Fixed and completed support for returning multiple values. This includes support for returning arrays, both fixed and dynamically sized. The way this is achieved is by storing an additional int value next to the dynamic parameter in the ParameterList stack which gets passed to the C function as an additional parameter. In the case of return parameters it is duty of the C function to set this int to the correct length value for the dynamic parameter (which makes sense). Note that for the dynamic output/return parameter it is assumed the function has allocated that memory (which gets freed automatically).
Also, I cleaned the makesrna's bridge function generation code a bit and renamed PROP_RETURN to PROP_OUTPUT, which represents better the reality now that there are multiple returns. The function now to mark multiple returns (outputs) is RNA_def_function_output.
For an example, look at Action.get_frame_range in rna_action_api.c, by the way Aligorith I removed the #ifdef for this function now that there's support for returning arrays, feel free to modify (the function seems to work).
- scene.render_data.frame_path(frame=num), returns the output path for rending images of video.
- scene.render_data.file_extension, readonly attribute, gives the extension ".jpg", ".mov" etc
- player support was guessing names, use the above functions to get the actual names used, accounting for #'s replacing numbers.
from the report...
# bug 1. UV properties are not raw editable but are reported
# as RAW_TYPE_INT by RNA causing wrong conversion
# internally (bpy_rna.c line 2205)
# bug 2. raw update of UV coordinates crash blender (rna_access.c line 252)
mtfaces.foreach_set("uv", rawuvs)
# workaround:
#for i in range(int(len(faces)/4)):
# mtfaces[i].uv = uvs[i]
# bug 3. raw update of non-array property fails (rna_access.c line 2270)
mfaces.foreach_set("material_index", mats)
# workaround:
# for i in range(int(len(mfaces))):
# mfaces[i].material_index = mats[i]
# bug 4. It is not possible to add a vertex color layer using mesh API.
me.add_vertex_color()
# no workaround...
patch from Elia Sarti, (vekoon) with some modifications mainly for the python api.
- multiple values are returned as a typle in the order that are defined.
- added support for registered types returning multiple arguments (untested).
- renamed func->ret --> func->c_ret, since this only defines what the C function returns.
The Animation System now respects the Transform Locks too (i.e. lock x-location, etc.) when writing settings. This means that it is no longer necessary to set up "constant drivers" to make sure some values don't get accidentally animated.
Internally, added a new callback for properties in RNA, which is responsible for checking if the item at some array-index is editable. This needs to be manually called for each place which uses rna to set settings for arrays (see the code changes in anim_sys.c for changes how to do this; the same thing needs to be done in the UI code too, and probably in py-api too)
eg.
scene.objects.link()
object.constraints.new()
mesh.verts.transform(...)
mesh.faces.active
PropertyRNA stores an StructRNA pointer where these can be defined.
* Enums with an _itemf callback now never get context NULL passed in,
rather a fixed list of enum items are defined which should contain
all items (if possible), from which the _itemf callback can then use
a subset.
Keymaps are now saveable and configurable from the user preferences, note
that editing one item in a keymap means the whole keymap is now defined by
the user and will not be updated by Blender, an option for syncing might be
added later. The outliner interface is still there, but I will probably
remove it.
There's actually 3 levels now:
* Default builtin key configuration.
* Key configuration loaded from .py file, for configs like Blender 2.4x
or other 3D applications.
* Keymaps edited by the user and saved in .B.blend. These can be saved
to .py files as well to make creating distributable configurations
easier.
Also, user preferences sections were reorganized a bit, now there is:
Interface, Editing, Input, Files and System.
Implementation notes:
* wmKeyConfig was added which represents a key configuration containing
keymaps.
* wmKeymapItem was renamed to wmKeyMapItem for consistency with wmKeyMap.
* Modal maps are not wrapped yet.
* User preferences DNA file reading did not support newdataadr() yet,
added this now for reading keymaps.
* Key configuration related settings are now RNA wrapped.
* is_property_set and is_property_hidden python methods were added.
% sign for percentage assuming it is between 0-100, while factor is
for values 0-1.
Move collision setting absorption from modifier to collision settings,
was inconsistent to have it there as the only one, and made it have
range 0.0-1.0 instead of 0-100.
* PROP_NEVER_NULL is now a flag instead of a subtype.
* It works for function parameters too now, so setting
this flag can help avoid NULL checks in the function.
* Renamed LocalLamp to PointLamp, making it consistent
with the UI name.
* Set icons for the different lamp struct types.
Appending and Linking
* Linking Operator, invokes filebrowser for Append/Link
* Separated the append/link function into three parts:
** BLO_library_append_begin finds main for appending
** BLO_library_append_named_part appends one Object,Group, Material, ...
** BLO_library_append_end actually reads and expands the libraries
NOTE 1:
I also changed the returned properties for the filebrowser operators to the following convention:
"path" - the full path to a file or directory, means what is in directory + filename buttons in filebrowser
"directory" - the content of the directory button in filebrowser
"filename" - the content of the filename button in filebrowser
Usually only path should be required, but in some cases it might be more convenient to retrieve the parts separately.
