Re: preserving forensic information when we freeze

On 2013-12-20 21:56:42 -0500, Robert Haas wrote:
> On Fri, Dec 20, 2013 at 7:51 AM, Andres Freund <andres(at)2ndquadrant(dot)com> wrote:
> >> I wondered that, too, but it's not well-defined for all tuples. What
> >> happens if you pass in constructed tuple rather than an on-disk tuple?
> >
> > Those should be discernible I think, t_self/t_tableOid won't be set for
> > generated tuples.
>
> I went looking for existing precedent for code that does things like
> this and found record_out, which does this:

I think there's plenty precedent for C code, but here the problem seems
to be that, when we pass a record that's originally a plain row, it's
coerced (via IO functions) into a record with typeid/typemod set.

So, for this to work via SQL, we'd need to add a notation that allows
passing table rows to functions without being modified. That might be a
good idea in general, but likely more work than it's work tackling in
this context.

> This appears to be a typical pattern, although interestingly I noticed
> that row_to_json() doesn't bother setting t_tableOid or t_self, which
> I think it's supposed to do.

Yes, that seems to be a minor bug. Andrew?

> Am I missing a trick?

No, don't think so, I wasn't thinking on the SQL level.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services