Regressions || Huston we have a problem

[hanna.skott at kogmbh.com]

On Sat, 16 Jul 2011 07:18:30 +0200, Cyrille Berger Skott
<cberger at cberger.net> wrote:
> On Saturday 16 July 2011, Jos van den Oever wrote:
>> On Friday, July 15, 2011 13:52:26 PM Cyrille Berger Skott wrote:
>> > On Friday 15 July 2011, Boudewijn Rempt wrote:
>> > >
>> > > And those as well. Jos' server checks whether we can actually load
> all
>> > > the kofficetests document, and even there we have failures.
>> >
>> > jos's server does check whether it saves (and validation too I think
> ?).
>> > I am trying to setup that on cruncher as well.
>>
>> Yes, please reuse/improve the script here:
>>  calligra/tools/scripts/profileOfficeFileLoading.py
>> It loads, saves and validates taking the koffice namespace extensions
> into
>> account with an rng that is extended:
>>  calligra/tools/scripts/OpenDocument-v1.2-cd05-schema-calligra.rng
> 
> Well at this point, I am using an other scripts, but now the biggest
> challenge
> I am facing is that it takes one hour to run the full test suite. (might
> be
> due to a lot of filters not respecting the --batch option...)

The script Cyrille is using on cruncher is Jos' script with a few changes
to it that I made. A lot of the failures reported from that script, when
related to round-tripping, are indeed due to the --batch option not being
respected by a lot of filters - it is a well known problem.  

//Hanna Skott