Ton, Brecht: If you have time to take a look, let me know if anything needs to be fixed.
* Disable setting array length of dynamic array for now, this was not
implemented correct, and it's not really needed now.
* Allow all dimensions to be dynamic size, not just the first.
* Change storage of multidimensional to be simpler.
* Rename API functions to be more compact.
* Fix some bugs in the implementation.
* RenderLayer.rect and RenderPass.rect use a multidimensional
dynamic array now.
* Added RNA subtype for layers.
* Shift-click works again.
* uiItemR can now also handle armature/bone layers.
* Also makes Move to Layer popup work as expected.
Multidim. arrays can now be modified at any level, for example:
struc.arrayprop = x
struc.arrayprop[i] = x
struc.arrayprop[i][j] = x
struc.arrayprop[i][j][k] = x
etc...
Approriate rvalue type/length checking is done.
To ensure all works correctly, I wrote automated tests in release/test/rna_array.py.
These tests cover: array/item access, assignment on different levels, tests that proper exceptions are thrown on invalid item access/assignment.
The tests use properties of the RNA Test struct defined in rna_test.c. This struct is only compiled when building with BF_UNIT_TEST=1 scons arg.
Currently unit tests are run manually by loading the script in the Text Editor.
Here's the output I have: http://www.pasteall.org/7644
Things to improve here:
- better exception messages when multidim. array assignment fails. Those we have currently are not very useful for multidim.
- add tests for slice assignment
Notifiers
---------
Various fixes for wrong use of notifiers, and some new notifiers
to make things a bit more clear and consistent, with two notable
changes:
* Geometry changes are now done with NC_GEOM, rather than
NC_OBJECT|ND_GEOM_, so an object does need to be available.
* Space data now use NC_SPACE|ND_SPACE_*, instead of data
notifiers or even NC_WINDOW in some cases. Note that NC_SPACE
should only be used for notifying about changes in space data,
we don't want to go back to allqueue(REDRAW..).
Depsgraph
---------
The dependency graph now has a different flush call:
DAG_object_flush_update(scene, ob, flag)
is replaced by:
DAG_id_flush_update(id, flag)
It still works basically the same, one difference is that it now
also accepts object data (e.g. Mesh), again to avoid requiring an
Object to be available. Other ID types will simply do nothing at
the moment.
Docs
----
I made some guidelines for how/when to do which kinds of updates
and notifiers. I can't specify totally exact how to make these
decisions, but these are basically the guidelines I use. So, new
and updated docs are here:
http://wiki.blender.org/index.php/BlenderDev/Blender2.5/NotifiersUpdateshttp://wiki.blender.org/index.php/BlenderDev/Blender2.5/DataNotifiers
Example code: http://www.pasteall.org/7332/c.
New API functions: http://www.pasteall.org/7330/c.
Maximum number of dimensions is currently limited to 3, but can be increased arbitrarily if needed.
What this means for ID property access:
* MeshFace.verts - dynamic array, size 3 or 4 depending on MFace.v4
* MeshTextureFace.uv - dynamic, 2-dimensional array, size depends on MFace.v4
* Object.matrix - 2-dimensional array
What this means for functions:
* more intuitive API possibility, for example:
Mesh.add_vertices([(x, y, z), (x, y, z), ...])
Mesh.add_faces([(1, 2, 3), (4, 5, 6), ...])
Python part is not complete yet, e.g. it is possible to:
MeshFace.verts = (1, 2, 3) # even if Mesh.verts is (1, 2, 3, 4) and vice-versa
MeshTextureFace.uv = [(0.0, 0.0)] * 4 # only if a corresponding MFace is a quad
but the following won't work:
MeshTextureFace.uv[3] = (0.0, 0.0) # setting uv[3] modifies MTFace.uv[1][0] instead of MTFace.uv[3]
* Popup menus now remember the last clicked item again.
* Modifier and File Format menus are now organized in multiple
columns with categories.
* Hook, explode, uv project modifiers have all their buttons
again with the relevant operators implemented.
* Modifiers that can't be added by the user, or don't work on
curves for example, are not in the menu anymore.
* Fix search menu overlapping buttons when near the bottom of
the screen.
* Fix uv layers search menu not working in some modifiers.
* Cleanup popup menu code a bit, layout engine is used in more
cases now instead of ugly position calculation code.
possible from python, but it's still work in progress.
Pointers and collections are restricted to types derived from
IDPropertyGroup (same as for operators), because RNA knows how to
allocate/deallocate those.
