Re: [PATCH] big test separation POC

>> +serial_%: parallel_%
>> + echo "# this file is generated automatically, do not edit!" > $@
>> + egrep '^(test|ignore):' $< | \
>> + while read op list ; do \
>> + for test in $$list ; do \
>> + echo "$$op $$test" ; \
>> + done ; \
>> + done >> $@
>> +
>
> This won't work on windows all that easily.

Hmmm. I made the assumption that a system with "gnu make" would also have
a shell and basic unix commands available, but it is possibly quite naive.

ISTM that there are perl scripts used elsewhere for building postgresql.
Would assuming that perl is available be admissible?

> Maybe we should instead add a "--run-serially" parameter to pg_regress?

I guess that it is quite possible and easy to do. I'm not sure whether it
is desirable.

>> -installcheck: all tablespace-setup
>> - $(pg_regress_installcheck) $(REGRESS_OPTS) --schedule=$(srcdir)/serial_schedule $(EXTRA_TESTS)
>> +# after installation, serial
>> +installcheck: all tablespace-setup serial_schedule
>> + $(pg_regress_installcheck) $(REGRESS_OPTS) \
>> + --schedule=serial_schedule $(EXTRA_TESTS)
>
> Why do we use the serial schedule for installcheck anyway? Just because
> of max_connections?

I asked myself the same question, and found no obvious answer. Maybe if
a check is run against an installed version, it is assumed that postgresql
is being used so the database should not be put under heavy load?

--
Fabien.