On Sat, 2005-12-03 at 00:15 -0500, Zack Weinberg wrote:
> On Fri, Dec 02, 2005 at 03:15:49PM -0600, Joe Wreschnig wrote:
> > On Thu, 2005-12-01 at 16:27 -0500, Zack Weinberg wrote:
> > > On Thu, Dec 01, 2005 at 02:27:03PM -0600, Joe Wreschnig wrote:
> > > >
> > > > Can you download the SVN version
> > > > (http://svn.sacredchao.net/svn/quodlibet/trunk/quodlibet) and see if it
> > > > still happens?
> > >
> > > Sure, I can try that. I'll give it several days to stay gone if the
> > > problem seems to be fixed.
>
> The SVN version has not yet hung up on me. However, it does occasionally
> pop up a completely unhelpful error box (the one generated by
> SongWatcher.error) and halt playback. I tried to hack up the logic in
> player.py to get a more helpful error message, hoping that the GStreamer
> 'error' signal would trigger and I could use that, but it didn't work --
> set_state() is returning gst.STATE_FAILED on one or other of the calls
> from PlaylistPlayer.__load_song, and I can't find any more detailed
> error indication.
As far as I can tell, the various error signals aren't going to be
useful until we switch to GStreamer 0.10. So, the unhelpful error, which
is at least more helpful than a crash.
> Tangentially, I'd like to suggest rearranging the logic in
> PlaylistPlayer.__get_song so that this bit
>
> config.set("memory", "song", song["~filename"])
> self.song = song
> self.info.song_started(song)
> self.volume = self.__volume
>
> all happens before the call to __load_song.
Done in r2386.
> > Another suggestion which seems to have worked for someone on the IRC
> > channel is to remove the "gst.use_threads(True)" line from player.py.
>
> I'll try that next. What IRC channel, by the way?
#quodlibet, irc.oftc.net.
--
Joe Wreschnig <[EMAIL PROTECTED]>
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
