https://bugs.winehq.org/show_bug.cgi?id=41107
--- Comment #16 from Sebastian Lackner sebastian@fds-team.de --- (In reply to katsunori.kumatani from comment #15)
Any news for this regression?
I know it's probably hard to fix considering that it's been up for so long after detected with it being a regression, but shouldn't the patch that caused it be reverted? Or is what the patch fixed very important? (forgive me if I ask stupid questions, I'm not familiar with wine-staging's policies)
Wine-Staging uses the same policy as Wine, which is that patches should not be reverted or removed unless they are shown to be wrong. Technically the patch makes Wine behave more correct and fixes bugs in other applications. In addition, TotalCommanders thread handling is clearly broken when it spawns million of threads just to enumerate the directory content. The correct way would be to use (or implement) a threadpool, which can deal with lots of small tasks.
After some discussions with Michael Müller I have updated the patch and reduced the timeout interval, which should improve performance a bit. However, of course it is still not as good as before, where waiting for thread exit was broken. In order to improve it even further it will be necessary to come up with a method which is faster, without breaking what the initial patch is supposed to fix.