next release and release pace in general

[Thomas Pfeiffer]

On Saturday 26 January 2013 09:42:57 Pierre Stirnweiss wrote:
> On Fri, Jan 25, 2013 at 6:16 PM, Yuë Liu <yue.liu at mail.com> wrote:
> > We talked about having a testing branch before commit into master
> > branch last year. We can restrict commit right to master, let the
> > features committed to and tested in testing branch thoroughly. So
> > master will always keep stable.
> 
> I don't think we have to be as extreme as restricting commit rights. We
> have become pretty good at putting up review requests, IMHO. I think we are
> at a fine balance of not putting up every little thing, but also getting
> review on potentially harmful changes.
> I am all in favor of a testing branch is however how do we ensure that it
> gets sufficient testing.

Plasma Active has switched to an "always stable Master, plus not-necessarily-
stable Integration branch" model in our current development cycle ( 
http://aseigo.blogspot.com/2013/01/plasma-active-stable-devel-and-
bleeding.html ) for the same reasons it is suggested here.

We suggest to people who want to have new features fast but still need 
stability to use the Master branch/packages, whereas the development team 
itself and some users who volunteered for testing use the Integration 
branch/packages.

Plasma Active currently isn't exactly "stable" even in development images 
using Master, but that is mainly because these images use Master form other 
KDE projects and those are currently causing most of the trouble, which isn't 
an issue for Calligra (Calligra users don't have to run Master of all KDE 
software for Calligra Master to work). Our Master/Integration tandem as such 
seems to work pretty well so far, so I can recommend that.