property_update(C, &iod->pprop.ptr, iod->pprop.prop);
}
if (iod->iuser) {
iuser = iod->iuser;
}
else if (area && area->spacetype == SPACE_IMAGE) {
SpaceImage *sima = area->spacedata.first;
ED_space_image_set(bmain, sima, obedit, ima, false);
iuser = &sima->iuser;
}
else {
Tex *tex = CTX_data_pointer_get_type(C, "texture", &RNA_Texture).data;
if (tex && tex->type == TEX_IMAGE) {
iuser = &tex->iuser;
}
if (iuser == NULL) {
Camera *cam = CTX_data_pointer_get_type(C, "camera", &RNA_Camera).data;
if (cam) {
LISTBASE_FOREACH (CameraBGImage *, bgpic, &cam->bg_images) {
if (bgpic->ima == ima) {
iuser = &bgpic->iuser;
break;
}
}
}
}
}
/* initialize because of new image */
if (iuser) {
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
/* If the sequence was a tiled image, we only have one frame. */
iuser->frames = (ima->source == IMA_SRC_SEQUENCE) ? frame_seq_len : 1;
iuser->sfra = 1;
iuser->framenr = 1;
if (ima->source == IMA_SRC_MOVIE) {
iuser->offset = 0;
}
else {
iuser->offset = frame_ofs - 1;
}
iuser->scene = scene;
BKE_image_init_imageuser(ima, iuser);
}
/* XXX BKE_packedfile_unpack_image frees image buffers */
ED_preview_kill_jobs(CTX_wm_manager(C), bmain);
BKE_image_signal(bmain, ima, iuser, IMA_SIGNAL_RELOAD);
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, ima);
MEM_freeN(op->customdata);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_open_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
SpaceImage *sima = CTX_wm_space_image(C); /* XXX other space types can call */
const char *path = U.textudir;
Image *ima = NULL;
Scene *scene = CTX_data_scene(C);
2015-11-23 15:44:15 +11:00
if (sima) {
ima = sima->image;
}
if (ima == NULL) {
Tex *tex = CTX_data_pointer_get_type(C, "texture", &RNA_Texture).data;
if (tex && tex->type == TEX_IMAGE) {
ima = tex->ima;
}
}
if (ima == NULL) {
PointerRNA ptr;
PropertyRNA *prop;
/* hook into UI */
UI_context_active_but_prop_get_templateID(C, &ptr, &prop);
if (prop) {
PointerRNA oldptr;
Image *oldima;
oldptr = RNA_property_pointer_get(&ptr, prop);
oldima = (Image *)oldptr.owner_id;
/* unlikely to fail but better avoid strange crash */
if (oldima && GS(oldima->id.name) == ID_IM) {
ima = oldima;
}
}
}
if (ima) {
path = ima->filepath;
}
if (RNA_struct_property_is_set(op->ptr, "filepath")) {
return image_open_exec(C, op);
}
image_open_init(C, op);
/* show multiview save options only if scene has multiviews */
PropertyRNA *prop;
prop = RNA_struct_find_property(op->ptr, "show_multiview");
RNA_property_boolean_set(op->ptr, prop, (scene->r.scemode & R_MULTIVIEW) != 0);
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
image_filesel(C, op, path);
return OPERATOR_RUNNING_MODAL;
}
static bool image_open_draw_check_prop(PointerRNA *UNUSED(ptr),
PropertyRNA *prop,
void *UNUSED(user_data))
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
{
const char *prop_id = RNA_property_identifier(prop);
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
return !(STREQ(prop_id, "filepath") || STREQ(prop_id, "directory") ||
STREQ(prop_id, "filename"));
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
}
static void image_open_draw(bContext *UNUSED(C), wmOperator *op)
{
uiLayout *layout = op->layout;
ImageOpenData *iod = op->customdata;
ImageFormatData *imf = &iod->im_format;
PointerRNA imf_ptr, ptr;
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
/* main draw call */
RNA_pointer_create(NULL, op->type->srna, op->properties, &ptr);
uiDefAutoButsRNA(
layout, &ptr, image_open_draw_check_prop, NULL, NULL, UI_BUT_LABEL_ALIGN_NONE, false);
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
/* image template */
RNA_pointer_create(NULL, &RNA_ImageFormatSettings, imf, &imf_ptr);
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
/* multiview template */
if (RNA_boolean_get(op->ptr, "show_multiview")) {
uiTemplateImageFormatViews(layout, &imf_ptr, op->ptr);
}
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
}
/* called by other space types too */
void IMAGE_OT_open(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Open Image";
ot->description = "Open image";
ot->idname = "IMAGE_OT_open";
/* api callbacks */
ot->exec = image_open_exec;
ot->invoke = image_open_invoke;
ot->cancel = image_open_cancel;
ot->ui = image_open_draw;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
/* properties */
WM_operator_properties_filesel(ot,
FILE_TYPE_FOLDER | FILE_TYPE_IMAGE | FILE_TYPE_MOVIE,
FILE_SPECIAL,
FILE_OPENFILE,
WM_FILESEL_FILEPATH | WM_FILESEL_DIRECTORY | WM_FILESEL_FILES |
WM_FILESEL_RELPATH,
FILE_DEFAULTDISPLAY,
FILE_SORT_ALPHA);
RNA_def_boolean(
ot->srna,
"use_sequence_detection",
true,
"Detect Sequences",
"Automatically detect animated sequences in selected images (based on file names)");
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
RNA_def_boolean(ot->srna,
"use_udim_detecting",
true,
"Detect UDIMs",
"Detect selected UDIM files and load all matching tiles");
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Match Movie Length Operator
* \{ */
static int image_match_len_exec(bContext *C, wmOperator *UNUSED(op))
{
Scene *scene = CTX_data_scene(C);
Image *ima = image_from_context(C);
ImageUser *iuser = image_user_from_context(C);
if (!ima || !iuser) {
/* Try to get a Texture, or a SpaceImage from context... */
Tex *tex = CTX_data_pointer_get_type(C, "texture", &RNA_Texture).data;
if (tex && tex->type == TEX_IMAGE) {
ima = tex->ima;
iuser = &tex->iuser;
}
}
if (!ima || !iuser || !BKE_image_has_anim(ima)) {
return OPERATOR_CANCELLED;
}
struct anim *anim = ((ImageAnim *)ima->anims.first)->anim;
if (!anim) {
return OPERATOR_CANCELLED;
}
iuser->frames = IMB_anim_get_duration(anim, IMB_TC_RECORD_RUN);
BKE_image_user_frame_calc(ima, iuser, scene->r.cfra);
return OPERATOR_FINISHED;
}
/* called by other space types too */
void IMAGE_OT_match_movie_length(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Match Movie Length";
ot->description = "Set image's user's length to the one of this video";
ot->idname = "IMAGE_OT_match_movie_length";
/* api callbacks */
ot->exec = image_match_len_exec;
/* flags */
/* Don't think we need undo for that. */
ot->flag = OPTYPE_REGISTER | OPTYPE_INTERNAL /* | OPTYPE_UNDO */;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Replace Image Operator
* \{ */
static int image_replace_exec(bContext *C, wmOperator *op)
{
Main *bmain = CTX_data_main(C);
SpaceImage *sima = CTX_wm_space_image(C);
char str[FILE_MAX];
if (!sima->image) {
return OPERATOR_CANCELLED;
}
RNA_string_get(op->ptr, "filepath", str);
/* we cant do much if the str is longer then FILE_MAX :/ */
BLI_strncpy(sima->image->filepath, str, sizeof(sima->image->filepath));
if (sima->image->source == IMA_SRC_GENERATED) {
sima->image->source = IMA_SRC_FILE;
BKE_image_signal(bmain, sima->image, &sima->iuser, IMA_SIGNAL_SRC_CHANGE);
}
if (BLI_path_extension_check_array(str, imb_ext_movie)) {
sima->image->source = IMA_SRC_MOVIE;
}
else {
sima->image->source = IMA_SRC_FILE;
}
/* XXX BKE_packedfile_unpack_image frees image buffers */
ED_preview_kill_jobs(CTX_wm_manager(C), CTX_data_main(C));
BKE_icon_changed(BKE_icon_id_ensure(&sima->image->id));
BKE_image_signal(bmain, sima->image, &sima->iuser, IMA_SIGNAL_RELOAD);
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, sima->image);
return OPERATOR_FINISHED;
}
static int image_replace_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
{
SpaceImage *sima = CTX_wm_space_image(C);
if (!sima->image) {
return OPERATOR_CANCELLED;
}
if (RNA_struct_property_is_set(op->ptr, "filepath")) {
return image_replace_exec(C, op);
}
if (!RNA_struct_property_is_set(op->ptr, "relative_path")) {
RNA_boolean_set(op->ptr, "relative_path", BLI_path_is_rel(sima->image->filepath));
}
image_filesel(C, op, sima->image->filepath);
return OPERATOR_RUNNING_MODAL;
}
void IMAGE_OT_replace(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Replace Image";
ot->idname = "IMAGE_OT_replace";
ot->description = "Replace current image by another one from disk";
/* api callbacks */
ot->exec = image_replace_exec;
ot->invoke = image_replace_invoke;
ot->poll = image_not_packed_poll;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
/* properties */
WM_operator_properties_filesel(ot,
FILE_TYPE_FOLDER | FILE_TYPE_IMAGE | FILE_TYPE_MOVIE,
FILE_SPECIAL,
FILE_OPENFILE,
WM_FILESEL_FILEPATH | WM_FILESEL_RELPATH,
FILE_DEFAULTDISPLAY,
FILE_SORT_ALPHA);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Save Image As Operator
* \{ */
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
typedef struct ImageSaveData {
ImageUser *iuser;
Image *image;
ImageFormatData im_format;
} ImageSaveData;
static char imtype_best_depth(ImBuf *ibuf, const char imtype)
{
const char depth_ok = BKE_imtype_valid_depths(imtype);
if (ibuf->rect_float) {
if (depth_ok & R_IMF_CHAN_DEPTH_32) {
return R_IMF_CHAN_DEPTH_32;
}
if (depth_ok & R_IMF_CHAN_DEPTH_24) {
return R_IMF_CHAN_DEPTH_24;
}
if (depth_ok & R_IMF_CHAN_DEPTH_16) {
return R_IMF_CHAN_DEPTH_16;
}
if (depth_ok & R_IMF_CHAN_DEPTH_12) {
return R_IMF_CHAN_DEPTH_12;
}
return R_IMF_CHAN_DEPTH_8;
}
if (depth_ok & R_IMF_CHAN_DEPTH_8) {
return R_IMF_CHAN_DEPTH_8;
}
if (depth_ok & R_IMF_CHAN_DEPTH_12) {
return R_IMF_CHAN_DEPTH_12;
}
if (depth_ok & R_IMF_CHAN_DEPTH_16) {
return R_IMF_CHAN_DEPTH_16;
}
if (depth_ok & R_IMF_CHAN_DEPTH_24) {
return R_IMF_CHAN_DEPTH_24;
}
if (depth_ok & R_IMF_CHAN_DEPTH_32) {
return R_IMF_CHAN_DEPTH_32;
}
return R_IMF_CHAN_DEPTH_8; /* fallback, should not get here */
}
static int image_save_options_init(Main *bmain,
ImageSaveOptions *opts,
Image *ima,
ImageUser *iuser,
const bool guess_path,
const bool save_as_render)
{
void *lock;
ImBuf *ibuf = BKE_image_acquire_ibuf(ima, iuser, &lock);
if (ibuf) {
Scene *scene = opts->scene;
bool is_depth_set = false;
if (ELEM(ima->type, IMA_TYPE_R_RESULT, IMA_TYPE_COMPOSITE)) {
/* imtype */
opts->im_format = scene->r.im_format;
is_depth_set = true;
if (!BKE_image_is_multiview(ima)) {
/* In case multiview is disabled,
* render settings would be invalid for render result in this area. */
opts->im_format.stereo3d_format = *ima->stereo3d_format;
opts->im_format.views_format = ima->views_format;
}
}
else {
if (ima->source == IMA_SRC_GENERATED) {
opts->im_format.imtype = R_IMF_IMTYPE_PNG;
opts->im_format.compress = ibuf->foptions.quality;
opts->im_format.planes = ibuf->planes;
}
else {
BKE_imbuf_to_image_format(&opts->im_format, ibuf);
}
/* use the multiview image settings as the default */
opts->im_format.stereo3d_format = *ima->stereo3d_format;
opts->im_format.views_format = ima->views_format;
}
BLI_strncpy(opts->filepath, ibuf->name, sizeof(opts->filepath));
/* sanitize all settings */
/* unlikely but just in case */
if (ELEM(opts->im_format.planes, R_IMF_PLANES_BW, R_IMF_PLANES_RGB, R_IMF_PLANES_RGBA) == 0) {
opts->im_format.planes = R_IMF_PLANES_RGBA;
}
/* depth, account for float buffer and format support */
if (is_depth_set == false) {
opts->im_format.depth = imtype_best_depth(ibuf, opts->im_format.imtype);
}
/* some formats don't use quality so fallback to scenes quality */
if (opts->im_format.quality == 0) {
opts->im_format.quality = scene->r.im_format.quality;
}
/* check for empty path */
if (guess_path && opts->filepath[0] == 0) {
const bool is_prev_save = !STREQ(G.ima, "//");
if (save_as_render) {
if (is_prev_save) {
BLI_strncpy(opts->filepath, G.ima, sizeof(opts->filepath));
}
else {
BLI_strncpy(opts->filepath, "//untitled", sizeof(opts->filepath));
BLI_path_abs(opts->filepath, BKE_main_blendfile_path(bmain));
}
}
else {
BLI_snprintf(opts->filepath, sizeof(opts->filepath), "//%s", ima->id.name + 2);
BLI_path_make_safe(opts->filepath);
BLI_path_abs(opts->filepath, is_prev_save ? G.ima : BKE_main_blendfile_path(bmain));
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
/* append UDIM numbering if not present */
if (ima->source == IMA_SRC_TILED &&
(BLI_path_sequence_decode(ima->filepath, NULL, NULL, NULL) != 1001)) {
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
int len = strlen(opts->filepath);
STR_CONCAT(opts->filepath, len, ".1001");
}
}
/* color management */
