This would previously break because begin/end functions for each tree type still have some checks of the ntree->execdata pointer in them, despite the intended use of execdata instances instead of trees themselves for execution data storage. This is an artifact of the old execution system that required these checks to be made in the functions to avoid multiple execution of top-level trees. Now these functions take an additional argument, so group nodes can prevent them from setting and checking the nodetree->execdata pointers.
Some transform operators (like push/pull, shrink/fatten, to sphere and so)
were creating "value" as single scalar value. This used to confuse
RNA_float_get_array used in initTransform.
Use RNA_float_get_array for array values and RNA_float_get for scalar value
in transform initi function.
This is currently only needed for displaying the up/down buttons of group sockets. All regular group sockets should have this flag to indicate they are added by the user. More complex "group-type" trees may use non-dynamic sockets in the future for sockets that are not supposed to be manipulated.
- It's like sketch mode for lines, but you're specifying line knots
by clicking on position you want to add next knot.
- View can be navigated between knots creation.
- Holding LMB down and sliding mouse will lead to new segment preview
so it can be created more accurate.
Additional change: fixed GP->Bezier conversion. Last point used to
be ignored in this operator.
The reason was that group nodes tried to reconstruct sockets from the template lists, which are empty. Now the verification function checks if there are any sockets in the template lists, which are always empty for group nodes.
- Implement own undo stack for grease pencil, so now there'll be no keymaps conflicts.
- Supported redo's during sketch session.
- Get rid of flag stored in Globals -- use undo stack to check if grease pencil session is active.
This branch adds mostly organizational improvements to the node system by renaming the node folders and files. A couple of internal features have been added too.
Detailed information can be found on the wiki page:
http://wiki.blender.org/index.php/User:Phonybone/Particles2010
OpenCOLLADA is a validating parser, so is pretty strict about document form. The added error handler will print out any errors the parser finds. A pop-up will be shown too, advising the user to check the console for the error log.
in graph editor > channel menu. Problem was these did not inherit operator
execution context correctly.
Fix found by Sergey, also needed to fix logic operators which were not working
when invoked instead of executed.
*Added Smooth/Sharp/Root/etc. mappings to WeightVGEdit modifier, in addition to custom curve one.
*Added Smooth/Sharp/Root/etc. mappings to WeightVGProximity modifier, without the custom curve one!
*Factorized the common mapping code into MOD_weightvg_util.
Such load/make edit structures introduced regression into iterators
via object's geometry (vertices, edges, control points and so) when
adding hooks in the body of this iterator.
Fix for wrong index should be non-destructable for geometry.
This will fix#28506: Unusual behavior in curves.
- was doing NULL checks on freeing memory in cases where the values were already accessed (blender would have crashed anyway), so remove the NULL checks.
- use deform.c api weight functions to replace inline weight lookups in some cases.
- change if checks in weightvg_do_mask() so its more obvious whats going on.
Multires doesn't store displacement for base mesh and reshaping when
multires subdivision level is set to zero is crappy.
Add report that reshape can't work with base level and cancel reshape operator.
Problems was caused by angle=2*pi and steps=2 in screw modifier.
Such configuration produced duplicated geometry to close object
and it was confusing for subsurf cache.
Restrict steps=2 for screw modifier now, so now 3<=steps<=512.
Memory Estimate is actually 31 characters length, str[31] is a null-terminator.
Return length of 31 for memory estimate property. Returning proper length
would lead to slowdown because of 2x iteration through vertices.
Use quite easy and stupid approach like it used for shape keys:
re-make editmesh (editcurve or editlattice) before creating index array
for hook or storing vertex index in parenting object.
Even if hook was added in "current" edit mode, it should be re-mapped on
loading edit data because topology could be changed after it was created.
Such kind of re-loading edit structures is the easiest way for now to
update keyindexes to relevant state.
Also, fixed bug with not re-mapping indices for vertex-parented objects.
Really old error, not sure why it wasn't noticed yet.
After reviewing this code, it seems that this case can work after all.
However, several things needed to be tweaked:
1) Removed check which stopped parented objects from getting the
visual keying coordinates determined. This actually wasn't doing
anything, given that this case would never occur as...
2) Tweaked the visualkey_can_use() function to also consider parenting
as a cause for visual-keying to be necessary.
the script has a unix-name <> real-name mapping which is not totally complete since I couldn't find everyones real names.
In this case the commit name is credited.
Also added a link to the credits page in the splash.
from Jesse Kaukonen (gekko)
--- text from the patch.
Recently Campbell Barton added callback functionality for Python's usage, but this only includes pre- and post-render callbacks. There are no callbacks for the duration of the render. This patch adds the few lines required for executing a callback while Blender Render is working. The callback resides in the rendering pipeline stats function, so whenever statistics are printed, the callback is executed. This functionality is required if one wants to:
1) Observe what is happening while Blender is rendering via the command line
2) Add custom statistics that Blender prints while the renderer works
3) The user wants to continue executing his Python script without the code halting at bpy.ops.render.render()
Personally I'm currently using this for printing out more detailed progress reports at Renderfarm.fi (such as CPU time, time spent rendering, total progress in regards to the entire rendering process). Tested on Windows, Linux and OS X.
Example on how to use the callback:
def statscall(context): print("Thanks for calling!")
bpy.app.handlers.render_stats.append(statscall)
bpy.ops.render.render(animation=False, write_still=True)
Now all sounds that are not mono but have 3D checked automatically get reduced to mono during BGE conversion time.
Also removed the now unneeded function sound_get_channels and added a missing header file to audaspace's CMakeLists.txt.
* replace by BLI_snprintf in various places, note _snprintf on windows
does not properly null terminate the string.
* fix overflow in sequencer proxy code due to buffer being smaller than
specified size.
* fix some usage of snprintf as strcpy, this is will go wrong if the
string contains % characters.
* remove BLI_dynstr_printf function in gpu module, use BLI_dynstr_appendf
Important note: I used rsync to do the local merge, as "svn merge --reintegrate ^/branches/soc-2011-pepper" doesn't work with our svn server right now!
It was introduced in rev33603 with not good patch -- release event
was catching by outliner after clicking on menu item.
Use KM_CLICK instead of KM_RELEASE to deal with icon drag/drop.
This not changes drag/drop behavior, but prevents unneeded event be handled.
Also make consistent behavior of activating and extending selection.
Added callback to drawMappedFaces which checks if two faces have got equal draw options.
After discussion with Brecht we found it's nicest solution for now:
- Disabling VBOs in edit mode for this case wouldn't be nicer for this case -
some additional flag stored in DM should be added in this case.
- Adding new callback in DM isn't nicer that this solution.
- Handling face selection in drawobject would lead to duplicated code
which is also not nice.
Hopefully, this callback could handle all cases in the future.
Also, Brecht mentioned current VBO implementation isn't perfect, so maybe
when we'll redesign this area dealing with edit mode wouldn't be so tricky.
the uniquename was never fully implemented for sensors and actuators, only for controllers.
at some point we either get rid of all of them, or bring them all on.
For now removing the "unique name" of controllers
from Tom Edwards (artfunkel), with minor edits.
This patch makes the following improvements to environment map scripting:
* Adds a "is_valid" RNA property to envmaps. True if the map is ready for use, False if it needs rendering.
* Adds a "clear" RNA function to envmaps. Deletes any envmap image data.
* Adds a "save" RNA function to envmaps. Writes the envmap to disc with a configurable layout. (Defaults to the current hard-coded layout.)
* Updates bpy.ops.texture.envmap_save with configurable layout support as above.
These changes, particularly configurable layouts, make exporting envmaps to other software much easier.
This patch adds:
* support for proxy building again (missing feature from Blender 2.49)
additionally to the way, Blender 2.49 worked, you can select several
strips at once and make Blender build proxies in the background (using
the job system)
Also a new thing: movie proxies are now build into AVI files, and
the proxy system is moved into ImBuf-library, so that other parts
of blender can also benefit from it.
* Timecode support: to fix seeking issues with files, that have
a) varying frame rates
b) very large GOP lengths
c) are broken inbetween
d) use different time code tracks
the proxy builder can now also build timecode indices, which are
used (optionally) for seeking.
For the first time, it is possible, to do frame exact seeking on
all file types.
* Support for different video-streams in one video file (can be
selected in sequencer, other parts of blender can also use it,
but UI has to be added accordingly)
* IMPORTANT: this patch *requires* ffmpeg 0.7 or newer, since
older versions don't support the pkt_pts field, that is essential
for building timecode indices.
Windows and Mac libs are already updated, Linux-users have to build
their own ffmpeg verions until distros keep up.
- particle.c wasn't setting all components of the vector when reading cache and setting dummy velocity values.
- some functions incorrectly took a float[3] argument when the 4th value was set.
- remove a few redundant lines of code.
Patch from Juha Maki-Kanto
- initgrabz was called with local space coord.
- Brush radiu was multiplying by 2.0 for grab and snake brushes.
Not sure why this was needed. Made some tests and didn't notice
any regressions without this multiplication.
This means you can import a BVH, rotate 90d and apply the rotation, the animation will still work as expected - thanks to Benjy's script for showing how obvious it is that this works :)
Description from the tracker:
"It's really handy to be able to prevent an object/material from casting a shadow. So, I made use of the Cast Buffer Shadows option in the material settings, and made it work in the viewport and the BGE."
- renaming a bone could crash if the area had to spaces in it (reported by Sebastian Koenig).
- renaming bones wouldn't update inactive 3d views locked bone names.
- selecting locked bones in the UI didnt work in editmode.
Final GSoC commit.
* Bugfix: Negative frames crashed
* Bugfix: JOS sample buffer size prediction error (wasted memory)
* Optimisation: for JOS upsampling (around 12 % difference measured here)
* Optimisation: Better filter for JOS resampling
* Bugfix: Error in relative 3D audio code.
* Removed Attenuation
* Bugfix: Multiple scenes in BGE lead to errors, BGE audio now all relative, to support multiple scenes.
to scale towards zero
This is an attempted bugfix for a bug which seems to be very fickle to
reproduce (it only happens sporadically after quickly jerking the
strips around in a certain way). So far when testing, I haven't had
any more problems after applying this fix, though it may just be
unreliable testing.
Use PageUp/Down for moving up/down, and Shift PageUp/Down for moving
to top/bottom. This is more comfortable than the old combinations
involving shift+ctrl.
Thanks pepeland and 3duan for the suggestions. I've been looking at
improving these for a while...
* Left/Right Arrow = Single Frame stepping as before
* Up/Down Arrow = Jumps to next/previous keyframe (used to be the
uncomfortable Shift PageUp/Down)
* Shift Up/Down Arrow = Jumps forward/back in 10 frame increments
(used to be Up/Down Arrows). 10 frame increment should get
customisable again as in 2.4, but need to find some UI space to put
that!
* Ctrl Shift Up/Down/Left/Right = Jump to start/end frame (used to be
Shift <Arrow Key>)
Change OURPLATFORM from "linux<major_version>" to simple "linux".
Since new policy for linux kernel versions that major version in
platform doesn't make much sense for building rules so the same
rules could be used for both of linux2 and linux3 now/
Tested on both of linux2 and linux3 systems.
- dont use hash lookups in this case because converting the string to unicode and doing a hash lookup is slower then looping over the keys and comparing (which avoids creating and throwning away a unicode string).
- operator strings were doing undo pushes (in fileselector text for example), this is dumb since the operators themselves handle undo.
- interface code checks rna props are arrays rather then checking the array length.
- disable properties window pin undoing.
- sequencer refresh was calling undo, disable since this is clearnign global data not handled by undo.
- added commented out code for drawing mesh vertex index/key index, useful for debugging shapekey - hook issyes.
Noticed clicking anywhere in the outliner was doing undo pushes, even in empty areas.
- check if any selection is made before redrawing.
- don't do an undo push when selecting outliner items since only screen data is touched here.
- Zkey to switch shading was pushing undo's.
- Wkey to interactively edit camera, lamp settings wasnt doing an undo push when it should.
- Toggling settings (such as bone boolean options) now skips an undo push if there are no items selected.
* The emitter object's inverse matrix wasn't in global coordinates during rendering, so the surface normals of the hair emission locations were transformed with the wrong matrix.
3d view header mode menu.
A recent bugfix was incorrectly hiding pose and particle mode when the object
data was library linked, but these modes edit object level settings so should
be available.
* Removing check if Node is in between, so in-/output nodes can be muted as well. Useful for example if you want to temporarily mute a file output node.
linked
Shapekey actions weren't getting copied when their owner data was.
This was due to the IMO totally convoluted way in which the duplicate
action flags have been set up:
- the function to copy animdata takes a param to specify whether
actions are copied or not, but this is never touched (i.e. this always
just gets FALSE passed in)
- instead, we jump around in hoops later figuring out whether the
userpref wants copying to occur, then fixing up the links
IIRC, part of this may be due to a desire/need to not duplicate
actions when dealing with NodeTree copies for multi-threaded
rendering, but at the expense of complicating everything else.
High quality resampling on mixdown, linear for playback.
* Lots of improvements and fixes for the JOS resampler, now it works fine!
* High quality filter coefficients for the JOS resampler (sorry for the 5 MB source file).
* Fix for GE orientation bug. Note: moto uses x,y,z,w quaternion storage, while rest of blender uses w,x,y,z.
* Minor changes/fixes.
There's no reason to have it really, this situation is totally normal, and it
means a terminal window is opened on Windows as long as you haven't saved any
default settings yet.
in a particular setup with two nested material nodes. Material active texture
was looking also recursively into material node, but this was already done
outside of this function.
* Split off code to refresh relative/builtin KeyingSets for the
current context before they get used to a separate function.
* Hooked this up to a new PyAPI/RNA function: KeyingSet.refresh().
Call this before checking the paths that a Keying Set has, especially
if it is not "absolute"
* Added option for "Select Grouped" operator (for Objects), which will
select all objects affected by the active Keying Set. This is probably
more useful for absolute KeyingSets, where changing the selection is
less likely to affect the result.
- The equivalent for bones is currently still in development, but is
likely to be more useful for animators, where rigs are the primary
animation entities they deal with
So apparently this was a regression from 2.4x, since vector handles
were one of the handle types which could be set independently for each
handle (vs both needing to be the same, for example, Auto Handles)
infact this is not really a bug, irrespective zero vertex group weights gave overlapping geometry which isn't useful, add an option to set the thickness factor for zero weighted verts.
While I originally made these so that they wouldn't be on by default
due to concerns over the filtering used for these leading to reduced
framerates/interactive speed, I'd since found while doing some
profiling that this isn't the case.
Rather, decreased framerates were more likely to stem from trying to
perform the checks necessary for the long-keyframe drawing when many
"child" channels are involved (affecting other channels too).
So, since these are generally useful, they are now enabled by default
:)
- add change data operator to strip panel next to image file properties since editing every image manually isnt really usable.
- added new sequencer operator "Clear Offsets" (Alt+O), useful to reset the start/end frames around the strip data.
* Code cleanup in the space_*.py files.
* Removed layout.column() statement in _MT_ panels, they are useless.
* Only define variables at the beginning of a function!
*BLENDER_VERSION_CYCLE set to alpha
Idea is to set it to alpha for the first 3 weeks (B-Con1)
Beta for B-Con 2 and 3
RC for B-Con 4
Release for B-Con 5
Causing a flurry of refresh file prompts post-commit,
Confusing local diffs and causing merge conflicts,
Stating the obvious; redundant and useless...
We shall not miss thou, blasted expand $keywords$
Currently, grease pencil conflicts with such operators
as undo and set object mode which makes behavior totally
unpredictable and crash for some cases.
The only way to solve this proper is to ger rid of pointers
to data which can chage stored in operator custom data.
Adding a mono flag to mixdown non-mono sounds for 3D audio.
* Added mono sound loading.
* Bugfix: AUD_COMPARE_SPECS usage was wrong.
* Bugfix: JOS resampler = instead of ==.
* Bugfix: Change of a sound should apply settings in AUD_SequencerHandle.
* Bugfix: Memory leak when canceling open sound operator.
if vec.dot(vec) > some_distance*some_distance:
do_something()
might not be quite as obvious looking as:
if vec.length_squared > some_distance*some_distance:
do_something()
As to why you'd want to use length_squared over length is that length uses a square root, which isn't always necessary for simple distance checks (e.g., closest object, checks like the ones above, ect).
Reported by Jean Francois Sarazin
Lack of normal normalisation caused scaling issues. SIGGRAPH fix by jahka and jesterKing. Thanks to host dfelinto with entertainment provided by slikdigit.
2 changes
- When writing OGG only allow Theora encoding, this fixes the crash.
- When setting the MPEG preset, dont allow the 'Codec' to be left as Theora, this is just confusing.
* note that this is highly confusing for users and devs - there are 4 places to set the codec/format, with both python and C presets :S.
Improved waveform drawing in the sequencer.
* Drawing the waveform of a sequencer strip is now independent from whether the sound is cached or not.
* Improved drawing of the waveform in the sequencer (especially speed!).
* Making it possible to vertically zoom more in the sequencer to better see the waveform for lipsync.
* Fixed a bug which crashed blender on loading a sound file via ffmpeg.
Speaker objects fully functional!
Minor changes:
* Fixed three memory bugs found via valgrind.
* Fixed bug with jack transport crashing after file loading.
* Sound NLA Strips now start at CFRA instead of 0.
* Blendin for Loop End works even after a negative pulse. Flipper could still use some work in this area.
* Continuous works a lot better.
* BL_Action::SetFrame() should work a little smoother.
- building with GHOST/SDL now adds the PREFIX
- image.external_edit operator errors out if the filepath isnt set (was annoying and loaded gimp while running tests)
Second part of previous commit. Now, when speaker objects are created,
they are created by default with an NLA sound strip so that it is easy
to just start immediately using this to do cool stuff (i.e. timing
when you want the sound to start).
These are basically just for specifying when a speaker should fire off
it's soundclip, and as such, many NLA operations are irrelevant for
it. They can only be specified on object-level for speaker objects.
I've still got some UI tweaks I'll need to work on in order for these
to be able to be added even when the speaker doesn't have any NLA
tracks yet. (EDIT: while typing this, I had an idea for how to do
this, but that'll be for next commit). In the mean time, you'll need
to add a single keyframe for the object, snowflake that action and
delete the NLA strip before you can start editing.
* Pepper depends on ffmpeg 0.7.1 or higher now, windows and mac build systems set to ffmpeg-0.8
* Fixed orientation retrieval in OpenAL device code.
* Added stopAll() method to AUD_IDevice (also for Python) and call it on BGE exit
* Changed BGE to use audaspace via native C++ instead over the C API.
* Made AUD_SequencerFactory and AUD_SequencerEntry thread safe.
* Changed sound caching into a flag which fixes problems on file loading, especially with undo.
* Removed unused parameter from sound_mute_scene_sound
* Fixed bug: changing FPS didn't update the sequencer sound positions.
* Fixed bug: Properties of sequencer strips weren't set correctly.
* Minor warning fixes.
Mixdown functionality.
* Mixdown possible via libsndfile and ffmpeg!
* Fixed some ffmpeg deprecation warnings
* Mixdown UI only shows working Container, Codec and Format combinations!
* Minor bugs and warnings fixed
added _prop suffix to ui functions which take a prop rather then a propname, may change this later since its not that nice but for gsoc branches this keeps existing UI functions working the same.
This commit makes some tweaks to the way that the "active keyframe"
settings in the Properties region in the Graph Editor work (for the
better, hopefully).
Basically, the problem was that previously, these were clunky and non-
intuitive to use, since they were just directly displaying the RNA
properties for those keyframes for editing purposes. But due to
limitations of RNA (i.e. from a RNA pointer to a keyframe, you
couldn't see which F-Curve you came from), several things were
impossible, notably:
1) Doing proper updates, including validating that the handles are in
a valid state - that requires access to the F-Curve to provide to the
F-Curve-based curve validity checking functions
2) Having the values of the keyframes display in whatever unit that
the property the F-Curve affects displays as - for this, you once
again need to know the F-Curve in order to resolve the property that
it affects; also the fact that only a single unit could be set for RNA
properties further limited things
This commit basically gets around these problems by moving away from a
layout-engine based approach to one where we attach custom update
callbacks and also override the units of the y-co widgets when
creating the widgets for these, thus allowing the buttons to work in
the ways that animators expect.
Diff Keymaps
User edited keymaps now no longer override the builtin keymaps entirely, but
rather save only the difference and reapply those changes. This means they can
stay better in sync when the builtin keymaps change. The diff/patch algorithm
is not perfect, but better for the common case where only a few items are changed
rather than entire keymaps The main weakness is that if a builtin keymap item
changes, user modification of that item may need to be redone in some cases.
Keymap Editor
The most noticeable change here is that there is no longer an "Edit" button for
keymaps, all are editable immediately, but a "Restore" buttons shows for keymaps
and items that have been edited. Shortcuts for addons can also be edited in the
keymap editor.
Addons
Addons now should only modify the new addon keyconfiguration, the keymap items
there will be added to the builtin ones for handling events, and not get lost
when starting new files. Example code of register/unregister:
km = wm.keyconfigs.addon.keymaps.new("3D View", space_type="VIEW_3D")
km.keymap_items.new('my.operator', 'ESC', 'PRESS')
km = wm.keyconfigs.addon.keymaps["3D View"]
km.keymap_items.remove(km.keymap_items["my.operator"])
Compatibility
The changes made are not forward compatible, i.e. if you save user preferences
with newer versions, older versions will not have key configuration changes that
were made.
* Tweaked order of handle types to make it easier to find Auto/Auto-
clamped in the list
* Fixed a number of places which were still just checking for auto-
handles when they should have included auto-clamped too, including
handle rotation
- icons were scaling by the sqrt(dpi)/8.48528, but infact they only need to be scaled by (dpi/72).
- UI_icon_get_width value was being used without multiplying by dpi scale.
dont show a popup anymore, was silly because you had to change the value for before anything was done, can use f6 redo popup instead, sequencer should eventually have a view3d operator redo panel.
handle/key
This used to be a weird per-curve setting which would happen to get
applied/work correctly if handles were set to "auto", and was a source
of constant confusion for both old and new animators. The main effect
of this handle-type/option was really to just ensure that auto-handles
stayed horizontal, instead of tilting as the keys were moved.
This commit simply changes this from a per-curve to per
keyframe/handle setting.
Implementation of Julius O. Smith's resampling algorithm for high quality audio resampling.
The filter currently is a sinc filter with a 0.9 cutt-off and windowed by a kaiser window (beta = 10).
Also includes minor changes in the linear resampler.
Some people like curved lines, other hate them. This commit will let the user change the level of curving.
In UserPreferences=>Themes=>Node editor=>Noodle curving the level can be modified. Allowed range is 0-10 with the default on 5
The patch will default everything to the way blender works ATM.
File subversion has been increased otherwise older 258 files got straight lines.
The data is stored in the ThemeSpace.noodle_curving
the bezierdrawing is done in the drawnode. Also tested the Line cut tool
existing check for driver to use GIL was not thread safe and could cause, details in the report.
This bug was caused by a check to avoid hanging, a fix for [#27683] that worked in 2.4x because the UI didn't use python to draw while rendering.
Apply a different fix for [#27683], when calling an operator, call PyEval_SaveThread(), then PyEval_RestoreThread() so the GIL can be aquired by threads started by the operator - in this case bake starting a thread that evaluates drivers.
This is for bug #28141
While not really a bug, it makes it a lot easyer to use if it
has the exension. (Isn't hidden from the user when they try to load it...)
Kent
to conficuration variable BF_FFMPEG_DDL.
This would allow to use different FFmpeg in buildbot.
- Added some 3DMOUSE variables to list of command line options.
Now 3dmouse related-settings can be set from command line.
* Minor audaspace library improvements.
* Considering location, velocity and orientation in AUD_SequencerReader and AUD_SequencerHandle.
* Bugfix: Maximum and Minimum volume weren't used before in the software device.
* Bugfix: Adding speaker objects via info space crashed.
* Listener settings now get updated in the audio system.
NDOF is disabled in CMake.
- Added "default" section to switch in sendMotionEvent.
It's what strict gcc rules don't like much and it's
And it's good practice in general, imo.
the Outliner
Channels which can be renamed include:
- Scenes, Objects, World, Material, Texture, etc. (i.e. "ID-blocks",
or the dark and light blue channels)
- Action Groups (green channels)
- Action expanders (i.e. "CubeAction", "WorldAction", etc.)
- Grease Pencil stuff
Channels which CANNOT be renamed, as they mostly use hardcoded values
or otherwise include:
- Drivers expander
- FCurves (they don't technically have a "name"; what is shown is just
a user-friendly representation of their rna_paths)
python scons\scons.py BF_BITNESS=32
python scons\scons.py BF_BITNESS=64
So from now on for Windows you don't have to run a specific win32 or win64 Python version
to get the Blender version you want. If you omit the BF_BITNESS flag the bitness as per
Python version will be used as before.
Note that this is an argument, so works *only* on the command-line, not in your user-config.py
- added includes for spnav
- added FindSpacenav.cmake which allows using spacenav from a nonstandard path.
- remove NDOF_LIBPATH, use a full library path instead.
use default when invalid settings given.
removed odd/annoying bit shifting of the flagu/v for such basic function made code hard to understand and would fail if new flags were added.
- The object ID was being passed to the data_delete_material_index_id() from object_remove_material_slot(), rather then the object data. (so the material slot fix wouldnt run in that case).
- add support for fixing text object materials too.
create a new parameter for materials.pop() to not remove material slot.
this way the mat_nr is still the old one.
for the default behaviour we now have material remapping (i.e. data_delete_material_index_id(id, index)).
This new function is brought from the material_slot remove function.
Not sure why this was enabled, possibly from copy/paste with bezier->nurbs code?
If you have meny poly lines there was no nice way to convert these into a smoothed nurbs curve.
Ran into this when trying to convert generated ivy into smooth nurbs.
OK, after that failure of committing a bunch of old junk, hopefully
this is what I actually meant to commit :)
* Added big comments to some of the fields in struct PBVHNode. I
always forget the details of these, so finally wrote it down
properly.
* Changed types of PBVHNode.face_vert_indices and PBVHNode.flag to
better reflect their contents.
* There should be no functional changes here.
I named all the BGE modules with their actual names (e.g. Rasterizer, Video Texture, ...). so in the API index.html page they look more like the other Blender modules.
I did the same for the bgl module.
For bge.constraints this patch exposes the constants values for debug mode and createConstraints (they were hardcoded innts before).
+ making all the "todo" and #comments into rst comments (.. comments)
Thanks Solano, it's great to get help to those tasks :)
Old issue with OSX Cocoa code: shift+scrollwheel should send
a 'horizontal wheel' event to Blender. Blender only recognizes
scroll events in general though. The old code then just didn't
send an event at all, not passing on shift+scrolls.
Now the scroll event is sent anyway, relying on Blender's
keymapping to define what to do with shift+scroll.
This fixes things like shift+scroll to scale ListBox widgets.
Implemented basic audio animation.
* AnimatableProperty: Propper cache writing and spline interpolation for reading (the solution for stair steps in audio animation)
* Animatable properties so far are: volume, pitch, panning
* Users note: Changing the pitch of a sound results in wrong seeking, due to the resulting playback length difference.
* Users note: Panning only works for mono sources, values are in the range [-2..2], this basically controls the angle of the sound, 0 is front, -1 left, 1 right and 2 and -2 are back. Typical stereo panning only supports [-1..1].
* Disabled animation of audio related ffmpeg output parameters.
* Scene Audio Panel: 3D Listener settings also for Renderer, new Volume property (animatable!), Update/Bake buttons for animation problems, moved sampling rate and channel count here
Enabled by default and also enabled for older filesm so
there should be no regressions.
In some cases it's useful to not use subdivided uvs for multires.
*Removed curve init code in readfile (no more needed since the split broke anyway compatibility with earlier WeightVGroup files…).
*Updated get_ob2ob_distance() code (much simpler – I’m not a matrices’ god!).
*Enhanced a few RNA names (Campbell has others in mind here, though, I think).
Main: Complete rewrite of the sequencer related audio code to support 3D Audio objects later and especially adressing the animation system problems (see mailing list if interested).
Note: Animation is not working yet, so with this commit volume animation doesn't work anymore, that's the next step.
Minor things:
* Changed AUD_Reference behaviour a little to be more usage safe.
* Fixed bug in AUD_I3DHandle: Missing virtual destructor
* Fixed enmus in AUD_Space.h
* Fixed a warning in rna_scene.c
* Removed an unneeded call in rna_sound.c
- Default size of Graph Editor handle vertices is now 4 (up from 3).
This "small" change seems to be enough to make a substantial
difference when it comes to actually differentiating between these
- "Only Selected" DopeSheet filter is enabled for new Graph Editor
instances by default. It helps hone in on the F-Curves of the data
most animators hope to just be refining the motion for (i.e. the
selected stuff)
- "Only Selected Keyframe Handles" is now enabled, to reduce clutter
from handles of nearby keys getting in the way.
Bezier curves are now drawn smoother (i.e. less segmented), especially
for curve segments where there is a very large vertical displacement
over a short period of time (i.e. 120 degrees rotation over 1 frame)
and/or often when zoomed in a bit too.
- Made the resolution calculation take the vertical distance into
account too, instead of just the horizontal distance.
- Segment multiplier changed from 3 to 5, as this seems to give better
zoomed-in performance.
- added menu templates
- move template menu into the header of the text editor (so users will find more easily)
- updated mathutils examples, switching the order of multiplication.
trying to fix blenderplayer startup warnings. It seems we do need to init
the SDL video subsystem even if we only want events, thanks Juha Maki-Kanto
for pointing this out.
This was never highly tested, that's why I never committed (my patch for this was from September 2010).
But once again I got a report that this bug was a deal-break and the patch seems to work for this artist.
I believe it's working, but I will keep my eyes open for this.
vector * matrix
vector *= matrix
vector * quaternion
vector *= quaternion
Use the reverse order instead, enable WITH_ASSERT_ABORT in cmake to promote the warnings into errors.
* where_is_object_time was called for every effector evaluation only to determine the object velocity in some rare cases.
* Calculating the effector velocity is now done in the effector precalculation stage.
* Removing this makes the code thread safe and also should give some nice performance boosts when simulating a lot of points.
* Thanks to MiikaH for noticing this problem.
This commit fixes very noticeable seams caused by margins
calculated incorrectly. This commit changes way margin is
calculated in and makes textures really seamless.
Also margin limited to 32 isn't good now -- artists are baking
really large textures nowadays so margin is now limited to 64px.
Thank you, Morten!
This fixes bug #26764 and several others like it, where modifier
properties (and others, but most visibly modifiers) would not do
anything when animated or driven, as modifier properties require the
RNA update calls to tag the modifiers to get recalculated.
While just adding a call to RNA_property_update() could have gotten
this working (as per the Campbell's patch attached in the report, and
also my own attempt #25881). However, on production rigs, the
performance cost of this is untenatable (on my own tests, without
these updates, I was getting ~5fps on such a rig, but only 0.9fps or
possibly even worse with the updates added).
Hence, this commit adds a property-update caching system to the RNA
level, which aims to reduce to the number of times that the update
functions end up needing to get called.
While this is much faster than without the caching, I also added an
optimisation for pose bones (which are numerous in production rigs) so
that their property updates are skipped, since they are useless to the
animsys (they only tag the depsgraph for updating). This gets things
moving at a more acceptable framerate.
* Fixed an error when "Custom" was used and the path was empty (variable reference before assignment)
* The Operator now raises an error if there is no path or the path given is not available.
not exactly a bug, but the option to copy individual properties was not working from the SPACE menu.
I believe this was happening because we are using dynamic enums.
This commit makes the "merge" option to be the default one. So if you call it from the SPACE menu it will be the one used.
* Reviewed and improved the linear resampler. Now it should work pretty good also for special cases that caused errors previously.
* Fixed a crash in the GE when a sound actuator doesn't have a sound assigned.
* Corrected the OpenAL device's threading code. This is a bugfix for #27913, thanks to Juha Mäki-Kanto for helping to resolve this.
* Added back icon to open the Splash Screen in the info header, next to version string info.
* Removed an unnecessary toggle argument for particle mode select buttons. The Toggle argument is only intended for booleans, not enums.
- Intersection code was using undefined vector
caused wrong lines to be picked
- Code now also copes with hidden sockets.
If all fails, is just unhides a good socket.
On dragging a non-connected node on a noodle, it will insert it.
Functionality tweaks are possible, but it already feels non-intrusive.
Rules:
- Insertion only when a single noodle is intersecting with node.
- Default connects first matching socket type.
- If no socket match, it connects the first.
Recoded side-chain reparenting step to fix (as far as I've been able
to tell) infinite looping problems which were a bit intermittent here
using the test file. The fix here involves some tighter checks to
prevent corrupting the parenting of bones in the run of bones being
merged but also of any ancestors of those.
Paste pose no longer just does a blind "replace all properties" on
bones that it pastes on. Instead:
* when properties exist on the target already - only change the
properties in common
* when properties don't already exist - copy all properties
Time-scale drawing wasn't respecting the time unit setting.
While working on this, I tried to tweak the grid drawing to a more
common setting. It's hardcoded to show lines at every 25 px = once
every 25 frames, which is only really fine when FPS=25. Anyways, this
works fine enough for the sequencer for now in general usage.
functions based on editor types
This could be split up further in future if there's such a need, but
this should already be sufficient. Most notably required since the NLA
recalc stuff was taking quite a few lines within that block
Blender render optimizes alpha=0 materials away, unless it has
a number of properties... but there wasn't a check for material
being ray-mirror, it then should be rendered always.
[#28032] Python Mathutils: Matrix Multiplication Error
Since they ended up reversing the order we better keep old code unless its proven to be incorrect.
also change Matrix.__repr__ function args to evaluate correctly (need to be inside a tuple).
From what I can tell there is no good fix for this bug, calculating the 2d/3d viewborder and then attempting to align them to be pixel perfect fails because of float imprecision.
Added a workaround, so the camera border is always drawn in 2d space,
since this workaround may cause problems later on its kept under the define VIEW3D_CAMERA_BORDER_HACK so we can get old behavior back easily.
Generated images would not be re-generated with a float buffer on load, even when selected on creation.
Now save the float buffer setting as a generated image flag.
This means you can enable before baking to enable baking to a float buffer.
rendering, to prevent any race condition problems
I've noticed some weird and random crashes recently while rendering,
which I suspect have been arising from having an Action Editor open
while rendering. Previously only the timeline was patched against
these problems, though the issues may be more widespread. Hence,
solving this problem at the root cause instead.
Issue was that the Shader tree execution changed the ShaderInput.
Changes are that the UI is updated that only the main material will have the pass_index this is displayed in the "render pipeline options" panel.
When the material is not a node material the pass_index will be shown at the "options" panel
To test enable nodes on the material
Add a new input material
change the pass_index of the material (render pipeline options)
Enable RenderPass material ID and use the compositor to read out the material pass
Jeroen
xsch and ysch were originally planed to replace the scene->r.xsch/r.ysch
however in blender/3dview we still need to use the r. values. Therefore we can't really run
from using those values even in bplayer. So removed the values in gamedata.
The way it's now, render values (xsch and ysch) are responsible for aspect ratio and gamedata xplay and yplay are responsible for the size of the window.
I believe this bug was there since we (me) moved the game settings to scene->gm
Since I was here I added support for x/y non square aspect pixels (i.e. anamorphic)
we were already using it for videotexture so I don't know why we were not here.
Tested in OSX, but it should be working in all OSs.
the overflow of the clock was causing crash in the game engine in Linux.
(on June 11 2011, 18:39:00 GMT)
running to the "where is waldo (wally)" bug award of 2011.
Addons are checked for their timestamps and reloaded when it changes but this failed when, 2 addons had the same name since different times caused 2 reloads on every redraw.
Now when duplicate addons are in the path now give a error message in the UI and print path conflict in the console and don't thrash reloading.
* Find first unused frame function was failing to correctly detect
conflicts with the lower bound due to the way that markers are not
stored in sorted order. Fixed by performing additional search passes.
* Fixed some update bugs where there were missing notifiers. Most
noticable when the poselib is being viewed in an Action Editor
- add back slicing for buffers, (was previously in 2.4x but not working in py3):
buf = bgl.Buffer(...)
ls = buf[:]
- fix for crash with negative index access not being clamped.
- improve repr() function for multi dimensional buffers.
- add back 'list' attribute, but print deprecation warning.
Now it behaves right on playback:
- Starting playback "Anim Player" button appears on header.
It used to appear only on mouse hover before.
- Stopping playback triggers refresh on compositor, so
actual result would be visible if image sequence/movie
is used in nodes.
Visual Keyframing was broken by r.34685, which used another method
which, at the time, appeared to work perfectly fine. Apparently not.
Also, extend/fixed visual keying to work for axis-angle rotations too.
Needs some testing, but should probably work
* Adding methods KX_GameObject.stopAction() and KX_GameObject.isPlayingAction().
* Made all layer arguments optional. This means I had to change setActionFrame(layer, frame) to setActionFrame(frame, layer=0). This seems a little backwards to me, but I guess that's what you get with optional arguments. Also, this will break existing scripts.
* Made sure to check user supplied layer values on all action methods. Previously this was only done for playAction().
* Fixed a few newline issues.
* New object pointers can't be loaded properly for library linked groups, so the weight groups now store an index to the group objects at save time. This index is used at load time to set the objects without relying on the old pointers.
* If the library linked group is modified the indices can be wrong, but this can't really be avoided easily as there's no way to relate objects in a linked group between loads.
- use getset rather then getattr
- remove verbose Buffer docstring, better use sphinx docs for this
- replace bge.Buffer with the Buffer class and add a __new__ function to the class so it can be called.
- improve error messages
* Fixed a warning in AUD_DoubleReader.cpp
* Removed some unneeded includes
* Fixed a bug resulting in a crash when stopping a sound
* Fixed a bug where a NaN resulted in a horrible memory error
* Fixed a typo bug which caused crackling in audio playback and export
* Added memory debugging code (ifdefed)
- use char rather then STR_String for the event printer.
- added option to build WITH_GHOST_DEBUG for cmake
- renamed WITH_SDL_GHOST --> WITH_GHOST_SDL
when using the node on a single white pixel on black background, the output should look like as the bokeh image.
being a round image, but it looked like a donut.
the make_gausstab used dist/rad and bokeh used (dist/rad)*2 - 1
I changed it to reflect the correct bokeh circular image
this will reconnect nodes as if the deleted node is muted.
Operation is added to the space_node node menu and to the keymap as CTRL-X
to test this just add some nodes to the space_node
select one or multiple nodes and press CTRL-X
It should reconnect the nodes as they were muted
limitations:
1. it performs a delete and reconnect per node. It does not evaluate all selected nodes as one whole
2. mute only supports Value, Vector and Color data types, so does this feature
3. not usable for nodes where input and output does not match (like colorToBW)
Where reconnect could not be preformed the links will be removed from the model.
Undo works with this delete with reconnect.
This opens up the option for blender to be more easily ported to other devices, OS's.
TODO
- continuous grab.
- text glitch with multiple windows (was a bug in X11 too for a while, will check on this)
keyframe lines are wrapped up nicely by it
Ideally it could be made so that it only became wide when it is on a
frame with a keyframe, though that could end up causing performance
problems, so this will have to do (if a bit "chunky" looking at
times).
* Keyframe lines were being drawn too short when frame number box was
enabled. The code for drawing this was modifying the View2D view-space
to get it's stuff in the right place, but the timeline code was not
accounting for this.
* In order to make the time ticks more visible outside the frame
range, I've moved the start/end frame drawing stuff in timeline to
occur after the grid drawing, and to draw semi-transparent, just like
the preview range curtains in the other animation editors
* When a track is being solo'd, all other channels for that block are
drawn darker
* Strips in non-solo tracks are drawn flat shaded instead of with
shading
* Mute toggles are hidden (they wouldn't affect the result)
Added some new star icons for the "solo" toggles in NLA editor.
Unfortunately they look a tad scruffy alongside some of the other
icons, although they should hopefully turn out to be more descriptive
(especially when combined with some drawing tweaks I've got in the
pipeline...)
wrong entries if obdata selected
In this case, the problem was that there were some lingering F-Curves
that were unselected by still had "active" flags set (a problem caused
by the old filtering channel visible vs list visible bug). Now,
"active" flag is treated separately from "selected" flag (bringing
this back into line with bones), leaving no confusion.
blocks and Objects from add_element
These two chunks were significantly large that they really needed to
be placed into their own functions to allow for easier source
navigation.
As per my proposal (http://lists.blender.org/pipermail/bf-
committers/2011-July/032553.html), I've split outliner.c into several
new files based on the purpose of the relevant code.
* outliner_tree.c - building outliner structure
* outliner_draw.c - outliner drawing (including toggle buttons and
their handling)
* outliner_edit.c - all operators for toggling stuff, and/or hotkey
accessed operators. Also KeyingSet and Driver operators go here
* outliner_tools.c - all operators and callbacks used for handling RMB
click on items
* outliner_select.c - stuff for selecting rows, and handling the
active/selected toggling stuff
In a few cases, the split hasn't been totally clear-cut due to cross-
dependencies and other spaghetti. However, in a few cases, I have
managed to remove the need for some of the prototypes that were needed
in the past by judicious reshuffling of functions, which also makes it
easier to actually find what you're looking for.
* Objects are now always rotated in the directions of the hair paths
* Secondary fix: particle size wasn't updated for hair particles, so dupliobject size couldn't be change after the hair was edited
* Noise is now considered an animated texture as it changes with every frame
* Converted a few places in particles code to use the particle system's own random table instead of BLI_frand.
It was a regression introduced in rev36301. Average normal calcilation
used to fail due to triangular faces which are too slight.
Do not use triangles with too small area for average normal calculation.
This patch attempts to clean up and document the GPU buffers
code. There are a few bug fixes as well.
Patch reviewed here: http://codereview.appspot.com/4631052/
Summary:
* Bugfix: make GPU_buffer_copy_normal convert from shorts to floats
correctly, also fixed the use of cached face normal CustomData.
* Bugfix: changed the `mat_nr' field of GPUBufferMaterial from char to
short.
* Changed color buffer setup to not alloc a temporary copy of color
data, just passes the MCol data in directly.
* Changed the GPU buffer pool code to make clearer what operates
specifically on the global pool.
* Lots of refactoring for GPU_drawobject_new; should operate mostly
the same (except got rid of one unecessary allocation), just split
into more functions and without macros now.
* Converted some #defines into enumerations.
* Made some stuff private, pulled out of header file.
* Deleted unused function GPU_buffer_pool_free_unused().
* Removed GPU_interleaved_setup and related #defines. (I think this
was used for editmode VBOs, but those were disabled.)
* Added lots of comments.
* Added a few comments in the code signed `--nicholas' to note places
where I am unsure about design or usage, would be good to address
these better.
* Code formatting changed to be more consistent with the rest of
Blender.
* Renamed some fields and variables to be more consistent with
Blender's naming conventions.
* Renamed some fields and variables to use more descriptive names,
e.g. renamed `redir' to `mat_orig_to_new'.
* Removed print outs with DEBUG_VBO -- don't feel too strongly about
this one, just not used elsewhere in Blender, could be easily added
back if others disagree though.
* Moved the PBVH drawing code down to the bottom of the file, before
was sitting in the middle of the other VBO code
Softbody was still using a flag to determine if it should use the final or
deform derivedmesh, but this wans't exposed in the UI. Others systems use the
collision modifier, now softbody uses it also to get vertices and faces, but
with own collision code.
+ some small fixes in other docs.
A topic for later(soon?), I think we should name the modules according to the rest of Blender modules. So instead of:
Game Engine bge.logic Module
We would have it:
Game Logic (bge.logic)
...
Extrapolation" tool again
Added "Make Cyclic" and "Clear Cyclic" options to "Set Extrapolation"
tool (found from Channels menu) in Animation Editors. These options
simply add or remove (respectively) Cycles FModifiers from the
selected F-Curves, making them have cyclic extrapolation with a single
click, instead of having to go through the FModifiers UI (or Graph-
Editor only "Add FModifier" operator), which should make it easier to
do this apparently common chore.
* Ported filtering code for Grease Pencil frames editing to the newer-
style refactored stuff
* Decoupled active status of layers from selection status, bringing
this into line with everything else again
view for transforming strips
When the "Include animation data blocks with no NLA data" toggle
(action icon) is off, action lines are only shown if they have
keyframes. So when this option is off, only NLA blocks that have NLA
tracks will be shown, and of those, only those which currently have an
active action with keyframes will have their red action lines shown.
Combined with the vertical-space tweak when show control curves is
turned off, this should be good enough for most cases.
Thanks Torsten Rupp (rupp) for the patch!
This patch adds the abilities to sort the bone group list in the
properties panel and to move bone groups up/down in the list (similar
like for vertex groups)
* When clicking on "Animation" items in the Outliner, there's now a
menu containing from which you can change the action used, and
refresh/delete all drivers.
