Isaac Clerencia wrote:
On Sunday 14 November 2004 02:10, Cedric Duval wrote:
Hmm. Then I think we should state the definition of .0 releases.
For me (and until now), it was a rather "stable" release, at least one
on which we concentrated exclusively on polishing and fixing. (like 0.8,
intended to be good enough to be presented on LSM).
If I remember correctly, some time ago we decided that our "stable" and
"development" versions didn't have much sense. In fact, AFAIK, we don't have
any server running the supposed stable version (0.8).
Well, I think the .0 releases were meant to be 'stable', however in
practice they didn't tend to be substantially more stable than any other
release.
I think if you want to call a release 'stable' it means you have to
remain committed to it until you release new stable version. i.e. if a
serious bug is found in the stable release, you have to backport a fix
to the stable release. This basically means you have to branch.
Also, I think that a big factor is that almost all our releases are
stable enough to be called 'stable' considering the version number is <
1.0. Which I think is something we should all be fairly happy about.
It's amazing how many times I tell users about the bug list on Savannah,
and they say "whoah! Wesnoth has bugs?? I haven't noticed any!" ;)
David
