Update on speech-dispatcher and pulse

[Nolan Darilek]

On 05/27/2009 04:29 PM, Luke Yelavich wrote:
> I think I hit the same crash the other day when doing some testing. However, 
> I think the latency issues currently plaguing speech-dispatcher and pulse are 
> more important, and fixing those latency issues may just fix the crasher as 
> well, although we won't know till latency is sorted.
>
>    
Hmm. I actually don't find the latency all that bad, especially since 
installing the version of ALSA from your PPA. They're certainly still 
there a bit, but I'm used to working on a Mac which has a bunch of 
horrible latency issues all its own, so comparatively these are nothing. :)

My issue, though, is that when SD crashes, I'm usually in an unknown 
state. Is an application taking a while to respond? Did I lose keyboard 
input for some reason? So ultimately I have to kill and restart SD. 
Sometimes this works, sometimes it doesn't. But I think that being in a 
known state is a bit more important than latency from my perspective. 
The latter is an annoyance while the former takes a bit of time to 
diagnose each time. Granted, it's my default assumption when things stop 
responding, but still, that's ten seconds of time lost.

BTW, has Canonical approved your proposal to work on SD? Really hope so, 
that'd be a great and much-needed project, particularly if it can be 
gotten working with Pulse.

SD crashed or locked up twice during the composition of this mail, which 
took five minutes. Scratch that, thrice. Twice gave me the dummy SD 
driver, so I knew it was a crash, and the third instance was one of the 
silent lockups. See my post a month or so back for SD logs of each of 
these scenarios, including errors from Pulse.

Would be really nice to have an issue tracker, and maybe VCS more modern 
than CVS. Lighthouse/Github, anyone? Or Launchpad?