Random note

People post: I forgot to mention that peavey now has a bouncing baby boy, who also seems to be heading well on the way to being a hacker.

Grats to peavey and his wife for their new arrival, and a wish of good health to all three from us

Module loader..

As those amongst our fans who are following 1.2 would have noticed -- the module loader rework is now under way, spearheaded by Om.

Module factories have already been removed, the new API is currently being worked on. It's looking good so far.

UID support has continued, with no servernames now being sent server to server. Server IDs are used instead, and are also TS6 compliant. Kudos to Brain for that.

In slightly non-insp related news, Brain is off to Vegas for a well earned break shortly. He'll be back sooner or later, though :p

Functors

Another thing that slipped my mind: functors!

Brain and Om have been doing some really neat work on this in trunk. It's interesting stuff, and could lead to some really neat modules in the future once it's use becomes more widespread.

For more information, see http://www.inspircd.org/wiki/Development/Functors

Roadmap

While I'm blogging, I suppose I should note that the roadmap on the wiki now has a number of Development/ subpages.

Interesting info if you want to look over 1.2 development.

http://www.inspircd.org/wiki/Roadmap

Holidays and UUID

What do the two have in common?

Well, a holiday lead to UUID support in trunk (1.2). Yeah, see, Brain and I got bored, so we sat down and now we have UUID. That's the sort of insane stuff we have.

Anyway..

Our implementation is somewhat similar to TS6. SID is 3 characters starting with a number, CID is 6 characters 0-9A-Z, UID is SID + CID.

All commands now support this, though modular S2S commands will require a TRANSLATE*() macro in their constructor. For an example, see m_remove. (for the curious, TR_NICK is automagically translated to UID before being sent to another server). Remote command handlers will recieve a UID instead of a nick, and InspIRCd::FindNick() finds by UID if the first character is a digit. Neat stuff.

Nick collisions have been given a workover too in amongst this: Now, if there is a nick collision upon linking, one (or both) clients will have their nickname changed to their UID, according to TS rules. Much better than having both killed.

A similar thing happens with nick collisions caused by SVSNICK failing: if the original SVSNICK fails, it attempts to use the client's UID as a last ditch effort. If that fails, then and only then is the client exitted.

All in all, some interesting work going on in trunk at the moment.

Recap

Lots has been happening, and I've had people bitching at me left, right and center to ..actually post something :p

So, in the spirit of goodwill, here goes. I'm blogging! :p

We've had a number of 1.1.x releases, the latest is 1.1.11 - 1.1.12 is probably still a little around the corner. Be careful when linking it back to 1.1.11, as a number of modules are now correctly marked common - so if you get messages about module mismatch on link, upgrade your other servers.

We're also cooking up a few new modules, I wrote m_nickflood the other day. It provides channelmode +F - set as +F nicks:seconds. If more than nicks occur in seconds, then further nickchanges in that channel will be denied.

SVN news: 1.1 has been branched, and we've started work on 1.2. Not too much to see that's different yet, but there is a reasonable amount of work going on under the hood. We'll keep you posted on that. One neat feature that Om and Brain cooked up was functors - the ability to override some of the core behaviour of the IRCd. This will make things like a module allowing UTF-8 nicknames for 1.2 theoretically possible. :)

While we're talking about 1.2, I'd like to direct everyone's attention to the newspost on inspircd.org about how to help out -- we're really in need of people in all areas, so come talk to us :).

Lots more has happened, but is escaping my thoughts at present. I'll be sure to try to keep posting this time. ;p