BKE_color_managed_display_settings_copy(&opts->im_format.display_settings,
&scene->display_settings);
2020-02-25 18:26:26 +11:00
BKE_color_managed_view_settings_free(&opts->im_format.view_settings);
BKE_color_managed_view_settings_copy(&opts->im_format.view_settings, &scene->view_settings);
}
BKE_image_release_ibuf(ima, ibuf, lock);
return (ibuf != NULL);
}
static void image_save_options_from_op(Main *bmain,
ImageSaveOptions *opts,
wmOperator *op,
ImageFormatData *imf)
{
if (imf) {
BKE_color_managed_view_settings_free(&opts->im_format.view_settings);
opts->im_format = *imf;
}
if (RNA_struct_property_is_set(op->ptr, "filepath")) {
RNA_string_get(op->ptr, "filepath", opts->filepath);
BLI_path_abs(opts->filepath, BKE_main_blendfile_path(bmain));
}
}
static void image_save_options_to_op(ImageSaveOptions *opts, wmOperator *op)
{
if (op->customdata) {
ImageSaveData *isd = op->customdata;
BKE_color_managed_view_settings_free(&isd->im_format.view_settings);
isd->im_format = opts->im_format;
}
RNA_string_set(op->ptr, "filepath", opts->filepath);
}
static bool save_image_op(
Main *bmain, Image *ima, ImageUser *iuser, wmOperator *op, ImageSaveOptions *opts)
{
opts->relative = (RNA_struct_find_property(op->ptr, "relative_path") &&
RNA_boolean_get(op->ptr, "relative_path"));
opts->save_copy = (RNA_struct_find_property(op->ptr, "copy") &&
RNA_boolean_get(op->ptr, "copy"));
opts->save_as_render = (RNA_struct_find_property(op->ptr, "save_as_render") &&
RNA_boolean_get(op->ptr, "save_as_render"));
WM_cursor_wait(1);
bool ok = BKE_image_save(op->reports, bmain, ima, iuser, opts);
WM_cursor_wait(0);
/* Remember file path for next save. */
BLI_strncpy(G.ima, opts->filepath, sizeof(G.ima));
WM_main_add_notifier(NC_IMAGE | NA_EDITED, ima);
return ok;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static void image_save_as_free(wmOperator *op)
{
if (op->customdata) {
ImageSaveData *isd = op->customdata;
BKE_color_managed_view_settings_free(&isd->im_format.view_settings);
Color Management, Stage 2: Switch color pipeline to use OpenColorIO Replace old color pipeline which was supporting linear/sRGB color spaces only with OpenColorIO-based pipeline. This introduces two configurable color spaces: - Input color space for images and movie clips. This space is used to convert images/movies from color space in which file is saved to Blender's linear space (for float images, byte images are not internally converted, only input space is stored for such images and used later). This setting could be found in image/clip data block settings. - Display color space which defines space in which particular display is working. This settings could be found in scene's Color Management panel. When render result is being displayed on the screen, apart from converting image to display space, some additional conversions could happen. This conversions are: - View, which defines tone curve applying before display transformation. These are different ways to view the image on the same display device. For example it could be used to emulate film view on sRGB display. - Exposure affects on image exposure before tone map is applied. - Gamma is post-display gamma correction, could be used to match particular display gamma. - RGB curves are user-defined curves which are applying before display transformation, could be used for different purposes. All this settings by default are only applying on render result and does not affect on other images. If some particular image needs to be affected by this transformation, "View as Render" setting of image data block should be set to truth. Movie clips are always affected by all display transformations. This commit also introduces configurable color space in which sequencer is working. This setting could be found in scene's Color Management panel and it should be used if such stuff as grading needs to be done in color space different from sRGB (i.e. when Film view on sRGB display is use, using VD16 space as sequencer's internal space would make grading working in space which is close to the space using for display). Some technical notes: - Image buffer's float buffer is now always in linear space, even if it was created from 16bit byte images. - Space of byte buffer is stored in image buffer's rect_colorspace property. - Profile of image buffer was removed since it's not longer meaningful. - OpenGL and GLSL is supposed to always work in sRGB space. It is possible to support other spaces, but it's quite large project which isn't so much important. - Legacy Color Management option disabled is emulated by using None display. It could have some regressions, but there's no clear way to avoid them. - If OpenColorIO is disabled on build time, it should make blender behaving in the same way as previous release with color management enabled. More details could be found at this page (more details would be added soon): http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.64/Color_Management -- Thanks to Xavier Thomas, Lukas Toene for initial work on OpenColorIO integration and to Brecht van Lommel for some further development and code/ usecase review!
2012-09-15 10:05:07 +00:00
MEM_freeN(op->customdata);
op->customdata = NULL;
}
}
static int image_save_as_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
Main *bmain = CTX_data_main(C);
Scene *scene = CTX_data_scene(C);
ImageSaveOptions opts;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
Image *image = NULL;
ImageUser *iuser = NULL;
ImageFormatData *imf = NULL;
if (op->customdata) {
ImageSaveData *isd = op->customdata;
image = isd->image;
iuser = isd->iuser;
imf = &isd->im_format;
}
else {
image = image_from_context(C);
iuser = image_user_from_context(C);
}
BKE_image_save_options_init(&opts, bmain, scene);
/* just in case to initialize values,
* these should be set on invoke or by the caller. */
image_save_options_init(bmain, &opts, image, iuser, false, false);
image_save_options_from_op(bmain, &opts, op, imf);
opts.do_newpath = true;
save_image_op(bmain, image, iuser, op, &opts);
if (opts.save_copy == false) {
BKE_image_free_packedfiles(image);
}
image_save_as_free(op);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static bool image_save_as_check(bContext *UNUSED(C), wmOperator *op)
{
ImageSaveData *isd = op->customdata;
return WM_operator_filesel_ensure_ext_imtype(op, &isd->im_format);
}
static int image_save_as_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
Main *bmain = CTX_data_main(C);
Image *ima = image_from_context(C);
ImageUser *iuser = image_user_from_context(C);
Scene *scene = CTX_data_scene(C);
ImageSaveOptions opts;
PropertyRNA *prop;
const bool save_as_render = (ima->source == IMA_SRC_VIEWER);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (RNA_struct_property_is_set(op->ptr, "filepath")) {
return image_save_as_exec(C, op);
}
BKE_image_save_options_init(&opts, bmain, scene);
if (image_save_options_init(bmain, &opts, ima, iuser, true, save_as_render) == 0) {
return OPERATOR_CANCELLED;
}
image_save_options_to_op(&opts, op);
Render & Compositing Thread Fixes * Rendering twice or more could crash layer/pass buttons. * Compositing would crash while drawing the image. * Rendering animations could also crash drawing the image. * Compositing could crash * Starting to rendering while preview render / compo was still running could crash. * Exiting while rendering an animation would not abort the renderer properly, making Blender seemingly freeze. * Fixes theoretically possible issue with setting malloc lock with nested threads. * Drawing previews inside nodes could crash when those nodes were being rendered at the same time. There's more crashes, manipulating the scene data or undo can still crash, this commit only focuses on making sure the image buffer and render result access is thread safe. Implementation: * Rather than assuming the render result does not get freed during render, which seems to be quite difficult to do given that e.g. the compositor is allowed to change the size of the buffer or output different passes, the render result is now protected with a read/write mutex. * The read/write mutex allows multiple readers (and pixel writers) at the same time, but only allows one writer to manipulate the data structure. * Added BKE_image_acquire_ibuf/BKE_image_release_ibuf to access images being rendered, cases where this is not needed (most code) can still use BKE_image_get_ibuf. * The job manager now allows only one rendering job at the same time, rather than the G.rendering check which was not reliable.
2009-09-30 18:18:32 +00:00
/* enable save_copy by default for render results */
if (ELEM(ima->type, IMA_TYPE_R_RESULT, IMA_TYPE_COMPOSITE) &&
!RNA_struct_property_is_set(op->ptr, "copy")) {
RNA_boolean_set(op->ptr, "copy", true);
}
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
RNA_boolean_set(op->ptr, "save_as_render", save_as_render);
Color Management, Stage 2: Switch color pipeline to use OpenColorIO Replace old color pipeline which was supporting linear/sRGB color spaces only with OpenColorIO-based pipeline. This introduces two configurable color spaces: - Input color space for images and movie clips. This space is used to convert images/movies from color space in which file is saved to Blender's linear space (for float images, byte images are not internally converted, only input space is stored for such images and used later). This setting could be found in image/clip data block settings. - Display color space which defines space in which particular display is working. This settings could be found in scene's Color Management panel. When render result is being displayed on the screen, apart from converting image to display space, some additional conversions could happen. This conversions are: - View, which defines tone curve applying before display transformation. These are different ways to view the image on the same display device. For example it could be used to emulate film view on sRGB display. - Exposure affects on image exposure before tone map is applied. - Gamma is post-display gamma correction, could be used to match particular display gamma. - RGB curves are user-defined curves which are applying before display transformation, could be used for different purposes. All this settings by default are only applying on render result and does not affect on other images. If some particular image needs to be affected by this transformation, "View as Render" setting of image data block should be set to truth. Movie clips are always affected by all display transformations. This commit also introduces configurable color space in which sequencer is working. This setting could be found in scene's Color Management panel and it should be used if such stuff as grading needs to be done in color space different from sRGB (i.e. when Film view on sRGB display is use, using VD16 space as sequencer's internal space would make grading working in space which is close to the space using for display). Some technical notes: - Image buffer's float buffer is now always in linear space, even if it was created from 16bit byte images. - Space of byte buffer is stored in image buffer's rect_colorspace property. - Profile of image buffer was removed since it's not longer meaningful. - OpenGL and GLSL is supposed to always work in sRGB space. It is possible to support other spaces, but it's quite large project which isn't so much important. - Legacy Color Management option disabled is emulated by using None display. It could have some regressions, but there's no clear way to avoid them. - If OpenColorIO is disabled on build time, it should make blender behaving in the same way as previous release with color management enabled. More details could be found at this page (more details would be added soon): http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.64/Color_Management -- Thanks to Xavier Thomas, Lukas Toene for initial work on OpenColorIO integration and to Brecht van Lommel for some further development and code/ usecase review!
