Patchwatcher online

Zachary Goldberg zgold550 at gmail.com
Mon Aug 11 20:12:21 CDT 2008


On Mon, Aug 11, 2008 at 8:34 PM, Dan Kegel <dank at kegel.com> wrote:
> Vijay Kiran Kamuju <infyquest at gmail.com> wrote:
>> Ok I was expressing my concern as it took around 2-3hrs to see my
>> patch in the patchwatcher.
>
> It's running on a 1GHz single core machine right now.
> I'll probably put it on something rather faster.
>
>> Also as you you running the wine tests all for each patch are you
>> cleaning the .wine directory ( I am bit confused here)
>
> No.  Probably should, but I'm not.
>
>> It would better if we have a parallelized version of the tests also
>> run on a fast m/c.
>
> I do have a patch that enables parallel execution of conformance
> tests, I hope Alexandre accepts it.   That will help on multicore
> systems.  Beyond that, I could fairly easily use multiple machines,
> e.g. assign all patches to machines based on md5sum.
>
>> Also can you improve the messages.
>
> Yes.  I already changed the success message to make more sense,
> and added background colors of green and red for success and failure.
>
>> If there are errors, Its possible to only show the test data that
>> failed rather than the complete test run.
>
> Yes.
>
>> Also put it in a public repository with you as sole commiter.
>
> Already there, see http://code.google.com/p/winezeug/
>
>> So If we have any suggestions/improvements, can mail you with the
>> changes (We will not flood ur mail box ;) )
>
> Please do.
> - Dan
>
>
>


Dan, how are you handling the case when Alexandre floods the list with
commits?  One way I can think of to avoid those patches is to ignore
any patches emailed by Alexandre but aren't written by him.  Its not
the worst thing in the world if the script doesn't skip commits, just
some wasted perf.



More information about the wine-devel mailing list