New PEP: The directive statement

[Tim Peters]

[Martin von Loewis]
> That was rather meant as a comment to PEP 236: Instead of giving a
> mandatory release, it appears that all you can specify is a
> is_mandatory (isMandatory, or mandatory_p) flag; this is also all you
> said an application might need.

I don't think I followed that.  Since the future isn't certain until it
happens (and half the time not even then <0.9 wink>), I don't see that it's
possible "even in theory" to do better than give a best guess, and one that
gets updated if the once-predicted future arrives in a form other than hoped
for.

> ...
> it is also a fact that I do like to see the possibility to specify
> source encodings.

Me too.

> It's just that I cannot hope to get that now - but I had the hope that
> the future statement would be removed before the 2.1 release. Now that
> the final beta still has it, that hope is vanishing...

Note that Guido didn't reject your patch, he postponed it.  He even said he
*liked* the idea of directives!  The "transitional" form of the patch didn't
show up until Tuesday (three tiny days before the 2.1b2 scheduled release:
the worst possible time to try to get mindshare), and it didn't come with
(for example) doc changes, or patches to IDLE or python-mode.el to colorize
stuff the way people will expect.  We couldn't possibly make time to take up
that slack, and 2.1b2 is supposed to be the last chance for *anything* new to
get in.  The PEP was also too late, and now that you've written it it's clear
enough that it hasn't been met with universal delight either.

So I agree w/ Guido's patch comment in all respects:

    I'm postponing this now.

    I like the idea of directives, but I don't think we should
    disrupt the 2.1 release for this.

    So let's look at this again after 2.1.

You have to work pretty hard to get depressed by that <wink>.