04:42:46 <stepcut> HugoDaniel: how about something like this, http://www.hpaste.org/fastcgi/hpaste.fcgi/view?id=16295#a16295
14:52:08 <HugoDaniel> stepcut
14:52:09 <HugoDaniel> thats nice
14:52:14 <HugoDaniel> and quite flexible
14:52:22 <HugoDaniel> but its not as easy as something like django
14:52:23 <HugoDaniel> w8
14:52:33 <HugoDaniel> http://docs.djangoproject.com/en/dev/topics/http/urls/
14:52:59 <HugoDaniel> the only thing with the django approach is that they have to specify the types, probably because python requires that
14:53:04 <HugoDaniel> in haskell that comes for free
14:53:12 <HugoDaniel> so it might be possible to do it better
14:55:25 <HugoDaniel> another approach is the bfg, which is also nice: http://docs.repoze.org/bfg/current/narr/urldispatch.html#path-pattern-syntax
14:57:46 <HugoDaniel> i just dont like using xml for routing :/
15:06:33 <stepcut> yeah
15:07:09 <stepcut> fortunately, adding new URL mapping schemes can be done trivially in 3rd party code.. so hopefully we can get some alternatives someday
15:08:29 <stepcut> I think it might be possible to implement something similar to the django one
15:09:58 <stepcut> or looking at your original example, something like this might be possible, "teacher-$id$/$day$-$month$-$year$" ->> myfunc
15:10:38 <stepcut> (->>) :: (MonadPlus m) => String -> (a -> m r) -> m r
15:10:52 <stepcut> except where 'a' is some varargs type thing
15:11:24 <stepcut> but I am not quite sure how to do it
15:12:04 <stepcut> similar to printf, http://cvs.haskell.org/Hugs/pages/libraries/base/Text-Printf.html
15:12:59 <HugoDaniel> i plan to start working on something like that
15:13:14 <stepcut> if you had, "$d$ - $d$" ->> \d1 d2 d3 -> ...
15:13:30 <stepcut> it would just never match, because the string only contains two ints, and the function requires three
15:13:39 <stepcut> but you can detect that at compile time..
15:13:56 <HugoDaniel> hmm, yes
15:13:59 <stepcut> yikes
15:14:07 <stepcut> you *can not* detect that at compile time
15:14:16 <HugoDaniel> even with template haskell ?
15:15:49 <stepcut> if you want to use template haskell, then you could just write something like, $(mkMatch "$d$ - $d$") and have it generate, (,) <*> int <* str " - " <*> int
15:15:54 <stepcut> or whatever it is supposed to generate
15:17:42 <HugoDaniel> hmm
15:17:52 <HugoDaniel> template haskell is temptative
15:18:04 <stepcut> :)
15:18:04 <HugoDaniel> another approach might be using something like dyre
15:18:17 <HugoDaniel> and use dynamic recompilation
15:18:25 <HugoDaniel> xmonad style
15:19:12 <stepcut> there are many possibilities ! just not enough hackers !
15:20:21 <stepcut> I personally prefer templating and url mapping solutions which utilize strong typing.. but we could definitely benefit from templating and url mapping solutions which would be friendly to ruby and python developers
15:23:09 <HugoDaniel> yes, strong typing is a great plus haskell gives us
15:23:49 <HugoDaniel> and using parsec ?
15:23:54 <HugoDaniel> maybe its a bit overkill...
15:25:17 <stepcut> why would parsec be overkill?
15:37:00 <HugoDaniel> it seems like a huge canon to kill a fly, i mean, using it to parse a single line
15:37:22 <HugoDaniel> probably regexp, like django does, is a more sensible choice
15:40:52 <stepcut`> in my experience regexp and parsec are not good substitutes for each other
15:41:19 <stepcut`> each does a job well, but the other job.. not as well
15:42:34 <stepcut`> what makes parsec huge and regexp not ?
15:44:14 <HugoDaniel> i dont know, it just feels like huge, being a full fledged parser
15:45:56 <HugoDaniel> i wouldn't define a grammar just to parse a single line... but im lazyyyy
15:46:56 <stepcut`> do { string "teacher-" ; int }
15:46:59 <stepcut`> ez :)
15:47:23 <stepcut`> don't have to remember where to put all those damn \ things
15:49:17 <stepcut`> also makes it easy to add structure..
15:50:42 <stepcut`> but, regex would be good for winning ruby and python converts, because they know it
15:51:20 <stepcut`> the parsec version gives you strong type checking though
15:52:55 <stepcut`> in your code, Parsec is just a few constructors and a few functions.. not really that huge
15:53:21 <stepcut`> probably the same as regexp
16:03:57 <stepcut`> hmm
16:09:41 <HugoDaniel> yes
16:09:53 <HugoDaniel> codewise parsec is very elegant
17:27:30 <sm> good to allow either
20:17:15 <HugoDaniel> http://www.coactivate.org/projects/deliverance/introduction
20:17:19 <HugoDaniel> i really like this app
21:46:17 <stepcut> koeien: is the buildbot offline?
21:46:36 <stepcut> koeien: also, 6.8 is failing now because readline.h is not installed..
