Horton

Re: installcheck-world concurrency issues

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Cc: pgsql-hackers@postgresql.org, Peter Geoghegan <pg@bowt.ie>
Date: 2022-10-05T17:20:27Z
Lists: pgsql-hackers 
Hi,

On 2022-10-05 08:16:37 +0200, Peter Eisentraut wrote:
> On 04.10.22 01:41, Andres Freund wrote:
> > BTW, shouldn't src/test/modules/unsafe_tests use the PG_TEST_EXTRA mechanism
> > somehow?  Seems not great to run it as part of installcheck-world, if we don't
> > want to run it as part of installcheck.
> 
> I think there are different levels and kinds of unsafeness. The ssl and
> kerberos tests start open server processes on your machine.  The
> modules/unsafe_tests just make a mess of your postgres instance.  The latter
> isn't a problem when run against a temp instance.

I agree - but I suspect our definition of danger is reversed. For me breaking
an existing cluster is a lot more likely to incur "real world" danger than
starting a throway instance listening to tcp on localhost...

Greetings,

Andres Freund

Commits

  1. tests: Rename conflicting role names

  2. tests: Restrict pg_locks queries in advisory_locks.sql to current database