Re: [Denemo-devel] Showstopper bug?

[Richard Shann]

Thank you for looking at all this stuff.  We have to balance the pain we
cause users leaving them with versions without the currently fixed bugs
against giving them a new version with those fixed and new ones
introduced... we have gotten really glacial with our releases (well, not
compared with some projects :), there are several serious issues that
have been fixed in git which are not available in the distro packages.

Richard


On Thu, 2012-05-10 at 21:49 +0200, Dominic Sacré wrote:
> Hi Richard,
> 
> On Thursday 10 May 2012 10:44:20 Richard Shann wrote:
> > Dominic Sacre came on irc #denemo to report problems with playback. He
> > pointed out that the page_for_time() is not thread-safe.
> > I felt torn about declaring this a showstopper for the release, as we
> > have been so long in getting this release out, but threading bugs are
> > so pernicious that I think we have to put in a fix for this and make
> > another release candidate.
> 
> I noticed more playback issues. Some of these may be specific to certain 
> backends (or combinations thereof), I'll have to do more testing and 
> debugging to be sure.
> 
> - Sometimes playback just doesn't start. The "playback only works once" 
> issue I mentioned before seems to be a common case of this problem, but 
> somehow this has become harder for me to (intentionally) reproduce.
> 
> - Occasionally playback stops after a few seconds, always at the same 
> position in the score.
> 
> - The JACK backends cause an xrun every time I stop playback. Every now 
> and then, Denemo gets zombified by JACK.
> 
> - PortMidi output is completely disabled, due to an unconditional return 
> statement in the process_midi() function. If, as the comment in that 
> function suggests, the PortAudio backend supports some feature that the 
> PortMidi backend doesn't, doesn't this apply to the ALSA and JACK backends 
> as well?
> 
> All in all, I don't think Denemo in its current state is quite ready for 
> release yet. The chance of users stumbling upon one or more of these 
> issues is just too high.
> If that's ok for you, I'd like to try and fix as many of these bugs before 
> the next release candidate. As far as I can tell none of these issues 
> existed half a year ago, so it shouldn't be too hard to figure out what 
> changed since then, and hopefully I'll have some time to do that this 
> weekend.
> 
> 
> Dominic