08:12:52 <donri> i wonder if you could compile acid-state with ghcjs and pipe the remote backend over websockets
08:13:30 <donri> maybe just maybe it could make sense to add an authorization api to remote transactions :p
13:22:57 <donri> man i'd love to finish that patch for safecopy generics, but it's a bit beyond me :(
13:30:21 <donri> i'd like an instance in a library that i'd like inferred Safe, which rules out TH :p
13:30:39 <donri> ironically it's less safe to hand write these instances ;)
13:31:43 <donri> hm need to add Trustworthy to safecopy too
13:32:09 <donri> assuming it doesn't export anything unsafe
16:59:45 <stepkut> donri: :)
20:27:46 <donri> stepkut: hey looks like the fix for newtype deriving has landed in ghc
20:27:54 <donri> http://ghc.haskell.org/trac/ghc/wiki/Roles
20:29:14 <stepkut> yay ?
20:29:24 <stepkut> does this fix a problem I actually care about?
20:29:57 <donri> well it's supposed to make generalized newtype deriving safe and sound
20:30:27 <donri> and we rely on it a lot in idiomatic happstack code :)
20:33:00 <stepkut> seems you can't do, [hsx| <% [jmacro| ... |] %> |]
20:33:12 <donri> in deed
20:34:15 <donri> nor can you do [jmacro| `([hsx| ... :P
20:34:39 <donri> i think we discussed that long ago when we were working on hsx-jmacro
20:36:13 <stepkut> yeah
20:36:16 <stepkut> someone should fix that
20:36:30 <stepkut> i think you can use let to work around it ?
20:36:53 <donri> in the jmacro case it could be fixed by enabling XmlSyntax in its own splice parser which i think uses haskell-src-exts
20:37:05 <donri> let certainly works
20:37:11 <donri> it's just a syntactical problem
20:37:16 <stepkut> i suspect so
20:37:40 <stepkut> I am going to leave that section of the crash course using hsx2hs for now
20:38:24 <donri> the preproc also has the benefit of working in pattern matching context
20:38:39 <donri> though don't think the crash course uses that :)
20:39:37 <donri> maybe a Q Pat QQ could be made do work using view patterns perhaps?
20:39:55 <stepkut> dunno
20:40:16 <stepkut> After these docs are updated I will not be focusing on HSP for a while hopefully
20:40:41 <donri> \o/
21:45:55 <stepkut> hey.. i think the, read "1e100" :: Int, bug might be fixed?
21:48:43 <donri> > read "1e100" :: Int
21:48:44 <lambdabot>   *Exception: Prelude.read: no parse
21:49:57 <stepkut> \o/