The Undo/Redo actions now tell you what kind of action will be
undone/redone. This is achieved by adding an "action text" field to the
ImageUndoCommand and having everyone who calls did_complete_action()
provide this text.
Each of these strings would previously rely on StringView's char const*
constructor overload, which would call __builtin_strlen on the string.
Since we now have operator ""sv, we can replace these with much simpler
versions. This opens the door to being able to remove
StringView(char const*).
No functional changes.
This algorithm utilizes a modified scanline method that takes advantage
of the fact that if you are filling rows starting from the top left and
going right, you do not need to check pixels very often except in
certain cases such as at the beginning or end of a row.
There are some tests on top of this that ensure correct filling in all
other cases. This leads to much-improved speed compared to the
4-directional queue method, and no heap allocations.
This is in preparation to support masking of Layers. We now distinguish
between the "display_bitmap" which will be the whole Layer with every
effect applied and the "content_bitmap" which contains the actual
unmodified pixels in the Layer.
