The issue was caused by the render engine loading edit mesh, which re-allocates
mesh array which might be referenced by other object's derived meshed.
Worst thing about this is that updating render engine happens from the end of
scene update function, after all the objects are updated and so. This is needed
so render engine gets the update objects which is correct.
The only proper way to solve the issue is to make it so viewport engine does not
leave objects in inconsistent state, meaning nobody will reference to freed data.
In order to reach this we do edit mesh loading before running objects update so
all the objects which uses that mesh will have proper references in the derived
mesh.
This also solves old creepyness which happened before when having single object
in edit mode. tweaking it will calculate derived mesh as a part of scene update,
then this derived mesh will be freed by edit mesh loading and viewport will be
creating derived mesh again.
Now render engine is expected to do nothing with meshes which are in edit mode,
but they still need to load edit data for non0meshes. It's not really easy to
do from the BKE level because needed functions are implemented in the editor.
Thanks Campbell for the review!
Differential Revision: https://developer.blender.org/D697
Summary:
Before this adding Surface type of force field and removing
this field would leave Surface modifier alive in the stack.
This might be really misleading and annoying.
Now removing force field will ensure no modifiers needed for
it are remained in the stack.
This also fixes missing notifier to redraw modifier stack
when changing force field type.
Reviewers: brecht, campbellbarton
Reviewed By: brecht
Differential Revision: http://developer.blender.org/D13
this was only done in some cases before and it was possible to enable weightpaint+sculpt at the same time when enabling sculpt by directly running the mode switching operator.
add generic function to ensure a compatible mode before entering the new mode (added to each operators exec function)
There was actually two different issues:
* mode and toggle options had "remanant" values, when e.g. you used MODE_EDIT/toggle, these would be used again with the newly defined keymap (which sets eg MODE_OBJECT/notoggle)... Defining those props as PROP_SKIP_SAVE fixed this.
* Toggling was not supported for MODE_OBJECT, fixed this (can indeed be handy to have a shortcut to toggle between current mode and Object one ;) ).
That ended up in tricky code trying to mimic depsgraph
branch behavior API-wise preserving texspace and bound
box calculation compatible with previous releases.
So for now bring cu->disp back to the trunk but keep
texpsace and boundbox APIs the same as in the branch.
This keeps texpsapce and boundbox behavior fully compatible
with previous releases and still makes API the same as
for meshes.
flush the selection on entering editmode instead (since the selection mode can be changed with another mesh).
is other tools leave the selection incorrectly flushed, those will need to be fixed so transform works as expected.
Fix turned out to remove as much "manual UI" from 3D view header as possible. Mode selector and all transform manipulators/orientations stuff are now RNA-based UI (leaving basically only edit mesh select modes with custom handlers, as they have some quite specific features).
To achieve this, four main modifications were done:
* enum-operator-generated menus are now MENU (i.e. dropdown lists) in headers too.
* All bit-flag enums expanded in ROW buttons now have a handling consistent with e.g. layers, or what we already have for transform manipulators, i.e. clicking select only one element, shift-click to select multiple ones.
* Consequently, the three RNA booleans manipulators flags are merged into a single bit-flag enum (yes, this is also an API change, though I doubt many scripts use it).
* Now the width of enum-based dropdown lists is computed from longest item name in enum, no more from a dummy place holder string (when no label/name is given).
All this allows to remove some code from 3DView/transform areas, that was actually mostly duplicating RNA/operator one.
Also done a few optimizations here and there (among others, do not pass &numitems to RNA_property_enum_items() when you do not need it, saves at least an iteration over enum items to count them).
Many thanks to Brecht for the reviews!
