Re: postgres.h included from relcache.h - but removing it breaks pg_upgrade

Andres Freund <andres(at)anarazel(dot)de> writes:
> On 2021-09-13 22:40:19 -0400, Tom Lane wrote:
>> As for the fix ... what in the world is pg_upgrade doing including
>> relcache.h? It seems like there's a more fundamental problem here:
>> either relcache.h is declaring something that needs to be elsewhere,
>> or pg_upgrade is doing something it should not.

> We could split visibilitymap.h into two, or we could forward-declare Relation
> and not include relcache...

Without having looked at the details, I think using a forward-declare
to avoid including relcache.h in visibilitymap.h might be a reasonably
non-painful fix. OTOH, in the long run it might be worth the effort
to split visibilitymap.h to separate useful file-contents knowledge
from backend function declarations.

>> No. If anything, I'd want to throw an error for "redundant" includes
>> of these files, because it's a pretty good red flag about
>> poorly-thought-out header modularization.

> I think we might be thinking of the same. What I meant with "avoid" was to
> raise a warning or error.

Ah, we are on the same page then. I misunderstood what you wrote.

> If we were to do that, it's probably worth doing the
> build system ugliness to do this only when building postgres code, rather than
> extensions...

As long as we do this in HEAD only, I'm not sure why extensions
need an exception. Perhaps it will result in somebody pointing out
additional poorly-thought-out header contents, but I don't think
that's bad.

regards, tom lane