[vlc-devel] Resignation as co-release manager

[Rémi Denis-Courmont]

Le samedi 9 août 2008 12:25:41 Felix Paul Kühne, vous avez écrit :
> Especially the win32 port needs final tests IMO, as it
> lacks the test3-release.

The reason why the Win32 port sucks is dead simply. No (active) developpers is 
willing and able to take care of it. The same thing goes for the Mozilla 
plugin on OSX, or for Skins2. Delaying a release because a developper needs 
some extra time to fix a problem is good. Delaying a release while there is 
no developper is plain dumb wishful thinking.

Unmaintained stuff is by definition not maintained. As the VLC code base 
grows, and as time passes, it is only natural that there be more and more 
unmaintained stuff. And unmaintained stuff tends to break and not get fixed.

Welcome to the real world. The *only* alternatives are:
- removing unmaintained broken things, or
- shipping them inspite of known bugs, however severe.
Waiting for unmaintained components to be fixed will achieve nothing but 
ensure that we NEVER make a release EVER again... and further discourage 
people, increase the ammount of unmaintained stuff, and so on.

Over the past year everybody was complaining about our slow release process, 
and how badly we needed to switch from feature-based releases to time-based 
releases. Now that the implications are apparent, everybody wants to go back 
to feature-based. Guys, we cannot have it both ways.


With this said, given that everybody disagrees with me, I have to give up on 
0.9.0 release management.

N.B.: It was great wasting so much of my time for this non-release over the 
last two months, thank you guys.

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