- comments to PyObjectPlus.h
- remove unused/commented junk.
- renamed PyDestructor to py_base_dealloc for consistency
- all the PyTypeObject's were still using the sizeof() their class, can use sizeof(PyObjectPlus_Proxy) now which is smaller too.
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).
This changes how the BGE classes and Python work together, which hasnt changed since blender went opensource.
The main difference is PyObjectPlus - the base class for most game engine classes, no longer inherit from PyObject, and cannot be cast to a PyObject.
This has the advantage that the BGE does not have to keep 2 reference counts valid for C++ and Python.
Previously C++ classes would never be freed while python held a reference, however this reference could be problematic eg: a GameObject that isnt in a scene anymore should not be used by python, doing so could even crash blender in some cases.
Instead PyObjectPlus has a member "PyObject *m_proxy" which is lazily initialized when python needs it. m_proxy reference counts are managed by python, though it should never be freed while the C++ class exists since it holds a reference to avoid making and freeing it all the time.
When the C++ class is free'd it sets the m_proxy reference to NULL, If python accesses this variable it will raise a RuntimeError, (check the isValid attribute to see if its valid without raising an error).
- This replaces the m_zombie bool and IsZombie() tests added recently.
In python return values that used to be..
return value->AddRef();
Are now
return value->GetProxy();
or...
return value->NewProxy(true); // true means python owns this C++ value which will be deleted when the PyObject is freed
Hiding faces is a editing option like selection and should not change rendering, it wasn't even working right because meshes without UVs ignored it.
I thought this was needed for compatibility with old files but just noticed this messes up 2 of the files in demos-2.42.zip
Other small changes...
- KX_Camera and KX_Light didnt have get/setitem access in their PyType definition.
- CList.from_id() error checking for a long was checking for -1 against an unsigned value (own fault)
- CValue::SpecialRelease was incrementing an int for no reason.
- renamed m_attrlist to m_attr_dict since its a PyDict type.
- removed custom getattro/setattro functions for KX_Scene and KX_GameObject, use py_base_getattro, py_base_setattro for all subclasses of PyObjectPlus.
- lowercase windows.h in VideoBase.cpp for cross compiling.
* Forgot to finish this code yesterday, dragging panels was
broken. This is fixed and new panels are now inserted after
the last added one, inbetween others rather than at the end.
Texture nodes hang when nodes have a cyclic case.
Added a (temp?) provision to tag node->need_exec zero for cyclic
nodes, and added check for this in texture nodes.
There was also a bug in 'tag changed' for texture nodes, which not
only tagged, but also called the tree exec (should not happen!).
In general the texture exec needs recode; it doesn't use the stacks
as provided per node, but recurses itself to previous nodes, giving
problems like this. Node execs should only do their own bizz, the
node system handles dependency and eval order nicely already.
* Added back backgrounds behind buttons for modifiers and driver targets
* 'Protect' toggle doesn't draw for F-Curves with no keyframes to prevent editing on anymore
* Inserting keyframes on objects using builtin keyingsets now use the right names
* For new buttons spaces, automatically set horizontal/vertical
align depending on size, instead of free.
* Cleaned up the UI panel API. There's now a new uiBeginPanel
function which takes a panel type, and a uiEndPanel which takes
the final size. uiNewPanel* functions will be phased out.
* Animate the re-alignment when a panel size changes, e.g. when
enabling dupliframes.
* Load ui scripts from the release/ folder first if it is
available. This makes it easier to edit ui scripts, since it
will directly use the original files which avoids having to
run the build system.
* Improve editing of panel types while blender is open. That
means fixing some issues with lacking updates, overlaps, strange
ordering. It even does an animation now when the panel resizes.
* Added extra parameter to generic unique name finding function BLI_uniquename() for specifying the delimeter between non-unique parts of the name and digits.
* Driver target variables now get unique names by default.
* The UI for adding and modifying drivers is now functional, and reflects the new way they work. It has also been tidied up a bit (from the previous incarnation), with RNA buttons starting to be used (currently limited to the 'Path' settings).
* Fixed errors in RNA wrapping for drivers
* Don't call generic layout hints templates anymore, i.e.
TemplateRow becomes Row, etc.
* Added more general layout nesting, using uiLayoutSplit()
and uiLayoutBox() functions, for which the sublayouts
can then be accessed using uiLayoutSub(), to put items
in those sublayouts.
* Some steps to make the layout decisions, like which items
to put in which columns, independent of the width of the
window or the text in the buttons. We want the layout to
be stable under resizes and translations.
* Added an "expand" parameter to uiItemR, used now to expand
enums into a row instead of using a menu.
Drivers now support multiple targets which act as 'variables'. The targets have a short 'name' (see later), and reference some property (in much the same way as F-Curves do, using RNA-Paths) which acts as the 'value'.
These named variables can then be used in a Python Expression which relates them to each other for more fine-grained control over the result of the driver. By using only the names of these variables in the expressions, we are able to define expressions/relationships in a much more readable way, as data access is separated from data use. This makes the underlying relationships easier to understand.
By default, if no Python Expression is given, the variables are simply averaged together, so old files won't break. :)
For example, check the following diagram (thanks Cessen/Nathan V from Peach team):
http://download.blender.org/ftp/incoming/250_drivers_mockup_cessen.png
TODO List:
* Depsgraph building for new driver relationships doesn't work yet. This needs to be recoded again, but this new system makes this much easier, since the targets are clearly defined (i.e. no need to parse py expressions to get list of objects)
* Graph Editor interface for editing these needs to be rewritten
* Python function for evaluating these expressions is needed (Campbell?)
* Replaced a quicky hack needed to get PoseLib working with a proper new group-naming option for KeyingSets. Now, all builtin KeyingSets will use the name of the data (i.e. Object or PoseChannel) as the name of the group new channels are added to
* Fixed a bug with LocRotScale builtin KeyingSet, which meant that scale keyframes were not getting added.
* TAB key (toggle original pose) now works again. Previously, events were flying past too quickly.
+bvhtree cache (if the derived model doenst gets destroyed then the same BVHtree can be used)
this was needed to allow shrinkwrap constraint to be usable.
It has been ready for a long time.. but only got merged now, for 2.49.
Free python modules defined within the blendfile between loading scenes since they would end up accessing old GameLogic, Rasterizer modules as well as old game engine data in the module namespace which can cause problems.
- Fun for testers: Added "Redo Operator" Panel in view3d 'nkey' region.
It's going to be part of tools UI I know, but this will give good
tests of what goes on with operators. I had to add small fixes in
Transform for it already. :)
One important issue to note is that it lists every operator, also from
non-3D-window ops. Needs checked or classified somehow.
- Fix: removed bad 2.4x hack from how pulldown menus were defined. Made
widget code cleaner, and will show menus like SHIFT+A correct now.
Imagepaint make sure ibuf exists before we use it. This is CID# 545
editmesh_tools.c make sure were not indexing array with -1 This is CID# 137
(also removed some trailing whitespace I found)
anim5.c totlen could be -1 so check its > -1 This is CID# 134
Kent
