Anne Schilling <[email protected]> writes:

> Here are several *ideas*:
>
> * We should run daily tests on the "needs review" section and pop patches
>   off that section if the tests do not pass (or people are not actively 
> working
>   on making them pass).
>
> * Everyone should try to get their patches into main-sage in a reasonable
>   time frame or remove their patches from the queue if they are outdated.
>
> * Perhaps we should have a second queue for only the patches currently
>   actively developed and in the review or final development process.
>   This queue would only be maintained on the very last version of sage
>   and patches are removed once the next version of sage comes out.
>   This would take the extra constraints off developers to, at the same time,
>   make their patch apply cleanly on a given version of sage and juggle
>   the tail of sage-combinat patches that lie on top of it.
>   This caused quite some frustration to both Franco and Mike and myself last
>   week trying to work on some large patches.

Although I am now relieved that I didn't break anything, let me add one
other idea: in the unittesting package for fricas, two things turned out
to be very useful:

1) being able to mark tests as "expected to fail"

2) a statistic 

   how many tests were run
   how many failed, but should have passed
   how many passed, but were expected to have failed.

with these two, I think one could actually require that no unexpected
failures happen at any time -- the only problem being that fixing
something somewhere might break something elsewhere, but that problem
exists now, too.

I do not know how much energy one should spend on that, however.

Best,

Martin

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Reply via email to