Alexander Korotkov <aekorotkov@gmail.com> writes:
> On Tue, Jan 30, 2024 at 8:46 PM Pavel Stehule <pavel.stehule@gmail.com> wrote:
>> I marked this patch as ready for committer.
> The last version of the patch still provides transform for builtin
> type in a separate extension. As discussed upthread such transforms
> don't need separate extensions, and could be provided as part of
> upgrades of existing extensions. There is probably no consensus yet
> on what to do with existing extensions like jsonb_plperl and
> jsonb_plpython, but we clearly shouldn't spread such cases.
Yeah, I think including this as part of "plperl[u] 1.1" is probably
the best way forward. The patch of record doesn't do that, so
I've set the CF entry back to Waiting On Author.
Taking a quick look at the rest of the patch ... I think the
documentation is pretty inadequate, as it just says that the transform
causes byteas to be "passed and returned as native Perl octet
strings", a term that it doesn't define, and googling doesn't exactly
clarify either. The "example" is no example at all, as it does not
show what happens or how the results are different. (Admittedly
the nearby example for the bool transform is nearly as bad, but we
could improve that too while we're at it.)
regards, tom lane