13:05:45 <McManiaC> stepkut: got that patch with getDataQuery/BodyFn
13:06:05 <McManiaC> gonna send it in a minute to the ML
13:09:34 <stepkut> McManiaC: sweet!
13:09:41 <stepkut> I'll apply it this afternoon
13:11:30 <McManiaC> I've added this for simplehttp and rqdata.hs now
13:14:28 <st3pcut> cool
13:14:50 <st3pcut> I working on getting RqData completely haddock'd up, and then I am going to make the switch
13:18:19 <McManiaC> ok
13:18:20 <McManiaC> mail is out
13:18:21 <McManiaC> :)
13:19:06 <bonobo> hi, will there be any changes to RqData?
13:19:45 <McManiaC> only additions
13:20:08 <bonobo> I'm interested!
13:21:06 <McManiaC> basicly, the current getData(Fn) just looks into *all* input fields (that is QUERY + POST + Cookies) if there is anything we'd like to
13:21:21 <McManiaC> so this patch adds the possibility to distinguish between post, get etc
13:21:35 <bonobo> perfect!
13:24:09 <st3pcut> bonobo: in darcs there is a new RqData module (not yet fully integrated). It includes plural versions of all the look functions (so you can find all the matches, instead of just the first match)
13:24:54 <bonobo> st3pcut: I need that once so I wrote my own version of those
13:25:29 <st3pcut> also, if you use the applicative instances, it will accumulate any failed matches and give you a report
13:25:46 <bonobo> hmm, how to use this?
13:25:49 <st3pcut> which is useful during development, but also useful if you provide a REST api and want to return a useful error message
13:26:14 <st3pcut> right now you explicitly import Happstack.Server.RqData and hide the instances from SimpleHTTP
13:26:30 <st3pcut> later this week (hopefully), you will just import SimpleHTTP like normal
13:27:06 <st3pcut> the new version should be largely backwards compatible
13:27:38 <st3pcut> except the type signature for getData* and withData* is Either instead of Maybe
13:28:21 <st3pcut> gotta shower and run. bbl.
21:06:47 <McManiaC> stepcut: had some time yet to take a look at my patch? :)
21:06:58 <stepcut> McManiaC: no, just got back
21:07:02 <McManiaC> kay
21:07:04 <stepcut> but I had some thoughts about it anyway
21:07:33 <stepcut> it seems like it allows you to either get stuff from the body or from the query string? But what if you want to get things from both, but specify what location to look ?
21:08:00 <McManiaC> how do you mean?
21:08:03 <stepcut> for example, get the value 'x' from the QUERY_STRING, and the value 'y' from the POST data
21:08:52 <stepcut> do x <- lookQuery "x"
21:08:52 <stepcut>    y <- lookBody  "y"
21:08:52 <stepcut>    return (x,y)
21:08:57 <stepcut> something a bit like that ?
21:10:19 <McManiaC> hmm
21:12:15 <stepcut> I don't think that creating lookBody, lookBodyRead, lookupBodyInput, lookBodys, etc. is a great plan, because then we will have dozens of those damn things
21:12:27 <stepcut> there are already too many look functions
21:12:52 <stepcut> I wonder if we can make it simpler and more flexible at the same time
21:13:02 <stepcut> (even if it is not backwards compatible)
21:18:32 <McManiaC> maybe you could change RqData from a "([(String, Input)], [(String, Cookie)])" into something more "complex" where you could use some kind of modifiers or sth
21:18:55 <McManiaC> do x <- body $ look "x"
21:19:00 <McManiaC>    y <- query $ look "y"
21:19:06 <McManiaC>    return (x,y)
21:20:14 <McManiaC> not exactly sure how that rqdata stuff works
21:20:14 <McManiaC> :>
21:23:03 <stepcut> I think RqData is fine as it is, just the lookup functions need to be changed..
21:23:17 <aavogt> what of laziness?
21:23:28 <aavogt> (in the parsing of multipart data)
21:24:04 <stepcut> aavogt: it's trouble
21:24:11 <aavogt> to add it?
21:24:31 <aavogt> because if there's any there, it's not enough to be useful
21:24:47 <stepcut> aavogt: you can't have laziness and lookup at the same time really.. if you try to look up a value that does not exist, you have to look through all the post data to tell that
21:35:17 <stepcut> after the timeout stuff is fixed, I want to look at the issue of parsing post data and uploading files
21:35:40 <McManiaC> stepcut: http://npaste.de/Wc/ maybe like this? no Idea if this even compiles, just a quick scetch
21:35:54 <McManiaC> (and I forgot the obvious return :) )
21:36:23 <stepcut> yeah