Resolver/identd will be done as separated modules in 3.1 (future release), added some TODOs for discuission

This commit is contained in:
Valery Yatsko 2008-04-11 00:39:49 +04:00
parent 73d6283cfc
commit a2001bb371

25
TODO
View file

@ -1,16 +1,14 @@
/ = in progress, x = done, ? = to be discussed, F = for future releases
/ = in progress, x = done, ? = to be discussed, F = charybdis3.1 or next releases
[/] finish legacy code removal
[x] remove 2.8 report_error() in ratbox imported stuff
[F] client.c, channel.c is very 2.8 style still. it'd be nice to pack them into their own
namespace and such. moreover, the other 2.8 code needs similar rewriting/reworking too...
[ ] rewrite s_auth.c -> split into three distinct parts
[ ] authentication state/lock manager
[ ] identd check
[ ] dns check
### this should be a nice framework to let modules hook the
### auth state imo. it'd also be cool to make the DNS checker and
### auth checker modules (loaded by default, of course). --nenolod
### well, I suggest working on it in 3.1 if you're not against ;)
[F] rewrite s_auth.c
[F] authentication state/lock manager
[F] move resolver/auth checker code into separated modules
(or maybe give user ability to run them as ratbox3-style helpers (./configure option)?)
[x] port to libratbox
[x] get it running
[x] clean up maxconnections kludges &c
@ -27,8 +25,15 @@
[?] +C (noctcp) channel/usermode
[ ] gk:line/akill syncing
[ ] drop non-TS6 (legacy protocol) support
[ ] module engine rework
[ ] more beautiful way of adding new channel modes by module
[F] module engine rework
[F] more beautiful way of adding new channel modes by module
[F] make nick/user/host validation functions/match tables able to work in separated modules,
this will help us making support for native characters sets/slashes in host etc
[F] auth checker module
[F] resolver module
[?] g:lines
maybe remove them completely or rework them a bit - ircu-style glines?
I don't think if anybody use them in the networks different from EFNet
[ ] other stuff
[?] internally split out +o/+v "ranks" into a series of permissions. this could allow for configure-defined
special access levels, halfops, etc. (would need to match globally, somehow. extra SVINFO param?)