Horton

Re: Small issues with CREATE TABLE COMPRESSION

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Dilip Kumar <dilipbalaut@gmail.com>, Postgres hackers <pgsql-hackers@lists.postgresql.org>, Jacob Champion <pchampion@vmware.com>
Date: 2021-05-06T00:05:03Z
Lists: pgsql-hackers 
On Wed, May 05, 2021 at 09:59:41AM -0400, Robert Haas wrote:
> TAP tests have a facility for conditionally skipping tests; see
> perldoc Test::More. That's actually superior to what you can do with
> pg_regress. We'd need to come up with some logic to determine when to
> skip or not, though. Perhaps the easiest method would be to have the
> relevant Perl script try to create a table with an lz4 column. If that
> works, then perform the LZ4-based tests. If it fails, check the error
> message. If it says anything that LZ4 is not supported by this build,
> skip those tests. If it says anything else, die.

There is a simpler and cheaper method to make the execution of TAP
test conditional.  As in src/test/ssl/t/002_scram.pl for channel
binding, I think that you could use something like
check_pg_config("#define HAVE_LIBLZ4 1") and use its result to decide
which tests to skip or not.
--
Michael

Commits

  1. Add support for LZ4 build in MSVC scripts

  2. Add more TAP tests for pg_dump with attribute compression

  3. doc: Fix some gaps with the documentation related to LZ4

  4. Fix incorrect error code for CREATE/ALTER TABLE COMPRESSION

  5. Additional doc fixes for configurable TOAST compression.

  6. docs: Clarify how ALTER TABLE .. SET COMPRESSION works.