Collections have .add() and .remove(number) functions that can be
used. The remove function should be fixed to take an other argument
than a number.
With the IDPropertyGroup restriction, pointers are more like nested
structs. They don't have add(), remove() yet, not sure where to put
them. Currently the pointer / nested struct is automatically allocated
in the get() function, this needs to be fixed, rule is that RNA get()
will not change any data for thread safety.
Also, it is only possible to add properties to structs after they have
been registered, which needs to be improved as well.
Example code:
http://www.pasteall.org/7201/python
- use RNA_struct_blender_type_set to set the type to NULL before its freed
There is a memory error here when reloading scripts - ui_handler_panel_region, need to look into theis further.
* Reviewed subtypes, making them more specific and adding new ones.
* Subtypes now have an associated type of units (length, area, volume,
mass, rotation, time, velocity, acceleration). These are not used
yet anywhere.
* Centralized code that decides the name of array items based on
subtype (XYZ, RGB), was copied in 3 places.
* RNA_def_float etc functions still need to be update, will do this
later together with another change.
Various fixes for painting, sculpting and particle edit, still
much to be done...
* Move RNA paint and sculpt structs into rna_sculpt_paint.c,
* Added Particle Edit RNA.
* Some tweaks to existing Paint RNA.
* Put texture paint and particle edit object in context.
* Fix some errors in the brush layout, properly doing None
checks, fixing some wrong property identifiers.
* Added tool enum for texture paint and particle edit in panels.
* Allow editing brush textures in the texture buttons, still with
a stupid toggle, ideas for how to make the connection better are
welcome.
* Changed update callbacks for several armature properties to do proper depsgraph flushing instead of sending redraw notifiers
* Wrapped preview-range settings for Scene
* Fixed bug with 'parent-type' enum menu for Object parent-type. The problem here was that PARSKEL was used and is still used for both Armatures and Lattices, but the lattice entry would always override armatures entry. Now, when dynamically building the list to display, the function to add these is given the specific item to use, and has been made to stop after the first match has been added. Hopefully this doesn't break anything else...
* ID blocks can now get RNA properties defined from python, e.g.:
bpy.types.Scene.BoolProperty(..)
* RNA structs/functions/properties can now get pointers duplicated
(mostly strings), since we can't point to some static string then.
* Added ExtensionRNA struct to add into *Type structs for subclassing,
is a bit more compact than defining the 4 variables each time.
Only disadvantage is it requires including RNA in more places.
* Added suppport for generating code without verifying with DNA,
this doesn't give good errors and is more error prone, but makes
it easier to wrap things like EditBone which are not in DNA.
* RNA_define_verify_sdna(0), and set to 1 again afterwards.
* RNA_enum_items_add_value and RNA_enum_item_add_separator utility
functions, to add an item from an existing array with a certain
value, and to add a separator.
* Enums can now be dynamically created in the _itemf callback,
using RNA_enum_item(s)_add, RNA_enum_item_end. All places asking
for enum items now need to potentially free the items.
* This callback now also gets context, this was added specifically
for operators. This doesn't fit design well at all, needed to do
some ugly hacks, but can't find a good solution at the moment.
* All enums must have a default list of items too, even with an
_itemf callback, for docs and fallback in case there is no context.
* Used by MESH_OT_merge, MESH_OT_select_similar, TFM_OT_select_orientation.
* Also changes some operator properties that were enums to booleas
(unselected, deselect), to make them consistent with other ops.
* Context panel now draws without header, with arrows, no scene name.
* Softbody vertex group search popup.
* Improve names for autogenerated shortcut keys in menus.
* Make most Select menus in the 3D view header work.
* Fix armature border select selection syncing.
* Add POSE_OT_select_constraint_target,
MESH_OT_select_by_number_vertices, MESH_OT_select_vertex_path.
* Merge mesh select similar into one operator.
* Don't give MESH_OT_select_random Space hotkey.
* Add DAG_object_flush_update to many mesh edit tools, not calling this
will crash with modifiers.
* RNA_def_enum_funcs for dynamic enums in operators, but not very useful
without context yet.
* Fix refresh issue with image window header + editmode.
* Fix drawing of shadow mesh for image painting.
* Remove deprecated uiDefMenuButO and uiDefMenuSep functions.
* Remove keyval.c, code is in wm_keymap.c already.
* Rename WM_operator_redo to WM_operator_props_popup.
Implementation of RNA side of foreach_get/foreach_set,
Campbell will do python code.
Three functions for efficiently setting some property for all
items in a collection. RNA_property_collection_raw_array gives
access to the properties as an array with length, stride, and
type specified, if this is possible, i.e. not when it uses a
ListBase, or if a manual get/set function is implemented.
Two other functions take a C array pointer and get/set it
using the a collection + property name, using efficient array
access if possible, and otherwise using slower RNA iterator.
