From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; ja-JP; rv:1.6) Gecko/20040510 Description of problem: IRC charset must be ISO-2022-JP by default for Japanese locales. (ja_JP.UTF-8, ja_JP.eucJP) So when user run xchat in Japanese locales, it should set character set as ISO-2022-JP (Japanese) by default. Users don't want use a pico seconds of their time to select the appropriate character set. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Run xchat in Japanese locale. 2. Join Japanese IRC channel. 3. Scrambled. Actual Results: Scrambled by default. Expected Results: ISO-2022-JP by default. Additional info: Red Hat/Fedora must always be responsible which Japanese codeset you are posting to the Internet.
Unfortunately, the IRC protocol has no standard support for internationalization. The protocol was not designed to support explicit negotiation of character sets between users, by channel, or by server/network, and it was not designed with a way to add such support as an authoritative extension (however, see below). Traditionally, all messages sent through IRC are assumed to be using a modified form of US-ASCII (possibly refered to as the "ascii character set" with "rfc1459 case mappings"), and it may be a feature that all messages received by IRC clients are displayed in that character set rather than their host session's character set (unless explicitly reconfigured by the user). An alternative, where the IRC client presumes that all incoming messages are in the character set of its host session, does not sound more reasonable than the traditional behaviour. Another alternative may be available on servers that use the non-standard 005 "extended version" numeric sent during connect. This numeric exists on many major IRC servers and includes both "CHARSET" and "CASEMAPPING" fields, which could be used to tell the client which character set and case map to use by default (instead of US-ASCII and the pseudo-Scandinavian "rfc1459" case map). This would be per-network, however, as it would not be reasonable to have some servers on a network advertize one character set and other servers advertize a different one, so it would only make sense for pure-Japanese IRC networks (it could not be used by a Japanese user using the Red Hat IRC network, the EFnet IRC network, etc.). The relevant documentation is RFC 1459 "Internet Relay Chat Protocol" section 2.2 "Character codes", which specifically states that no thought was given to character code specification in the protocol.