21:56:00 <HugoDaniel> can someone help me to find out what encoding im getting in this http://hpaste.org/fastcgi/hpaste.fcgi/view?id=16323#a16323 ?
21:56:22 <HugoDaniel> 'á' is \195\131\194\161
21:57:18 <stepcut> what happens if you use toString from utf8-string
21:57:50 <HugoDaniel> im not able to use toString because this is a regular ByteString, and not a utf8-string ByteString
21:58:00 <stepcut> ??
21:58:20 <HugoDaniel> oh wait
21:59:02 <HugoDaniel> for á i get 'á'
21:59:28 <stepcut> I still fail to understand what you mean by 'regular ByteString' and 'utf8-string ByteString'
22:00:03 <HugoDaniel> sorry, it was my mistake
22:00:39 <HugoDaniel> because im not able to use the encode functions of utf8-string in bytestrings
22:00:46 <HugoDaniel> it complains about that in ghc
22:01:08 <HugoDaniel> but im not using encoding or decoding, this is straight from the database
22:01:40 <stepcut> one moment
22:02:14 <HugoDaniel> which is a mssql db, strangely if i set my konsole to utf8 and do a select on the db i get the same output that happstack is giving me (that is the 'á' for 'á')
22:02:41 <stepcut> well, perhaps you db is corrupt?
22:02:45 <HugoDaniel> i tried out diff encoding schemes on konsole, but to no success...
22:03:08 <HugoDaniel> hmm, i dont think so, because the characters show up properly if...
22:03:17 <HugoDaniel> let me see if i can find a case where they show up properly...
22:03:44 <HugoDaniel> i already did make them show up properly, just the page was all messed up
22:04:26 <HugoDaniel> ok, so
22:04:28 <HugoDaniel> on freebsd
22:05:05 <HugoDaniel> they show up all messed up no matter what, on linux they show up correctly if i change the browser from utf-8 to iso-8859-1 (this also makes all the other characters on the page to show up messed up)
22:06:16 <stepcut> ok, so the encoding is iso-8859-1 ?
22:07:19 <stepcut> what does the database say the encoding is ?
22:07:52 <HugoDaniel> no
22:07:58 <HugoDaniel> i still haven't found out
22:08:02 <HugoDaniel> i dont know
22:08:12 <HugoDaniel> w8, ill give you two examples of the same code
22:08:15 <HugoDaniel> running on diff OS
22:11:29 <HugoDaniel> okey, so the first is linux, arch linux
22:11:31 <HugoDaniel> http://dev.fcsh.unl.pt/a3es/
22:11:32 <HugoDaniel> here
22:11:36 <HugoDaniel> login with id: -100
22:11:39 <HugoDaniel> passwd: a
22:11:54 <HugoDaniel> then chose the 3rd link on the table, and go to the bottom of the page
22:11:59 <stepcut> proxy error
22:12:05 <HugoDaniel> oh :(
22:12:07 <stepcut> oh, worked this time
22:12:19 <HugoDaniel> hmm ok
22:12:33 <HugoDaniel> those the 3rd one, the "Adriano duarte rodrigues"
22:12:36 <HugoDaniel> and go to the end of the page
22:13:01 <HugoDaniel> the last table im getting the data from that weird encoding
22:13:08 <HugoDaniel> and the "á" shows up correctly
22:13:27 <HugoDaniel> now, the freebsd version, its the exact same code, on the same db
22:13:57 <HugoDaniel> http://serv.fcsh.unl.pt/a3es/
22:14:30 <HugoDaniel> oh well... :/
22:14:39 <stepcut> it looks to me like they are ISO-8859-1
22:16:24 <HugoDaniel> ok
22:16:55 <stepcut> is this some existing database you have no control over?
22:17:19 <HugoDaniel> sort of
22:17:29 <HugoDaniel> this data comes from diff databases
22:18:06 <HugoDaniel> the login part comes from one db, the middle part from another, and the final table in the bottom cames from another one... i can't change these db's
22:18:57 <stepcut> then you probably need to use iconv
22:18:59 <stepcut> http://hackage.haskell.org/package/iconv
22:19:16 <stepcut> and convert the ISO-8859-1 stuff into a String
22:19:34 <HugoDaniel> then i have my own db, where i can change stuff around, this form is stored in my db, the next part is generate some pdf's from it... but thats another war
22:20:08 <HugoDaniel> allright
22:20:14 <HugoDaniel> this seems just the tool i need, thanks :)
22:20:25 <stepcut> I would use iconv to convert the strings from the ISO-8859-1 database to utf-8, and then toString to convert them into a String
22:20:43 <stepcut> and then everything should work automatically from there
22:21:02 <stepcut> and then flame the people that created the ISO-8859-1 database ;)
22:24:16 <HugoDaniel> stepcut: thats my boss :)
22:25:31 <stepcut> even better !
23:03:49 <HugoDaniel> its not iso-8859-1
23:03:49 <HugoDaniel> but oh well
23:03:49 <HugoDaniel> tomorow ill try to put somebody correcting all those special characters... :/