[Python-Dev] Decoder functions accept str in py3k

[Terry Reedy]

Guido van Rossum wrote:
> OK, ignore my previous comment. Sounds like the inidividual codecs
> need to tighten their type checking though -- perhaps *that* can be
> fixed in 3.0.1? I really don't see why any codec used to convert
> between text and bytes should support its output type as input.
> 
> --Guido
> 
> On Wed, Jan 7, 2009 at 10:26 AM, M.-A. Lemburg <mal at egenix.com> wrote:
>> On 2009-01-07 16:34, Guido van Rossum wrote:
>>> Sounds like yet another remnant of the old philosophy, which indeed
>>> supported encode and decode operations on both string types. :-(
>> No, that's something I explicitly readded to Python 3k, since the
>> codecs interface is independent of the input and output types (the
>> codecs decide which combinations to support).

My memory is that making decode = bytes -> str and encode = str-> bytes 
was considered until it was noticed that there are sensible same-type 
transforms that fit the encode/decode model and then decided that 
reusing that model would be better than adding a transcode module/model.

The bug of Unicode de/encoders allowing wrong inputs and giving weird 
outputs confuses people and has come up on c.l.p, so I think fixing it 
soon would be good.

tjr