* Moved action-setting logic for AnimData actions to a single utility
function in anim_sys, since this was starting to be done in too many
places already.
* Fixed Outliner refresh bug after changing the active action
Reported by David Roy
Patch by Brecht van Lommel
UV import code wasn't taking possible stride into account (always assuming stride==2), thus reading UV coords totally wrong.
The information was written in the temp exr files, but was not read back.
After checking I saw that the pass was not merged back in the rendercore.
After adding this it worked. tested with all FSAA settings.
Old light-blue colouring for "keyed" bones is no longer applied, even
if the flags were set in earlier versions of Blender.
This was a legacy feature used to get around some ancient issues,
which isn't needed anymore. Instead, it ends up causing confusion, so
removing.
... and I can't test rst in windows so for now let's pretend it builds. If someone can generate the docs and see how it goes please let me know.
(plus a small fix for bge.logic rst)
disabled, the strips are drawn so that they take up less vertical
space.
Originally, the primary reason why these were taller than those in the
other animation editors was really so that these control curves could
be visualised adequately. So, when these aren't shown, we can afford
to collapse the strips vertically.
This should make it possible to fit more strips on screen to retime
them. in some staggered fashion.
Wavelet tile generation code was reading outside of array bounds, and as a
result could give a noise.wavelets file with nan values. Now that problem
is fixed, and existing files with nan are ignored.
* Horrors from the ancient world of deprecated code: object animation offset can't really work correctly with particles unless point cache takes full control of particle system timing.
* Disabled the non-working offset control from effecting particles so that for now particles will work consistently and the offset is only applied to the object.
didn't perform updates
* This problem was caused by a typo when adapting old code
* Fixed crash where keyframes-update was being called in Grease Pencil
transforms too
Todo:
Outliner/Datablocks Viewer doesn't update that nicely when these
keyframes get modified. Outside of gdb, I managed to get a few non-
repeatable crashes here; while debugging though, there was only some
lagging oddness if panning before the tree updated.
I can't test this here, somehow blender+scons is failing in my building env.
but it should work. Basically it skips the copy of the script folder for the blenderplayer.app
I will test it once buildbot get pass this review ;)
===== Situation before this patch =====
in the current situation inside the node editor there is a properties panel (press 'n'-key). This pabel displays some information about the node, backdrop and grease pencil. The UI of the property panel is typically vertical oriented. Nodes in the other hand are not oriented in a direction. Both area's are draw via the same draw function.
With some nodes this will create not user-friendly UI. Try the color-balance for instance). The 3 color circles are drawn next to each other, it would be better to draw them below each other.
When creating more complex nodes you don't want to display all handles in the node-panel and in the properties panel. For instance fine-tuning handles you only want to appear in the property panel to reduce place in the node itself.
===== Situation after this patch =====
This patch separates the draw functions of the property panel and the node panel.
When no special draw function is created for the property panel, the draw function of the node will be used as 'fallback'
===== Impact =====
==== BKE_node.h ====
add a new uifunc (called uifuncbut) to the bNodeType struct. The definition is the same as the uifunc.
==== node_buttons.c ====
if the uifuncbut is set, call it. currently calls the uifunc method
==== drawnode.c ====
static void node_composit_set_butfunc(bNodeType *ntype). set the uifuncbut function where needed. When at the end of the method uifuncbut is still empty, set uifuncbut to the uifunc.
===== Final note =====
! PS. this is not limited to the compositor it also works for Materials and Textures !
! PPS. For other branching creating their own node-tree. Please make sure that your uifuncbut is set NULL or a valid draw function !
A mesh can consist out of multiple material. Take a character with clothing's. the skin can be a different material as the different clothing's. During compositing it is a common use-case to only do a part of the composit on only a specific material. Currently this can not be done.
In blender movies this feature is known to be implemented, but until now it never got integrated into trunk.
Proposal
With material index the Blender internal renderer will be capable of creating a buffer containing the material indexes of the first pixel-hit. This will be implemented in the same manner as the object index.
In the compositor the ID Mask node can be used to extract the information out of the Render pass.
Impact
User interface
On the properties-space the next changes will be done
Scene⇒Render layer⇒Passes⇒Material index will be added
Material⇒Options⇒Pass index will be added
DNA
Material struct will get an new field called “index”. this will be a short-type.
Material struct the field pad will be removed.
A new Render-layer pass will be added (bit 1«18)
RNA
Material RNA is updated (based on “pass index” from object)
Render layer RNA is updated (based on IndexOB)
Blender internal renderer
The Blender internal renderer will process the render pass as a copy of the Object index.
Blender compositor
The render layer input will get a new output socket called “IndexMA”
Usage
An example on how to use material index can be found at:
https://svn.blender.org/svnroot/bf-blender/trunk/lib/tests/compositing/composite_materialindex.blend
This is also example of a commit message longer than the commit itself :)
The manipulator was always using vertex selection flags, but those are only
valid in vertex mode, as selection flag flushing only happens in the direction
vertex -> edge -> face. Now use edge/face selection flags when needed.
Original patch by Jan Diederich, adapted by Pelle Johnsen. Review assistance by Daniel Tavares.
This patch adds an option to export only the selection.
It is now possible to create "scripted expression" drivers by simply
clicking on some property, and typing some short Python expression
prefixed with a '#'. This will result in a scripted expression driver,
with the typed-in text being created.
For example, you can click on X-Location of the default cube, and
type:
#sin(frame)
and a new driver will be created for the x-location of the cube. This
will use the current frame value, and modulate this with a sine wave.
Do note though, that the current frame is a special case here. In the
current implementation, a special "frame" driver variable, which
references the current scene frame is created automatically, so that
this simple and (assumed) common case will work straight out of the
box.
Future improvements:
- Explore possibilities of semi-automated extraction of variables from
such expressions, resulting in automated variable extraction. (Doing
away with variables completely is definitely 100% off the agenda
though)
- Look into some ways of defining some shorthands for referencing
local data (possibly related to variable extraction?)
Objects with drivers are now treated as needing updates when the
current frame changes. This assumption has been documented in the
code, and should at least mean that users who try to use drivers for
creating simple time-based expressions that this should work.
Note:
- It is still recommended to create a "cfra" driver variable instead
of actually inlining bpy.context.scene.frame_current into the
expressions. Not only does the latter look rather nasty to type/have
in the expression, but it is also less future-proof for when I get
around to actually working on a beefed-up depsgraph (nothing official
on that front yet...)
* Removed frame-number display from NLA strips. Indeed doing so makes
things look cleaner/easier to identify.
* When transforming NLA strips, the "temp-metas" (purple strips) get
their frame extents drawn on either end, like in the sequencer, which
seems to be easier to read than the ones inside the strips.
---
The downside of this tweak is that there is no longer any visual
feedback for which strips run reversed instead of forwards, as that
used to be shown using the frame extents stuff.
The Keying Set paths list now shows the icon of the type of ID-block
that a path item refers to. This make it easier to make snese of the
paths shown in the list.
Animation Editor toggle tweaks:
* By popular request, curve visibility toggles in the Graph Editor are
now represented using the eyeball icons
* Muting is now represented by a speaker icon (a speaker for this
purpose seems fairly common?)
New icons:
* Keying Sets now have their own icons (as found in a proposal on
jendrzych's "Pixel Sized" blog)
* Drivers also have their own icon now. This is just a hacky one I've
devised which doesn't look that great. Suggestions on this are very
welcome.
Closed regions didn't always draw the (+) icon right place, confusing
for users.
Next to that, I think this icon is using a bad metaphor or visual language,
Illustrated best if you close a header in outliner or buttons. Icons are
UI widgets, for screen/editor layouts different controls can be stylized.
My preference is something that aligns visually to the seperators between
regions; for testing and hacking pleasure I've added two quick versions,
a small tabbish thing and a triangle. Enable these with debug menu,
ALT+CTRL+D, values 1 or 2.
This is simply drawn with opengl now. An image for it can be made as well.
Previews:
http://www.blender.org/bf/closed_regions1.pnghttp://www.blender.org/bf/closed_regions2.pnghttp://www.blender.org/bf/closed_regions3.png
There's other design ideas to explore as well, like making region deviders
8-10 pixels wide, with a 'drag me' dot on it or so. That takes some screen
estate though, and will require to add big editor-dividers too...
Fun stuff for the mockup-mafia to check on, we have time :)
Channels can now be used as "animation containers" to be filtered
further to obtain a set of subsidiary channels (i.e. F-Curves
associated with some summary channel).
The main use of this is that object and scene summary channels can now
be defined without defining the filtering logic in three different
places - once for channel filtering, once for drawing keyframes in
action editor, and once for editing these keyframes.
An indirect consequence of this, is that the "Only selected channels"
option in Timeline will now result in only the keyframes for a
selected bones getting shown (when enabled), instead of all keyframes
for the active object. This was requested by Lee during Durian, and is
something which has only become possible as a result of this commit.
* After changing driver target settings, the driver F-Curves now have
their "disabled" flags cleared, so that they will be updated
immediately instead of needing a manual "Update Dependencies" flush
* Little comment tweak to appease my text editor
render strands use the window matrix and window size which were both zero while baking, this caused divides by 0 and eternal malloc loop.
So set unit window matrix and dummy view size.
This is more a workaround then a fix but avoids crashing.
* F-Curves no longer show the name of the datablock of the property
they affect if this is an ID-block. For example, transform curves for
a cube won't get the "... (Cube)" suffix anymore. In these cases, it's
relatively clear that these belong to the parent datablock, so doing
this should be fine (and reduces clutter).
However, for non-id data (i.e. subsurf modifier settings) or bones,
this info still gets shown. In these cases, there is some ambiguity.
* "ActiveAct: <...>" is no longer shown for NLA action channels (i.e.
just the name of the action gets shown). This should make it easier to
see at a glance what action is being used.
== Datablock filters in the headers are now hidden by default ==
This has been done because users were generally not frequently
toggling these, so quick access vs screen-estate cost wasn't really
worth it to have these always showing and taking up space on the
header.
Usage notes:
- To show these again, click on the "Filter more..." toggle.
- The "Filter more..." button DOES NOT affect whether those filters
apply.
Design notes:
- I tried many other button/icon combinations, but those were either
too space-hogging, vague, or had wrong button order.
- I also tried putting a box around these, but there was too much
padding.
- The ordering of the filters has also been modified a bit so that the
group/fcurve-name filters occur earlier in the list, given that
they're used more frequently
== Graph Editor - Use Fancy Drawing ==
Renamed this option to "Use High Quality Drawing" as suggested by
Matt. "Fancy" isn't really descriptive enough.
== Icons for Mode Dropdowns ==
The mode dropdowns in the DopeSheet and Graph Editors now have icons.
- These were important enough (compared to the auto-snap mode) that
some visual decoration was perhaps warranted.
- It makes it easier to see at a glance what mode the view is in
Icon choices:
- In some cases, the icons seem like quite a natural fit IMO (i.e.
outliner<->dopesheet, key<->shapekey editor, grease pencil, fcurve
editor)
- Action Editor uses an "object" icon to indicate that this is object-
level only for now (though I hope to find a way to address this
soon/later). This will be kept like this until then.
- There isn't any icon for drivers, so after trying a few
alternatives, I settled on area-link icon, since it ties together two
entities using some link.
modified auto-completion, though this may need to become a preference.
The problem is:
- including _all_ text as a prefix can take a lot of space, and isnt too readable.
- including only the previous word is error prone because detecting delimiters can fail when editing strings.
so I've set it to only include the last part of the string but align to the cursor to make it more readable.
* Use the same method as from unlinking actions to do this.
* Split off the make single-user code used for the ID-browser into a
function in blenkernel which can be used elsewhere. Getting materials
to also work using this method proved to be a bit too tricky (due to
the whole messy ob vs obdata situation), so I haven't done that.
this bug was introduced in 2.58 (r37342), when adding filtering support to imagewrap(), the problem is boxsample was getting float values which were not wrapped as int values are.
It is now possible to use the Outliner for managing the active action
of an ID-block (provided that it appears in the Outliner), which
should be a bit better than having to go through the NLA Editor.
So far, this only allowing unlinking actions, using some existing
operators. To use:
1) Navigate through the Outliner tree to find the
Object/Material/Lamp/etc. that the animation belongs to. (NOTE: this
doesn't work in Datablocks mode, but should in the normal "All Scenes"
and related modes)
2) Expand the "Animation" entry below this
3) Right-click on the Action entry below this, and select "Unlink"
from the RMB menu
In the process, I've fixed problems with some data-blocks not showing
their animation data in Outliner.
* Mesh Animation-Data was not shown. Other data types would get this
shown.
* Added attempted fix for the problem where when you try to expand the
last item in a RNA list or so, you often end up expanding the first
item (and then have to close and try again, at which point the expand
works as you expected the first time round). More testing needed, but
seems to work better already
Fix for bug found by psy-fi.
* gpu_buffers.c was using GL_ARB_vertex_buffer_object to check for VBO
support, should be using GLEW_ARB_vertex_buffer_object.
error messages on start. These were coming from initializing the SDL video
subsystem as part of Joystick init. We do not need this, as video stuff is
covered by GHOST, most likely this was conflicting.
* Unlike blender, the game player draws only on windows update callbacks,
and those wer not implemented.
* Going fullscreen for player was not implemented correct, it expected an
existing window but actually it should create one.
* Removing the last of the owner/ownertype stuff. The bulk of this
stuff was removed in Part3 of the refactor, but it seems I forgot to
actually remove these struct members at the end of that.
* Texture datablocks without animdata aren't skipped immediately
anymore. This could lead to texture nodetrees on animdata-less
textures getting skipped.
On clicking in a non-active Blender window (when you activated others),
the mouse position of the first click was still the old position.
Problem is in GHOST; it sends out the 'activate window' event after the
mouseclick event itself. Code now checks for this case and reads the
correct mouse position.
Toggling options on the selection is better done as a generic operator.
Replace ARMATURE_OT_flags_set and POSE_OT_flags_set with WM_OT_context_collection_boolean_set and use menus to access it with specific settings.
This way its easy make a key shortcut which toggles any boolean on any collection - sequences, metaballs, objects, bones etc.
Fix for bug #27710, 'Multires lost from 2.49 file in 2.5x'
Reported by Gaia Clary.
Problem was that the old multires data didn't flush changes to
vertices out to the Multires structure on filesave. So, recent bits of
sculpting could be lost if the multires level wasn't changed before
filesave.
We already had code to deal with missing multires vertex data, which
simply copies the Mesh vertex data into the multires vertex data if it
matches the number of vertices in the highest level. Moved this code
up a bit so that we always make this copy if the numbers match up.
Was able to reproduce the bug fresh in 2.49b, and confirmed that the
fix works. However, this does not help if changes were sculpted on a
multires level other than the highest level and saved without a
subsequent level change.
- comment/remove assignments from values to themselves.
- add case break statements (no functional change but some source code checkers notice).
- fix python errors when the sculpt brush is None.
This commit is aimed at cleaning up the filtering code by changing the
filtering idiom/pattern used. While the old code used a "check then
do" approach, the new code does a "grab then assimilate".
The main benefits are that:
* the code duplication that used to exist has now been removed, making
it easier to add new channel types for data
* a recursive "peeking" ability now means that the old problems with
data existing deep in the tree (i.e. figuring out whether a channel
should be shown based on whether it will have any descendents) should
now work much better than before.
In the process, I've found and fixed a few previously unnoticed bugs
with how some channels were constructed, so hopefully things work a
bit better now.
TODO's:
* Action-Group filtering stuff hasn't been refactored yet. This was
causing some grief in the past, so I still need to check this
carefully.
* Material Nodes support (missing in trunk) should be easy to slot in
now :)
lambda_cp_line --> line_point_factor_v3
lambda_cp_line2 --> line_point_factor_v2
correction to previous commit function name
isect_seg_sphere_v3 --> isect_line_sphere_v3
... since its not clipped.
added a clip argument to the python version of the function.
Thumbnail save for .blend crashed, when being in editmode for a mesh
that has other object users as well. Derivedmesh confusement...
Thanks to Sergey for finding the cause!
Default startup theme used same node header color for "in/out" nodes
as for "inactive or undefined" nodes. This made it impossible to see
which of the output nodes in node setup was 'active'. Made the active
in/out color a slight blueish shade of grey now.
Python:
* adds bpy.app.handlers which contains lists, each for an event type:
render_pre, render_post, load_pre, load_post, save_pre, save_post
* each list item needs to be a callable object which takes 1 argument (the ID).
* callbacks are cleared on file load.
Example:
def MyFunc(scene): print("Callback:", data)
bpy.app.handlers.render_post.append(MyFunc)
C:
* This patch adds a generic C callback api which is currently only used by python.
* Unlike python callbacks these are not cleared on file load.
Usage currently is limited to:
- Panel text, widget text and label text style:
point size, shadow effects
Setting individual fonts to these is not possible yet, it uses the
default for it.
Access goes via outliner now; check "User Preferences". UI team could
add this in userpref scripts :)
This was causing a lot of backward compatibility problems and side effects.
It has to be done properly, linked with Material ObjectColor shader_flag or at least the TF OBCOLOR (which apparently worked at some point so I've been told).
If you need to use OBCOLOR simply keyframe the obcolor and it will work as before.
* BL_DeformableGameObject is no longer responsible for handling keys, BL_ShapeDeformer is
* BL_ShapeDeformer also creates a copy of the key on construction and puts it back on the mesh when destructed. This avoids us permanently modifying Blender data.
* I'm not too fond of clearing out the key every frame, but this works and I can't think of another alternative at the moment (something may be possible with some key juggling)
In addition the billboards can be scaled by the particle velocity with optional head and tail factors (similar to line drawing options). This allows for pseudo-motionblur effects.
The rna set function clamps to the property range however properties with range functions were ignored when set by python or the animation system.
Now call the range function for ints and floats when setting.
* Insert Key on Selected Channels in Action Editor was broken
* Transform/Select All tools in Action Editor were broken as result of
filtering changes.
* Set Visibility operator, when used from Graph Editor now does
similar things to the TabKey lock/unlock operator with regards to the
flags it uses for filtering
the list of animated curves is closed
At long last, this old bludger can be put out to pasture. I figured it
would involve some of the visibility-filtering stuff I added, but this
required a bit extra effort than anticipated.
* This (big) commit is aimed at cleaning up the filtering flags used
by the animation channel filtering code. The list of filtering flags
has been growing a bit "organically" since it's humble origins for use
in the Action Editor some 3 years (IIRC) ago now during a weekend
hackathon. Obviously, some things have ended up tacked on, while
others have been the product of other flag options. Nevertheless, it
was time for a bit of a spring clean!
* Most notably, one area where the system outgrown its original design
for the Action Editor was in terms of the "visibility" filtering flag
it was using. While in the Action Editor the concept of what channels
to include was strictly dictated by whether the channel hierarchy
showed it, in the Graph Editor this is not always the case. In other
words, there was a difference between the data the channels
represented being visible and the channels for that data being visible
in the hierarchy.
Long story short: this lead to bug report [#27076] (and many like it),
where if you selected an F-Curve, then collapsed the Group it was in,
then even after selecting another F-Curve in another Group, the
original F-Curve's properties would still be shown in the Properties
Region. The good news is that this commit fixes this issue right away!
* More good news will follow, as I start checking on the flag usage of
other tools, but I'm committing this first so that we have a stable
reference (of code similar to the old buggy stuff) on which we can
fall back to later to find bugs (should they pop up).
Anyways, back to the trenches!
- Sequencer dynamics: Now it's possible to change the output channels and the resampling quality also increased (previously maximum quality was 44,1 kHz)
- Changed two buffers to use ffmpeg allocation, not sure if that helps somehow.
- Implemented a nice rechanneling solution with unofficial speaker arrangement standards similar to what OpenAL soft has
- Renamend AUD_Channel in the C API to AUD_Handle
- Removed the unlogical 7.2 speaker configuration, that's a hardware only config
for default DPI. Problem now is that icons/text are scaling
and drawing with pixel units, whilst other items draw subpixel.
This makes not every dpi size result in perfect aligning yet.
* Hair combing now uses substeps to apply the combing when the mouse movement exceeds 0.2 of the brush radius.
* This could make combing a bit slower on fast mouse movements, but the increase in combing quality is definitely worth it.
* Changed all int's to size_t's, where the int's were used for size of
channel list returned
* Object vs Base is now passed to filtering functions - was relic from
old owner/ownertype code which required access to bases
* Found bug in NLA code where filter was being overwritten and then
used again as input for some other function unintentionally
* Found bug where trying to select a NLA strip would crash if lamp
data was around
* Removed list-expanders for Materials, Textures, and Particles. So
instead of:
Object
Materials
Material 1
... material 1 anim data ...
we now have
Object
Material 1
... material 1 anim data ...
This makes it faster+easier to get to these items. If you don't want
to see all of these, you can still use the data-block filters from the
header to hide these.
* Internal cleanup - removed "owner" and "ownertype" settings from
bAnimListElem. The purpose of these was muddled, and more of a hassle
to maintain than doing anything useful - it was only really used for
the stuff above.
* Removed need for "sa->spacedata.first" casts all over the show for
animation editor tools which needed access to editor data. This can
now be retrieved directly.
Render + compositing error:
When adding renderlayer nodes in a composite, without having own
scene render, the renderlayer nodes were not tagged as changed,
causing compositing to give previous result.
reported by Rebin Cornelius.
This needs patch from upstream report http://code.google.com/p/opencollada/issues/list?thanks=164 applied to OpenCOLLADA. lib/windows/collada and lib/win64/collada have already been updated. Ensures we don't read past array boundaries.
regarding issues in the report
1) fixed in previous commit.
2) intentional leaving as is.
3) left 'handle size' as is IMHO acceptably general, edited 'hotshop' tooltip/
4) corrected default values.
- enable manipulator option, applies to new 3d views (not sure why it was disabled) - reported in [#27664]
- remove rigify and netrender scene ID properties
- set console scrollback to 256, was 128 which could sometimes cut off output of help()
- enabled syntax highlighting and line number in the text editor for game logic and python screens
- add support for IDProp array slicing, but not resizing.
- rename array attribute type to typecode and use chars 'f', 'd', 'i' which match pythons array module. (was using int's which only have a meaning internally).
- rename function 'convert_to_pyobject' to 'to_dict' and 'to_list' for IDProp group and array types respectively.
- remove 'len' array attribute, calling len(array) is fine.
This commit adds an experimental effect strip to the sequencer: "Title
Card".
This is useful for adding simple one-line text section headers or
"title cards" (i.e. title + author/contact details) to video clips,
which is often seen in demo videos accompanying papers and/or
animation tests.
See http://aligorith.blogspot.com/2011/06/gsoc11-simple-title-
cards.html for some more details on usage.
Code notes:
- There are a few things I've done here which will probably need
cleaning up. For instance, the hacks to get threadsafe fonts for
rendering, and also the way I've tried to piggyback the backdrop
drawing on top of the Solid Colour strips (this method was used to
keep changes here minimal, but is quite fragile if things change).
This patch adds anisotropic filtering of textures in the viewport and the BGE. The quality of the filtering is adjustable in the user preferences under System. For more information on anisotropic filtering:
http://en.wikipedia.org/wiki/Anisotropic_filtering
One current limitation of this setup (having the option a user preference) is it makes runtimes more troublesome. Runtimes don't have user preferences set, so for now the blender player defaults to 2x AF. Options will be added later to change this value (probably a command line option).
Vertex parents were not requesting the original index layer, now do this as
part of depsgraph building, and make constraints with vertex groups use the
same system. Fix is based on patch by Campbell, but with some changes.
- fix python error when the object had an empty material slot
- initialize start frame from the current frame
- set frame range to 300000 max (which is blenders own maximum)
- mesh order was arbitrary, instead use selected -> active, removed invert option.
also fix for missing include in bpy_extras.view3d_utils
Sculpt and Paint undo steps kept hanging in the sculpt/paint modes.
Now undo will switch back to global undo, and redo enter the mode
again. Just like weight/vertex paint.
##########
original name: "Allow to change the strenght of the "go behind" constraint of the camera actuator"
The camera actuator is an actuator that drive the camera to follow an object, with a set of constraint.
Currently, when the object followed rotate on himself (like a person, or an helicopter), the camera is really slow to go behind (at least 10 seconds).
This patch gives the UI to tweak the strenght of the 'go behind'[named damping] constraint.
###########
epydocs (rst) updated too
Reported by Pelle Johnsen
Fix falloff import. Point light and Spot light always were set to inverse quad, instead of choosing the proper one based on imported values.
The
By popular demand, the "Transformation Channel" driver variable type
now has a "local space" transform space option which uses the same
magic that constraints use for defining local-space. This is what many
bug reporters and feature requesters have moaned about for a while
now, so after reviewing several of the bug reports which lead to the
current situation, here is what has been much-wanted for so long!
In order to implement this, I've:
- renamed the old "Local Space" option here to "Transformation Space",
in order to prevent old rigs breaking. This has also been kept, as it
is useful for #21384 (though perhaps with this new option it isn't
needed anymore)
- reviewed my fix for #20870 (IIRC, a Durian-era bug), which related
to the non-uniqueness of matrix->euler decomposition
The problem was that SCA_2DFilterActuator was defining and using a
variable called "m_gameObj", when it should be using "m_gameobj" as
defined by SCA_ILogicBrick. The way it was, reparenting did only half
the work required to duplicate the actuator.
Patch by Alex Fraser, thanks!
a) Enable the possibility to remove the "air bubble" around submerged collision object. This feature is enabled as standard for new files. The code was found in elbeem by nudelZ, coded and provided by Nils Thürey (thanks!)
b) Old baked files gets deleted if a new bake gets started (were overwritten before and resulted in weird old bake + new bake mixture) (idea by nudelZ)
Full Sample AA (FSA) was failing in cases. Bug report was an empty
scene (with compo nodes) linking in another .blend scene (with render).
That case gave warning "FSA not supported with rendering". That now is
allowed.
Then I noticed FSA was giving corrupt sample buffers or crashes in cases,
especially on first buffer, this appeared to be a missing compo tag on
first sample buffer.
Lastly, to make FSA render a tiny bit less frustrating: added render window
statistic to show which of the FSA steps is being done.
* Sets up a particle system and an explode modifier.
* In "blend" mode requires two selected objects and creates "crossed keyed" particle systems between the objects for a simple blend effect where the first object explodes and then recombines as the second object.
* Also renamed the other quick effect operators as "Quick ..." to bring some consistency to the operators (also nice that now you can just write "quick" to the operator search and get all these operators).
The text datablock was linked to the controller. So even if the script was set to 'module' and saved, once linked/appended the object the script would come together.
If someone wants to implement this "clear" only once the file is saved, please go ahead. But I believe it's ok to loose the script if you change it for module (and with the new datablock lookup it's straightforward/quick to reassign a textblock)
-- bug not reported anywhere, from my own list
* Adding BL_Action::Play() and BL_Action::Stop()
* Making BL_ActonManger reuse BL_Actions instead of recreating them all the time
* Making the Property play type work for the Action actuator.
Ending localview - which is similar to layer change - should send
update similar to change layers too. Needed for example when loading
a file saved in local view or when changing time in localview.
I was going to include this change along with support for moving
FModifiers around on the stack, though that looks like it might be a
bit more involved than first though. To be dealt with later...
"entire array" property on by default, making it easier to add
transforms to Keying Sets.
This doesn't affect Keying Set paths added via Python or any other
means.
Displacement mapping didn't do linear interpolation between pixels,
causing render artefacts. Now it uses for image render without AA
a default interploation filter of 1 pixel size.
Fix provided by Miika Hamalainen. Thanks!
ED_view3d_init_mats_rv3d was calling glMultMatrixf() which was mostly harmless but could also lead to confusing bugs (2 reported previously).
Looked into this and every call to ED_view3d_init_mats_rv3d except for object drawing, doesn't need this so made a second version of ED_view3d_init_mats_rv3d - ED_view3d_init_mats_rv3d_gl which does the matrix multiplication, remove confusing checks in selection code.
- was using un-initialized stack memory if the source / target object had no vertex group.
- if the target object had no vertex groups it would fails silently (not a bug but not very good functionality)
- added an error message if any copying fails.
there is a minor problem with this commit:
0.00001 --> 0.00001 # good
0.000015 --> 0.000015 # good
0.0000199 --> 0.00002 # ok
0.00002 --> 0.000020 # wrong, has trailing 0
Tried to fix this but the case is hard to check for without more calculations which Id like to avoid.
This fix also allows for partial update of the image, speeding up painting.
The different code path implemented will be used to upload high resolution images to OpenGL when onion branch is merged.
Due to conversion of float textures to/from sRGB, corrections made to brush color sampling to take account of the image profile. This is not 100% correct yet as texture images used for projection painting strokes are not converted to/from sRGB yet(This has been decided due to loss of precision for 8-bit formats). It will have to do for now, though.
last-minute update, exr image loading is broken, will fix asap
This means script authors can perform actions using these callbacks rather then on drawing which puts blender in a readonly state.
Simple example:
import bpy
def up_func(self, context):
print("test")
bpy.types.Scene.testprop = bpy.props.FloatProperty(update=up_func)
bpy.context.scene.testprop = 11
# prints -> test
operator names instead of the internal names. This wasn't really noticeable,
expect that it broke automatically looking up shortcuts for display in menus.
added 'INTERNAL' operator flag so operators which are only meant to be called by other operators or internal use are not displayed to the user.
Currently only use this flag for the operator search toolbox, is ignored in debug mode.
=======================
Added option to baked named "Bake From Multires" which is avaliable for
normals baking and displacement baking.
If this option is enabled, then no additional hi-res meshes and render
structures would be created . This saves plenty of memory and meshes
with millions of faces could be successfully baked in few minutes.
Baking happens from highest level against viewport subdivision level,
so workflow is following:
- Set viewport level to level at which texture would be applied
during final rendering.
- Choose Displacement/Normals baking.
- Enable "Bake From Multires" option.
- You're ready to bake.
Displacement baker had aditional option named "Low Resolution Mesh".
This option is used to set if you want texture for realtime (games)
usage.
Internally it does the following:
- If it's disabled, displacement is calculated from subdivided
viewport level, so texture looks "smooth" (it's how default
baked works).
- If it's enabled, dispalcement is calculated against unsubdivided
viewport levels. This leads to "scales". This isn;t useful for
offline renders much, but very useful for creating game textures.
Special thanks to Morten Mikkelsen (aka sparky) for all mathematics
and other work he've done fr this patch!
- Bone animations are also exported as f-curve animations now. As a result euler rotaions of bones are also exported. All animations with BEZIER, LINEAR or STEP ipo are exported.
- Quaternion rotations export.
- Object parented with armatures, animations to Armature Objects as a whole are also exported.
flag on ED_area_initialize(). This however was causing 2 problems;
- the view state got reset (popping window view back)
- the view2d operator polls failed (sliders didnt work)
This re-init was only needed for the headers though, limiting it
to these types of regions solves it.
- fix: user pref, window title was reset to 'Blender' on tab usage
- Undo history menu back:
- name "Undo History"
- hotkey alt+ctrl+z (alt+apple+z for mac)
- works like 2.4x, only for global undo, editmode and particle edit.
- Menu scroll
- for small windows or screens, popup menus now allow to display
all items, using internal scrolling
- works with a timer, scrolling 10 items per second when mouse
is over the top or bottom arrow
- if menu is too big to display, it now draws to top or bottom,
based on largest available space.
- also works for hotkey driven pop up menus.
- User pref "DPI" follows widget/layout size
- widgets & headers now become bigger and smaller, to match
'dpi' font sizes. Works well to match UI to monitor size.
- note that icons can get fuzzy, we need better mipmaps for it
Following on from my commit to introduce frame ranges for FModifiers,
those frame ranges can now have blend in/out values. By setting a
blendin or blendout value, you're specifying the number of frames for
the modifier's "full influence" to take effect or fade out relative to
the start/end frames.
The "full influence" above needs a little clarification.
When the "use influence" setting is enabled, "full influence" is taken
from the "influence" slider (a new setting). Otherwise, it uses 1.0
(i.e. unmodified influence, same as old behaviour before the
introduction of influence controls). The influence slider basically
says how much the modifier's effects are allowed to contribute to the
final result.
---
Notes:
- This opt-in "Use Influence" approach is really forced upon us
because there are heaps of old files for which we cannot easily
version patch without spending some effort going through all the data
in the file, hunting out the F-Modifiers.
- interpf() seems to use a backwards order compared to everything else
Due to overwhelming support from animators, Actions are no longer
created with fake users by default. If you're mainly creating action
libraries (the primary use case and argument for having this, mostly
used for creating a set of motions for games or perhaps to use in
NLA), you're really in the minority here.
For the most part, fake users just lead to heaps of "dangling" actions
in files which newbies (and even experienced users) may often be
unaware of. Since Fake Users are really more of an "opt-in" system
everywhere else (i.e. when creating Material Libraries), the same
should applied for Actions and creating Action Libraries.
file
This is just a more formalised version of a local hack I've been
running locally for the past year now. It's especially useful when you
want to maintain your own set of recently opened test files (or
perhaps current project files), but then be able to quickly open some
.blend files downloaded from the web (i.e. checking out some bug
report, or how someone else sets up some node setup) without
loosing/polluting your existing recent files list as a result of doing
so, and having to either resort to some nasty methods to get it back.
Of course, this is still really hacky, as for instance, it means that
the currently opened file will not show up in the recent files list
for quick reload. However, that's why this is a userpref :)
Using this feature, it is now possible to for example have different
noise-profiles for different parts of a curve, which makes it possible
to do animate camera shake for example.
Or perhaps, for having greater control of mixing and matching
different parts of F-Modifier effects, such as combining several
generator modifiers to get multi-case functions for instance.
See http://aligorith.blogspot.com/2011/06/gsoc11-fmodifier-range-
masks.html for details.
Now I think we are all good. We still have a few actuators that were using TOGN before but that I didn't make as negative_boolean.
All fixed now:
- parent actuator
- edit object actuator
- action actuator
- shape actuator
* they were too slow for dynamic python enums, calling the callback
to list the items for each button, to get a tooltip
* enum tooltips sometimes were showing the same description twice
.blend file.
* From operator search, find "Clear Useless Actions"
* "Action library" actions are preserved by this operator. It targets
actions without any F-Curves
* By default, only actions which are single-user (where that user is a
Fake user) will be targeted. This can be changed to have it target any
"dangling" action that doesn't have any F-Curves
* A save/reload cycle is still required to fully remove such Actions
from the current file. Though at least now it's a simpler process to
have these semi-automatically removed ;)
- moved do_history into WM_write_file after successful write of .blend@ temporary file
- Added new file flag, to avoid writing history on writing the startup.blend, autosave files and undo.
Thanks Campbell, Brecht for review!
There won't be dangling "empty" actions left behind anymore in the
DopeSheet channel list after you've deleted all their keyframes (and
don't want to add keyframes to them anymore).
Of course, this poses problems with more actions getting created if
you then go and keyframe those objects again. If this does turn out to
be an equally bad problem, then another approach from the channel
filtering code side (probably aided by the restructed code) will help
(though doesn't solve the problem where people complain of having
heaps of "empty" actions dangling from objects they no longer want
animated).
Dunno how long this has been broken for (*), but the Time Slide
transform tool in DopeSheet no longer did anything most of the time.
It appeared to be be caused by some blotched indexing code from ages
ago. I've fixed this problem, as well as preventing the case where it
would also give errors when only a single key was selected.
(*) Does anyone actually use this tool? IIRC, this was added during
Orange, though I can't find the commit for this anymore or why it was
added. Probably it might be better to just let it go...
Face's totdisp was set to correct value, but memory hasn't been
allocated for disps. Handle this in multires_topology_changed(),
so the whole MDISPS layer wouldn't be totally re-allocated when
applying displacement.
when installing to a folder that already existed. Now the uninstaller
will remove only the files it has installed, and leave any other files
in the installation directory intact.
Adding/Rename markers (M/Ctrl-M) were restricted to only being
available when the mouse was hovering just over the time scroller at
the bottom of animation editors, as otherwise we'd get nasty keymap
conflicts where markers keymap would block all the primary function
keymaps.
However, in the case of Adding/Renaming markers, there are no other
keys which currently conflict with these in such cases. Hence, it is
fine to let these ones be able to be run from anywhere within the
animation editors, which should make it easier to add markers for
lipsyncing purposes again for example.
* Converted BL_ActionActuator::Update() to use the new action api (still just armatures)
* Not all of the functionality of the old Update() have been ported (Lood end, continued animation, etc)
* Things are still pretty messy. Once have things more flushed out, I'll start stripping more of the old actuator out.
BL_ActionManager:
* IsActionDone(short layer) - Checks to see if the animation on the given layer has finished.
KX_GameObject:
* PlayAction(...) - Adds an action to the object's action manager
* StopAction(short layer) - Remove an action from the object's action manager
* IsActionDone(short layer) - Check if an action has finished playing
blender_add_lib now takes a separate include argument to suppress warnings in system includes (mostly ffmpeg & python).
also only build wm_apple.c on apple+carbon configuration.
It was lag of optimization logic, which always retuns object's derivedMesh
if one of of boolean operation's meshes has got no faces.
Actually, result depends on operation and which mesh has got no faces.
Added small utility function to handle this.
Do not indent if there's any non-space character after colon.
This only makes life a bit easier, but it's still not 100% correct indentation
strategy. For example when colon is inside non-closed string or so.
Also there's not indentation for { and un-indentation for }.
Handling such cases would require much smarter strategy..
This was caused because a small part of sculpt's radial control code did not make it into the new version. The old code would set a new object-space size by scaling it proportional to how much the new screen-space size was changed.
The solution I implement here is to do the same scaling inside the RNA callbacks. This way, users of those properties do not have to worry about inconsistency.
I added a comment warning that brush_set_size, brush_set_unified_size, brush_unprojected_radius, and brush_set_unprojected_radius do not guarantee consistency because it is not always possible to precisely know what the new unprojected radius is in all contexts where you might set the size.
I would implement the consistency check at the lower level (in those listed functions) but at this time I think it needs to be looked at to make sure that won't cause problems. In addition, I am not sure that scaling by the ratio of change is strictly correct in all cases.
In any case, this at least fixes the immediate problem.
[the problem also affected OSX]
PhysicsId are Long, not ints (see PyObject* KX_GameObject::PyGetPhysicsId() )
There is a reference in the code to use PyCapsule instead of int. I'm not sure
about that. This patch at least stops the crashes
(update: I talked with Campbell and he repeated that PyCapsule are better, but if long is working it's fine for now).
Revert of "SVN commit: /data/svn/bf-blender [36957]
trunk/blender/source/gameengine/ VideoTexture/VideoFFmpeg.cpp: fix for
ffmpeg linking in BGE ( patch by Jens Verwiebe (jensverwiebe) over IRC)"
Sorry folks, that patch breaks current ffmpeg GIT version.
Good news: it's all handled now automagically by ffmpeg_compat.h in
intern/ffmpeg
so: everything should be fine and dandy for very old and very new versions.
Added central compatibility header file, which enables blender to compile
against very old ffmpeg versions as well as very new versions using the
*NEW* API. (Old API functions are simulated using macros and inline functions)
Added a whole lot of additional checks, tested against 6 different versions
down the timeline, hopefully, now finally all is well.
fixed frame colors for letterbox drawing (happens when in camera view).
Cause is unclear, seems some sort of strange graphics driver thing on 32 bit.
Changes are a fix for the incorrect usage of glViewport, and avoiding the extra
clear if it's not needed.
* Windows installer not working for non-admin users and multiple users
* Addon scripts not installing next to user configuration
* Portable install not being taken into account in all places
The main problem was the windows installer was installing system scripts in
AppData next to the user configuration directory, which is not shared between
users. Now these are installed in ProgramFiles, and only addon scripts added
by the users go to AppData.
On all platforms, addon scripts were sometimes getting installed between
system scripts, because the scripts folder in the executable directory was
given precedence over the user configuration folder, that is no longer done
now. So addons now behave like user configuration, they are preserved even
if you download a newer build of the same blender version.
If you have an installation of 2.57 on windows, the addon install location
will not change until we do the version bump to 2.58, to avoid conflicts with
the existing the installed 2.57 version.
The old behavior of giving precedence to the local folder was done to support
portable install, where all configuration is written to the local folder. This
is now implemented differently: if and only if a "config" folder exists in the
local folder, portable install will be assumed, and files will only be written
to that local folder.
Added some API compatibility code again, since some API-changes weren't even documented
(they even didn't do a proper version-bump, arghh!)
If it breaks again, please tell!
* removed a lot of old cruft code for ancient ffmpeg versions
* made it compile again against latest ffmpeg / libav GIT
(also shouldn't break distro ffmpegs, since those API changes
have been introduced over a year ago. If it nevertheless breaks,
please send me an email)
This is equivalent to using B-Bones which are all scaled to have
xwidth=zwidth=0, which can be useful to see how some limbs will bend,
without the overhead of blocks blocking the view or having to scale
down bone sizes first.
In the case of this bug e.g. material.new became MATERiAL_OT_new, due to
different capitalization of "i" in Turkish. Fixed by not using the locale
dependent toupper/tolower functions.
When trimmed the seq->len was being reduced from the offsets (initial and final). This was the length passed to the elements.
This had two problems:
1) it would not allow you to change the elements not visible (although you likely want to change them as well).
2) the element[0] was always the seq->strips[0].stripdata[0], regardless of the initial trim.
Anyhoo it's all working now.
Thanks Campbell for helping out with this one.
this is more commonly used that the TimeSlide tool (which is DopeSheet
only).
Noticed that this does bring this out of line with the hotkey for
setting extrapolation, but then again, extrapolation is a per-curve
setting.
few cases, especially if the object and target are at the same
location when the constraint is created.
This manisfested as the constrained object disappearing when the
constraint was added, only reappearing after transforming it a bit.
The Limit Distance Constraint now has a "For Transform" option just
like all the other Limit constraints. This option controls whether the
constraint gets applied to interactive transforms in the 3D View too,
preventing controllers from getting large values without the animator
knowing.
Additional code changes:
* Split code to get constraint targets and grab their matrices for
solving out to a separate helper function:
get_constraint_targets_for_solving()
* Fixed a bug where "found constraint ...." prints would appear in the
console. Looks like some warning print that was forgotten
TODO:
* While coding this, I noticed potential division by zero bugs with
the Limit Distance constraint. Looking into these after this commit.
Added operator to convert animation for standard object transforms
(i.e. loc/rot/scale) to delta transforms.
This can be accessed from the Object -> Transform -> Animated
Transforms To Deltas menu entry in the 3D View.
Since the situation which causes this is quite common (especially for
motion-graphics type applications), where users animate some object
first and then decide to duplicate this and place it around the place
in different locations, it's probably important that we have some
support for this kind of thing. Newbies with the "help, all my anmated
duplicates disappear" problem are recommended to use this operator
from hereon in.
For reference of rationale, see:
http://blenderartists.org/forum/showthread.php?219126-Move-Existing-f
-Curve-to-delta-equivalent
"On Selected Only" option is now on by default. Stored poses only get
pasted on to bones that are selected when pasting, instead of on the
bones that were selected when copying.
(First GSoC11 commit. Yay!)
* Adding a BL_Action and a BL_ActionManager
* Each KX_GameObject has a BL_ActionManager, which can control up to for BL_Action objects at a given time
* Currently, the only interface to BL_ActionManager is through KX_GameObject via Python
* Only armature animations are currently supported
when loading them. Custom cursor shown during load was not freed correctly
when closing the old window, which resulted in unpredictable crashes later on.
- don't allow building if the LIBDIR is not found on mac/windows.
- by default use -O2 rather then -O3 for GCC release flags, was crashing some GCC versions and blender releases are supposed to use -O2.
for durian we had camera rigs which needed to have the parent transformed rather then the camera, for this reason I made fly mode fly the parent rather then the camera its self.
Make this a preference and use this for view camera/view locking too.
previous fix unintentionally made linking groups also add objects to the scene,
tested this more throughly, its ensured only to run on append and not to conflict with group linking/appending.
Linking in groups also ended up adding those objects to the current
scene that we're linking the group in to. This is a regression. It
leads to more work (and/or confusion) from animators when just trying
to get a character into a shot file to start animating it (see my
notes on my blog for the proper workflow regarding this).
Reverting r36762, which caused this mess.
Committing here a patch by Bastien Montagne (mont29), a more understandable Translation Constraint UI.
Before: http://www.pasteall.org/pic/12578
Now http://www.pasteall.org/pic/12258
From the description:
"When you set “X” under the Destination’s “Z”, it does not mean that the Z transform of the source should affect the X transform of the destination, but rather that the X transform of the source should affect the Z transform of the destination…"
The new UI should make it a bit more clear.
*In Particle Mode, there was no way to see what p-sys is being edited in the 3D View Tool bar, when having multiple ones.
