[XML-SIG] Dropping Python 1.5 support in PyXML

[Dinu Gherman]

"Martin v. Loewis" <martin@v.loewis.de>:

> > Well, that is enough to make new versions of PyXML incompatible with
> > Python 1.5, isn't it? If so, this measure probably does qualify as
> > "actively dropped support for 1.5".
> 
> Not necessarily. Someone could volunteer to back-port all changes to
> 1.5. Also, it may be that users of Python 1.5 don't care about new
> features, as long as it still installs.

Well, someone could volunteer to rewrite CPython in K&R-style C. Does 
that mean Python is actively supported on dinosaur PDP-11 boxes?;-) 
(Just an example, I'm not sure there were ANSI-C compilers for them, 
too.)

> > Isn't that a highly subjective argument? For others list comprehen-
> > sions qualify as a more or less efficient obfuscation mechanism.
> 
> To modify a Linus Torvalds quote: He who writes the code choses the
> syntax. If the authos, of, say, c14n.py would like to use list
> comprehensions in his code, I had no problems with that. If somebody
> contributed a patch to xmlproc to change all loops to list
> comprehensions, I'd check with the xmlproc author first.
> 
> Being able to use the latest language features is attractive for some
> developers, and I'd like to attract developers to contribute to PyXML.

I make a distinction between "really" new features and other ways 
(sometimes more convenient ones) to do the same thing. Maybe some
of you still remember the times when Python was taking pride for
its orthogonal design? But that's an entirely different story and 
should not be discussed here, I agree...

Regards,

Dinu