[pulseaudio-discuss] [PATCH 0/3] fix PA test suite failure

[Felipe Sateler]

On Mon, Aug 18, 2014 at 5:11 AM, Peter Meerwald <pmeerw at pmeerw.net> wrote:
>
>> OK, so it now seems that with the hurd we still are missing 1 test:
>> once-test[1]. Looks like the autobuilder is killing the build after a
>> timeout is reached.
>>
>> [1] https://buildd.debian.org/status/fetch.php?pkg=pulseaudio&arch=hurd-i386&ver=5.0-10&stamp=1408316471
>
> daemon/pulseaudio-daemon-conf.o: In function `parse_rtprio':
> /«PKGBUILDDIR»/src/daemon/daemon-conf.c:481: warning: sched_get_priority_max is not implemented and will always fail
> /«PKGBUILDDIR»/src/daemon/daemon-conf.c:481: warning: sched_get_priority_min is not implemented and will always fail
>
> #ifdef OS_IS_WIN32
>     pa_log("[%s:%u] Realtime priority not available on win32.", state->filename, state->lineno);
>
> maybe add a similar line for hurd?
>
> #else
> # ifdef HAVE_SCHED_H
>     c = state->data;
>
>     if (pa_atoi(state->rvalue, &rtprio) < 0 || rtprio < sched_get_priority_min(SCHED_FIFO) || rtprio > sched_g$
>         pa_log("[%s:%u] Invalid realtime priority '%s'.", state->filename, state->lineno, state->rvalue);
>         return -1;
>     }
>
>     c->realtime_priority = (int) rtprio;
> # endif
> #endif /* OS_IS_WIN32 */
>
>
> daemon/pulseaudio-main.o: In function `main':
> /«PKGBUILDDIR»/src/daemon/main.c:1017: warning: mlockall is not implemented and will always fail
>
>     if (conf->lock_memory) {
> #if defined(HAVE_SYS_MMAN_H) && !defined(__ANDROID__)
>         if (mlockall(MCL_FUTURE) < 0)
>             pa_log_warn("mlockall() failed: %s", pa_cstrerror(errno));
>         else
>             pa_log_info("Successfully locked process into memory.");
> #else
>         pa_log_warn("Memory locking requested but not supported on platform.");
> #endif
>
> something similar here?

I don't think I can comment on this. However, as long as pulseaudio
works with those functions always failing perhaps it is better to
leave them enabled, so that if Hurd starts supporting them no
additional porting is needed.

In other news, it looks like the test failure is a bug in pthread:
http://thread.gmane.org/gmane.linux.debian.devel.pulseaudio/4502



-- 

Saludos,
Felipe Sateler