Changed List type to normal (with limit to 3 rows). Request by venomgfx. :)
* Removed redundant text info about disconnected hair from Particle UI.
some functions had vague names, I even ended up re-writing some of these functions by accident!
also added doxy comments.
* ED_view3d_win_to_3d (was window_to_3d)
* ED_view3d_win_to_delta (was window_to_3d_delta)
* ED_view3d_win_to_vector (was window_to_3d_vector / viewvector)
* ED_view3d_win_to_segment_clip (was viewline)
* ED_view3d_win_to_ray (was viewray)
- when saving blend file with 'Remap Relative' enabled, don't try make paths absolute if the internal filename is invalid.
- use case insensitive path comparison on windows when checking if path remapping is needed & for comparing next/prev dirs in the file selector.
- rename 'name', 'dir' --> 'filepath' where these actually represent a file path to avoid confusion.
- bugfix for possible (but unlikely) uninitialized string.
- remove commented script append function, now we have a python api for this.
This fixes one part of [#27353] VSE crashes on large M4V
StripData was alloced in full length for MOVIE and SOUND-tracks, which only
use the first element for filename storage. (StripData as an array is only
used in IMAGE strips).
Fixed the crash and documented accordingly.
This patch adds adjustment layer tracks to the sequencer and does some cleaning
up of the code.
What's an adjustment layer?
Think of it as an effect track, which takes no explicit input, but alters
the output of everything down the layer stack.
So: you can add several stages of color correction with it.
And: you can even use it with metastrips to group several adjustments together.
Reviewed by Tom Musgrove and myself.
From the patch description:
ValterVB on #blendercoders submitted a long list of missing tooltips in Blender, and I went through the list and added all I knew. After that I crowdsourced the rest by putting a spreadsheet on Google docs and having people fill in the missing ones that I didn't know. So if there's some weird tooltip in there that doesn't make sense, that's why.
Thanks to Wolter, spacetug and others on BlenderArtists for contributing tooltips.
and there are still another multires modifiers in the stack.
Helps to prevent loosing sculpt data when you occasionally added another
multires and reomved it with "X" button.
This fixes frame colors not showing up right when using letterbox in the embedded player. Frames are drawn by clearing the whole canvas and then changing the viewport to be within the frames. The problem is that the embedded player's canvas is setup to be within the frames. This means that the extra that would normally be cleared and filled with the frame color is instead the gray color of Blender's region since nothing is actually drawn there by the BGE. To solve this, I just handle the frames in BL_KetsjiEmbedStart.
from the current shape mix. The old behavior is still accessable from the
menu as "New Shape From Mix".
Checked with Sergey and Bassam that this is a good change. New users
expect the add shape button to simply add a new blank shape, and get
confused when that is not the case. It is also really easy to
accidentally have other shape information in a new shape when the
"from mix" behavior is default.
- follow rotate/pan/zoom/dolly operators.
- auto-depth preference works.
- smooth view navigation supported.
- view selected, all & numpad operator work too.
TODO
- deal with camera transform locked axis
- find a way to move/zoom the frame while the camera is locked (if it turns out to be a problem).
Smooth brushes at requires face map which haven't used to be created for
simple meshes (without modifiers).
Use corrected need_fmap flag at stroke initialization, so face map
is created properly for "temporary" smooth brushes and simple meshes.
The main purpose for this is to allow rendering motion blurred blender fluids in external renderers (eg. http://vimeo.com/21870635 ).
Python code snippet for interpreting this data here: http://www.pasteall.org/21577 . Cleaned up some ugly hacks in this area too
* Also added read-only access to scene.subframe to RNA - setting current frame and subframe should still go via scene.frame_set()
Actual problem was caused by insufficient buffer size
in ui_text_leftclip()
Also fixed possible invalid memory write in GHOST_SystemWin32::getClipboard
which was caused by accessing clipboard buffer after closing
clipboard. This mustn't happen.
Also fixed possible crush when buffer was failed to be locked.
mouse coords would with cont. grab would wrap at short.
use mouse coords as int rather then short.
this problem still happens on linux because of XTranslateCoordinates
Patch to make the radial control more generic with RNA. Patch was
reviewed here: http://codereview.appspot.com/4280080/
Prior to this update, the radial control code in trunk had generic
parts of the radial control implemented as an incomplete operator
within WM. Then each different user of the radial control had to
implement a separate operator to actually pass in specific brush data
-- e.g. sculpt's brush size, vpaint's brush size, etc.
This patch removes all the extra operators and makes the WM operator
do everything. It now takes several RNA path strings as its properties
-- the only required property is data_path, which specifies the data
to be modified by the radial control. The other paths affect display
in various ways, e.g. rotation, color, etc.
In addition to decreasing some duplicate paint brush code, these
updates make it pretty easy to enable radial control for other
purposes (and it can be set up entirely though python or keymaps, no
extra C code needed.)
Re-calculate handles after applying transform matrix on curve object
This commit is'n fixing changing of curve shape whe nyou're deforming it (as
it was intially reported) but just adds needed handles recalculation.
Handles are calculated correct in that "unneded deformed" curve.
This commit fixes boundbox for empty curve-typed objects. Because of strange
reason (-1,-1,-1) and (1,1,1) was used as boundbox. Now it uses zero-sized
boundbox (as it's done for meshes).
This commit makes nothig with that single space character text from the report.
from Andy Braham (andybraham)
This adds support for empties to reference images and draw in the 3D view.
Modifications from the original patch.
- use an empty draw 'image' type
- use image aspect ratio for non-square-pixels
- when the image is not found, still draw the frame.
some WMs not only send 'crossing' events when really moving from one
window to another, but also when mousewheeling.
distinguishing those events 'mode' property fixed this.
brecht++ for figuring out the details.
- Recreate mipmaps if they're dirty when drawing background image
(this prevents "delayed" update of background image when you're paiting on it
in image editor).
- Mark mipmaps as dirty when inverting image channels.
TODO: there's memory leak caused by IMB_remakemipmap, but it'll be
anuther bugfix commit after discussion with other devs.
- In Drivers mode of Graph Editor, expanders for sub-object data were
not working at all. This was because they were getting overriden by an
errant ob-level drivers check
- Adding drivers from py-api didn't update Graph Editor
The following script would fail:
#ob = bpy.context.active_object
pb = bpy.context.active_pose_bone
pb.bone.driver_add("hide") # <--- exception here
The RNA-path function for Bone assumed that when it got called, it's
"id_data" (or owner-idblock-pointer) would only be ID_AR (i.e. an
armature). However, in the above example, pb.bone has ob as its
id_data, resulting in an invalid RNA path getting created. Added check
for this case, since it's likely to be common
The tracker description:
"This patch fixes this bug:
http://projects.blender.org/tracker/index.php?func=detail&aid=25487&group_id=9&atid=306
This was accomplished by making sure the KX_GameObject's object color gets set during the conversion process in the method gameobject_from_blenderobject. Otherwise all the values for object color default to 0. The reason adding an IPO worked as a work around is because it set the object color."
The bug report mentioned is bug [#25487] "BGE: Object Color only works when it has a keyed frame"
(patch co-reviewed by Nathan Letwory)
Overview: GHOST using rawinput for keyboard input. GHOST window receives WM_INPUT only when it is the active window. Child window cannot be active, so when embedding blenderplayer, WM_INPUT is consumed by top level parent window (for Burster it is the web browser window). Patch register raw input device as 'inputsink' - it makes GHOST window receives all keyboard messages. Window procedure check if GHOST window is active or focused.
::TranslateMessage(&msg) generates WM_CHAR etc. messages from WM_KEYDOWN, WM_KEYUP etc. Because of using RawInput only WM_INPUT messages are processed, so we doesn't need WM_CHAR, WM_KEYDOWN etc. [this is why ::TranslateMessage is no longer getting called].
Note: It's responsibility of the parent window (aka the wrapper) to send WM_SETFOCUS to child window (embedded blenderplayer).
However some parent windows (e.g. webbrowsers) will not send WM_SETFOCUS to the child window when someone clicks on it.
In those cases the blenderplayer needs to be patched to call setFocus(&msg); in the event of WM_LBUTTONDOWN (see GHOST_SystemWin32.cpp)
Submitted by Amorilia.
DXT1 .dds textures with 1-bit alpha channel have their alpha channel imported in Blender.
The patch also makes change to contact info for the patch submitter.
I left the sync with upstream nvtt for another commit.
Added new option to find panel of space text which toggles
case-esensitive search.
Additional changes:
- Send NC_TEXT|NA_EDITED when removing markers in find_and_replace modifier
this prevents "sticked" markers which disappears on first redraw when
search text wasn't found
- Do not show "Text wasn't found" error when text to be searched is contained
in the end of buffer and it's selected. Replacing/marking used to happen, but
this popup message was really annoying for this case.
TODO: It's incorrect to use UI_GetThemeColor4ubv from this operator
this isn't a fix but GetModuleFileName is returning a name that doesn't exist, so finding python after this isn't working.
Show a popup that the path has an invalid name so at least the user can move it.
- Constructive modifiers are enabled by default in sculpt mode.
- There's option to disable all constructive modifiers in the "Options"
panel of toolbox in sculpt mode,
- Use one column in options panel to make strings easier to read
- No modifiers would still be applied on multires
- installing an addon which creates a new script directory didn't add this to the sys.path.
- installing the addon was meant to set the search string to the addon name but was broken.
The patch can also be found in http://codereview.appspot.com/4431072/
##############
This patch fix anti-aliasing (multisampling) implementation for win32 platform. It also gives opportunity to embed blenderplayer inside parent window.
Usage:
blenderplayer.exe -i 123456 -m 16 file.blend
where:
123456 - parent window handler (integer, default: 0)
16 - multisample level (integer, default: 0, max: 16. Put there maximum level you want. If not supported, player will automatically try 15,14,13,...,3,2,1)
##############
This patch was originally created as part of the Burster (aka webplugin) project but benefit any one embedding the bge in a custom OpenGL context. By the way, to embed the BGE in a .Net application is really straightforward now =)
The Multisampling work for blenderplayer as a whole.
Missing functionalities:
- to expose the multisampling to the ui (so far it only works in console)
- window focus and keyboard messages for embedded blenderplayer (supported in their previous patch for 2.49, yet to be ported over)
- handle resizing (to be investigated, indeed the changes in getState() in GHOST_WindowWin32.cpp are going to get in the way of that if I'm not mistaken. To be addressed together.
Doxygen documentation to be added whenever I sort out how to do so. Sorry Nathan too many stuff to deal with at the same time. The sooner this patch gets in, the sooner the missing functionalities can be patched on top of that.
Reported by Thomas Engel
Fix [#26938] Blender Zoom not working after startup (Windows)
Reported by Ilija Boshkov
by applying patch [#26881] Fix for console disappearing in debug mode [Windows]
Submitted by Alexander Kuznetsov (AlexK)
The patch moves console toggling code into GHOST and improves on the toggling behaviour.
The patch changes handling of WM_SYSCOMMAND so that alt-key toggling isn't a problem anymore.
simple modifier, almost like a hook, except it can deform with 2 object source -> target, has option to preserve rotation and use different falloff types.
added a label for image sequence images showing the image file's frame, even when not able to load, this also gives realtime feedback to the user while dragging the frame offset/start/duration buttons about so they can better understand how these settings work.
Simple incorrect property accessing fix (enum_items vs. items).
Additional change: removed hotkey to toggle airbrush in sculpt mode
TODO: incorrect hotkeys are shown in stroke method menu, but it's how
hotkey string detecting works now.
Made some improvements to the point density texture. Added support
for tweaking the falloff with a custom curve. Also coded new
falloff types based on the age or velocity of particles.
Also added a test break check to the volumetric shade cache code,
to avoid nasty hangups from the preview render (on render, exit,
etc).
Poll function was too strict for this case that's why there was
no options in operator panel.
Additional changes:
- Added 'direction' parameter to operator so now extruding
could be made from script by providing direction vector.
- Fill this direction vection in operator's invoke functions
so abjusting offset/steps in operator panel gives better visual
feedback -- direction stays unchanged so user could easily see
final result.
- Made some tweaks to soft limits, so adjusting values by mouse
drag isn't such confusing now.
Tested in normal mode (from 3d view), as script and in background mode.
Haven't noticed any regressions.
some bones made un-selectable
Pose Bone selection (like object selection) used OpenGL to draw
wireframes with different colours into an offscreen buffer and then
from that identify which item was clicked on. The bug here was that
unselectable bones were getting drawn for this step too, so they were
getting caught instead of the control bones that were selectable.
When in quad split view, operators that only apply to the unlocked region can now be accessed from menus and when the mouse is over a locked view.
Applied to:
- VIEW3D_OT_object_as_camera
- VIEW3D_OT_view_persportho
- VIEW3D_OT_view_orbit
- VIEW3D_OT_viewnumpad
"glow" that was appearing on bones drawn using "octahedral" from
certain angles. It seems to have been caused by a typo in the code
used to "clear spec colour"
Nothing is changed by default but some linux distributions want to have executing python be opt-in.
This keeps the same functionality but disables auto-run from factory settings and in background mode unless its enabled as a command line argument.
This CMake option is marked as advanced and wont show in the regular options list so its less likely to be enabled by people that like to turn everything ON without reading descriptions :)
icon is set'
* Main problem was calling BKE_icon_changed too much, blocked previews
from updating
* Also fixed clearing the icon when it's not a valid image path
Review link: http://codereview.appspot.com/4356045/
blocks via PyAPI/RNA
For example:
ob = bpy.context.active_object # assumes default cube has some drivers
added already before running script
dst = bpy.data.objects["Camera"]
adt = dst.animation_data_create()
for driver in ob.animation_data.drivers:
new_driver = adt.drivers.from_existing(driver)
- added notes to release todo's.
- renamed view3d view transform matching functions.
- added assert in edge split modifier to make a certain bug easier to spot.
- making local object data - Curve/Mesh/MBall lost references to linked materials.
- joining a linked mesh object into a local one lost the link.
As well as these reported bugs, checked all local functions for consistency/correctness and found other cases which would also fail.
- making local metaball didn't ensure unique ID name.
- make_local_armature() was missing check for object users - main body of code would never run.
- local particles didn't set the dupli-group or textures to extern.
checked all local functions for consistency/correctness.
* Implemented a new operator "WM_OT_properties_context_change" to switch to a different tab inside the properties window.
* This is used now inside the Modifier tab for Simulation Modifiers. Based on a mockup by Janne Karhu:
http://www.pasteall.org/pic/11261http://www.pasteall.org/pic/11262
Rather than having a delete button there anymore, the button changes the context to Physics/Particles, where you can edit the settings and delete the actual simulation.
- When strength is 0, there's no need to perform any of the
calculations at all
- When the vertexgroup weight for a vert is set to 0, skip evaluating
the modifier for that vertex as it should result in no-change to the
final result
- allow RegionView3D.view_matrix to be set.
- RegionView3D.view_rotation was inverted.
- add C function view3d_settings_from_mat()
note, intentionally removed NULL checks, double checked this is ok with callers.
- made EXPANDED_AGRP take bAnimContext as an argument.
- remove unneeded NULL check drawFacesColored functions.
- comment some vars which are set but not used.
It's not real fix, just patch which makes things better by using average
normal for each displist separately. There are still some artifacts with
quite highly deformed letters "O" or "g", but correct fix would need
to calculate derivative of modifiers, which isn't implemented yet.
- ghost C api, BLI_get_folder_version() could assign garbage values.
- pointcache ptcache_find_frames_around() had a superfluous NULL check which would have crashed anyway if actually NULL.
Keymaps for subdivision_set operator were only added for Object mode.
Since this is useful for meshes in particular, added these in EditMode
keymap for mesh editing too.
Vertex offset, which was used to update referenced keys was calculating between
editmesh (which represents shapekey data) and base mesh (ob->mesh) which represents
Bases key.
This commit fixes bug with incorrect ofsset calculation for case when some keys
got other (not Basis) keys as relative key by calculating offset using EditMesh
(new shapekey data) and keyblock data (which was used to create EditMesh when
entering edit mode).
This commit shouldn't lead to regressions, but maybe there's something else
which should be fixed for such kinda complicated cases -- more testing would
be welcome.
======================
All this work with sculpting on armatured/deformed mesh allowed to
implement sculpting on non-locked keys very easy -- just use the same
approach of propagating offsets from deformed PBVH to "sculpting layer".
- If key is locked, then old logic would be used.
- If there's multires modifier enabled, sculpting would happen on multires.
Recently there were 2 reports about zoom not working right because at some point you can't zoom in any further.
This is not actually a bug with zoom but a limitation in blender that there is no way to move the viewpoint forward (unless you count rotate 90d, pan, rotate back which is a crummy workaround).
So adding view dolly operator:
- Supports zoom to mouse position setting.
- Supports dolly hoz/vert, invert setting.
- Moves by a fraction of the view 'dist', so the zoom distance my be roughly in proportion to the scale of objects in the scene.
only used in perspective view since this is not useful in camera/ortho view.
- continue zoom now uses the same options as dolly (hoz/vert & invert).
- remove event mouse coord hack to bypass touchpad zoom invert, instead pass invert as an argument.
Ctrl+Click on mesh or curve view was using the selected points location or the cursors.
if either of these was behind the view it would add the point at (0, 0, 0).
now fallback to the view orbit pivot, added this option as an argument to view3d_get_view_aligned_coordinate().
B-Bone resizing was acting more like translation than resizing when using numeric input. Added the flag to set all xyz values to same value when using numeric input
- remove some warnings
- fix typos
- cmake allow in-source build (when WITH_IN_SOURCE_BUILD is defined)
- cmake, use an explicit list of rna files (don't glob)
A change in the poll callback that Available KeyingSet used to use
restricted its use to Object-mode only, while this could also be
useful in Pose Mode (though it would only affect all channels there).
Made this use a custom poll callback now that tests for whether the
active object has an action. This does mean that if you select a bunch
of objects with animation data, but the active object doesn't have it,
then the keyingset will fail to fire, but that's been marked as a todo
in the code.
* Explode modifier was flagged as "nonconstructive", so the "apply as shape" option was shown. And yes I know exploding things probably isn't usually considered as very constructive, but.. :P
Outliner objects showed "active" or "selected" in confusing
ways. Now the display is simply binary, and more in line with
how other items in outliner draw:
- text is white = item is active
- color circle behind icon: shows selection state
reported by Michael R
Don't print the error message, as it isn't very useful. If there is no Quicktime in Blender GUI in a build that has support enabled, it means that the user hasn't installed Quicktime.
Top bar: Add -> Mesh -> UV Sphere + Enter crashed.
It didn't crash with leftmouse, but that was coincidentally working.
Menus were freeing modal handlers in Window, while handlers were still
in use. Fix provides to tag handlers for being freed now.
Will add on my attention list for more elaborate checking work here, for
upcoming 2.57a I rather stick to minimal change in code here.
This bug was caused by tabs->spaces conversion. Change pate-ing logic to
paste buffer AS-IS (without any conversions).
This commit also fixes undo-ing block deletion which contains tabs when
"Tabs as spaces" is toggled on. Also, markes shouldn't be moved after
pasteing new buffer.
Never memset(&ob, 0,sizeof(class)) when there is a constructor, it overrides all memory.
The problem was that the memset(0) was setting the scaling to (0,0,0), the height of the cone became 'infinity'
so GJK would iterate 'MAX_ITER' without converging due to this #NAN value
* Collision modifiers are disabled here by intention, but particles and softbody were copied the wrong way over :)
* On a further note I don't really get this whole "link modifiers" thing as it just copies the modifiers. As modifiers aren't ID blocks there's no sense in calling this linking!
* Secondly I don't think particles, smoke etc should be considered as modifiers here at all, meaning they shouldn't be linked/copied. These "modifiers" only read the mesh data at a certain location of the stack, but don't actually modify the mesh in any way (more info here http://wiki.blender.org/index.php/User:Jhk#Modifier_Stack_proposal).
EditMode mesh: tool "Mesh Rip Move" shouldnt be in the toolbar.
The tool was coded to use the mouse position next to the selection.
The rip then happens correctly after pressing V and move mouse away
from selection.
Actions now get tagged with an ID-code, which is used to determine
what ID-blocks they can be assigned to. This ensures that material
actions cannot be assigned to the object-level for example.
* Action lists in general will now show only the actions that can be
set for that particular slot. This prevents selection of invalid
actions, and helps cut down the list of actions.
** An exception here is the Add Action Clip in NLA Editor, which will
show all actions but will only add where appropriate. This is because
it's not easy/possible to tell in advance which blocktypes to filter
for when building this list. (TODO?)
* The "Action Editor" is now strictly for object-level action
editing+setting now. This avoids repeateded confusion by people who
try using this to view their shapekey actions, which should go to the
Shape Key Editor instead!
** A context switcher for the legitimate times where this capability
might come in handy is still being investigated.
* "Floating" actions (i.e. actions in some action_library.blend) are
NOT able to be automatically tagged until they are assigned to some
datablocks (i.e. loaded onto the rig + played back once). It is
possible to write scripts that check for certain RNA-paths and "guess"
what datablocks they work on, but it is recommended that you load up
the Datablocks Viewer, and go through such actions by hand, setting
the "ID Root Type" property as appropriate per action.
Using search for operators showed ambigious names like "Duplicate" or "Delete".
Default names should give at least a descriptive label. In case operators
are collected in a group name-shortening should be handled separaly.
- the frame from the current scene wasn't used - whereas with rendering it is, set the current frame as is done when node rendering.
- camera switching also failed, added a call to it.
also found curve click-extrude was always aligning the new points depth to (0,0,0), now work the same as mesh edit - align to the selected point or the cursor if none are seleted.
FFMPEG was reallocating buffers it didn't own and wasn't allowed to. This workaround should work now flawlessly.
Also fixing a bug regarding unpacking sounds, the UI stated unpacking to //audio/filename while it was unpacking to //sounds/filename
Compositing, lens distortion node: it delivered RGBA out with alpha zero,
making it not draw previews, violates both premul or straight alpha usage.
Now it just sets alpha to 1 for entire output.
Cleaned code readability a bit as well.
- allow building blenderplayer with redcode.
- when ffmpeg is enabled remove strict compiler errors for imbuf and blenkernel since its hard to avoid these warnings across ffmpeg versions.
Clamped the maximum displacement distance to 4 x the input image dimensions - prevents hanging when vary large values are mistakenly plugged in, such as Z buffers,
Fixes Fix for [#25713] VSE shows and renders wrong straight alpha gradient even after convert to pr
(see revision: 34540, fix by Janne)
By not breaking the seqcache interface API.
Added comments to header file, so that it is easier to understand, how the
cache API is supposed to work.
spotlights halo and intensity tooltip had "(buffer shadow)" in there,
while it works just fine with no-shadow and ray-shadow.
the volumetric "step" property is only visible with buffershadows
anyways.
- use cores+1 threads
- do not clean build directories
- separate build directories for blender and blenderplayer
This should make compilation much faster and buildbot rebuild tag could
happen more often
Text editor used to add extra indentation when inserting new line from
line with dictionary. Also, fixed extra indentation when comma is inside string.
- render check for ortho/panorama combination wasn't working since the flags were not initialized at the time of checking.
- disable panorama button in ortho mode.
Redraw issue was caused due to different redraw rectangles used for 3d view redraw and
gathering PBVH nodes to be re-drawed. I moved redraw rect expansion with rect from
previous step into sculpt_get_redraw_rect, so now redrawing works as it was planned
some commits ago -- redraw everything to which is inside currect rectangle and rectangle
from previous stroke step -- this still prevents artifact caused by fast strokes but
mesh doesn't disappear.
Brecht, Nicholas: it's the simpliest fix i could suggest atm. I've got some more
ideas with additional node flags, but it looked more complicated for me and
made code more difficult to understand. If you could see something better (like
revert all this redraw fixes for fast strokes) please tell me.
Image Editor.
There's already a flag in use for tagging the "special drawing in
Image Editor" case, so make use of that now.
Most of the changes here are just whitespace tweaks...
discussed with Janne, Ton, Nathan and we agreed this kind of change at least needs discussion with module owners.
Its also too close to release to be making these kinds of changes.
commands used:
# reverse merge
svn merge -r36073:36072 .
# for some reason this gave a lot of property changes
svn revert `svn st | grep "^ M" | awk '{print $2}'`
# reverse merging didn't work here, removing while dir.
svn rm extern/eltopo/
# manually fixed conflict in
# ./source/blenderplayer/CMakeLists.txt
#
# also manually removed 2 lines from
# ./CMakeLists.txt
It was simply missed crazyspace correction for shape keys when mesh is
deformed by modifiers and current tool is smooth/layer (they need special
approach due to they don't use proxies)
* ghost frame 1 did nothing (alpha 0.0).
* was subtracting alpha from the existing alpha which could end up with negative alpha.
* with larger frame ranges the frames on each size would get too close to being the same as the active frames alpha, use 66% alpha for all non-active frames as well as time based falloff.
- only attempt to restore old 'user' settings (not local), since bundled blender's always use their own settings.
- only automatically run 'bpy.ops.wm.read_homefile()' after copying files if the user hasnt alreadt started making changes in the blend file.
This crash was discovered by Dalai and this happened because of
unset current context (as result of call wglMakeCurrent(NULL, NULL)).
In this case glGetString(GL_VENDOR) returns NULL. Rather than add check
for vendor != NULL before string comparison, I've changed a bit logic of
context creation:
- Create context and set it as current
- If it's crappy Intel card -- delete this context and
share the only one context between all Windows
- Otherwise, use initial logic (with sharing lists and so on)
This could also fix crash when opening userprefs from a menu with Intel card.
* memory corruption when skipping over long marker (was attempting to read over end of file)
* also updated internal jpeg macros to be the same as in jpeg lib
Plugged the eltopo library into the cloth solver.
I was playing with it earlier, and it's so easy to
use I decided to quickly put it in (trunk's) cloth.
See http://www.cs.ubc.ca/labs/imager/tr/2009/eltopo/eltopo.html
. The authors are on the bleeding edge of continuous collision
detection (one of them did ILM's cloth sim).
I
don't really have to time to plug it into softbody, particles,
bullet, fluid, etc, but doing so would be pretty straightforward.
I'll leave that up to someone else.
To use, turn on USE_ELTOPO (in cmake) or WITH_BF_ELTOPO in scons.
Intel video cards don't work fine with multiple contexts and
have to share the same context for all windows. This could work
incorrect with multiple video cards configuration, so suppose
there'll be no such situation (Intel cards are mostly in portable
devices like notebooks and laptops, where there's actually no
dual video cards). Anyway, it should work much better now.
Non-Intel cards behavior was kept unchanged.
Thanks to Ton for debug session :)
if not defined (first run) these are now set blank but can be defined later.
the problem is that scons & cmake builds would link against different libraries since cmake added its own defaults.
now, by default, scons & cmake have the same libraries.
This fixes an obscure crash in MinGW where cmakes default linking with -ladvapi32 would crash on string formatting which used float precision as an argument, eg:
printf("%.*f", 3, value);
...without giving a useful backtrace or pointing to the line of code doing the string formatting.
- Assigning local materials to library objects disabled
(crashes on undo/redo cases)
- Disabling options in Material buttons to add/remove slots
on library data
- Drawing Object ID template in Object properties, this
allows browse active Object, but especially shows library
status then.
submitted by Tom Edwards
This patch introduces a switch -con and its longer version --start-console. When giving this on cmd-line you'll get the black console window. The new behaviour is to hide it by default. We'll still see briefly the console at startup and during exit, but that's something that cannot be changed.
If you start blender from a cmd.exe, the console will not be hidden.
was no longer wrapped by python as a vector. now fixed size float arrays
with PROP_NONE subtype are wrapped as vectors since its convenient to
have x/y access.
On anim-render, a click in timeline stopped render completely.
The reason for this was a bit wacko code to cope with frame-step
feature (steps of multiple frames).
I thought of fixing that, but instead decided to block any operator
in Blender to change a frame while a render is in progress.
Both render engine and UI are accessing (writing to) the same
data then, which is a bad conflict.
Still a serious weakness of threaded render, but I'll keep
trying to allow this as far as possible :)
Rotation and Location don't have Keying Sets
Added Keying Sets for Delta Loc/Rot/Scale settings (aka dLoc/dRot).
These settings could already be found in the Object properties, under
the collapsed "Delta Transforms" panel.
I've added these to the end of the Keying Sets list, since adding any
earlier will end up breaking active Keying Set setting in older files.
Besides, these settings aren't that frequently used either...
Two part bug:
Part 1) NLA Editor menu was calling wrong operator
Part 2) r35829 broke NLA scaling, since it only checked that the
transform mode used was allowed in the Action Editor (probably
confused by Part 1)
Disable this for pointer buttons since this was intended for vgroups & uvlayers.
Pointer buttons often reference ID's which can be in the 1000+, lookups like this on each button draw is not useful since invalid pointers here will likely crash on scene evaluation.
Small fix for recent folders in filebrowser, didn't save correctly when too many folders in the list (more than the currently allowed 10), now skips the oldest entries.
Reported on IRC.
Force link to zlib for all platforms/tragets to be sure there'll
be no gzopen64 symbol undefined (it was in some dependend library, so
blenderplayer wasn't working in fedora 13 32bit)
Problems was in 0-based numbering vs. 1-based. Fonts are getting converted to
displists when they're displaying and here base conversion happens, so no need
in conversion when assigning material to selected text.
This makes indices more inconsistant, but saves compatibility in both of
directions. Maybe in further releases we'll make indices consistent.
Join redraw rectangles from previous step and current during one stroke
and reset previous rectangle when stroke is finished. This shouldn't cause problems
with viewport changes during the sculpt session and this should emulate
object boundbox merging for states "before step" and "after step".
At leastm it's more logical because boundbox could not only grow up during
stroke, but it could also shrink down.
Texture preview render now doesn't increment/decrement material->texture
user counts anymore. Blenders library.c code could use some overhaul
once to support data relinkage and copying around better. :)
on smooth/flat flag on faces. This does give better results for low poly
game models, but there's just too much functionality that depends on this
(modifiers, displacey, editmode tools, extrude, ...), that there's not
enough time to fix these before the release.
- make outliner search unrestricted
- clip-end not on 5000 (zbuffer errors), but on 1000
- disabled continious grab
- "duplicate action" default on
- addon svg import default on
It's not the most efficient solution, but this can be optimized later. It's
best to copy out all the pixels at once into a list, rather than accessing
them one by one.
It's obviously much better than the two previous behaviors, since you don't
get confusing linked materials, and you don't have useless added materials.
- Remove packed to buildbot_archive.zip archive file to keep scons install
directory as clean as possible (if i understood correct, this file isn't
actually needed in the future, so we could remove it just after packing rather
than cleaning install directory before packing as it's done for cmake)
- No need in scripts installation for linux slaves when preparing archive
Operator "New Image" cannot be redone simply, undo system keeps images
around, and each redo will create new block. Removed it from being
registered for redos.
Also: fix up for default operator names to be complete, and not just
"new" or "load". If we want contextual shortening of operator names
(like in Image menu), that should be handled seprately, automatically,
or anything we can sanely come up with that will suit any possible
language as well.
- Use the same config as we're using for release builds
- Added stipping (maybe it should be generalized)
- Not software gl packing, bot would be easy to add
Note, that you should have special environment to create
proper linux builts (which would work for everybody)
*Brush option "size" had different naming in 3D View / Image Editor.
Ported back name "Radius" to RNA (Brush and ParticleBrush)
Discovered by Bart Crouch. Thanks!
A lot of UV tools only work in Image editor, but the operator polls
allowed them to be used in 3d window too, causing crashes in cases.
Also added a poll() check in F6 redo menu.
I'm finally yielding to months of feature requesting, and adding
support for filtering F-Curves by name, where the "name" here is the
text which is displayed for each F-Curve in the Animation Editor
channel lists.
To use, just enable the magnifying-glass toggle on the DopeSheet
filtering settings, and enter a snippet of text to find within the
names of channels you wish to filter. This is case insensitive, and
currently doesn't support any wildcard/regrex fanciness.
Some examples:
loc <--- location curves only
x loc <--- x location curves only
x eul <--- x rotation curves only
rot <--- rotation curves only
etc.
was not recursively restoring sound strips on paste.
also found many duplicate functions were defining the transform mode as in int but getting as an enum, use enum for both now.
Operator for switching brushes based on type, cycling through brushes when multiple exist.
This has the advantages over the old method that it doenst rely on hard coded brush names and if there are multiple brushes of the same type it cycles between them.
also fix error in previous commit which broke number buttons changing brushes.
If greasepencil became active, making Blender area full or restore
screen from full, the greasepencil handler was still running, using
outdated area pointer. This crashed Blender.
Now greasepencil modal() ends on the case its own stored area does
not exist anymore.
* Fluidsim has to be before any constructive modifiers.
* Also a bit nicer domain size calculation + a warning message for using flat objects as fluid objects.
* Some code cleanup and clarification too.
was missing array cap ends, wave map object and shrinkwrap objects.
use modifiers_foreachIDLink() rather then having to list all modifiers ID's in this function.
also add foreachIDLink() for smoke domain.
This fixes a bug where a linked object has as a modifier using an indirectly linked object for the missing cases mentioned above.
- Pose Propagate and Pose Sliding tools now work in the same way as
Pose Library previewing, with regards to selections. If some bones are
selected, then only those will be affected. But if no bones are
selected, then the whole rig gets affected.
- Added a "On Selected Markers" option, which only propagates poses to
frames where there's a selected marker. Animators can combine this
with a "select markers whose name contains..." operator to get an
effective way to manage hand-keyed walk cycles, etc.
- Renamed "Last Keyframe" mode to "Before End". This mode still just
copies the pose to all keyframes starting from the current frame until
the last one encountered per F-Curve
- "Last Keyframe" mode (new one) now copies the pose to the last
keyframe. This is useful for making animations cyclic (i.e. go to
first keyframe, edit, then Pose->Propagate->To Last Keyframe (Make
Cyclic))
this does not fix the underlying problem that once in this state you cant exit editmode, only that the layer key bindings dont let the state come about.
Looked into fixing properly but its not so simple since currently object mode operators all use "active_object" which is supposed not to be set when in a hidden layer.
* "Make fur" now creates a basic fur material for each object (smaller strand tip width and a little bit of surface diffuse)
* If fluid/smoke emitters aren't rendered the objects are also set to draw only in wire mode so that the fluid/smoke can be clearly seen.
* simplify export and import, now that we have blender profiles for lights. The vanilla import is now more in line with the specs. If a blender profile is found, skip normal import, use the profile data instead.
* multiply energy into color rgb export (common profile).
* recalc distance taking metrics in account
* A couple of operators to quickly create effects that would otherwise take some time to set up.
* Nice to use for demoing functionality or as a starting point for more complex effects.
* "Make Fur" - Gives every selected mesh object particle fur with a desired density and length.
* "Make Smoke" - Makes each selected object a smoke emitter and creates a new domain object around the emitters with the correct material to render the smoke.
** Has style options for "stream": constant smoke flow, "puff": only create smoke once from the volume of the emitter object, "fire": enable high resolution smoke and set a secondary fire color texture for the domain object.
* "Make Fluid" - Makes every selected object a fluid object (normal/inflow) and has the option to start fluid baking immediately.
* This should provide a nice base for extending these / adding more operators for different effects.
provided by Mike Sloman
from the patch report:
When building on Mac OSX with SCONS the Info.plist file is zero length and thus OSX cannot identify icons and file type associations.
The problems was in the execution of the "date" command to add times in the modified bundle name strings passed to sed. This failed and thus the info.plist file was overwritten with 0 length file.
Noted that a floating point exception caused by this r28953.
With the render engine using DerivedMesh normals this came up as a memory error when rendering some files.
for now zero the normals, could interpolate between vertex normals if needed.
* Too strict poll functions for render & display panels.
* Like said in the report cache & vertex groups are a part of the particle system (not particle settings) they can't be shown when the settings are pinned.
submitted by Tom Edwards
Fix [#25473] 64bit Windows installer for version 2.56 is not working
patch submitted by Caleb (Dobz)
The thumbnail patch adds a thumb handler DLL that adds .blend thumbnail support in Windows Explorer. A -r option is added to do registration in background. The patch also improves icon building and metadata for blender.exe.
Caleb fixes and cleans up our installer to an acceptable state. The patch uses the new -r option to do the .blend extension and thumbnailer registration.
Thanks to both Caleb and Tom for their efforts!
Missing null check for "improper" driver variables.
------
Also, this commit includes some WIP NLA evaluation reshuffling which
should be harmless. I had originally intended this for a later
commit...
Bone Color Sets weren't set in the call used here. This seems to be a
newish operator that didn't exist in the past, so this error didn't
really show up in the past.
Gave the report-timer (in top header) an own TIMER define, to not
confuse with timers in use for UI. Easier for debugging. :)
Also gave better malloc names to regions in 3d window.
All this for testing proper buffer drawing in Intel cards... which
is still unsolved.
reported by Keith Boshoff (Wahooney)
Instead of a confusing backtrace popup, tell the user the image editor cannot be found, and where to set the path to it.
- timer = WindowManager.event_timer_add(time_step, window=None)
- WindowManager.event_timer_remove(timer)
Still TODO, is a way for python to check the timer identity.
Add cancellation checks for time* transforms that only run in specific spaces.
Hide Transform Mode operator property (it shouldn't be modified after the operator is run), made default mode Translate, not the useless Dummy.
[#25045] User Preferences Input, not enough menu choices
[#26525] Saving the Maya preset and another key configs
Untangle keyconfig presets and interaction presets (maya keyconfig would set interaction setting, blender interaction would set keyconfig).
Preset menu on slash screen now displays a list of available keyconfigs (as before) but will also set the correspondant interaction presets, if it exists)
* Driver F-Curves were not getting fixed by the "FCurve/Driver Version
Fix" tool. This was causing problems such as shapekey drivers from
older (2.56 compatible rigs) failing to run. As well as renaming the
paths for these Driver F-Curves, the "disabled" flags also get cleared
from these drivers so that they can be run again
* "Revive Disabled F-Curves" operator in Animation Editors can now be
used to revive disabled drivers too.
into account that some tools use normals for things other than display. Now
we properly initialize vertex normals at flat faces too.
Also fixed a normal refresh issue, and deduplicated CDDM/mesh normal
calculation code.
Recent renaming of Key.keys to Key.key_blocks broke converter for old
files.
Campbell Alert: However, more seriously, it has broken many current
rigs, especially those with heavy usage of (driven) Shape Keys
The purpose was to set a wait-cursor draw on 3d windows. I tried for
but it keeps failing... inside threads you can't do UI stuff. Needs
further thinking, probably something via jobs system.
When rendering, during processing scene data, drawing in 3d window
is now locked. Can get extended to more areas in UI easily.
At least this solves all crashes with conflicting memory access in
render && 3d drawing. Deleting objects via operators or delete
modifiers isn't locked yet.
Also fixed: crash on quitting a renderwindow when it was rendering.
Changed selection flag used for the selected_nodes context property from SELECT to NODE_SELECT. These are just incidentally the same, but NODE_SELECT should be used for nodes.
* Copying/pasting sequence strips didn't properly check for unique names between the copied/pasted strips, so the rna paths of copypasted strips couldn't always be checked properly.
Now this gives the line number of the scripts thats running, eg:
uiItemFullO: unknown operator 'some.operator'
/c/bin/2.56/scripts/startup/bl_ui/space_view3d_toolbar.py:73
Ported joeedh's Euler Filter code from Python to C so that this is
more in line with the other Graph Editor tools - i.e. joeedh's version
only worked on the active bone's curves, while standard tools could
work with multiple bones/objects at the same time.
To use this new version of this operator:
1) Select all the F-Curves for all 3 of the components (XYZ) for the
euler rotations you wish to clean up. In the Graph Editor, they must
be one after the other (i.e. you can't have "RotX, RotY, something
else, RotZ")
2) Activate the operator from the Key menu in the Graph Editor
In an old test file I have floating around, this method did not appear
to be good enough to fix a very clear discontinuity in the middle of
the action, so I'll test some additional methods too
[#25725] Transform bone constraint & 'local with parent' as owners
space
[#26014] ChildOf Bone Constrain (influence) works in wrong workspace
Moved the influence calculation stuff outside of the space conversions
(i.e. so that it is done in worldspace only) fixes these problems,
which seem to arise when a constraint doesn't work in worldspace AND
doesn't need to apply inverse correct for this space conversion when
it's done, hence resulting in mismatch between spaces for old and new
matrices resulting in all the weird behaviour.
Patch to fix this from Jahka. Cheers!
I still would love to see a change in the action actuators (Action, Shape and FCurve) even if only a rename or unifying them in a single actuator (I do have some patch for that). Anyways
- Colormanagement code was gamma correcting non-RGBA buffers in composite.
(Like vector buffers).
- Crash on using "use Color Management" button during composites, because
it was freeing all node images. Added code to stop first jobs before
freeing. It sends notifier for recomposites after free anyway.
tried doing this smarter by validating the property exists in UI code before access but this ended up making it too complicated and also hard to account for possible access without checking every time.
for now just redraw all areas when a user property is changed.
Envelope FModifier data points were getting saved in a non-portable
way. Apparently this broke when trying to read files with this data on
PPC machines if the files were saved using Intel machines.
This tool automates the process of copying a pose to successive
keyframes, making it easier for animators to go back and change the
pose for some controls which remain "static" for periods of time.
Previously, animators would need to do a "{Ctrl-Pageup Ctrl-V} *
number_of_static_keyframes" dance for each set of controls that this
happened on, which is not too good ergonomically speaking.
There are two modes exposed via the menu (Pose->Propagate):
- "Pose Propagate" - also known as the 'WHILE_HELD' mode, which
propagates to all keyframes that are holding the same value
- "To Next Keyframe" - which only propagates the pose to the closest
keyframe in the occurring after (but not including) the current frame
Additionally, there are a few other modes that can be used, though
they are less useful for direct use from the UI, though they can be
used via the PyAPI as need be.
---
Also, I did some cleanups in the "Pose" menu to bring it more into
line with the Object mode one. There are some more tweaks that could
still be done here, such as bringing the keyframing operator entries
under a submenu too (as in the Object mode version) to get the length
of this under control.
* added check for valid files a bit too eagerly in Rev.35654. Solved now by
setting correct type for .blend file contents. Objects are now tagged as
regular files, while the list (Mesh, Material, ...) (no idcode given) is
tagged as directory.
(bugfix #26573, #26574 and #26551)
Layer visibility code for hotkeys uses different code still, added same
new DAG_on_visible_update() in this operator.
restore, would not get their dependencies updated when they became visible.
It happend with a shrinkwrap modifier in these reports, but could happen with
other modifiers too.
Now we keep track of which layers have ever been updated since load, and tag
objects on them to be recalculated when they become visible.
this is useful for the register() class method which is called before the class gets structRNA assigned.
eg:
class MyClass(bpy.types.PropertyGroup):
@classmethod
def register(cls):
cls.name = StringProperty() # assigned but registration is delayed.
reported by Juan Linietsky
The export was not really wrong, but Maya expects their profile to be present inside <geometry>. Added this for mesh with ME_TWOSIDED flag set.
Ideally this will change in the future to be controllable through options to the exporter (like current Google Earth and 3DS Max extensions for <double_sided>).
reported by Juan Linietsky
Export <specular> for <phong> and <blinn> shaders, <shininess> was already being written for these.
<lambert> shader doesn't have <shininess>.
Right now we write <phong> when blender spec is phong, <blinn> when blender spec is blinn. When spec is
any other shader, and diffuse shader set to lambert, we export as <lambert>. Any other combination defaults
right now to <phong>. This will change when Blender specific profiles have been created for the shader
combinations in Blender.
* Adding keyword arguments to LibLoad
- load_actions (Default: False) Ensures that all actions are loaded. Otherwise, just actions used by objects in the scene are loaded.
- verbose (Default: False) Turns on extra prints (eg SceneName: MyScene).
* Also making error reporting better for when an invalid group/idcode is given.
move calls to the classes register/unregister function into register_class() / unregister_class() and add docs.
also other minor changes:
- remove face sorting keybinding, was Ctrl+Alt+F, this is quite and obscure feature and face order normally doesn't matter, so access from Face menu is enough.
- add commented out call to mesh.validate() in addon template since its useful to correct incomplete meshes during development.
Using environment map type "load" increased user counter on each
preview render.
Also noticed that this type of envmap use wasn't threadsafe, causing
imbufs being allocated for all threads. Also fixed that.
ui/ --> startup/bl_ui
op/ --> startup/bl_operators
scripts/startup/ is now the only auto-loading script dir which gives some speedup for blender loading too.
~/.blender/2.56/scripts/startup works for auto-loading scripts too.
* Particle emission now updates all parent objects too to the exact emission time.
* This only does object level animation as updating the object data for every particle would be too slow.
