KaiGai Kohei wrote:
> Alvaro, could you check the patched code on reloptions.h, reloptions.c
> and rel.h? It is a working example of string reloptions, and I could
> found a few strange codes.
I'm intending to revisit the string code ... I was thinking yesterday
night that I shouldn't have committed it at all, and left it for a
subsequent patch that I had more chance to test properly :-(
> 1. HANDLE_STRING_RELOPTION() always put an empty string when
> optstring->default_isnull is true, even if user gives a
> valid string reloption.
This is a plain bug, sorry.
> 2. HANDLE_STRING_RELOPTION() cannot handle an offset style.
> The patched one enables to put reloption string on the
> tail of StdRdOptions structure, and adjust offset value.
I'll look at it, thanks.
> 3. Why the "StdRdOptions lopts;" is necessary?
It is like this because the autovacuum patch adds a few more options and
I want to have the chance to not allocate the part belonging to
autovacuum when none of the options are present.
> And, I have a request.
> 4. Is it possible to support a call-back to validate a given
> string reloption? I want to check whether the given default
> Row-level ACLs has a valid format, or not.
Hmm, why a callback and not just call the validation function in
heap_reloptions?
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support