Re: Adding CI to our tree

Hi,

On 2022-01-13 13:06:42 -0600, Justin Pryzby wrote:
> Question: are data-checksums tested at all ? The only thing I can find is that
> some buildfarm members *might* exercise it during installcheck.

There's some coverage via src/bin/pg_basebackup/t/010_pg_basebackup.pl and
src/bin/pg_checksums/t/002_actions.pl - but that's not a whole lot.

Might be worth converting one of the "additional" pg_regress runs to use
data-checksums? E.g. pg_upgrade's, or the one being introduced in the "replay"
test?
https://postgr.es/m/CA%2BhUKGK-%2Bmg6RWiDu0JudF6jWeL5%2BgPmi8EKUm1eAzmdbwiE_A%40mail.gmail.com

> From b67cd05895c8fa42e13e6743db36412a68292607 Mon Sep 17 00:00:00 2001
> From: Justin Pryzby <pryzbyj(at)telsasoft(dot)com>
> Date: Sun, 9 Jan 2022 22:54:32 -0600
> Subject: [PATCH 2/7] CI: run initdb with --no-sync for windows

Applied this already.

> From 885becd19f630a69ab1de44cefcdda21ca8146d6 Mon Sep 17 00:00:00 2001
> From: Justin Pryzby <pryzbyj(at)telsasoft(dot)com>
> Date: Tue, 11 Jan 2022 01:30:37 -0600
> Subject: [PATCH 4/7] cirrus/linux: script test.log..
>
> For consistency, and because otherwise errors can be lost (?) or hard to find.

> - make -s ${CHECK} ${CHECKFLAGS} -j${TEST_JOBS}
> + script --command "make -s ${CHECK} ${CHECKFLAGS} -j${TEST_JOBS}" test.log
> EOF

I'm not following this one? all the output is in the CI run already, you can
download it already as well?

The only reason to use script as a wrapper is that otherwise make on
freebsd/macos warns about fcntl failures?

> only_if: $CIRRUS_CHANGE_MESSAGE !=~ '.*\nci-os-only:.*' || $CIRRUS_CHANGE_MESSAGE =~ '.*\nci-os-only:[^\n]*linux.*'
> @@ -183,7 +178,7 @@ task:
> mkdir -p ${CCACHE_DIR}
> chown -R postgres:postgres ${CCACHE_DIR}
> echo '* - memlock 134217728' > /etc/security/limits.d/postgres.conf
> - su postgres -c "ulimit -l -H && ulimit -l -S"
> + su postgres -c "ulimit -l -H && ulimit -l -S" # XXX

Hm?

Greetings,

Andres Freund