On 2018-05-10, Jeroen Demeyer <j.deme...@ugent.be> wrote:
> On 2018-05-10 17:20, Thierry wrote:
>> Run more patchbots, with various combinations of optional packages
>> included.
>
> -1
>
> That would more the problem more visible at best, but it wouldn't 
> *solve* anything.

-1: When a change of SageMath results in breaking an optional package's tests,
then this is a bug in SageMath and releasing bugs should be avoided. Making a
bug visible and thus preventing it from being released does solve a problem
(for some definition of of "solving" at least).

I guess that most optional packages do not interfere with each other.
Hence, a single patchbot that installs, tests, and builds documentation
of *all* optional packages at once, would probably be very valuable.

Alternatively or in addition, we might return to the old idea that each
package has a maintainer. So, when a new release takes shape (but before
an official release takes place), the release manager may notify the
maintainers (by an automated mail) and ask them to test their package with
the current beta. If one of the maintainers reports that his/her package
got broken, it will be a critical bug or blocker for the SageMath release.

I reckon it wouldn't be much additional work for the release manager, as
the testing is assigned to the package maintainers. It would probably be
doable for the package maintainers, as it is just running one test every
now and then (we currently do not do releases every fortnight, do we?).
And after all it would be in the interest of the package maintainers if
SageMath wouldn't break their optional package.

Best regards,
Simon


-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to