Re: Some oversights in query_id calculation

Hi Aleksander,

On Wed, Apr 28, 2021 at 01:19:36PM +0300, Aleksander Alekseev wrote:
> Hi Julien,
> 
> > I'm attaching a patch that fixes those, with regression tests to reproduce each
> > problem.
> 
> I believe something could be not quite right with the patch. Here is what I did:
> 
> $ git apply ...
> # revert the changes in the code but keep the new tests
> $ git checkout src/backend/utils/misc/queryjumble.c
> $ ./full-build.sh && single-install.sh && make installcheck-world
> 
> ... where named .sh scripts are something I use to quickly check a patch [1].
> 
> I was expecting that several tests will fail but they didn't. Maybe I
> missed something?

I think it's because installcheck-* don't run pg_stat_statements' tests, see
its Makefile:

> # Disabled because these tests require "shared_preload_libraries=pg_stat_statements",
> # which typical installcheck users do not have (e.g. buildfarm clients).
> NO_INSTALLCHECK = 1

You should see failures doing a check-world or simply a make -C
contrib/pg_stat_statements check