On Tue, Nov 15, 2022 at 04:29:08PM -0800, Jeff Davis wrote:
> Committed with some significant revisions (ae168c794f):
>
> * changed to insert a deterministic message, rather than a random
> one, which allows more complete testing
> * fixed a couple bugs
> * used a static initializer for the RmgrData rather than memset,
> which shows a better example
>
> I also separately committed a patch to mark the argument of
> RegisterCustomRmgr as "const".
This is causing the CI job to fail for 32-bit builds. Here is one
example in my own repository for what looks like an alignment issue:
https://github.com/michaelpq/postgres/runs/9514121172
[01:17:23.152] ok 1 - custom WAL resource manager has successfully registered with the server
[01:17:23.152] not ok 2 - custom WAL resource manager has successfully written a WAL record
[01:17:23.152] 1..2
[01:17:23.152] # test failed
[01:17:23.152] --- stderr ---
[01:17:23.152] # Failed test 'custom WAL resource manager has successfully written a WAL record'
[01:17:23.152] # at /tmp/cirrus-ci-build/src/test/modules/test_custom_rmgrs/t/001_basic.pl line 56.
[01:17:23.152] # got: '0/151E088|test_custom_rmgrs|TEST_CUSTOM_RMGRS_MESSAGE|40|14|0|payload (10 bytes):
payload123'
[01:17:23.152] # expected: '0/151E088|test_custom_rmgrs|TEST_CUSTOM_RMGRS_MESSAGE|44|18|0|payload (10 bytes):
payload123'
[01:17:23.152] # Looks like you failed 1 test of 2.
[01:17:23.152]
Not many buildfarm members test 32b builds, but lapwing does.
--
Michael
