sed and encodings

Julien ÉLIE julien at trigofacile.com
Thu Jan 29 21:46:17 UTC 2009


Hi Russ,

>> Then, why not use LC_CTYPE=C instead of LC_ALL?
>> It should also fix the encoding problem.
>
> Hm, we could do that.  Do we also need to set LC_COLLATE=C?

If we do that, I think that we may break sorting.
For instance, in a French locale, we want to have:

    enfant
    été
    ex æquo

and not:

    enfant
    ex æquo
    été

("é" comes after "e" in a C locale -- whereas it is the /same/ letter
in French sorting)

If LC_CTYPE is set to "C", will there be an error when a string badly
encoded according to LC_COLLATE is given to a sorting function?
(I do not know.)

-- 
Julien ÉLIE

« -- Prends un peu de potion magique, Jolitorax ?
  -- Mais ça va être l'heure de l'eau chaude ! » (Astérix) 




More information about the inn-workers mailing list