* A better fix could be to interpolate the emission location directly from the current particle emission location and the location from the previous frame, but for this some point cache changes have to be made, so it will have to wait.
Modified to only do one lookup.
from Martin:
"Basically, what it does is allow you to add register and unregister class methods to rna types, this way you don't have to rely on module register/unregister methods to setup your types properly (and it makes them easier to move around when reorganizing code and easier to understand what a type does when reading code). This is especially nice for PropertyGroup classes that are added as properties to existing types, you can easily see in their register methods where they are added and removed in their unregister method. Obviously, those two methods are optional, so current code still works fine."
weighted normals as the render engine, and the render engine will copy
normals from the mesh rather than always recalculating them.
Subsurf/multires still use regular vertex normals, but they are expected
to be sufficiently high resolution to not need this.
This means that normal maps displayed in the viewport actually match the
render engine exactly and don't have artifacts due to this discrepancy.
It of course also avoids unexpected surprises where your render normals
look different than your viewport normals.
Subversion bumped to 4 for version patch to recalculate normals.
Patch by Morten Mikkelsen, with some small changes.
* Separate out selection flags from file type flags, was bothering me since forever ;)
* Also renamed ACTIVEFILE to SELECTED_FILE to better reflect actual usage in code.
* Fix crash introduced with last commit, better check for valid selection
Patch from Alexander Kuznetsov: Toggle selection rather than just extending.
Implements behaviour that unintendedly was available with previously using macro operator for selection. This was removed and now the functionality is properly implemented.
Patch accepted with minor changes:
1. Used enum rather than #defines and added value for removing from selection (deselect)
2. Moved if (select) outside file_select_do and improved check for whether last file in selection is actually selected. (Necessary since toggle can deselect and toggle select should still make file active)
3. Additionally fixed check in file_select_invoke to be consistent with border select.
Patch from Alexander Kuznetsov: Real-time File Selection, thanks for the contribution.
Still made a few minor changes from latest patch:
1. Rename SELECTEDFILE to HILITED_FILE, since we are not actually selecting the file, but previewing/highliting the possible selection.
2. Also made this clearer by not drawing the files as selected, but just highlight them.
3. Removed the Select/Deselect toggle when clicking on file, will be committed separately soon.
Improving UI visualization of using Material nodes:
- Property editor: the context top line now shows the linkage path
(in material and texture view mode)
- Activating new material node refreshes preview in property editor
when in texture mode.
This should make editing textures with node materials a bit more clear.
Tried to resolve todo item "Layer change in 3d view header doesn't update
dependencies on changed previously visible objects". Code is too much spread
around now though, and the 3d view header shares RNA code for changing layers.
I'll let this be a "glitch" for now.
- script_load_addons: loads and onloads all addons, error if any cant be enabled or disabled.
- script_load_modules: loads all blenders modules, error if any modules raise an exception or if any modules in the scripts dir are not tested (except for presets and templates).
These are useful because lazy module loading means a module can have an error which isnt raised until the tools used. This gives a way to detect basic errors that used to happen on startup or when enabling the addon.
reported by Georg K
with patch [#26469] Windows mouse fix
by Alexander Kuznetsov
The amount of mouse grabs wasn't properly balanced with ungrabs, thus preventing from using proper mouse input outside client area.
* The selected uv-vertices (the ones being transformed) weren't being disregarded as snapping targets, so moving two verts that were in the same location together was impossible.
* The old collisions code detected particle collisions by calculating the
collision times analytically from the collision mesh faces. This was
pretty accurate, but didn't support rotating/deforming faces at all, as
the equations for these quickly become quite nasty.
* The new code uses a simple "distance to plane/edge/vert" function and
iterates this with the Newton-Rhapson method to find the closest particle
distance during a simulation step.
* The advantage in this is that the collision object can now move, rotate,
scale or even deform freely and collisions are still detected reliably.
* For some extreme movements the calculation errors could stack up so much
that the detection fails, but this can be easily fixed by increasing the
particle size or simulation substeps.
* As a side note the algorithm doesn't really do point particles anymore,
but uses a very small radius as the particle size when "size deflect" isn't
selected.
* I've also updated the collision response code a bit, so now the particles
shouldn't leak even from tight corners.
All in all the collisions code is now much cleaner and more robust than before!
old bug from r6568 (or before) in BKE_write_ibuf(), when passed TAGA format it would use whatever format the image buffer already had.
expect this is exposed by updated image saving code in 2.5x.
really old one!, since initial commit blender would crash scaling down large sizes eg: 60962 -> 128 (width or height).
the problem is scaledownx/y doesn't check buffer endpoints, with really large images theres a loop on a float value which can fail with large image sizes.
previous commit added asserts if the buffer runs over (assuming it doesnt crash),
This commit changes an epsilon value, tested this with random small images as well as images over 200,000 px, and it works fine, this is still flakey though and for really really big images it probably still fails.
Not really a "bug", but it was on my todo anyways. Based on patch
[#26508] by Campbell, with a few modifications including extending
this to the Action/DopeSheet editor too.
This manisfested as an inability to draw any more strokes after trying
to draw a stroke while the active layer was locked and then unlocking
that layer to try and draw on it again.
People had trouble finding the 'Create New Directory' operator button in the file browser. Changed this to show the text as well and make button bigger.
For cmake users, you should now use "make install" instead of "make". This was
already changed for Linux and Windows. Same for Xcode, use the install target.
Changes:
* CMake install mechanism, resulting bundle was verified to be indentical here.
* For cmake, include Info.plist in bundle using builtin mechanism for that.
There was some code in packaging.cmake, but it wasn't correct as these
properties need to be set on the executable.
* For scons, fix app bundle version, was still using removed release/VERSION.
* Remove unused blendercreator.app and blenderpublisher.app.
* Fix Info.plist being set as a binary file in svn, should be plain text.
Object.is_modified(scene, 'PREVIEW') function for python exporters to check if any modifiers or shape keys are applied (weather the original mesh can be used for exporters).
- look for icons in datafiles/icons (was looking in datafiles)
- was loading all images in datafiles/ on startup to check if they were the correct icon size, commented this since its unnecessary disk overhead on startup & images are checked for correctness when used anyway.
when running blender from the source dir would load splash.png every time.
also add missing NULL pointer check if the icon couldn't be loaded and ensure no buffer overflow check on icon path creation.
Discussed with Campbell, it's not actually bug and it's more about limitation
of topology mirror. It will work properly when both sides of mesh have matching
unique topology.
Added note to tooltip, so now unpredictable behaviour shouldn't confuse users
so much.
Also gray out "Topology mirror" when "X Mirror" is disabled.
- opening a file with blender by passing it as an argument would and loading it once in blender left script auto execute flag in a different state.
- command line args --enable/disable-autoexec were being overridden by the user prefs.
creating RNA within draw functions can free existing RNA, crashing blender when this is already used in the UI.
disallowing this so it raises a python exception.
This was being used to dynamically generate addon categories so for now they are hard coded and we need proper enum-functions for python to do this.
* Don't use "make install", doesn't work yet.
* Use total number of cores rather than number of processors.
* Quicker number of cores query, by not asking for all system info.
- Sync Markers option works for local markers (or any other list of
markers in future) too now.
- Apply Pose to Restpose operator now displays a warning if an action
was found (warning about the action now being invalid)
useful.
Instead of only working for the "extend" transform mode, standard
transforms now work too now (i.e. grab and scale).
TODO:
This currently only works on Scene markers, though it should be
possible to make this work on a provided list of markers instead...
'create' was used as prefix and suffix, change dupli list functions to use as suffix, this matches obj.animation_data_create() & obj.animation_data_clear().
obj.create_dupli_list() --> obj.dupli_list_create()
obj.free_dupli_list() --> obj.dupli_list_clear()
Don't use 'create' for object to mesh function since other uses of this are for createing data which stays attached, instead use mathutils style naming convention.
obj.create_mesh() --> obj.to_mesh()
In addition don't store unfinished links in the nodetree->links list any more. This makes code a bit safer because all links in that list can be considered valid now. The temporary bNodeLinkDrag structs used by the modal linking operator are now also stored in a list in SpaceNode, so these links can be drawn too (this separation also allows different display of temporary links, e.g. currently they are drawn on top of all nodes).
- Custom properties are now affected by the Pose Sliding tools too.
This is now more important to support, given that modern rigs use
these a lot for facial expressions/posing. By and large, this should
work fine, though discrete integer values may experience a bit of
trouble
- Fixed potential bugs with the code which detects which F-Curves are
relevant to a PoseBone's transforms (+ custom props). This was prone
to being tricked by certain setups if the names of the bones contained
some of the keywords these were searching for.
- Shuffled some code around: moved bulk of logic out of vec3 case into
new function for single-value, since it was really doing per axis
already
Made changed to linux2-config and freebsd*-config. Windows/MacOS are kept
unchanged.
I'm not sure which flags should be used for freebsd systems, so flags for
them also kept empty. Maintainers of this platform, please check which
flags should be used and make changes to rules.
in RNA as "translation" values, which implies dimensionality is
incorrect here as F-Curves do not specifically "know" about the type
of the data they affect
add PhysicsConstraints.exportBulletFile(char* fileName) python command
I'll be checking the bf-committers mailing list, in case this commit broke stuff
scons needs to be updated, I'll do that in a second.
this is not well suited to RNA so this is a native python api.
This uses:
bpy.data.libraries.load(filepath, link=False, relative=False)
however the return value needs to use pythons context manager, this means the library loading is confined to a block of code and python cant leave a half loaded library state.
eg, load a single scene we know the name of:
with bpy.data.libraries.load(filepath) as (data_from, data_to):
data_to.scenes = ["Scene"]
eg, load all scenes:
with bpy.data.libraries.load(filepath) as (data_from, data_to):
data_to.scenes = data_from.scenes
eg, load all objects starting with 'A'
with bpy.data.libraries.load(filepath) as (data_from, data_to):
data_to.objects = [name for name in data_from.objects if name.startswith("A")]
As you can see gives 2 objects like 'bpy.data', but containing lists of strings which can be moved from one into another.
the same, but big changes have happened both on the outside and on the inside.
New UI:
* The old parameters were quite true to the underlying algorithm, but were quite obscure
from a users point of view. Now there are only a few intuitive basic parameters that
define the basic fluid behavior.
** By default particle size is now used to determine the interaction radius, rest
density and spring rest lengths so that it's easy to get stable simulations by simply
emitting particles for a few frames and adjusting the particle size (easy when the
particle size is drawn) so that the fluid appears continuous (particles are touching
eachother).
** Stiffness - in reality most fluids are very incompressible, but this is a very hard
problem to solve with particle based fluid simulation so some compromises have to be
made. So the bigger the stiffness parameter is the less the fluid will compress under
stress, but the more substeps are needed for stable simulation.
** Viscosity - how much internal friction there is in the fluid. Large viscosities also
smooth out instabilities, so less viscous fluids again need more substeps to remain
stable.
** Buoancy - with high buoancy low pressure areas inside the fluid start to rise against
gravity, and high pressure areas start to come down.
* In addition to these basic parameters there are separate advanced parameters that can
either be tweaked relative to the basic parameters (or particle size) or defined
independently.
** Repulsion - the stiffness parameter tries to keep the fluid density constant, but this
can lead to small clumps of particles, so the repulsion keeps the particles better
separated.
** Stiff viscosity - the normal viscosity only applies when particles are moving closer to
eachother to allow free flowing fluids. Stiff viscosity also applies smoothing to
particles that are moving away from eachother.
** Interaction radius - by default this is 4 * particle size.
** Rest density - by default this is a density that the particles have when they're packed
densely next to eachother.
** Spring rest length - by default this is 2 * particle size.
* There are also new options for 3d view particle coloring in the display panel to show
particle velocity and acceleration. These make it easier to see what's happening in the
fluid simulations, but can of course be used with other particles as well.
* Viscoelastic springs have some new options too. The plasticity can now be set to much
higher values for instant deletion of springs as the elastic limit is exeeded. In addition
to that there is an option to only create springs for a certain number of frames when a
particle is born. These options give new possibilities for breaking viscoelastic fluids.
New in the code:
* Most of the fluids code is now thread safe, so when particle dynamics go threaded there
will be a nice speed boost to fluids as well.
* Fluids now use a bvh-tree instead of a kd-tree for the neighbor lookups. The bvh-tree
implementation makes the code quite a bit cleaner and should also give a slight speed
boost to the simulation too.
* Previously only force fields were calculated with the different integration methods, but
now the fluid calculations are also done using the selected integration method, so there
are again more choices in effecting simulation accuracy and stability. This change also
included a nice cleanup of the whole particle integration code.
As the internals are pretty stirred up old particle fluid simulations will probably not
work correctly straight away, but with some tweaking the same level of control is still
available by not using the "relative versions" of the advanced parameters (by default these
are not used when loading old files).
* Removed some fields from struct SculptSession:
- Fields drawobject, projverts, and previous_r were completely
unused
- Field `ob' was really unnecessary, changed sculpt functions
to pass the object rather than the SculptSession
This removal of `ob' from SculptSession should should make it a little
easier to continue generalizing paint/sculpt functionality.
There should be no visible changes from cleanup.
type and data
Recoded Keyframe Smoothing operator to work better for continuous
curves (i.e. ones with monotonically increasing slopes in sections) as
opposed to hypothetically jagged ones.
The old method assumed that handles should be flat as otherwise, you'd
often get unsmooth curves just because you went and tilted some of the
handles for local extrema, causing some unkeyframed overshoots, which
also leads to changes in timing, which in turn often means unsmooth
motion. Hence, the code took advantage of this to do things with less
extra data.
However, now we have a proper "flatten handles" tool (under snap ->
horizontal) so this functionality is not needed in the general case
where it will lead to stair-stepping artifacts.
for sampling were not consistent, now the RNG is seeded per strand, and some
tweaks were done to make the jittered sampler cache return consistent sample
numbers for strands.
Pressing ESC on material icon preview changes made it stop updating
the icons. Was caused by default 'break' callback testing ESC and
not resetting it. Now it uses same break callback as other previews.
Double click in File Window also loaded a file on double-click at
a directory.
The operator for it was using a Macro mistakingly here. On any double
click, it selected the item first and then executed load.
By default, any double click action has to rely on the first click
being properly handled before. Simply removing this macro, and assign
the "exec operator" to the double-click event works as expected.
A double click on a directory then just shows the dir (because the exec
requires an 'active' file item).
One node update call (for nodes within group nodetrees) was using the
wrong nodetree (node-editor's nodetree, not the group) which meant
that the wrong RNA context for such nodes would get used, resulting in
errors when trying to keyframe such nodes.
Hopefully this is the last time I have to fix these bugs...
* Using the job system for physics baking is not yet in the near future, so here's some good old console based progress info to all point cache based physics baking.
* The info contains current total bake time, baking time for the current frame, and a simple estimate of completion time.
* The info is only shown if the estimated total time for the bake is higher than one minute, so quick bakes don't suffer any performance hits due to console printing.
does exit cancelling the translation (as expected), while RMB doesn't,
it works same as LMB
This should fix the remaining issues with marker tweak-grab mappings
(hardcoded for right-mouse for now)
The sequencer made it's own copy of the Markers keymap, which was
inconsistent with the rest of Blender, making things confusing to use.
I've removed these duplicate keymap entries, and also changed the
conflicting hotkeys for Metastrips.
Metastrips now use the same hotkeys that their NLA cousins use:
Shift-G to add, Alt-G to remove; These were chosen since in user-
terms, metastrips are more like "strip groups"
Node compositor crash:
When you very quickly unlink a socket *right* after connecting it
(within 0.1 second), the compositor started with a noodle without
valid connections, which was not foreseen to happen.
More problems with Undo and Render Slots (Image editor)
- Undo storage for operator is now back, but only when new
buffers were added (not when viewing existing)
- A real bug: On undo/redo, the stored buffers were never
retrieved, but always freed entirely.
Note however that when you undo back to a state before you
rendered (or added slots), the render buffers that didn't
exist back then also get freed. A redo doesn't bring it back.
ASCII lookup table wasn't being freed when when the rest of the glyphs were.
Also found own bug where mesh stat text was using wrong sized array (char[3] as char[4]).
Very bad crashing in using "Not AA fonts" and drawing length info
on EditMesh. This uses the call BLF_draw_ascii() which apparently
corruptes then badly. Disabled the call for now, and wait for fix.
In the course of testing, added proper strlen storage for draw
in this code.
Added strlen
- Tooltip-hanging fix made the 'active' button be removed when mouse
enters another subwindow. Caused by commit of a week ago.
- Reverted the eventsystem change that sets 'active subwindow' before
it calls modal handlers. This made editors become active on using
menus, buttons, or other modal ops.
Side effect is that for transform operators called via toolbar, the
overlay-extra draw with helplines now gets skipped. Will check on
good ways for it.
Option for tagging creases (Ctrl+RMB) to also re-unwrap the mesh.
In 2.42 this could be done by setting rt==8 (very hidden), now its a little less hidden (in the toolbar).
change how the extraction commands are presented (old BUILD_TYPE was empty all the time here) and
ensure also that python/lib subdir under versioned dir gets actually created.
Submitted by Alexander Kuznetsov
Fixes [#25279] Shift-Numpad Combinations fail to align view to selected
and addresses [#26328] Blender uses global keyboard message hook which hurts system responsiveness on Windows
A whole new way of handling keyboard input improves greatly both code readability and event handling. Thanks for the great patch, Alexander!
Hanging Tooltips solved!
It appeared to be that an active button remained in that state when
another region/editor became active. It then kept the button-activate
state, and therefore also the optional tooltip.
This only happened on fast moves, when a mousemove event was not passed
on anymore to the previously active subwindow.
It has been solved with a new notifier (SWINACTIVE), which gets sent on
new active regions. The screen listener then calls uiFreeActiveButtons()
to find out if buttons were still active somewhere else.
* NLA timing was only applied to fcurve keys, but not handles, so strange things happened.
* This time tweaking was missed in selections too, so fcurve handles couldn't be selected properly either if the NLA strip was moved from frame 1.
* Grid distribution isn't really suited for hair, so this is now disabled.
* Setting a jittered distribution with particles/face = 1 now creates particles on the center of faces.
* Quite a bit of cleanup of the whole particle distribution code.
Reported by David Roy
When <instance_node>s where read, their transformation matrix got overwritten with the transform matrix
of their own node, not taking into account the parent node transformation. Instead of doing
that we now get the parent node transformation matrix and apply it to its own, and prevent
caller from overwriting this new transformation matrix.
Submitted by David Roy
Multiple nodes can reference the same geometry, and specify the same materials. This lead
to the import code overwriting material mappings of faces in a destructive way. Instead of
just writing the material bindings always we now keep book of what geometry+material mapping
we've already handled.
big spacing. Not sure if it's the best solution or it should be fixed when
calculating new areas.
It'll be cool if somebody else familiar with this area will check this.
* Tsk! Particle mirroring was working fine for the case where subsurf modifier was before particles (orcos are transformed in this case), but not the other way around (unnecessary inverse transform of orcos).
* Not really satisfied with having to check for the CD_ORIGINDEX layer, so if Brecht or somebody else knows a better way then please change this :)
Patch from Miika Hämäläinen.
The old Material "Only Shadow" used an ancient 'best guess'
formula using Lamp Distance and some averaging for converting
shadow values to alpha.
A couple of bug reporters already complained about the not
very predictable renders. Miika fixed this by adding two
new options, to only give the true shadow factor exclusively,
or to give a result including light intensity values.
More info:
http://projects.blender.org/tracker/index.php?func=detail&aid=26413&group_id=9&atid=127
Submitted by Filiciss Muhgue
Selecting COLLADA export after starting blender into default scene Blender would crash, due to misreading the path string (no file saved yet, so it was zero).
found crash while changing operator string size.
Shrinking arrays never worked right.
rather then "newlen * sizeof(...)", it would memcpy "newlen * oldlen * sizeof(...)" which always goes over the array bounds.
---------------------------------------------------------------------------
The Rasterizer code was relying in the values defined on TF_ DNA files. I'm working in the recode of TexFace, bringing the options to the material panel and ran into those cases. They are hard to spot and add a lot of the "magic" effect to the code. Hardcoded values are at least easy to spot. We (still) have a few defines duplicated, relying on each other (a flag previously defined in the code is checked later on but using a different define (although with same value. (e.g. TF_BMFONT and RAS_RENDER_3DPOLYGON_TEXT). It's hell =)
I'm adding some comments to help on that.
Things will be revamped anyways, but it's nice to keep the code a bit more coherent before the real feature commit. That's all, thanks for listening.
Using Marker menu (dopesheet) didn't work, the operators themselves
were checking Y coordinate of event.
Handlers also support boundbox checks. For this case it needed a bit
special handling. But works :)
Arrows-move-cursor is back!
It now works for any running modal operator that doesn't handle own
arrow keys. Might need to become more restricted though, some modal
ops don't need it. Want to investigate that still where conflicts are.
we cant ensure that a requested buffer can be allocated so report opengl errors when failing to allocate the buffer (rather then printing to console).
this is common enough and generic error isn't too helpful to users.
Make edge crease additive rather then overwriting existing values.
There are other problems pointed out in this report but they are unrelated to solidify.
Not sure why proportional editing used to disable crazyspace corrections,
but from that time hwn it was done transform stuff was changed a lot.
I've made several tests (with file from report and sintel from durian
data files) and it worked nice.
use 'install' target rather then copying files with post-build commands.
Main change is that you need to run 'make install', which copies files to ./bin by default
unless WITH_INSTALL_PORTABLE is OFF, where CMAKE_INSTALL_PREFIX is used for a typical unix-system install.
Transform started with button in Toolbar was drawing the helper
line badly.
This needed two fixes:
- helper line now has poll() callback to check for correct region
- event system needs to set 'subwinactive' for modal handlers too
The latter might fix issues with cursor/overlay drawing in other
cases?
Shift+D in graph editor was an operator calling internally an operator.
Better is to make it a Macro, then Undos and Esc work nicely.
Note for API users: the operator "graph.duplicate" will now just
copy the selection and not run transform. Nicer too :)
- "need exec" flag was cleared in wrong tree for preview renders
(it should clear it in the copy, i did in original).
This fixes projection paint on images with previews open
- Previews for nodes were copied always, now only for previews
invoked by node editor itself.
* Render level derived mesh (needed for proper hair rendering) was only calculated before duplication code if the particle object was a level 0 duplicator, so particles creating objects inside a dupligroup didn't work.
* I really don't see the need for the level == 0 check, so I just removed it and after some quick tests can't see any negative effects. If the check is needed for some case though then a more complicated test is needed to handle all cases.
Blender crashed on repeatedly pressing "reload image" or "unpack" and
other image options, with a material/texture preview running.
Solved with a signal to stop preview render before doing the action.
Do not overwrite coord of vertices in mapped vertex array used for crazyspace
corrections. This should make stuff use position of mesh vertex after
deformation, not possible generated images of this vertices.
fix for crash when iterating over a collection which allocates the collection and frees on when finished.
The ability for BPy_StructRNA to hold a reference to other PyObject's was added to support this.
Previously the api just converted the collection to a list and got the iterator from the list to return.
This has the advantage that it uses minimal memory on large collections where before it would make an array.
Though the main reason for this change is to support a bugfix for collections which free memory when they are done, this currently crashes the python api since once the list is built, the data is freed which is used by the list items in some cases (dynamic enums for eg).
the old FH setting was blended with the other physics settings (friction and elastic)
Also in the Physics panel it was saying "Use Material Physics" but the button is only for Force Field.
Since I was here I decided to change the Constraint FH ui name from Fh to Force. I don't think users really understand what FH is (I for once don't).
Thanks to Carsten Wartmann for pointing that out.
in dope sheet/ graph editor leads to duplicated keys
The old hack using the transform "undostring" didn't work anymore, as
this wasn't set. Instead, I've added a special mode transform mode for
this that the duplicate operators can set to get this functionality.
-------------
Given the code perspective this is a hacky solution. However it's self-contained and documented one.
Ended having a talk with Ton and he agreed that this can acceptable in this case.
Zero chances of crashes (I've been using this myself for more than 6 months now).
Give it a shot. Although the code is +- for the user it's a really nice addition.
At the moment it makes more sense to put energy into the future Nodal Logic.
Yet, it's nice to give the current Logic Editor the love it deserves.
The Logic Bricks that use angles (Radar Sensor and Constraint Actuator) do so in degree.
Reverting for the time being. Ideally we want to make them to use radians internally (that requires at least a doversion and a convert from radians to degree in BGE Converter routine).
If someone feels like helping on that, please go ahead.
* The fcurve points changed their order when rotating, but the transform data wasn't updated properly for this change so some curve handle pointers got lost.
* Also found a second bug while fixing this: The fcurve handle type pointers weren't updated at all when the order changed, so some auto handles could turn into aligned handles even if the transform was canceled (no undo possible in this case!).
Right Mouse on area edges has menu to Split or Join. Works like 2.4.
Code needed cleanup and upgrade; operators were hardcoded tied to using
the area corner widgets only. In theory this is getting py ready even,
but that might need some additional testing. :)
The "Script" was added in the original rna commit but for no apparent reason the controller was renamed from the original 2.49 name.
Spotted/reverted in time.
- Added option "Fixed Texture" to the UI. Because of strange reason,
this feature was implemented but hidden from users.
Would be cool, if somebody familiar with 2d texture paiting check.
- Fixed some issues in existing code of fixed texture paiting.
It now handles brush radius and curve correct.
- Also fixed issue with paiting with texture from node tree - it used
to be painted with regular brush color instead of texture.
Paint strokes now can be mapped to any key. The operators now store the
event it was started with, so it ends with a release. Even hotkeys work
(while hold).
Second method for not having python crash blender on invalid access (ifdef'd out ATM, so no functional change).
This uses a weakref list per ID, and invalidates all members of that list when the ID is freed.
the list is not stores in the ID pointer but using a hash table since storing python in DNA data is not acceptable.
This is more correct then the previous method but shows down execution of scripts significantly since its always adding and removing from lists when data is created and freed.
Reported by Edwin Amsler
The function for extrapolating the template name with the frame number didn't properly NULL-terminate the
resulting string, this causing broken names and weird paths (if you were lucky - totally unusuable names
detected while stepping through the code).
- Linking logic buttons had draw glitch, caused by reading
data from free'd buttons
- Doing an undo with mouse-over Curve Widget caused crash or
corruption.
Both related to code that tries to keep the "active" button in
memory. On each redraw of UI, buttons are re-created, which makes
it all tricky... hope we're safer now.
Also: added back undo pushes on adding links in logic editor.
This uses pythons GC so its no overhead during runtime but makes removing ID's slower.
Commented definition 'USE_PYRNA_INVALIDATE_GC' for now, so no functional change.
* For rendering every object's inverse matrix "ob->imat" needs to
be in render view coordinates, but for drawing in 3d view it needs
to be in global coordinates. Originally (way back in historical
times) ob->imat was only used for rendering, but over the years
other uses came up too. Before 2.5 this wasn't a problem as
rendering was a totally blocking operation, but with the new
"interactive ui while rendering" the problems started. Basically
any update that redrew the 3d view while rendering (like rotating
the scene with mouse) updated the inverse matrix into global
coordinates thus invalidating it for the render, leading to all
kinds of strange discontinuities with textures and volumetrics.
* Problems were very easy to achieve using orco/object coordinates
for pretty much any textures (bump, point density, volume, sky),
for examples see bug reports 24906 and 25229. Render baking normals
in object coordinates was wrong most of the time too.
* Now there is a separate inverse matrix ob->imat_ren that's
calculated at the beginning of rendering and is used in all places
inside render code where it's needed. This way the original ob->imat
can change at will during rendering without causing problems.
- use own OrderedDictMini class, pythons collections.OrderedDict is overkill, 179 sloc. replaced with own, 11 lines.
- remove code which stored the class file & line per RNA subclass, this was useful but would raise its own exception every time to generate a stack trace to get the class info so we could use of the class failed to register. the class stores its module & name which can be enough to find where it was defined.
Finally, the Blender icon can be used for the splash screen menu entry, on own todo for almost 1 year!
Thanks to elubie for fixing this in the code. :)
* Moved ICON_BLENDER away from (0,0) spot in blenderbuttons, since ICON_BLENDER=0 define is conflicting ICON_NULL define as well as with logic checks for nonzero icon id.
* This solved bug where ICON_BLENDER can't be set from Python as well as when using new UI functions from within Blender.
- Invalid shape when deformation mesh was disabled befoe undoing
- Crash when deformation modifier was enabled before undoing
Still got problems with undoing multires changes when toggling modifiers.
From the tracker:::
Issues fixed:
- ConeTwist-constraint's params weren't making it to the CcdPhysicsEnvironment, also added Hinge's params.
- UI wasn't using angles where applicable.
- btHingeConstraint's constructor can create frame-matrices which don't align so the hinge doesn's start at 0 degree tilt.
This is an issue when setting limits.
Changes:
- UI: Hinge limits can be set (and disabled).
- UI: ConeTwist only has max-limits and only the twistX can be disabled
- PyApi via rna_constraint.c: added the functions limit_xyz_min, limit_xyz_max (for 6dof), limit_angle_xyz_min,
limit_angle_xyz_max (for 6dof), limit_angle_x_min, limit_angle_x_max (for hinge).
- PyApi: dropped python-function limit_cone_min.
.:. Extra:
UI Changes:
- renamed "RigidBody Joint" to "Rigid Boidy Joint"
- reorganized UI to conform with other parameters (e.g. Limit Rot)
- added dis/active all over the place :)
to sequential error by adopting a two-pass smoothing strategy instead.
This seems to give some minimal reductions in the amounts of unwanted
shrinkage that occurs
- Fixed problem where just trying to replace existing keyframes would
result in the intepolation set on that keyframe to get lost. This was
mostly an issue if trying to re-block some animation in the middle of
a shot, with the rest of the keys set to Bezier, but the first
keyframe in this new segment needing to be Constant so that we don't
get sloppy automatic interpolation in the way
- Hooked up Media-Play/Stop/Next/Prev controls to animation playback
and keyframe jumping functionality in default keymap in addition the
existing controls. I'm also considering whether to migrate Next/Prev
Keyframe key mappings off the Ctrl-PageUp/Down keys for a more
ergonomic option (i.e. shift <, shift >)
* More code cleanup.
When removing the narrow_ui feature, a lot of split functions were kept. Changed them to rows, which are more easily readable. Check out properties_data_lattice.py for a good example.
Two bugs in one:
- Lukas commit monday for new group/socket handling accidentally removed
to set socketype in stacks, which as used by (texture) nodes to detect
whether value, color or vector had to be read.
Result was that all texture nodes were rendering as B&W
- Old 2.5 bug: preview renders for texture nodes didn't call a
NodeEndExec function, which gave crashes on deleting nodes.
The change in interface_hanlers.c is only a comment to explain
how keymaps are being found.
markers.
This is useful for when working with lipsync shots, where you've used
markers for noting down key syllables and want to separate these out
into chunks to manage things better.
Layer height used to be controlled with brush radius, quite confusing decision.
Added new property for brushes - height for adjusting affectable brush height
(it could be not only layer height in the future).
the length check was running sequence checks on every number which would fail, small speedup by avoiding this.
should eventually get this working faster by reading once into an allocated array.
* This is basically a total rewrite of the edge split algorithm. The old one didn't handle tris at all and quads were cut wrong in some cases too with the addition of not handling uv coordinates at all.
* This new algorithm uses a flag system to categorize different splits and the identical but rotated cases in a similar way to how marching cubes indexes different cases.
* It cuts quads and tris and creates proper uv's for the new faces too.
* I also renamed the option to "edge cut" to differentiate if from the edge split modifier and added an option to override a uv-channel in the exploded mesh with particle age as x-coordinate so that the shrapnel can be faded out nicely etc.
Having 2 or more windows open, a render in another window made the
other (active) window fail badly on F11 key (show render).
Now things should go smoothly. If a render is being shown in
inactive windows, they get popped to the front nicely.
Invert Image Colors (RGB)
--
Invert Red Channel
Invert Green Channel
Invert Blue Channel
Invert Alpha Channel
mostly because of the recent changes in normalmap channels,
so users can adopt old bakes quickly.
though they might aswell prove useful in other situations.
was exporting normal maps with red and green channel inverted
relative to the geometry it actually exports.
This change makes blender export normal maps
which are very similar to most tools out there.
patch by Morten S. Mikkelsen
* In some cases fluid particles could be born at the exact same locations. Now these cases are just ignored and such particles don't effect each other.
note: BPY_class_validate() could come in handy later if we need to check classes for properties/functions but for now there is no point in keeping it in.
auto-ik updates are done by notifiers at the moment which are ignored when running bpy.ops.transform.translate(), so add an special_aftertrans_update() check for this case.
Now we have the old UI again, as suggested by several developers on the Mailing List.
At least for simple materials. For Node Materials we keep the new Pipeline panel.
This should make both sides happy. :)
Was trying to free audio data from sequencer strips that don't even have audio. Corrected the error in several ways so this will definitely not happen again :-)
cmake here.
Somehow in OSX a function call doesn't return but enters a next function.
Will find around if someone else can make an OSX preprocessed version
of this C file. With this commit things go fine, and it's a sane check
anyway.
* Tsk! Using sizeof(string) to determine string length works if the string is a char array, but not if it's a pointer to a char array!
* Now the fluid code uses the actual size of the string directly.
adding meshes in C does:
Add Empty Mesh -> Enter Editmode -> Create Mesh
while python does:
Add Generated Mesh -> Enter Editmode
problem with this is there is no empty undo state for undo-redo to use so it always gave a duplicate mesh on redo-ing.
workaround by adding an empty mesh, do an undo push, and join the generated mesh into the empty one.
this would be fixed if undo stack spanned modes.
X_Material's mScene member variable wasn't properly set on merge if it didn't have an mBlenderShader defined. This patch always makes sure it gets set on merge.
--- from the tracker
It's annoying that build_debian.sh downloads the source tree to create tarball every time we build a deb package.
fix_debian_rules_avoid_downloading.patch modifies debian/rules so that the tarball will be created from local repository if possible.
problem was on installing a file which was already installed, when the source and target paths match python deleted the file.
now check if the selected file is inside any of the addon search paths and abort with an error.
this also means we can have a version string like '2.56a-beta' without using buildinfo.
release/VERSION was only used by scons, NSIS installer.
Possibly helps to fix bug [#26062] too.
CMake build option WITH_PYTHON_MODULE, will build ./bin/bpy.so
This allows 'bpy' to be imported from python or other applications/IDE's which embed python, eg:
python -c "import bpy ; bpy.ops.render.render(write_still=True)"
This runs in background mode and has similar restrictions to running a script:
blender --background --python test.py
TODO:
- install to site-packages with blender scripts
- add support for imp.reload()
- Fix for crash with Keying Sets when a Keying Set path has no ID-
block to target
- Info window now indicates the types of the reports shown by
colouring their entries if the entries are not selected.
- WITH_OPENJPEG wasn't defined for creator.c with CMake.
- remove shadowed/redefined vars.
- remove some unused RNA report args.
- re-arrange IMB_FILE_TYPES so IRIS is not the first format tested, since its not very common test JPEG and PNG first.
New render output option "No Output", which renders without
forcing an editor to show an image. Nice for people who
prefer to setup composites with background image view.
Shader nodes: option "backface" failed for materials using
full oversampling.
The usage of shi->flippednorm in render code is heavily messed
up... also flipping vertexnormals themselves feels wrong.
Marked it in code with XXX for future fixups.
A very Bad Bug!
On every draw-object, a function free_old_images() was called which
was freeing "unused" images during renders/bakes
This was a left-over from 2.4x code, missed it altogether.
I'm sure this fix will solve a lot of render crashing... :)
* Adding play method to the device classes to play back a reader (not used yet, preparation for a later feature).
* Using a linear resampler in case SRC is disabled.
available on many keyboards these days, so that they can be used for
animation playback (giving more options over alt-a and alt-a ad-
infinitum).
Currently, this is Windows only as I don't have a Linux/Mac system to
test on (it should compile with both mingw and msvc, at least using
scons). Maintainers for those systems can probably easily add this in
once they find out the relevant mappings for those systems.
Option "Align view to selected" only aligned to the last activated
item in editmode. This whilst code comment even said "use alignment
from Manipulator in normal mode".
Now it does what it's (apparently) meant to do, and what brazilians
think it should! ;) It aligns to selected!
User Preferences
Using "no AA font" crashed, caused by commit to make static variables.
However, how this was ever meant to work... using such globals has to be
treated carefully. Fixed for now with XXX warning for Diego :)
Thanks to reporter Susanne H. to point at the right code!
* In windows the old button was tested to be equal to the new button (ui_but_equals_old()) even if their optype wasn't the same. Adding a check for optypes fixes all three reported issues.
* For some strange reason this didn't happen on other platforms.
- Bring back cursor set to PRESS event - block selection wouldn't
have correct start position;
- Undo stack push was missed in cursor_set operator;
- Remove unneeded cursor moving at set_selection operator;
- Fixed bug with scroll bar - it shouldn't use EVT_TWEAK;
There could be still small issues with selecting single character by
mouse (due to EVT_TWEAK threashold), but this operator is for block
selection, not single char. So shouldn't be big pain here.
Compositor/texture nodes: math node now allows to use pow() for
negative raising too, but only when that value is near-integer.
For other negative cases result is zero.
Patch provided by Aurel W
example of usage:
0) Game Properties: text (String) and log (Boolean=True)
1) Keyboard Sensor set to AllKeys with log as logging and text as Target
2) Expression Controller: text=="quit\n"
3) Game Actuator: Quit Game
[1] <-> [2] <-> [3] .:. this will quit the game when you write quit + Enter
4) Keyboard Sensor: set to Return
5) And Controller
6) Property Actuator: Assign text property to ""
[4] <-> [5] <-> [6] .:. this will reset the string everytime you press Enter
# # # # # # # # # # # # # # # # # # # # # # #
Since the change is in the InputParser.cpp it actually affects all the text
input fields in the Logic Editor. So for instance you can use it in the
assign Property Actuator.
# # # # # # # # # # # # # # # # # # # # # # #
Connect an expression controller: text="idclip\n" with an actuator to disable
the Collision of your walls and you can re-create Doom with only Logic Bricks (:
- Change cursor changing from mouse press to mouse click. A bit "delayed"
UI feedback, but otherwise we'll be unable to use double-click events here
- Change clinebacks to callbacks in comment for select_line operator.
Or it was supposed to be clinebacks?
- cursor_set which just sets cursor position immediatelly
- selection_set which could be used to select block of text
Now there're could be operators binded to mouse double-click event
Some reorg of modules/pages, start makesdna and makesrna.
In many places license block needs to be changed to not start with /**, because otherwise documentation will go weird.
Use shift-del to cut, ctrl-ins to copy and shift-ins to paste
Most of editors support this hotkeys and it's annoying when you
trying to use them in text space.
Annoyance was that operators that defined and __init__ function would need to do...
def __init__(self, another_self):
....
py/rna was calling the class directly with PyObject_Call() but needed to pass the pre-allocated object only so __init__() would run .
This works OK internally but becomes messy since __new__ and __init__ always get the same args there was no way to avoid a superfluous self argument to __init__.
WM_modal_tweak_exit() was making incorrect use of the user-pref option
"Release Confirms Transform", indicated by confused coder comment
(<quote>"XXX: WTH is this?"</quote>).
This manisfested when moving markers by just click-dragging and
existing marker, and having it "drop" whereever the mouse was released
regardless of the user-pref option. This was quite confusing as it was
inconsistent with the way that all other transforms worked when this
option is off, where you would usually start the transform (click-
drag), release the button, move around a bit, and then finally click
to end.
problem and crash
- It turns out we still need the "copyob" still, if for nothing other
than making sure that the unkeyed transforms can get restored. This
was removed originally as I thought that just reevaluating the
animation would work.
- Removed a buggy line of logic that was causing crashes when there
was no animation data. It's better to just assume that if animation
data exists, that something exists there.
- Make Duplicates Real was not clearing data such as the new animation
data or constraints.
I have no idea why this hasn't been implemented before (the DEFINES were even there already). But since the Ipo (Fcurve) Actuator supports it I don't see why Action and ShapeKey shouldn't.
More than a new feature or a bugfix this is actually a step forward into unifying them.
From the patch description:
"A new panel is proposed to bring togheter all the properties of a material that belong to the render pipeline level.
Such properties are currently not mixable with node materials, as nodes operate on a shader level."
Commiting this patch as approved in the sundy meeting.
Switching layers reveiled hidden objects in wrong positions, changes
are currently not being handled for hidden objects.
Only way to fix it is by completely update newly visible objects...
(Also fixed a typo and a compile warning)
Halo texture: disabling texture slot with [] button didn't work.
Note: halos only allow 1 texture slot, UI allows to add more. This
rna template stuff for texture slots I rather don't touch though...
Removed superfluous blend-mode change which was causing some weird
white lines to appear, especially when using some markers that don't
have any names.
* Particle system's are comparable to texture slots, which can
only exist within an id block. Particle settings on the other
hand are idblocks which should be pinnable just like textures.
* When particle settings are pinned only properties that make
sense without the actual particle system are shown in the
particle panel.
* Now texture pinning works consistently for all textures.
* The only problem is that there is no way to pin textures if there are
no textures in the context path, so it's for example impossible to pin
the texture of a displace modifier without adding a texture somewhere
else and selecting that texture first. I my opinion this is quite
counter productive so perhaps a dedicated "only textures" texture
context option is needed after all.
So, it turns out that dupliframes weren't that bad to restore... the
old version didn't do truly accurate transform freezing as it didn't
update ancestors too. However, as a modelling tool, this will probably
suffice.
* All the various index-related issues should finally be sorted now.
It seems you cannot just partially implement some of these active
index getter/setters...
* Standardised the call used by PoseLib operators to get the Object
from which they get the active PoseLib data from
* PoseLib operators which require some existing data to work now use a
poll() which checks for this
* Added back the operator used to make standard actions into ones
usable by PoseLib
* Added a dummy operator for the apply active pose button which really
just calls the same backend functions as "Browse Interactive", but
which has a nicer frontend (naming + description) for the purpose of
being used in this way
* Also, removed some useless code from here
- Added operator button to show the current pose. This runs on top of
the interactive pose browse functionality, so the tooltip may be a bit
obtuse...
- Specifying a pose for the Pose browse functionality to set now works
correctly. Previously, some old depsgraph hacks were actually
interferring with correct updating (only the armature updated
correctly, but deforms didn't occur)
- Fixed a case of accessing freed memory, which I'm surprised no
static checkers have reportedly picked up on yet, or that nobody has
really had issues with (probably due to low memory "turnaround" so
far)
r34883).
Full description:
When defining an operator button in the UI layout code, trying to set
the value for such an operator's enum properties, where said enum uses
a dynamically generated list of items (which depends on using context
info), will "fail". No context info will be passed to the callbacks
used to generate this list of items, as PROP_ENUM_NO_CONTEXT is still
set on the operator properties (it seems these will only get cleared
when the operator actually runs, which is far too late already for
this usage) so RNA_property_enum_items() will pass NULL instead of a
context pointer *even* when one exists!
I'm not sure of why we even need this flag. It seems to have caused a
few other rounds of problems already, from quick searches I did on
this matter...
This presents a UI from which PoseLibs can be assigned/removed from
Objects. From here, it is also possible to see the list of poses and
add/remove poses from this list.
Known Issues:
- [Py/RNA/Operators BUG ALERT!] If after immediately starting Blender
you try to remove a pose from the PoseLib using the UI buttons, you'll
get a an error the first time you do so (but not for subsequent
attempts). This seems to be caused by the "pose" enum (dynamically
generated) of the POSELIB_OT_pose_remove operator, which does not seem
to be getting initialised when the operator's exec gets called without
the invoke having been called previously
- Changing the active Pose Library still seems to be broken (to be
fixed soon)
Todos:
- Operator button to make the selected pose get shown in the 3d view
- Restore the "validate action" operator and add that to this panel
- Rename pose access
unawaliable state.
- Also changed dynamic space bar menu, so it now also wouldn't allow to
set brush stroke flags to unavaliable state.
Also tried to remove use_anchor, use_space, use_restore_mesh and
use_airbrush, so user would be unable to set flags to unavaliable state,
but it was quite difficult for non-sculpt paint modes. They don't support
all of stroke methods, so brush.stroke_method can't be used for them.
Keep this area unchanged for now.
first and last NURBS' control points (toggle cyclic flag)
- Show U/V selection menu only for surface objects. There's no
much logic in asking U/V direction for simple curves.
- Set active spline when two segments were joined.
- Disallow setting cyclic flag for single-point segment when this lonely
segment is selected
- Clear cyclic flag for single-pointed segments when finding segments which
can be merged
Custom properties are now supported by Pose Library. They will get
saved and restored correctly during previewing and pose adding.
