[petsc-dev] plans for PETSc release
Barry Smith
bsmith at mcs.anl.gov
Sun Apr 24 16:18:55 CDT 2016
> On Apr 24, 2016, at 4:11 PM, Jed Brown <jed at jedbrown.org> wrote:
>
> Barry Smith <bsmith at mcs.anl.gov> writes:
>> There is still no excuse for not during regular testing
>> there. Maybe not daily but at least once a week and automating it
>> as much as possible. For example I can envision each Wednesday
>> morning at 9 am Satish receiving an email that says: Log into
>> mira.alcf.anl.gov and run the script testmira to start the test
>> build. Then that script runs all the compiles and ships off to the
>> dashboard all results without requiring further actions by a
>> human. Each Thursday Jason gets a similar email for hopper etc. It
>> is not perfect but it is better than someone manually doing tests
>> when they are inspired. The timing of these pseudo automated
>> testing could be increased around releases.
>
> That would improve reliability, but it doesn't exist yet and given an
> open dev community assumption, will never cover all such machines. For
> example, we've gotten substantial contributions from Europe during this
> release cycle. Those developers have packages that depend on PETSc, but
> may not have updated in the final days before an (vaguely announced)
> release. Not all of them will test if we announce a feature freeze, but
> some probably will. After all, it means a lot to their users to have an
> easy install and be functional with the latest PETSc.
After our release the developers of each of those packages needs to (possibly update and then) test against against the PETSc release and report problems with the PETSc release. We fix those problems in a patch release and then they can make their package release. Is it really crucial that all the packages that depend on PETSc have to make their releases exactly when PETSc makes a release? Probably they are in the middle of their implementation cycle and are not ready to make a new release anyways. I don't think we should even consider having the goal that other packages that use PETSc have the same date of release as PETSc. Yes we want them to be within a couple of months at most, but I don't see setting a goal of the exact same day.
More information about the petsc-dev
mailing list