https://bugs.winehq.org/show_bug.cgi?id=49852
--- Comment #3 from Martin Storsjö martin@martin.st --- Thanks for having a look!
(In reply to Jacek Caban from comment #2)
Avoiding using tty code path reduces that time to 2.5s (something like wine ihateeh-x86_64.exe |cat)
Oh, that's interesting. So pipeing stdout in that case makes it a non-tty, avoiding the slow code, but redirecting stdout to /dev/null still considers it a tty?
The application takes about 14s to run on Windows for me, so our performance is not really that bad in comparison.
Fair enough - I guess the testcase is rather extreme in the amount of output it produces. (The only sensible way to run it is by piping the output to /dev/null anyway.)
I guess one way to avoid it altogether would to have a look if I could cut some of the amount of IO done in the testcase, while keeping the functionality...