Re: pgsql: Provide a TLS init hook

On 3/26/20 9:50 AM, Tom Lane wrote:
> Andrew Dunstan <andrew(dot)dunstan(at)2ndquadrant(dot)com> writes:
>> On 3/25/20 9:28 PM, Tom Lane wrote:
>>> jacana has just exposed a different problem: it's not configured
>>> --with-openssl, but the buildfarm script is trying to run this
>>> new test module anyway. I'm confused about the reason.
>>> "make installcheck" in src/test/modules does the right thing,
>>> but seemingly that client is doing something different?
>> Ugh. I have put in place a hack to clear the error on jacana. Yes, the
>> client does something different so we can run each module separately.
>> Trawling through the output and files for one test on its own is hard
>> enough, I don't want to aggregate them.
> Well, I'm confused, because my own critters are running this as part
> of a single make-installcheck-in-src/test/modules step, eg
>
> https://buildfarm.postgresql.org/cgi-bin/show_stage_log.pl?nm=longfin&dt=2020-03-26%2002%3A09%3A08&stg=testmodules-install-check-C
>
> Why is jacana doing it differently?

longfin is also running it (first) here
https://buildfarm.postgresql.org/cgi-bin/show_stage_log.pl?nm=longfin&dt=2020-03-26%2014%3A39%3A51&stg=ssl_passphrase_callback-check

That's where jacana failed.

I don't think this belongs in installcheck, we should add
'NO_INSTALLCHECK = 1' to the Makefile.

cheers

andrew

--
Andrew Dunstan https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services