[Python-Dev] [PATCH] Fix dictionary subclass semantics whenused as global dictionaries

[Crutcher Dunnavant]

Okay, but is there any reason not to include this in 2.5? There
doesn't seem to be any noticeable performance impact, and it does add
consistancy (and opens some really, really cool options up).

Does anyone have objections to 1402289?

On 1/12/06, "Martin v. Löwis" <martin at v.loewis.de> wrote:
> Crutcher Dunnavant wrote:
> > I sorta disagree about it not being broken. Adding a feature which
> > works for eval but not for exec seems pretty broken.
>
> You "seem" to have a different notion of "to be broken", then.
>
> Python is broken, if and only if
> - the interpreter crashes, for any Python input
> - the implementation does not do what the documentation says
>   it would do
> - the BDFL pronounces it is broken
>
> In this specific change, the change did precisely what the requestor
> of the feature wanted it to do (that eval could accept non-exact
> dicts was a new feature back then also)
>
> > It's difficult to
> > reason about what will happen in the exec context, so I can't see what
> > fixing it would endanger; but I'd deffinately like to see it for 2.5.
>
> It would make Python code run which is currently rejected with an
> exception. Therefore, it is a new feature (a behaviour change).
>
> Applications relying on that feature would have to specify that
> they require "2.4.3"; people would find that code that runs fine
> in 2.4.3 fails in 2.4.2. This is not acceptable.
>
> Regards,
> Martin
>


--
Crutcher Dunnavant <crutcher at gmail.com>
monket.samedi-studios.com