[Testtools] Adding new Cbc branch to nightly builds
Stefan Vigerske
stefan at math.hu-berlin.de
Mon Nov 9 05:23:57 EST 2009
Hi,
>> Why is this a clean solution?
>
> Two things, which is why I created the separate project.
>
>
> 1) For PROJECT_EMAIL_ADDRS John Forrest is listed as the project
> manager. Do we want him to get emails if something goes wrong with the
> branch of Cbc? I thought perhaps no.
OK, that might be your point.
On the other hand, since he is the project manager, he should then
forward to whom he thinks is responsible.
I do not send mails out automatically. Maybe I should, but you wouldn't
like to get complains about Couenne caused memory leaks in OS from me
every day ;-).
> 2) If you look in NBProjectConfig.py there is a special feature for the
> unitTest for a regular Cbc run.
You mean the NBcbcRunTimes.cbcSaveRuntimes ? I don't know if someone use
it, but I would argue that it is the task of the cbcSaveRuntimes routine
to decide whether it wants to collect the timings from the reeng branch.
> But sure just doing
>
> { 'SvnVersion': 'branches/reeng', 'OptLevel': 'Default',
> 'Distribute': 'No', 'ThirdParty':'Allowed', 'AdditionalConfigOptions':
> '--enable-parallel'}
>
> will work, but won't John get an email if something goes wrong?
See above.
Stefan
>
> Cheers
>>
>>> We do a similar thing for CoinAll.
>>
>> But that had a different reason. Here the problem is that the CoinAll
>> project is located in the CoinBinary repository.
>>
>>> Then in my
>>> NBuserParameters.py I have right now
>>>
>>>
>>> 'Cbc_reeng' :
>>> [
>>> { 'SvnVersion': 'branches/reeng', 'OptLevel': 'Default',
>>> 'Distribute': 'No',
>>> 'ThirdParty':'Allowed' , 'AdditionalConfigOptions': '--enable-static
>>> --disable-shared'}
>>> ]
>>>
>>
>> I'll just add
>> { 'SvnVersion': 'branches/reeng', 'OptLevel': 'Default',
>> 'Distribute': 'No', 'ThirdParty':'Allowed', 'AdditionalConfigOptions':
>> '--enable-parallel'}
>> to the other testcases in 'Cbc'. Why should that not work?
>>
>> Stefan
>>
>
>
--
Stefan Vigerske
Humboldt University Berlin, Numerical Mathematics
http://www.math.hu-berlin.de/~stefan
More information about the Testtools
mailing list