[Libreoffice-qa] Addressing regressions within a release

[Bjoern Michaelsen]

On Wed, Apr 24, 2013 at 10:45:36AM +0200, Bjoern Michaelsen wrote:
> On Tue, Apr 23, 2013 at 10:21:31PM -0700, Joel Madero wrote:
> > Apologies for my lack of technical skills, but...what does this mean
> > exactly? In terms of QA being able to use it, is it ready? Almost
> > ready? 
> 
> The stuff that I pushed would allow you to build the bibisect repo if you would
> understand how it is intended to be used -- which is the trouble, as it is
> severelt underdocumented right now.
> 
> > And for the bibisect repo, that would be up to most recent
> > commit?
> 
> The one I have sitting on my local disc? The last version it contains is
> 4.0.0.0.beta1. It does not contain any master commits -- so e.g. you can see
> were it broke between libreoffice-4-0-branch-point and the libreoffice-3.6.4.3.
> But it has all the microreleases e.g. 3.6.2.1, 3.6.2.2, 3.6.3.1 ... so you can
> see if and where is broke _in_ one major release. So its a bibisect repo
> without any of the intermediate steps on master, but with all released versions
> (including rcs and alpha/betas).
> 
> Im packing what I have right now and will try to upload it today, I guess just
> playing around with it might make it easier to see what I mean. ;)

I uploaded the tags repo as:

 http://people.canonical.com/~bjoern/bibisect-4.0-tags1.tar.xz
 md5sum: e9e3119777cce4020b310dd99fef0458

its a bit bigger than usual as I did not repack it and it ends with 3.6 at
3.6.4 and 4.0 at beta1. Still, if you want to play with it, go ahead! ;)

Best,

Bjoern