Started cleaning up FModifiers in preparation for allowing them to be used on NLA Strips.
This commit separates the 'Built-in Function' mode for the Generator modifier out into its own modifier, since it was being quite frequently used (and the RNA wrapping for this used to be quite hackish).
BACKWARDS COMPATABILITY WARNING:
Old files with FModifiers saved (i.e. old 2.5 files, but not any others) will not load correctly as a result of these changes (the wrong modifiers will be shown). I've decided that there are not likely to be many files affected by this yet, but doing this will result in a much nicer modifiers-define list in the long run.
While trying to fix the mapping conversions for repeat, I came across some limitations with the current (soon to be previous) mapping methods.
Now the mapping conversions should work nicely for all places that use them.
* Adding/removing particle systems to an object.
* Changing of particle settings.
* Currently showing an object's particle systems in a list (like materials).
* Based on what happens during simulation the cache is marked (also in cache panel, this could possibly be extended to 3d view as well) as:
- exact (not marked)
- outdated (simulation is not done completely with current settings)
- non-exact (frames were skipped during simulation)
* The parameter "cache step" effects the number of frames between saved cache frames.
- This can save a lot of memory (or disk space) if absolutely frame accurate simulation is not required.
- Speeds up the "quick caching" very much.
- Frames between cached frames are interpolated from the cached frames.
- Current default value of 10 frames works nicely with up/down-arrows (skip 10 frames forwards/backwards on timeline), but can be changed if wanted.
* The caching can work in normal or "quick" mode:
[Normal cache]
- Basic: Calculate what even happens (settings change, big frame steps etc.) and cache results, if possible try to use "cache step" when saving cache frames.
- Becomes non-exact: After larger than 1 frame steps.
- Becomes outdated: After any change effecting the simulation other than frame steps.
- Pros/cons: Freedom of doing anything and playing with particles, but exact results have to calculated from the beginning.
[Quick cache]
- Basic: Calculate simulation up to current frame automatically on changes with cache step sized jumps in simulation. With multiple "quick cached" simulations the smallest cache step is used.
- Becomes non-exact: Always from frame 1 (unless cache step = 1).
- Becomes outdated: Never.
- Pros/cons: Not very accurate, but super fast!
- Todo: Transform of any animated (non-autokeyed) object is locked! Probably needs some tinkering with anim sys overrides.
* The simulation can be run forwards or backwards even if it's cache is outdated or non-exact, the following rules apply in these situations:
- step forwards (to unknown) -> simulate from last exact frame, store result
- step backwards (to known) -> result is interpolated from existing frames, store result, clear cache forwards if current frame is after last exact frame
* "Calculate to current frame" runs the simulation from start to current frame with a frame steps of 1.
- Baking does the same, but runs the simulation all the way to the end of simulation.
- Rendering does this automatically if the simulation is outdated of non-exact, so all rendered simulations will always be updated and exact.
* Every cache panel also holds buttons to "Bake all dynamics", "Free all dynamics" and "Update all dynamics to current frame".
* Cloth simulation supports the new cache too.
* Strip evaluation now takes into account repeats
* Increasing the number of repeats lengthens the strip, while decreasing the number of repeats does the opposite.
TODO:
- inverse correction doesn't take into account repeats != 1, so tweaking strips with repeats is currently not recommended!
* Added support for passing collections to/from RNA functions,
this is done using a ListBase of CollectionPointerLink, which
is a standard ListBase link + PointerRNA.
* Added editable active uv/vcol layer to Mesh.
* Armature.bones now includes all bones, not only the ones without
parents.
* Modifier UV layer fields now are allowed to be empty, previously
this would set the name during modifier evaluation if there was
none.
* RNA_struct_name_get_alloc function to get the name from a
pointer, instead of having to deal with name property.
* CTX_data_pointer_get_type to get data from context with
a check for the RNA type.
* Transition tracks are now drawn in a different colour (bluish)
* Action line now shows keyframe indicators for the action it is representing. These are drawn as small empty diamonds to show that they're not directly editable (and never ever will be) via the NLA Editor.
* Action line is also drawn with borders to help differentiate it from the other tracks.
* As an experiment, removed the 'TweakMode' buttons from the header. I had originally added this as a way of indicating that this functionality was there (and to allow for keeping track of it), though it seems to have caused far too much confusion and frustration about 'modality' instead.
Added AnimData to ParticleSettings so that this will work. This seems to be ok in the brief tests I did, but be warned that this may not be optimal for the Particles System...