RNA_property_collection_raw_get
RNA_property_collection_raw_set
The number of type conversion required here got a bit out of
hand, it could be more efficient still if checking for more
cases, but function is already long enough.
Example: http://www.pasteall.org/6362/c
check Save Image or Export PLY operator for example.
Also these code changes:
* Added some RNA collection iterator macros to simplify code.
* Fix bpy.props.BoolProperty not working correct.
* Merge uiDefAutoButsRNA/uiDefAutoButsRNA_single into one.
* Optimized RNA property lookups and path resolving, still can be
much better, but now the 1000 IPO example on bf-taskforce25
runs at reasonable speed.
* Also an optimization in the depsgraph when dealing with many
objects, this was actually also a bottleneck here.
Merging changes made by Arystanbek in the soc-2009-kazanbas branch,
plus some things modified and added by me.
* The API files now all in the makesrna module, convention is to
call them e.g. rna_mesh_api.c for rna_mesh.c. Note for visual
studio build maintainers, the rna_*_api.c files are compiled as
part of "makesrna", but do not have rna_*_gen.c generated as part
of the library. SCons/cmake/make were updated.
* Added function flags FUNC_USE_CONTEXT and FUNC_USE_REPORTS, to
allow RNA functions to get context and error reporting parameters
optionally. Renamed FUNC_TYPESTATIC to FUNC_NO_SELF.
* RNA collections now have a pointer to add/remove FunctionRNA's, this
isn't actually used anywhere yet, purpose is to make an alias
main.meshes.add() for main.add_mesh() in python.
* Fixes to make autogenerating property set/get for multidimensional
arrays work, though a 4x4 matrix will be exposed as a length 16
one dimensional RNA array.
* Functions and properties added:
* Main.add_mesh()
* Main.remove_mesh()
* Object.matrix
* Object.create_render_mesh()
* WindowManager.add_fileselect()
* Added icon to property and enum property items. The latter is
responsible for the large number of files changed.
* For RNA functions, added PROP_RNAPTR flag to ask for a PointerRNA
as argument instead of a C pointer, instead of doing it implicitly
with the AnyType type.
* Material: properly wrap diffuse/specular param variables, and
rename some things for consistency.
* MaterialTextureSlot: added "enabled" property (ma->septex).
* Image: make animated property editable.
* Image Editor: make some things editable, notifiers, respect state.
* Context: fix issue with screen not being set as ID.
* Accept None as NULL pointers through python function calls.
* Added type callback for pointers back, it's useful still in
some cases. Made Object.data editable using this, the pointer
type varying based on object type.
* Wrap pin ID pointer in buttons space.
* Added subclasses for text and surface curve ID blocks, to
organize data better and get proper icons.
* Added RNA_type_to_ID_code and ID_code_to_RNA_type functions.
* Update RNA_access.h with new RNA types.
* Added PROP_ID_REFCOUNT flag to control if refcounting should
be done on ID pointer properties. All ID pointers are refcounted,
by default, with the exception of Object, Scene and Text.
* Also made TextureFace image pointer editable, with the special
refcounting behavior.
* Automatically do us++ and us-- reference counting in ID pointer
set functions.
* Added an enum property callback to dynamically vary the list of
available items.
* Added some functions to do removes on pointers and collections
runtime defined for RNA and using ID properties.
* Constraints now have owner/target space wrapped, and most
pointers made editable. They can be ported to use python layouts.
* Also other pointers made editable that I think are see now with
the automatic reference counting.
* Any Struct can now have ID properties, by creating a callback
function to create/return an IDProperty.
* Wrapped PoseChannel ID properties.
* Note there is still no way to create ID Properties in 2.5, though
the callback to get/create the initial group is now exposed through
RNA_struct_idproperties.
The Python API to define Panels and Operators is based on subclassing,
this makes that system more generic, and based on RNA. Hopefully that
will make it easy to make various parts of Blender more extensible.
* The system simply uses RNA properties and functions and marks them
with REGISTER to make them part of the type registration process.
Additionally, the struct must provide a register/unregister callback
to create/free the PanelType or similar.
* From the python side there were some small changes, mainly that
registration now goes trough bpy.types.register instead of
bpy.ui.addPanel.
* Only Panels have been wrapped this way now. Check rna_ui.c to see
how this code works. There's still some rough edges and possibilities
to make it cleaner, though it works without any manual python code.
* Started some docs here:
http://wiki.blender.org/index.php/BlenderDev/Blender2.5/RNATypeRegistration
* Also changed some RNA_property and RNA_struct functions to not
require a PointerRNA anymore, where they were not required (which
is actually the cause of most changed files).
* Added REQUIRED flag for function parameters.
* Made dupliframes/verts/faces/groups an enum, and make it editable.
* Enum bitflags were broken, fixed.