To do this, I've changed PoseLib over from using the "LocRotScale"
Builtin Keying Set to the "Whole Character" one instead. One
consequence of this, is that now entire poses are stored in the Pose
Lib, irrespective of whether you only selected part of a pose to store
(i.e. facial controls only, or hand poses only). If this is enough of
a problem, I'll make a second Keying Set that does take selections
into account, and use that here instead.
intended as a standalone library for use in other applications that
want the same tangent space as Blender.
This also keeps blenkernel clean(er) from extra math functions.
popup appears, saving an extra click
I've separated out the "XXX"-'d event-adding-hack section from the
search-menu code into a separate API function (as recommended there).
This call is used to make sure that textboxes in popups can get
activated by default, to allow typing immediately.
the two files mikktspace.h and mikktspace.c. These are standalone files
which can be redistributed into any other application and regenerate the
same tangent spaces. The implementation is independent of the ordering
of faces and the vertex ordering of faces.
Very nasty UI code issue: since every button is re-defined on a
redraw, having UI redraws while using a button was not possible.
This was solved long ago by copying over data from previous button.
However, this fails when buttons have callbacks with its own (or
a parent button) pointer.
This bug reporter found crashes in draw-overlap UI mode, this
draws entire UI over for every menu redraws, making previous button
pointers invalid. (for triple buffer, the UI is not redrawn, only
the menus).
In general: all systems falling back to old swapbuffers would have
suffered some instability because of this.
Fix is that now the old button gets lifted out from the previous
list and inserted in the new list. Works fine, but needs some tests!
Also in this commit: TIFF endian switching not needed for 16 bits tiff.
brushes. They aren't using proxies, so deformation wouldn't be propagated
in sculpt_combine_proxies().
So, maye idea of getting rid of "extra" memory allocaiton/disposing was not
such cool due to it lead to such exception ways?
- rename rna collection structs Main prefix to BlendData: eg, MainObjects --> BlendDataObjects
- printing python collection now prints its type (when available)
- renamed shadowed vars in bpy_rna.c.
- when making functions static I also made debugging/test functions static, reverse and add definitions to headers instead.
left/right to current frame in dope sheet
This commit separates out this selection functionality out of the
click-selection operator into a separate operator, so that hotkeys and
menu entries can be assigned to it.
This is based on an idea+patch (#23738) submitted by Torsten Rupp
(rupp), though I've ultimately decided not to go with the suggested
implementation as I don't think this fits that well under the "column"
select operator.
Todo: Graph Editor support will be coming shortly...
Crash was caused by keeping active segment index even when this
segment had been deleted. Fixed this and also changed behaviour
of creating new CV for case nothing is selected: new segment with
BEZIER type would be created.
- fix mistake with grease pencil UI (&& was intended but & used).
- use (void) rather then () across _all_ blenders code.
- a few minor edits, don't shadow stack variables in roll calculation & avoid running memset() for VBO vertex map.
most local modifier,GPU,ImBuf and Interface functions are now static.
also fixed an error were the fluid modifier definition and the header didnt have the same number of args.
* New option to distribute particles in a hexagonal grid.
* This is much more stable for fluids than normal grid distribution and looks quite nice otherwise too :)
* Also some small scale code cleanup of grid distribution code.
- use NULL rather then 0 where possible (makes code & function calls more readable IMHO).
- set static variables and functions (exposed some unused vars/funcs).
- use func(void) rather then func() for definitions.
Cicrcle and border selection doesn't add entry to EditSelection. Use cycle
through vertices rather than EditSelection stack to find two selected vertices.
No slowdown for case two vertices are selected (use this cycle to clear temporary
flag too), minor slowdown for unsupported selections.
The changes here were breaking old rigs, for example the right eyeball
in the Sintel rig (which uses the Mirror Modifier and its vertex-group
mirroring functionality)
* Renamed the old split uv's animate option "time" to "age" and added a new option to change the used split frame by frame. These are good changes were suggested/implied by Hannu Hoffren over 3 years ago in his tutorial video! (oh my how time flies)
* Also cleaned up the billboard ui quite a bit. For example now the uv channels can be properly chosen from the existing channels.
* Particle duplis are now always created with render percentage if G.rendering is set.
* This is not yet a perfect solution (hair for example won't yet work correctly), but
it's good to have even partial functionality here until a proper way to handle this
is implemented.
* Greetings from farsthary: particle rotation is now taken into account for particle effector direction.
** This gives all kinds of new possibilities as he shows in his blog http://farsthary.wordpress.com/2011/02/08/vortex-particle-simple-tut/.
**The only modification I made to his patch was to use the actual rotated particle direction as the effector direction as this defaults to the particle velocity vector, so no actual new options are needed.
* I also added an "effector amount" setting for particle effectors so that only a part of the particles can be considered as effectors. This makes it possible to create simple "farsthary vortexes" with only one particle system.
* Also some tiny reorganization of the falloff min/max values for a nicer ui.
==========================
Removed limitation of armatured-only objects for sculpting -- now all
deformation modifiers are allowed in sculpt mode. Use crazyspace corrections
like from transformation modules was used to support all deformation modifiers.
Internal change: all crazyspace-related functions were noved to crazyspace.c
P.S. Brush could make quite unexpected deformation for meshes which are
deformed in specified way. Got patch for this and discussing with Brecht
if it's really needed or maybe it could be done in better way.
deformed PBVH to the base mesh.
It's not real bottleneck, but crazyspace corrections looks nicer now.
Real bottleneck is normals re-calculation, which calls plenty of
fsqrt's. We could avoid this for some modifiers (which don't use
normals), but such checking would make code less controllable.
- modifier code was using sizeof() without knowing the sizeof the array when clearing the modifier type array.
- use BLI_snprintf rather then sprintf where the size of the string is known.
- particle drawing code kept a reference to stack float values (not a problem at the moment but would crash if accessed later).
* Effecting particle properties with textures was possible in 2.49,
but not in 2.5 anymore.
* Now particles have their own textures (available in texture panel
for objects with particle systems), which are totally separate from
the material textures.
* Currently a basic set of particle properties is available for
texture control. Some others could still be added, but the whole
system is not intended as an "change anything with a texture" as
this kind of functionality will be provided with node particles in
the future much better.
* Combined with the previously added "particle texture coordinates"
this new functionality also solves the problem of animating particle
properties through the particle lifetime nicely.
* Currently the textures only use the intensity of the texture in
"multiply" blending mode, so in order for the textures to effect
a particle parameter there has to be a non-zero value defined for
the parameter in the particle settings. Other blend modes can be
added later if they're considered useful enough.
* External caches didn't load for smoke straight away. Smoke caches store all necessary info in every file, so no need to try to look for an info file.
would stop the entire menu from drawing if they raised an exception.
now print the exception and continue.
Also added a verbose argument for bpy.utils.(un)register_module() to help test whats being registered.
- KeyingSetInfo classes are now collected like Panels, Operators etc so bpy.utils.register_module() can be used.
- move bpy.types.register() to bpy.utils.register_class
after discussion over the mailing-list it seems that it really makes no sense to restrict that. Specially since we have the "Property" playback mode that usually will be a float.
Since there is no need for do_version or whatoever I don't see a problem. Any Thoughts? Maube to allow it only through the Python API?
This removes auto-registration, committed by Martin r30961.
Realize this is a contentious topic but Brecht and myself both would rather opt-in registration.
TODO:
- addons need updating.
- class list will be modified to use weakrefs (should have been done for existing system too).
- will move bpy.types.(un)register functions into bpy.utils.(un)register_class, currently including these functions in a type list is internally ugly, scripts which loop over types also need to check for these.
Good discovery this one.
It appeared that "pose mode" changes were not registered in our undo system.
That way the first operator you try to redo after exit posemode fails.
Adding a byte rect to float ImBuf was always freeing the mipmap levels.
Removed this convention since it crashes renders + image texture draw.
Proper ownership handling of ImBuf is high on the wish list :)
The backdrop zoom factor for new node-editor instances was not set
(i.e. was default initialised to 0). Now, this gets set to 1.0.
Also, set the property default in RNA to match this.
conections
It seems that some of the Outliner hacks used while building the tree
was causing problems, as Make Single User (and potentially other code
working with ID-data, specifically with the "newid" value there) was
making use of the variable used there for other purposes, leading to
memory corruption.
This bug also occurred in 2.4x, though when I tested there, it crashed
immediately.
Ton, you may want to double-check this bug!
tested that correcting invalid meshes works by generating random meshes and checking that only the first call to mesh.validate() makes changes.
found 2 bugs in mesh validation.
- face sorting array wasn't assigned correct indices.
- removing invalid edges used wrong comparison.
Recoded animation channel sorting code. In particular, the old code
didn't handle "islands" of selected items well (i.e. a chain of
several connected items in a row), with some of these cases having
unpredictable results.
There were also some bugs in the way some of the rearranging methods
worked, allowing some invalid operations to be performed. Some of
these probably triggered errors such as some channels getting stuck,
and so on.
from Shinsuke Irie (irie) with some minor edits.
Shinsuke's description from the tracker:
---
I have implemented GVFS framework support of blender-thumbnailer.py which allows some file managers like Nautilus and Thunar to show thumbnails in trash or network directories. If Python's gio module is available, the thumbnailer uses it to access to filesystems mounted via GVFS. This change shouldn't affect desktop environments other than GNOME and XFCE.
A function gvfs_open() in this patch is defined to solve a stupid incompatibility between Python file object and GIO Seekable object.
On Ubuntu 10.10, I confirmed thumbnails can be generated for file://, trash://, sftp://, and smb://.
detects...
- invalid vertex range for edges/faces
- duplicate indices in edge/face
- duplicate edges/faces in mesh
- missing edges data in faces
At the moment it doesn't correct errors, but eventually it will do this.
- Tidying up some inconsistent formatting
- Names of old IPO blocks are now included in the names used for new
actions. These are included after a "CDA:" prefix, (i.e. "_C_onverted
_D_ata _A_ction:"), which makes it easier to browse through these
actions later.
One-liner fix - a missing "OPERATOR_FINISHED" on the select operator
was causing problems renaming markers and potentially with other
operations too!
To find this bug, I added debug method to dump the list of markers to
console. This has revealed some troublesome things about the way
markers are organised, which IMO need to be addressed.
-> 2.50
Actionified ShapeKey IPO-blocks (i.e. "Shape Key Actions") would have
an action channel with the hardcoded name, "Shape", and this action
would be assigned to Object level (although ShapeKey blocks had their
own IPO-block slot, only Objects could have actions, so actionifying
ShapeKey IPO-blocks would wrap a ShapeKey block's IPO's to an Object-
level action).
Hence, the path conversions code would wrongly interpret this action
channel as referring to a Pose Channel instead, thus creating some
invalid paths with a 'pose.bones["Shape"]' prefix wrongly getting
tacked on. To ensure that the converted animation can work out of the
box, a 'data.shape_keys' prefix is now used instead so that these
actions can still be Object-rooted while still being able to correctly
control the Shape Keys. This is because there's no easy way to
identify and then shift such action from Object-level to ShapeKey-
level within the conversion code. The consequence though is that such
converted ShapeKey actions CAN ONLY BE USED THROUGH OBJECT LEVEL (i.e.
via Action NOT ShapeKey editor).
Secondly, the Action/ShapeKey editor version patching code has been
modified so that if a ShapeKey editor view was active when loading an
old 2.4x file, the action gets cleared from the view. This is because
of this didn't make semantic sense: the ShapeKey editor is for
ShapeKey-rooted actions, while the Action Editor is for Object-rooted
actions. The converted files though let Object-level actions be shown
in either one.
* Texture context was previously determined by going to the appropriate panel, for example "world panel -> texture panel" to access world textures. Additionally there was a separate button to access brush textures.
* Now the texture context can be selected directly through an expanded icon menu, which shows the available context options.
* This context selector is now at the top of the texture panel, but this could later be perhaps integrated to the context path somehow to be more intuitive.
RMB click in InfoWindow "console log" was missing call to set subwindow,
needed because it uses opengl picking to find active lines.
Result was that entire UI was glCleared, for some buffer swap methods.
This was reverted by Brecht r22541 because its inefficient but there is no way to work around this with povray at the moment so adding it back as optional args.
The poll() callbacks used in the Text Editor for scrolling and text-
block unlinking operators were too restrictive when the text block was
lib-data.
- Scrolling lib-linked texts is useful for just checking out parts of
the linked-in file that aren't visible on screen already. For example,
checking the specific rig that some UI panels script will work on, or
reading a "README.txt" linked in with notes on which layers various
controls are on. It should be fine that this temporarily modifies the
linked text-block (but for view-settings which will can be reset
later/on file load without any real consequences).
- Unlink operator should be able to be run, otherwise it would be very
difficult to remove linked files from a file (?)
- python 3.2 does 'import site' on startup which now tries to parse pyconfig.h which isn't copied. so for now just run without importing 'site', alternative would be to copy the header file for posix systems.
- cache PYTHON_VERSION variable so it can be set to 3.2, needed for copying python installation's other then 3.1.
mode with drivers
ChildOf constraints added using the PoseBone.constraints.new() method
via Python scripts instead of using the operator (this latter method
is still the preferred/recommended method) were not getting some
critical flags set, causing errors arising from space conversions
being performed more than once.
- negative ray casts would invert the offset direction.
this meant if positive and negative were enabled at once and the mesh was slightly inside & outside the object it wrapped,
the offset would be applied in opposite directions.
This way the offset is always along the vertex normal.
- allow negative offset from RNA, could be useful and no benefit to disable.
While testing Apply Visual transforms last night, I noticed that
setting a constraint to use its owner as its target was allowed and
didn't trigger any warnings. This clearly doesn't do any good and is
different from the old behaviour.
After talking with Ton Roosendaal we agreed on making those changes.
I wonder if we should go for "Start Game" instead of "From File". But Space is not a constraint here so be it.
- Armature Actuator now only shows the "Secondary Target" option when the Bone Constraint supports it (IK only now).
-- that may be overkill I don't know. It shouldn't slow down the UI considerably, so it should be fine. Easy to revert if needed though.
- renaming things such as "Start frame" to "Start Frame"
When using masks or other simple 3D elements in composites, doing
a layer re-rendering on a node is a bit clumsy all the time.
This commit does two things to help:
- new hotkey "Z" in node editor automatically finds render layer
that changed and re-renders it + composites
- option "Auto Render" does same, but then after every transform
edit in 3D window
The latter is experimental; real & proper system for this requires
full threaded render support (like previews). But it works!
Demo file:
http://download.blender.org/demo/test/auto_composite.blend
Important fix:
After any render, all the render layers were tagged "changed", which
caused any edit to first totally recomposte everthing. Now it only
composites changes.
Implementation notes
- DAG scene flush now sets 'changed' flags in render layer nodes
- Added notifier for 'transform finished' to trigger the update,
this is temporarily.
Drivers for Scene, World, and Compositing Nodes now "work" (well,
sort-of)! Previously they were strictly restricted to object-
accessible data only; now they can function across the board (give or
take some weak spots).
Although there is still no depsgraph support so that these properties
update properly when their source controls are changed (this will
probably require a lot more work), they can still update under other
circumstances (i.e. frame change and/or manual refresh flushing via
mouse movement, etc.)
As the depsgraph tagging support is lacking, these just get always
executed for now, which might potentially be quite sluggish, though it
is hoped that there are so few of these top-level datablocks with
drivers hooked up that this is barely an issue in practice. At least I
haven't noticed any substantial slowdowns for animation playback, so
it should probably be fine.
* There were a lot of settings in the particle panels that made no sense for simple hair and only cluttered up the ui.
* Now these settings are hidden by default unless "advanced" hair options are shown.
* Without advanced options the particle velocity controls are replaced by a simple "hair length" value, which actually corresponds to the grown hair length in blender units.
* Some hair effector options that are actually very useful were not shown in ui. These are now found in the "field weights" panel.
Moving this code to the same method as is used for Apply Visual
Transform (for bones case), which uses standardised decomposition
methods instead of the improvised ones used here before. This was in
an attempt to get IK chains Visual Keyed correctly, though this
doesn't seem to have improved the situation much.
Also, the decompositon for objects here now takes care of euler
rotation modes.
TODO: other rotation modes should get included here too...
Simplified the code for Apply Visual Transform as Pose.
This makes it work for IK chains now, since there aren't any
recalculations of the pose involved now, which were contributing to
the pose getting incorrect as the applying process went on, since the
IK chain was getting calculated differently with each successive
change to the chain.
Probably wouldn't cause a problem but manually editing these types through python could easily crash blender.
also changed cmake, stub-makefile default build dir to be lower case and leave out architecture string, easier for documentation.
Use ../build/linux/ rather then ../build/Linux_i686/
When using Pose Lib with a search string, if the search fails after
typing a character and then you try to backspace to get back to the
previous search string, the pose does not refresh until you type in
another valid letter such that the search works again.
Did a logic reshuffled in the get next preview pose function so that
this is dealt with properly
- [#25951] Hardcoded Keys in PoseLib modal operator register twice per
use
Now only the KM_PRESS events are handled. KM_RELEASE events for these
are ignored...
- While checking the above, I also found that the view-manipulation
while previewing functionality was broken. This was one of the
benefits of this UI approach over a fixed screenshot-only listview.
Fixed now :)
Fix in crashing displacement last friday, assumed wrongly that
the old 'new bump' failed for displacement too. It still did
work (somewhat). This restores the situation again.
- Bugfix #25937
Child-of constraint now behaves like regular parent-child
relationship when all options are set. This prevents the
errors that can happen when decomposing non-uniform matrices.
- Todo item
The area corner hotspots for splitting/merging were far too
narrow. Now it uses a circular distance to detect whether
the hotspot is active. Also cleaned up drawing code for it.
* After countless different bugs particles should now render correctly inside dupligroups.
* Only particles with metaball visualization are still problematic, this is mostly due to the ancient metaball code.
* I'll also add a test file for some of the situations, so that hopefully these cases stay fixed :)
- 'axis' arg was not coerced from a tuple like other args now do.
- 'axis' arg was being modified in-place (VERY BAD).
- also made new function matrix_3x3_as_4x4().
3D View:
In quad-view mode, the options to set back/bottom/left views were
blocked. Now they work again.
Note: the oparator polls don't have a check for properties... so it
cannot be finegrained here. Checks are inside code.
These should not have any effect on render results, except in some cases with
you have overlapping faces, where the noise seems to be slightly reduced.
There are some performance improvements, for simple scenes I wouldn't expect
more than 5-10% to be cut off the render time, for sintel scenes we got about
50% on average, that's with millions of polygons on intel quad cores. This
because memory access / cache misses were the main bottleneck for those scenes,
and the optimizations improve that.
Interal changes:
* Remove RE_raytrace.h, raytracer is now only used by render engine again.
* Split non-public parts rayobject.h into rayobject_internal.h, hopefully
makes it clearer how the API is used.
* Added rayintersection.h to contain some of the stuff from RE_raytrace.h
* Change Isect.vec/labda to Isect.dir/dist, previously vec was sometimes
normalized and sometimes not, confusing... now dir is always normalized
and dist contains the distance.
* Change VECCOPY and similar to BLI_math functions.
* Force inlining of auxiliary functions for ray-triangle/quad intersection,
helps a few percentages.
* Reorganize svbvh code so all the traversal functions are in one file
* Don't do test for root so that push_childs can be inlined
* Make shadow a template parameter so it doesn't need to be runtime checked
* Optimization in raytree building, was computing bounding boxes more often
than necessary.
* Leave out logf() factor in SAH, makes tree build quicker with no
noticeable influence on raytracing on performance?
* Set max childs to 4, simplifies traversal code a bit, but also seems
to help slightly in general.
* Store child pointers and child bb just as fixed arrays of size 4 in nodes,
nearly all nodes have this many children, so overall it actually reduces
memory usage a bit and avoids a pointer indirection.
Patch #25901 by Tobias Oelgarte.
Bone transformations would be converted back and forth between different
representations when changing modes, which due to numerical errors could
lead to bone transformations slowly changing as you edit the armature.
Now the editmode head, tail and roll values are stored in bones and used
directly when entering edit mode. Head and tail were already there but
now we ensure they are the exact same value, roll was not yet there, so
we have a version patch for it.
The sub version was incremented to 1 for the version patch.
Using displacement with new bump caused NaN values or crashes.
Since displace code doesn't provide derivatives, it now falls
back to "old bump" by default.
- comparing eulers was ignoring the order.
- printing Euler()'s now prints the order too.
- un-orderable types (all except for Vector's), were not raising an exception when compared with >=, >, <, <=.
each accept Euler/Quaternion/Matrix types.
eg:
Euler.rotate(Quaternion(axis, angle))
Vector.rotate(Euler((pi/2, 0, 0)))
matrix.resize_4x4() and euler.make_compatible() were still returning an instance of themselves, now return None.
1) damping factor is not percentage (it's in frames, who would know that)
2) angles as degrees (cleaning the tooltips for those too)
3) setting Direction to None when changing from Loc and Fh Consts to Ori - negative axis is not supported in Ori Const. Buggy since ever (2.49).
Manually disable jitter usage for anchored and drag dot brush stroke metdhods.
Jitter slider is hidden in UI for this strokes so users can't set it to 0 by hand
and even if this slider would be visible in UI jitter gives wierd result for
this stroke methods.
* Particles that aren't shown are now actually deleted (huge memory savings for flat objects).
* Grid distribution for flat objects is now done on the surface object surface without offset.
* Invert grid option wasn't in ui and it didn't work for non-volume grids.
* New parameter to randomize the grid point locations.
* Resolution soft/hard limits changed to even 50/250.
reported by Micael Dias (and many others, see duplicates list)
On closing the first ("main") Blender window the very first OpenGL context
got deleted too. This context needs to be retained, since we share quite a
bit of OpenGL data through it to the newly created contexts (new windows).
Thanks to Ton Roosendaal for thinking out loud while trying to figure out
what the actual cause was.
After discussions with ZanQdo, it was agreed that the current workflow
for making a pose symmetrical was a bit too cumbersome, especially
when auto-keying was enabled, requiring pasting the flipped pose on
another frame so that the changes could be merged back in without
overwriting the "good" half of the rig.
This option for the Paste Pose operator makes things easier, by adding
an option which will make the pose only get pasted on to selected
bones instead of overriding the entire pose. By default this option is
turned off, but can be easily enabled either from the toolbar
(operator properties) or through the F6 popup.
The intended workflow with this option for making a rig symmetrical is
now:
1) Copy pose
2) Select "bad" bones
3) Paste Flipped
4) Enable "On Selected Only" for the operator
If there is sufficient interest, this option can even be enabled by
default. But, we'll see about that later
Scene (Toolsettings, i.e. alongside "layered" option for using NLA
while doing auto-keying)
This option makes all Auto-Keying operations use the active Keying Set
to carry out keyframing operations instead of picking and choosing
their own Keying Sets to use, thus cutting down on the number of
unwanted keys.
Warning: if the older userpref option was enabled in an old
startup.blend, it may be difficult to turn this option off.
mode for case there was no animation at object before enter editmode
- Fixed memory corruption for case of fcurve groups
- Fixed bug with incorrect re-naming "spline[%d].smth" pathes
TODO:
- Also update drivers' curves
- Enable I-key menu and autokeying
Will do this after consultation with Aligorith, shouldn't harm atm.
Render bug in spothalo (weird horizontal line).
Appeared to be precision error... even for doubles,
before doing division a check for FLT_EPSILON works
better than DBL_EPSILON (if there's an alternative).
* Tracing objects with IOR < 1.0 (like air bubbles under water) wasn't working correctly as a refraction was always assumed to be the first thing that happens for transparent materials.
* This fix is ok, but the fact that the internal renderer is not a physically based one is starting to show, as for example blurred reflections in this case are not really possible nicely without some slightly heavier modifications to the ray code.
* Also some cleaned up logic and better comments for my previous total internal reflection commit.
* [#25638] 'Insufficient thread locking' for sounds
- Actually a workaround for an msvc bug, msvc STL containers are buggy
* [#25922] Sound does not play in BlenderPlayer(r34579)
- Windows OpenAL doesn't seem to have clear context error state on initialising
Running 'make' starts an out of source build with CMake in ../build/OS_CPU, creating the dir and configuring CMake if necessary.
So casual users can just run this without changing dirs:
'svn up ; make'
* This was wrong since 2006!
* Raytrace code assumed refractions to go "air -> glass -> air -> glass -> air.." so actually only the first total internal reflection was calculated correctly, but everything after that was wrong.
* Now after a total internal reflection there needs to be an actual refraction before the ray escapes the "glass" object.
Marker selection operators now use a special poll() callback which
checks that there are some markers before trying to run them. This
means that when there are no markers, the full screen range is
available, instead of just those areas above the region masked off for
markers which can be quite slim when trying to save space by only
showing the summary channel.
- use BLI_math functions for removing rotations from objects and pose channels.
- add unit_axis_angle() to avoid setting the Y axis inline anywhere rotation needs removing.
Migrating "redraws" settings from TimeLine view data to per Screen.
The options are now still shown in the TimeLine "Playback" menu
though.
This means that whatever redraw settings you set in a TimeLine editor
will be used throughout a screen (i.e. editor layout) to determine
which editors will get updated during playback, instead of only
certain editors doing certain things at vague times.
---
Also, I moved some version patches pre 2.56 version bump into a
version-check for 2.56. These must've been missed when doing the
release...
While animating, I realised that actually the 'only selected' and
'include hidden' DopeSheet filtering options are also useful in the
Action Editor, especially while tweaking the animation for some
characters, where you'd like to focus only on some of the character's
control (i.e. eyelid tweaks, hand tweaks, etc.).
The other DopeSheet filtering options aren't so relevant here, so I've
excluded them from this.
Rotation brush used to be crazy because it depends on normal of area
under brush, which is used as axis to rotate vertexes around and
this normal used to change during stroke.
Calculate normal from original mesh state for rotation tool too.
* Only the first object (in scene object order) that instanced a dupligroup was rendered properly, because particle instances were checked for only after creating the render object (and even this check was done wrong).
* Now the actual render object is created only after both object and particle instances have been checked.
Info header: option to switch scenes or add new ones has to be hidden
for "full screen" mode. Current full-screen code doesn't allow to
browse screens or switch scenes nicely.
This comes in handy for testing importers against 100's of files, quickly showing breakages and easier to setup then unit tests.
Example usage:
blender.bin --background --python source/tests/batch_import.py -- \
--operator="bpy.ops.import_scene.obj" \
--path="/data/testfiles/obj" \
--match="*.obj" \
--start=0 --end=50 \
--save_path="/tmp/test"
Also found my name was spelt wrong in some places :)
calling
bpy.ops.wm.read_factory_settings()
... would clear a scripts namespace if running directly, not in a module.
Fix by backing up and restoring the __main__ module.
Also found BKE_reportf wasnt printing all reports in background mode as BKE_report() was doing.
While doing some animation work over the past few days, I found that
it would be useful to be able to tag keyframes as a another type yet:
"jitter" for fill-in secondary motion (i.e. blinks/hand-keyed camera
shake). Of course, what you really use these for is up to you, but
this is one more option when you need it :)
- "CONSTRAINT_IK_AUTO" flag for targetless IK constraints was being
set in the wrong place. This is for the IK constraint data's flag, not
the generic constraint's flag
- Converting stack var "targetless" from type bConstraint to
bKinematicConstraint (i.e. constraint baseclass -> specialised data),
since it was only used in one place with a cast used there.
- When using targetless IK with no specified chain length, bone
rotation locks are taken into account too, saving a bit of extra setup
work
- bpy.data.*.load() functions were only accepting UTF-8 paths.
- rna functions/properties now accept byte values rather then strings for file paths.
- bpy.path.resolve_ncase now supports byte objects.
- operators which reload G.main would crash blender if called from python and then accessed bpy.data.*
- WM_read_homefile_exec was setting the contexts Scene to NULL as a signal for the event system, this didnt work in background mode, crashing when property update functions expected scene to be set.
- now writing to RNA is disabled when inside render() call.
- disallow calling operators when writes are disabled.
Rendering runs in a thread so running operators from the thread is not safe unless rendering becomes a blocking operator again.
Used a crazyspace approach (like in edit mode), but only modifiers with
deformMatricies are allowed atm (currently shapekeys and armature modifiers only).
All the rest modifiers had an warning message that they aren't applied because
of sculpt mode. Deformation of multires is also unsupported.
With all this restictions users will always see the actual "layer" (or maybe
mesh state would be more correct word) they are sculpting on.
Internal changes:
- All modifiers could have deformMatricies callback (the same as deformMatriciesEM but
for non-edit mode usage)
- Added function to build crazyspace for sculpting (sculpt_get_deform_matrices), but it
could be generalized for usage in other painting modes (particle edit mode, i.e)
Todo:
- Implement crazyspace correction to support all kinds of deformation modifiers
- Maybe deformation of multires isn't so difficult?
- And maybe we could avoid extra bad-level-stub for ED_sculpt_modifiers_changed
without code duplicating?
Area split error: if the first split position was exactly aligned
with another 'edge' it merged the edges, causing the subdivision
layout to go haywire. Only happens in rare occasions, good find
this report :)
* Added Backdrop Panel inside the "Properties" area.
* Hard limit for the "zoom" value. Negative values doesn't make sense.
* Added a missing notifier for the Node backdrop move operator.
This was a two-part bug: a user error + API error.
* User Error: before calling bpy.ops.constraint.childof_set_inverse()
for a constraint defined on a bone, you firstly need to explicitly
make that bone the active bone. To do that, you do
armature.bones.active = posebone.bone # or something similar
* API Error: active bone setting was a bit too strict. It only allows
setting the active bone if the new bone comes from the same armature,
but was overlooking the fact that RNA pointers may have been created
through the object using the armature instead.
In 2.49 we had this implemented in the logic_windows.c. I think I skipped this when porting the new UI. I probably didn't know how to do it back then or was waiting for something. Anyhoo, it's all good now.
For the records, the Logic Bricks that I haven't gone over to make sure everything is 100% are:
Filter 2D, Constraints and Animation (action, shapekey, fcurve)
I'm going over them this week, so the Logic Bricks are close to receive a "revisited review check-up" ISO 9000 :)
New bump code: option "Texture space" only works for image textures.
It appears this option is some kind of "fake normal map" actually.
Needs nicer name... maybe "Image space".
Now it ignores the option when no image texture is used.
UI bug: when a button has an open menu, the menu closed on any
mouse-over of other buttons in this panel. That's not too bad,
but it didn't check for whether the mouse was already inside the
menu itself (respecting safety region).
The bug showed error on zoomed in UI, using FPS presets, in case
the menu-button was drawing aligned with other buttons. A real
boundary case... :)
.blend1 etc backups.
Proves again that lazy coders only make bad code :)
Implementation note:
The filewindow now recoginizes .blend version backups as
a special type, so filtering for .blend files themselves
ignores it. However, they're recognized correctly as valid
.blend files, and draw an icon as .blend file when filtering
is off. Can become a distinct icon if we want...
older GLSL versions (< 1.3)
Thanks Matthew M:
- adding mat3 from ma4 function
- removal of transpose()
And I've hacked in myself a textureSize() replacement, the image
size gets passed on to function now.
oldbump -> original
newbump -> compatible
*new* -> default (3tap)
*new* -> best quality (5tap)
the latter two have an option to apply bumpmapping in
viewspace - much like displacement mapping
objectspace - default (scales with the object)
texturespace - much like normal mapping (scales)
were getting formed wrongly
Although the RNA paths for the custom properties could get evaluated
correctly, keyframe status highlights in buttons didn't always work
correctly, and would lead to a duplicate F-Curve for the same setting
getting created.
This commit introduces a new Keying Set: "Whole Character", which is
specially designed for character animators blocking out their
animation. It should make animating with rigs such as the Sintel rigs
(and other "mainstream" setups, though others may also work with a few
modifications) much easier.
It automatically determines which properties on every bone in the
active rig should be keyframed, avoiding an initial set up step where
properties may be missed, or non-animatable properties are also
needlessly keyframed. To do this, it relies on several rules:
1) All bones in the armature, regardless of visibility status are
considered, so that hiding some layers on some keyframes then
keyframing them later won't create problems with earlier poses
changing
2) Bones starting with certain prefixes, i.e. DEF, MCH, VIS, etc. (the
full list is available in the code for this, and can be/is meant to be
modified by riggers in their own versions as they see fit), so that
some bones on hidden layers which shouldn't be seen by animators are
not keyframed
3) Locked transforms AREN'T keyframed
4) All custom properties ARE keyframed - currently this is the best we
can do, as it's hard to tell if they're needed or not, or even if
they're already driven.
* Caching of the start and end stills were just referencing the original imbuf (which got premultiplied after the caching), so as a result most of the time the premul was applied twice.
* Now the start and end stills are stored in the cache as duplicates of the original (non modified) imbuf.
disable getattr metaclass forwarding attributes from the python class, eg:
bpy.types.Scene.foo != bpy.types.Scene.bl_rna.properties['foo']
... This was convenient but too tricky to properly maintain with attribute assignment and attributes defined within the class.
avoid doubles in dir() by converting to a set and then back to a list.
they have some rotation-affecting constraint (i.e. Track To and Copy
Rotation) targetting, transforming the objects (directly, using GKEY
-> grab) becomes unreliable.
This was caused by a typo in some code checking for some
OB_NO_CONSTRAINTS under "flag" instead of "transflag"
The console was getting flooded with output like
g
i
i
...
all as a result of what looks like a debugging print. Whoever put this
in, you can get back your debugging prints by enabling BF_GHOST_DEBUG
in your local config :)
For Constraints, there's now a working "Local" Space for Objects
without parents. This is defined as relative to the object's rotated
set of axes which results from rotation that gets set via "rotation"
transform properties.
I'm not sure whether this different behaviour between parented and
unparented objects will be too confusing (and thus require separate
settings + a round of version patching), so I'll wait until we get
proper testing from experienced riggers first.
Outliner was drawing icons sometimes blurred. Happens when subpixel
positions vary. UI code doesn't suffer this, it makes own ortho for
it.
This fix re-uses an un-used v2d flag (V2D_PIXELOFS_X) to force a 2d
view on pixel exact positions. It's set for outliner only, I need
testing feedback first.
Submitted by Alexander Kuznetsov.
This adds the progress bar to the taskbar item of Blender for Windows 7.
Small change in version retrieval, otherwise as is.
only case that doesn't fully work yet is the one where quats/axis-
angle are converted to eulers first.
- Fixed incorrect comment
- Removed obsolete "armature_clear" var from clear origin operator.
This was some ancient stuff from 2.4x code that ended up getting
ported across...
- values were added to both the classes __dict__ as well as the internal StructRNA.
- made properties available from the type since this is where the python api assigns them:
>>> bpy.types.Scene.frame_start
<bpy_struct, IntProperty("frame_start")>
- rename RNA_struct_type_properties() -> RNA_struct_type_properties(), added RNA_struct_type_find_property()
Added OPTYPE_GRAB_POINTER flag to scrolling operator, changed a bit logic of
delta calculation and removed unused members from TextScroll structure.
Got issue with very-fast mouse moving (when mouse pointer could go out of window,
but it's general issue for all grab operator).
Was unable to check MOUSEPAN event (two-finger scroll). Please poke me if
there are any issues with this event.
Text space's scroll bar now uses quite the same strategy of handling
drag and click events as view2d.
I tried to share some code between this two spaces. but there was no big
luck with it: text sace uses it's own constants (like scroll bas size and offsets),
so the simpliest way was to define zone stuff for text space without sharing it
with view2d.
----------------------------------
While we are more and more moving towards enabling features in the Python API, it's also important to have Logic Bricks working with no scripts.
This option allows you to start the game with the mouse cursor on (it's on Render Buttons). The defalt is still off (no do_version needed here).
Support for physics is done by skiping the modifiers that
don't support mapping to original mesh. This mapping is
required to report the hit polygon to the application
by the rayCast() function.
Support for graphics is done by using the same render
function that blender uses for the 3D view. This guantees
equal result.
Limitation: there is still a known bug if all these conditions are met:
- Display list enabled
- Old tex face with a several textures mapped to the same material
- no armature or shape keys
- active modifiers
In this case, only a part of the mesh will be rendered
with the wrong texture. To avoid this bug, use the GLSL
materials or make sure to have 1 material=1 texture in
your old tex face objects.
* Before the different simulations all had a panel with an "add this" button making the whole tab look really messy. It also rarely makes sense to have more than one or two physics things enabled for a single object, so having all the panels in the tab just added a great deal of visual clutter.
* Now there is a single "enable physics for" panel at the top that allows for enable/disable of any simulation. All actual physics panels are hidden until a simulation is enabled.
* There was no "add" button for force fields before, but I added a toggle between "none" and "force" to unify the ui even further.
- move test -> tests, this name is used elsewhere in lib/tests.
- change interface code not to loop on a float value (clang warning), harmless, but with extreme cases an eternal loop would still be possible though unlikely.
- multiplying a 2D vector by a 3x3 or 4x4 matrix would use un-initialized memory, now throw an exception.
- use more variable length array BLI_math functions.
ParentActuator: sub options only available when mode==setParent + deactive ghost option when compound is true (isntead of saying that in the tooltip)
PropertyActuator: I never knew that you could pass a property name to the Add and Assign options of the Property Actuator. Where it's by design or a side effect is cool. And a tooltip explaining that does help.
* Argh, particles tab was showing the whole "non applicable settings for fluid particles"-galore as the particle type "fluid" can't be checked from rna using the settings type value. Now the ui is a lot cleaner and only settings that actually effect the fluid particles are shown.
"Modal cursors" now remain visible, until reset. These cursors
were reset already on edges. Example: 3d view toolbar, grease
pencil. Do note, only use WM_cursor_modal() on real modal tools,
and restore it at end.
* Unpack menu option "Remove Pack" didn't get a pointer to the image to be unpacked. This worked from image view by using the active image, but not from texture panel.
* The whole case of lib linking and pointcaches is not very well defined currently, but this fix sets the behavior of sb to the same as other physics currently.
* A proper fix will be easy to implement after a good physics baking ui is added.
* fix for armature sensor (Status Changed doesn't use any value to evaluate itself) + renamed "Test Type" to "Test"
* visibility tooltip was wrong - now we can change visibility from the physic buttons instead of the outliner
* bonus * when you change the visibility from the physics panel the icon in the outliner wasn't changing
vgroup functions were mixing up active group and one passed as an argument.
also made other changes.
- removed superfluous call to defvert_find_index() in vgroup_delete_object_mode(), was also doing unnecessary NULL check on each loop.
- remove paranoid NULL check from ED_vgroup_vert_remove, callers all check for valid 'ob'
The plane which defined the shear had the factor applied to each axis equally.
This meant that the shear for any 3x3 or 4x4 matrix would be diagonal on the positive values of each axis.
Only being able to create diagonal shear matrices seems stupid, now take a pair of floats for the shear factor corresponding to the plane axis values, so its possible to shear on only one axis of the plane.
Will need to write full docs on this on the wiki.
basic info.
- 21 tests, OBJ/3DS/X3D/FBX, 3 tests per format import export. STL, PLY, BVH are TODO.
- uses files in ../lib/tests (checkout separate)
- run with CMake Makefiles "make test" or "ctest"
- currently checks against basic MD5 hash on scene import and file MD5 hash on export (realize this wont work predictably on binary formats *TODO*).
- currently uses a generic script for all tests with arguments to specify command to run, expected output, testing method, files to check against etc.
Has already proved useful, found a number of bugs in import export and some in blender too.
It seems that rna_def_effect() is missing a rna_def_input(), resulting
in:
rna_uiItemR: property not found:
EffectSequence.animation_offset_start
rna_uiItemR: property not found: EffectSequence.animation_offset_end
in the console. I'm not sure whether these settings are used for
"effect" strips or not, but I'll leave that to sequencer guys to
figure out if/when they get a report about this.
Some window manager can set default value of this to be the
screen size, so running blender with -p or --window-geometry
don't work with value bigger than that.
This commit try to "avoid" the bug #25709, but at the end
depend on the window manager, so maybe work or maybe not.
Fixed one of oldest annoyances in Blender: the text input button!
It always behaved stupid when you had clipped text in a button.
Now while using arrows the cursor will move as expected, and only
internally shift contents when cursor reaches edge of button.
- Swap strips operator crashed when it encountered an empty track
- Duplicating NLA strips was renaming the wrong strips - the old ones
would always get renamed, which was annoying
Added because CTest has no convenient way to set environment vars for commands it runs.
--env-system-config -> BLENDER_SYSTEM_CONFIG
--env-system-datafiles -> BLENDER_SYSTEM_DATAFILES
--env-system-scripts -> BLENDER_SYSTEM_SCRIPTS
--env-system-plugins -> BLENDER_SYSTEM_PLUGINS
--env-system-python -> BLENDER_SYSTEM_PYTHON
Report was that move-to-layer menu failed. The real cause was
more complex; had to dive deep in the dungeons of the interface
code that handled undos and operators. Found several issues:
- popup menus (like redo operator, color picker) executed again
on a mouse-exit
- far too many buttons were sending undo pushes; even worse, in
the operator redo-panel each button action was pushed twice
- in case operator redo-buttons have own callbacks (like layer
buttons) the redo wasn't working
- layerbutton menu was called without creating a proper undo/redo
case
Things should all work smoother now!
On todo:
- better definition and handling of all versions for operator menus
(four types now, not fun)
also: make operator "do" menu, which on first action does operator
and then switches to redo-ing
- bring back Undo menu, to list the undo stack and jump in it.
Although strictly-speaking, this only applies to a certain type of NLA
Strip (there are 3 types), but since only the classic "Action Clip"
gets much attention, it's probably worth exposing this here too to
make batch editing of strips a bit easier.
1) Added a new operator to swap the order of strips within a track
(Alt-F).
This makes it possible to select two strips (or more precisely, two
islands of consecutive + selected strips) in a single track and change
the order in which the appear without needing a extra tracks to
perform the move through.
As usual, the non-overlapping rules apply, so there may be some cases
where swapping in this way is not possible without adjusting the
intermediate strips first manually. Otherwise, everything just gets
too tricky to manage deciding what adjustments should be done to the
obstructing strips to make a fit.
2) Freeing meta-strips didn't free their local data properly (i.e.
modifiers they may have had).
3) Adding strips to tracks, where the endframes for the strips
overlapped would cause problems with incorrect ordering of strips. I
still need to double-check whether evaluation works ok in this case...
Since rev.2 (hans ftw) we have a strange situation where horizon color was being used at Convert time for the WorldInfo background color (and for the fog). However through the Python API only the Rasterizer background color was being updated. On top of that the KX_KetsjiEngine.cpp::SetBackGround was using the WorldInfo bgcolor when render mode was the potato one (TEXTURED). Bottomline, when in potato mode the glClearColor used was the original one in worldinfo, not the API updated one in Rasterized.
* When a simulation is baked no dynamics calculations should happen anymore (even outside of baked frame range) since these can be very time consuming and baked simulations are supposed to be fast!
enter weight paint, hopefully fully fixed this time
* The texture selector for brushes wasn't updating. Seems that preview
images have two sizes, small (icon) and big, but it was only
updating if the icon size was set to update. Now both are checked.
* Also changed the previewimage arrays to use the already-existing
PREVIEW_MIPMAPS define, makes it a little clearer what the arrays
are for.
- invalid dimension type could be passed without raising an error.
- negative dimensions could crash blender, now they raise errors.
- zero length dimension arg was not detected.
- floating point lengths were accepted, now only allow ints.
also comment unused vars.
Sketching Sessions.
Fix for secondary issue noted in this bug report. Some of the code for
the fancy 3d view sketching options needed to be moved around so that
there is a clearer "paint/sketch session" and "stroke" separation
between various stages. There's still some more stuff that needs to be
done here before I think it's totally cleaned up.
bring back nice shortening of filenames in thumbnail view, shortening at front and end is confusing.
- adds ... at the end to indicate that the filename has been shortened
- also cuts off at the end and not at both front and end now
Note: maybe this can be moved into button code or so in the future, but for now should work.
Compositor operators now have better poll check, so they won't
show up in operator search or hotkeys for other node views.
- backimage move, size
- read renderlayers
Three code fixes for 1 report. User experienced crashes while
painting on float buffer + having preview renders on.
- Texture Nodes: Image was re-allocated without using
proper thread lock
- Paint code: old convention to free the byte rect from
a float image as signal to re-create now is a proper
flag. This keeps image memory unchanged. Nice for render.
- Imbuf: call to make a byte rect from float was freeing
mipmaps unnecessary.
misleading (?)
"X-Axis Mirror" option didn't work in Pose Mode, so hiding this option
for now. Perhaps one day we could properly support this for all
operators there, but that is more of a long term todo...
- metaball tessellation functuion was calculating density when it didn't need to.
- image drawing was using a float as a loop counter, in extreme cases this could cause an infinite loop.