In TweakMode, the keyframes of the Active Action are now shown (and can be edited) in NLA-mapped time, with appropriate corrections applied when editing. This works in the DopeSheet and Graph Editors :)
To do this, got rid of the old wrappers/API-methods, replacing them with new-style ones. A few methods previously (in this branch) used only for evaluation are now used for this purpose too. As the same code is used for editing + evaluation, this should now be much better to work with.
I've only done a few brief tests now, but I think I might've muddled the invert-flags on one or two cases which I'll need to check out tomorrow. So, beware that there may be some weird and critical bugs for the next few days here...
Also, added proper license headers to new NLA files.
TODO:
- testing + bugfixing due to this commit
- show range of keyframes in NLA Editor active-action line
* Moved proportional edit, snap, autokey mode, and a few others
from Scene to ToolSettings.
* RNA wrapped properties in ToolSettings for the UV editor:
proportional edit, snap settings, selection modes.
Mostly harmless ones, except for one about "gzopen64" being
undeclared. This needs some defines in BLI_storage.h to be set
before <unistd.h> is included. Might fix a crash in compressed
file reading, though I'm not sure since it's hard to repeat
the crash consistently.
* removed radiosity render code, DNA and RNA (left in radio render pass options), we'll get GI to replace this probably, better allow baking to vertex colors for people who used this.
* removed deprecated solid physics library, sumo integrations and qhull, a dependency
* removed ODE, was no longer being build or supported
* remove BEOS and AMIGA defines and references in Makefiles.
* Particles support larger than 1 frame changes, bigger frame changes can result in inaccurate results, but it's super fast and you get a nice feeling of how the particles behave!
* "Cache to current frame" button calculates the exact result of particles at current frame.
* Current state of cache can be protected by making it a bake.
* Cache is now in memory by default, disk cache is an option.
* Only "viewport %" number of particles are calculated and cached in viewport, baking and rendering calculate all particles.
* Info on cached frames and memory usage given in ui.
* Support for exact "autocaching" of changes and large frame changes(disabled for now until exact place in event system is decided)
* "Continue physics" is probably deprecated after this and should be removed once sb & cloth use the new cache code.
Todo:
* Make softbody & cloth use the new cache things.
Other changes:
* Some cleanup of particle buttons.
These fixes get the 'pathJumper.blend' file from our testing suite workable in 2.5 (with a few minor tweaks still needed *)
Changes required:
- Added a 'ctime' var to curve structs for storing the value that used to be obtained by specially evaluating the 'speed' curve when evaluating objects parented to the curve. This can now be animated as a 'proper' var as per normal.
- Added a special hack for detecting constraint blocks, as the old method resulted in paths for Objects instead...
(*) Issues:
- Unfortunately, the paths still don't work out of the box. For some reason, the constraint names in the paths are spelt incorrectly - "Ar" and "Br" instead of "Ap" and "Bp". I'm not sure where this problem is coming from, but changing the paths manually in the Datablocks viewer fixes this error...
- I noticed that in the buttons view, only 1st of the constraints gets shown. This seems a bit like some of the intermittent problems I've had with some arrays/lists not expanding properly in Datablocks view.
Old NLA-data now gets mostly ported converted over to the new system, with strips and their respective Actions being handled correctly in the test cases I've got.
The conversion procedure now tries to fit multiple strips into since tracks as it is assumed that quite a few old setups tried to do. However, some old setups may be adversely affected by this (i.e. if they depend on a certain order of holding adds for example).
For now, there are no complete replacements for the NLA-Modifier/Auto-Walking stuff yet, so that info is currently just ignored (but correctly freed). The current plan here is to get Armature-level pose-offset system + F-Modifiers where appropriate. This should be one of the major causes of file breakage now...
Also, I've yet to restore some patching for group instancing NLA stuff, since more trickery here is required. This is probably the second major cause of file breakage...
* Strips using the same action as the 'tweaking action' now get the error flag cleared after tweakmode is exited. (These strips draw with red shading)
* The direction in which strips get played (as a result of the 'reversed' option) now gets indicated on strips by the direction of the arrow text printed on each strip
* The active strip flag is now cleared after duplicating/splitting strips.
Updated Mathutils.Vector/Euler/Quaternion/Matrix so these are types rather then module methods, each type now has a tp_new function, matching python builtins float/int/str.
Also cleaned up float conversion and arg passing.
Changed buttons_objects.py...
if ob in groups.objects: # no longer works
if ob.name in groups.objects: # is the new syntax
...its more dict like and a lot faster (avoids python iterating over each item and comparing each, use a single rna lookup instead).
* 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.
