2010/9/28 Tom Lane <tgl@sss.pgh.pa.us>:
> Pavel Stehule <pavel.stehule@gmail.com> writes:
>> 2010/9/28 Tom Lane <tgl@sss.pgh.pa.us>:
>>> Yes, there is. The syntax you propose is flat out ambiguous: there are
>>> two possible legal interpretations of some commands.
>
>> what are you thinking? The subquery cannot be interpreted different.
>
> Sure it can: it could be a parenthesized top-level query. In fact,
> that's what plpgsql will assume if you feed it that syntax today.
no - there are not any legal construct FOR r IN (..)
I believe so we can find more than one similar undocumented features,
like this - so it means so plpgsql will be a buggy?
>
> regards, tom lane
>