alpha-channels broken for RGB 8 bit precision images
As the title says, it seems that alpha-channels management with GIMP-2.10.34 and babl-0.1.104 is broken when working with RGB 8 bits integer precision.
The problem can be seen transfering a layers's alpha channel to a layer mask and then doing something with it like "Disable Layer Mask", "Show Layer Mask", copy/pasting it to a new layer and/or image, etc.
It simply shows an empty alpha image or sometimes a flat white alpha making it impossible to work with alpha channels.
This does not occur when working with 16/32 bits integer or float, or using the 0.1.102 version of babl instead of 0.1.104
Slackware-Current, GIMP-2.10.34, babl-0.1.104, gegl-0.4.44
Regards, Julius.