- remove/comment unused vars.
paint
A couple underlying issues:
* Paint icon was looking only at the object mode to determine what the
"current" mode is, but that gave problems when the object mode was
anything other than texpaint, but 2D image paint was turned on. Fix
was to also look at what space is being drawn, and only if it's in
the 3D view does it look at the ob mode.
* The brushes lists weren't getting filtered correctly in the same
case where 2D image paint was on but a different object mode is
enabled. Fixed by changing the brush rna poll to look at the paint
source, rather than the object mode.
Texture nodes: including output node in group crashed when it got
connected. Note that this feature is not supported to do anything
else than temporarily giving an output when editing node groups.
Texture nodes, the Image node didn't get initialized on load.
This caused the node to refuse accepting a new Image if a
file couldn't be found on saving.
- enabling/disabling no longer prints in the terminal unless in debug mode.
- remove 'header' struct from BLI_storage_types.h, from revision 2 and is not used.
- Add GCC property to guardedalloc to warn if the return value from allocation functions isn't used.
Quite silly fix, not sure if it could be smarter with current events/brushes design.
Use pressure_value from first brush step for brushes which don't support strokes -- thumb.
brush, brushes with anchored stroke method.
Should be fixed in nicer way after events redesigning.
P.S. Tried to place pressure saving into invaliants update fuunction, but it seens
that this function wouldn't know about pressure yet.
On using loop-selects, the active vertex got cleared, it could still
simply indicate one of the vertices of indicated edge to become active.
That keeps the Ui for vertex-groups drawing correct, for example.
Also made loop-ring select store active edge for tools, like the loop-select.
Added new option for applyong modifiers on splines' points. This moves
tesselation point and path would be affected by modifiers which are applied on
splines' points.
2D Editors using the standard translation transform tool no longer
have their values converted. For instance, it simply didn't make much
sense in such editors to be showing frame numbers/timing in terms of
meters/inches.
* When euler-rotation F-Curves had a single keyframe only, the view
would be artifically extended to fill up to 57 (this comes from the
radians to degrees calculations) due to a combination of the bounds-
finding function enforcing a minimum separation of 1 unit between
min/max. This has now been moved to the operator-level where it gets
applied AFTER these conversions have taken effect
* F-Curves with samples only (i.e. baked F-Curves) would be ignored by
these operators. Was caused by using a poll calback that only
considered whether there were keyframes. Hopefully this is sufficient;
otherwise a hybrid poll method will be needed.
Drivers created from the Properties Editor for Materials and Textures
will now be created on Object-level instead of on their owner
Material/Texture as for their keyframes.
The intention of this hack is to allow users to be able to easily set
up drivers for materials and textures. Without this hack, users would
have had to do this manually via the Datablocks editor (I've described
this method a few times in detail, though this still attracts
complaints), as the way the depsgraph works does not allow ID blocks
other than Objects and directly-linked Object data to be driven. As a
result, although this hack can be done for these two cases, there are
no workarounds possible for Scene and Scene-linked settings.
There are 2 issues that will be noticed with this approach:
1) There may be confusion over why the drivers are found under Object
level and not Material/Texture level.
2) Driver status will not be shown in the buttons, leading to attempts
to try to keyframe the properties directly later and subsequent
confusion when finding that that won't work.
However, these are the sacrifices we'll need to make to get easy-setup
working in the meantime until the proper fixes can be done.
#25554. As found by Daniel Lara (Pepeland) in #25554, the original
problems in #21825 all stemmed from the ChildOf constraint there
having an incorrect inverse set, which was altering the final results.
Now Auto-IK works nicely for old rigs again (i.e. FK arms on BBB rigs)
Crazy users! Who on earth would try undoing in the middle of an action
(here the crash occurred while a mouse button was still held)?!
(At least it wouldn't have been possible for those using the DKEY +
mousedrag method of invoking strokes though...)
- files blender exports (blender uses unusual array formatting).
- scene needed to be rotated on import.
- lamp spot size was half as big as it needed to be.
x3d export typo broke image export.
- rna array parsing was using PySequence_Size() in a loop, this can be slow to run so only call once.
- assigning a single value to a multi-dimensional array was missing type check.
- improve exception messages for rna array type errors.
- simplify vector slice assignment by using mathutils_array_parse(...)
Reporter saw weird fgons and edge creases on spin-mesh.
Appeared that the edge-flag copying code happened after
freeing edges. Already since May 2007 or so... weird!
Also in this commit a couple of simple cleanups.
The problem was flag-enums were being treated as regular enums, a default value of 0 was using the first enum item, whereas with flag enums we want to be able to use 0 as a default value to specify all flags are off.
* In some rare cases gzread has problems with the fluid files. This could be minor file corruption or some strange thread issue, but checking the amount of read bytes always after read seems to give a graceful way out.
Session
As the key combination for undo was unhandled by Grease Pencil
operator and allowed to execute, some of the lingering Grease Pencil
data would get corrupted by undo as some flags may still have been
set.
This commit attempts to fix.workaround this problem by catching undo
events, using the internal "delete last stroke" functionality to
emulate undo-like behaviour as expected but without the associated
risks. The underlying functionality used was already part of the
original 2.4 implementation, but was exposed via the GUI instead there
where it was less useful.
---
Other tweaks related to Grease Pencil:
1) Spacebar can be used to end Sketching Sessions too now
2) Grease Pencil animation editor now displays GP datablocks in light
blue (i.e. "sub-id") colours as per dopesheet instead of them being
presented like groups. This better reflects their true nature.
The tool-redo depends on a working undo system, so it can rewind
a step and then redo operator with new settings. When a user
disables undo, this won't work.
Now the properties for redo operator (toolbar, F6) will grey out
when a redo isn't possible.
now its basically usable, next step is to add menu item.
update for changes in mathutils api.
- fix radians/degrees usage
- correct matrix multiplication order
- primitives were being imported twice the size they should be.
- uv transformations were not working.
Problem was using va_list value more then once, fix by using va_copy().
Note, va_copy() is c99 spec but only alternative I can see is to turn BLI_dynstr_vappendf() into a macro which calls va_start/end inline.
also minor functional changes
- OBJECT_OT_make_links_data() type property is now assigned to the operator property (so popup menu can find it)
- removing BG image now returns cancelled if no image is removed.
Submitted by: Peter Tarasenko (pit)
This patch switches the screen real-estate given to the "open/close"
and "drag" zones for panels.
- Now, the entire header is used for resizing (open/close) panels
instead of just the triangle widget being used for this. This makes it
a larger target for clicking on, making it easier to do so (Fitt's Law
in action!)
- Dragging panels now is relegated to the "grabber" zone on the top-
right corner only now. This reflects the lesser importance of panel
reordering in the panel designs now over being able to open/close
them.
Submitted by: Dan Eicher (dna)
<quote>
Adds:
AnimData.nla_tracks.new(prev)
* (optional) prev -- add new track after this track
AnimData.nla_tracks.remove(track)
AnimData.nla_tracks.active(track)
* (optional) track -- track to set active
* returns active track
NlaTrack.strips.new(name, start, action)
NOTE: fails if the strip can't fit in the track as opposed to the
operator which will create a new track and add the strip to that.
* name -- name for new strip
* start -- start frame of new strip
* action -- action to assign to strip
NlaTrack.strips.remove(strip)
</quote>
---
I've resolved the issue (noted in the original patch) regarding the
validation of the created strip by creating and using a "dummy
AnimData" block to solve the missing dependencies.
when using the Numpad0 feature to reset properties to their default
values.
While this implementation here is not a full/proper implementation, as
you cannot truly specify a default value for some pointers that may
require something other than NULL (i.e. nothing), this should be good
enough for the vast majority of (editable) cases which are fine if set
to NULL.
This is most noticeable with the Active Keying Set field in the
TimeLine header, where it's now possible to simply use Numpad0 to
clear it instead of using a confusing click+backspace+enter dance to
do the same thing.
That primitives, which used ri crash blender, flips normals just after creation
and this normals flipping calls layers interpolation, but MDISPS layer
contains no data still.
Just added checking to layerInterp_mdisps.
"Show All Layers"). This has been mapped to Ctrl-Accentkey
If necessary, you can alter your keymaps so that this operator is
invoked with its "all" property disabled. This will only toggle the
first row (first 16) layers, which is useful in most rigs for only
enabling all the layers with rig controls and not showing the layers
with rig mechanics.
Raytracing didn't show soft shadow in reflections, nor did it do any
derivative even. Added a basic version for it in raytracer now, still
needs improvement on heavily curved surfaces. But it's better!
Examples:
Glass sphere, mirror cube and sphere, look how it ignores bump and shadow
http://www.blender.org/bf/derivative256.png
in svn now:
http://www.blender.org/bf/derivative-svn.png
When pressure was zero, a sculpt brush was still being executed
with step amount divided by zero, and thus entering eternal loop.
Maybe tablet-specific this but I wonder how this never got reported...
it easier to reset a rig to its default pose again
- Refactored clear pose operators to separate out the common parts,
and made sure that they all had descriptions
rotation range clamping used a while loop which would run forever when the value was so big subtracting a full revolution didnt change the value.
Solve by using fmod() and double precision angle.
(rna flags were not available when this was originally written)
action.fcurves.add(frame, val, 1,1,1) --> .add(frame, val, {'REPLACE', 'NEEDED', 'FAST'})
When creating new actions using the "new" button the Action Editor
header databrowse, or the NLA editor's "Animation Data" databrowse,
the existing action will get copied (if it exists) instead of an empty
action getting added everytime.
Apparently this behaviour is very good for being able to "version"
actions within a single .blend file (Bassam?)
that used markers. I might've missed a few still, but at least a few
more cases will work now
- Accidentally broke keyframe selection on group channels in gpencil
commit, after misreading a call name.
This commit restores some basic functionality for retiming Grease
Pencil sketches. Some of the functionality that existed before still
hasn't been restored (namely snap/mirror tools as well as copy+paste),
though it should be possible to use this for basic retiming and
sketch-frame management again.
- There's still a lot of work required to get this up to the standard
of the rest of the animation editor code, as some of this code was
originally just hacked in based on the old-style code.
- Work is already required to not have to directly access the main db
global to get the list of Grease Pencil datablocks to show, but that
can come along with pending cleanups of the filtering code.
Added Ctrl-T to Curves Tilt button in the toolbar (edit mode), currently was reporting Alt-S for tilt
(reported by Rickyx here http://www.kino3d.com/forum/viewtopic.php?f=21&t=8485&start=0)
Also added the label "Tilt" to the tilt transform in the Curve > Control points header submenu, currently was just "Transform"
This builtin Keying Set is supposed to insert keyframes for every
F-Curve that exists for the selected data (usually objects and/or
bones only). However, as coded, it was only useful for objects, since
it would just go through all the F-Curves for the object's action,
instead of just the F-Curves relevant to a selected bone.
Tweaked the code to make this case (and similar ones) hopefully work
better by default.
- report an error if adding a pose constraint with no active pose channel.
- dont run update_pose_constraint_flags() when adding an object constrant to an armature object.
These were markers which belonged to an action instead of the scene,
and are used by PoseLib to keep track of where poses are.
To restore this, I've made this only available in Action/Shapekey
Editor modes, and only when an action is being shown and the "Show
Pose Markers" option in the Markers menu has been enabled. Other than
that, all the standard marker operators apply now (instead of using a
separate set of special operators).
Old value of 30 px was perhaps a bit too strict, given that scrollbars
took up 16px and that standard icons are 16px, which in total would
have been 32px from the bottom of the region. Having said that, the
marker icons are only about half-size, though it seems that in general
people (can) only aim quite loosely...
* Viscoelastic springs between the fluid particles can simulate all kinds
of viscous and elastic substances, such as jelly and honey. This is
achieved by creating springs dynamically between neighboring particles
and adjusting their rest length based on stretching/compression.
* This nearly completes the currently intended functionality for particle
fluids. The last missing thing is a surfacing extraction algorithm,
which is needed for a proper representation of a sph fluid.
* I also cleaned up and renamed some of the fluid parameters to make the
ui a bit easier to understand.
* One addition to the patch is an option to use "initial rest length" for
the springs, which uses the lengths between the particles at the time of
spring creation as the spring rest lengths instead of interaction radius/2.
This makes the fluid keep it's original shape better (good for very
viscoelastic materials), but can create large density differences inside
the fluid (not really physically correct for a fluid).
* Viscoelastic springs are stored in point cache as extra data.
+ changed lines connecting nodes:
they now use a linewidth of 1.5px for the light foreground and 4px for the dark background.
this should fix node-lines not being visible on almost black or all white backdrops.
+ muted nodes now also show a red tinted header if they are hidden (collapsed)
+ both active and selected nodes show a (now properly antialiased) highlighting frame
+ fixed a small error in dropshadow code resulting in a gap at borders
+ fixed a tiny error for the collapsing indicators (triangles) - they were not symmetrical.
Ton will add proper theme colors for the node-editor in the coming days.
* Not strictly necessary right now, but better for future.
* Struct data (only boids at the moment) is now written as structs (with dna) so they work between 64 and 32 bit machines too.
Actually a minor tweak:
Auto-IK fails on correcting for constrainted bones in chains.
For that reason it stops including constrainted bones for it.
Now it does include constraints with influence zero, or which
were disabled by user input.
(should have made this change along with the others).
Matrix([1, 2], [3, 4]) --> Matrix(([1, 2], [3, 4]))
This is so adding initialization args works right.
Also simplify initialization code (re-use slice assignment).
Improved report print for 'remove doubles' to make it correct english.
Unsure if this kind of code can survive though, static strings do much
better for future translation efforts.
Instead of providing nice grammar sentences, we could make it more
declarative, like:
"Removed vertex amount: %d".
Transform crash on autokey during animplay.
Transform event handling requires Context, for inserting keys.
Marked it with XXX warning; we have to be more careful with which
API calls require context; like inserting keys shouldn't need it?
Originally Context was meant for operators; to define user level
or scripted context, not for underlying APIs to work.
- this just toggled between different rotations, I can't find any references to this as a common operation to have with eulers.
- wasn't working at all nobody noticed, not used by any blender scripts/addons either.
* I've getting bad feelings about the point cache index_array for a while (cause for this bug too), so from now on memory cache uses a simple binary search directly on the index data to handle queries to specific data points.
* This is a bit slower than just checking from a dedicated array, but it's much less error prone, uses less memory and makes the code more readable too, so it's not a tough choice.
Downloading results for jobs from blender now uses the current output settings, it doesn't just download the multilayer exr as it used to.
Render output panel now visible under the jobs panel in client mode.
- Outliner: new scroll operator, PageUp PageDown scroll entire page now.
- 2D views (like buttons) PageUp PageDown now also scroll entire page.
(they used same step as scrollwheel before)
Python console crashed on opening.
Too quick code cleaning for "unused variables" caused a complete line
of code to disappear that was needed anyway :)
- Proportional circle size is printed in header
Allows you to find out if you make it smaller when it's large
- Proportional size is clipped with view3d clip-end now
- Added the size to rna, so you can inspect values via UI and py.
also removed unused vars.
can_pbvh_draw() had a NULL check which is never needed (callers check for this), a NULL ob would have crashed the function anyway.
More cleanups: moved a function declaration to the correct module,
removed old/incorrect comments, marked more things with TODO where
appropriate, refactored copy-pasted function, de-duplicated code.
More cleanups to the cursor drawing code; factored out another
sculpt-related function, replaced float calculations with bitwise
flags, removed unnecessary GL changes, de-duplicated some lines,
removed unused parameters, and added more comments.
This adds the "Apply Base" feature from my gsoc2010 branch.
Apply Base partially applies the modifier, in that the mesh is
reshaped to more closely match the deformed mesh. The upper-level
displacements are recalculated so that the highest multires level
appears unchanged.
Multires does not currently deal well with too large displacements.
An easy-to-reproduce example: create any mesh type, add multires,
subdivide a few times, then use the sculpt grab brush to drag the
entire mesh over a few units. At the highest level, and at level 0,
the mesh looks fine, but all of the intervening levels will have ugly
spikes on them.
This patch doesn't help with situations where you can't modify the
base mesh, but otherwise works around the problem fairly well (albeit
with a heuristic, not an exact solution.)
This problem was in both new bump and new new bump. The derivatives used for
bump mapping did not include the mapping, which gave mismatched offset for the
taps, now just do the mapping beforehand.
Many thanks to them!
For comparison, see here:
http://kishalmi.servus.at/3D/bumpcode/
Based on algorithm in: Mikkelsen M. S.: Simulation of Wrinkled Surfaces Revisited.
http://jbit.net/~sparky/sfgrad_bump/mm_sfgrad_bump.pdf
This fixes bugs:
#24591: Artefacts/strange normal mapping when anti-aliasing is on
#24735: Error at the Normal function.
#24962: Normals are not calculated correctly if anti-aliasing is off
#25103: Weird artefacts in Normal
This will break render compatibility a bit, but fixing this bugs would have also
done that, so in this case it should be acceptable.
Patch committed with these modifications:
* Bump method Old/3-Tap/5-Tap option in UI, 3-Tap is default
* Only compute normal perturbation vectors when needed
* Fix some middle of block variable definitions for MSVC
* Renamed children to "simple" and "interpolated" as this is
easier to explain and more descriptive than "from particles"
and "from faces".
* Also shuffled the child ui around a bit to make it clearer.
* Child seed parameter allows to change the seed for children
independent of the main seed value.
* Long hair mode for interpolated children:
- Making even haircuts was impossible before as the child
strand lengths were even, but their root coordinates were
not similar in relation to the parent strands.
- The "long hair" option uses the tips of the parent strands
to calculate the child strand tips.
* Hair parting options:
- Hair parting can now be calculated dynamically on the fly
when in 2.49 there was a cumbersome way of using emitter mesh
seams to define parting lines.
- For long hair parting can be created by a tip distance/root
distance threshold. For example setting the minimum threshold
to 2.0 creates partings between children belonging to parents
with tip distance of three times the root distance
((1+2)*root distance).
- For short hair the parting thresholds are used as angles
between the root directions.
* New kink parameters:
- Kink flatness calculates kink into a shape that would have
been achieved with an actual curling iron.
- Kink amplitude clump determines how much the main clump value
effects the kink amplitude.
- The beginning of kink is now smoothed to make the hair look
more natural close to the roots.
* Some bugs fixed along the way too:
- Child parent's were not determined correctly in some cases.
- Children didn't always look correct in particle mode.
- Changing child parameters caused actual particles to be
recalculated.
* Also cleaned up some deprecated code.
All in all there should be no real changes to how old files look
(except perhaps a bit better!), but the new options should make
hair/fur creation a bit more enjoyable. I'll try to make a video
demonstrating the new stuff shortly.
* Two separate bugs, with very similar symptoms.
* The distribution binary search didn't work correctly in cases where there were a lot of faces with 0 weights.
* Maximum distribution sum should have been exactly 1, but due to the wonderful nature of floats this wasn't the case at all.
from Dan Eicher (dna)
From the tracker (with minor edits)
========================
cube = bpy.data.objects['Cube']
foo = cube.vertex_groups.new('foo')
foo.add([1,3,5,7], 1.0, 'ADD')
for i in range(len(cube.data.vertices)):
try:
weight = foo.weight(i)
print('vert: %i weight: %f' % (i, weight))
except:
pass
foo.remove([1,3])
cube.vertex_groups.remove(foo)
from Alexander Kuznetsov (alexk) with edits.
From the report:
Blender assumed that all files are .blend as retval = 0;
Now retval is initialized as file cannot be open (-1) for gzopen fail and directory case
retval = -2; is defined for not supported formats
This must be assigned before #ifdef WITH_PYTHON because this part can be missing
Finally retval = 0; if it is a .blend file
---
also made other edits.
- exotic.c's blend header checking was sloppy, didn't check data was actually read, only checked first 4 bytes and had a check for "blend.gz" extension which is unnecessary.
- use defines to help readability for BKE_read_exotic & BKE_read_file return values.
- no need to check for a NULL pointer before calling BKE_reportf(). (will just print to the console)
- print better reports when the file fails to load.
- avoid looping over the entire collection unless a negative index is used.
- dont use the get index function for building the slice list, instead loop over the collection until the stop value.
don't draw the image if the size is 0.
Crash was actually an assert() so debug builds only, replace assert() with BKE_assert() so crash is opt in build option.
- Toobar views were reset on hide/unhide. Now they keep the view
and zoom level.
- Added operator to delete all unused 'space data', this to make
clean startup.blend files, remove unused editors, and to be
able to test starting defaults for editors.
No hotkey, use search for "Clean-up space-data"
Compositor: Texture Node now behaves like an image.
- Image always in render output size
- Buffer outputs RGBA and Value both supported
- Works for filter and blur and scaling too.
- Mixing 2 textures works
Implementation note:
The texture node was meant to be 'procedural', not a buffer
but a color-sample method. Unfortunately the node editor
didn't support this well, blur/filter/scale ignored it too.
For now, its better to drop this procedural concept, then
things work at least as expected. :)
Auto-texture space now is more responsive and correct.
- on transforming it, the buttons get redrawn to show option was reset
- on enabling option, texture space is recalculated/reset
Issues:
- It looks a bit crowded though, so perhaps we could do without, and
just leave it for the TimeLine only?
- Due to the way the invoke() testing works, the rename operator
called from the menu currently fails. Will need to experiment more
with execution options to find a workaround.
Speedup for getting collection indices, avoid getting the collection length unless a negative index is given. This avoids a loop over the entire collection in many cases.
Speedup for getting collection slices by detecting collection[:] and internally calling collection.values(), this gives a big speedup with some collections because each slice item would loop over the list until that index was found.
Rough test with 336 objects.
- getting index of listbase collection ~ 5.0x faster
- getting index of array collection ~ 1.15x faster
- getting slices of listbase collections ~ 34.0x faster
- getting slices of array collections ~ 1.5x faster
animation editors (DopeSheet, Graph Editor, NLA, Sequencer) ==
=== Usage Notes ===
In animation editors, marker operators will only be considered while
the mouse is hovering near/over the horizontal scrollbar (i.e. where
the markers usually appear). That means, in order to do something to
the markers, just position your cursor in line with the row of
markers, and then use the same hotkeys you'd use in the TimeLine (so,
unlike in 2.4x, no more need to hold down extra modifier keys for this
case). In the TimeLine, nothing changes, so you don't need to worry
about mouse placement there :)
=== Technical Details ===
Since early 2.5 versions, this functionality has been disabled, as the
markers were always getting evaluated first, and hence "swallowing"
all the events before the editor's own keymaps could access them.
In order to get this working again, I've had to give every marker
operator a "wrapper" invoke callback which performs some checking to
ensure that the mouse is close to the markers (vertically) before the
operator will try to be run. This wrapper also makes sure that once
the operator has finished running, that if it didn't manage to do
anything, then the editor's own keymaps get to have a go.
The vertical tolerance used is currently 30 pixels (as was used for
the borderselect operator).
=== Other Assorted Changes ===
* Gave marker operators dependent on having selected markers to
operate on suitable poll() callbacks. These new poll callbacks ensure
that there are selected markers for the operator to operate on,
further cutting down the number of places where markers may override
standard hotkeys (and avoiding calls to the wrappers too)
* Simplified some of the selection code
* Made some formatting tweaks for consistency, and in one case so that
my text editor's function-list display doesn't get confused
- convertblender.c, remove assignments to unused vars.
- readfile.c, fix 2 possible crashes. null pointers were being checked for then used later without checking.
- space_graph.c, use switch statement for automatic color assignment rather then a float array.
I've been asked by Giuseppe Ghibò (joeghi) for more icons for Mandriva Linux distro.
I've noticed that some folders had a blender.svg different from the one in scalable/, so made them all the same derived from
scalable/blender.svg
Now each svg file has "document properties" of the right dimensions according to the folder it's in, and the svg is saved so that people
just have to press "export bitmap" in Inkscape in case they need it.
Outliner: using the object-data icons to enter editmode, didn't
reset the toolbar Operator redo. It was not using an operator
call here. Note to self: more outliner tools don't...
Error was that operator-redo then would repeat an action incorrect.
Code changes are minimal, re-using the code as already was there.
Options are in pulldown menu, or SHIFT+T and SHIFT+ALT+T
Might be that Martin likes to see it different... do we need
a special operatortype for it?
Object color option now is incorporated in preview render, and
in display in 3D window.
Note that Object-color is "modulating" the render result, it just
multiplies... which is very limited for practical use. This was
added for Apricot game project. The original meaning however was
to replace diffuse only... so specular and light work still correct.
ALso made header for properties editor start on top, to reflect
default (report in tracker).
mostly rewrote mesh export and added support for multiple materials/images, using texface or material images depending on the materials TexFace option.
Slightly modified to better fit in architecture (moved to related GHOST SystemPaths)
Thanks to Harley Acheson for the research and for providing the original patch.
Note: I added empty function for X11(Linux) and Mac (Carbon and Cocoa) to be implemented still.
we add extra creasing in the UV mesh, to keep it from shrinking, leading to
distorted UVs, but this wasn't always working right, so tweaked the conditions.
Some tweaks for this operator, found while scanning through the
current state of the markers code. These seem to be correct judging
from the overall direction the code is going in, though the intention
of this operator seems a bit unclear from the naming.
- Fixed potential issue where with only some markers selected (and the
last one selected not being the last marker in the source scene), you
could end up chaining the source and destination marker lists together
improperly.
- Clarified description/error messages...
DNA reconstruct now terminates strings, when the new string
in a file is longer than in current definition. This will
enable a safer ID and other name lengthening, keeping at
least some forward compitability.
Use object's displists for storing deformed tesselated curve. Was unable to
totally get rid of curve's displist because of how texture space is calculating.
- mesh creaseAngle was being exported as degrees, should be radians.
- remove unused vars & code.
- indenting was using a for loop, can just multiply a string instead.
bpy.data.meshes.tag = True
But this was only useful for setting so make it a function for 2.5x.
bpy.data.objects.tag(False)
X3D: use tagging rather then a name dictionary, this fixes a bug where library name overlaps could mix up names.
implemented the support for animated colors and a workaround for non animated colors. To be cleaned if/when this bug is fixed:
[#25487] BGE: Object Color only works when it has a keyed frame
* also increased the size of the text. Now it supports 280 (or 140 unicode) chars.
New option to start threaded wmJobs, with flag WM_JOB_SUSPEND.
This makes the job wait 1 timer step before running.
Used now for Material Icon render renders, which makes the
big preview to be always rendered first while using UI.
- File Window: when opened with operator (save, load, etc), you
couldn't start a new one, causing memleaks. Now it nicely
refreshes file window for new operator.
Also means you can make CTRL+F3 screenies of filewindow now.
- CTRL+F3 screenshot had memleak on cancel.
Due to popular request and usability considerations, this commit
reintroduces functionality similar to 2.4's "Draw Mode" for Grease
Pencil.
In the toolbar under the Draw/Line/Eraser buttons, you can find the
"Use Sketching Sessions" toggle, which enables this feature. This is a
per-scene setting, and defaults to off, so that the current 2.5
behaviour is still the default (i.e. the Grease Pencil operator will
only do a single stroke at a time).
With this option enabled, drawing with Grease Pencil will enter a
semi-modal state where you can draw multiple strokes without needing
to keep holding the DKEY throughout (though you'll still need to do so
to start the strokes, unless you use some toolbar buttons), while
still being able to manipulate the viewport. Header help-text prints
show the appropriate keybindings (i.e. press ESCKEY or ENTER to end
the sketching session).
Notes:
- To aid maintainability of the 3D-View toolbar code, I've taken the
liberty to factor out the groups of widgets which commonly occur in
most of the toolbars into separate functions (namely "Repeat" and
"Grease Pencil"). Perhaps it might make it slightly harder to newbies
to the toolbar code to grasp, though the physics panels are far worse
;)
- I've reshuffled some code in the Grease Pencil code to separate out
the various states of operation again more clearly, though some more
work is still needed there (TODO)
- There can now be only one Grease Pencil operator running at a time
- Redoing Grease Pencil operations where sketching sessions was
enabled still needs work. Namely, a way of delimiting the set of
points recorded into strokes is still needed (TODO)
- Ultimately, it should be possible to switch tools midway through a
session. Currently sessions are limited to only being able to be used
with a single drawing mode (TODO)
- After ending a drawing session, the titlebar contols may not work on
Windows without manually making the main window lose focus and then
regain (i.e. click on some other window in toolbar, then come back).
This may be related to (bug #25480)
2D view scrollers were drawing over background contents, making it
look somewhat messy (like for text, nodes, fcurves). Now it clears
first the back in the region edge.
Drag/drop now has a User preset for dragging threshold.
Noticed this was set to 3 pixels even, made it 5 as default.
Tablet owners can put it larger too :)
Note: the tweak-threshold (3d win) is 10 pixels now, I think
this needs another preset too, leave this for now.
Also: fixed crash in filewindow: drag .blend icon and drop it.
You can't test anything in Blender or you get a bug :)
- On each re-render, the node image was cleared. Skipping this gives
nicer pictures
- Node render was using AA, but unfortunately only 1 sample for Nodes
is being stored. Disable AA render for now, nice speedup too.
- New Node editor now opens with larger view, the default
was like zoomed in factor 2.
- Add node via menu now shows new node more visible
- Click on (material) node now doesn't re-render the entire
tree anymore, much nicer.
- Duplicate node creates preview image immediate
After loading file, the Undo-push happened too early, causing an
undo for the first action to show animated setups wrong.
(material.c: removed old crap)
This commit partially fixes the problems with Shapekeys from older
files, as seen from the Regression suite (relative.blend and
dolphin.blend in particular).
In older files, keyblock->slidermax was never truly set to 1.0 even
though the UI may have shown such a value (which was bizzarely being
sourced from somewhere else). Hence, after loading the files in 2.5,
the shapekeys wouldn't animate, as the value would get clamped between
0 and 0.
To fix this, I've added a version patch which corrects these
situations in old files, and I've adjusted the slider-RNA code so that
it is not possible to set up such clamping anymore.
TODO:
The fixes detailed here only make it possible for these files to work
again in 2.5. However, I haven't been able to find a way to get the
files to actually work in 2.5 without manually changing the active
shapekey (per object) after loading the files with these patches
applied. Possibly it's just some depsgraph magic needed, unless
there's still some other evil voodoo in the shapekey code
move Object.update(...) to ID.update(). since depsgraph update function can now be called on ID types.
also changed how update flags work.
obj.update(scene, 1, 1, 1)
... is now
obj.update({'OBJECT', 'DATA', 'TIME'})
Don't pass scene anymore. This was used for recalculating text but I think this is better dont in a different function.
Also added bpy.path.display_name_from_filepath(), since filepaths are not ensured to be utf8.
same as calling: os.path.splitext(os.path.basename(name))[0].encode("utf8", "replace").decode("utf8")
Keyframing operators now use a dynamically-generated enum for their
"type" property, which determines the Keying Set to use for keyframing
instead of the obscure "index" values which were determined
internally. Internally though, these same indices are still being used
:)
Notes:
- I've kept the menu-building function and the special "menu" operator
for now, since it's better to not "pollute" the actual insert
keyframes operator with hardcoded menu-showing logic. Instead, the
menu operator does that, so that if you like, you could write another
such wrapper that works differently.
- The 'type' properties could have the PROP_HIDDEN flags removed,
though I think it's cleaner to leave these without this option for
now.
Experimental tweak for the "Object Constraints" tab so that it now
shows a warning message instead of the "Add Constraints" button when
the active object is in Pose Mode. Hopefully this will further
alleviate any confusion over Object vs Bone level constraints.
Todo:
While coding this, I noticed that we currently don't have any way of
making help-text labels in UI panels which can span multiple lines
(word-wrapped or manually split). Probably not a critical issue, but
it would be nice for completeness...
Setting the group for F-Curves from the Py-API (i.e. to move F-Curves
from one group to another) was failing. After debugging this, this
functionality should now work correctly, while still prohibiting the
setting of groups on "F-Curves that aren't in actions" (i.e. driver-
fcurves)...
The icons for materials were always lagging or not updating
at all. I also found it suspicious slow...
It appeared that the icons now store a "mip level", where for
every change in Materials 2 render jobs for icons were started,
one for 32x32 pix, one for 96x96. The latter was cancelling out
the first job almost always.
Also made preview renders detect size, to set amount of tiles
to be rendered. Small icons use 1 part, larger previews 16 now.
All in all, behaves much smoother now! But, will also update
the thread Jobs manager to allow "delayed jobs" like for icons,
these are aggressively put as first in the jobs list.
* Removed the unused fuctions spotted by Campbel.
* Removed the gotos to make Ton happy.
* Added better debug prints to disk cache operations.
* Fixed a memory unmap error that seemed to happen on a test file.
work correct with sculpting data now.
Joining two triangles could give incorrect sculpting result for
special topologies, but it's that case that can't be nicely handled
with our layers architecture.
* Massive reorganization of pointcache code, things are now cleaner than ever.
* For all but smoke the cache is first written to memory when using disk cache and after that written to disk in one operation. This allows less disk operations and the possibility to compress the data before writing it to disk.
* Previously only smoke cache could be compressed, now the same options exist for other physics types too (when using disk cache). For now the default compression option is still "no compression", but if there aren't any problems this can be set to "light compression" as it's actually faster than no compression in most cases since there's less data to write to the disk. Based on quick tests heavy compression can reduce the file size down to 1/3rd of the original size, but is really slow compared to other options, so it should be used only if file size is critical!
* The pointcache code wasn't really 64bit compatible (for disk cache) until now, so this update should fix some crashes on 64bit builds too. Now all integer data that's saved to disk uses 32 bit unsigned integers, so simulations done on 64bit should load fine on 32bit machines and vice versa. (Important disk cache simulations made on 64bit builds should be converted to memory cache in a revision before this commit).
* There are also the beginnings of extradata handling code in pointcache in anticipation of adding the dynamic springs for particle fluids (the springs need to be stored as extradata into point cache).
* Particles were being read from the cache with a slightly wrong framerate. In most cases this probably wasn't noticeable, but none the less the code is now correct in every way.
* Small other fixes here and there & some cosmetic changes to cache panel, but over all there should be no functional changes other than the new disk cache compression options.
* This whole re-organization also seems to fix bug #25436 and hopefully shouldn't introduce any new ones!
* Argh my bad, sorry about this!
* Now only the actual data array is saved to avoid constant re-allocations, but no relations to active data are kept.
* Also reverted Ton's quick fix for the crash as it's not needed anymore.
Floor constraint didn't work: the defines for the enums were using
the wrong ones, the right ones are not logical... but code and dna
and old files assume these. Now it works :)
Crash in Bezier animation (inserting keys on control points in
curve object). The animation rna paths were not fixed after an
editmode session, which got fixed 2 weeks ago, but for all older
binaries the issue can still pop up.
The crash happened because the RNA array-itterator was not doing
a boundary check, even whilst the array size was passed on to the
itterator callbacks. With rna then writing far outside of valid
memory, very bad and unpredictable corruptions happen.
I've added a range check now, and a decent print to denote the
issue. An assert quit is useless, since a tab-tab on curve objects
will fix the channels nicely.
Example of warning print:
Array itterator out of range: Spline_bezier_points_lookup_int (index 30 range 2)
Serious *bad* crash in undo introduced by commit Janne dec 21st.
Time window now stores some kind of cache for fluids/cloth, but
it's pointing (in SpaceTime) to data inside Objects. (Not ID).
That's really not allowed... this commit fixes crashes but the
cache code really needs to be redesigned. I'm also afraid this
crash is going to frustrate everyone using physics...
- the invert flag was only being used for multi-modifier, but there is no reason not to use this in normal cases as well.
- Armature modifier RNA name 'vertex_group' was incorrectly named 'vertex_group_multi_modifier' (own fault), confusion was caused by 'invert_vertex_group_multi_modifier' which was correct.
Having edit operations select the mirrored bones means you cant so easily continue editing on one side of the armature.
This is also inconsistent with other mirror editing operations in armature and mesh modes which don't adjust selection like this.
The bug was 'write_still' was incorrectly being initialized to 'view_context'.
'write_still' should always write an image, so failing silently here is bad behavior.
Also, opengl render and internal render engine operator should use this option the same way.
this missed some cases, now also disallow ints to be wrapped as floats.
This commit also exposed a number of cases where ints/floats were incorrectly wrapped.
Bugs like [#25416] wont slip through the cracks anymore.
Fixed by patch attached by submitter. Cheers Jacob F (racoon)!
<quote>
If you write something like...
def do_something:
...and press return it makes the new line with a tab which is good,
but if you want to move the def line down to make room above it by
pressing return while your caret is at the start of the line it
indents the new line which has the def statement on it.
Attached patch just checks if it's checking for a colon after the
caret and returns if it is.
</quote>
"Numpad 1" shortcut to set preview view zoom to 1:1 (i.e. 100%) did
not exist in View menu. While investigating this, I found that the
operator was missing a description/tooltip, so added one too.
Clobbered together some new code to do this based on the code for Edit
Bone circle select as this case doesn't seem to have been implemented
yet.
NOTE: the code in view3d_select.c needs some urgent love and tidying
up (like a room full of flaking paint).
Lamp shadows for offscreen render (opengl anim) had to be remade
to cope with animated objects. Fix proved by Alexander Kuznetsov
in the tracker log. Thanks!
- rather then use unlink="None", just don't pass unlink as a keyword. This is more pythonic.
- added an RNA flag for properties which cant be unlinked by setting to None.
Bugfix for job cancellation (reported by Carsten in email)
Ended up recoding part of the communication pipe (use json more consistently)
Fix bpy data modifications where it shouldn't happen (as a bonus, thumbnailing is now done out of process)
Frame Mapping (map old, map new) didn't set the the framelen
variable. Note that this feature is half-working, and on the
to. Might be removed/replaced with something better.
Keyframes for locked channels are now shown faintly so that it is
possible to easily distinguish between keyframes for locked channels
and unlocked channels. Hopefully this solves the problem where you
have some keyframes selected, and try to move them but forget that
those channels are locked (without any feedback other than a single
icon).
Thanks for pointing out this problem Ronan Zeegers!
compositor
Now ungrouping grouped-nodes copies the animation back to the main
nodetree. This means that it is now possible to successfully roundtrip
group/un-group nodes and their animation data.
---
This should also be done for the Separate Armature operator... hmm...
compositor
This commit fixes the original bug reported here by adding some
methods to move the relevant F-Curves (and drivers) over to the new
Node-Tree's (i.e. group-node) AnimData. Animated nodes which
subsequently get grouped will still be able to animate as a result of
this commit.
TODO's:
- Ungrouping now will not yet merge the animation back (or at least
copy it)
- Buttons for nodes freshly grouped do not correctly show animated
status indicators for some reason, yet normal animation does
Loading a vector font (.ttf) didn't have option for Relative Path
in the file window. Now it does, and it sets it to the default.
Todo note: the actual path is not visible in the UI yet, nor does
outliner RNA view allow to edit path.
Some AnimData looping functions were not updated to take into account
the addition of AnimData for some additional datatypes (i.e. meshes
and lattices) as well as the changes to make nested nodetrees (i.e.
for Material and Texture nodes).
Auto-IK: applying back the IK transform on bones fails for constrainted
bones themselves. This would require the post-transform of IK to be
converted to a pre-transform for the constraint. That's too much of
a headache for bugfixing days.
To keep the autoIK feature work it now stops IK chains at constrainted
bones. Added todo item for a matrix wizzard to fix this one day.
(What would be really cool is a matrix-api for armature to easily
get all the required pre- and post transforms and apply. There's a lot
of duplicate exception handling for hinge bones for example.)
Mesh "Add grid" was quite broken:
- it added an edge to a previous (not selected) vertex
- it ignored cursor location
- and it didn't use 'align to view'
Now it does behave :)
Reported by Nils Austa
There was double handling of modifier keys. With my autumn work on the modifier keys I had moved state checks for
modifier keys into convertKey(), but left the call to handleModifierKeys. That caused problems with proper
modifier key handling in the entire key handling code of GHOST.
UI name buttons with Search callback show red when item couldn't be
found. The check was too simple though, since the searches also find
partial matches. This caused KeyingSet browse to show red for "RotScale".
Now 10 matches are stored, and each match is being checked. Still a
bit weak for cases with more than 10 matches, but in that case it
won't show red-alert.
Reshuffling order of "sync" and "Auto-Keying" buttons so that they are
grouped more appropriately. Hopefully this isn't too much of a muscle-
memory breaker...
Compositor: "Read & Compo Full Sample Buffers" option stopped after
1 pass; the break callback was returning '1' falsely.
Also made it show "wait cursor" again, since it doesn't run in
thread nor updates UI while compositing buffers. Check console for
progress for now.
again)
- Graph Editor "Active Keyframe" panel now displays more descriptive
error messages. In particular, hopefully this helps to alert users of
the default generator modifier for Driver F-Curves
- The first F-Modifier added to a list is now set to be active one for
that list.
- Adding drivers from the UI (not from py-scripts though) will now
automatically add a "Transform Channel" driver variable to the newly
created drivers. This makes setting up drivers a bit more convenient
for the most commonly used case.
- Drivers now report their errors using the Reports system instead of
writing these directly to the console.
- Clarified some comments to be more insightful about the "why's" of
some design decisions, and related formatting/cleanup tweaks
- Reduced scope of "path" vars to just the scope they're required in
- Removed some unused defines from a failed experiment in the original
Keying Sets code ("templates" and "template flags") which was
superseeded by the more flexible + nicer "Builtin KeyingSets"
adjust dependency calculations for cache files of smoke simulations (no more separate paths for low and high). That was changed a while ago in the api, just didn't update yet.
Compositor: Texture node didn't use texture-nodes itself.
Now composites initialize texture nodes correctly.
Also reviewed the fix for crashing texture nodes for displace.
It appears texture nodes also are used for sculpt/paint
brushes, in these cases it can be allowed again. But, don't
do this during rendering for now!
- OpenGL still render failed when output format was set to Movie.
Now it just doesn't save a file and renders anyway.
- Bone heat weight was missing 'wait cursor'.
- Waitcursor for Mac Cocoa is back! Made all OS's use the same
nice hourglass cursor.
(Note: this violates Mac UI guidelines, we should rely on the
spinning wheel of death instead. Highly disputable that.)
runs in user mode, not elevated to admin. Otherwise users can't install to the default which used to
be Program Files, which causes more problems than necessary. Although unconventional location, probably
currently a better solution.
Network admins etc. can override install location on command-line. NSIS docs tell how.
Normal Map bool and the normal_space menu are only used for material based textures, not for textures used by lamp/world etc. Accidentally ported this over from 2.49, where these buttons are drawn even they don't work in that context.
Node editor
Some nodes - when muted - didn't draw a red line to indicate
a pass-on action. Stil, the node then is muted and either passes
on a buffer of same type, or just black.
To more clearly indicate muting, it now draws the header in a
reddish blended color too.
Hair render: using strand "Blender Unit" size didn't correctly
clip for larger/wider strands. Now code clips strands based
on the maximum width.
Also found bad code for using clipping flags, which was mixed up,
and probably caused hair strands to be missing in cases.
VBO enabled, GLSL shading and bevel modifier leads to crash when entering edit mode
Bevel modifier doe not support remapping still, so added check about original index
in cdDM_drawFacesTex_common. Now it works like VBO-less drawing for that special case.
action editor and the NLA editor
This commit fixes the situation where if a NLA Track was on 'solo'
mode and then deleted, NLA evaluation (including Active Action) would
stop working as a flag wasn't getting cleared.
Issue caused because previously sliders used a separate mask and now they're
drawing inside of region.
Thanks to AlexKu for point and Ton for checking :)
Sculpt: brush type "Grab" with option "Frontface only" was flipping
back and forth when grabbing outside of original brush area.
Reason: it was calling a function that used deformed coordinates
for getting normal, clipped by brush area. Now it uses original
coordinates, so it always finds normal.
Nicholas can determine better if tag "cache->original" should be
set as default for this brush type, this far I don't dare to go :)
Having 2 windows open, rendering to image editor, activate other
window and press F11 -> UI went haywire. Simple fix prevents
render output to be closed in this case.
Property buttons: the "ID template" allowed to unlink data from object,
like Mesh or Curve etc, which is not supported. Button now is hidden.
Thanks Dan Eicher for patch.
Also: improved tooltip for 'data browse'. It was generic, now it gives
specific per-type information.
- operators need to enable the option OPTYPE_PRESET or 'PRESET' in python.
- hidden properties are not written.
- currently this only works in the file selector (enabled for FBX/OBJ export)
- currently the menu label doesnt change when presets are selected, this is a TODO, not so simple since the UI is defined in C and the label in python.
- presets save in "scripts/presets/operators/*id*/*.py"
The main advantage with this is that its close to twice as fast to do 'vertex.co[:]' then 'tuple(vertex.co)', this is common for writing a vertex array.
the correct python behavior in this case is to return a copy of the original type, however euler and quats don't support different sizes so we cant do so easily.
Child-of constraint issue: on adding, it wasn't checking owner
correctly for Bones, resulting in a constraint working in wrong
space; it looked as if transform was applied double when moving
the object. Only adding via Py API went wrong btw.
