http://bugs.winehq.org/show_bug.cgi?id=33632
--- Comment #12 from Henri Verbeet hverbeet@gmail.com 2013-09-16 03:50:43 CDT --- (In reply to comment #11)
This issue happens to me regardless of the Desktop Manager in use. So far I have also confirmed it on KDE 4 and xfce. So it's not just a GNOME issue.
Well, that's a start at least.
As for nvidia, wine would be practicaly useless for most users (read: home users) with nvidia cards without nvidia's binary driver.
Perhaps, but that's not really the point. NVIDIA's RandR is known to be broken on some configurations, NVIDIA has indicated they don't intend to fix / change this, and we obviously can't fix it ourselves. Even with that consideration aside, it would be very useful to know if anyone can reproduce this with one of the Nouveau, Radeon or Intel drivers. In the worst case, if it turns out to really be specific to the NVIDIA implementation, we could try to fallback to RandR 1.0 if we can reliably detect affected configurations.
For me personally, what I can say for sure is that it works fine for me with both the Radeon and the Nouveau driver on both FVWM and KWin on various hardware. I've also had reports from some people on IRC that had issues with this on discrete NVIDIA cards where it went away after switching to either Nouveau or their integrated Intel GPU with corresponding driver.
Then there's a report at http://bugs.winehq.org/show_bug.cgi?id=31245#c5 that indicates there's potentially a GNOME 3 specific component to this bug as well, but so far nobody else has confirmed that workaround. Really, what this bug needs is for someone that can reproduce this to try a few different combinations of drivers and window managers, and get a good overview about which configurations are affected and which ones aren't.
Another problem is that wine won't restore the original resolution in some instances. Requiring the user to manually reset it. (Which they may not be able to due to the low resolution without resetting X.)
Yes, that's a known issue that unfortunately hasn't received the amount of attention that it perhaps deserved over the years. There's a chance it's related to CDS_FULLSCREEN handling (i.e. the same issue that prevents switching the display mode back when a fullscreen application loses focus), which would make this pretty hard to fix, although that's mostly just speculation. It's not really related to this bug either way though.