Closed Baklap4 closed 2 years ago
In a nutshell, I asked @Baklap4 the open this.
@Tronic I have a hunch that you added this ages ago because it was hanging for you, yes? If you were on mac at the time, that behavior was reported at least once upstream but it seems it could have been a CoreAudio limitation, rather than a Portaudio bug.
I was hoping we could try to get rid of the audiokiller and see how portaudio 19.7 behaves. If I'm correct in the motivation for it but the error persists, maybe we can keep this behavior only when CoreAudio is being used as backend?
It was probably hanging on Linux, where I recall such problems being common. Could be related e.g. to unplugging of USB sound card while in use. I am not very confident of such problems being fixed so that the audiokiller could be entirely removed, but perhaps it could be disabled for debug builds (like we already disable some exception catching in debug).
And then make sure that releases are actually built in release mode, not in debug with optimizations which is the default for developers.
I am not very confident of such problems being fixed so that the audiokiller could be entirely removed
I guess that's to be tested
And then make sure that releases are actually built in release mode
This should always be done when releasing, so yes!
Just tested out removing those std::async
calls within the destructors. This doesn't seem to resolve the problem that we still encounter an error on exit. Getting a Debug Assertion Failed
statement within debug_heap.cpp
from the appcrt
so i guess it won't be an easy/fast fix ghehe
This might also be specific to Edirol M16DX, a soundcard-mixer suitable for karaoke use but out of production for many years now. IIRC, no-one was willing to fix driver bugs on that even ten years ago or so, and the hanging of devices might've been one of those issues.
I still got one in use, though, but I haven't used it on Linux. Would need quite a bit of testing to confirm whether the hang issue is truly gone, and frankly I don't feel like doing that.
I think keeping that "in case of" is a bad idea. Actually, if the hang is still present in some way somewhere, we'll get some bug open and it will be time to see what kind of setup causes it and see if the former workaround is still relevant.
I think keeping that "in case of" is a bad idea.
Agreed. If it's still there we'll get a bug report. Else the problem is long gone ;)
I'll re-check #676 and see if the suggestion/note by @OznOg there fixes stuff
Do you want to request a feature or report a bug?
Bug
What did you do?
Start performous Stop performous Hit a breakpoint in external code
Error message on console:
What did you expect to see?
I expected performous to actually close without coughing up errors.
What did you see instead?
Errors as explained above.
Output of
performous --version
: (What version of Performous are you using?)Latest git master
What is your environment & configuration (arguments, platform, ...)?
Windows - with
-DCMAKE_BUILD_TYPE=Debug
(different than the not defined build type). I don' t get this error when running inRelWithDebInfo
orRelease
If applicable, please paste the log output in DEBUG level (
--logLevel=DEBUG
switch)https://media.discordapp.net/attachments/273434633138470913/884818022852861962/unknown.png?width=1417&height=676
Notes
@OznOg wrote:
@Lord-Kamina wrote: