Horton

Re: pg_basebackup's --gzip switch misbehaves

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Daniel Gustafsson <daniel@yesql.se>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers@lists.postgresql.org
Date: 2022-09-16T02:22:46Z
Lists: pgsql-hackers 
On Wed, Sep 14, 2022 at 10:26:42AM +0200, Daniel Gustafsson wrote:
> Maybe the creation of $tempdir should take PG_TEST_NOCLEAN into account and not
> register CLEANUP if set?

Agreed.  It sounds like a good idea to me to extend that to temporary
paths, and then check those rmtree() calls where the tests would not
retain too much data for small-ish machines.

By the way, should we document PG_TEST_TIMEOUT_DEFAULT and
PG_TEST_NOCLEAN not only in src/test/perl/README but also doc/?.  We
provide something in the docs about PROVE_FLAGS and PROVE_TESTS, for
example. 
--
Michael

Commits

  1. Make PG_TEST_NOCLEAN work for temporary directories in TAP tests

  2. doc: document the TAP test environment variables

  3. doc: update metacpan.org links to avoid redirects

  4. Use min/max bounds defined by Zstd for compression level

  5. Fix incorrect value for "strategy" with deflateParams() in walmethods.c

  6. Simplify handling of compression level with compression specifications