[petsc-dev] broken nightlybuilds (next vs next-tmp)

[Jed Brown]

Satish Balay <balay at mcs.anl.gov> writes:

> On Sat, 11 Nov 2017, Jed Brown wrote:
>
>> 
>> Merging is synchronous.  If I do
>> 
>>   git checkout master
>>   git pull
>>   git merge jed/risky-business
>>   make alltests  # works on my machine
>
> Note:  this is my recommendation for the *currentK next model.
>
> If you are running 'make alltests' on your laptop - then you don't need to test on es - before merge to maint.

alltests takes hours and doesn't catch weird configurations -- you need
different PETSC_ARCH for that.  It is normal to at least compile and run
a couple local tests.

> My suggestion on using es is for Matt - who says he can't run tests on his laptop.
>
>> Satish and Barry, is this really the workflow you're advocating?
>
> What Barry is advocating is throwing away next. Perhaps he'll explain
> what his thoughts are..  [I've expressed my interpration]
>
> In the new model [since most folks are not doing any local testing] -
> the above merge and (multiple) tests would be done with some automated
> test suite we would cook up - similar to what we are doing currently
> with next
>
> Note: If it can't do the merge - it will throw back the isssue back to
> the author.  If the tests are successful - it [or one of us] would do
> the merge to master.

As I propose in my other email, let's do the basic automation before
going to 'next'.  Synchronous tests that take hours are just not
workable for developers and will be very resource-hungry when we have
people pushing to a dozen open pull requests.