02:49:03 <stepkut> McManiaC: yeah I think something like what you have, x <- body $ lookRead "foo", should work
02:49:09 <stepkut> where, body :: RqData Input -> RqData Input
02:49:26 <stepkut> and body locally rewrites the RqData so that it only include values from the body
02:49:39 <stepkut> or query_string, or whatever
02:50:00 <stepkut> I'll hack something up tomorrow
02:50:29 <stepkut> body :: RqData a -> RqData a, rather
06:12:15 <Gracenotes> is there a means to create state that's only persistent within a session? I'm thinking of reverse dependencies, actually - can set them up at the beginning, update them throughout, and throw them away at the end
06:13:20 <Gracenotes> the other solution is to make a global IORef, or an effectively global one, and adorn it with various concurrency safeguards. but macid already does that...
06:14:19 <Gracenotes> and does a good job too
06:24:07 <Gracenotes> of course there's not much harm keeping it and doing an onLoad, perhaps
15:19:12 <stepkut> hopefully the bot will auto-reconnect more reliably now
18:11:07 <stepcut> McManiaC: I pushed a patch for the query_string vs body stuff. But it will certainly conflict with your patch.. so you'll want to rollback your local patch or something..
20:48:07 <Gracenotes> does happstack have any functions for manipulating/matching on uri fragments?
20:49:06 <Gracenotes> or.. if that's even possible... I need to read up on how browsers deal with them nowadays
20:51:44 <Gracenotes> right, they don't. never mind. :/
20:51:57 <aavogt> Gracenotes: this is for routing urls?
20:52:10 <aavogt> to make sure that a link you produce actually leads somewhere
20:53:01 <Gracenotes> well, I mean http://...#thesethings
20:53:51 <Gracenotes> and, I almost forget, normal web browsers just send a request without them and then try to match
20:54:27 <Gracenotes> i.e. look for an anchor name or html element with their id... which is somewhat useless if you're not actually using html
21:50:03 <stepkut> Gracenotes: yeah, they don't :)
21:51:24 <stepkut> Gracenotes: in theory, you can seek into various media, http://www.w3.org/TR/media-frags/
21:52:02 <Gracenotes> I just thought it would be nice to POST url#fragment
21:52:13 <Gracenotes> and have the server see *that*
21:52:23 <stepkut> yeah
21:52:51 <Gracenotes> can it? I'd expect the browser to just use fragments on the resulting document, alas..
21:53:10 <stepkut> as far as I know, the browsers never send the #fragment
21:54:19 <Gracenotes> after all, if you can view a specific portion of a document as url#distros, you'd expect to modify that specific portion of the document by posting to url#distros
21:58:22 <stepkut> A fragment identifier is only meaningful when a URI reference is
21:58:22 <stepkut>    intended for retrieval and the result of that retrieval is a document
21:58:22 <stepkut>    for which the identified fragment is consistently defined.
21:58:33 <stepkut> from rfc2396 section 4.1