Horton

Re: Transform for pl/perl

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
Cc: Andrew Dunstan <andrew.dunstan@2ndquadrant.com>, Dagfinn Ilmari Mannsåker <ilmari@ilmari.org>, Pavel Stehule <pavel.stehule@gmail.com>, Nikita Glukhov <n.gluhov@postgrespro.ru>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2018-04-29T18:28:03Z
Lists: pgsql-hackers 
Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes:
> On 4/24/18 14:31, Andrew Dunstan wrote:
>> There is the routine IsValidJsonNumber that helps - see among others
>> hstore_io.c for an example use.

> I would need something like that taking a double/float8 input.  But I
> think there is no such shortcut available, so I just wrote out the tests
> for isinf and isnan explicitly.  Attached patch should fix it.
> jsonb_plpython will need a similar fix.

I looked this over, it looks fine to me.  I first questioned the use
of ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE for rejecting NaN, but I see
that we are doing that in lots of comparable places (e.g., dtoi4())
so I'm on board with it.

			regards, tom lane

Commits

  1. Fix jsonb_plperl to convert Perl UV values correctly.

  2. Fix contrib/hstore_plperl to look through scalar refs.

  3. Allow plperl_sv_to_datum to look through scalar refs.

  4. Fix excessive enreferencing in jsonb-to-plperl transform.