13:40:51 <mekeor> weird. there's no instance for (ToMessage Html). why not? -- that's really annoying me atm :/
13:44:57 <mekeor> please help ;-(
13:50:23 <mekeor> the error i get is:   No instance for (ToMessage Html)arising from a use of `toResponse' 
13:51:29 <mekeor> ANY ideas?
14:01:21 <mekeor> i mean, i don't understand this. i didn't modify the code of my happ and just tried to compile it when i got this error. but why?
14:38:22 <stepkut> mekeor: you likely have installed blaze-html 0.5, but happstack server is built against 0.4
14:38:29 <mekeor> ah!
14:38:56 <mekeor> how can i use a specific version?
14:39:07 <mekeor> i have both 0.4.* and 0.5 installed.
14:39:14 <stepkut> it is easy to updated happstack-server to support blaze-html 0.5.. but it causes a chain reaction that would force us to upgrade to disgestive-functors 0.3
14:39:40 <stepkut> you can specify the version in the .cabal file or use -hide-package blaze-html-
14:39:44 <stepkut> or whatever the version number is
14:40:24 <mekeor> "-hide-package blaze-html-" for GHC?
14:40:28 <mekeor> in command-line?
14:40:29 <stepkut> yeah
14:40:37 <stepkut> i think you can put, :set -hide-package blaze-html-, in a .ghci file in the directory and then it will be added automatically
14:40:54 <stepkut> looks like the version is just
14:41:05 <mekeor> yes.
14:41:08 <stepkut> you might also need to hide blaze-markup
14:41:29 <stepkut> alterantively you might be able to use qualified package imports
14:41:52 <stepkut> import "blaze-html-" Text.Blaze.Html
14:42:05 <stepkut> needs, {-# LANGUAGE PackageImports #-}, or something like that
14:42:10 <mekeor> stepkut: thank you very much!!!! it work yippiee yaaaay
14:42:17 <stepkut> or.. you could uninstall blaze-html
14:42:51 <mekeor> stepkut: thank you, really! =)
14:42:52 <stepkut> I am trying to solve the underlying problem that is preventing us from upgrading to so hopefully soon
14:43:09 <mekeor> awesome. great. super, brilliant! yayayayay
14:44:26 <donri> stepkut: can't we support a range including 0.4?
14:45:01 <stepkut> nope. the API changed
14:45:05 <donri> ah
14:45:19 <stepkut> hence the bump in the .Y component :p
14:45:32 <donri> true
17:10:55 <donri> stepcut: what is the plan re df anyway?
19:01:25 <stepcut> donri: I am wrapping up a fork of 0.2
19:03:32 <donri> stepcut: so no view separation eh
19:04:07 <stepcut> donri: no.. the fork includes view separation .. one reason it has been taking a while is making that was supportable with both blaze/hsx and also heist
19:04:21 <donri> oh, you solved that type safely?
19:09:51 <stepcut> of course!
19:10:09 <donri> nice
19:10:25 <donri> jasper made it sound all difficult
19:11:28 <stepcut> I just make it look easy :p
19:11:42 <donri> :)
21:37:06 <stepkut> ok.. I need a name for types that are isomorphic to IdentityT/Identity but are specifically intended to be used with HSP to generate XML
21:44:52 <donri_> Xml vs XmlM?
21:45:06 <donri_> XmlT
21:45:31 <stepkut> we already have a type named XML
21:45:36 <stepkut> having Xml and XML seems a bit confusing
21:45:40 <donri_> haha
21:45:54 <stepkut> we used to have *two* types named XML.. but now we are down to one :)
21:45:57 <donri_> XMLGenerator... ohwait
21:46:06 <stepkut> GenXML is also taken already
21:46:14 <donri_> XMLGen!
21:46:18 <stepkut> AnXMLMonadTransaformer is too long
21:46:23 <stepkut> no, there is an XMLGen class
21:46:26 <donri_> ;)
21:47:38 <donri_> maybe rename XMLGen to MonadXML
21:47:54 <stepkut> heh
21:48:21 <donri_> backwards compat was thrown out long ago anyway
21:48:28 <stepkut> so.. here is something interesting.. trhsx/hsx-qq do not really depend on the XMLGen class
21:48:52 <donri_> ah yea just what's in scope
21:48:56 <stepkut> while trhsx does generate specific names like, genElement, asChild, asAttr, etc.. there is no requirement that those names map to that type class
21:49:19 <stepkut> that could make it easier to use trhsx formlets-like stuff
21:49:22 <donri_> i think the do notation translation works similarly? just whatever >>= is in scope?
21:49:39 <stepkut> only if NoImplicitPrelude is enabled I think
21:49:46 <stepkut> or some extension like that
21:50:53 <donri_> (kinda serious about MonadXML, or is it semantically wrong? having XMLGen and XMLGenerator is just damn confusing, and now we're talking about more "xml generating types"...)
21:51:30 <donri_> also XMGenT is something else entirely right?
21:51:34 <stepkut> yes
21:51:50 <stepkut> it is isomorphic to IdentityT.. but plays a different role
21:53:07 <donri> I think: rename XMLGen to MonadXML; make XMLGen the Identity version of XMLGenT; possibly rename XMLGenerator to something more meaningful (not sure)
21:53:22 <stepkut> in order to embed a child into a parent you need an instance like, instance (EmbedAsChild m c, m ~ n) => EmbedAsChild m (n c). But that instance is so generally it would then overlap with everything
21:53:55 <donri> (maybe rename XMLGenerator to simply HSX in lieu with the Happstack class)
21:55:05 <stepkut> unfontunately, my proposal to remove undecidable instances is flawed
21:55:17 <donri> overlapping you mean?
21:55:24 <stepkut> it didn't support simple things like, <p><% "foo" %></p>
21:55:34 <donri> i'm not sure about the difference between overlapping and undecidable
21:56:39 <stepkut> yeah.. i guess it was the overlapping that was causing trouble
21:56:46 <stepkut> or maybe both
21:58:14 <stepkut> in theory, we might be able to replace the XMLGenerator (or Happstack) class with a type constraint synonym now...
21:58:16 <stepkut> http://www.haskell.org/ghc/docs/latest/html/users_guide/constraint-kind.html
21:58:54 <stepkut> type Happstack m = (Functor m, Monad m, MonadIO m, etc)
21:59:57 <stepkut> not sure what the advantage is aside from not having to explicitly write, instance Happstack (AppT m)
22:02:59 <donri> and it requires ghc 7.{2,4} something
22:03:03 <stepkut> yup
22:03:59 <stepkut> hmm. we can't make XMLGen the Identity version of of XMLGenT.. XMLGenT and the XMLT monad transformers serve different purposes
22:05:14 <donri> :/
23:21:10 <donri> stepkut: oh hey gabriel pushed stuff to pipes repo a week ago https://github.com/Gabriel439/Haskell-Pipes-Library (including much older commits)
23:21:40 <stepkut> interesting
23:21:49 <stepkut> i wonder what that means
23:23:13 <stepkut> seems unlikely that the version number 1.0.2 follows PVP
23:23:32 <donri> you mean the pipes on hackage?
23:23:50 <stepkut> oh.. duh
23:24:03 <stepkut> I guess he has not bumped the version in github yet
23:24:38 <donri> i think it's still in development
23:24:44 <stepkut> right
23:24:54 <stepkut> i would have expected to see something on reddit if it was released
23:31:21 <donri> bedtime!
23:31:30 <stepkut> good night!