On 2 December 2013 04:55, Pavel Stehule <pavel.stehule@gmail.com> wrote:
> Hello
>
> it looks well, thank you
>
> Regards
>
> Pavel
>
I've been thinking about this some more, and there's another case that
concerns me slightly. We're now making some of the DROP...IF EXISTS
commands tolerate non-existent types as well as non-existent schemas
--- functions, aggregates, casts and operators all have type names in
their specifications. Of course it's possible that the type is missing
because it was in a schema that was dropped, so this change seems to
be in spirit of what was discussed, but it seems like a change that
might catch some people out.
I think that, on balance, it is a sensible change, since if the type
doesn't exist, the dependent object can't exist either, so DROP...IF
EXISTS shouldn't be raising an error. However, I wonder if we should
be issuing a more specific NOTICE in this case too --- i.e., check for
non-existent types in the same way as we check for non-existent parent
objects --- type_does_not_exist_skipping() and
type_list_does_not_exist_skipping().
Regards,
Dean
