01:07:06 <stepcut> KBme: happstack should support http 1.0 clients. if it doesn't it is a bug
05:30:39 <rlpowell> stepcut: Random FYI: updateIx will update *or* insert on the given key.  So the PasteID == 0 stuff in the CtrlV example isn't really necessary.
05:30:53 <rlpowell> I checked teh code; it works be deleting and then inserting.
05:37:49 <stepcut> uh
05:38:30 <stepcut> in insertPaste?
05:41:11 <stepcut> when PasteId is 0, we give the paste an entirely new id and increment the nextPasteId counter, and add the new entry to the database.. when it is not zero, we just update the existing entry in the ixset.
05:41:45 <stepcut> we still need to check for the case when the PasteId is 0 and give the paste a new Id..
05:54:33 <stepcut> I'm going to sleep while you think about that :p
06:59:52 <rlpowell> stepcut: You can just give the new paste nextPasteId, and run updateIx on that.  I find that more natural, but it's a matter of taste I suppose.
10:59:29 <LambdaDusk> who can I ask about more tags for refrom-hsp? just <div> would be great
11:00:34 <LambdaDusk> and setting class and ids of these elements
13:13:12 <donri> another thought re: our use of mtl-style for pluggability: we use a lot of generalized newtype deriving to make this not suck, but newtype deriving isn't allowed by SafeHaskell, which we might like to use for e.g. CMS plugins
13:13:41 <donri> not sure it matters for that (e.g. maybe all the deriving happens in trusted packages) but worth thinking about
14:51:10 <stepcut> rlpowell: yes, but we also have to increment the nextPasteId counter… seems clearly to me to have to distinct branches because things are fairly different rather than try to unify things so that you can merge the two code paths after the branch ?
14:52:35 <stepcut> donri: I'm hoping they just fix newtype deriving. I think the SafeHaskell speaker at ICFP said that they know how
14:53:18 <stepcut> but, in general, I would like to consider alternatives to the leaning tower of newtypes
15:48:37 <rlpowell> *snrk* at that last.
18:24:09 <donri> hs-suite slides http://www.haskell.org/wikiupload/9/94/Hiw2012-niklas-broberg-haskell-suite.pdf
18:32:13 <rlpowell> donri: Did you see stepcut's last comment to you?
18:32:33 <donri> ah, no, reading now
18:33:53 <rlpowell> stepcut> donri: I'm hoping they just fix newtype deriving. I think the SafeHaskell speaker at ICFP said that they know how stepcut> but, in general, I would like to consider alternatives to the leaning tower of newtypes
18:34:49 <donri> rlpowell: hey btw ghc 7.8 is getting "type holes", which makes underscore an expression of any type, and throws a warning at compile-time showing the inferred type. you might find that useful when confused what type something should be.
18:35:22 <rlpowell> Neat, thanks!
18:35:26 <rlpowell> When's that out?
18:35:50 <donri> hm, late 2012 or early 2013 i think they said
18:36:00 <rlpowell> 'k.
18:36:00 <donri> 7.6 just came out
18:36:03 <rlpowell> *nod*
18:36:27 <donri> https://github.com/ghc/ghc/commit/8a9a7a8c42da3adb603f319a74e304af5e1b2128
18:39:50 <donri> stepcut: is this related to the fixable issues? http://joyoftypes.blogspot.se/2012/08/generalizednewtypederiving-is.html
18:42:34 <donri> bbl dinner
19:41:27 <stepcut> donri: no idea
19:58:38 <donri> stepcut: did you and nibro get to do anything about hsp?
19:59:26 <stepcut> nope
20:00:01 <stepcut> he was only there for a short time *and* got sick :)
20:00:07 <donri> aha
20:02:28 <stepcut> and my hard drive crashed
20:03:30 <donri> new logging package looks at a glance potentially like not complete utter crap http://hackage.haskell.org/package/simple-log
20:04:04 <stepcut> :)
20:04:55 <donri> although...
20:05:34 <donri> it seems to work by putting this IO-centric datatype in a ReaderT https://github.com/mvoidex/simple-log/blob/master/src/System/Log/Base.hs#L169
20:05:50 <donri> which i thought was better than having global iorefs, but still IO-centric...
20:06:32 <donri> would be easy to make that parametric on the monad though
22:00:01 <Lemmih> donri: Did I apply your patch?
22:00:38 <donri> Lemmih: not that I know, or can see with darcs changes
22:34:03 <donri> bedtime