From: Tom Lane <tgl@sss.pgh.pa.us>
To: Kevin Burke <kevin@meter.com>
cc: pgsql-admin@lists.postgresql.org
Subject: Re: Getting started notes and trouble running the tests
In-reply-to: 
 <CAAJ6v9CymXJMdgKaRx7=rKvCdhGbQhSStvfhEFayomr5cbwyZQ@mail.gmail.com>
References: 
 <CAAJ6v9CymXJMdgKaRx7=rKvCdhGbQhSStvfhEFayomr5cbwyZQ@mail.gmail.com>
Comments: In-reply-to Kevin Burke <kevin@meter.com>
	message dated "Sat, 13 Feb 2021 15:52:05 -0800"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1878103.1613263014.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Sat, 13 Feb 2021 19:36:54 -0500
Message-ID: <1878104.1613263014@sss.pgh.pa.us>
Precedence: bulk

Kevin Burke <kevin@meter.com> writes:
> On Mac, I get this error when I run "./configure" and then "gmake check"=
:

> [Output omitted]
> sh: line 1: 51711 Abort trap: 6           "psql" -X postgres < /dev/null=
 2>
> /dev/null

Yeah, this is a known issue with Apple's SIP (System Integrity Protection)
breaking our use of DYLD_LIBRARY_PATH to point the executables at the
as-yet-uninstalled libraries.  The two known workarounds are
(1) do "make install" before "make check";
(2) turn off SIP.
We've made occasional efforts to develop a less fragile solution.
If you're interested in helping with the latest such effort, see

https://www.postgresql.org/message-id/flat/ecec88aa-c982-a473-1b0f-cea596d=
405cf@enterprisedb.com

> Is there a way to run just the tests for a particular part of the code? =
The
> docs on running the tests weren't clear, and starting "gmake check" seem=
s
> slow in the best case. In my case I'd like to run just the tests related=
 to
> OpenSSL, and src/backend/libpq/be-secure.c.

The core regression tests aren't subdivided, but on the other hand
they only take a few seconds on any modern hardware (hint: use the
parallelized make targets).  There are a lot of other test suites
scattered around the tree, and those can be run individually.
In your case I'd venture that "make check" in src/test/ssl/ might
cover most of what you care about (but see the README there for
caveats --- it transiently opens ports).  Also, for repeated test runs,
using "make install" and then "make installcheck" can save time.

			regards, tom lane