2012-09-15 10:05:07 +00:00
ImageSaveData *isd = MEM_callocN(sizeof(*isd), __func__);
isd->image = ima;
isd->iuser = iuser;
memcpy(&isd->im_format, &opts.im_format, sizeof(opts.im_format));
op->customdata = isd;
/* show multiview save options only if image has multiviews */
prop = RNA_struct_find_property(op->ptr, "show_multiview");
RNA_property_boolean_set(op->ptr, prop, BKE_image_is_multiview(ima));
prop = RNA_struct_find_property(op->ptr, "use_multiview");
RNA_property_boolean_set(op->ptr, prop, BKE_image_is_multiview(ima));
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
image_filesel(C, op, opts.filepath);
Render & Compositing Thread Fixes * Rendering twice or more could crash layer/pass buttons. * Compositing would crash while drawing the image. * Rendering animations could also crash drawing the image. * Compositing could crash * Starting to rendering while preview render / compo was still running could crash. * Exiting while rendering an animation would not abort the renderer properly, making Blender seemingly freeze. * Fixes theoretically possible issue with setting malloc lock with nested threads. * Drawing previews inside nodes could crash when those nodes were being rendered at the same time. There's more crashes, manipulating the scene data or undo can still crash, this commit only focuses on making sure the image buffer and render result access is thread safe. Implementation: * Rather than assuming the render result does not get freed during render, which seems to be quite difficult to do given that e.g. the compositor is allowed to change the size of the buffer or output different passes, the render result is now protected with a read/write mutex. * The read/write mutex allows multiple readers (and pixel writers) at the same time, but only allows one writer to manipulate the data structure. * Added BKE_image_acquire_ibuf/BKE_image_release_ibuf to access images being rendered, cases where this is not needed (most code) can still use BKE_image_get_ibuf. * The job manager now allows only one rendering job at the same time, rather than the G.rendering check which was not reliable.
2009-09-30 18:18:32 +00:00
return OPERATOR_RUNNING_MODAL;
}
static void image_save_as_cancel(bContext *UNUSED(C), wmOperator *op)
{
image_save_as_free(op);
}
static bool image_save_as_draw_check_prop(PointerRNA *ptr,
PropertyRNA *prop,
void *UNUSED(user_data))
{
const char *prop_id = RNA_property_identifier(prop);
return !(STREQ(prop_id, "filepath") || STREQ(prop_id, "directory") ||
STREQ(prop_id, "filename") ||
/* when saving a copy, relative path has no effect */
((STREQ(prop_id, "relative_path")) && RNA_boolean_get(ptr, "copy")));
}
static void image_save_as_draw(bContext *UNUSED(C), wmOperator *op)
{
uiLayout *layout = op->layout;
ImageSaveData *isd = op->customdata;
PointerRNA imf_ptr, ptr;
const bool is_multiview = RNA_boolean_get(op->ptr, "show_multiview");
/* image template */
RNA_pointer_create(NULL, &RNA_ImageFormatSettings, &isd->im_format, &imf_ptr);
uiTemplateImageSettings(layout, &imf_ptr, false);
/* main draw call */
RNA_pointer_create(NULL, op->type->srna, op->properties, &ptr);
uiDefAutoButsRNA(
layout, &ptr, image_save_as_draw_check_prop, NULL, NULL, UI_BUT_LABEL_ALIGN_NONE, false);
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
/* multiview template */
if (is_multiview) {
uiTemplateImageFormatViews(layout, &imf_ptr, op->ptr);
}
}
2018-07-02 11:47:00 +02:00
static bool image_save_as_poll(bContext *C)
{
2019-10-02 01:16:36 +10:00
if (!image_from_context_has_data_poll(C)) {
return false;
}
if (G.is_rendering) {
/* no need to NULL check here */
Image *ima = image_from_context(C);
if (ima->source == IMA_SRC_VIEWER) {
CTX_wm_operator_poll_msg_set(C, "can't save image while rendering");
return false;
}
}
return true;
}
void IMAGE_OT_save_as(wmOperatorType *ot)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
/* identifiers */
ot->name = "Save As Image";
ot->idname = "IMAGE_OT_save_as";
ot->description = "Save the image with another name and/or settings";
/* api callbacks */
ot->exec = image_save_as_exec;
ot->check = image_save_as_check;
ot->invoke = image_save_as_invoke;
ot->cancel = image_save_as_cancel;
ot->ui = image_save_as_draw;
ot->poll = image_save_as_poll;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
/* properties */
RNA_def_boolean(ot->srna,
"save_as_render",
0,
"Save As Render",
"Apply render part of display transform when saving byte image");
RNA_def_boolean(ot->srna,
"copy",
0,
"Copy",
"Create a new image file without modifying the current image in blender");
WM_operator_properties_filesel(ot,
FILE_TYPE_FOLDER | FILE_TYPE_IMAGE | FILE_TYPE_MOVIE,
FILE_SPECIAL,
FILE_SAVE,
WM_FILESEL_FILEPATH | WM_FILESEL_RELPATH | WM_FILESEL_SHOW_PROPS,
FILE_DEFAULTDISPLAY,
FILE_SORT_ALPHA);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Save Image Operator
* \{ */
static bool image_file_path_saveable(bContext *C, Image *ima, ImageUser *iuser)
{
/* Can always repack images. */
if (BKE_image_has_packedfile(ima)) {
return true;
}
/* Test for valid filepath. */
void *lock;
ImBuf *ibuf = BKE_image_acquire_ibuf(ima, iuser, &lock);
bool ret = false;
if (ibuf) {
Main *bmain = CTX_data_main(C);
char name[FILE_MAX];
BLI_strncpy(name, ibuf->name, FILE_MAX);
BLI_path_abs(name, BKE_main_blendfile_path(bmain));
if (BLI_exists(name) == false) {
CTX_wm_operator_poll_msg_set(C, "image file not found");
}
else if (!BLI_file_is_writable(name)) {
CTX_wm_operator_poll_msg_set(C, "image path can't be written to");
}
else if (!BKE_image_buffer_format_writable(ibuf)) {
CTX_wm_operator_poll_msg_set(C, "image format is read-only");
}
else {
ret = true;
}
}
BKE_image_release_ibuf(ima, ibuf, lock);
return ret;
}
static bool image_save_poll(bContext *C)
{
/* Can't save if there are no pixels. */
2019-10-02 01:16:36 +10:00
if (image_from_context_has_data_poll(C) == false) {
return false;
}
Image *ima = image_from_context(C);
ImageUser *iuser = image_user_from_context(C);
/* Images without a filepath will go to save as. */
if (!BKE_image_has_filepath(ima)) {
return true;
}
/* Check if there is a valid file path and image format we can write. */
return image_file_path_saveable(C, ima, iuser);
}
static int image_save_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
Main *bmain = CTX_data_main(C);
Image *image = image_from_context(C);
ImageUser *iuser = image_user_from_context(C);
Scene *scene = CTX_data_scene(C);
ImageSaveOptions opts;
2020-02-25 18:26:26 +11:00
bool ok = false;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (BKE_image_has_packedfile(image)) {
/* Save packed files to memory. */
BKE_image_memorypack(image);
return OPERATOR_FINISHED;
}
BKE_image_save_options_init(&opts, bmain, scene);
if (image_save_options_init(bmain, &opts, image, iuser, false, false) == 0) {
return OPERATOR_CANCELLED;
}
image_save_options_from_op(bmain, &opts, op, NULL);
Render & Compositing Thread Fixes * Rendering twice or more could crash layer/pass buttons. * Compositing would crash while drawing the image. * Rendering animations could also crash drawing the image. * Compositing could crash * Starting to rendering while preview render / compo was still running could crash. * Exiting while rendering an animation would not abort the renderer properly, making Blender seemingly freeze. * Fixes theoretically possible issue with setting malloc lock with nested threads. * Drawing previews inside nodes could crash when those nodes were being rendered at the same time. There's more crashes, manipulating the scene data or undo can still crash, this commit only focuses on making sure the image buffer and render result access is thread safe. Implementation: * Rather than assuming the render result does not get freed during render, which seems to be quite difficult to do given that e.g. the compositor is allowed to change the size of the buffer or output different passes, the render result is now protected with a read/write mutex. * The read/write mutex allows multiple readers (and pixel writers) at the same time, but only allows one writer to manipulate the data structure. * Added BKE_image_acquire_ibuf/BKE_image_release_ibuf to access images being rendered, cases where this is not needed (most code) can still use BKE_image_get_ibuf. * The job manager now allows only one rendering job at the same time, rather than the G.rendering check which was not reliable.
