Closed fidergo-stephane-gourichon closed 4 years ago
This sounds useful, are there any plans to merge this PR?
Unfortunately, at the moment, I'm so busy that I can only merge things that are quick and easy to test.
The extents calculations are the polar opposite and have no test suite, so, until I can write a test suite for them, I've developed a bit of an "It works for me... DON'T @#$% WITH IT!" attitude.
...however, since you expressed interest, I'll see if I can make time to at least get started on a test suite for them in the near future.
Since the fix for #108 should fix this for the GTK 3.x version and this is for the GTK 2.x version, has merge conflicts, and Python 2.x is out of support, I'm closing this.
Sorry that I never got around to merging it. (Hopefully, the more solid automated testing that I've been growing on the GTK 3.x version will prevent that from ever happening again.)
Computation assumes that removed rectangles cover full width or height. Actual computation for left and right was performed on rectangles that had correct height but were "off" by the space left on top by actual panel, making computation result wrong,
usable_rect
wrong (whole screen). Net result would be panel ignored and windows covered by panels or overlapping each other. Fixed by positioning the trimmed rectangles correctly.