I just realised the entry for pg_file_rename is duplicated in pg_proc.h.
Unless someone can say it's a good thing (it was in the original
patch..), please remove one of those entries before applying. It breaks
the opr_sanity test.
//Magnus
> -----Original Message-----
> From: pgsql-patches-owner@postgresql.org
> [mailto:pgsql-patches-owner@postgresql.org] On Behalf Of
> Magnus Hagander
> Sent: Saturday, July 30, 2005 4:39 PM
> To: PostgreSQL-patches
> Subject: [PATCHES] Updated instrumentation patch
>
> Per recent discussion, here is yet another updated version of
> the instrumentation patch. Changes:
>
> * Added guc option "disable_remote_admin", that disables any
> write operations (write, unlink, rename) even for the
> superuser. Set as PGC_POSTMASTER so it cannot be changed remotely.
> I put this under "file locations", because that's where all
> the other config file information is. Though that doesn't
> feel completely right, I couldn't find a better place without
> creating a whole new category (it's not *connection*
> security, after all), and if that's to be done I think it's
> better if one of the committers pick name etc for it :-)
>
> * Make sure pg_file_stat() can only be used by superuser. It
> lacked this check previously.
>
> * Updated so it applies to current cvs. This means all oids
> have changed, since they were all used for other things now.
> Also added a required header that had moved with the datetime stuff.
>
>
> Actual code changes against the previous patch are very small.
>
> //Magnus
>