will python 3000 break my code?

Daniel Berlin dan at cgsoftware.com
Wed Mar 1 20:54:04 CET 2000


> 
> > Sorry, that's just not the way it works.
> > This is a *BUG*. NOT A FEATURE.
> then why is the 1.5.2 standard library (huge parts of which
> are written by Guido) using this all over the place?
> 

Let me expand a bit.
Why would something that is a bug suddenly become a feature simply because
a lot of code happened to be affected by that bug?
It doesn't make it any less of a bug.
When turning on strict aliasing in gcc broke the linux kernel, because
Linus violates the ansi rules for for aliasing all over the place, it
didn't make it a bug in gcc.
It just meant that the linux kernel is *very* buggy.
(I picked this example because it dealt with a very very large base
of broken
code)
If you don't follow the language rules, accidently, or on purpose, your
code is  bound to lose.
When it does, why should you get to complain because the implementation
actually follows the rules more stricly now?
--Dan





More information about the Python-list mailing list