[Python-Dev] Re: PEP 318 trial balloon (wrappers)

Paul Moore pf_moore at yahoo.co.uk
Thu Mar 11 17:59:12 EST 2004


"Jewett, Jim J" <jim.jewett at eds.com> writes:

> I am hoping to post a new draft of PEP318 soon.
>
> Unfortunately, the solution that currently looks 
> best to me has not been mentioned before.

[...]

> class Foo:
>     [transform] from:
>         def bar():
>             pass
>

Ack, no.

I'm very strongly -1 on this particular suggestion, but I'd also make
the point that we *really, really* don't need a new draft of the PEP
opening up issues again. Please, only collate what's already been
stated.

My general impression is:

1. Semantics are pretty clear, but not documented explicitly yet. The
   PEP should document them. There's an open issue over the order in
   which decorators are applied.
2. Syntax is coming down to a few contenders. The version implemented
   in mwh's patch, Guido's variation with the [...] in front of the
   args, and variations with "as" (with a few other suggested
   keywords).

Not much more than this. (I know, that's very over-simplified...)

Please don't take this the wrong way, but I got the impression that
you were one of the relatively few people still suggesting more
radical alternatives. While I respect your motives, I hope the revised
PEP will document the overall consensus, with a clear listing of the
basic alternatives. By all means add more radical suggestions, but
please keep them separate, and make it clear that they have not had
the same level of discussion as the more "mainstream" suggestions.

The revised PEP needs to consolidate and summarise the discussions,
not start them up again!

Paul.
-- 
This signature intentionally left blank




More information about the Python-Dev mailing list