Re: Review: Typed Table - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Review: Typed Table
Date
Msg-id 12882.1264708873@sss.pgh.pa.us
Whole thread Raw
In response to Re: Review: Typed Table  (Peter Eisentraut <peter_e@gmx.net>)
List pgsql-hackers
Peter Eisentraut <peter_e@gmx.net> writes:
> The right thing would probably be SELECT FOR SHARE on the pg_type row,
> but I don't see that sort of thing used anywhere else in system catalog
> changes.

If we were to do it the right thing would just be to define a locktag
for type OIDs and add appropriate locking calls all over the system.
But that would be a large, invasive change that seems far outside the
scope of this patch, and certainly much beyond what can get done for
9.0.

(Actually, if memory serves there is some notion of locking on arbitrary
catalog objects already in the DROP code, so there probably is a
suitable locktag defined already.  But getting ALTER and generic type
references to play along is still a major project, and I'm not convinced
about the cost/benefit ...)
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tim Bunce
Date:
Subject: Re: Add on_trusted_init and on_untrusted_init to plperl [PATCH]
Next
From: Tim Bunce
Date:
Subject: Re: Add on_trusted_init and on_untrusted_init to plperl [PATCH]