- Camera to 3D view didnt check for rotation order.
- Fly mode didnt check for rotation order.
added util functions.
- object_apply_mat4(ob, mat4); applies a 4x4 matrix to an objects loc,scale,rot (accounting for rotation modes)
- object_mat3_to_rot(ob, mat3, use_compat); apply a 3x3 matrix to the objects rotation, option to use a euler compatible with the existing euler.
- appending scenes would not append the sound and scene ID's for sequence strips
- reload button in sequence header now reloads sounds as well.
- redrawing the sequence image view didnt work while plaing (unless play was activated from that region)
- generic functions for running a callback on sequence strips recursively. seqbase_recursive_apply() and seq_recursive_apply()
- bind marker with camera was set to home key, use Ctrl+B instead.
This commit attempts to fix some of the bugs which were causing grief with some Durian animation tests.
In one of those files, the order in which F-Curves were stored was seriously messed up; causing problems with some F-Curves still existing but unable to be edited (i.e. still showing up in the Object/Action summaries but nowhere else) since the standard assumptions for the way the data was stored had been violated.
I've recoded the code that ensures that when F-Curves get added to Action Groups (and the Action that contains these) it ends up in the right places, since it was very likely that all the F-Curves would only ever get added near the end of the list.
Hopefully this is enough to prevent these problems reoccurring, though I have a feeling there may still be one or two buggy tools which caused the problems in the first place.
Finished wrapping Action Groups in RNA to help debug some bugs showing up in the Animation Editors for some files from the Durian team. Access is strictly read-only for these added settings, given the trouble already caused by these problems.
Takes into account the hierarchical structures of keymaps as well as wildcards (KM_ANY) in event definitions, user remaps (emulate numpad, action/select mouse buttons, ...) and event values that overlap (click, press and release)
For now, doesn't do anything other than print conflicts in the console.
As a result, I cleaned up a lot of keymaps that had double definitions, moved some keymap items in more appropriate places, fixed wrong definitions and removed kmi that were added for testing a long long time ago.
Out of all the remaining conflicts, after removing obvious non-issues, here's what remains: http://www.pasteall.org/9898
Saving back orientation in operator didn't take into account that constraint orientation can be different than user selected orientation.
Also simplify the switching logic a little.
Until we have proper multi-object editing, this adds a Copy To Selected
option to the right mouse button menu for Object and Bone properties, to
copy the value from the active object to the selected objects.
Also includes some implementation changes to reset to default operator.
Fix# 20043 & 20392
The issue is that OSX lib does not implement TLS (Thread Local Storage), so libgomp uses pthread functions to read/write thread specific vars.
But this implementation is currently (gcc 4.2) buggy : the write function is called only at lib start (in main thread), and the var is undefined for background thread.
The workaround is to perform this gomp_tls_key var write at beginning of background threads that use openMP. (Currently: render & fluidsim)
- multiple pastes after copying
- clear the sound handle when copying (was crashing)
- allow seq freeing without a scene (assumes seq strip isnt active and sound handle isnt set)
- free clipboard sequences on exit
- paste sequence strips using the relative playhead location from when they were copied.
TODO
- check scene pointers on paste
- detect overlaps after paste
Added a new operator for properties which resets RNA-based settings to their 'default' values, as defined in RNA. This currently only works for floats, ints, enums, and booleans (strings and pointers still need to be implemented).
The current extensions to the RNA API that I've made here seem a bit excessive, and can be toned down if necessary. In short, I've just added accessor functions for the default-values of the property definitions.
For this to be really useful, many properties in RNA will need to get defaults defined, since the current defaults for quite a few properties tested were less than ideal.
Former implementation was to populate the "system" paths with the mounted paths. The issue is that OSX mounts some paths that are of no use such as /home (The home dirs reside in /Users).
When compiled with 10.5 SDK, it retrieves the folders you have in the finder and OSX standard open/save dialog.
When compiled with older SDKs, it assumes these folders are the default ones (home, Desktop, Documents, Pictures, Music, Movies).
New unique ID per keymap item (unique inside their keymap) for default and configuration keymaps.
This allows restoring a single user defined kmi to its previous (default or config) values instead of having to restore the whole keymap.
The restore item button is disabled for kmi added by the users (they don't have an ID).
Also fixes a bug in the rna function for add keymap item (parameter order was incorrect, messing adding back saved configurations).
Display job status in web interface
Better use of Reports api for errors and success notifications
Don't show some Client options if server address is default (hasn't been scanned or entered manually yet)
Currently access by selecting a marking and binding with the active camera from the view menu.
Note:
after long discussion we decieded there is no nice way to do this.. animate pointers? animate multiple camera visibility?, use sequencer? use NLA?.... have a kind of event system (like framechange scriptlinks)... etc
so this is ifdef'd with DURIAN_CAMERA_SWITCH
