Re: About sodipodi

Lauris Kaplinski <[email protected]> Mon, 01 Mar 2004 18:58:25 +0200
Newsgroups gmane.comp.graphics.sodipodi
Message-ID <[email protected]>
Hello!

Frans Englich kirjutas E, 01.03.2004 kell 14:57:
> On Sunday 29 February 2004 23:18, Lauris Kaplinski wrote:
> > I think usually the one creating daily builds would be the
> > one detecting most breakage ;-) The vast majority of problems
> > are minor header incompatibilites (strcasecmp/stricmp etc.)
> > that will be detected compile-time.
> 
> I think there are clear advantages with doing frequent releases -  a lot of 
> major projects do it. I think it will speed up development, get more 
> contributors etc. ESR describes the dangers of "mages hacking on their 
> cathedral".
> One possibility is to have a stable "tree"/version and then having daily 
> builds which people can check out if they like.

Now, actually... there is no stable sodipodi versions at all ;-)
Everything is still 0.x, so there shouldn't be any reason to
not to RERO - but as I said, even with the current state of
things, there is the problem of distributions/propagating
binaries. And then there is CVS, public to all and really easy
to set up IMHO.

> But as usual, someone has to rig it up..

What qualifies as 'release'?
I can see the reason of releasing Win32 binaries, because most
potential users are not able to rebuild them. Under Linux/Unix
binaries are not of much importance, if you are living on the
bleeding edge. But is there much reason in packaging CVS tarballs
other than internal testing (that it builds), or should we
teach and encourage people to use CVS more?

Best wishes,
Lauris Kaplinski





-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click