2009-09-30 18:18:32 +00:00
if (BLI_exists(opts.filepath) && BLI_file_is_writable(opts.filepath)) {
if (save_image_op(bmain, image, iuser, op, &opts)) {
/* report since this can be called from key-shortcuts */
BKE_reportf(op->reports, RPT_INFO, "Saved Image '%s'", opts.filepath);
2020-02-25 18:26:26 +11:00
ok = true;
}
}
else {
BKE_reportf(
op->reports, RPT_ERROR, "Cannot save image, path '%s' is not writable", opts.filepath);
}
2020-02-25 18:26:26 +11:00
BKE_color_managed_view_settings_free(&opts.im_format.view_settings);
if (ok) {
return OPERATOR_FINISHED;
}
return OPERATOR_CANCELLED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
static int image_save_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
{
Image *ima = image_from_context(C);
if (!BKE_image_has_packedfile(ima) && !BKE_image_has_filepath(ima)) {
WM_operator_name_call(C, "IMAGE_OT_save_as", WM_OP_INVOKE_DEFAULT, NULL);
return OPERATOR_CANCELLED;
}
return image_save_exec(C, op);
}
void IMAGE_OT_save(wmOperatorType *ot)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
/* identifiers */
ot->name = "Save Image";
ot->idname = "IMAGE_OT_save";
ot->description = "Save the image with current name and settings";
/* api callbacks */
ot->exec = image_save_exec;
ot->invoke = image_save_invoke;
ot->poll = image_save_poll;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Save Sequence Operator
* \{ */
static int image_save_sequence_exec(bContext *C, wmOperator *op)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
Main *bmain = CTX_data_main(C);
Image *image = image_from_context(C);
ImBuf *ibuf, *first_ibuf = NULL;
int tot = 0;
char di[FILE_MAX];
struct MovieCacheIter *iter;
if (image == NULL) {
return OPERATOR_CANCELLED;
}
if (image->source != IMA_SRC_SEQUENCE) {
BKE_report(op->reports, RPT_ERROR, "Can only save sequence on image sequences");
return OPERATOR_CANCELLED;
}
if (image->type == IMA_TYPE_MULTILAYER) {
BKE_report(op->reports, RPT_ERROR, "Cannot save multilayer sequences");
return OPERATOR_CANCELLED;
}
/* get total dirty buffers and first dirty buffer which is used for menu */
ibuf = NULL;
if (image->cache != NULL) {
iter = IMB_moviecacheIter_new(image->cache);
while (!IMB_moviecacheIter_done(iter)) {
ibuf = IMB_moviecacheIter_getImBuf(iter);
if (ibuf->userflags & IB_BITMAPDIRTY) {
if (first_ibuf == NULL) {
first_ibuf = ibuf;
}
tot++;
}
IMB_moviecacheIter_step(iter);
}
IMB_moviecacheIter_free(iter);
}
if (tot == 0) {
BKE_report(op->reports, RPT_WARNING, "No images have been changed");
return OPERATOR_CANCELLED;
}
/* get a filename for menu */
BLI_split_dir_part(first_ibuf->name, di, sizeof(di));
BKE_reportf(op->reports, RPT_INFO, "%d image(s) will be saved in %s", tot, di);
iter = IMB_moviecacheIter_new(image->cache);
while (!IMB_moviecacheIter_done(iter)) {
ibuf = IMB_moviecacheIter_getImBuf(iter);
if (ibuf->userflags & IB_BITMAPDIRTY) {
char name[FILE_MAX];
BLI_strncpy(name, ibuf->name, sizeof(name));
BLI_path_abs(name, BKE_main_blendfile_path(bmain));
if (0 == IMB_saveiff(ibuf, name, IB_rect | IB_zbuf | IB_zbuffloat)) {
BKE_reportf(op->reports, RPT_ERROR, "Could not write image: %s", strerror(errno));
break;
}
BKE_reportf(op->reports, RPT_INFO, "Saved %s", ibuf->name);
ibuf->userflags &= ~IB_BITMAPDIRTY;
}
IMB_moviecacheIter_step(iter);
}
IMB_moviecacheIter_free(iter);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
void IMAGE_OT_save_sequence(wmOperatorType *ot)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
/* identifiers */
ot->name = "Save Sequence";
ot->idname = "IMAGE_OT_save_sequence";
ot->description = "Save a sequence of images";
/* api callbacks */
ot->exec = image_save_sequence_exec;
2019-10-02 01:16:36 +10:00
ot->poll = image_from_context_has_data_poll;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Save All Operator
* \{ */
static bool image_should_be_saved_when_modified(Image *ima)
{
return !ELEM(ima->type, IMA_TYPE_R_RESULT, IMA_TYPE_COMPOSITE);
}
static bool image_should_be_saved(Image *ima, bool *is_format_writable)
{
if (BKE_image_is_dirty_writable(ima, is_format_writable) &&
ELEM(ima->source, IMA_SRC_FILE, IMA_SRC_GENERATED, IMA_SRC_TILED)) {
return image_should_be_saved_when_modified(ima);
}
return false;
}
static bool image_has_valid_path(Image *ima)
{
return strchr(ima->filepath, '\\') || strchr(ima->filepath, '/');
}
bool ED_image_should_save_modified(const Main *bmain)
{
ReportList reports;
BKE_reports_init(&reports, RPT_STORE);
uint modified_images_count = ED_image_save_all_modified_info(bmain, &reports);
bool should_save = modified_images_count || !BLI_listbase_is_empty(&reports.list);
BKE_reports_clear(&reports);
return should_save;
}
int ED_image_save_all_modified_info(const Main *bmain, ReportList *reports)
{
GSet *unique_paths = BLI_gset_str_new(__func__);
int num_saveable_images = 0;
for (Image *ima = bmain->images.first; ima; ima = ima->id.next) {
bool is_format_writable;
if (image_should_be_saved(ima, &is_format_writable)) {
if (BKE_image_has_packedfile(ima) || (ima->source == IMA_SRC_GENERATED)) {
if (ima->id.lib == NULL) {
num_saveable_images++;
}
else {
BKE_reportf(reports,
RPT_WARNING,
"Packed library image can't be saved: \"%s\" from \"%s\"",
ima->id.name + 2,
ima->id.lib->filepath);
}
}
else if (!is_format_writable) {
BKE_reportf(reports,
RPT_WARNING,
"Image can't be saved, use a different file format: \"%s\"",
ima->id.name + 2);
}
else {
if (image_has_valid_path(ima)) {
num_saveable_images++;
if (BLI_gset_haskey(unique_paths, ima->filepath)) {
BKE_reportf(reports,
RPT_WARNING,
"Multiple images can't be saved to an identical path: \"%s\"",
ima->filepath);
}
else {
BLI_gset_insert(unique_paths, BLI_strdup(ima->filepath));
}
}
else {
BKE_reportf(reports,
RPT_WARNING,
"Image can't be saved, no valid file path: \"%s\"",
ima->filepath);
}
}
}
}
BLI_gset_free(unique_paths, MEM_freeN);
return num_saveable_images;
}
bool ED_image_save_all_modified(const bContext *C, ReportList *reports)
{
Main *bmain = CTX_data_main(C);
ED_image_save_all_modified_info(bmain, reports);
bool ok = true;
for (Image *ima = bmain->images.first; ima; ima = ima->id.next) {
bool is_format_writable;
if (image_should_be_saved(ima, &is_format_writable)) {
if (BKE_image_has_packedfile(ima) || (ima->source == IMA_SRC_GENERATED)) {
BKE_image_memorypack(ima);
}
else if (is_format_writable) {
if (image_has_valid_path(ima)) {
ImageSaveOptions opts;
Scene *scene = CTX_data_scene(C);
BKE_image_save_options_init(&opts, bmain, scene);
if (image_save_options_init(bmain, &opts, ima, NULL, false, false)) {
bool saved_successfully = BKE_image_save(reports, bmain, ima, NULL, &opts);
ok = ok && saved_successfully;
}
}
}
}
}
return ok;
}
static bool image_save_all_modified_poll(bContext *C)
{
int num_files = ED_image_save_all_modified_info(CTX_data_main(C), NULL);
return num_files > 0;
}
static int image_save_all_modified_exec(bContext *C, wmOperator *op)
{
ED_image_save_all_modified(C, op->reports);
return OPERATOR_FINISHED;
}
void IMAGE_OT_save_all_modified(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Save All Modified";
ot->idname = "IMAGE_OT_save_all_modified";
ot->description = "Save all modified images";
/* api callbacks */
ot->exec = image_save_all_modified_exec;
ot->poll = image_save_all_modified_poll;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Reload Image Operator
* \{ */
static int image_reload_exec(bContext *C, wmOperator *UNUSED(op))
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
Main *bmain = CTX_data_main(C);
Image *ima = image_from_context(C);
ImageUser *iuser = image_user_from_context(C);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
if (!ima) {
return OPERATOR_CANCELLED;
}
/* XXX BKE_packedfile_unpack_image frees image buffers */
ED_preview_kill_jobs(CTX_wm_manager(C), CTX_data_main(C));
BKE_image_signal(bmain, ima, iuser, IMA_SIGNAL_RELOAD);
DEG_id_tag_update(&ima->id, 0);
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, ima);
return OPERATOR_FINISHED;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
}
void IMAGE_OT_reload(wmOperatorType *ot)
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
{
/* identifiers */
ot->name = "Reload Image";
ot->idname = "IMAGE_OT_reload";
ot->description = "Reload current image from disk";
/* api callbacks */
ot->exec = image_reload_exec;
2.5: Space Image ported back Organized as follows: uvedit/ uv editing related code uvedit_draw.c: drawing code uvedit_ops.c: operators, just a few done uvedit_unwrap_ops.c: will be operators for unwrapping uvedit_paramatrizer.c: lscm/abf/stretch/pack space_image/ space_image.c: registration and common getter/setters image_draw.c: drawing code, mostly functional image_panels.c: panels, all commented out image_render.c: render callbacks, non functional image_ops.c: operators, only view navigation done image_header.c: header, menus mostly done but missing buttons Notes: * Header menus consist only of Operator and RNA buttons, if they are not implemented they're displayed grayed out. Ideally the full header could work like this, but std_libbuttons looks problematic. * Started using view2d code more than the old code, but for now it still does own view2d management due to some very specific requirements that the image window has. The drawing code however is more clear hopefully, it only uses view2d, and there is no switching between 'p' and 'f' view2d's anymore, it is always 'f'. * In order to make uvedit operators more independent I move some image space settings to scene toolsettings, and the current image and its buffer is in the context. Especially sync selection and select mode belonged there anyway as this cannot work correct with different spaces having different settings anyway. * Image paint is not back yet, did not want to put that together with uvedit because there's really no code sharing.. perhaps vertex paint, image paint and sculpt would be good to have in one module to share brush code, partial redraw, etc better.
2009-01-15 04:38:18 +00:00
/* flags */
ot->flag = OPTYPE_REGISTER; /* no undo, image buffer is not handled by undo */
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name New Image Operator
* \{ */
#define IMA_DEF_NAME N_("Untitled")
enum {
GEN_CONTEXT_NONE = 0,
GEN_CONTEXT_PAINT_CANVAS = 1,
GEN_CONTEXT_PAINT_STENCIL = 2,
};
typedef struct ImageNewData {
PropertyPointerRNA pprop;
} ImageNewData;
static ImageNewData *image_new_init(bContext *C, wmOperator *op)
{
if (op->customdata) {
return op->customdata;
}
ImageNewData *data = MEM_callocN(sizeof(ImageNewData), __func__);
UI_context_active_but_prop_get_templateID(C, &data->pprop.ptr, &data->pprop.prop);
op->customdata = data;
return data;
}
static void image_new_free(wmOperator *op)
{
if (op->customdata) {
MEM_freeN(op->customdata);
op->customdata = NULL;
}
}
static int image_new_exec(bContext *C, wmOperator *op)
{
SpaceImage *sima;
Object *obedit;
Image *ima;
Main *bmain;
PropertyRNA *prop;
char name_buffer[MAX_ID_NAME - 2];
const char *name;
float color[4];
int width, height, floatbuf, gen_type, alpha;
int stereo3d;
/* retrieve state */
sima = CTX_wm_space_image(C);
obedit = CTX_data_edit_object(C);
bmain = CTX_data_main(C);
prop = RNA_struct_find_property(op->ptr, "name");
RNA_property_string_get(op->ptr, prop, name_buffer);
if (!RNA_property_is_set(op->ptr, prop)) {
/* Default value, we can translate! */
name = DATA_(name_buffer);
}
else {
name = name_buffer;
}
width = RNA_int_get(op->ptr, "width");
height = RNA_int_get(op->ptr, "height");
floatbuf = RNA_boolean_get(op->ptr, "float");
gen_type = RNA_enum_get(op->ptr, "generated_type");
RNA_float_get_array(op->ptr, "color", color);
alpha = RNA_boolean_get(op->ptr, "alpha");
stereo3d = RNA_boolean_get(op->ptr, "use_stereo_3d");
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
bool tiled = RNA_boolean_get(op->ptr, "tiled");
if (!alpha) {
color[3] = 1.0f;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
ima = BKE_image_add_generated(bmain,
width,
height,
name,
alpha ? 32 : 24,
floatbuf,
gen_type,
color,
stereo3d,
false,
tiled);
if (!ima) {
image_new_free(op);
return OPERATOR_CANCELLED;
}
/* hook into UI */
ImageNewData *data = image_new_init(C, op);
if (data->pprop.prop) {
/* when creating new ID blocks, use is already 1, but RNA
* pointer use also increases user, so this compensates it */
id_us_min(&ima->id);
PointerRNA imaptr;
RNA_id_pointer_create(&ima->id, &imaptr);
RNA_property_pointer_set(&data->pprop.ptr, data->pprop.prop, imaptr, NULL);
RNA_property_update(C, &data->pprop.ptr, data->pprop.prop);
}
else if (sima) {
ED_space_image_set(bmain, sima, obedit, ima, false);
}
BKE_image_signal(bmain, ima, (sima) ? &sima->iuser : NULL, IMA_SIGNAL_USER_NEW_IMAGE);
WM_event_add_notifier(C, NC_IMAGE | NA_ADDED, ima);
image_new_free(op);
return OPERATOR_FINISHED;
}
static int image_new_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
{
/* Get property in advance, it doesn't work after WM_operator_props_dialog_popup. */
ImageNewData *data;
op->customdata = data = MEM_callocN(sizeof(ImageNewData), __func__);
UI_context_active_but_prop_get_templateID(C, &data->pprop.ptr, &data->pprop.prop);
/* Better for user feedback. */
RNA_string_set(op->ptr, "name", DATA_(IMA_DEF_NAME));
return WM_operator_props_dialog_popup(C, op, 300);
}
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
static void image_new_draw(bContext *UNUSED(C), wmOperator *op)
{
uiLayout *col;
uiLayout *layout = op->layout;
PointerRNA ptr;
