[Python-Dev] RE: Ill-defined encoding for CP875?

[Tim Peters]

[M.-A. Lemburg]
> The problem is: which part would raise the exception -- the
> encoder or the decoder ?

Since I don't yet use any of this stuff for real, I have no idea:  seems
mostly a question of pragmatics, and I don't have any feel for how cp875
users would view it.

> Here are some more options:
>
> * sort the items before creating the encoding table from the
>   decoding one (makes the mapping stable)

If users don't care that round-trip can fail silently, fine.

> * map keys which have multiple mappings in the encoding table
>   to None -- this causes their usage to raise an exception
>   (undefined mapping)

If users don't care that they'll get an exception when they try something
that can't be round-tripped, fine.  Or would this depend on the value of the
"errors" argument too?  Then it's easier to impose.

There's a theme here <wink>:  I have no idea how important roundtrip is in
Unicode Practice, or even that it's a constant across apps and encodings.  If
I write a codec to map all ASCII consonants to u"k" and vowels to u"a",  I
wouldn't care that I can't get "love" back from u"kaka" <wink>.