py3k s***s

[Rhamphoryncus]

On Apr 16, 12:52 pm, Aaron Watters <aaron.watt... at gmail.com> wrote:
> On Apr 16, 2:33 pm, Rhamphoryncus <rha... at gmail.com> wrote:
>
> > The point is, you can't have it both ways.  Either you evolve the
> > language and break things, or you keep it static and nothing breaks.
>
> I disagree.  You can add lots of cool
> stuff without breaking the existing code base, mostly.
> For example the minor changes to the way ints will work will
> effect almost no programs.

What changes are minor though?  Eliminating old-style classes should
be minor, but I'm not sure it is.  Applications & libraries have a way
of depending on the most obscure details - even if trivially fixed, it
still requires a fix.  Consider "as" becoming a keyword, or True and
False.

In hindsight, it would have been better to add future imports for
unicode literals and print-as-a-function back in 2.5.  I guess the
time machine was out of service.  2.6 will have to do (and that's what
it's for.)

I'm personally not too worried about the syntax changes though,
they're superficial(!).  What I am worried about is the library APIs
changing to use unicode instead of bytes.  It's not something you
could do incrementally without providing two of every lib or two of
every API - having .write() and .writeex() would suck.