The kernel does not use AVX2 vectorization, and trying to use BVH8 was
leading to an empty scenes.
Fixes T64624: Ctest : Win32 + AVX2 fails virtually all cycles tests
Main goal here is to make it obvious and predictable about
what is going on.
Summary of changes.
- Access to dependency graph is now only possible to a fully evaluated
graph. This is now done via context.evaluated_depsgraph_get().
The call will ensure both relations and datablocks are updated.
This way we don't allow access to some known bad state of the graph,
and also making explicit that getting update dependency graph is not
cheap.
- Access to evaluated ID is now possible via id.evaluated_get().
It was already possible to get evaluated ID via dependency graph,
but that was a bit confusing why access to original is done via ID
and to evaluated via depsgraph.
If datablock is not covered by dependency graph it will be returned
as-is.
- Similarly, request for original from an ID which is not evaluated
will return ID as-is.
- Removed scene.update().
This is very expensive to update all the view layers.
- Added depsgraph.update().
Now when temporary changes to objects are to be done, this is to
happen on original object and then dependency graph is to be
updated.
- Changed object.to_mesh() to behave the following way:
* When is used for original object modifiers are ignored.
For meshes this acts similar to mesh-copy, not very useful but
allows to keep code paths similar (i.e. for exporter which has
Apply Modifiers option it's only matter choosing between original
and evaluated object, the to_mesh() part can stay the same).
For curves this gives a mesh which is constructed from displist
without taking own modifiers and modifiers of bevel/taper objects
into account.
For metaballs this gives empty mesh.
Polygonization of metaball is not possible from a single object.
* When is used for evaluated object modifiers are always applied.
In fact, no evaluation is happening, the mesh is either copied
as-is, or constructed from current state of curve cache.
Arguments to apply modifiers and calculate original coordinates (ORCO,
aka undeformed coordinates) are removed. The ORCO is to be calculated
as part of dependency graph evaluation.
File used to regression-test (a packed Python script into .blend):
{F7033464}
Patch to make addons tests to pass:
{F7033466}
NOTE: I've included changes to FBX exporter, and those are addressing
report T63689.
NOTE: All the enabled-by-default addons are to be ported still, but
first want to have agreement on this part of changes.
NOTE: Also need to work on documentation for Python API, but, again,
better be done after having agreement on this work.
Reviewers: brecht, campbellbarton, mont29
Differential Revision: https://developer.blender.org/D4834
Note that the same issue actually showed without a frame node, when
trying to click-drag on a socket on its 'inside node' part, you would get
same behavior.
Only solution I can see here is to prevent Node selection to go on when
user clicks on one of its sockets, there is no way afaik to make
drag-from-socket to start if we keep select-node operator running
modal, since both operators (NODE_OT_select and NODE_OT_link) use the
same shortcut, if select (which is checked first, being a Tool) returns
modal, then event is considered handled it seems, even though
Passthrough is also returned...
* Align category icons and color use with Default theme.
* Set colors for new properties and regions.
* Fix several issues.
Needs further testing to tackle all editors but this puts it at much more
usable level compared to how it was.
When normally editing curves, deleting the last keyframe also
deletes the curve. Thus if for some reason it didn't happen,
e.g. maybe due to removing keys directly via Python, skip the
bad curve instead of resetting the channel to zero.
This constraint requires full bone data to evaluate, so it can't
use the generic single target matrix system. Calculating it is
thus useless waste of CPU time.
Helps to:
* Make a connection between what we see in Outliner, Viewport and Properties editor
(clicking on the obdata icon takes us to Edit mode, which shares colors).
* Quickly tell which mode we are in.
* Armatures have now distinct color when in Object, Edit, and Pose modes.
Missing is Pose mode which is currently cyan in the viewport and obdata category in properties,
it would probably need its own special category though as it doesn't match the others.
The choices are now World, View and 3D Cursor.
This breaks Python API compatibility, add-ons that add objects with this
parameter will need to be updated.
Differential Revision: https://developer.blender.org/D4706
Both windows.h and wm_types.h define HKEY, moving wm_types after windows.h
sidesteps the issue, since HKEY isn't used in fsmenu.c it doesn't cause
problems.
We need to make those paths validation async, since some infamous OS is
unable to provide a quick way to check whether a path is valid or not...
Not much to say here, this is basic wmJob stuff really... We don’t even
need complex threaded coordination locks or atomics, given how simple
the changes are here.
Fake-tested with some `sleep()` calls to add atificial delay, seems to
work as expected.
This is the angular diameter as seen from earth, which is between 0.526° and
0.545° in reality. Sharing the size with other light types did not make much
sense and meant the unit was unclear.
Differential Revision: https://developer.blender.org/D4819
This removes the ability to hit X or Del to delete data from inside the File Browser.
This was extremely dangerous because it didn't use the system trash/recycle bin. This made it a very dangerous operation with no way to get data back.
Later on, we should add support for the OS trash through native API's, but until then we will remove this feature from the keymaps.
This addresses T61412 and T64104
