Hello Dan,

> The big problem with this is character set conversion. If the URL is in UTF-8
> and the user's filesystem is in CP1252, this will produce a garbage file name.
> Or even worse, the other way around which could result in a file that's not
> even accessible by some software. Since there's no way to tell what character
> set the URL is in, this seems impossible to do correctly. Unless you're
> suggesting just handling characters in the range 0x20-0x7e or something along
> those lines, but that could result in confused users when some files are
> converted and some are not.

Those are very good points, I would like to check how wget is doing this today
to see if their approach makes sense.

Also, I will hold off implementing this for wcurl for a bit, at least until we
understand exactly how feasible it would be to have it on curl directly (and
which behavior exactly we could have, e.g.: only decode a set of characters).

Cheers,


-- 
Samuel Henrique <samueloph>
-- 
Unsubscribe: https://lists.haxx.se/mailman/listinfo/curl-library
Etiquette:   https://curl.se/mail/etiquette.html

Reply via email to