[Web-SIG] Python 3.0 and WSGI 1.0.

[P.J. Eby]

At 02:39 PM 5/8/2009 -0400, James Y Knight wrote:
>On May 8, 2009, at 1:37 PM, Robert Brewer wrote:
>>If Python3 WSGI apps emit unicode strings (py3k type 'str'), you're
>>effectively saying the server will always call
>>"chunk.encode('latin-1')". That negates any benefit of using unicode
>>as
>>the type for the response. That's not "supporting unicode"; that's
>>using
>>unicode exactly as if it were an opaque byte string. That's seems
>>silly
>>to me when there is a perfectly useful byte string type.
>
>Agreed. Accepting py3k "str" and always encoding in latin-1 is
>basically just undoing the separation of unicode&byte-strings that was
>one of Py3k's major design goals.
>
>Probably there should be nothing in WSGI should be allowed to be given
>as either bytestring or character string. The spec should choose one
>or the other for each circumstance. And for body content it's clear
>that the only sane thing is a bytestring.

With the amendments as written (and previously discussed here), 
accepting latin-1 (or ASCII-only) strings allows backward 
compatibility with code converted via 2to3.  Otherwise, you would 
have to track down every string-returning function in your program 
that *might* be used to generate a response or a yielded portion thereof.