When an F-Curve doesn't contain any keyframes anymore but it still exists, the F-Curve's value is not recalculated and flushed anymore if the F-Curve's value will not change. That is, if the F-Curve doesn't have any other data, i.e. drivers or generator-modifiers, which would still change its value, it wouldn't be recalculated to have a value of zero.
This solves the problem of deleting all scale keyframes, whereby objects/bones would appear to 'vanish'
== Transitions ==
Transition strips are now able to be created + evaluated. Transitions allow for interpolation between the endpoints of two adjacent strips in the same track (i.e. two strips which occur in the same track one after the other, but with a gap between them).
- The current behaviour when only one endpoint affects some setting is non-optimal, since it appears somewhat inconsistently extend/replace values...
- Transform code needs a few fixes still to deal with these
== Strip Adding Operators ==
* New strips referencing Actions can be added using the Shift-A hotkey while in the strips-area. You must have a track selected first though.
The new strip will get added, starting from the current frame, in the selected track(s) only if there is enough space to do so. Otherwise, the new strip gets added at the top of the stack in a new track.
* New transition strips can be added with the Shift-T hotkey while in the strips area. You must have two adjacent strips selected for this to work.
== New Backend Methods ==
* Recoded the strip/track adding API to be more flexible
* Added a new method for testing whether F-Curve has any modifiers of with certain attributes. Will be used in a later bugfix...
== Bugfixes ==
- Fixed bug with strip-blending which caused the blending modes to be useless.
- NLA buttons now use proper poll callbacks instead of defining checks
- Commented out missing operator in menus, silencing warnings in console
- Removed obsolete/incorrect comments
* extract_pose_from_pose only checked one of the list items for NULL when looping over them yet its possible they are different sizes.
* game_free_pose needed to be used rather then MEM_freeN, channels would never be freed leaking memory.
* setChannel() would make a new pose that wasnt aligned with the existing pose, the lists are assumed aligned so when extracting the channels its unlikely this was ever useful.
* Added getChannel() - returns pose loc/size/quat
* Added option args for setChannel(channel, matrix) or setChannel(channel, loc, size, quat)
* Added option for panel to be closed by default.
* Added support for RNA property and enum icons in buttons.
* Remove some deprecated RNA menu code.
* Fix issue with newly created panels not being inserted in the
right place.
* Fix issue with 3-split layout not being divided correctly.
* FIx issue with menu items not drawing correct using python UI.
* 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.
actually two modifier datamask optimizations that were never done.
* Don't use modifier data mask for disabled modifiers.
* Check if UV data is needed for particle system instead of always
requesting it.
Big proxy fix (addressing hopefully most complaints on mailing list and
in tracker)
* proxy render settings are now independent of render size settings.
That means: which proxy size is used, is controlled by two parameters
now: the proxy size of the files, can be controlled with additional
buttons within the proxy panels. What is shown in a specific preview
window depends on the header settings of the preview panel.
So: proxies are _only_ used in those windows, that are switched to a
specific proxy resolution.
* output rendering is always done _without_ proxies.
* proxy generation now shows a waitcursor with numbers.
(closing Bug: [#18909] Building Proxies doesn't give any feedback
which was rather a feature request, since the problem mentioned there
was always the case :) )
Checker: FORWARD_NULL (help)
File: base/src/source/blender/blenkernel/intern/modifier.c
Function: meshdeformModifier_deformVerts
Description: Variable "dm" tracked as NULL was passed to a function that dereferences it.
fix provided by Brecht.
Kent
Checker: REVERSE_INULL (help)
File: base/src/source/blender/blenkernel/intern/BME_tools.c
Function: BME_split_face
Description: Pointer "example" dereferenced before NULL check
Check for null before we try to use it.
Kent
* Missing changes to release/windows/installer
* Sequencer fixes in source/blender/src/seqaudio.c dont apply to 2.5
* brechts fix for #18855 r20763 wasnt merged, does this apply to 2.5?
New tracks can be added in the following ways (with the mouse hovering over the channels-list):
* Shift-A - this will add a new track at the top of the stack (i.e. above all the existing NLA-tracks but below the Active Action) for every AnimData block where there was a selected NLA-Track
* Ctrl-Shift-A - this will add a new track above every selected one
* Strip scaling was being evaluated in the wrong way, resulting in scaled strips not being played too fast (for lengthened strips) or too slow (shortened strips)
* Also, verified that the 'reversed' option works correctly (no changes needed to be made here)
Joshua Leung
Janne Karhu
Brecht Van Lommel
Guillermo S. Romero
Campbell Barton
Nicholas Bishop
Martin Poirier
Ton Roosendaal
Peter Schlaile
Kent Mein
Robin Allen