[vlc-devel] [PATCH] darwin/thread: Ignore vlc_mutex_unlock failures
ajanni at videolabs.io
Tue Oct 29 14:28:56 CET 2019
I agree for the need of a smart and better fix.
In the mean time, it might make sense to include this fix
in any case as it seems that it's either no assertion at
all or this assertion removed ?
On Tue, Oct 29, 2019 at 02:14:25PM +0100, Hugo Beauzée-Luyssen wrote:
> Apology accepted.
> Now, about the patch. I agree that this sucks, and that if a mutex is still locked while it's expected to be unlocked, all hell breaks loose. That being said, the operating system is at fault here, so we have limited options to circumvent the issue.
> The current state of things is that macOS developers can't realistically use --enable-debug in their builds. I think we can all agree that this is not sustainable.
> If I understood your answer correctly, you're suggesting to do something like win32, where the cleanup handlers are not passed to pthread_cleanup_push/pop (not that it exists on win32), but rather stored in a TLS value, and we can run the cleaners ourselves when the thread terminates, thanks to pthread_key's destructor, or manually when we pop them, if the thread wasn't cancelled
> If that's what you had in mind, I like this solution better, even though it's indeed more complex to implement.
> On Tue, Oct 29, 2019, at 1:38 PM, Rémi Denis-Courmont wrote:
> > And I never claimed that you were biased against Thomas. I expressed
> > that I did not want Thomas to influence you on non technical arguments.
> > Though that was anyway an insensitive statement that I shouldn't have
> > made.
> > Le 29 octobre 2019 14:29:49 GMT+02:00, "Hugo Beauzée-Luyssen"
> > <hugo at beauzee.fr> a écrit : Funny how you're complaining about being
> > singled out after explicitly naming me and accusing me of
> > > - Not doing my job
> > > - Being biased to Thomas
> > > while I haven't stated an opinion yet.
> > > I seriously can't fathom how you can twist things and claim that you're the one being attacked here.
> > >
> > >
> > > On Tue, Oct 29, 2019, at 12:39 PM, Rémi Denis-Courmont wrote:
> > >> Yeah and I would really appreciate if TC members did their job and
> > >> considered the implications and alternatives of a patch instead of
> > >> support blindly removing assertions and singling out another TC member
> > >> for actually doing his job.
> > >>
> > >> Le 29 octobre 2019 12:49:53 GMT+02:00, "Hugo Beauzée-Luyssen"
> > >> <hugo at beauzee.fr> a écrit : On Tue, Oct 29, 2019, at 11:39 AM, Rémi
> > >> Denis-Courmont wrote:
> > >>>> How do you even know that a crash report is related to this issue? What
> > >>>> I see is that this patch makes two assumptions,neither of them are
> > >>>> documented nor backed in any way. But sure send it to the TC.
> > >>>> Obviously, you will provide explanations what the assumptions
> > >>>> underlying this patch are, how they are verified, and why no proper
> > >>>> workaround can be made instead, so that the TC can make an informed
> > >>>> decision, and you won't abuse your ties with Hugo to distort the vote,
> > >>>> hmm...
> > >>>>
> > >>> I would *really* appreciate that you'd refrain from this kind of implications.
> > >> --
> > >> Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser
> > >> ma brièveté. vlc-devel mailing list
> > >> To unsubscribe or modify your subscription options:
> > >> https://mailman.videolan.org/listinfo/vlc-devel
> > --
> > Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser
> > ma brièveté.
> > _______________________________________________
> > vlc-devel mailing list
> > To unsubscribe or modify your subscription options:
> > https://mailman.videolan.org/listinfo/vlc-devel
> Hugo Beauzée-Luyssen
> hugo at beauzee.fr
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
More information about the vlc-devel