ubuntu-audio-dev team mailing list archive

[David Henningsson <david.henningsson@xxxxxxxxxxxxx>]

On 09/23/2011 04:25 AM, Luke Yelavich wrote:
> Hey folks,
> So Beta 2 is released, and we have some bugfixes from David in the queue for pulse. I was going to upload them today, however the upload queue is still ocked and being reviewed, so I think it worth while waiting a few days, in case we have more fixes that need to go in.
>
> In addition, if there are any other fixes from git master that people think we need, please speak up now. I don't mind doing the legwork to get them into the packaging, but I hope to upload one more revision of pulseaudio no later than Wednesday afternoon Sydney time next week, so we have until then to get final fixes together. Of course we will be able to SRU stuff, but the criteria for that are more strict than the small time we have now for getting other general fixes in.
>
> Luke

Thanks for the heads up. I went to ask Colin and Arun on IRC (qouting 
conversation below), and unsurprising, they believed in the recent 
commits and claimed we should take 0.99.4 + the two patches from git 
master on top. My gut feeling says that it's probably okay to do so, but 
yeah...there's always regression risk I guess. OTOH, should we end up 
with 10-20 patches on top of 0.99.3, it feels like we could just as well 
take 0.99.4 anyway.

I could read the patches through, of course, and take 0.99.4 if I can't 
find anything troublesome. But there are quite a few of those which are 
about PulseAudio functionality/code I don't know much about, so I don't 
know how much that would help.

What do you think?

<diwic> Ford_Prefect, coling, hi there :-)
<diwic> I just got a heads up from Luke that it's time to prepare the 
final PulseAudio version for Ubuntu 11.10.
<diwic> Right now we're basically shipping 0.99.3 with just my patches 
on top. Is there anything between 0.99.3 and 0.99.4 that strikes you as 
more important to actually get in?
<coling> diwic, Looking at the shortlog I'd say you probably want most 
of the patches in 0.99.4... some are build related (i.e. prevent 
overlinking which will be a pain for version updates), some are 
migration from old version fixes (i.e. stream-restore data would have 
been mostly lost).
<coling> diwic, Oh and you *really* want the rename of sync_volume stuff 
otherwise you could end up wtih broken config files.
<diwic> coling, hmm, I was afraid so.
<coling> diwic, I'd just have a look through the shortlog and decide 
yourself. Anything that relates to public API (like the sync volume 
stuff) and build fixes and migration fixes you likely want.
<coling> If I were you I'd basically update to .4 and pull in the final 
patch from Arun that is in master (as well as your one obviously) that's 
been submitted since 0.99.4.
<coling> But I know the general trepidation you must feel!
<coling> Most of the patches are quite small tho', so you can probably 
have a read over most of them and satisfy yourself they are "safe".
<coling> (and if they are not it would be nice to get a code review 
anyway :p
<coling> )
<diwic> A few of them are related to shared volume stuff, dunno how 
important / regression likely that stuff is
<diwic> and there is this modargs triple-quote stuff
<Ford_Prefect> diwic: by diwic I'd say you want everything
<Ford_Prefect> Because we've been pretty particular about what we're 
merging and what we're not
<Ford_Prefect> diwic: btw, do you know what version of Empathy is shipping?
<diwic> Ford_Prefect, empathy --version => 3.1.92
<Ford_Prefect> diwic: you want the shared volumes stuff then, so that 
the echo cancellation behaves properly :)
<Ford_Prefect> And you want the crasher fix that comes after that as well
<Ford_Prefect> Well, volume behaviour with echo cancellation, to be precise
<diwic> hmm
<diwic> well, at least thanks for your opinions. But there is always 
regression risks as well
<Ford_Prefect> Fair enough :)
<coling> On balance, if I were in your position, I'd have a good read 
through the patches then include them all (assuming all is well).



--
David Henningsson, Canonical Ltd.
http://launchpad.net/~diwic