Age | Commit message (Collapse) | Author |
|
The calls were missing BONE_UNKEYED flags, so the options looked like they were not working.
|
|
constraint
Fixed the code to be able to handle cases where the size of the chosen axis was so small (i.e. close to zero) that it is likely to cause division-by-zero errors and also result in infinite loops in places... Fix will be ported to trunk at some point.
|
|
Thanks Martin!
|
|
Remaining conflicts for Aligorith to solve:
source/blender/blenkernel/BKE_blender.h
source/blender/include/transform.h
source/blender/makesdna/DNA_armature_types.h
source/blender/src/editarmature.c
source/blender/blenloader/intern/readfile.c
|
|
Object was visible or not
|
|
I've added a check for selected bones in the function which assigns/creates the Vertex Groups.
|
|
commit.
|
|
* Keyframes in IPO's now get considered when doing this
* Vertical placement of view has been improved, though it's still not optimal
|
|
The buggy cull keyframes option in DopeSheet/Action Editor was always on. Disabled it.
|
|
* Added pivot-point option for rotation/scaling. Care needs to be taken with the 'mouse cursor' option, which currently uses a fixed constant scale factor (Theeth: if you have time, could you see if there's a better alternative ;)) as the scaling would be 'way' too sensitive (as starting mouse position = transform center/pivot).
* Excluded IPO-handles from autosnap
|
|
Currently, this only works for PoseChannels (i.e. bones in Pose Mode). Also, added these settings to the toggle/enable/disable setting tools.
|
|
included for ALL node files. This should mean that I don't have to sit through the nodes being recompiled after a simple animation-related change (or something in another unrelated place).
|
|
This setting can be found under 'Edit Methods'. By default, 'Bezier' interpolation is still used, but this can be changed to another option.
|
|
* Added scaling support to the column-select tools
* Fixed the Action keyframe summary that's included in the Object summaries, so that it shows with NLA scaling
* Made a few irrelevant entries in the menus be hidden too
|
|
* Added support for NLA scaling throughout most of the Dopesheet functions, although for a few it is still difficult to achieve. (Notably in column-select tools, but also Time-Slide tool)
--> 'Object' summaries still need some corrections to get the 'action' summaries they include to be scaled too...
* Improved the NLA-Editor's display of keyframes from scaled actions
--
* Also, bumped up subversion number to 3, as a few defaults were changed:
- Dopesheet is now default view (instead of Action Editor). I may review this again later, but it seems OK.
- Turned on transform-autosnapping for Action/IPO/NLA Editors by default and also for old files
- Tweaked Grease Pencil sensitivity settings again... they were still too crude, as could be seen from trouble people often had when making demo vids.
|
|
|
|
* Joining mesh without shapekeys to mesh with shapekeys now works.
--> Joining mesh with shapekeys to mesh without still doesn't work ok though.
* Prevented unlikely crashes in key_get_named_keyblock()
|
|
Now the base mesh doesn't need to have shapekeys for keys to get joined. However, there's still a bug where vertex coordinates still seem to be wrong (but I haven't figured out where it's going wrong yet!) Help is welcome on this issue... ;)
|
|
It is now possible to join several meshes that have shapekeys (using Ctrl-J). This is still a 'first draft' version, as there are still a few cases things that need ironing out still (*see notes).
Be aware that this may very well crash or screw things up in weird and wonderful ways.
Where possible, I've attempted to add some comments and general cleanups to this code to make it easier to follow.
Notes (i.e. things that will be fixed in upcoming commits):
- Joining of shapekeys will only happen if the destination mesh has some shapekeys. If one of the meshes being merged into it has shapekeys but it doesn't, these are ignored/lost.
- IPO's are currently not transferred across yet...
|
|
|
|
|
|
Now the IPO-Editor has the Auto-snapping setting that the Action/NLA Editors have, which mean that the keyframes will always be snapped to the nearest frame. This means that keyframes can now be moved without having to constantly hold the shift/control keys.
Notes:
- Due to the difference in implementation, there is no "frame/second step" option here, but that was not too useful anyway.
- As there are values as well as times for keyframes, this frame snapping can only be turned on/off using the menu (and not any hotkeys)
|
|
* Fixed the long-standing problem where a selected Beztriple's handles would not be able to be scaled when if one of the handles wasn't moved first. The cause of this was that auto-handles will keep handles clamped in place.
To solve this, I've added a special hack, which will result in auto-handles being converted to aligned handles whe rotating/scaling BezTriples that have all its points selected. This should behave more in line with expected behaviour.
* Setting per-segment interpolation in curve EditMode was broken. Was a misplaced check that I forgot to move.
|
|
This was rather simple to accomplish, as it was pretty much the same as for recentering meshes.
Now, to get Join and Separate working... (no wonder nobody has bothered to touch those ;) )
|
|
* Added support for Material, Lamp, and Camera IPO's to DopeSheet
* Added appropriate filters for the added datatypes
* Fixed issues with Constraint Channels caused by a typo in an earlier commit
|
|
* Made current frame indicator waste less space
* Minor formatting tweaks
|
|
This now uses the same method used for armature bones points, with only a few core lines being drawn (one per axis).
|
|
* Moved the context-checks to a centralised place.
* Disabled a few menus which will not work with the dopesheet
|
|
Improved the Object channel selection behaviour to be more consistent with the rest of Blender
|
|
Fixed the OpenGL voodoo needed to get the frame indicator showing in the right place in the IPO Editor.
|
|
Added a few basic options for filtering what gets displayed in the Dopesheet - Selected, Non-Armature Objects, Armatures, Shapekeys, IPO blocks, Action blocks, Constraint Channels.
