[Wine] Regression testing - compilation error after second bisect

[Austin English]

On Mon, Apr 20, 2009 at 1:27 AM, James Huk <huk256 at gmail.com> wrote:
> 2009/4/18 vitamin <wineforum-user at winehq.org>
>
>>
>> James Huk wrote:
>> > Can I somehow- using git - fetch all patches from date x, for custom dll?
>>
>> That's exactly what git does. You literally cloning the entire repository
>> all patches since day 1.
>>
>>
>> James Huk wrote:
>> > I mean can I get all patches from wine-0.9.15 onward for dsound.dll? If
>> so, will I be able to apply them for wine-0.9.15 source?
>>
>> No. There were lots of wholesale changes in Wine source tree. You can't
>> backport big changes to each individual dll, especially if it relies on lots
>> of other things.
>>
>>
>>
>>
>>
>>
> I see. I tried again with:
>
> git bisect start
> git bisect good wine-0.9.15
> git bisect bad
>
> and... well:
>
> 1st compilation successful - bug present
> 2nd compilation successful - bug present
> 3th compilation successful - well now even winecfg doesn't want to start –
> some kind of “can't create window” error (sorry I didn't write it down). Now
> I tried again with git bisect bad and recompilation – but wine was never
> able to create window. Few compilations later it failed to compile at all.
>
> My questions are:
>
> 1.Is there a way to compile and test wine using above git bisect?

http://wiki.winehq.org/ReverseRegressionTesting

> 2.If I found regression between 0.9.15 and 0.9.16 – should I report it at
> all? This is after all very old wine version...

Yes. The bug can't be fixed if it isn't known.

-- 
-Austin