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#if 0
Scene *scene = CTX_data_scene(C);
const bool is_multiview = (scene->r.scemode & R_MULTIVIEW) != 0;
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#endif
RNA_pointer_create(NULL, op->type->srna, op->properties, &ptr);
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
/* copy of WM_operator_props_dialog_popup() layout */
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
uiLayoutSetPropSep(layout, true);
uiLayoutSetPropDecorate(layout, false);
col = uiLayoutColumn(layout, false);
uiItemR(col, &ptr, "name", 0, NULL, ICON_NONE);
uiItemR(col, &ptr, "width", 0, NULL, ICON_NONE);
uiItemR(col, &ptr, "height", 0, NULL, ICON_NONE);
uiItemR(col, &ptr, "color", 0, NULL, ICON_NONE);
uiItemR(col, &ptr, "alpha", 0, NULL, ICON_NONE);
uiItemR(col, &ptr, "generated_type", 0, NULL, ICON_NONE);
uiItemR(col, &ptr, "float", 0, NULL, ICON_NONE);
uiItemR(col, &ptr, "tiled", 0, NULL, ICON_NONE);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#if 0
if (is_multiview) {
uiItemL(col[0], "", ICON_NONE);
uiItemR(col[1], &ptr, "use_stereo_3d", 0, NULL, ICON_NONE);
}
Multi-View and Stereo 3D Official Documentation: http://www.blender.org/manual/render/workflows/multiview.html Implemented Features ==================== Builtin Stereo Camera * Convergence Mode * Interocular Distance * Convergence Distance * Pivot Mode Viewport * Cameras * Plane * Volume Compositor * View Switch Node * Image Node Multi-View OpenEXR support Sequencer * Image/Movie Strips 'Use Multiview' UV/Image Editor * Option to see Multi-View images in Stereo-3D or its individual images * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images I/O * Save/Open Multi-View (OpenEXR, Stereo3D, individual views) images Scene Render Views * Ability to have an arbitrary number of views in the scene Missing Bits ============ First rule of Multi-View bug report: If something is not working as it should *when Views is off* this is a severe bug, do mention this in the report. Second rule is, if something works *when Views is off* but doesn't (or crashes) when *Views is on*, this is a important bug. Do mention this in the report. Everything else is likely small todos, and may wait until we are sure none of the above is happening. Apart from that there are those known issues: * Compositor Image Node poorly working for Multi-View OpenEXR (this was working prefectly before the 'Use Multi-View' functionality) * Selecting camera from Multi-View when looking from camera is problematic * Animation Playback (ctrl+F11) doesn't support stereo formats * Wrong filepath when trying to play back animated scene * Viewport Rendering doesn't support Multi-View * Overscan Rendering * Fullscreen display modes need to warn the user * Object copy should be aware of views suffix Acknowledgments =============== * Francesco Siddi for the help with the original feature specs and design * Brecht Van Lommel for the original review of the code and design early on * Blender Foundation for the Development Fund to support the project wrap up Final patch reviewers: * Antony Riakiotakis (psy-fi) * Campbell Barton (ideasman42) * Julian Eisel (Severin) * Sergey Sharybin (nazgul) * Thomas Dinged (dingto) Code contributors of the original branch in github: * Alexey Akishin * Gabriel Caraballo
2015-04-06 10:40:12 -03:00
#endif
}
static void image_new_cancel(bContext *UNUSED(C), wmOperator *op)
{
image_new_free(op);
}
void IMAGE_OT_new(wmOperatorType *ot)
{
PropertyRNA *prop;
static float default_color[4] = {0.0f, 0.0f, 0.0f, 1.0f};
/* identifiers */
ot->name = "New Image";
ot->description = "Create a new image";
ot->idname = "IMAGE_OT_new";
/* api callbacks */
ot->exec = image_new_exec;
ot->invoke = image_new_invoke;
ot->ui = image_new_draw;
ot->cancel = image_new_cancel;
/* flags */
ot->flag = OPTYPE_UNDO;
/* properties */
RNA_def_string(ot->srna, "name", IMA_DEF_NAME, MAX_ID_NAME - 2, "Name", "Image data-block name");
prop = RNA_def_int(ot->srna, "width", 1024, 1, INT_MAX, "Width", "Image width", 1, 16384);
RNA_def_property_subtype(prop, PROP_PIXEL);
prop = RNA_def_int(ot->srna, "height", 1024, 1, INT_MAX, "Height", "Image height", 1, 16384);
RNA_def_property_subtype(prop, PROP_PIXEL);
prop = RNA_def_float_color(
ot->srna, "color", 4, NULL, 0.0f, FLT_MAX, "Color", "Default fill color", 0.0f, 1.0f);
RNA_def_property_subtype(prop, PROP_COLOR_GAMMA);
RNA_def_property_float_array_default(prop, default_color);
RNA_def_boolean(ot->srna, "alpha", 1, "Alpha", "Create an image with an alpha channel");
RNA_def_enum(ot->srna,
"generated_type",
rna_enum_image_generated_type_items,
IMA_GENTYPE_BLANK,
"Generated Type",
"Fill the image with a grid for UV map testing");
RNA_def_boolean(
ot->srna, "float", 0, "32 bit Float", "Create image with 32 bit floating point bit depth");
RNA_def_property_flag(prop, PROP_HIDDEN);
prop = RNA_def_boolean(
ot->srna, "use_stereo_3d", 0, "Stereo 3D", "Create an image with left and right views");
RNA_def_property_flag(prop, PROP_SKIP_SAVE | PROP_HIDDEN);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
prop = RNA_def_boolean(ot->srna, "tiled", 0, "Tiled", "Create a tiled image");
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
}
#undef IMA_DEF_NAME
/** \} */
/* -------------------------------------------------------------------- */
/** \name Invert Operators
* \{ */
static int image_invert_exec(bContext *C, wmOperator *op)
{
Image *ima = image_from_context(C);
ImBuf *ibuf = BKE_image_acquire_ibuf(ima, NULL, NULL);
SpaceImage *sima = CTX_wm_space_image(C);
const bool is_paint = ((sima != NULL) && (sima->mode == SI_MODE_PAINT));
/* flags indicate if this channel should be inverted */
const bool r = RNA_boolean_get(op->ptr, "invert_r");
const bool g = RNA_boolean_get(op->ptr, "invert_g");
const bool b = RNA_boolean_get(op->ptr, "invert_b");
const bool a = RNA_boolean_get(op->ptr, "invert_a");
size_t i;
if (ibuf == NULL) {
/* TODO: this should actually never happen, but does for render-results -> cleanup */
return OPERATOR_CANCELLED;
}
ED_image_undo_push_begin_with_image(op->type->name, ima, ibuf, &sima->iuser);
if (is_paint) {
ED_imapaint_clear_partial_redraw();
}
/* TODO: make this into an IMB_invert_channels(ibuf,r,g,b,a) method!? */
if (ibuf->rect_float) {
float *fp = (float *)ibuf->rect_float;
for (i = ((size_t)ibuf->x) * ibuf->y; i > 0; i--, fp += 4) {
if (r) {
fp[0] = 1.0f - fp[0];
}
if (g) {
fp[1] = 1.0f - fp[1];
}
if (b) {
fp[2] = 1.0f - fp[2];
}
if (a) {
fp[3] = 1.0f - fp[3];
}
}
if (ibuf->rect) {
IMB_rect_from_float(ibuf);
}
}
else if (ibuf->rect) {
char *cp = (char *)ibuf->rect;
for (i = ((size_t)ibuf->x) * ibuf->y; i > 0; i--, cp += 4) {
if (r) {
cp[0] = 255 - cp[0];
}
if (g) {
cp[1] = 255 - cp[1];
}
if (b) {
cp[2] = 255 - cp[2];
}
if (a) {
cp[3] = 255 - cp[3];
}
}
}
else {
BKE_image_release_ibuf(ima, ibuf, NULL);
return OPERATOR_CANCELLED;
}
ibuf->userflags |= IB_DISPLAY_BUFFER_INVALID;
BKE_image_mark_dirty(ima, ibuf);
if (ibuf->mipmap[0]) {
ibuf->userflags |= IB_MIPMAP_INVALID;
}
ED_image_undo_push_end();
/* force GPU reupload, all image is invalid */
GPU_free_image(ima);
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, ima);
BKE_image_release_ibuf(ima, ibuf, NULL);
return OPERATOR_FINISHED;
}
void IMAGE_OT_invert(wmOperatorType *ot)
{
PropertyRNA *prop;
/* identifiers */
ot->name = "Invert Channels";
ot->idname = "IMAGE_OT_invert";
ot->description = "Invert image's channels";
/* api callbacks */
ot->exec = image_invert_exec;
ot->poll = image_from_context_has_data_poll_no_image_user;
/* properties */
prop = RNA_def_boolean(ot->srna, "invert_r", 0, "Red", "Invert Red Channel");
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
prop = RNA_def_boolean(ot->srna, "invert_g", 0, "Green", "Invert Green Channel");
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
prop = RNA_def_boolean(ot->srna, "invert_b", 0, "Blue", "Invert Blue Channel");
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
prop = RNA_def_boolean(ot->srna, "invert_a", 0, "Alpha", "Invert Alpha Channel");
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
/* flags */
ot->flag = OPTYPE_REGISTER;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Scale Operator
* \{ */
static int image_scale_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
{
Image *ima = image_from_context(C);
PropertyRNA *prop = RNA_struct_find_property(op->ptr, "size");
if (!RNA_property_is_set(op->ptr, prop)) {
ImBuf *ibuf = BKE_image_acquire_ibuf(ima, NULL, NULL);
const int size[2] = {ibuf->x, ibuf->y};
RNA_property_int_set_array(op->ptr, prop, size);
BKE_image_release_ibuf(ima, ibuf, NULL);
}
return WM_operator_props_dialog_popup(C, op, 200);
}
static int image_scale_exec(bContext *C, wmOperator *op)
{
Image *ima = image_from_context(C);
ImBuf *ibuf = BKE_image_acquire_ibuf(ima, NULL, NULL);
SpaceImage *sima = CTX_wm_space_image(C);
const bool is_paint = ((sima != NULL) && (sima->mode == SI_MODE_PAINT));
if (ibuf == NULL) {
/* TODO: this should actually never happen, but does for render-results -> cleanup */
return OPERATOR_CANCELLED;
}
if (is_paint) {
ED_imapaint_clear_partial_redraw();
}
PropertyRNA *prop = RNA_struct_find_property(op->ptr, "size");
int size[2];
if (RNA_property_is_set(op->ptr, prop)) {
RNA_property_int_get_array(op->ptr, prop, size);
}
else {
size[0] = ibuf->x;
size[1] = ibuf->y;
RNA_property_int_set_array(op->ptr, prop, size);
}
ED_image_undo_push_begin_with_image(op->type->name, ima, ibuf, &sima->iuser);
ibuf->userflags |= IB_DISPLAY_BUFFER_INVALID;
IMB_scaleImBuf(ibuf, size[0], size[1]);
BKE_image_release_ibuf(ima, ibuf, NULL);
ED_image_undo_push_end();
/* force GPU reupload, all image is invalid */
GPU_free_image(ima);
DEG_id_tag_update(&ima->id, 0);
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, ima);
return OPERATOR_FINISHED;
}
void IMAGE_OT_resize(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Resize Image";
ot->idname = "IMAGE_OT_resize";
ot->description = "Resize the image";
/* api callbacks */
ot->invoke = image_scale_invoke;
ot->exec = image_scale_exec;
ot->poll = image_from_context_has_data_poll_no_image_user;
/* properties */
RNA_def_int_vector(ot->srna, "size", 2, NULL, 1, INT_MAX, "Size", "", 1, SHRT_MAX);
/* flags */
ot->flag = OPTYPE_REGISTER;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Pack Operator
* \{ */
static bool image_pack_test(bContext *C, wmOperator *op)
{
Image *ima = image_from_context(C);
if (!ima) {
return 0;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
if (ELEM(ima->source, IMA_SRC_SEQUENCE, IMA_SRC_MOVIE, IMA_SRC_TILED)) {
BKE_report(
op->reports, RPT_ERROR, "Packing movies, image sequences or tiled images not supported");
return 0;
}
return 1;
}
static int image_pack_exec(bContext *C, wmOperator *op)
{
struct Main *bmain = CTX_data_main(C);
Image *ima = image_from_context(C);
if (!image_pack_test(C, op)) {
return OPERATOR_CANCELLED;
}
if (BKE_image_is_dirty(ima)) {
BKE_image_memorypack(ima);
}
else {
BKE_image_packfiles(op->reports, ima, ID_BLEND_PATH(bmain, &ima->id));
}
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, ima);
return OPERATOR_FINISHED;
}
void IMAGE_OT_pack(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Pack Image";
ot->description = "Pack an image as embedded data into the .blend file";
ot->idname = "IMAGE_OT_pack";
/* api callbacks */
ot->exec = image_pack_exec;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Unpack Operator
* \{ */
static int image_unpack_exec(bContext *C, wmOperator *op)
{
Main *bmain = CTX_data_main(C);
Image *ima = image_from_context(C);
int method = RNA_enum_get(op->ptr, "method");
/* find the supplied image by name */
if (RNA_struct_property_is_set(op->ptr, "id")) {
char imaname[MAX_ID_NAME - 2];
RNA_string_get(op->ptr, "id", imaname);
ima = BLI_findstring(&bmain->images, imaname, offsetof(ID, name) + 2);
if (!ima) {
ima = image_from_context(C);
}
}
if (!ima || !BKE_image_has_packedfile(ima)) {
return OPERATOR_CANCELLED;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
if (ELEM(ima->source, IMA_SRC_SEQUENCE, IMA_SRC_MOVIE, IMA_SRC_TILED)) {
BKE_report(
op->reports, RPT_ERROR, "Unpacking movies, image sequences or tiled images not supported");
return OPERATOR_CANCELLED;
}
if (G.fileflags & G_FILE_AUTOPACK) {
BKE_report(op->reports,
RPT_WARNING,
"AutoPack is enabled, so image will be packed again on file save");
}
/* XXX BKE_packedfile_unpack_image frees image buffers */
ED_preview_kill_jobs(CTX_wm_manager(C), CTX_data_main(C));
BKE_packedfile_unpack_image(CTX_data_main(C), op->reports, ima, method);
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, ima);
return OPERATOR_FINISHED;
}
static int image_unpack_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
{
Image *ima = image_from_context(C);
if (RNA_struct_property_is_set(op->ptr, "id")) {
return image_unpack_exec(C, op);
}
if (!ima || !BKE_image_has_packedfile(ima)) {
return OPERATOR_CANCELLED;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
if (ELEM(ima->source, IMA_SRC_SEQUENCE, IMA_SRC_MOVIE, IMA_SRC_TILED)) {
BKE_report(
op->reports, RPT_ERROR, "Unpacking movies, image sequences or tiled images not supported");
return OPERATOR_CANCELLED;
}
if (G.fileflags & G_FILE_AUTOPACK) {
BKE_report(op->reports,
RPT_WARNING,
"AutoPack is enabled, so image will be packed again on file save");
}
unpack_menu(C,
"IMAGE_OT_unpack",
ima->id.name + 2,
ima->filepath,
"textures",
BKE_image_has_packedfile(ima) ?
