[Python-Dev] urllib.quote and unicode bug resuscitation attempt

[Mike Brown]

Stefan Rank wrote:
> Well, originally, I would have expected it to return a byte str(ing),

I'd assume unicode in, unicode out, and str in, str out, but since it's
always going to produce ASCII-range characters, it wouldn't matter.
Moot point anyway.

> BUT
> I am now converted and think it is best to raise a TypeError for 
> unicode, and leave the encoding decisions to higher level code.
> 
> So I'll repeat the "patch" #1, slightly modified::
> 
>   if isinstance(s, unicode):
>       raise TypeError("quote expects an encoded byte string as argument")
> 
> Is it safe to assume that code that breaks because of this change was 
> already broken?

Yes. The patch seems fine to me, although maybe

if not isinstance(s, str)

would be better?