* The new default particle size is quite small, so exploded pieces didn't match the original pieces
* There's now an option to use the particle size (useful for some effects), but it isn't used by default
* This commit will change how some old files look (explode modifier and not 1.0 particle size), but the exact old behavior is achieved with the new "size" option
Cleaned up a code mismatch that I think got there by accident before.
Now area swap is assigned to Ctrl LMB drag to prevent any 'emulate mmb' conflicts.
Removed extra brush features that aren't applicable in some painting modes from the UI.
Would be great to have this properly cleaned up and made consistent.
Also tweak to startup blend, hiding brush tools panel.
Submitted by: Harley Acheson
Harley writes:
The normal convention in Windows is to show drive letters in uppercase, as in "C:\", not "c:\" as
seen in the File window.
Although Windows file paths are generally case insensitive, drive letters have been shown as uppercase since the early
DOS days. Seeing them in lowercase in the Blender File window looks a bit strange.
Keymap conflict. Remapped all shrink/fatten or 'fatness' resizing tools to use Ctrl-Alt-S instead of Alt-S, since Alt-S needs to be kept for consistency with clear transforms operators.
- Added an optional string arg to struct.keyframe_insert() and struct.keyframe_delete() for the name of the group to add the keyframes to (for newly created F-Curves), instead of doing this as post process.
- Added error prints to the RNA function for setting an F-Curve's group. The old way of setting the groups afterwards couldn't be used anymore, since there was no way to find the action the F-Curve belonged to. This is necessary if the F-Curve list is to be kept in a valid state, since adding to any random group that may not be in the same Action does not work well. There were other issues with the list being iterated over changing while it was still being iterated over too...
TODO:
Find a way to allow the iterator there to still work ok?
by Xavier Thomas
This adds the waveform monitor and vectorscope to the image editor 'scopes'
region, bringing it inline (plus a bit more) with sequence editor functionality,
and a big step closer to the end goal of unifying the display code for image/
comp/sequence editor. It's non-intrusive, using the same code paths as
the histogram.
There's still room for more tweaks - I modified the original patch, changing
the openGL immediate mode drawing of the waveform display to vertex arrays for
speed optimisation. Xavier can look at doing this for the vectorscope now too.
Thanks very much Xavier!
create_mesh, create_dupli_list, make_display_list and is_visible.
This is done in order to make these context independent as the RNA API should
be as much as possible, and to fix#21297 and #21719, where there was an
assumption from these functions that there is a scene in the context, which
does not work for external render engines exporting in a separate thread.
Also avoided using context in a number of other functions, ideally only UI/WM
type functions should use context.
I've updated the scripts in trunk, but the addons and external ones in
development will need updates too.
Now, when the mode is 'replace', no F-Curves are created during keyframing (i.e. only existing F-Curves are used).
---
Also, fixed missing line in previous commit for Select Linked.
Just make the operator cancel if the region in the context is not a RegionView3d.
This is not an issue isolated to this operator, it affects other ones that require a
RegionView3d too, like knife. One to solve for later, at least this doesn't crash now.
* Fixed crash with when/after undoing Grease Pencil sketches in the Image Editor. To get this working, I've had to sacrifice the backwards-compat loading for old .blend files from 2.49 which were using the same pointer as the new data (local data vs datablock data). If anyone has any ingenious hacks to fix this, go ahead, but otherwise, some caution is advised here...
* Grease Pencil block used for Image Editor instances can now be changed
* Enabled proper active polls on the advanced 3D-view projection methods for drawing. These are not valid in views other than the 3D-View, and so should not be shown as available for use elsewhere (though may still be shown just so that users know what mode an 'imported' datablock may be using).
* Enabled proper updates for Grease Pencil editing in the Image Editor. Was missing notifiers for this stuff.
* When drawing in the Image Editor with no image active, the system now (quietly) switches to using screen-space 'view' vs 'cursor' space for the strokes, since the previous behaviour was confusing with strokes disappearing after drawing them. IMO this is a much better option than the confusion that occurred before, even though it does change the user's settings under their feet!
Didn't crash for me, but did export incorrectly since editmode changes weren't flushed.
Fixed this by setting all exporters to go back to object mode before exporting.
Also cleaned up some naming in exporter creator strings, it's "Blender" not "Blender3D".
Added back the 'corner cut pattern' as an operator property for subdivide operator.
Not added for knife yet, since operator redo doesn't seem to be working at all -
already logged in the tracker.
Early when implementing the Graph Editor in 2.5, a key complaint that was levelled at the old 'IPO Editor' was that it was a constant annoyance that adjacent handles were getting selected in addition to the keyframes, when only the keyframes were intended. I solved this by making this default to only selecting keyframes and ignoring the handles, but this means that it isn't possible to batch move several handles at once.
I've now improved this situation by adding an option to the border select operator (involved using Ctrl-B instead of B) which makes the handles get treated separately (as if they were separate verts, as in 2.4x). The default is still to only select keyframes, to have consistency with the DopeSheet...
Also performed some more renaming work in the code...
R 27991 broke all script-based Keying Set support, including all the Built-In Keying Sets. This meant that it nearly impossible to still keyframe anything (there are other less convenient ways, but none work quite as well).
by Harley Acheson (harley)
note from the submission
-snip-
While in there I removed the Win32 conditionals that made "_strdate" used to
get date on Windows. "localtime" works fine on Windows so the results are consistent on all platforms.
Also I noticed there's a check on strip.type == 'EFFECT', which can never be true because strip.type actually contains the type of effect, e.g. "GLOW", "ADD" etc. Not a big problem currently because it's not used.