((ImagePackedFile *)ima->packedfiles.first)->packedfile :
NULL);
return OPERATOR_FINISHED;
}
void IMAGE_OT_unpack(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Unpack Image";
ot->description = "Save an image packed in the .blend file to disk";
ot->idname = "IMAGE_OT_unpack";
/* api callbacks */
ot->exec = image_unpack_exec;
ot->invoke = image_unpack_invoke;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
/* properties */
RNA_def_enum(
ot->srna, "method", rna_enum_unpack_method_items, PF_USE_LOCAL, "Method", "How to unpack");
/* XXX, weak!, will fail with library, name collisions */
RNA_def_string(
ot->srna, "id", NULL, MAX_ID_NAME - 2, "Image Name", "Image data-block name to unpack");
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Sample Image Operator
* \{ */
/* Returns color in linear space, matching ED_space_node_color_sample(). */
bool ED_space_image_color_sample(SpaceImage *sima, ARegion *region, int mval[2], float r_col[3])
{
if (sima->image == NULL) {
return false;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
float uv[2];
UI_view2d_region_to_view(&region->v2d, mval[0], mval[1], &uv[0], &uv[1]);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
int tile = BKE_image_get_tile_from_pos(sima->image, uv, uv, NULL);
void *lock;
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
ImBuf *ibuf = ED_space_image_acquire_buffer(sima, &lock, tile);
bool ret = false;
if (ibuf == NULL) {
ED_space_image_release_buffer(sima, ibuf, lock);
return false;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
if (uv[0] >= 0.0f && uv[1] >= 0.0f && uv[0] < 1.0f && uv[1] < 1.0f) {
const float *fp;
uchar *cp;
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
int x = (int)(uv[0] * ibuf->x), y = (int)(uv[1] * ibuf->y);
CLAMP(x, 0, ibuf->x - 1);
CLAMP(y, 0, ibuf->y - 1);
if (ibuf->rect_float) {
fp = (ibuf->rect_float + (ibuf->channels) * (y * ibuf->x + x));
copy_v3_v3(r_col, fp);
ret = true;
}
else if (ibuf->rect) {
cp = (uchar *)(ibuf->rect + y * ibuf->x + x);
rgb_uchar_to_float(r_col, cp);
IMB_colormanagement_colorspace_to_scene_linear_v3(r_col, ibuf->rect_colorspace);
ret = true;
}
}
ED_space_image_release_buffer(sima, ibuf, lock);
return ret;
}
void IMAGE_OT_sample(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Sample Color";
ot->idname = "IMAGE_OT_sample";
ot->description = "Use mouse to sample a color in current image";
/* api callbacks */
ot->invoke = ED_imbuf_sample_invoke;
ot->modal = ED_imbuf_sample_modal;
ot->cancel = ED_imbuf_sample_cancel;
ot->poll = ED_imbuf_sample_poll;
/* flags */
ot->flag = OPTYPE_BLOCKING;
PropertyRNA *prop;
prop = RNA_def_int(ot->srna, "size", 1, 1, 128, "Sample Size", "", 1, 64);
RNA_def_property_subtype(prop, PROP_PIXEL);
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Sample Line Operator
* \{ */
static int image_sample_line_exec(bContext *C, wmOperator *op)
{
SpaceImage *sima = CTX_wm_space_image(C);
ARegion *region = CTX_wm_region(C);
Scene *scene = CTX_data_scene(C);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
Image *ima = ED_space_image(sima);
2012-03-28 11:53:18 +00:00
int x_start = RNA_int_get(op->ptr, "xstart");
int y_start = RNA_int_get(op->ptr, "ystart");
int x_end = RNA_int_get(op->ptr, "xend");
int y_end = RNA_int_get(op->ptr, "yend");
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
float uv1[2], uv2[2], ofs[2];
UI_view2d_region_to_view(&region->v2d, x_start, y_start, &uv1[0], &uv1[1]);
UI_view2d_region_to_view(&region->v2d, x_end, y_end, &uv2[0], &uv2[1]);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
/* If the image has tiles, shift the positions accordingly. */
int tile = BKE_image_get_tile_from_pos(ima, uv1, uv1, ofs);
sub_v2_v2(uv2, ofs);
void *lock;
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
ImBuf *ibuf = ED_space_image_acquire_buffer(sima, &lock, tile);
Histogram *hist = &sima->sample_line_hist;
if (ibuf == NULL) {
ED_space_image_release_buffer(sima, ibuf, lock);
return OPERATOR_CANCELLED;
}
/* hmmmm */
if (ibuf->channels < 3) {
ED_space_image_release_buffer(sima, ibuf, lock);
return OPERATOR_CANCELLED;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
copy_v2_v2(hist->co[0], uv1);
copy_v2_v2(hist->co[1], uv2);
/* enable line drawing */
hist->flag |= HISTO_FLAG_SAMPLELINE;
BKE_histogram_update_sample_line(hist, ibuf, &scene->view_settings, &scene->display_settings);
/* reset y zoom */
hist->ymax = 1.0f;
ED_space_image_release_buffer(sima, ibuf, lock);
ED_area_tag_redraw(CTX_wm_area(C));
return OPERATOR_FINISHED;
}
static int image_sample_line_invoke(bContext *C, wmOperator *op, const wmEvent *event)
{
SpaceImage *sima = CTX_wm_space_image(C);
Histogram *hist = &sima->sample_line_hist;
hist->flag &= ~HISTO_FLAG_SAMPLELINE;
if (!ED_space_image_has_buffer(sima)) {
return OPERATOR_CANCELLED;
}
return WM_gesture_straightline_invoke(C, op, event);
}
void IMAGE_OT_sample_line(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Sample Line";
ot->idname = "IMAGE_OT_sample_line";
ot->description = "Sample a line and show it in Scope panels";
/* api callbacks */
ot->invoke = image_sample_line_invoke;
ot->modal = WM_gesture_straightline_modal;
ot->exec = image_sample_line_exec;
ot->poll = space_image_main_region_poll;
ot->cancel = WM_gesture_straightline_cancel;
/* flags */
ot->flag = 0; /* no undo/register since this operates on the space */
WM_operator_properties_gesture_straightline(ot, WM_CURSOR_EDIT);
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Set Curve Point Operator
* \{ */
void IMAGE_OT_curves_point_set(wmOperatorType *ot)
{
static const EnumPropertyItem point_items[] = {
{0, "BLACK_POINT", 0, "Black Point", ""},
{1, "WHITE_POINT", 0, "White Point", ""},
{0, NULL, 0, NULL, NULL},
};
/* identifiers */
ot->name = "Set Curves Point";
ot->idname = "IMAGE_OT_curves_point_set";
ot->description = "Set black point or white point for curves";
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
/* api callbacks */
ot->invoke = ED_imbuf_sample_invoke;
ot->modal = ED_imbuf_sample_modal;
ot->cancel = ED_imbuf_sample_cancel;
ot->poll = space_image_main_area_not_uv_brush_poll;
/* properties */
RNA_def_enum(
ot->srna, "point", point_items, 0, "Point", "Set black point or white point for curves");
PropertyRNA *prop;
prop = RNA_def_int(ot->srna, "size", 1, 1, 128, "Sample Size", "", 1, 64);
RNA_def_property_subtype(prop, PROP_PIXEL);
RNA_def_property_flag(prop, PROP_SKIP_SAVE);
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Cycle Render Slot Operator
* \{ */
2018-07-02 11:47:00 +02:00
static bool image_cycle_render_slot_poll(bContext *C)
{
Image *ima = image_from_context(C);
return (ima && ima->type == IMA_TYPE_R_RESULT);
}
static int image_cycle_render_slot_exec(bContext *C, wmOperator *op)
{
Image *ima = image_from_context(C);
const int direction = RNA_boolean_get(op->ptr, "reverse") ? -1 : 1;
if (!ED_image_slot_cycle(ima, direction)) {
return OPERATOR_CANCELLED;
}
WM_event_add_notifier(C, NC_IMAGE | ND_DRAW, NULL);
/* no undo push for browsing existing */
RenderSlot *slot = BKE_image_get_renderslot(ima, ima->render_slot);
if ((slot && slot->render) || ima->render_slot == ima->last_render_slot) {
return OPERATOR_CANCELLED;
}
return OPERATOR_FINISHED;
}
void IMAGE_OT_cycle_render_slot(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Cycle Render Slot";
ot->idname = "IMAGE_OT_cycle_render_slot";
ot->description = "Cycle through all non-void render slots";
/* api callbacks */
ot->exec = image_cycle_render_slot_exec;
ot->poll = image_cycle_render_slot_poll;
/* flags */
ot->flag = OPTYPE_REGISTER;
RNA_def_boolean(ot->srna, "reverse", 0, "Cycle in Reverse", "");
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Clear Render Slot Operator
* \{ */
static int image_clear_render_slot_exec(bContext *C, wmOperator *UNUSED(op))
{
SpaceImage *sima = CTX_wm_space_image(C);
Image *ima = image_from_context(C);
if (!BKE_image_clear_renderslot(ima, &sima->iuser, ima->render_slot)) {
return OPERATOR_CANCELLED;
}
WM_event_add_notifier(C, NC_IMAGE | ND_DRAW, NULL);
return OPERATOR_FINISHED;
}
void IMAGE_OT_clear_render_slot(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Clear Render Slot";
ot->idname = "IMAGE_OT_clear_render_slot";
ot->description = "Clear the currently selected render slot";
/* api callbacks */
ot->exec = image_clear_render_slot_exec;
ot->poll = image_cycle_render_slot_poll;
/* flags */
ot->flag = OPTYPE_REGISTER;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Add Render Slot Operator
* \{ */
static int image_add_render_slot_exec(bContext *C, wmOperator *UNUSED(op))
{
Image *ima = image_from_context(C);
RenderSlot *slot = BKE_image_add_renderslot(ima, NULL);
ima->render_slot = BLI_findindex(&ima->renderslots, slot);
WM_event_add_notifier(C, NC_IMAGE | ND_DRAW, NULL);
return OPERATOR_FINISHED;
}
void IMAGE_OT_add_render_slot(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Add Render Slot";
ot->idname = "IMAGE_OT_add_render_slot";
ot->description = "Add a new render slot";
/* api callbacks */
