Re: 9.1 plperlu bug with null rows in trigger hash

On Mon, May 23, 2011 at 20:08, Greg Sabino Mullane <greg(at)endpoint(dot)com> wrote:
> On Mon, May 23, 2011 at 05:04:40PM -0600, Alex Hunsaker wrote:
> ...
>> Greg, can you confirm the attached fixes it for you?
>
> Yes, seems to have done the job, thank you.

Thanks for testing!

[ Does a little dance to try and attract a -commiter ]

This was broken as part of:
commit 87bb2ade2ce646083f39d5ab3e3307490211ad04
Author: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
Date: Thu Feb 17 22:11:50 2011 -0300

Convert Postgres arrays to Perl arrays on PL/perl input arguments

More generally, arrays are turned in Perl array references, and row and
composite types are turned into Perl hash references. This is done
recursively, in a way that's natural to every Perl programmer.

To avoid a backwards compatibility hit, the string representation of
each structure is also available if the function requests it.

Authors: Alexey Klyukin and Alex Hunsaker.
Some code cleanups by me.

Which also means it only breaks HEAD/9.1 (I did test and review 9.0 and down.)

Per http://search.cpan.org/~jesse/perl-5.14.0/pod/perlguts.pod#AVs,_HVs_and_undefined_values,
we do not want to use &PL_sv_undef for undef values in hashes and
arrays. I (inadvertently) broke this in the above commit. As perldoc
mentions &PL_sv_undef seems like the intuitive thing to use. But its
wrong in certain cases.

We have 6 other uses of &PL_sv_undef, 2 &PL_sv_no and 1 &PL_sv_yes.
These are all ok as none of them use the HV/AV store interface.

I elected _not_ to add any regression tests. (If we forget about this
in the future, it will likely be in other code paths). Instead I added
comments to the places that used &PL_sv_undef noting that we
explicitly avoid it on purpose.