some tests with high poly mesh data in hashes.
- empty buckets before 4-5%, after 1-2%
- speedup for hash lookups, in my tests lookups take approx ~60% of the time they did before.
here is not efficient for such cases, a ray depth can give up to 2^depth rays due
to the ray splitting in two at each depth. A proper solution requires a better
algorithm, for now I've ensured that you can at least cancel such renders. The
overhead from the extra test_break is negligible.
- any error in cloth_build_springs wasn't freeing the edge-hash.
- was checking BLI_edgehash_haskey on matching vertices.
- was looping over setting NULL for all elements of a calloc'd array.
* fix old bug: wrong layout that could happen when switching between thumbnail view and list view. This caused the layout to be recalculated sometimes and the items being moved. Reason was that the layout was wrongly initialized without the scroll bars, so calculated wrongly.
Actually, happened in any view2D... Just added a call to WM_event_add_mousemove() in view_pan_apply, so that overed/active button is refreshed in this case.
Object.imat isn't always the inverse of Object.obmat, needs to be set before usage
as mentioned in DNA_object_types.h.
Thanks to Miika for tracking down the cause of this bug.
Use scene's GPencil when active object is deselected. Else it can be tricky and not user-friendly to access to the scene's GPencil once some objects have GPencil data (you have to select/active a non-gpencil object, or switch to a layout without active object...).
Notes:
* Made those edits by full checking of py files, so I should have spoted most needed edits, yet it remains quite probable I missed a few ones, we'll fix if/when someone notice it...
* Also made some cleanup "on the road"!
In previous optimization in outliner I assumed that order in treehash was not important.
But testing outliner in datablocks mode revealed a problem: when user expands multiple recursive levels and then closes any element, it always closed the top level of recursion.
Now it should work fine with recursive trees.
Now treehash contains groups of elements indexed by (id,nr,type). Adding an element with the same (id,nr,type) results in appending it to existing group. No duplicates are possible in treehash.
This commit should also make lookups a little bit faster, because searching in small arrays by "used" is faster than searching in hashtable with duplicates by "id,nr,type,used".
col = layout.column(True)
row = col.row(False)
Items in row would be 'aligned' in the same group as those in col. Now to get this happening, you have to set row as aligned as well.
Please note that fixes for py UI scripts will follow in another commit.
Also fixed labels of RNA pointers searchboxes, which were missing the colon!
builtin classes. This was done in Python 2.3 to prevent changes to the 'object' type definition and similar issues. As explained by Guido van Rossum in the following mail, the python check will look for
the *closest* base class, which fails for RNAMetaPropGroup because its first base is RNAMeta, which is in turn a subclass of 'type'.
http://code.activestate.com/lists/python-dev/34489/
The easiest and safest way to prevent this issue therefore seems to be
to swap the base class order for RNAMetaPropGroup, so that StructMetaPropGroup is the first base, which has a perfectly valid setattr implementation.
