16:38:34 <stepkut> donri: you said the default BotConf should have some "." instead of "" ?
16:39:04 <donri> stepcut: you already pulled my patches for that
16:39:12 <stepkut> :-/
16:39:26 <donri> and it's the nullUser i think
16:39:33 <stepkut> ah
16:39:35 <stepkut> that explains it
18:08:15 <stepkut> hmm, there is a lot of state-machine like stuff in an ircbot
18:17:16 <stepkut> not sure how to model that more purely
18:21:43 <stepkut> like.. you can't join a channel until you have completed nick/user registration. But that is a multi-step process which requires checking for error codes, etc
19:04:56 <stepkut> #dice 3d6+5
19:04:56 <stepbot> You rolled 3 6-sided dice with a +5 modifier: [6,3,4] => 18
19:08:32 <stepkut> donri: I uploaded a new ircbot that should do a better job of registering the nick/user and joining channels
19:08:53 <stepkut> donri: but you will need to change your init code to include some extra parts now
19:08:57 <stepkut> donri: see the updated demo.hs
19:09:02 <donri> roger
19:09:17 <stepkut> needs more love still, but that is all I have time for today
19:10:30 <stepkut> so, a normal BotPartT only gets run when an irc message arrives. But sometimes you things to happen for other reasons..
19:10:59 <donri> ah yea that's another problem
19:11:04 <donri> if you have something like hackagebot
19:11:11 <stepkut> yeah
19:11:30 <stepkut> the new channelsPart sort of handles it.. but it could probably be a lot better
19:17:05 <donri> hm, is there any point to using text with web-routes?
19:17:57 <stepkut> as opposed to what?
19:17:59 <donri> string
19:18:16 <stepkut> it used to use String :)
19:18:22 <donri> i got an error for text not having a PathInfo instance
19:18:27 <stepkut> but Text is more better..
19:18:30 <donri> which i guess wouldn't be too difficult to add?
19:18:44 <donri> or could even use derivePathInfo?
19:18:59 <donri> i'll probably get a warning for orphan instance right
19:19:34 <stepkut> there are already PathInfo instances for Text and [Text]
19:19:37 <stepkut> oh
19:19:52 <stepkut> but not Lazy Text
19:19:56 <stepkut> perhaps that is your problem?
19:20:05 <donri> don't think i tried to use lazy
19:20:35 <stepkut> oh I see the problem
19:20:42 <donri> web-routes doesn't even depend on text
19:21:18 <donri> bbl dinner
19:21:20 <stepkut> 0.25 doesn't
19:21:22 <stepkut> 0.26 does
19:21:27 <donri> ah :)
19:21:31 <donri> i'll wait then
19:21:34 <stepkut> just uploaded it
19:21:43 <donri> nice!
19:22:06 <stepkut> gotta upload the whole family though
19:29:35 <stepkut> ok, all uploaded
20:03:17 <donri> is -regular an alternative to -th?
20:03:54 <stepcut> yes.. using generics instead of template haskell
20:04:01 <donri> which do you recommend?
20:04:04 <stepcut> -th
20:04:12 <donri> motivation?
20:04:32 <stepcut> why wait till runtime to do what you can do at compile time?
20:04:37 <stepcut> the regular library uses TH anywy
20:04:47 <donri> oh
20:05:03 <stepcut> I think -th and -boomerang are probably the best options
20:11:21 <donri> looks like implSite wants Text now too? may need to udpate crashcource
20:13:41 <donri> also had to change implSite "/" "" to implSite "" ""
20:14:17 <donri> all working with text now \o/
20:15:36 <donri> oh there's already "pack" in the crash cource
20:15:40 <donri> OverloadedStrings works too
20:25:46 <stepcut> yup
20:26:05 <stepcut> but OverloadedStrings often causes problems because it makes things like, toResponse ambiguous
20:26:57 <donri> ok
22:19:25 <donri> stepcut: why isn't happstack-extra on hackage?
22:19:39 <donri> waiting for happstack 7 maybe?
22:19:39 <stepcut> donri: because it is evil?
23:11:29 <Entroacceptor> stepcut: again, you can authenticate on connect ;)