Attempted to try to run Qt Nethack for first time. qt-3.1.2-r4 and nethack-3.4.2 1 in 3 times or so it manages to start with corrupted graphics, and often non-functional keyboard input. The other times it segfaults, usually as shown in strace attachment. Two worrying things about the segfault that I can see offhand. Fairly clear memory has been corrupted from the filename it tries to open. And, even more worrying, it seems to have been linked against /var/tmp qt files. Reproducible: Always Steps to Reproduce: 1. run /usr/bin/nethack or /usr/games/bin/nethack 2. If it doesn't crash immediately, see if you have keyboard. 3. Clicking mouse to move adventurer seems to quickly eliminate keyboard responsiveness. Actual Results: Nethack is either unresponsive or crashed. If it managed to start up, portions of the theme (scrollbars, radio buttons, the like) are corrupted. Expected Results: Started up cleanly.
Created attachment 20183 [details] Strace of attempt to start nethack
Created attachment 20186 [details] Emerge log of nethack
At recommendation from lu_zero I updated to ~x86 qt-3.2.2 Works fine now. My vote for fixing the qt-3.1.2-r4 ebuild or pushing qt-3.2.2 to ARCH.
Dang it. It does seem to work now, no corrupted graphics, no crashing. *but* I still see searching for /var/tmp qt files in strace. Quite odd.
the searching of $S is a 'feature' of QT ... dont worry about it, it wont match the files and ultimately settles on the real lib (/usr/lib/libmng.so.1) ... in terms of stability, you'll have to file a new bug to push QT-3.2.2 to stable ... for now i've added a little message to the ebuild saying that qt-3.1.2 may be unstable ...
Don't quite follow... Why is /var/tmp in the path at all? Also, is there any way, if there is a list of search dirs, to push those near the bottom or remove entirely? Seems kind of silly to have that happen every time.