There are still a few things that will need corrections before everything works with this; namely, the keyframe summary for Object channels.
|
|
Added an (optional) frame number indicator beside the current frame indicator (green line) in Action/NLA/IPO editors. This can be turned on/off from the View menu. It draws as a green box containing the current frame number, and is attached to the bottom indicator line.
Notes:
* The placement in the IPO Editor is still quite non-optimal
* The aim of this was to make it easier to see what frame is the current frame. However, this may be too obstructive for some people.
|
|
As an experiment, handle vertices in IPO Editor are now drawn as unfilled circles. This is more in line with what other packages are doing, and is aimed at making the handles more visibly different from the actual keyframes.
Comments are welcome on whether these contribute more visual clutter than worth, any performance hits on more intensive shots, or OpenGL incompatabilities (there shouldn't be issues, as other parts of Blender does this too).
|
|
Shapekey keyframes didn't show up in the 'object' summary channels
|
|
* Added option to only flatten the handles
* Added to toolbox
|
|
* Added ShapeKeys to DopeSheet
* Sliders now work in the DopeSheet
|
|
|
|
* Insert Key options now work in the Dopesheet
* Fixed selection filtering code for Object channels (nothing much to really see here)
|
|
* Made colours for Object and its sub-channels themeable colours (set by default to this bluish colour)
* Backdrops behind keyframe summary for Object and its subchannels now use the theme colours to help differentiate them a bit. This is still not enough (same applies to group keyframes), but it is a start!
* Auto-merge keyframes option now works. Also, transforming keyframes now validates the order of the keyframes too (as it should).
|
|
Fixed selection problems and crashes with Object->IPO keyframes
|
|
Made the Action Channel expand widgets easier to click on. Previously, clicks were only detected left of the widgets.
|
|
This commit introduces the 'DopeSheet' mode in the Action Editor. It shows the animation data of multiple objects divided into the relevant categories (IPO, Action, Constraints), and allows the editing of that animation data.
A picture says a thousand words:
http://aligorith.googlepages.com/dopesheet_03.png
Most parts of this new editor should work OK, however, there are some tools (most notably channel-related ones) that still require attention before they are usable here. Also, please be aware that there may still be many bugs in this which I have not yet unconvered during testing.
Additional Notes - Further work:
* Although this may sound very similar to what the NLA Editor currently does, it is necessary as the NLA Editor doesn't show the animation data in sufficient detail (and with fewer and less well maintained tools) as that's not what it was meant to do. In the not-too-distant future, the NLA Editor will be refreshed to refocus on its core purpose, so this functionality needs to be moved here first.
* Currently, colour of Object/Sub-object channels are hardcoded defaults which will be turned into themeable colours soon.
* Keyframe 'Summaries' for Object/Sub-object channels should probably get a different coloured backdrop (like for groups) to help indicate the difference between them and standard channels
* Sub-object channels are currently collapsed by default. Feedback would be welcome as to whether these should be unfolded instead.
|
|
A rough attempt at giving the IPO-editor a toolbox for 'quicker' access to commands. I've tried to only include those options which are most likely to be changed, while for the rest, the menus will suffice for now.
|
|
* Formatting and style tidyups for most of the code (some parts haven't been touched yet)
* Patched up some areas where new code hasn't been integrated yet
|
|
* Cleaned up more of the IPO-code. There's still a LOT of messy code lying around that needs attention
* Restored the double recalculation for curve handles when inserting keyframes, in light of some comments I discovered in messy code
* 'Alpha' sliders are now available when setting the colour of curve handles to help differentiate them from the keyframes.
* Silenced compiler warnings in transform code due to changes in types for BezTriple vars
|
|
* AutoIK did not work with euler rotations yet. The conversions were still only being done for quats.
* Fixed spelling errors in tooltips
|
|
Added a new option to the View menu of the IPO Editor - 'Show Handles'. This is turned on by default, but when disabled, all handles on BezTriples will be hidden, making it easier to see the keyframes.
|
|
|
|
There seems to be a consensus that the behaviour of the IPO-borderselect tool was far too messy.
1. "It's too easy to accidentally select tangents of other keys".
2. "When moving multiple keys, the handles for the selected keys (and only those keys) should move automatically with them"
I've modified the behaviour so that the following occurs:
* Only keyframes can get selected by borderselect (i.e. only the coordinates of the keyframe but not its handles are tested if they lie within the box)
* Handles of a keyframe which lies within the borderselect region will inherit whatever selection status gets applied to that keyframe. This is consistent with clicking on the keyframes individually.
|
|
In this commit, I've attempted to improve the situation of bone roll problems that occur with the recalculate bone roll tool, and also when extruding/moving bones under certain conditions.
What's changed:
* When transforming bones, as part of the recalculations done on every update, the roll of bones who have their tip selected will be recalculated using the 'z-axis up' method. Reported by Glenn Melenhorst on BA.org
* Split the 'z-axis up' and 'align z-axis to cursor' roll recalculation methods into separate functions that can act as callbacks. Refactored the auto_align_armature() function to support this. It was needed for the first major change to occur.
* As an experiment, I've changed the final step of the z-axis method to use mat3_to_vec_roll instead of atan2 directly. It currently seems to work better, but I'm not totally sure yet. Will check on this again another day...
|