Re: plperl - make $_TD global

From: Andrew Dunstan <andrew(at)dunslane(dot)net>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: "Patches (PostgreSQL)" <pgsql-patches(at)postgresql(dot)org>
Subject: Re: plperl - make $_TD global
Date: 2006-05-22 22:57:18
Message-ID: 447241CE.3060306@dunslane.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-patches

Tom Lane wrote:

>Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
>
>
>>The attached tiny patch will fix the problem Greg Sabino Mullane had
>>with a shared lexical $_TD, by making it a global and just pushing a
>>local value in the trigger function.
>>...
>>I don't think a docs change is needed.
>>
>>
>
>Are you sure this doesn't cause any user-visible semantics change
>(ie, something that might break someone's code)?
>
>
>
>

I think it should be mentioned in the release notes.

I would be fairly stretched to come up with an example that it breaks.
Essentially it's a way around the sharing violation that Greg got from
using $_TD in a nested subroutine. All we are doing is moving $_TD from
the local namespace to the global namespace. It still gets a per
trigger-call value (that's what the "local $_TD = $_[0];" does) and that
will work correctly even under recursive calls, so I think it's fairly safe.

Maybe it is worth putting somethng in the plperl Trigger docs about the
nature of the beast. I will do that if you like.

cheers

andrew

In response to

Responses

Browse pgsql-patches by date

  From Date Subject
Next Message Bruce Momjian 2006-05-23 01:25:42 Re: [HACKERS] Duplicate definition of LOCALEDIR
Previous Message Tom Lane 2006-05-22 22:45:28 Re: plperl - make $_TD global