Re: Broken handling of lwlocknames.h - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: Broken handling of lwlocknames.h
Date
Msg-id CAB7nPqS4N3vV1LbvcLWpZsnPupJMifKcFDgpcs1cOkjziUdQFQ@mail.gmail.com
Whole thread Raw
In response to Re: Broken handling of lwlocknames.h  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Sat, Jul 2, 2016 at 4:13 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Michael Paquier <michael.paquier@gmail.com> writes:
>> Yes that would be indeed cleaner this way. I have poked a bit at that
>> and finished with the attached that defines some rules to generate all
>> the files needed.
>
> I made some mostly-cosmetic changes to this and pushed it.  One thing
> to note is that it seemed to me you'd broken the rule for schemapg.h:
> by removing the phony target, I think you removed the behavior that
> we'd always go and recheck schemapg.h's dependencies.  To do it correctly
> without that target, we'd need src/backend/Makefile to know all of those
> dependencies, duplicating the rather long list in catalog/Makefile.

Thanks. Yes the bug that I was mentioning yesterday was that I missed
one place where the target of schemapg.h was referenced, and I didn't
take the time to fully cover this set of dependencies..
-- 
Michael



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: pg_replication_origin_xact_reset() and its argument variables
Next
From: Robert Haas
Date:
Subject: Re: [sqlsmith] ERROR: plan should not reference subplan's variable