ot->exec = image_add_render_slot_exec;
ot->poll = image_cycle_render_slot_poll;
/* flags */
ot->flag = OPTYPE_REGISTER;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Remove Render Slot Operator
* \{ */
static int image_remove_render_slot_exec(bContext *C, wmOperator *UNUSED(op))
{
SpaceImage *sima = CTX_wm_space_image(C);
Image *ima = image_from_context(C);
if (!BKE_image_remove_renderslot(ima, &sima->iuser, ima->render_slot)) {
return OPERATOR_CANCELLED;
}
WM_event_add_notifier(C, NC_IMAGE | ND_DRAW, NULL);
return OPERATOR_FINISHED;
}
void IMAGE_OT_remove_render_slot(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Remove Render Slot";
ot->idname = "IMAGE_OT_remove_render_slot";
ot->description = "Remove the current render slot";
/* api callbacks */
ot->exec = image_remove_render_slot_exec;
ot->poll = image_cycle_render_slot_poll;
/* flags */
ot->flag = OPTYPE_REGISTER;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Change Frame Operator
* \{ */
2018-07-02 11:47:00 +02:00
static bool change_frame_poll(bContext *C)
{
/* prevent changes during render */
if (G.is_rendering) {
return 0;
}
return space_image_main_region_poll(C);
}
static void change_frame_apply(bContext *C, wmOperator *op)
{
Scene *scene = CTX_data_scene(C);
/* set the new frame number */
CFRA = RNA_int_get(op->ptr, "frame");
FRAMENUMBER_MIN_CLAMP(CFRA);
SUBFRA = 0.0f;
/* do updates */
DEG_id_tag_update(&scene->id, ID_RECALC_AUDIO_SEEK);
WM_event_add_notifier(C, NC_SCENE | ND_FRAME, scene);
}
static int change_frame_exec(bContext *C, wmOperator *op)
{
change_frame_apply(C, op);
return OPERATOR_FINISHED;
}
static int frame_from_event(bContext *C, const wmEvent *event)
{
ARegion *region = CTX_wm_region(C);
Scene *scene = CTX_data_scene(C);
int framenr = 0;
if (region->regiontype == RGN_TYPE_WINDOW) {
float sfra = SFRA, efra = EFRA, framelen = region->winx / (efra - sfra + 1);
framenr = sfra + event->mval[0] / framelen;
}
else {
float viewx, viewy;
UI_view2d_region_to_view(&region->v2d, event->mval[0], event->mval[1], &viewx, &viewy);
framenr = round_fl_to_int(viewx);
}
return framenr;
}
static int change_frame_invoke(bContext *C, wmOperator *op, const wmEvent *event)
{
ARegion *region = CTX_wm_region(C);
if (region->regiontype == RGN_TYPE_WINDOW) {
SpaceImage *sima = CTX_wm_space_image(C);
/* Local coordinate visible rect inside region, to accommodate overlapping ui. */
const rcti *rect_visible = ED_region_visible_rect(region);
const int region_bottom = rect_visible->ymin;
if (event->mval[1] > (region_bottom + 16 * UI_DPI_FAC) || !ED_space_image_show_cache(sima)) {
return OPERATOR_PASS_THROUGH;
}
}
RNA_int_set(op->ptr, "frame", frame_from_event(C, event));
change_frame_apply(C, op);
/* add temp handler */
WM_event_add_modal_handler(C, op);
return OPERATOR_RUNNING_MODAL;
}
static int change_frame_modal(bContext *C, wmOperator *op, const wmEvent *event)
{
switch (event->type) {
case EVT_ESCKEY:
return OPERATOR_FINISHED;
case MOUSEMOVE:
RNA_int_set(op->ptr, "frame", frame_from_event(C, event));
change_frame_apply(C, op);
break;
case LEFTMOUSE:
case RIGHTMOUSE:
if (event->val == KM_RELEASE) {
return OPERATOR_FINISHED;
}
break;
}
return OPERATOR_RUNNING_MODAL;
}
void IMAGE_OT_change_frame(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Change Frame";
ot->idname = "IMAGE_OT_change_frame";
ot->description = "Interactively change the current frame number";
/* api callbacks */
ot->exec = change_frame_exec;
ot->invoke = change_frame_invoke;
ot->modal = change_frame_modal;
ot->poll = change_frame_poll;
/* flags */
ot->flag = OPTYPE_BLOCKING | OPTYPE_UNDO;
/* rna */
RNA_def_int(ot->srna, "frame", 0, MINAFRAME, MAXFRAME, "Frame", "", MINAFRAME, MAXFRAME);
}
/* Reload cached render results... */
/* goes over all scenes, reads render layers */
static int image_read_viewlayers_exec(bContext *C, wmOperator *UNUSED(op))
{
Main *bmain = CTX_data_main(C);
Scene *scene = CTX_data_scene(C);
SpaceImage *sima = CTX_wm_space_image(C);
Image *ima;
ima = BKE_image_ensure_viewer(bmain, IMA_TYPE_R_RESULT, "Render Result");
if (sima->image == NULL) {
ED_space_image_set(bmain, sima, NULL, ima, false);
}
RE_ReadRenderResult(scene, scene);
WM_event_add_notifier(C, NC_IMAGE | NA_EDITED, ima);
return OPERATOR_FINISHED;
}
void IMAGE_OT_read_viewlayers(wmOperatorType *ot)
{
ot->name = "Open Cached Render";
ot->idname = "IMAGE_OT_read_viewlayers";
ot->description = "Read all the current scene's view layers from cache, as needed";
ot->poll = space_image_main_region_poll;
ot->exec = image_read_viewlayers_exec;
/* flags */
ot->flag = 0;
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Render Border Operator
* \{ */
static int render_border_exec(bContext *C, wmOperator *op)
{
ARegion *region = CTX_wm_region(C);
Scene *scene = CTX_data_scene(C);
Render *re = RE_GetSceneRender(scene);
RenderData *rd;
rctf border;
if (re == NULL) {
/* Shouldn't happen, but better be safe close to the release. */
return OPERATOR_CANCELLED;
}
rd = RE_engine_get_render_data(re);
if ((rd->mode & (R_BORDER | R_CROP)) == (R_BORDER | R_CROP)) {
BKE_report(op->reports, RPT_INFO, "Can not set border from a cropped render");
return OPERATOR_CANCELLED;
}
/* get rectangle from operator */
WM_operator_properties_border_to_rctf(op, &border);
UI_view2d_region_to_view_rctf(&region->v2d, &border, &border);
/* actually set border */
CLAMP(border.xmin, 0.0f, 1.0f);
CLAMP(border.ymin, 0.0f, 1.0f);
CLAMP(border.xmax, 0.0f, 1.0f);
CLAMP(border.ymax, 0.0f, 1.0f);
scene->r.border = border;
/* drawing a border surrounding the entire camera view switches off border rendering
* or the border covers no pixels */
if ((border.xmin <= 0.0f && border.xmax >= 1.0f && border.ymin <= 0.0f && border.ymax >= 1.0f) ||
(border.xmin == border.xmax || border.ymin == border.ymax)) {
scene->r.mode &= ~R_BORDER;
}
else {
scene->r.mode |= R_BORDER;
}
DEG_id_tag_update(&scene->id, ID_RECALC_COPY_ON_WRITE);
WM_event_add_notifier(C, NC_SCENE | ND_RENDER_OPTIONS, NULL);
return OPERATOR_FINISHED;
}
void IMAGE_OT_render_border(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Render Region";
ot->description = "Set the boundaries of the render region and enable render region";
ot->idname = "IMAGE_OT_render_border";
/* api callbacks */
ot->invoke = WM_gesture_box_invoke;
ot->exec = render_border_exec;
ot->modal = WM_gesture_box_modal;
ot->cancel = WM_gesture_box_cancel;
ot->poll = image_cycle_render_slot_poll;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
/* rna */
WM_operator_properties_border(ot);
}
/** \} */
/* -------------------------------------------------------------------- */
/** \name Clear Render Border Operator
* \{ */
static int clear_render_border_exec(bContext *C, wmOperator *UNUSED(op))
{
Scene *scene = CTX_data_scene(C);
scene->r.mode &= ~R_BORDER;
WM_event_add_notifier(C, NC_SCENE | ND_RENDER_OPTIONS, NULL);
BLI_rctf_init(&scene->r.border, 0.0f, 1.0f, 0.0f, 1.0f);
return OPERATOR_FINISHED;
}
void IMAGE_OT_clear_render_border(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Clear Render Region";
ot->description = "Clear the boundaries of the render region and disable render region";
ot->idname = "IMAGE_OT_clear_render_border";
/* api callbacks */
ot->exec = clear_render_border_exec;
ot->poll = image_cycle_render_slot_poll;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
}
/** \} */
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
2020-03-25 21:24:59 +11:00
/* -------------------------------------------------------------------- */
/** \name Add Tile Operator
* \{ */
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
static bool do_fill_tile(PointerRNA *ptr, Image *ima, ImageTile *tile)
{
float color[4];
RNA_float_get_array(ptr, "color", color);
int gen_type = RNA_enum_get(ptr, "generated_type");
int width = RNA_int_get(ptr, "width");
int height = RNA_int_get(ptr, "height");
bool is_float = RNA_boolean_get(ptr, "float");
int planes = RNA_boolean_get(ptr, "alpha") ? 32 : 24;
return BKE_image_fill_tile(ima, tile, width, height, color, gen_type, planes, is_float);
}
static void draw_fill_tile(PointerRNA *ptr, uiLayout *layout)
{
uiLayout *split, *col[2];
split = uiLayoutSplit(layout, 0.5f, false);
col[0] = uiLayoutColumn(split, false);
col[1] = uiLayoutColumn(split, false);
uiItemL(col[0], IFACE_("Color"), ICON_NONE);
uiItemR(col[1], ptr, "color", 0, "", ICON_NONE);
uiItemL(col[0], IFACE_("Width"), ICON_NONE);
uiItemR(col[1], ptr, "width", 0, "", ICON_NONE);
uiItemL(col[0], IFACE_("Height"), ICON_NONE);
uiItemR(col[1], ptr, "height", 0, "", ICON_NONE);
uiItemL(col[0], "", ICON_NONE);
uiItemR(col[1], ptr, "alpha", 0, NULL, ICON_NONE);
uiItemL(col[0], IFACE_("Generated Type"), ICON_NONE);
uiItemR(col[1], ptr, "generated_type", 0, "", ICON_NONE);
uiItemL(col[0], "", ICON_NONE);
uiItemR(col[1], ptr, "float", 0, NULL, ICON_NONE);
}
static void initialize_fill_tile(PointerRNA *ptr, Image *ima, ImageTile *tile)
{
ImageUser iuser;
BKE_imageuser_default(&iuser);
if (tile != NULL) {
iuser.tile = tile->tile_number;
}
/* Acquire ibuf to get the default values.
