[dcc2] DCC Chat suggestions

Tue May 25 16:39:02 EDT 2004

Oh now that's a very good idea ;) With CTCP support over Chat as
discussed, it shouldn't be hard at all. The only problem I see is
differentiating between CTCP and CTCP Reply as there is no
PRIVMSG/NOTICE.

Moving back a reply or two, I see problems with NICK as a CTCP relying
on the other end to allow or deny. It'd be trivial (though quite
silly) to disallow any nick changes on the part of the other side. It
might be better to have clients just silently ignore invalid NICK
requests.

Another thing to think about: beware of treating DCC CTCPs with the
same code as IRC CTCPs... DCC Chats do not have the length
restrictions imposed on IRC and therefore potential buffer overflows
may pop up (too-long messages and such).

I don't see why 'KEEPALIVE' couldn't just be 'PING'.

-myndzi

On Tue, 25 May 2004 11:29:54 -0400, Dave Johnson
<dave-dcc2-org at iroffer.org> wrote:
> 
> 
> Was this suggested intended to be in-band over the DCC2 connection
> using an escape sequence or through the IRC server?
> 
> In-band may be nice for IRCChat, but not IRCFile.  I dont think we
> should have an in-band control channel for IRCFile as clients
> would then need to search for escape sequences in the file stream.
> 
> Dan Smith writes:
> > Hey Dominick,
> >
> > I think your suggestion of a command channel could be very useful.  Another
> > item I would suggest adding is DCC2 negotiation.  With DCC2 negotiations
> > for new connections taking place over an established chat session, a user
> > could spawn a whiteboard, file transfer, or video/voice session.
> >
> > Cheers!
> > Dan
> >
> > At 06:08 PM 5/21/2004 -0400, you wrote:
> > >Well, no one seems to have any real thoughts on any additional features for
> > >DCC2's DCC Chat feature, so I thought I might make some suggestions myself.
> > >I think the new DCC Chat should encompass everything the current DCC Chat
> > >has, but add some more useful features as well.
> > >
> > >
> > >Current non-standard features:
> > >CTCP ACTION is maintained and handled as it would be for regular IRC text
> > >A client should format a DCC Chat message just as one from IRC (font codes,
> > >color codes, etc.)
> > >
> > >New stuff:
> > >CTCP is the means of control within a DCC Chat session.
> > >
> > >CTCP NICK
> > >Currently, there is no way to change the nickname you appear as in a DCC
> > >Chat. Most clients I've seen just display the nick that was given in the
> > >CTCP response. If I, as codemastr, send a DCC chat request, no matter what I
> > >change my nickname to, I still appear as codemastr in the DCC session. The
> > >CTCP
> > >NICK feature changes this:
> > >
> > >\1NICK newnick\1
> > >
> > >I think it is rather useful. First off it makes things uniform. If I change
> > >to "codemastr-away," why shouldn't I appear that way to people in DCC Chat
> > >too? Also, think of it differently, DCC is independant of the IRC server.
> > >So, for example, when DCC Server support is addressed, it could be rather
> > >useful to allow users to change their nicknames. To do this, a legal set of
> > >characters would need to be defined. Probably those from RFC1459 or RFC2812,
> > >though, if we wanted to look the the future, we could again realize that DCC
> > >is seperate from IRC and use a Unicode scheme. I'm sure that would please
> > >people in non Latin-charset areas. The only other problem I see is the issue
> > >of trying to change your nickname to the nick of the other party. To address
> > >these, a new CTCP ERROR would be introduced.
> > >
> > >CTCP KEEPALIVE
> > >This is something similar to a PING (command, not CTCP) on IRC. It's just
> > >something I was thinking of as a means to try and make sure the connection,
> > >hasn't been, and won't be, dropped. Basically, when I send a CTCP KEEPALIVE,
> > >you respond back with a CTCP KEEPALIVE.
> > >
> > >CTCP ERROR
> > >As I said, this would be used with CTCP NICK, but it can also be used for
> > >other things. The syntax is:
> > >\1ERROR type text\1
> > >
> > >The type can be one of the following:
> > >NICKINUSE (the nickname specified in the NICK CTCP is already in use. Not
> > >only is this needed for the other party saying "that's my nickname!" it also
> > >works for partyline scripts so that the "server" can tell the client that
> > >the nick specified is already in use by another partyline member)
> > >
> > >NICKILLEGAL (the nickname specified in the NICK CTCP contains illegal
> > >characters)
> > >
> > >x-* (The x- prefix is used for private error types. This can be used for
> > >individual implementations that wish to add extended features and may need
> > >to report errors for those features).
> > >
> > >Lastly, the client SHOULD treat all CTCPs (FINGER, VERSION, PING, TIME,
> > >etc.) sent through DCC as it would for a regular IRC connection. This is for
> > >2 reasons. One, clients support DCC Server (which as I said, I'd like to see
> > >standardized as well). This means, I can start a DCC chat with someone
> > >without ever being connected to an IRC server! Doesn't it make sense to
> > >allow such users to get this information? Also, even if you forget about DCC
> > >Server, I can initiate a DCC Chat over IRC, then disconnect from IRC and the
> > >DCC remains open. It'd be nice to still be able to get this information.
> > >
> > >I'm sure some people will disagree with the ideas, but I'm throwing them out
> > >there anyway :)
> > >
> > >P.S. Are there any plans to begin discussion of the other proposed DCC
> > >extensions (voice/video/whiteboard)? For filesends and chat, it's pretty
> > >easy. We have "standardized" existing implementations to work from.
> > >Voice/video/whiteboard are going to need much more fundamental thought work
> > >done (several implementations exist, and it's possible none are truly
> > >perfect), and so, imho they should probably be something people begin to
> > >think about sooner rather than later...
> > >
> > >-- codemastr
> > >
> > >
> > >_______________________________________________
> > >dcc2 mailing list
> > >dcc2 at dcc2.org
> > >http://six.pairlist.net/mailman/listinfo/dcc2
> >
> > ---------------------------
> > Dan Smith
> > +1 608-213-2867
> > Algenta Technologies, LLC
> >
> > _______________________________________________
> > dcc2 mailing list
> > dcc2 at dcc2.org
> > http://six.pairlist.net/mailman/listinfo/dcc2
> 
> --
> Dave
> 
> _______________________________________________
> dcc2 mailing list
> dcc2 at dcc2.org
> http://six.pairlist.net/mailman/listinfo/dcc2
>