[MPlayer-users] Next rc or 1.0 anyone
Nan Null
hovh03 at yahoo.com
Tue Jul 1 05:06:22 CEST 2008
----- Original Message ----
> From: Raimund Berger <raimund.berger at gmail.com>
> To: mplayer-users at mplayerhq.hu
> Sent: Monday, June 30, 2008 7:46:31 PM
> Subject: Re: [MPlayer-users] Next rc or 1.0 anyone
>
> Don't know if you guys are aware of it, but what you're really
> fighting about (resp. against) here is svn/cvs, and the hassle it
> involves to maintain a release and development branch, including
> carrying patches over amongst them. That's why some of you guys so
> desperately try to keep all on one branch, although it involves
> various kinds of compromises.
>
> I'm not here to advocate anything, but that's really one of the
> stinkers git stepped up to solve. And regarding scm and release
> practices, linux kernel development clearly became a role model thanks
> to it.
>
> So while I don't propose anything here, I do recommend keeping in mind
> that at the bottom of the disagreement just lies the currently used
> toolset's lacking support of proper practices. Remembering that might
> help taking a little heat out of the discussion.
No, that's not a problem. It takes a lot more to have a release. One of them being understand how important it is.
Saying there should be no expectation of a release, of quality, of when it's done, it's done when it's done, etc. is just reckless in software engineering.
svn/cvs may not be the best tool, but I have seen them used and have used them in big projects without problem. It takes a lot more to do a good release than just the SCM thing. It's planning, coordination between different developers, organizing features, evangelizing developers and helpers to put out documents, and do marketing (if they feel their product and mind share of developers are important, after all users become developers in open source are the norm).
The biggest problem is not seeing a value in a good coordinated release. There's a science to release a piece of engineer. It's no secret. One way maybe better than another, but not doing it at all is just reckless.
More information about the MPlayer-users
mailing list