2008/12/9 Tom Lane <tgl@sss.pgh.pa.us>:
> "Pavel Stehule" <pavel.stehule@gmail.com> writes:
>> 2008/12/9 Tom Lane <tgl@sss.pgh.pa.us>:
>>> If you could prove that it were *only* being used by this contrib module
>>> then I might hold still for replacing it. But you can't. The odds are
>>> good that people have custom data types using similarly-named operators.
>
>> it means, so we must not implement any new operator?
>
> No, it doesn't mean any such thing. If we invented, say, "int4 => int4"
> it would not break someone's use of => for their own custom datatype.
> What you're proposing would be a global redefinition of the meaning of =>.
it's not true, because anybody could to define own operator on buildin
types - so every new operator is risk and carry problems. So only new
operator on new types are safe. All others shoud be problem - an using
of any well know world carries risks.
>
> This is closer to creating a new reserved word, which as I'm sure you
> know we try hard to avoid, even for keywords that the spec says we can
> reserve. The bar for making a new fully-reserved word that isn't in
> the spec is *very* high.
>
what is problematic on GUC? We use it actually for it? So we should
disable or enable named_params, and when this feature will be
disabled, then pg will be 100% compatible. It's better then creating
some strange syntax.
regards
Pavel
> regards, tom lane
>