[Audacity-devel] Releasable state for 2.0.5 WAS Re: Libraries

[Gale A. <ga...@au...>]

| From Richard Ash <ri...@au...> 
| Fri, 27 Sep 2013 21:40:52 +0100
| Subject: [Audacity-devel] Libraries
> On Thu, 26 Sep 2013 20:56:40 -0700
> Vaughan Johnson <va...@au...> wrote:
> 
> > PS: But related, and higher priority, is updating the FFmpeg
> > interface, as we have a P2 bug about it. Top of my dev list, but I
> > have other admin things to do first. 
>
> > And anybody else who wants to do any of this, please let us know. I'm 
> > not territorial about it!
> 
> Before we dive into any of this, are we wanting to get a quick 2.0.5
> release out to test the fixes we have made for 2.0.4 crashes? This
> would argue against messing with the library stack in the short term
> (there is also a libsndfile release promised soon, which will include
> our patches, so might be worth waiting that one on). This really
> depends how quickly we think a solution to the WDM-KS problem might be
> available, as we seem to have a fix for the Mac issue.

I'm guessing a new topic is perhaps better for this content. 

At the moment I feel we aren't very close to a releasable state. 

AFAIK we have no fix for the Windows WDM-KS BSOD's (which 
comprise about one-third of the reported cases). 

The only tested patch that fixes the hangs (Leland's) kills MME,
so the next step is to try Lasse's solution out.    

Although the Mac import crash looks like it is fixed, that problem
was covering up another problem introduced somewhere in 
2.0.4-alpha that you can't double-click an AUP, use Finder "Open
with" or drag a file to the Audacity icon and have the file open  -
nothing happens.

I've started a bug (initially at P1) for this:
http://bugzilla.audacityteam.org/show_bug.cgi?id=655 .

This seems to be repeatable. 40 people or so have already complained
about it and are mostly opting to go back to 2.0.3. 

I can't see that we could easily release 2.0.5 with that issue after 
the 2.0.4 problems. I know you can use the Audacity menus, but
the clear user reaction to that has been "no". I think this could only
be P2 if the fix is difficult and time-consuming, meaning it's better
to get the crash fix out.  



Gale