For the commands that just take a new and old image (or set of images, in the case of layers), we should see if it's possible to reduce them down to more generic commands that just take an image (or set of images). To do this we have to check that the commands can still be accurately represented in a history panel (undo view). To do that we need a model that can represent undo actions:
As we use macros (for reasons I don't remember), it should be possible to use the text of each macro to identify the command, even after removing the duplicate classes.
For the commands that just take a new and old image (or set of images, in the case of layers), we should see if it's possible to reduce them down to more generic commands that just take an image (or set of images). To do this we have to check that the commands can still be accurately represented in a history panel (undo view). To do that we need a model that can represent undo actions:
https://bugreports.qt.io/browse/QTBUG-95640
As we use macros (for reasons I don't remember), it should be possible to use the text of each macro to identify the command, even after removing the duplicate classes.