Ray Satiro <raysat...@yahoo.com> writes: > Hello, > > I notice that idna_to_ascii_lz can return success even if the string > could not be converted from the current locale due to being built > without libiconv [1][2], and that doesn't appear to be documented. The > result appears to be libidn without libiconv treats the locale encoded > string as utf-8 and attempts to convert from utf-8 to punycode. Could > someone explain if that is intended or not and whether there are any > implications I should look into.
Hello. Thanks for analysis. You are indeed correct. Why would you use libidn without libiconv? And in particular the idna_to_ascii_lz function? I see some solutions here: 1) Make libiconv a hard requirement of libidn. 2) Document the behaviour you have found, as a negative side-effect of building libidn without libiconv. 3) Make locale-related functions return an error when libidn was built without libiconv. I don't have a strong opinion here. I think we should understand better why libidn is used without libiconv in the real world before making decisions. /Simon
signature.asc
Description: PGP signature
_______________________________________________ Help-libidn mailing list Help-libidn@gnu.org https://lists.gnu.org/mailman/listinfo/help-libidn
