Comment 16 for bug 197673

I've reviewed the changes up to Mar 19th, and I see the Mar 28 changes for clone mode, but haven't reviewed those yet. In any case I do plan on updating our code to newer upstream stuff once this revert dialog is done.

I look forward to seeing your xevent test code. Sounds like it still requires creating a test config file, but maybe it'd be a cleaner code path.

Fwiw, I spent the day yesterday testing the dialog (and xrandr in general) on 10 different hardware configurations (3 intel/dell systems, 3 ATI's, 4 nvidia's) using the (default) open source drivers. I haven't written up the testing results yet, but I found a wide range of bugs, mostly non-catastrophic. By and large, resolution changing works, although there are some low-to-medium severity issues on a few cards. Rotation is the thing that causes severe issues. On Intel it works without issue. On ATI it causes catastrophic failure. On nVidia, it isn't even offered as an option. I also tested monitor hotplugging support and found its support varies pretty widely.

Anyway, it looks like it's pretty much the rotation case on ATI that is going to be exercising the revert dialog. ;-) (Maybe it would be good to not allow rotation on ATI at all to begin with?)