Also found a silly check for drawing constraints, which caused
constraint initialization to happen for every object on every
redraw!
Implementation note: con->flag CONSTRAINT_SPACEONCE was only used
for child-of constraints in Bones, so I've patched it on file
reading to always set the flag. Marked with XXX, so it can be
removed one day. Now at least things get corrected well for
imported armatures.
In some cases, a LMB-drag was drawing a stippled line. Leftover of
WIP code for support of gestures in Blender.
Disabled drawing for now, until gestures is back in control.
Since 2.5x blender has been using CD_MASK_BAREMESH for updating objects since object_handle_update() no longer has access to G.curscreen to calculate the mask from viewports.
The problem with this is after an initial calculation, CD_MASK_MTFACE may be required on draw, so it would recalculate the modifier stack multiple times per frame.
One case which caused this is armature animated mesh with texface in a dupligroup.
Fix this by having customdata_mask member in the scene, this isn't great design but at least fixes the bug and only changes a few files.
Loopcut check for deformed meshes was far too wide, it even disabled
loopcut on subsurfs.
Now added a check for armature/lattice only, and only give a warning,
not a return from the tool.
- adding template of a add_object addon from Florian Meyer (testscreenings)
- this template is using the new add_utils module in bf-extensions (also from Florian)
Porting the fix when Fullscreen flag is on (~ scene.gm->fullscreen)
I still see some issues with that, a difference between width/height. To be addressed later.
Porting the fix when Fullscreen flag is on (~ scene.gm->fullscreen)
I still see some issues with that, a difference between width/height. To be addressed later.
view rotate precision was lost over time, make view3d rotate normalize rv3d->viewquat.
note, this no longer crashes blender, just prints.
also commented some RegionView3D struct members.
* Timeline didn't listen to file read notifier, so the pointcache frames indicator didn't get updated.
* Also added listens to particle & modifier (cloth, sb & smoke) notifiers as changes that cleared a pointcache weren't shown directly in the timeline either.
* The timeline display was also always one frame behind the actual state, since the notifiers are handled before the actual dynamics are calculated.
** This is now fixed too, by creating the actual drawn data always at drawtime.
An operator Error throws up a menu, a Warning only a flashy header print.
In mesh editmode the menus for simple failures got very annoying, like
"Already a face" for Fkey on a face.
Proposal is to use warning for contextual failures, like:
- wrong selection
- unsupported combination
- wrong modes
And use errors for cases you really need user attention, for example when
issues are invisble or potentially damaging work.
- Memory failures
- Files not found
List can grow in future :) let's test this for mesh now.
I'll tackle this for other ops later after review.
(Also changed: loopcut disabled when editmode shows deformed result)
- SSS Presets were not working on pinned materials.
- added ability for save-presets to define variables to stop them becoming too verbose.
- remove object.active_node_material
ED_curve_updateAnimPaths now updates fcurves from animation data rather than
from local copy of curves: nurb and point index from keyIndex data gets
updated in this function and for correct handling undo stuff copy of
fcurves is stored in UndoCurve.
Additional changes:
- Revert of rev33640 and more proper fix: do not copy remained
spline-related rna fcurves after renaming pathes but all the rest
fcurves woulddbe copied back to animation data
- Refresh graph and dopesheet editor after updating animation data
"Make Local" option "Objects and Data" made materials get
lost on save, the tagging was incorrect. Now it makes
Materials local too.
Implementation note: the code is new (not 2.4) so it misses
some of the hairy goodies we had. It's something for another
time to really make this reliably work.
File Window draw error:
On start Blender in smaller sized window, a scalled down screen
causes the File Window main area to draw too high, clipping off
half of the top line. This case (scroll horizontal only) is not
handled as view2d type.
Material Node trees:
SSS wasn't displaying in Nodes.
Implementation notes:
- Currently copying local data for preview render, only copies
the base material. Node materials are re-used.
- This causes data to be in 2 "main" databases... complex.
- To make it work for preview, I had to add another loop in the
SSS code that checks the original Main dbase. That's marked
as warning in code to be fixed up.
Another bug:
- Material properties "SSS presets" copied settings to the
active material, not to the displayed one. Added RNA call
to retrieve this from Objects.
(Next commit will fix preset scripts)
Code cleanup to allow switching active output nodes in Compositor
made shader nodes output not set correctly.
Now you can have multiple output nodes in shaders too, and switch
on click-activate.
* Show unborn was on by default, so smoke got emitted from all particles regardless of their birth time, not a good default in my opinion.
* What made things worse was that particles weren't shown in viewport, so you didn't even know the particles were considered alive from the very first frame! (Not rendering is a good default, but they should still be visible in viewport!)
- made theme colors for mesh edge len & face angle/area display.
- use %g rather then %f for float display, trims unneeded zeros.
- store cached 2d and 3d text color as bytes rather then floats, compare when drawing to avoid setting the context.
- use unsigned char for more color functions, avoids casting to glColorubv().
Bug reported by Dominique Lorre
Fix submitted by Jeroen Bakker
When importing COLLADA files, the name of a custom data layer can be longer than 32 bytes. Make sure only 32 bytes are copied.
Preview render for node shaders broke, caused by localizing
materials last week, to prevent thread crashes. Fixed now.
Also added a temp fix to draw color-management corrected
node previews default. Will follow scene setting tomorrow.
Also: SSS in nodes doesn't render yet. Was issue in 2.4 too...
when nothing is selected but active spline is set or when "middle"
control point is selected new spline would be created.
Nothing wull happen when there is no selection and no active spline
due to it's unclear which kind of spline should be added.
Pulldown and other popup menus: this button type exits on release.
While holding mouse you then can move around, but the highlight of button
didn't disappear if you were outside item. The menu then doesn't close
when you release the mouse.
Now highlight goes on/off on mouse moves to show this better.
Implementation note: menu items are coded similar to regular activate
buttons (like for tools). There's no provision to make highlights go
to the next item while holding mouse in menus. That I rather not mess
with now.
SSS preview render didn't happy.
Was caused by using a Material copy for preview renders (to prevent
crashing in other cases). There were actually 2 errors:
- material copy should be added to preview-main database
- the render was invoked using wrong main database even!
allowed
Infinite recursion (manisfesting as a crash) occurred when trying to
set the dupli-group setting on an object, when the object is a member
of the group being set. Added a check and warning for this in RNA to
prevent such setups from occurring in future.
Todo:
The warning report is currently only printed to console as I can't
quite remember how to grab reports pointer without context/operator
pointer available.
In sculpt mode, Object transform still worked, which was:
1) Not undo-able (sculpt undo stack)
2) Clearing the entire sculpt undo stack on operator redo
3) Had keymap conflict for R
Simply added check for this case and return trans operator.
[#25284] Traceback error on "System Info" script
- Align was only working on mesh objects, now operate on all objects, missing boundbox's are treated as single points.
- obj.bound_box was returning all nan's for object types with no boundbox.
- ENUM_FLAG type enums were showing no text when displayed in operator redo panel.
in crash.
This commit fixes the crash (missing null check for nu==NULL).
However, there is still a problem here with Curve Ctrl-Click not
behaving the same as Mesh EditMode Ctrl-Click, which adds a new vert
no matter what (i.e. no previous selection required). Then again,
that's a separate "todo" issue, so we can close this report now :)
RotoBeizer
The jump to keyframes operator was being a bit too strict with its
checks for where it should check for keyframes.
In this particular case, RotoBezier keyframes were on Curve AnimData,
so the check for Object AnimData would return false, thus overlooking
this possibility. However, it should be safe enough to just open
things up a bit more.
it (eg cameras)
When trying to add a modifier to non-geometry objects, warnings are
now shown instead of just adding some (useless) modifiers that cannot
be removed later.
Bugfix #25280
Image: sequence option was hardcoded to assume "first frame" was always
picture "001". Made it impossible to have a sequence of images starting
with picture like "000"
Note that by allowing to render a first frame as 000 in Blender, things
mess up a bit here. Things work now as follows:
- Start Frame = 1 : Image 001 on frame 1
- Start Frame = 0 : Image 001 on frame 0
- Start Frame =-1 : Image 000 on frame 0 ;)
This is of course the lack of proper control for image sequences.
Definite something to work on; best idea I have now is a new setting
that defines the Image Number to be "first frame". That way you can
map that number on any Blender frame. Or it makes it more confusing? :)
For the doc department: the proper meaning of "Start Frame" now is:
"The blender frame a sequence starts playing, assuming the sequence
starts with image #1"
Tooltop was fixed accordingly
(Also fixed 'remove doubles' to show more precision in toolbar)
* Pointcache code was quite ugly looking and complicated, so here are mostly just cosmetic adjustments, but some improved logic also.
* Slight cleanup of pointcache ui too.
* Shouldn't have any functional changes what so ever, so poke me right away if something seems off.
makesrna was often generating source every build, but not updating the files because the contents wasn't changed.
this happened because makefiles would check makesrna.c and rna_*.c files were newer then rna_*_gen.c and force a re-generation.
Now ensure updating the files even if they dont change when makesrna.c or rna_*.c are newer then rna_*_gen.c files.
Another solution for this would be to run makesrna program for each C file for finer grained deps.
or remove file comparison checks but that would mean a change to any rna_*.c file would rebuild all.
loop and increasing memory usage.
Modifiers should never call mesh_get_derived_final or similar, only
use ob->derivedFinal if it exists, if the dependencies are set correct
and there are no cycles, it will be there.
- Fixed some wierd cursor placements when clicking in special text position
(mostly when line segments had null-terminator at the max allowed position)
- Also subtract top padding, so centering is looks a bit better
You now can drop a .blend inside blender window to open it.
Implementation notes:
- Added call to extract icon type for files. Code re-used from
space_file
- External files that get dropped set icon types too.
Drop box polls can check for this.
- Also enabled setting op-context for drop operators, this was
needed to prevent filewindow to open.
was actually incorrect location for everything!
- x3d's global scene rotation value wasnt converted from deg to radians.
- camera viewport was also incorrectly exported.
use mathutils rather then inline math for rotations.
Makes adding new flags give ambiguous results and also makes it less easy to tell whats intended.
In some places it looks like OB_RECALC_TIME should be left out too.
When compositing without render (Enable composite, no renderlayer nodes)
the option Full Sample AA caused havoc. Added warning for this case
and gracefully stop render.
Sequence effect rendering of color generator did some rather strange
fall through, because of a missing break statement. (Noone got hurt, but
better add that break, just in case)
[#25011] Opacity IPO not refreshing with still images
(should be better named: animation of any prefiltering parameters using
still images didn't work out as expected)
And this issue by private mail by Ton:
"I tried to debug a memory-free error; very simple case:
- add image strip
- click on strip at 2 places
- quit blender"
Surprising this wasnt noticed in a much more obvious case:
- Key Location, Move, Rotate, Undo-Rotate >> Resets to keyed location as well.
This was happening because DAG_on_load_update() was called on read_undosave(), flagging 'ob->adt->recalc |= ADT_RECALC_ANIM;'
Fix by adding an option to DAG_on_load_update(), not to recalculate time flags.
Using scrollwheel on properties window halted when it was over
a list-button. Now it disables handling wheel events when you use
scrollwheel outside of lists first.
Implementation note: UI event handlers have priority over other
keymaps. That's OK. In this case it's simple conflict, for which
no design solution exists... wouldn't know how to do this nicer!
Code is nice local and can be improved when we look into other
conflicts...
After using loop-cut with multiple edges, transform widget
disappeared. Reason was a variable being set in function
not being called on cancelled operators. Now it's in Posttrans()
Multi-layer images: clicking on the "Source" popup freed all memory
for multilayers, even when choosing the same "File" entry again.
Now it should work :)
Problem/Bug:
------------
There were no way to have proper unicode characters (e.g. Japanese) in Blender Game Engine. Now we can :)
You can see a sample here: http://blog.mikepan.com/multi-language-support-in-blender/
Functionality Explanation:
--------------------------
This patch converts the Blender Font Objects to a new BGE type: KX_FontObject
This object inherits KX_GameObject.cpp and has the following properties:
- text (the text of the object)
- size (taken from the Blender object, usually is 1.0)
- resolution (1.0 by default, maybe not really needed, but at least for debugging/the time being it's nice to have)
The way we deal with linked objects is different than Blender. In Blender the text and size are a property of the Text databock. Therefore linked objects necessarily share the same text (and size, although the size of the object datablock affects that too). In BGE they are stored and accessed per object. Without that it would be problematic to have addObject adding texts that don't share the same data.
Known problems/limitations/ToDo:
--------------------------------
1) support for packed font and the <builtin>
2) figure why some fonts are displayed in a different size in 3DView/BGE (BLF)
3) investigate some glitches I see some times
4) support for multiline
5) support for more Blender Font Object options (text aligment, text boxes, ...)
[1] Diego (bdiego) evantually will help on that. For the time being we are using the "default" (ui) font to replace the <builtin>.
[2] but not all of them. I need to cross check who is calculating the size/dpi in/correctly - Blender or BLF. (e.g. fonts that work well - MS Gothic)
[3] I think this may be related to the resolution we are drawing the font
[4] It can't/will not be handled inside BFL. So the way I see it is to implement a mini text library/api that works as a middlelayer between the drawing step and BLF.
So instead of:
BLF_draw(fontid, (char *)text, strlen(text));
We would do:
MAGIC_ROUTINE_IM_NOT_BLF_draw(fontir, (char *)text, styleflag, width, height);
[5] don't hold your breath ... but if someone wants to have fun in the holidays the (4) and (5) are part of the same problem.
Code Explanation:
-----------------
The patch should be simple to read. They are three may parts:
1) BL_BlenderDataConversion.cpp:: converts the OB_FONT object into a KX_FontObject.cpp and store it in the KX_Scene->m_fonts
2) KetsjiEngine.cpp::RenderFonts:: loop through the texts and call their internal drawing routine.
3) KX_FontObject.cpp::
a) constructor: load the font of the object, and store other values.
b) DrawText: calculate the aspect for the given size (sounds hacky but this is how blf works) and call the render routine in RenderTools
4) KX_BlenderGL.cpp (called from rendertools) ::BL_print_game_line:: Draws the text. Using the BLF API
*) In order to handle visibility of the object added with AddObject I'm adding to the m_scene.m_fonts list only the Fonts in a visible layer - unlike Cameras and Lamps where all the objects are added.
Acknowledgements:
----------------
Thanks Benoit for the review and adjustment suggestions.
Thanks Diego for the BFL expertise, patches and support (Latin community ftw)
Thanks my boss for letting me do part of this patch during work time. Good thing we are starting a project in a partnership with a Japanese Foundation and eventual will need unicode in BGE :) for more details on that - www.nereusprogram.org - let's call it the main sponsor of this "bug feature" ;)
Using RMB on menus to change hotkeys was broken.
- the input button was on a weird place outside menu, assign
would close pulldown, so you had to reopen to check
- ESC didn't close the button, but assigned ESC as hotkey.
This key is a protected key, and always should be escaping.
- Worst bug: if you used this on a 'user keymap' it removed
all entries from the map...
* Depsgraph wasn't updated properly for smoke flow collision object dependencies.
* Smoke also wasn't properly using the actual emission frame of the flow particles.
* There was a lot of bloated logic in some parts of particle code so this fix turned into a small scale cleanup operation.
** As a result particle updating and cache usage should be a bit more stable too.
Selecting graph keys would toggle channel selection if shift was held.
this was annoying when selecting 2+ unselected keys to have the channel change color each click.
Now set the channel based on the selection state of the point, as long as points are being selected the channel will stay selected too.
"Continuous grab": the boundary is now set to Area (editor), making it work
for operators started in other regions (like toolbar, or in quad view case)
Added support for working on delta transformations instead of plain
transform. this should help the fact that you cant randomize animated
objects (still need to check into that)
Removed minimun scale since it was not well done, need a stronger
version of this
Changing the edge crease median value often wouldn't result in the median value entered because of clamping from 0-1.
Now the median crease is applied by scaling the values up/down.
also add some simple checks to speed up updates,
- don't move verts or recalculate normals if only crease changes.
- don't apply crease changes if location is being edited.
Problem is is with operator redo which click-extrude exposed.
Check if redo operator can run, otherwise lock the UI and add a label that the operator doesn't support redo.
This is clunky but IMHO better then failing silently and leaving the user confused.
- Merged redo functions into ED_undo_operator_repeat(), code was duplicated in a few places.
- added WM_operator_repeat_check to check if WM_operator_repeat() can run, avoids an undo call when redo work.
Unrelated changes
- GHOST_SystemWin32.cpp set to utf8 encoding.
- cmake_consistency_check.py now checks source files are utf8.
Noise texture moved the edges of multires grids different, causing
tears in the mesh. Fixed with a call to re-stitch grids (but only done
if the brush texture is set to noise)
Submitted by Martijn Berger.
Make DocumentImporter class the actual IWriter implementation and move prototype to the header.
Group together functions that we should move out of the class.
No functional changes.
Graph Editor: "make selected channels visible" VKEY didn't
make unselected channels invisble.
What is left is that the active channel remains visible still.
Not sure if that's by design, for Joshua to answer.
Old bug report:
Image Editor, Painting: crash when texture was visible in
Material or Texture preview. Was caused by paint code
freeing mipmaps. Now replaced with a mipmap tag (to be done
again), and a new mipmap function that doesn't re-allocate.
Sequencer: Properties -> Panel "Strip Input" -> file browse button
next to directory name assigned the full file name.
Added on todo: unported code still for "Change Sequence", allowing
to rebrowse strip contents.
- Offset, so you can paste at the same time of the original frames, at the current frame or relative to the current frame at time of copying.
- Merge method, so the pasted keys can overwrite keys in their range or the entire curve.
Currently there is no redo panel for these space types so the only way to access these options is with F6 redo popup.
* Smoke is now only drawn/rendered after the simulation frame range starts.
* This does not apply to simulation end frame though, so that any remaining smoke can for example be faded away nicely through material settings.
Previous commit fixing rest of handle type keymaps renamed this
operator, but that brings it out of line with rest of animation editor
operators, so reverting that fix.
Also, made Action Editor/DopeSheet use same hotkey as Graph Editor for
setting handle types now (which is same as one used in 3D-View).
broke since 2.49b, highly confusing function.
Stick bones dont draw solid so cant run glLoadName() in this case.
Tested pose/object mode with all armature draw types and view draw modes.
Root cause is that some drawing modes don't work with PBVH
drawing. Worked around by adding a call to update mesh normals from
the PBVH so that sculpted changes appear correctly in those
"unsupported" modes. (They'll still draw much more slowly than solid,
but should at least appear correct now.)
Automatically assign menu keys based on name, alternative to pressing number 0-9 on menus items.
keys are assigned by first giving each menu item the first character of any word, if that fails any key in the name is used.
- active key is shown underlined.
- only ascii keys are assigned currently.
- can run operators, open menu items.
- currently this only works in cases where number buttons were used (UI_BLOCK_NUMSELECT), but could be enabled for file menu, splash etc by removing this check.
Was another problem caused by each brush being allowed in more than
one paint mode.
Added a new field to the brush struct to indicate what mode the icon
was last set for; if it's changed then reset it. Not sure if it's
really worth it to cache this, could remove it for simplicity.
Sculpt data shouldn't be lost when making topology changes without
quads<->tris face converison.
General idea:
- Go through all grid points of each corner and convert per-corner
coordiante to per-face cooredinate
- Apply weights and convert new point to per-corner coordinate
- Use bilinear interpolation to get needed displacement vector
Some additional work was necessery:
- Two neighbour corners could have different displacements along common
boundary. multires_mdisp_smooth_bounds() makes displacement "symmetrical"
- Point could change it's corner, so displacement vector should be flipped
in some way. In some cases it's not only flipping, because corner could
be mapped with some rotation. It's not solved for triangular faces yet,
so only z-axis displacement would be interpolated for tris.
More limitations:
- Interpolation will give incorrect result after quad<->triangle
face conversion.
- When face normal was fillped displacement would change it's direction too.
Compositor/Image viewer
In 2.4x viewer nodes had animation playback. Not restored yet.
However, when loading such older files the viewer never showed
any result.
Graph editor: submenu for handle type had drawing error.
Seems to be the collumns-separator feature for enum properties?
No time for debugging that, a 5-item menu doesn't need 2 collumns.
Armature edit mode: x-mirror: "switch bone direction" now flips the
mirror bone too. It leaves the mirrored bones selected too, so
you get good visual feedback things happened there.
* New option to "Regrow hair" for each frame.
* This was perhaps more a feature request, but there was a similar useful feature called "animated hair" in particles at some point.
* The previous behavior for hair growing was inconsistent to say the least, so this is a nice option to have.
This was set to being not-animatable in r33397. However, I think there
are some valid use cases where controlling layer visibility may be a
"too heavy handed" approach.
This allows manual (point by point) animation of their control verts,
although many other settings cannot really be animated with any
visible effects yet. Interestingly, lattices also had IPO block
pointers, though they were never really used (AFAIK).
Todo:
- Animation Editor support has yet to be added. I've got a few other
things to add to, so will group those changes together.
eg, "matrix[3][2]" could be used as a driver.
Test from python
bpy.context.object.path_resolve("matrix_world[2][3]")
Before this was always treated as a 1 dimensional array.
- equals_v2v2
- project_v2_v2v2
- isect_seg_seg_v2_point
which would be necessery for my further multires interpolation commit
M_Geometry_LineIntersect2D now uses isect_seg_seg_v2_point(). Behaviour of this
function was changed a bit -- it haven't returned intersection point in several
cases when two segments are making angle.
Dropping in image/movie/sound strips was broken; it dropped
things in wrong location (frame 0), and without giving images proper
length.
The file path setting code for the operators here is complex...
Outliner icons of ID blocks can be dragged, but it also caused an
activate of the item. This prevented using outliner for dragging
into name buttons. Now outliner activates on mouse-release, waiting
for potential drags to happen.
For constraints that have 'disabled' flag (because it has invalid
input) the name was drawn in red. Easy to recognize constraints that
stopped working.
* Close event probably happens when drawable is set to other window and this messes up drawing.
* Also fixes#25071 and possibly #22477 (somebody with osx has to re-test)
graph editor:
- option "extend mode" requires channel-selection. Moved pulldown menu
from Key to Channel instead
- hotkey SHIFT+E for same operation now works both in channel list as in
main view.
Text edit mode (3d): brought back the ALT+Backspace mode for
typing accented characters. Works with an operator property,
so the hotkey for it is free to define.
Example: type 'a', alt+backspace and then 'o'
works to combine characters with ' ` ~ o / and ^
* Calling update_children(..) is very light if there's nothing to update, so it doesn't matter if it's called every time the particle system is updated.
[#25159] Vertex locations dont read correctly and are not labeled correctly in the properties bar.
- non rna buttons can now have units set.
- calls with invalid units system now raises an assert().
- include .mxf in filter.
This is need to properly handle 3d text (dalai work on GE), before
the BLF_aspect only take one argument, and the result was a call to:
glScalef(aspect, aspect, 1.0)
Now the three value are store in the font (x, y and z) and also
need to be enable using BLF_enable(BLF_ASPECT).
By default all the code that don't have BLF_ASPECT enable work with
a scale of 1.0 (so nothing change to the current UI).
I also remove all the call of BLF_aspect(fontid, 1.0) found in
the editors, because is disable by default, so no need any more.
Campbell the only thing to check is the python api, right now
I modify the api to from:
BLF_aspect(fontid, aspect)
to:
BLF_aspect(fontid, aspect, aspect, 1.0)
This is to avoid break the api, but now you need add the BLF_ASPECT
option to the function py_blf_enable and in some point change
py_blf_aspect to take 3 arguments.
I still have to learn more of the recent changes in code :)
Didn't know the handle type options became a menu for Curve edit mode.
Providing much-used tools non-modal (direct) really should have
preference. Pull down is not very accessible here though, will
check on it later. :)
This restores H, Shift+H, V and alt+H for handle setting.
Graph editor: hotkeys for handles restored; they work immediate
and non modal now (menu was ugly and slow). Uses similar options
as 3d curve editing, but not toggling 'aligned'
HKEY: sets aligned
SHIFT+H: sets auto
ALT+H: sets free handle
VKEY: vector handle
In windows this was producing Linking dependence errors because we have BLF_api.h in the /blenfont/ and blf_api.h in /python/generic/. It doesn't produces crash out of the box but I was trying to link both "api" files to the same project (Ketjsi folder). For the linking order was determining what header to use. A workaround is to "include" the file using some ../../ relative folder. But renaming it is less error prone.
Probably Ketsji folder shouldn't link to BLF_api.h anyways, but this is something I will look better later before another commit. In the mean time it's not a good idea to have 2 api files with the same name (for non case-sensitive systems).
Two isses:
- Material assigning to MetaBalls used wrong pointer
(copy paste error, casting badly).
- Checking for node-material used wrong RNA pointer
(confusing void stuff going on here!)
The error leads to corruption in data and/or random
crashes. Better update svn now!
- Pasting from 1 fcurve to any other fcurve now works (skip index and rna path checking).
- Pasting multiple fcurves between bones now works.
- If path checking fails, pasting matches indices so Scale XYZ can be pasted into Location XYZ for eg.
This option allow the user to set a 4x4 matrix to be
multiplied before draw the text, for example:
double *m;
/* Get the matrix or build it! */
BLF_matrix(m);
BLF_enable(BLF_MATRIX);
/* set color, size, etc and draw! */
BLF_disable(BLF_MATRIX);
You don't need the last line (disable), but remember
that if you use the font to draw in any other place,
the matrix will be used!.
The GL code is:
glPushMatrix();
glMultMatrixd(m);
glTranslatef();
glScalef();
glRotatef();
glPopMatrix();
Let's Dalai test this!!! :D
* Textures applied to halo materials showed influence option for normal particles. This was really confusing, and with the cleanup I revealed a couple of hidden features too!
** Particles actually allow for textures to change halo size and hardness, but my guess is that nobody knew since the names were wrong in the ui!
** I also added the option to change the "add" value with a texture, since it was just silly not to have it.
* Halo material properties are also a bit cleaner now.
The texture node is procedural; like RGB node it has no real buffers,
but allows per-pixel reading. The compositor uses nodes that directly
access buffers too, which conflicts with it... needs more design here.
Restored old functionality that just passes on preview size buffers
for nodes to prevent crashes. Giving it a render-size buffer is not
nice; the resolution-independence of texture nodes is interesting to
keep.
Solution could be:
- visually tag input/output sockets for this case (sockets with buffers,
vs sockets with values), so users know what to expect.
from the tracker:
"""The required functionality is provided by the localOrientation property: setting this value will change the dynamic object orientation. This is because dynamic object have no parent and thus the local and world orientation are identical. However, setting worldOrientation will only change the scenegraph node, which has no effect as the physics controller will reset the orientation on next physics synchronization."""
- This is working correctly but confusing, made more confusing by UI problems, corrected these and updated tooltop for 'tex.use_rgb_to_intensity'
- bad RNA arguments used for lamps (left over from RNA renaming).
- use isinstance(...), rather then comparing type() directly, this failed with lamp type checks.
- removed redundant argument to internal texture UI function factor_but().
- the texture color was drawn inactive when it was used in some cases.
Note. AFAIK its not possible to do a general check to see if a texture is colored or not, eg: its possible a plugin texture returns color in some cases and greyscale in others.
so for now always have color button active.
Remove pose channel select, this is for internal use on read/write only.
Its possible to have a convenience attribute but rather not fake bone data being in the pose channel. just access pose_bone.bone.select
- active is no longer assumed to be selected.
this fixes a simple bug - eg: Adding a new armature, entering pose mode and toggling selection failed.
- outliner editbone selection now works like object and pose mode.
- mouse selection sets the bone active even when the tip is selected.
- active, unselected bones draw as wire color with a 15% tint of the selected color.
this is actually an assert not a crash, when the lamp and view axis were aligned, the quat could not correctly be converted into a matrix.
Now fallback to the X axis in this case.
Cone can't be drawn for duplicated lamps, because lamps with cone are drawn after
main draw function (from view3d_draw_transp) and list of duplicated object gets
freed to this moment.
Disable cone draw for lamps which are from dupli.
* Particle age can now be used as the texture x-coordinate, and location in a particle trail as the y-coordinate.
* This finally enables particles in 2.5 to change their color (or any other texturable material property) by their age.
* In 2.4x this was accomplished with the "100 frames == particle age", but this was both non-intuitive and slow as the animation system had to be recalculated for every particle.
* Currently these are 2d coordinates (age/lifetime == x-coordinate, trail particle index/number of trail particles == y-coordinate), but other particle properties or possibly even a user definable property can be added as coordinates in the future.
* On the code side this uses the same coordinate definition number (for halo materials) as strand coordinates (for surface materials). This is also nice as they intuitively mean nearly the same thing, i.e. along strand or during particle life.
(This commit doesn't fix the original bug reported in the report, but does fix one that was discovered while investigating that one)
Trying to insert keyframes for nodes within group nodes was failing.
This was caused by the ID-block for those UI widgets being set to the Node Editor's current ID-block (i.e. NodeTree) vs the NodeTree that those nodes lived in. The net result was that the paths couldn't be resolved, as the paths obtained for those widgets could only work up to the group's nodetree.
An error seems to have been introduced to the node-tree building at some point, which means that the ID-type for data-attached node trees was incorrect (i.e. scene->nodetree->id.name = NTREE_COMPOSIT instead of ID_NT).
This in turn meant that the ID AnimData availability poll would fail, as the ID-type could not be determined.
- with the NLA on a small strip text was drawn under the scroll bar, now draw with same alignment as rectangle constrained text.
- single alloc per text item.
- was using opengl context rather then passing color value.
This would have made bug #25003 very simple to find.
- Objects had their quats normalized when calculating their matrix, this is inconstant with pose bones and isn't useful for animation.
Also it wasn't normalizing the delta rotation so these would give bad rotations.
- Converting between rotation modes BKE_rotMode_change_values() assumed normal length quat. changing quat to euler rotation for eg could change the bone.
- Clear rotation and transform were not normalizing the quat when 4d loc was disabled on quat rotation, corrected and also made it so the quat scale is restored after conversion so animations curves dont jump.
There is 1 case in mat3_to_quat_is_ok() where quat_to_mat3 on an unnormalized quat is needed, for this I had to add an ugly static function quat_to_mat3_no_assert(), but overall its worthwhile IMHO to be able to find incorrect use of rotation conversion.
This is a bug as old as the ability to change the actuator values through Python.
For the records: although Blender supports floats as frame values BGE doesn't. It could but it doesn't. So only integers (longs) will be valid start/end frames.
when starting playback after undo-ing topology changes
The simpliest way was to change data at which undoCurve_to_editCurve and
editCurve_to_undoCurve works from curve to object. Maybe it's not very
cool, but still better than moving curve animation handlers outside of
editcurve module.
===================
Update fcurves RNA pathes after making changes to curve topology, so
animation will no longer lost after edit and wouldn't lead to crash.
Will be very useful for RotoBezier addon.
Also NodeKeyIndex renamed to CVNodeIndex due to node is an entity from
another module and better not used in others.
- Exceptions from reading vector values in the game engine where not being caught.
- Also wrote specific KX_GameObject exceptions, without these the errors are quite confusing.
- Object actuator rotation was being scaled by: (1/0.02)*2*PI/360 == 0.872, since revision 2.
- Remove scaling and use do_versions to adjust existing files.
- Part of this report is a misunderstanding, but there was no access to bold/italic fonts.
- Added rna access and changed the operators to use only rna properties.
- skip fixing file paths on undo.
- simplify bpath alloc and free functions, also pass Main structure so as not to rely on G.main, (needed for file load).
- loop over sequencer plugin and texture voxel paths.
- fix leak in python bpy.utils.blend_path() and use PyUnicode_DecodeFSDefault() to ensure correct paths with different encodings.
- operators to make paths absolute & relative now redraw the view.
only tags the ID and does the actual flush/update delayed, before the next
redraw. For objects the update was already delayed, just flushing wasn't
yet.
This should help performance in python and animation editors, by making
calls to RNA property update quicker. Still need to add calls in a few
places where this was previously avoided due to bad performance.
this is needed because some UTF8 names didnt resolve on windows when using surrogateescape'd strings.
This meant you couldn't export to models some directories on windows.
* Old boids didn't have all of the necessary data for the new system.
* Changed the particles code so that a check for all necessary data is always done before starting actual dynamics calculations.
Issue with Library linking: if you already have libraries linked in,
and you attempt to link more data, using internally already linked
stuff (like a group with group links) the linkage would fail. It did
resolve on save-load though.
There's a long commented history in this code explaining it. I found
out it was because of manual step-by-step linking versus normal file
reads. In the first case, the "lib pointer map" was missing references.
Python Operator "Delete Edge Loop" (which keeps mesh connected)
changed selection if sliding operator failed. Simply added a
check for return value. Gosh, that Python code is not bad ;)
Crash when deleting Shapekeys, while Graph or Dopesheet was visible.
Caused by RNA collection lookup, which is not checking against index
out of range errors.
Brecht might have to give blessings for this though :)
Texture face options like "two side" were invisible, unless the
top bar "render engine" was set to Game Engine.
This is confusing, since 3d window does draw these options well.
Added back panel to show default, but with a label warning that
the options are not supported for render.
This is ifdef'd and may be disabled later on, or only enabled in debug mode.
This applies to setting any RNA value that has an ID and is not a screen or window-manager datablock.
Some addons break this rule and need fixing but from my tests blender UI scripts are ok.
The old blocking "time cursor" wasn't working anymore.
Commit 32798 overlooked that the initialization was
needed.
Now bakes show it again. Note to self: it seems to flash
slightly (like 2.49), need to check on it one day.
Material Node shaders: displaying GLSL + rendering previews crashed.
Reason was non-safe threaded access to material nodes. Now, on a preview
render, a full copy of Material is being made.
* Rendering a scene strip updated all animation data to it's frame, so fcurves were left with the wrong value.
* Now the animation data is recalculated to original frame after rendering each scene strip.
from Dan Eicher (dna)
This allows automated node tree creation and editing.
developer note, made some changes to the patch.
- removed incorrect use of container_of(), GCC only also, search for the scene instead.
- commented socket min/max access, these are internal use only as far as I can tell.
- commented group/ungroup functions, these use the selection context and are not really data level functions.
- use remove() rather then delete()
Array Modifier: edge indices could be equal to numVerts, causing
bad crashes in derivedmesh. Similar fix to previous here with
face indices. Should poke the coder of this... for now it survives OK.
Particle cache reading: crash when loading .blend on a different
endian system, code was dumping arrays in .blend without DNA.
General warning for devs: avoid generic write_data and dynamic
arrays in DNA.
GRLESS key was missing in RNA key list, so it didn't show up in
keymap editor. Internally support for it worked already.
Thanks to Milos Zajic for showing the fix.
* Cloth now uses the original derived mesh for simulation and a new dm only for applying the simulation result.
* Reverted Campbell's temporary workaround r33406.
this report raised a number of problems with rna paths, while we still dont have multi-dimensional array access, invalid paths were being accepted which confused things.
rna path resolving code was accepting all sorts of invalid input because atoi() just returns 0 for non numeric input.
now check if 0 number == '0' character.
- a linear float buffer was being created and saved into a non-linear DPX/Cineon file.
- removed the UI since the settings are not used at the moment.
added a utility function IMB_float_profile_ensure(), which returns a float buffer in the requested profile, using the existing if needed or returning an allocated buffer if the profile is different to that of the ImBuf. - Useful this case where the save function has its own linear setting.
also changed notify functions for new bone, hide/reveal, switch direction and parent.
these used ND_POSE or ND_DRAW, replaced with ND_BONE_SELECT since this changes editmode bone hierarchy/selection.
check for alpha channel while saving images that have been painted onto.
It would be nicer to do this while in paint mode except this isn't so simple with project paint using multiple images at once.
- remove report argument from console functions.
- don't update the scroll area while drawing, do this within operators instead.
- dont redraw while selecting text unless selection changes.
* Separate caches were causing quite a lot of problems both in principle and practice.
* For example it doesn't really make sense to have different frame ranges for normal and high resolution smoke, but this was fully possible before.
* Also to fully bake the smoke you had to do a "Bake All Dynamics", which completely defeats the whole point of the feature!
* As a result of this change the smoke cache usage is much much simpler and less error prone.
* This is quite a big change, but hopefully there should be less rather than more problems as a result :)
Some other related changes:
* Changing the cache name now works for disk caches properly too, it now just renames the cache files so should be faster too!
* Smoke is now always forced to disk cache with step 1 on file load as there were some strange cases where smoke was trying to use memory cache.
* Disabled smoke debug prints from console.
* Disabled changing smoke parameters when smoke is baked.
Note to users: The unfortunate side effect of this is that old high resolution simulations have to be baked again, but in return you get much better and more logical functionality. Sorry none the less!
Compositor: MultiLayer image node: on using layer menu, the entire
multilayer file got freed, causing node to collapse and open full
again after a composite.
- commented unused View3D->flag's
- popup dialog now centers over the mouse
- only overwrite image alpha with render settings on save if saving the render result.
When duplicating a mesh that has a mask modifier on it,
the invert option of that modifier dis not copy over.
Report + fix provided by Jacob F. Thanks!
Outliner: if bottom slider was hidden, you still could not click there
on items. Code was using region-mask to clip input, but the mask is not
adjusted for sliders now (keeps view same).
[#24170] Camera inside volume error
[#24838] Light inside Volume material drops on it's walls - it may be double
Problem in a previous bugfix commit, reverted back, the original bug is still fixed too.
Compositor: Texture node only allowed 1 user, with more nodes using it
there was a thread conflict, using same memory for writing values.
Also: brought back the original intention for texture nodes, which is to
be using a "procedural image", not allocating memory for a buffer, but
only allowing to read per pixel. Commit in 2007 (!) allocated full buffers
for texture nodes, without using them even.
This fixes Orig Dimension display (mostly).
* orx, ory both didn't get calculated, if dimension already matched
* putting them into Strip instead of StripData ment, that using images
of different dimensions in one strip could lead to incorrect results
Still TODO: on file open, timeline display happens before preview
display which means: orig_width and height are calculated after the
first draw of N-keys dialog. You have to hit refresh (or scrub one
frame) to get the right values displayed.
After discussing with campbell, I'm adding a menu in the addons panel to help people to find the addons development pages
This should give more visibility to the bf-extensions project so that people know where they can share their scripts "officially".
We had 3 reports of not being able to add images in editmode which was a workaround for redo in editmode bug.
Rather then having it use operator redo, just give an OK button.
- pep8 script was giving an error on non utf8 scons source files.
- use PyList_SET_ITEM macro when list type is ensured.
- all mathutils types use subtypes to create new types when available.
- use defines MAT3_UNITY, MAT4_UNITY to initialize unit matrices.
- bpy.app.driver_namespace, read-only dictionary, edit in-place.
- reset on file load and new file.
- on errors the namespace used to be reset, this should not be needed.
simple example.
def driverFunc(val):
return val * val
bpy.app.driver_namespace['driverFunc'] = driverFunc
... now all drivers can access this function.
(userpref->system), with a great help from brecht (its been way too long
for me).
However as brecht pointed out that the non-AA text is slightly lareger
then AA'ed Text :S, i did not do anything about this as this commit was
just the option not the text drawing.
this commit also makes it possible to do all kinds of UI textrender options
Crash was caused by invalid utf8 sequence pasteing from the clipboard.
Prevent memory corruption by giving utf8slen() the same rules of bytes checking
as utf8towchar() does.
added r31826.
This is valid python syntax but I rather be strict with data path format else it becomes harder to parse them if we try to support this.
it means checks like fcurve.data_path.startswith('pose.bones["SomeBone"]') isn't ensured to work, since blender uses "" quotes everywhere for keyframe paths I dont think its an advantage to allow users to do it differently.
- glReadPixels() was running to get the depth on each pixel, this works fine one some cards but was locking up on OSX.
- Replace glReadPixels() call with a single call to view3d_update_depths() right after view3d_validate_backbuf(), so the depths are only read once.
- Unrelated to the changes above, but should improve performance: view3d_validate_backbuf() was being called on every redraw while combing, now only call once when combing starts.
we were using SENSOR_RAY for the radar sensor axis. However the Ray axis is inverted (God knows why) so I created a set of defines only for radar sensor.
Also I thought it was a good idea to replace some hardcoded values in Radar and Ray codes by their defines in DNA_sensor_types.h (similar to what Benoit did for Armature Sensor, so I see no problem on that).
from Josh Wedlake (joshwedlake), who provided a reference script used to apply changes in ED_rollBoneToVector().
- Obvious bug fixed where Z-Up didnt work right.
- More align axis options to Recalculate Roll operator: X/Y/Z/View Axis & Negate.
- Axis Only option, ignore the axis direction, use shortest rotation to align bones.
ED_rollBoneToVector() changes:
- would give bad roll when the axis wasn't normalized or perpendicular to the bone.
some callers accounted for this but not all.
- option to align to the axis but not the direction.
Binding code had off-by-1 error, which meant that when "Even Divisions" was disabled the length of the wrong bone would get used.
This error was most noticeable when the lengths of the bones were quite different - for example, a chain with 3 bones of increasing length. Thanks to "Julius" on BlenderArtists for catching this. Cheers!
---
Also, simplified the binding code loop a bit to prevent this sort of error in future.
Crash in snapping, bvh tree. Wrong check for numFaces here.
Fixed the crash by adding test for face pointer, but not sure
what the coder intended here. Needs investigation, left XXX remark.
When report error was drawn in info header, using border select
or paint brushes flickered. Was caused by Triple Buffer method
not checking for correct redraw case.
Also made report redraws less aggressive, it was drawing the info
header all over with 50 FPS for 10 seconds. Made it 20 FPS, and
added code to only send notifiers for actual changes.
As todo note for future: animated UI options could get better
caching to cope with slower refreshes.
Logic Window: Actuator type options in menu didn't show special options
for Armature or Mesh when 'show selected' was used.
Dalai proposed a nice new RNA feature for inherited types, but with
only two exceptions handled, it was easy to code this check.
* Fluid baking (using the job system) didn't update the "lastgoodframe" anymore, so reversing the frames didn't work. Now the last valid frame is checked by going through all fluid bake files when "reverse frames" is selected.
* There was all kinds of fancy checks done in the fluid modifier for reading a different frame in different cases, but as the "lastgoodframe" was really not working I don't see the point of this whole code, so removed it for now. The new functionality is: if the fluid data for current frame exists use it, otherwise just return unmodified domain mesh without any fancy backup plans.
* There were also some errors on reading uncompleted files (scrubbing timeline while bake was running), so I made the fluid file reader just return null if the number of faces didn't correspond to to actually read data. Previously this just printed an error to the console.
This is an annoying but which isn't a problem for Python because they don't execute multiple scripts, one after another (there is one __main__ and everything else is a module).
So when the __main__ module in sys.modules is overwritten, it decref's the module and clears the dictionary with _PyModule_Clear(), even though the modules dictionary is still in use.
Strangely this problem only happens with Python3.1.1 and Python3.2x svn but not 3.1.2
This commit restores the namespace after _PyModule_Clear() sets all its values to None.
This happened when removing vertex normal flipping in the render engine. But
actually the normal of the first face in the curve was flipped on rendering,
because it used a different vertex order than other places. Also removed the
hack that computes a face normal only from a triangle instead of a quad.
On entering editmode Curve/Nurbs, the undo buffer was not
correct; storing old Curve/Nurbs object name. This caused
for example the Operator tool redo options to fail; it added
another curve on first use (with user pref "enter editmode).
Solved with adding good undo push + nicer code for curve and
nurbs naming.
- ortho grid now draws scaled by the view3d 'Scale' setting, venomgfx noticed this was missing.
- so as not to confuse add scale next to unit display text, so rather then "Metres" it shows "Metres x 1.5" otherwise its confusing that grid lines are not in exact units.
- changed grid spacing to grid scale (needed for more logical behavior with units)
- when units are enabled grey out subdivisions.
- Python calling operators didn't run WM_operator_properties_sanitize() so enum functions called from python were given a NULL context.
- PROP_ENUM_NO_CONTEXT and PROP_NEVER_NULL used the same value in the enum (possible conflict).
Playback now works. The problem was that material/texture nodetrees were getting ignored completely, as I was assuming that all of these existed in the main->nodetree collection when in fact only the "group" nodetrees lived there. I don't really agree with this, but that's the way it is...
TODO: animation editor support still to come
Color picker: RMB "insert single value" was inserting all three values.
Same case as yesterday, made code recurse into open menu to check for
active button there.
[#24849] changing objects to another layer causes segmentation fault
[#24848] Using an operator outside of edit mode crashes blender
[#24844] Crash related to the subdivision (aka subsurf) modifier
[#24843] ctrl+z crashes blender
* documented and rewrote the render interface of the sequencer.