* If the specified tile has no ibuf, try acquiring the main tile instead
* (unless the specified tile already was the main tile).*/
ImBuf *ibuf = BKE_image_acquire_ibuf(ima, &iuser, NULL);
if (ibuf == NULL && (tile != NULL) && (tile->tile_number != 1001)) {
ibuf = BKE_image_acquire_ibuf(ima, NULL, NULL);
}
if (ibuf != NULL) {
/* Initialize properties from reference tile. */
RNA_int_set(ptr, "width", ibuf->x);
RNA_int_set(ptr, "height", ibuf->y);
RNA_boolean_set(ptr, "float", ibuf->rect_float != NULL);
RNA_boolean_set(ptr, "alpha", ibuf->planes > 24);
BKE_image_release_ibuf(ima, ibuf, NULL);
}
}
static void def_fill_tile(StructOrFunctionRNA *srna)
{
PropertyRNA *prop;
static float default_color[4] = {0.0f, 0.0f, 0.0f, 1.0f};
prop = RNA_def_float_color(
srna, "color", 4, NULL, 0.0f, FLT_MAX, "Color", "Default fill color", 0.0f, 1.0f);
RNA_def_property_subtype(prop, PROP_COLOR_GAMMA);
RNA_def_property_float_array_default(prop, default_color);
RNA_def_enum(srna,
"generated_type",
rna_enum_image_generated_type_items,
IMA_GENTYPE_BLANK,
"Generated Type",
"Fill the image with a grid for UV map testing");
prop = RNA_def_int(srna, "width", 1024, 1, INT_MAX, "Width", "Image width", 1, 16384);
RNA_def_property_subtype(prop, PROP_PIXEL);
prop = RNA_def_int(srna, "height", 1024, 1, INT_MAX, "Height", "Image height", 1, 16384);
RNA_def_property_subtype(prop, PROP_PIXEL);
/* Only needed when filling the first tile. */
RNA_def_boolean(
srna, "float", 0, "32 bit Float", "Create image with 32 bit floating point bit depth");
RNA_def_boolean(srna, "alpha", 1, "Alpha", "Create an image with an alpha channel");
}
static bool tile_add_poll(bContext *C)
{
Image *ima = CTX_data_edit_image(C);
return (ima != NULL && ima->source == IMA_SRC_TILED && BKE_image_has_ibuf(ima, NULL));
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
static int tile_add_exec(bContext *C, wmOperator *op)
{
Image *ima = CTX_data_edit_image(C);
int start_tile = RNA_int_get(op->ptr, "number");
int end_tile = start_tile + RNA_int_get(op->ptr, "count");
if (start_tile < 1001 || end_tile > IMA_UDIM_MAX) {
BKE_report(op->reports, RPT_ERROR, "Invalid UDIM index range was specified");
return OPERATOR_CANCELLED;
}
bool fill_tile = RNA_boolean_get(op->ptr, "fill");
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
char *label = RNA_string_get_alloc(op->ptr, "label", NULL, 0);
bool created_tile = false;
for (int tile_number = start_tile; tile_number < end_tile; tile_number++) {
ImageTile *tile = BKE_image_add_tile(ima, tile_number, label);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
if (tile != NULL) {
ima->active_tile_index = BLI_findindex(&ima->tiles, tile);
if (fill_tile) {
do_fill_tile(op->ptr, ima, tile);
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
created_tile = true;
}
}
MEM_freeN(label);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
if (!created_tile) {
return OPERATOR_CANCELLED;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
WM_event_add_notifier(C, NC_IMAGE | ND_DRAW, NULL);
return OPERATOR_FINISHED;
}
static int tile_add_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
{
Image *ima = CTX_data_edit_image(C);
/* Find the first gap in tile numbers or the number after the last if
* no gap exists. */
int next_number = 0;
LISTBASE_FOREACH (ImageTile *, tile, &ima->tiles) {
next_number = tile->tile_number + 1;
if (tile->next == NULL || tile->next->tile_number > next_number) {
break;
}
}
ImageTile *tile = BLI_findlink(&ima->tiles, ima->active_tile_index);
initialize_fill_tile(op->ptr, ima, tile);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
RNA_int_set(op->ptr, "number", next_number);
RNA_int_set(op->ptr, "count", 1);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
RNA_string_set(op->ptr, "label", "");
return WM_operator_props_dialog_popup(C, op, 10 * UI_UNIT_X);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}
static void tile_add_draw(bContext *UNUSED(C), wmOperator *op)
{
uiLayout *split, *col[2];
uiLayout *layout = op->layout;
PointerRNA ptr;
RNA_pointer_create(NULL, op->type->srna, op->properties, &ptr);
split = uiLayoutSplit(layout, 0.5f, false);
col[0] = uiLayoutColumn(split, false);
col[1] = uiLayoutColumn(split, false);
uiItemL(col[0], IFACE_("Number"), ICON_NONE);
uiItemR(col[1], &ptr, "number", 0, "", ICON_NONE);
uiItemL(col[0], IFACE_("Count"), ICON_NONE);
uiItemR(col[1], &ptr, "count", 0, "", ICON_NONE);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
uiItemL(col[0], IFACE_("Label"), ICON_NONE);
uiItemR(col[1], &ptr, "label", 0, "", ICON_NONE);
uiItemR(layout, &ptr, "fill", 0, NULL, ICON_NONE);
if (RNA_boolean_get(&ptr, "fill")) {
draw_fill_tile(&ptr, layout);
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}
void IMAGE_OT_tile_add(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Add tile";
ot->description = "Adds a tile to the image";
ot->idname = "IMAGE_OT_tile_add";
/* api callbacks */
ot->poll = tile_add_poll;
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
ot->exec = tile_add_exec;
ot->invoke = tile_add_invoke;
ot->ui = tile_add_draw;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
RNA_def_int(ot->srna,
"number",
1002,
1001,
IMA_UDIM_MAX,
"Number",
"UDIM number of the tile",
1001,
1099);
RNA_def_int(ot->srna, "count", 1, 1, INT_MAX, "Count", "How many tiles to add", 1, 1000);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
RNA_def_string(ot->srna, "label", NULL, 0, "Label", "Optional tile label");
RNA_def_boolean(ot->srna, "fill", true, "Fill", "Fill new tile with a generated image");
def_fill_tile(ot->srna);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}
2020-03-25 21:24:59 +11:00
/** \} */
/* -------------------------------------------------------------------- */
/** \name Remove Tile Operator
* \{ */
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
static bool tile_remove_poll(bContext *C)
{
Image *ima = CTX_data_edit_image(C);
return (ima != NULL && ima->source == IMA_SRC_TILED && ima->active_tile_index != 0);
}
static int tile_remove_exec(bContext *C, wmOperator *UNUSED(op))
{
Image *ima = CTX_data_edit_image(C);
ImageTile *tile = BLI_findlink(&ima->tiles, ima->active_tile_index);
if (!BKE_image_remove_tile(ima, tile)) {
return OPERATOR_CANCELLED;
}
/* Ensure that the active index is valid. */
ima->active_tile_index = min_ii(ima->active_tile_index, BLI_listbase_count(&ima->tiles) - 1);
WM_event_add_notifier(C, NC_IMAGE | ND_DRAW, NULL);
return OPERATOR_FINISHED;
}
void IMAGE_OT_tile_remove(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Remove tile";
ot->description = "Removes a tile from the image";
ot->idname = "IMAGE_OT_tile_remove";
/* api callbacks */
ot->poll = tile_remove_poll;
ot->exec = tile_remove_exec;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
}
2020-03-25 21:24:59 +11:00
/** \} */
/* -------------------------------------------------------------------- */
/** \name Fill Tile Operator
* \{ */
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
static bool tile_fill_poll(bContext *C)
{
Image *ima = CTX_data_edit_image(C);
if (ima != NULL && ima->source == IMA_SRC_TILED) {
/* Filling secondary tiles is only allowed if the primary tile exists. */
return (ima->active_tile_index == 0) || BKE_image_has_ibuf(ima, NULL);
}
return false;
}
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
static int tile_fill_exec(bContext *C, wmOperator *op)
{
Image *ima = CTX_data_edit_image(C);
ImageTile *tile = BLI_findlink(&ima->tiles, ima->active_tile_index);
if (!do_fill_tile(op->ptr, ima, tile)) {
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
return OPERATOR_CANCELLED;
}
WM_event_add_notifier(C, NC_IMAGE | ND_DRAW, NULL);
return OPERATOR_FINISHED;
}
static int tile_fill_invoke(bContext *C, wmOperator *op, const wmEvent *UNUSED(event))
{
initialize_fill_tile(op->ptr, CTX_data_edit_image(C), NULL);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
return WM_operator_props_dialog_popup(C, op, 15 * UI_UNIT_X);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}
static void tile_fill_draw(bContext *UNUSED(C), wmOperator *op)
{
PointerRNA ptr;
RNA_pointer_create(NULL, op->type->srna, op->properties, &ptr);
draw_fill_tile(&ptr, op->layout);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}
void IMAGE_OT_tile_fill(wmOperatorType *ot)
{
/* identifiers */
ot->name = "Fill Tile";
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
ot->description = "Fill the current tile with a generated image";
ot->idname = "IMAGE_OT_tile_fill";
/* api callbacks */
ot->poll = tile_fill_poll;
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
ot->exec = tile_fill_exec;
ot->invoke = tile_fill_invoke;
ot->ui = tile_fill_draw;
/* flags */
ot->flag = OPTYPE_REGISTER | OPTYPE_UNDO;
def_fill_tile(ot->srna);
Add support for tiled images and the UDIM naming scheme This patch contains the work that I did during my week at the Code Quest - adding support for tiled images to Blender. With this patch, images now contain a list of tiles. By default, this just contains one tile, but if the source type is set to Tiled, the user can add additional tiles. When acquiring an ImBuf, the tile to be loaded is specified in the ImageUser. Therefore, code that is not yet aware of tiles will just access the default tile as usual. The filenames of the additional tiles are derived from the original filename according to the UDIM naming scheme - the filename contains an index that is calculated as (1001 + 10*<y coordinate of the tile> + <x coordinate of the tile>), where the x coordinate never goes above 9. Internally, the various tiles are stored in a cache just like sequences. When acquired for the first time, the code will try to load the corresponding file from disk. Alternatively, a new operator can be used to initialize the tile similar to the New Image operator. The following features are supported so far: - Automatic detection and loading of all tiles when opening the first tile (1001) - Saving all tiles - Adding and removing tiles - Filling tiles with generated images - Drawing all tiles in the Image Editor - Viewing a tiled grid even if no image is selected - Rendering tiled images in Eevee - Rendering tiled images in Cycles (in SVM mode) - Automatically skipping loading of unused tiles in Cycles - 2D texture painting (also across tiles) - 3D texture painting (also across tiles, only limitation: individual faces can not cross tile borders) - Assigning custom labels to individual tiles (drawn in the Image Editor instead of the ID) - Different resolutions between tiles There still are some missing features that will be added later (see T72390): - Workbench engine support - Packing/Unpacking support - Baking support - Cycles OSL support - many other Blender features that rely on images Thanks to Brecht for the review and to all who tested the intermediate versions! Differential Revision: https://developer.blender.org/D3509
2019-12-12 16:06:08 +01:00
}