Damn - sorry, I just found a bug in this. I'll resubmit soon.
Chris
Christopher Kings-Lynne wrote:
> Hi,
>
> This patch fixes pg_dump for the existing ownership and acl problems. It
> does not address the known damaged acls after an ownership change.
>
> It's hacky (although officially hacky :) ) in regards that it parses the
> dropStmt on each object to get the objects name. This is necessary
> because of an oversight in the initial design of the binary format.
>
> I've done quite a bit of testing with it, and it seems to work OK. The
> main thing to review would be my C string manipulation stuff that
> extracts the object from the drop statement. I've been careful to avoid
> overruns and segaults, but I may have made an error.
>
> I've resurrected the old --use-set-session-authorization flag for sql
> standard dumps using SET SESSION AUTHORIZATION instead of OWNER TO.
>
> After my other pg_dumpall patch is committed, we might need to add this
> new flag to pg_dumpall as well.
>
> Chris
>
>
> ------------------------------------------------------------------------
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