(now, the geometry / render_type / etc. settings are stored within a
seperate structure called SeqRenderData that is passed within the code.)
* that fixes
* cache problems, since the caching system didn't keep track of
proxy files vs. final renders.
* is a necessary step, to bring back frame blending in speed effect
(the SeqRenderData structure elements are already there)
* will make motion blur render options available within the sequencer!
* this patch also fixes:
* "easy retiming" using speed effects. (in Blender 2.49, you could
add a speed effect and resize the source track to retime it to that
length)
* adds labels for the Original dimensions for Image + Movie tracks
(worked in 2.49, too)
With some own changes:
- Select pick moved to invoke()
- Used editsurfcurve_region_view3d as poll function for this operator
due to ogl dependency
Thanks to Elia Sarti (vekoon)!
(Well, while testing this report I found this fix!)
Blender could crash after rendering a 2.4 or 2.5 file
The context before/during/after file reads is still fishy.
Need a more clear & clean mind to really check it from
scratch again.
[#24652] Project vertices button showing in object mode and leads to wrong behavior.
Removing soft redraw (code is still there in case we need it eventually) on mouse move, now that events are compressed.
Color Picker: when inserting keys (right mouse menu), the colors didn't
change for buttons to indicate such. The function that sets the flags
returned too early.
Fixed bug #20620, "VertColors and Flat/Soft imported from 2.49 are wrong
(MultiRes)" reported by Manuel R.
* Added function to load level-0 vertex colors
* Added function to load level-0 face flags
* Warning: the 2.5 multires modifier doesn't support multires vertex colors
or multires face flags; that data will be lost if you import it into 2.5.
- Image Editor, "New", gave for each tweak in redo menu a new Image
when Object is in Editmode. This is a limitation of our current
undo system. Marked this issue in the wiki todo.
Solved by adding a poll in operator that prevents this to be called.
(a bit annoying, but the error is worse!)
- On assigning a new image texture to faces in Editmode, no redraw
happened in 3D window. Added notifier for it.
- when python operators fail to execute they were returning RUNNING_MODAL, now return CANCELLED
now when an operator fails it gives an error popup as well as a message in the terminal.
- Reading 2.4x files could cause print "missing region type".
Appeared the 2.50 do_versions patch differed from default region
definitions
- Sometimes editors showed wrong button for type browsing.
Was because variable wasn't correctly reset on saves.
* Added the options in UI and RNA
Funny Note: This was on my own to-do list (see commit 20577 in the py file from 2009-06-02 when I ported the UI from C to Python) :P
- missing break in operator context switch meant RGN_TYPE_CHANNELS was always being overwritten with RGN_TYPE_PREVIEW when calling operators (from r26692)
- reload() has been removed from pythons builtins. use imp.reload() instead (still need to apply globally).
- fixed own mistake, not ignoring 'filter_glob' as a keyword argument (broke fbx, obj, 3ds export)
also added note that adjusting bone radius changes the parent bone for connected child bones, and fix typo on failing to read startup.blend (both pointed out by MikeS on IRC)
- no rotation resulted in NAN location.
- subtraction of pivot done in wrong order made the constraint give odd results when rotating on more then 1 axis.
Removing a scene would set the active scene, assuming the scene removed was the current scene. This broke UV Export which used a temp background scene.
* The SPH fluid particle algorithm was implemented a bit wrong. This problem could for example result in the fluid moving sideways after being dropped straight to a horizontal collision surface, a very big no-no as far as real world physics are concerned!
* After some extensive code shuffling the algorithm is now much more true to the paper it was implemented from, and more importantly now the physics should be correct too!
* The main thing was that fluids calculations can effect many particles simultaneously, so just a single loop through all particles can't work properly. As a side note this also means that the actual fluid algorithm can't be made threaded :(
* To make things work I also had to reshuffle some general particle physics code, but there should be no functional changes what so ever to other physics types, so poke me immediately if something strange happens.
Note to users: these changes will most probably effect the way previously done sph fluid simulations look, so some parameter tweaking will be needed to get things back looking the way they were.
OpenGL viewport render gave squeezed results in cases.
Reason: some graphics cards only give offscreen buffers in multiples
of 256 or 512 (my case).
Current fix uses the actual size returned by graphics card, which
is also safe for too large renders.
More elaborate cropping or matching is for another time.
(Added printf for feedback on this, might disappear)
When using Material Nodes, there was no indication in Material
Property window which node was active. The context/channel widget
now shows this.
Better would be to be able to browse nodes in this list, and
to have option to preview only this material (not node tree
result). Enough todo for future :)
this is a can of worms, at the moment blender depends on broken behavior for metaballs:
find_basis_mball() can return a metaball object that fails a is_basis_mball() check which makes this logic very confusing (added note about this in mball.c).
Metaballs needs a refactor however at least make drawing fail consistently,
For wire draw is_basis_mball() wasn't being checked, for solid drawing it was (hence the strange wire frame).
For now the motherball needs to exist in the main scene else it wont work.
scene.frame_set() was sending out a frame change notifier which would update the scene loosing any un-keyed changes after a script ran.
for now use a window redraw notifier.
Operator redo in 3D window now moves context to the main region,
makes knife cut work (and probably others)
Todo: this callback should be shared with other redo panels. On the
todo list.
Editmode undo operator was using the operator OPTYPE_UNDO flag as well as its own EM_DO_UNDO option.
This gave 2 problems.
- python suppressed undos would fail (as in the report).
- Undo push was called twice on exiting editmode for no reason.
Use the operator undo in this case.
Tweaked keymap for Graph Editor Ctrl-LMB so that the popup menu for displaying reports won't get dismissed immediately. Changing keymap here from KM_PRESS to KM_CLICK
* Keyframing operators now use the reports system for displaying all its error messages.
- The benefit of this is that users no longer need to check the console for error messages if keyframing fails.
- Unfortunately, reports are not currently viewable in any space/view in Blender, so...
* Added a temporary operator (UI_OT_reports_to_textblock), which can be accessed in the UI from the button which appears in place of the icon when more than one report exists. This dumps the current list of reports to a textblock "Recent Reports", from which they can be viewed.
This isn't really nice, but at least we now have a way to view these again, which makes debugging some things a pain.
* Bugfix #24606 - when trying to add keyframes to F-Curves with F-Modifiers already which alter the curve significantly enough that the keyframes will have no effect, there are now warnings which aim to alleviate any confusion.
Also use const char in many other parts of blenders code.
Currently this gives warnings for setting operator id, label and description since these are an exception and allocated beforehand.
Disable projection painting for soften tools.
Also it looks like smear tools support projection in the code, so
enabled project paint menu for such tools
Here is a image of it in action:
http://www.pasteall.org/pic/show.php?id=6351
What it monitors:
* VRAM used by textures created via bf_gpu and BL_Textures
What it does not monitor:
* VRAM used by the Blender ui
* VRAM used by 2d filters
* VRAM allocated by the user via KX_Scene.pre_draw and KX_Scene.pre_draw
from Dan Eicher (dna)
Made some edits to the patch because only the first spline was getting the updates.
also added path functions for splines and spline points (request from ZanQdo so splines can be keyframed).
Very old bug: Material/Render: the "filter" option for transparent
absorption also works for alpha==0, but in that case another part
of render code just didn't add the faces in render list.
The main reason to have this is so renders can be scripted to write to a specific file without having to do annoying tricks like set a dummy start/end frame range, render an animation and work out the current frame image will be written to, then rename after rendering.
Also made some 'char *' args into 'const char *'
Using texture nodes for displacement crashes, a conflict between using
nodes in render and for displacement at same time. Code simply disables
texture nodes for displacement for now, until we have better handling for it.
(on our long todo)
- don't rename the image when saving (problematic if the image is used with linked libraries)
- file writing and path updating was duplicated for multilayer images.
It was called "map on length", but it actually makes the UV
values of curves/nurbs to become used as "Generated" texture
map input.
Sorry; this might break own local UI py script saves... :)
- Warn for armature deformed meshes which are scaled, these don't work quite the same as in blender, reported as [#24663].
- Use matrix.decompose() to convert a matrix to loc/rot/scale.
- get vert/edge/face lists for each mesh only once.
- faster euler rad -> deg conversion function.
- incorrect range check broke ZYX euler rotations, use MIN/MAX constants so this doesn't happen again.
- BGE Armature PyAPI also wasn't using correct min/max with rotation modes.
- clamp on file read rather then when calling the rotation functions, so developers don't use invalid args without realizing it.
- added assert() checks for debug builds so invalid axis constants don't slip through.
Doing F1-load a lot of times on same .blend could crash.
Janne karhu provided a potential fix, which is good to add
anyway. Will ask him to verify too.
Added XXX warning for these lines, after filesel exec no
context variables should be re-used. Is for later investigation.
Added convenience function to operators, 'as_keywords()', so operator settings can be passed directly to a function as keyword arguments.
The problem in this case was that dictionary access to operator properties was not returning rna-property defaults, so as_keywords() ensures all defaults are set.
reported by Eclectiel L
When working with very heavy scenes Blender can seem to 'hang' (not responding). Key events that happen
during this period may get lost, especially for modifier keys.
Adding extra handling to account for these situations.
Found and fixed a few problems here, but strangely I don't recall seeing any of these a few months ago when this conversion (probably last) worked well ...
Reported by Ruud van Eeghem
Instead of running vcredist, which can fail when installing as normal user, have the DLLs
like our zip distribution bundled already inside install dir.
- VBO UV Edge display wasn't allocating a large enough array for drawing.
- VBO UV Edge drawing was using an edge flag with faces.
- notifier for modifiers updating the UV window.
- pose bone 'matrix_local' wasn't well named since it didn't work like object or regular bones.
- pose bone matrix values for rna had array access rather then 4x4 matrix access.
note: for pose bones update scripts by renaming 'matrix_local' --> 'matrix_basis'
This added a poll function to the Panel base class but this means that all panels register as having a poll() function which is called a lot and does nothing, raising an exception and removing it because of how python implements hasattr().
Found when checking why exceptions were continuously raised on redraw.
Martin, can you check on netrender?
Rather then applying the proposed fix, enabling 'Apply Transform' works in almost all cases including the reported one.
There are some files that don't work with 'Apply Transform', which is why the option is worth keeping.
also remove unneeded calls to paint_facesel_test()
Fixed old annoyance in Search menu, with a load of object-mode
operators showing up in editmode.
It's much cleaner now, but it will take further work and
investigation to have context & polls work satisfying for all cases.
ZCombine node in compositor ignored alpha values. Enabling it
would break compatibility though, so it's a tickbox option now.
Patch provided by reporter, Martin Lubich. Thanks!
Now object_apply_mat4() can be used as the reverse of object_to_mat4().
Noticeable result is fly mode and 'Apply Visual Transform' dont jump when deltas are used, also means setting matrix from python works as expected.
- Reverse vector * matrix multiplication order. now this matches how numpy works.
- Disallow 'matrix * vec' and 'quat * vec', now it raises an error.
- Add missing in-place multiply 'vec *= quat'
Many scripts will need to be updated for this but at least it will error rather then failing silently.
Was causing compile troubles with MSVC:
wm_operators.c
C:\Program Files\Microsoft Visual Studio 9.0\VC\INCLUDE\io.h(322) : error C3163: '_write': attribute
s inconsistent with previous declaration
C:\Program Files\Microsoft Visual Studio 9.0\VC\INCLUDE\io.h(224) : see declaration of '_wri
te'
scons: *** [C:\blenderdev\b250\buildvc\source\blender\windowmanager\intern\wm_operators.obj] Error 2
Define WINVER=0x0501 for MinGW due to some stuff requres this winver, but
MinGW's default is 0x0400 (thanks AlexKu for point)
Better to replace all #include <windows.h> with #include "BLI_winstuff.h" to
avoid possibility of different functions prototypes in different modules.
* In addition to fixing the memleak it's much better to always copy the voxeldata to the texture. Smoke data can change at any time due to some changes, so we can't depend on that data.
* Thanks to MiikaH for finding this!
Texture Nodes didn't work for normal output anymore anymore.
It was caused by new handling of bump code using derivatives,
but that's not how texture nodes work now. In case a texture
has nodes, it uses the old (2.4) method.
Crash in Array Modifier. The merging-vertices option
(mysteriously named "First Last") caused invalid indices in
faces to exist. This commit solves the crash, but not the
cause. For that Ben Batt's assistance is needed.
Ton will work on moving the File menu out of the Info space before release.
notes.
- reply Operator isn't working anymore.
- UI for reports is commented out so its not mixed with the file menu.
F-Curves tagged as "disabled" now have their channels drawn with a red line underlining their names.
"Disabled" F-Curves are skipped for evaluation, and typically result (for example) from assigning an action from one armature to another, but the new armature does not have some of the bones the action's F-Curves need in order for the datapaths (referring to the property the F-Curves affect) to be resolved. This is to prevent heaps of invalid channels slowing down animation playback.
I've also added a new operator, found by:
Channels -> Revive Disabled F-Curves
in the Graph Editor and DopeSheet/Action Editors, which will clear all the disabled tags for all the F-Curves in the animation editor at the time (based on the filtering criteria).
Use this operator to clear the disabled tags, allowing such channels to be able to be evaluated again (perhaps after adding the offending bones for example, or when using the action on the original armature again).
This was a 2-part bug:
1) "Track index" (i.e. "index of NLA track within NLA Stack") was calculated incorrectly when the transform data is created. A typo meant that this was actually "index of strip within original track"
2) Strips were flying around madly vertically (especially for downward movements, and with extreme difficulty for upwards) as the track index was being updated incorrectly. I suspect at the time I was misreading it as trackIndex = ... vs trackIndex += ..., though it obviously looked wrong looking at it now!
Transforming strips into locked tracks meant that they were not unlocked properly (i.e. they stayed as temporary "meta" strips). This is now taken into account as part of the step which checks if there's any space for them in those tracks.
NURBS surfaces always used resol{u,v} from spline and never used curve's render resolutions.
Now, if curve's render resolution is non-zero then it'll override resolution for all
splines when rendering (in needed direction only, ofcource).
When using a startup.blend without user prefs (themes etc) saved, Blender
crashed. Added a provision to then fall back to the compiled in startup,
including a print in console.
View2d custom keymap to put zoom in RMB didn't work, the modal handler
still has a hardcoded check for mouse/keys. Fixed it for this case quickly,
added wiki todo and warning in code.
* This null check just prevents the crash, but the real bug is that 3d view toolbar redo even shows the image view operator..
* ..and since the redo is in the wrong view it doesn't work. Both of these are known issues.
By just removing these vars (commenting them out), we were getting alignment errors which were causing crashes when loading old files with armatures (crash in armature drawing code).
Fortunately I happen to be working on a little project right now, and caught this :)
- Totally get rid of old active_brush_index stuff
- Use operator to set n-th brush as active in specified paint mode
thans to Cambo for it's optimization :)
- report errors when join fails.
- toggle editmode with internal functions rather then ED_object_enter_editmode / ED_object_exit_editmode, which rely on context too much and didnt work properly when calling join from python.
Issue: in user preferences window, using file selecting caused the
the userpref window to be saved, and not closing.
Reason: design error (by me) in using screen->full tag for denoting
a temporarily screen (like file window). Fixed by using a new
screen->temp variable for it.
System remained unstable though, noticed another issue with freeing
temp screens in wrong places. Seems nice stable now! Will check on
the wiki for relarted issues now.
* Particles didn't want to stay cached, even if there were no actual chages.
* Particle states weren't set properly for times before actual simulation start.
Was trying to decode an image when quicktime wasn't initialized.
There are a few other reports on windows about thumbnails crashing which this probably fixes.
This also fixes a bug where DisposeHandle wasn't called if decode was called with a NULL pointer.
[#24601] Net rendering master node fails to send/retrieve files to/from slaves
Both Reported by Dan McGrath
The second might not actually be fixed with this, but now there is better error checking and reporting.
Misc: add poll method in Panel base class, to better support mixins with different inheritance order.
- object updates were not being flushed, so children weren't updating.
- apply the matrix relative to the parent, added this as an option to object_apply_mat4() which allows assigning the worldspace matrix in python without worrying about the parent.
Image editor & texture properties, add new Image, inits 'start' now
to frame 1 for sequences.
Added triple-X warning in Image Open operator, it uses SpaceImage whilst
it can be called from other editors. Code is safe but not correct.
- Home or Numpad Period with smoothview disabled were not syncing up the other views.
- Disabling clip only disabled clip syncing but left clip enabled for all views.
- Clip was being calculated for every update even when not enabled.
- The perspective view was being used to copy settings from when changing box & clip settings, resetting the distance each time. Now use one of the aligned views instead.
disallow cloud depth over 30, it causes a lot of floating point exceptions (and nan value pixels), because of hitting the int limit with 1<<(oct+1) in BLI_gTurbulence
* image name is now set when the file is actually saved if it is empty. Otherwise it is only changed if we're not saving a copy.
* If previous image name is empty, filename will be 'untitled.png' for png saving.
NOTE: if we saved a copy of 'render.png' as render1.png, the save image dialog will be re-opened with the image name that Blender has (render.png) rather than render1.png.
Quad view split 3D window: didn't work for "box lock" option when
using "view all" or "view selected".
Made it work nicely for animated 'smooth' view too.
This commit restores the ability to reorder channels in the animation editors (DopeSheet/Action/etc., Graph/Drivers, NLA). The hotkeys for this are:
- Shift-PageUp = Move Up
- Shift-PageDown = Move Down
- Ctrl-Shift-PageUp = Move to Top of List
- Ctrl-Shift-PageDown = Move to Bottom of List
Do note that only animation data can get reordered using this.
So, do not expect to be able to change the object order (that is actually taken from the order that Blender actually evaluates them per update/frame).
---
In the process, I've fixed a couple of other bugs:
* Removed 'optimisation step' check in anim_filter.c for dopesheet with no filtering options modified, since this meant that the ANIMFILTER_ANIMDATA data filter was not getting processed (and potentially there were other maintenance problems with that).
* Made NLA Editor's channel list not totally duplicate the basic Animation Channels keymap. Instead, the "NLA Channels" keymap now only defines the parts that are different, and this then gets specified before the standard one so that these different parts will override the standard ones.
* Attempted to fix BorderSelect on NLA Channels list. Still not totally working correctly yet though.
* Moved "Euler Discontinuity" menu entry from "Channels" to "Keys". It really belongs in the latter, since it affects the keyframe values, rather than some aspect of the channel (i.e. ordering of channels or how they're displayed)
Added two user preferences for filebrowser:
1. Hide Recent Locations
Requested by Kernon Dillon. Reason: People doing video tutorials might not want to expose the name of other projects they are working on.
2. Open Filebrowser in thumbnail view for images and movies
Requested by Sebastian König at the Blender Conference. Reason: In the past (and most likely still) some corrupt images or movies could crash the filebrowser when generating thumbnails. (crashes in ffmpeg, libtiff, libjpg). While many of those were solved, artists in a production environment might want to use this setting to prevent any crashes.
NOTE: the second setting should probably be set next time we create a new default startup.blend. Until then users need to switch this on manually in the User Preferences.
Bringing back missing feature: Create new directory by typing a not existing name into the directory button.
Note: Small issue still with autocomplete -> if typing the new directory directly after autocomplete, it doesn't execute the operator yet.
Also fixed some minor compile/cleanup issues with warning about signed/unsigned comparison and missing header.
- Revert of my old change in curve->mesh conversion
- Do not ignore DL_POLYs for surfaces -- they will never be filled,
but ignore them for 2d curves -- they'll be filled with INDEX3 parts.
- use sizeof() in more places.
- fixed some off by 1 bugs copying strings. setting curve font family for instance was 1 char too short.
- replace strncpy and strcpy with BLI_strncpy
when selecting paths from the user preferences always default relative option to off.
otherwise you get paths like '//..\..\..\..\Program Files\GIMP-2.0\bin\gimp-2.6.exe'
Now the active strip doesn't just get cleared on fileload, but is relinked properly.
I had originally intended that files shouldn't be able to be saved with NLA data still in Tweakmode, but this turns out to be a bit more troublesome to get working as that would make undo keep popping out of this mode too.
Also reverting 32743 (bugfix for 24418), which was a hack around this.
* WM_timecursor is broken somehow, so pointcache baking makes the cursor disappear.
* For user this seems like blender has frozen although it's just a matter of no progress indication.
* This fix just sets the default "busy" cursor for the whole duration of baking and reports progress in the console.
* If there's and easy fix for this then it should probably be done straight away, but I want to re-implement baking using the job system soon anyways, so a proper fix for this is not strictly necessary.
WM_operator_poll() could fail in cases WM_operator_name_call() would succeed because calling the operator would setup the context before calling poll.
this would result in python raising an invalid error or menu items being greyed out.
now python can also check with an operator context:
bpy.ops.object.editmode_toggle.poll('INVOKE_SCREEN')
Comment from source...
/* re-calculating the frame positions means we loose our original transform if its not auto-keyed [#24451]
* this hack re-applies it, which is annoying, only alternatives are...
* - dont recalc paths.
* - have an object_handle_update() which gives is the new transform without touching the objects.
* - only recalc paths on auto-keying.
* - ED_objects_recalculate_paths could backup/restore transforms.
* - re-apply the transform which is simplest in this case. (2 lines below)
*/
Martin, if you think this workaround is unacceptable, then automatic recalculating of paths after transform should probably be disabled since it looses data on non transform un-keyed values.
fix included in report from Martijn Berger (mberger)
made some small changes.
- use ints rather then unsigned long for printing, values are not likely to be very large.
- CMake remove strict flags from collada build dir since I had warnings in the collada headers.
- added xml2 to collada libraries else I couldnt get collada building.
the brush system matches the brush mode with the object mode, but this doesn't work for 2D image view paint.
since the poll() function doesnt have access to the context, for now just check if no paint modes are active, default to texture paint.
The scenes 'Render' is kept by blender while blender runs but the callbacks were not cleared when the render was done.
In this case the callback would reference a freed render job.
This isn't normally a problem because on re-rendering new callbacks are set, however the sequencer can render a previously rendered scene without setting up callbacks.
Simple fix is to to dummy callbacks applied onto the scenes 'Render' struct once its finished.
Added extend option to lasso.
also...
- selecting bones wasn't checking their layer of if they were hidden in a number of places.
- fixed memory leak.
small unrealed changes
- added PBONE_VISIBLE macro
- renamed functions used for paint selectoin from *_tface to paintface_*. sine they no longer have anything todo with tface's.
- removed scanfill include from BLI_blenlib.h, this is only used in very few places and quite specific.
Noticed lasso select is broken for metaballs and face mask mode but this has been the case for a while, will look into it next.
* This caused nearly all particles to leak through the collision surface if simulation subframes were used and the collision object was moving.
* In addition to fixing this I also did some more cleanup of the collision code and refined some of the comments.
- use first selected if non active (clears confusion since this isn't at all clear)
- dont take unselected into account when calculating connectivity, would mess up in many cases.
- Setting operator properties was broken because of bpy_types.py meta-classing (surprising this wasn't noticed before!)
- Presets now use a dialog with an OK button.
- Presets use a check function on the filename so invalid chars are replaced editing.
- Submit docs operator was broken.
also known as: broken multicam strip caused by other fix.
Calculated render_size where it belongs (within the glow effect) and
restored old functionality.
also: renamed render_size to preview_render_size at all relevant places,
where the naming wasn't used correctly.
Hopefully it's now a little bit more clear.
render_size := render size from scene (just rescales width/height)
preview_render_size := preview render size from sequencer preview,
controls the resolution and the use of sequencer proxy sources
This fixes a rather subtle seeking issue with ffmpeg and Sony
XDCAM-footage.
Problem is: MPEG2 streams within an MP4 container can contain a start
time - at several places. There is a starttime within the video
and audio streams and one within the container.
FFMpeg commandline tool only uses the container starttime and we used
the stream starttime.
The world would be a better place, if those two timestamps always match
up, since in XDCAM-footage those two starttimes differ in 4
frames - and the container has the right one.
We now always use the container start time as ffmpeg commandline tool
does (in the hope, that there is a good explaination for this and this
is the right thing(tm) to do).
I tested this also with HDV footage, which seems to work with the new
code, too.
Additional fix: disabled seek_by_bytes again, since it will only work
correctly, if ffmpeg guessed the HDV bitrate right (which it doesn't).
If you have seeking issues with HDV and have an older version of ffmpeg
installed, please upgrade, newer versions have some fixes in them.
* FSA and motion blur can't work nicely together the way they're currently implemented, so I disabled this in the ui and code.
* FSA is used if both are selected.
* Also changed the name "Full Sample Motion Blur" to "Sampled Motion Blur" to avoid confusion with full sample anti-aliasing.
- fixed this error 7 different functions (deform groups, uv layers & similar).
- support for numbers over 999.
- renamed splitIDname() to BLI_split_name_num(), moved to BLI_path_utils
* Built-in filters were not being analyzed, which means no depth or luminance textures for them
* Removed an unnecessary if that becomes really hairy when new built-in filters are added (ie, when filters are defined beyond the value used for custom filters)
The undo problem was caused by python operators returning 'RUNNING_MODAL' rather then the return value from wm.invoke_props_popup(self, event) - 'FINISHED'.
This was done because returning FINISHED would free the operator causing the buttons redo handler to try and run a freed operator and crash.
So the real fix is to disallow any operators to use wm.invoke_props_popup(self, event) if they dont have the REGISTER option enabled, fixing the crash and redo problem.
from mathutils.geometry import PolyFill
I couldn't find a way for python's inittab to do this so just inserting mathutils.geometry into sys.modules manually.
lamps require a depsgraph update.
There is no more need to ED_render_id_flush_update because this is called by the DAG_id_flush_update.
help (and the right fix) from Brecht.
Patch by Dan Eicher. In 3dview properties you can enable this. Rotating view then uses cursor as pivot point. Note that with this option enabled just relocating the 3d cursor also changes the 3dview.
Sebastian König: "Now the 3d cursor is actually useful" (sitting next to me while applying patch at bconf hostel)
- ascii text drawing functions, slightly faster since they dont have to do hash lookups & utf8 conversions for each char.
- used ascii drawing functions for the view3d's number display.
- each text item was using fixed 128 chars, now only allocate the string length needed.
* The logic in some parts of the sequencer code was rather cryptic, so I cleaned it up a bit.
* There should be no functional changes what so ever from these changes.
mat4_to_eul & mat3_to_eul are already working this way.
Without this we get problems with constraints, eg:
rotation on the Y axis over 90d can be represented by setting the X and Z to -PI, Y would decrease to 0 (infact 180d).
* The axis draw function was changed by Campbell recently, but the brightness value was forgotten?
* Solved currently by mapping the brightness value to axis alpha.
Hello, from the bconf 2010 from Jeroen and Luca. Our first combined commit :)
Automatically create sdna documentations from Trunk.
Usage:
blender2.5 -b -P BlendFileDnaExporter_25.py [-- [options]]
Options:
--dna-keep-blend: doesn't delete the produced blend file DNA export to html
--dna-debug: sets the logging level to DEBUG (lots of additional info)
--dna-versioned' saves version informations in the html and blend filenames
--dna-overwrite-css' overwrite dna.css, useful when modifying css in the script
Examples:
default: % blender2.5 -b -P BlendFileDnaExporter_25.py
with options: % blender2.5 -b -P BlendFileDnaExporter_25.py -- --dna-keep-blend --dna-debug
* Related to bug report [#24329] scene length not updated in VSE
* This data is nice to have in the sequencer without having to go the actual scene that's added as a strip.
* No way currently to know the order of effect inputs, so I added a swap operator for the inputs.
* Also added the effect inputs to the strip property panel (weren't even in rna before). These are not yet editable, but can be very helpful in determining what the inputs are if the strip is too short to see the name in the timeline.
screen verts were being rounded so resizing a window would end up moving the border in one direction.
for this to work properly we should store floats internally but for now compensate for this.
The remains of a RawInput mouse attempt are included, but disabled. RawInput will still be used for multi-axis devices.
Eliminated the need for several #defines by requiring WinXP or newer.
(BoolVariable('WITH_BF_PYTHON','Compile with python',True)),
(BoolVariable('WITH_BF_PYTHON_SAFETY','Internal API error checking to track invalid data to prevent crash on access (at the expense of some effeciency)',False)),
('BF_PYTHON','base path for python',''),
('BF_PYTHON_VERSION','Python version to use',''),
('BF_PYTHON_INC','include path for Python headers',''),
Saving complex scenes in Blender is done within seconds. Blender
achieves this by saving data in memory to disk without any
transformations or translations. Blender only adds file-block-headers to
this data. A file-block-header contains clues on how to interpret the
data. After the data, all internally Blender structures are stored.
These structures will act as blue-prints when Blender loads the file.
Blend-files can be different when stored on different hardware platforms
or Blender releases. There is no effort taken to make blend-files
binary the same. Blender creates the blend-files in this manner since
release 1.0. Backward and upwards compatibility is not implemented when
saving the file, this is done during loading.
</p>
<p>
When Blender loads a blend-file, the DNA-structures are read first.
Blender creates a catalog of these DNA-structures. Blender uses this
catalog together with the data in the file, the internal Blender
structures of the Blender release you're using and a lot of
transformation and translation logic to implement the backward and
upward compatibility. In the source code of blender there is actually
logic which can transform and translate every structure used by a
Blender release to the one of the release you're using [ref: <ahref="http://download.blender.org/source/blender-2.48a.tar.gz">http://download.blender.org/source/blender-2.48a.tar.gz</a>
This section explains how the global file-structure can be read.
</p>
<ul>
<li>A blend-file always start with the <b>file-header</b></li>
<li>After the file-header, follows a list of <b>file-blocks</b> (the default blend file of Blender 2.48 contains more than 400 of these file-blocks).</li>
<li>Each file-block has a <b>file-block header</b> and <b>file-block data</b></li>
<li>At the end of the blend-file there is a section called "<ahref="#structure-DNA"style="font-weight:bold">Structure DNA</a>", which lists all the internal structures of the Blender release the file was created in</li>
<li>The blend-file ends with a file-block called 'ENDB'</li>
The first 12 bytes of every blend-file is the file-header. The
file-header has information on Blender (version-number) and the PC the
blend-file was saved on (pointer-size and endianness). This is required
as all data inside the blend-file is ordered in that way, because no
translation or transformation is done during saving.
The next table describes the information in the file-header.
</p>
<table>
<caption>File-header</caption>
<thead>
<tr><th>reference</th>
<th>structure</th>
<th>type</th>
<th>offset</th>
<th>size</th></tr>
</thead>
<tbody>
<tr><td>identifier</td>
<td>char[7]</td>
<td>File identifier (always 'BLENDER')</td>
<td>0</td>
<td>7</td></tr>
<tr><td>pointer-size</td>
<td>char</td>
<td>Size of a pointer; all pointers in the file are stored in this format. '_' means 4 bytes or 32 bit and '-' means 8 bytes or 64 bits.</td>
<td>7</td>
<td>1</td></tr>
<tr><td>endianness</td>
<td>char</td>
<td>Type of byte ordering used; 'v' means little endian and 'V' means big endian.</td>
<td>8</td>
<td>1</td></tr>
<tr><td>version-number</td>
<td>char[3]</td>
<td>Version of Blender the file was created in; '254' means version 2.54</td>
<td>9</td>
<td>3</td></tr>
</tbody>
</table>
<p>
<ahref="http://en.wikipedia.org/wiki/Endianness">Endianness</a> addresses the way values are ordered in a sequence of bytes(see the <ahref="#example-endianess">example</a> below):
</p>
<ul>
<li>in a big endian ordering, the largest part of the value is placed on the first byte and
the lowest part of the value is placed on the last byte,</li>
<li>in a little endian ordering, largest part of the value is placed on the last byte
and the smallest part of the value is placed on the first byte.</li>
</ul>
<p>
Nowadays, little-endian is the most commonly used.
Writing the integer <codeclass="evidence">0x4A3B2C1Dh</code>, will be ordered:
<ul>
<li>in big endian as <codeclass="evidence">0x4Ah</code>, <codeclass="evidence">0x3Bh</code>, <codeclass="evidence">0x2Ch</code>, <codeclass="evidence">0x1Dh</code></li>
<li>in little endian as <codeclass="evidence">0x1Dh</code>, <codeclass="evidence">0x2Ch</code>, <codeclass="evidence">0x3Bh</code>, <codeclass="evidence">0x4Ah</code></li>
</ul>
</p>
</div>
<p>
Blender supports little-endian and big-endian.<br>
This means that when the endianness
is different between the blend-file and the PC your using, Blender changes it to the byte ordering
This hex-dump describes a file-header created with <code>blender</code><code>2.54.0</code> on <code>little-endian</code> hardware with a <code>32 bits</code> pointer length.
This hex-dump describes a File-block (= <spanclass="header">File-block header</span> + <spanclass="data">File-block data</span>) created with <code>blender</code><code>2.54</code> on <code>little-endian</code> hardware with a <code>32 bits</code> pointer length.<br>
<codeclass="block"><spanclass="descr"> file-block
identifier='SC' data size=1404 old pointer SDNA index=150
0000 4430: <spanclass="header">[01 00 00 00]</span><spanclass="data">[xx xx xx xx xx xx xx xx xx xx xx xx</span> .... xxxx xxxx xxxx<spanclass="descr">
| |
count=1 file-block data (next 1404 bytes)</span>
</code>
</p>
<ul>
<li>The code <code>'SC'+0x00h</code> identifies that it is a Scene. </li>
<li>Size of the data is 1404 bytes (0x0000057Ch = 0x7Ch + 0x05h * 256 = 124 + 1280)</li>
<li>The old pointer is 0x0BFB3468h</li>
<li>The SDNA index is 150 (0x00000096h = 6 + 9 * 16 = 6 + 144)</li>
<li>The section contains a single scene (count = 1).</li>
</ul>
<p>
Before we can interpret the data of this file-block we first have to read the DNA structures in the file.
The section "<ahref="#structure-DNA">Structure DNA</a>" will show how to do that.
This hex-dump describes the file-block 'DNA1' header created with <code>blender</code><code>2.54.0</code> on <code>little-endian</code> hardware with a <code>32 bits</code> pointer length.<br>
structures in <ahref="#DNA1-data-array-types">types</a> in <ahref="#DNA1-data-array-types">types</a> in <ahref="#DNA1-data-array-names">names</a> in <ahref="#DNA1-data-array-types">types</a> in <ahref="#DNA1-data-array-names">names</a></span><spanclass="fade">
.... (398 structures, each one describeing own type, and type/name for each field)</span>
</code>
</p>
</div>
<p>
The DNA structures inside a Blender 2.48 blend-file can be found at <ahref="http://www.atmind.nl/blender/blender-sdna.html">http://www.atmind.nl/blender/blender-sdna.html</a>.
If we understand the DNA part of the file it is now possible to read
information from other parts file-blocks. The next section will tell us
how.
</p>
<aname="reading-scene-information"href="#reading-scene-information"><h2>Reading scene information</h2></a>
<p>
Let us look at <ahref="#example-file-block-header">the file-block header we have seen earlier</a>:<br>
</p>
<ul>
<li>the file-block identifier is <code>'SC'+0x00h</code></li>
<li>the SDNA index is 150</li>
<li>the file-block size is 1404 bytes</li>
</ul>
<p>
Now note that:
<ul>
<li>the structure at index 150 in the DNA is a structure of type 'Scene' (counting from 0).</li>
<li>the associated type ('Scene') in the DNA has the length of 1404 bytes.</li>
</ul>
</p>
<p>
We can map the Scene structure on the data of the file-blocks.
But before we can do that, we have to flatten the Scene-structure.
<codeclass="block">struct Scene {
ID id; <spanclass="descr">// 52 bytes long (ID is different a structure)</span>
AnimData *adt; <spanclass="descr">// 4 bytes long (pointer to an AnimData structure)</span>
Object *camera; <spanclass="descr">// 4 bytes long (pointer to an Object structure)</span>
World *world; <spanclass="descr">// 4 bytes long (pointer to an Object structure)</span>
...
float cursor[3]; <spanclass="descr">// 12 bytes long (array of 3 floats)</span>
...
};
</code>
The first field in the Scene-structure is of type 'ID' with the name 'id'.
Inside the list of DNA structures there is a structure defined for type 'ID' (structure index 17).
<codeclass="block">struct ID {
void *next, *prev;
struct ID *newid;
struct Library *lib;
char name[24];
short us;
short flag;
int icon_id;
IDProperty *properties;
};
</code>
The first field in this structure has type 'void' and name '*next'. <br>
Looking in the structure list there is no structure defined for type 'void': it is a simple type and therefore the data should be read.
The name '*next' describes a pointer.
As we see, the first 4 bytes of the data can be mapped to 'id.next'.
</p>
<p>
Using this method we'll map a structure to its data. If we want to
read a specific field we know at which offset in the data it is located
and how much space it takes.<br>
The next table shows the output of this flattening process for some
parts of the Scene-structure. Not all rows are described in the table as
there is a lot of information in a Scene-structure.
<ulclass="noprint"><li><ahref="http://www.blender.org/development/coding-guides/">Coding Guides</a></li><li>•</li><li><ahref="http://www.blender.org/development/report-a-bug/">Report a Bug</a></li><li>•</li><li><ahref="http://www.blender.org/development/submit-a-patch/">Submit a Patch</a></li><li>•</li><li><ahref="http://www.blender.org/development/release-logs/">Release Logs</a></li><li>•</li><li><ahref="http://www.blender.org/development/current-projects/">Current Projects</a></li><li>•</li><li><ahref="http://www.blender.org/development/architecture/">Architecture</a></li><li>•</li><li><ahref="http://www.blender.org/development/building-blender/">Building Blender</a></li><li>•</li><liclass="subnav-active"><ahref="http://www.blender.org/documentation/250PythonDoc/contents.html">PyAPI</a></li><li>•</li><li><ahref="http://wiki.blender.org/index.php/Main_Page">Wiki</a></li></ul>
<ulclass="noprint"><li><ahref="http://wiki.blender.org/index.php/Dev:Contents">Documentation</a></li><li>•</li><li><ahref="http://www.blender.org/development/report-a-bug/">Report a Bug</a></li><li>•</li><li><ahref="http://wiki.blender.org/index.php/Dev:Doc/Process/Patches">Submit a Patch</a></li><li>•</li><li><ahref="http://www.blender.org/development/release-logs/">Release Logs</a></li><li>•</li><li><ahref="http://www.blender.org/development/building-blender/">Building Blender</a></li><li>•</li><li><ahref="http://wiki.blender.org/index.php/Dev:Doc/Projects">Current Projects</a></li><li>•</li><li><ahref="http://wiki.blender.org/index.php/Dev:Source/Architecture">Architecture</a></li><li>•</li><li><ahref="http://www.blender.org/documentation/250PythonDoc/contents.html">Python API</a></li><li>•</li><li><ahref="http://wiki.blender.org">Wiki</a></li></ul>
(when accessing it from the Game Engine use Geometry instead of Blender.Geometry)
This new module provides access to a geometry function.
"""
defIntersect(vec1,vec2,vec3,ray,orig,clip=1):
"""
Return the intersection between a ray and a triangle, if possible, return None otherwise.
@type vec1: Vector object.
@param vec1: A 3d vector, one corner of the triangle.
@type vec2: Vector object.
@param vec2: A 3d vector, one corner of the triangle.
@type vec3: Vector object.
@param vec3: A 3d vector, one corner of the triangle.
@type ray: Vector object.
@param ray: A 3d vector, the orientation of the ray. the length of the ray is not used, only the direction.
@type orig: Vector object.
@param orig: A 3d vector, the origin of the ray.
@type clip: integer
@param clip: if 0, don't restrict the intersection to the area of the triangle, use the infinite plane defined by the triangle.
@rtype: Vector object
@return: The intersection between a ray and a triangle, if possible, None otherwise.
"""
defTriangleArea(vec1,vec2,vec3):
"""
Return the area size of the 2D or 3D triangle defined.
@type vec1: Vector object.
@param vec1: A 2d or 3d vector, one corner of the triangle.
@type vec2: Vector object.
@param vec2: A 2d or 3d vector, one corner of the triangle.
@type vec3: Vector object.
@param vec3: A 2d or 3d vector, one corner of the triangle.
@rtype: float
@return: The area size of the 2D or 3D triangle defined.
"""
defTriangleNormal(vec1,vec2,vec3):
"""
Return the normal of the 3D triangle defined.
@type vec1: Vector object.
@param vec1: A 3d vector, one corner of the triangle.
@type vec2: Vector object.
@param vec2: A 3d vector, one corner of the triangle.
@type vec3: Vector object.
@param vec3: A 3d vector, one corner of the triangle.
@rtype: float
@return: The normal of the 3D triangle defined.
"""
defQuadNormal(vec1,vec2,vec3,vec4):
"""
Return the normal of the 3D quad defined.
@type vec1: Vector object.
@param vec1: A 3d vector, the first vertex of the quad.
@type vec2: Vector object.
@param vec2: A 3d vector, the second vertex of the quad.
@type vec3: Vector object.
@param vec3: A 3d vector, the third vertex of the quad.
@type vec4: Vector object.
@param vec4: A 3d vector, the fourth vertex of the quad.
@rtype: float
@return: The normal of the 3D quad defined.
"""
defLineIntersect(vec1,vec2,vec3,vec4):
"""
Return a tuple with the points on each line respectively closest to the other
(when both lines intersect, both vector hold the same value).
The lines are evaluated as infinite lines in space, the values returned may not be between the 2 points given for each line.
@type vec1: Vector object.
@param vec1: A 3d vector, one point on the first line.
@type vec2: Vector object.
@param vec2: A 3d vector, another point on the first line.
@type vec3: Vector object.
@param vec3: A 3d vector, one point on the second line.
@type vec4: Vector object.
@param vec4: A 3d vector, another point on the second line.
@rtype: (Vector object, Vector object)
@return: A tuple with the points on each line respectively closest to the other.
"""
defPolyFill(polylines):
"""
Takes a list of polylines and calculates triangles that would fill in the polylines.
Multiple lines can be used to make holes inside a polyline, or fill in 2 separate lines at once.
@type polylines: List of lists containing vectors, each representing a closed polyline.
@rtype: list
@return: a list if tuples each a tuple of 3 ints representing a triangle indexing the points given.
@note: 2D Vectors will have an assumed Z axis of zero, 4D Vectors W axis is ignored.
@note: The order of points in a polyline effect the direction returned triangles face, reverse the order of a polyline to flip the normal of returned faces.
I{B{Example:}}
The example below creates 2 polylines and fills them in with faces, then makes a mesh in the current scene::
me.faces.extend(fill) # Add the faces, they reference the verts in polyline 1 and 2
scn = Blender.Scene.GetCurrent()
ob = scn.objects.new(me)
Blender.Redraw()
"""
defLineIntersect2D(vec1,vec2,vec3,vec4):
"""
Takes 2 lines vec1, vec2 for the 2 points of the first line and vec2, vec3 for the 2 points of the second line.
@rtype: Vector
@return: a 2D Vector for the intersection or None where there is no intersection.
"""
defClosestPointOnLine(pt,vec1,vec2):
"""
Takes 2 lines vec1, vec2 for the 2 points of the first line and vec2, vec3 for the 2 points of the second line.
@rtype: tuple
@return: a tuple containing a vector and a float, the vector is the closest point on the line, the float is the position on the line, between 0 and 1 the point is on the line.
"""
defPointInTriangle2D(pt,tri_pt1,tri_pt2,tri_pt3):
"""
Takes 4 vectors (one for the test point and 3 for the triangle)
This is a 2d function so only X and Y are used, Z and W will be ignored.
@rtype: int
@return: 1 for a clockwise intersection, -1 for counter clockwise intersection, 0 when there is no intersection.
"""
defPointInQuad2D(pt,quad_pt1,quad_pt2,quad_pt3):
"""
Takes 5 vectors (one for the test point and 5 for the quad)
This is a 2d function so only X and Y are used, Z and W will be ignored.
@rtype: int
@return: 1 for a clockwise intersection, -1 for counter clockwise intersection, 0 when there is no intersection.
"""
defBoxPack2D(boxlist):
"""
Takes a list of 2D boxes and packs them into a square.
Each box in boxlist must be a list of at least 4 items - [x,y,w,h], after running this script,
the X and Y values in each box will be moved to packed, non overlapping locations.
Example::
# Make 500 random boxes, pack them and make a mesh from it
This script shows how to use the classes: :class:`Device`, :class:`Factory` and
:class:`Handle`.
"""
importaud
device=aud.device()
# load sound file (it can be a video file with audio)
factory=aud.Factory('music.ogg')
# play the audio, this return a handle to control play/pause
handle=device.play(sound)
# if the audio is not too big and will be used often you can buffer it
factory_buffered=aud.Factory.buffer(sound)
handle_buffered=device.play(buffered)
# stop the sounds (otherwise they play until their ends)
handle.stop()
handle_buffered.stop()
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
