RE: [NeoStats-Devel] NeoStats 3.0 IRCd support

Going back a bit but some additional thoughts for consideration...

Justin Hammond 
> There is still about 20 networks on secure.irc-chat.net using 
> Unreal 3.1...
> I think maybe worthwhile keeping it around for a while longer 
> at least...

IIRC there are still a number of networks using a very old SecureServ
release as well. Until you actually implement the "cutoff" these will likely
remain on that version and the same with the NeoStats core. As such,
Unreal3.1 would remain available in 2.5.x but as an unsupported IRCd do we
really want to launch 3.0 with unsupported IRCds? 

I know of networks that still use early Unreal3.2 betas in preference to the
final versions since it still had some form of the total invisibility stuff.
I would never dream of trying to support the 20 odd beta versions of 3.2 as
well as the 3 release versions so I think we should consider this wrt
Unreal. 

Personally I think we drop Unreal 3.1 from the 3.0 codebase unless and until
there is a proven demand for it. Unreal 3.1 users still have a NeoStats
solution in the 2.5.x branch and if they still have not upgraded to Unreal
3.2, why would they upgrade to NeoStats 3.0?

> Ultimate2 - Also still being distributed and used, though not 
> widely, this might change, as they just made the first 
> offical win32 port on Ultimate2.... Its like NeoStats 2.5, 
> its in feature freeze mode... But users are encouraged to 
> move to 3.0..

Again I think this similar to the Unreal solution from our point of view and
based on what Eggy said, if we dropped support in NeoStats 3.0 for Ultimate
2.x, this would likely please ShadowMaster since it would offer another
incentive to upgrade. Again existing users have a path via NeoStats 2.5.x
and since Ultimate is likely to go to 3.0 final before us and will likely
drop support for 2.x at that time, we would be supporting an unsupported
IRCd so I think for 3.0 we drop and only provide if there is a proven
demand.

> I think its best to aim to cover the major IRCd's 
> available... Hybrid, Unreal, Bahamut, Ultimate, IRCu as most 
> other IRCd's are based off one of these, so porting new 
> IRCd's should be easier. (eg, the Plexus one should be 
> easier... If we have the base coded first!)

I agree on this but it seems counter productive to provide support for known
unsupported ircds when we do not have an active user base on them so no real
testing capability. Those I have removed so far are in the graveyard for
future reference and Unreal3.1 and Ultimate2.x could be moved here as well
for any future need. 

Once 3.0 goes final, 2.5.x will join the ranks of the unsupported but will
still include support for the now unsupported ircds. I would rather we drop
support in 3.0 since anyone who is forced to download 2.5.x to use an
unsupported ircd will likely request support for it in 3.0 so at least we
can be guaranteed feedback and testing of it rather than just hoping it will
perform in a production environment. 

Thoughts?

Mark.