Bright backdrop would get lost when having bright elements/background in the viewport.
This makes it use a darkened/lighten version of the theme space header color.
Since icons are colored using the 'text' value of the Theme Space, it's likely that any
theme is going to have a contrasting background color, but just in case darken/lighten it slightly.
Make edge decoration a bit thinner and try to reduce missing edge segment
due to triangle barycentrics.
This invert the "edge fix" strip offset direction, meanning there is now
the possibility that these triangles poke through nearby geometry depending
on the viewport near/far clips distances.
On some platform does not support line width > 1.0 and can even throw and
error. Better check an at least display something rather than no lines at
all.
This is sued by both object joinig and object apply scale operations.
Currently only uniform scale is working correct. Non-uniform gets averaged
and will produce slightly distorted results. This is something we should
fix, but priority of this particular case is not so high.
User do not care about that 'virtual' user, it only makes it more
confusing. So now, that template always shows actual number of real
users, we already have own dedicated button to show fake user status.
Using FakeUser on an ID sill put its usercount to 2 if it is actually
also 'really' used once. But we do not want to show that to user
(especially not since it also allows to 'make single user' of an ID
already only used once...).
We want to undo the initial usercount from low-level ID duplication, not
to erase all possible extra processing (like e.g. setting Fake User...).
Fixes part of mess in brush usercount.
First of all, follow our naming convention to use module prefix.
Second of all, mesh is being created and is to be freed, we also
have convention for such function names.
This only works thing around, just so we can unlock production here.
Joining objects of a different scale will not yet work correctly.
Proper fix is coming later (the code needs to be ported to new
Subdiv API).
The toolbar now shows brush types, the brush selector now
only shows brushes matching the current tool type.
Details:
- Add's Paint.tool_slots (used by the toolbar).
- Removed custom grease pencil brush tool code.
- Bumped subversion.
See T57526 for details.
- Draw on back buttons moves near mode
- zDepth offset moved from View Panel to Popover
- Target for Stroke mode moved to popover
- New First point snap mode
Center is misleading since it could mean the center of the objects, selection, etc.
Python API is left as is to not break compatibility. Maybe it could be renamed as well?
Just tooltips and RNA labels when possible. Node outputs use their name to keep connections,
so they haven't been changed, otherwise it will break old files.
Node group inputs should behave like cycles now.
----
We create dummy nodes that contains the default values for the nodegroup
inputs and link them to the sockets. This way the uniform buffer gathering
function can read them.
But we also need to disconnect all the internal links to sockets that have
hidden values. Theses sockets must not get the values from the nodegroup
input sockets if there is no link to them otherwise we get broken results
(i.e.: normals for a bsdf node).
- Extruding edges/vertices on normals gave bad/strange results.
- The adjust gizmo used the orientation axis instead of the
last-orientation (which could get out of sync).
Needed for situations when we can't use the orientation.
With extrude the initial extrusion recalculates normals for edges
and vertices which then don't give a useful axis.
