Horton

Re: [HACKERS] Regression tests vs existing users in an installation

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
Cc: Stephen Frost <sfrost@snowman.net>, Robert Haas <robertmhaas@gmail.com>, Alvaro Herrera <alvherre@2ndquadrant.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2019-11-26T22:32:16Z
Lists: pgsql-hackers 
Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes:
> I found this directory, and it seems like a good place to add some more 
> tests (other thread), but I'm puzzled why it's hidden under modules/, 
> since it's not, well, a module.  Why is it not in src/test/?

Doing it like this allowed it to be picked up automatically by the
buildfarm script.  If we'd put it in a new src/test directory, the
buildfarm script would have had to be taught about that, and we'd
have no coverage until owners updated their machines.  So it
seemed expedient to do it like this.

			regards, tom lane

Commits

  1. Move rolenames test out of the core regression tests.

  2. Add an enforcement mechanism for global object names in regression tests.

  3. Fix regression tests to use only global names beginning with "regress_".

  4. Disallow user-created replication origins named "pg_xxx".