Re: Another nasty pg_dump problem - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Another nasty pg_dump problem
Date
Msg-id 12921.1059672504@sss.pgh.pa.us
Whole thread Raw
In response to Another nasty pg_dump problem  ("Christopher Kings-Lynne" <chriskl@familyhealth.com.au>)
Responses Re: Another nasty pg_dump problem  (Peter Eisentraut <peter_e@gmx.net>)
List pgsql-hackers
"Christopher Kings-Lynne" <chriskl@familyhealth.com.au> writes:
> On my 7.3 server:
> REVOKE ALL ON TABLE exercise_activities FROM PUBLIC;
> GRANT ALL ON TABLE exercise_activities TO chriskl;
> GRANT SELECT ON TABLE exercise_activities TO "au-diary";
> GRANT SELECT ON TABLE exercise_activities TO "au-php";

> Now if you load that into 7.4CVS, you get:

> REVOKE ALL ON TABLE exercise_activities FROM PUBLIC;
> GRANT ALL ON TABLE exercise_activities TO chriskl;
> GRANT SELECT ON TABLE exercise_activities TO "\""au-diary\""";
> GRANT SELECT ON TABLE exercise_activities TO "\""au-php\""";

I've repaired this in CVS tip.  While testing it, though, I notice that
CVS-tip pg_dump puts out useless commands
REVOKE ALL ON SCHEMA public FROM PUBLIC;GRANT ALL ON SCHEMA public TO PUBLIC;

which are not generated when dumping from 7.3.  The reason evidently is
that this check in pg_dump.c no longer works:
       /*        * If it's the PUBLIC namespace, don't emit a CREATE SCHEMA record        * for it, since we expect
PUBLICto exist already in the        * destination database.  And emit ACL info only if the ACL isn't        * the
standardvalue for PUBLIC.        */       if (strcmp(nspinfo->nspname, "public") == 0)       {           if (!aclsSkip
&&strcmp(nspinfo->nspacl, "{=UC}") != 0)               dumpACL(fout, "SCHEMA", qnspname, nspinfo->nspname, NULL,
              nspinfo->usename, nspinfo->nspacl,                       nspinfo->oid);       }
 

since the default ACL for public no longer looks like that.  Can we fix
this?
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Proof-of-concept for initdb-time shared_buffers selection
Next
From: Josh Berkus
Date:
Subject: Re: followup on previous