[vlc-devel] 0.9.0-test2 status

[Rémi Denis-Courmont]

	Hello,

I have sunk a huge amount of my free time fixing and trying to fix bugs (most 
of them not caused by me, which is very demotivating) over the past 6 weeks. 
My level of motivation is now down to approximately zero. And most if not all 
of the remaining bugs are outside of my competencies.

The 0.9.0-test2 bugs count is huge. There is seemingly no time resources to 
address most of them (no surprise, volunteers are very bad at fixing bugs). 
Some bug fixes don't even get reviewed on the mailing list. Over the months, 
I kept warning that a feature freeze that lasts more than a few weeks, is a 
failure. We have started the feature freeze for almost a month, yet I have 
never seen a "frozen" VLC tree being so far from releasable state, which is 
utmostly worrying to me.


Now for the retrospective since 0.9.0-test1...
There was quite a lot of bug fixing, but even more bug reporting. That was to 
be expected from the first test release after two years of "unstable" 
development. Boringly, the two main offenders are the OSX port and the Qt4 
interface, in this order. I cannot figure out why OSX is consistently the 
release blocker, even though it has much more developpers attention than 
Windows and should have much less portability issues...

I will be building the tarballs for 0.9.0-test2 monday European morning, since 
it should be no worse than 0.9.0-test1 anyway. But I feel very bad about 
0.9.0. I suspect the only solution will involve having a fair amount of known 
regressions, and tell anyone who complains about them to send a patch. As bad 
as it may be in term of publicity, I think that is the only way a strained 
volunteer open-source project can go forward.

Best regards,

-- 
Rémi Denis-Courmont
http://www.remlab.net/