Re: parallel.c is not marked as test covered

On 9 May 2016 at 14:26, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
> David Rowley wrote:
>> On 9 May 2016 at 13:20, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
>
>> > It's not a buildfarm machine, but a machine setup specifically for
>> > coverage reports.  It runs "make check-world" only.  I can add some
>> > additional command(s) to run, if somebody can suggest something.
>>
>> I'm not familiar with what's possible with make check-world, but would
>> it be reasonable to make that just do another regression test run with
>> force_parallel_mode set to regress?
>
> My suggestion isn't to change what "make check-world" does; it's just to
> change the script in the coverage machine so that it runs some other
> command after "make check-world" and before "make coverage-html".

I understand what you meant. I was just taking the suggestion one step
further as I thought that if you were adding that for the coverage
test then it might actually also be a good idea that it occurred in
make check-world. The parallel tests are quite thin as it is, so
perhaps its a good idea to get more machines running through the
parallel code.

pg_regress seems to support --temp-config, so we could just have a
config file which has; max_parallel_degree = 2 and force_parallel_mode
= regress.

-- David Rowley                   http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services