The code now requires Python 3.7 or newer, as a side-effect of the changes
required for compatibility with 3.9 (as used in Blender 2.93). As a result,
Blender Cloud Add-on now requires Blender 2.80 or newer.
- No longer list Attract's RNA classes, these haven't changed in a long
time and it's not interesting to see.
- Reduced log level when updating internal state. The result of the update
is already logged at INFO level.
When the to-be-rendered blend file is contained in the job storage
directory, it is now assumed that all files are already reachable by the
Flamenco Workers. This supports environments working directly on shared
storage.
This assumes that the paths are already correct for the Flamenco
Workers. No detection of missing files is done (as BAT doesn't run).
The properties are still declared in the Python 3.5 compatible assignment
notation, and a class decorator that converts those to class annotations
as preferred by Blender 2.80.
The `Image.gl_load()` call was changed in Blender commit
7ad802cf3ae500bc72863b6dba0f28a488fce3d1; the two parameters we were using
were removed.
This commit fixes the exception and makes the texture browser usable again,
but doesn't properly fix everything. The textures are drawn in the wrong
colour space, which will be fixed in another commit once I know how.
When using Blender Cloud Add-on 1.12 or older, Flamenco Server will
automatically convert the Manager settings to version 1. As a result,
upgrading is recommended but not required to keep working with a newer
Flamenco Server.
The frame chunk size has a slightly different meaning when rendering
progressively (Flamenco Server can choose to chunk more frames together
when rendering a low number of samples).
Any 'Create Video' Flamenco task that's part of the job will pad the video
with black pixels to make the dimensions even, and this warning notifies
the artist about this.
