SUMMARY Regression introduced in 4.2.8: When a filter mask is 100% white and is copy-and-pasted to a new layer, the pasted filter mask will be 100% black. This does not seem to occur if the filter mask is not 100% white. STEPS TO REPRODUCE 1. Add a filter mask to a paint layer. 2. Copy filter mask. 3. Add a new paint layer. 4. Paste filter mask. OBSERVED RESULT The pasted filter mask will be an inverse of the copied filter mask. EXPECTED RESULT The pasted filter mask should be 100% white. SOFTWARE/OS VERSIONS Linux/KDE Plasma: Debian sid KDE Plasma Version: 5.20.5 KDE Frameworks Version: 5.78.0 Qt Version: 5.12.9 (Appimage)
Amother related issue introduced in the same version: If the copied filter mask is not 100% white (i.e. includes a small stroke of black), then the full filter mask won't be pasted, instead only pasting the coverage of the black stroke of the filter mask.
After more experimenting, I don't think it's pasting 100% black, but rather, the pasted mask seems to hold nothing... The more common scenarios of making a selection, creating a filter mask from it, then copy-and-pasting seems to work as expected. Also, this seems to affect other masks besides filter masks.
This is very likely a duplicate of 430474, especially as it's introduced in the same version and the same workaround by Ahab works. *** This bug has been marked as a duplicate of bug 430474 ***