Re: pg_regress: stat correct paths

Alvaro Herrera wrote:
> Jorgen Austvik - Sun Norway wrote:
>
>> This patch will also make pg_regress fail if the directories are
>> missing, instead of silently ignoring the missing directories. If you
>> think this fail-fast behavior is undesirable, I can create a patch
>> that just ignores the files (todays behavior), or ignores them with a
>> warning.
>
> This silent-skip behavior is there to avoid noise in other tests. Try
> the ecpg tests, contrib tests and src/pl tests.

OK, thank you for the pointers, I will look into this.

I will send out a new patch that is tested with those, and that does the
silent-skip.

> I didn't check the rest of the patch, but did you verify that it still
> works in VPATH builds?

I saw two comments in the file mentioning VPATH, but I have no idea
about what that is. Based on reading the code, I belive this patch
should work better with VPATH (setting srcpath) than without, as it
actually checks those folders instead of other folders.

-J
--
Jørgen Austvik, Software Engineering
Database Technology Group