https://bugs.winehq.org/show_bug.cgi?id=56459
Eric Pouech eric.pouech@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |eric.pouech@gmail.com
--- Comment #1 from Eric Pouech eric.pouech@gmail.com --- Did you test your program on windows, so that we can really tell what's wrong or right? That's IMO the first thing to do (see below).
Two global comments: - most of potential issues are likely not really about "flushing" but around: 1) (file) handle inheritance. Wine and Proton 9.0 have a bunch of changes regarding how they are inherited. Esp. handling GUI processes (as opposed to CUI processes) where Wine's behavior now closer mimics Windows. And GUI processes (like your test program) don't inherit console nor std file handles (unless explicitely told to) 2) several operations (like console creation / deletion, killing process...) happen to close handles (to the underlying object, which could even no longer be available). But your tests still expect some internal buffers to be flushed under these circonstances.
To move forward, could you please update your tests with: - running your test under windows and mark where there's a difference from Wine 9.0. It's very likely that your expectation would be closer to what Wine 8.0 was doing, unfortunately, it's not exactly how Windows behave. - when testing on Unix, you should also consider two cases: A) running from the Unix shell B) running from a windows-like console (using wineconsole) Results in B) should (that's the theory) be identical to Windows (if, not, it's a bug). Results in A) have no equivalent under Windows, so we had to make compromises between the two worlds, yet we have more flexibility on what's possible or not.
Depending on the outcome, it's likely we could open different tickets for different tests (as they could concern rather different part in Wine).
TIA