Unreal3.2.7 release and important strategy change

[Carsten M. <cvm...@da...>]

[ Strategy change ] - This is important, please read.

Unreal3.2.7 will be released on 14th July 2007 - marking 8 years since 
the first UnrealIRCd release. With this, we've decided to do a rather 
big change in strategy. Fact is, the Unreal3 code base has gone stale 
and is not suitable to service your IRC networks for 8 more years.

We've been working to improve this code base but run into serious issues 
constantly due to old code laying about and design decisions in the past 
by previous coders that has locked us into an IRCd that can hardly evolve.

In light of this, and through that people have worked on a IRCd from 
scratch, in C++ - InspIRCd. InspIRCd was made by people who initially 
used Unreal, but foresaw that something radical was needed to fix the 
IRCd code base so many IRCds share.

InspIRCd was partly inspired by Unreal and this means that it will be 
easier for us to move on to the logical step - we're forking InspIRCd 
(through a tight cooperation with them), to give
you Unreal4.0. For more information on the new core, see 
http://www.inspircd.org - any documents
dealing with module coding, server protocol, etc, works for Unreal4 too 
- as we will be the project that extends from bare (InspIRCd) to 
colourful and featureful (UnrealIRCd) and the goal is to be able to move 
modules from Unreal to InspIRCd without problems.

The target is for both development teams to work what's their speciality 
- we're dedicated to give our users service, implement features and 
maintain stability and innovate, where InspIRCd also has some of these 
qualities, but focuses more on a base IRCd that can be extended
with new features through modularity and maintaining a stable core that 
just works and performs well.

The fork will not be a hostile one - the idea is to have a shared core 
(InspIRCd) and we work together on inventing and implementing new ideas 
that can be added through the module system, and we will work alone on 
features that InspIRCd coders may not want to touch and reverse.

We hope to shape InspIRCd in our image and provide an Unreal4 that feels 
like an Unreal3 - same configuration format, same quality documentation, 
support, etc, and in the progress give you users the ability to have a 
"modern" IRCd that can handle loads and loads of users, and still give 
you the choice to choose what features your network wants to use.

The Unreal3.2 branch will continue mostly as bug fixes - and adding 
ability to link to Unreal4, so we're not giving up on you people who 
want to keep on running Unreal3, but when this link is stable (in 3.2.8 
- based off 3.3), and Unreal4 is as feature complete with Unreal3, we 
stop supporting Unreal3.

So, how will this affect you? We could need a hand. Enter the 
development wiki on http://dev.unrealircd.com/unreal4_development - aid 
out and find out what we're missing in Unreal4 to make it an IRCd you'd 
switch to. Test the SVN, find bugs we introduce, and we hope this 
cooperation between InspIRCd and UnrealIRCd will lead to a new age in 
IRCd development and for IRC networks in general - since we have many 
new ideas we can try out together.

Hang out in #unreal-devel on irc.unrealircd.com - we're interested in 
having you along. Making modules should be slightly easier from now on 
as well.

[ New Website ]

Our webmaster, nate, has redesigned UnrealIRCd.com for us - if you have 
comments on the design or any aspect of the new website, please comment 
in the IRC channel or in the comments on this new item.


[ Unreal3.2.7 ]

As a minor sidenote, we've also released Unreal3.2.7, with some minor 
changes yet important - elaborated in Changes. The IRCd can be 
downloaded through http://www.unrealircd.com

- Updated c-ares to 1.4.0, TRE to 0.7.5
- chmode +L does no longer require chmode +l
- Oper blocks now can have CIDR, as in "userhost *@127.0.0.1/32";
- Services coders: SVSNOLAG/SVS2NOLAG (described in Changes) will allow 
a user to avoid fake lag (ie, flood as much as he/she wants).
- More intelligent accept() handling - that is, take in multiple times 
at a time instead of one per I/O loop
- A lot of bug fixes, basically.

[ Known issues found in testing ]

- CHROOTDIR does not work nicely together with the new patch where you 
can do hardcoded
IRC_USER / IRC_GROUP (setuid / setgid with names), due to having to look 
up password files.
- Spamfilter warn does not work with /setname
- Documentation is not complete
- There is a problem with bans getting truncated we are still trying to 
investigate. Happens in 3.2.6 too.

/Stskeeps