Re: Preventing abort() and exit() calls in libpq

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Noah Misch <noah(at)leadboat(dot)com>
Cc: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, Michael Paquier <michael(at)paquier(dot)xyz>, Jacob Champion <pchampion(at)vmware(dot)com>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>, Daniel Gustafsson <daniel(at)yesql(dot)se>, info(at)cspug(dot)cz
Subject: Re: Preventing abort() and exit() calls in libpq
Date: 2021-07-03 14:45:59
Message-ID: 1233492.1625323559@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I wrote:
> I'm now wondering about applying the test to *.o in libpq,
> as well as libpgport_shlib.a and libpgcommon_shlib.a.
> The latter would require some code changes, and it would make
> the prohibition extend further than libpq alone. On the bright
> side, we could reinstate the check for abort().

After consuming a bit more caffeine, I'm afraid that won't work.
I'd imagined leaving, e.g., psprintf.c out of libpgcommon_shlib.a.
But if someone mistakenly introduced a psprintf call into libpq,
it'd still compile just fine; the symbol would be resolved against
psprintf in the calling application's code. We'd only detect a
failure when trying to use libpq with an app that didn't contain
that function, which feels like something that our own testing
could miss.

What I'm now thinking about is restricting the test to only be run on
platforms where use of foo.a libraries is deprecated, so that we can
be pretty sure that we won't hit this situation. Even if we only
run the test on Linux, that'd be plenty to catch any mistakes.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2021-07-03 15:00:42 Re: OpenSSL 3.0.0 compatibility
Previous Message Tom Lane 2021-07-03 14:14:38 Re: rand48 replacement