Originally Posted By: MeiR_ct
Even after you enable UTF encode/decode in all those major clients, basic commands like "nick" and "join" will [or should] be sent with codpages, as the way it was in mIRC until 6.35, since the vast majority of servers just won't understand anything else.


Err, no, this is not true. It's more likely that these clients will simply use a single encoding for all of their IO. As I pointed out, all of the clients I've seen on OSX only understand one encoding: Unicode. I highly doubt they behave the way you're describing (or wishing). Furthermore, there's no such rule that "commands must be sent using codepages". It's not in the RFC, it doesn't exist in reality either.

Originally Posted By: MeiR_ct


Those statistics are vastly oversimplified. Many networks will use one of those ircds as a base, but perform many customizations on it that don't get detected as a separate version. The number of networks that understand unicode commands are growing daily.

Besides, this isn't even a valid comparison. Even if most servers don't yet understand unicode commands-- what are we comparing this to? Codepaged commands? I don't know of *ANY* server that accepts nicknames using codepages, specifically because it's impossible to know what codepage you want to be using (actually, RusNet might support this, but RusNet has a robust encoding system, so internally it's probably not storing codepage data at all). I think you're confusing this with servers accepting *BINARY DATA* for *CHANNELS* (not nicknames, those are usually strictly defined in the ASCII range). This is indeed an issue, but it's not a server issue. It has to do with the fact that mIRC assumes everything the server is sending is UTF-8. This could be fixed without strictly supporting codepages, mIRC just needs to be more aware about how it reads channel names from a server. Occasionally this is impossible, but for the most part, mIRC could be smarter about the way it does things with channel names. Again, this has nothing to do with supporting codepages, though.

Originally Posted By: MeiR_ct
You said that implementing utf nicks support in ircd is technically possible, and the reason it's not on the go, is developers' laziness.
If talking technically, let's take the simplest examples, such as nick prefixes, or colons. blah blah blah majority of methods.


Are you an IRCd developer? You don't sound like one. Colons are part of the IRC protocol. Enabling UTF-8 doesn't change the fact that colons are not allowed in nicknames. Furthermore, the ":" is a single byte character in UTF-8, so nothing would change here, even if it was magically allowed in the spec. "Supporting unicode" does not mean "any character is allowed". Also, please don't put words in my mouth, because I never claimed it was easy. I actually know that it's not easy, but I don't really care. Why should I care that it's not easy? Is that some kind of valid excuse? I wouldn't think so. So let's move on.

Originally Posted By: MeiR_ct
As soon as it gets to a release of higher version, aka upgrade, users will almost always update their software


Ahh, the irony. You were paying attention to my previous response, weren't you? I've gathered data on this very issue before when investigating the number of users still on codepages. The vast majority in channels with codepage encodings are users on mIRC. Remember, 7.x users can't join these codepage channels, so we're talking 6.x. In fact, the data I collected showed that most users on this channel averaged an mIRC version of ~6.2 (I saw enough 5.x's to make you cry). People aren't upgrading. That is precisely the problem. If every 6.x user upgraded to 7.x *TODAY*, this discussion would be over, because there would be no codepage problem. mIRC is virtually the last client to still be using codepages. The other clients that use codepages are doing so for compatibility with mIRC only. If you visited that PJIRC link in my last response you would have seen that the conversation was exactly about trying to communicate with mIRC users who could not use UTF-8 (because PJIRC defaults to UTF-8). The conversation took place in 2004.

For what it's worth, mIRC 6.16+ supports UTF-8, and mIRC 7.x supports disabling UTF-8 outright. That means 2 things:

1) Most 6.x users can start using UTF-8. Tell your friends to enable it.
2) 7.x users can disable UTF-8 if they really want to talk to their 6.x friends.


- argv[0] on EFnet #mIRC
- "Life is a pointer to an integer without a cast"