Re: Bug in triggers

On Fri, Mar 5, 2010 at 2:32 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>>> On Wed, Mar 3, 2010 at 9:53 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>>> It does seem weird that assigning NEW to var changes the value; I'm
>>> not sure why that happens.  Is that what you're asking about?
>
>> Anyone else have an opinion on whether this is a bug?
>
> It's arguably a bug, but since we lack consensus on whether NULL and
> ROW(NULL,NULL,...) are the same thing, it's difficult to make a
> bulletproof case either way.  In any case nothing is likely to get done
> about it in the near term because it's wired into plpgsql's
> implementation.  Changing from row to record representation of such
> variables is possible but would probably have side effects, ie, it would
> create new compatibility issues of unknown seriousness.  I'm not too
> optimistic about the performance implications either.

I don't know if it is a bug. Different textual representations could
easily happen due to intermediate conversions of datatypes....

For example: I wouldn't expect timestamp::date::text to equal
timestamp::text. Textual representations are not necessarily
consistent.

I guess a better question for Oleg might be:

"Why is it important to you to get this fixed? What are you trying to
do that you can't do without fixing this?"

Best Wishes,
Chris Travers