[mythtv] Trunk Aborts

[Gavin Hurlbut]

On Wed, Sep 29, 2010 at 9:50 PM, Jean-Yves Avenard <jyavenard at gmail.com>wrote:

> > Disabling pulse audio has caused problems for a while (many, many
> > months) - something that has been effectively ingored by most people
> > by just removing pulse audio. Recent changes in trunk have however
> > made it far worse than before, to the extent that it's aborting on a
> > regular basis.
>
> Which changes are they?
>
> Is that the new system command?
>
> The disabling of pulse had been working reliably for me , though I
> haven't upgraded my main frontend for several weeks now
>
>
I had put in what should fix that a few days ago in r26526, and I have not
heard any reports since.  To be able to debug this, we will need a viable
backtrace of the crashed process.  As it's aborting, this may not be the
easiest to track down.

Also, if this is still occurring with recent trunk, and you can reproduce
it, it's in all of our best interests for a ticket to be created so we can
fix it.  Innuendo on the mailing list isn't nearly as effective.

I also have a potential fix for #9005 that I hope to try out shortly as
well, which could fix several other crashes.

As for the original post, it seems that there is an "assert" somewhere in
the code that is causing this. A quick git grep doesn't seem to find that
assert.  No idea where that code might be.

>> Assertion 'e->mainloop->n_enabled_defer_events > 0' failed at
>> pulse/mainloop.c:287, function mainloop_defer_enable(). Aborting.

The text "defer_events" and "mainloop_defer_enable" is not anywhere in the
mythtv source tree that I can find.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mythtv.org/pipermail/mythtv-dev/attachments/20100929/5d6844da/attachment.htm>