[Wine] Re: Screwy guide on the regression testing wiki page

[André H.]

Bruni wrote:
> I'm confused by this:
> 
> > 
> > ... then you need to reset your bisect and use that "Release x.x.xx" as your first good bisect (wine-x.x.xx). This particular bisect output is just a version tag, meaning that no real code was changed (and so cannot have made your problem appear). All the "good" bisects you've set tells GIT that all that code is good, until wine-x.x.xx . GIT eliminates all that code except for the version tag, which would be the latest revision before the next version tag.
> > 
> 
> What does "meaning that no real code was changed (and so cannot have made your problem appear)" phrase mean?


If you have a look at the change of Release 1.3.27 then you see that this can't introduce a regression as it only changes text, no code:
http://source.winehq.org/git/wine.git/commitdiff/6a6aab27633405aeb5e464943110960f0099dffe


> Same about
> > All the "good" bisects you've set tells GIT that all that code is good, until wine-x.x.xx
> 
> 
> Also, I am just curious about why ccache is not specified as being used when doing regression testing at intermediate steps:
> 
> > This will reposition the source code in the middle of the remaining patches. You will then recompile Wine by running 
> > 
> > Code:
> > ./configure --verbose && make
> > 
> > 
> > and test for the bug again.
> 


Be free to fix this, further i also would add "-C" for configure as that caches the checks so configure runs much faster.