[RFC] Platform champions

[Nicholas Allen]

> In principle, yes. In practice, things always have a chance to break on
> everything *except* whatever PQM is running on. I don't think a process
> where PQM must pass all tests on 10+ platforms each pqm-submit is
> sensible either. Heh, just running all tests on both Python 2.4 and
> Python 2.5 would be nice :-) , let alone doing multiple Python's on
> multiple platforms.
>   

Just an outsider making a suggestion here.

I would have thought it would be possible to have PQM only running on 
one platform as it does now but to have other machines and platforms 
perform the automated test suit as often as possible (most likely not 
every merge request but as soon as they finish they pull again and check 
the next set of changes they got from the pull). When it finds a problem 
it could inform the dev team automatically. So the merge request has to 
pass the current PQM test suite which gives it a pretty good chance of 
passing on other platforms too. I don't think it would even be necessary 
to purchase the hardware if virtual machines were used.

So PQM becomes a kind of sane gateway that reduces the likelihood of 
problems on other platforms and another machine (or machines) check the 
trunk test suit as well. I know it's a simple idea but wouldn't 
automating the process like this help and massively increase the chance 
that a given release works on all platforms?

Nick