Experimental IRC log happs-2007-07-16

Available formats: content-negotiated html turtle (see SIOC for the vocabulary)

Back to channel and daily index: content-negotiated html turtle

These logs are provided as an experiment in indexing discussions using IRCHub.py, Irc2RDF.hs, and SIOC.

14:59:50<shapr>whee!
15:00:00<shapr>blackdog_: How's HAppS treating you?
15:01:54<kaol>shapr: could we have common log format logging support for happs? (as in http://www.bacuslabs.com/WsvlCLF.html)
15:01:55<lambdabot>Title: Common Log Format
15:02:27<kaol>I would provide a patch if Haskell didn't still seem so scary to me
15:02:39<shapr>Go for it!
15:02:40<kaol>and probably will when I get that far
15:02:42<shapr>Patch it up!
15:03:23<kaol>Haskell's already a lot less scary than what it used to be :-)
15:03:49<kaol>someone must have made it easier while I wasn't looking
15:03:59<shapr>hiya Lemmih!
15:04:03<Lemmih>Yo!
15:04:11<Igloo>Haskell is just a lovely fluffy bundle of joy
15:04:12<shapr>How's life in the big city?
15:05:04<shapr>ACTION cheers!
18:56:29<shapr>hiya msouth
18:56:32<msouth>hi
18:56:39<msouth>newbie questions ok here?
18:56:47<shapr>Sure
18:58:51<msouth>can you store data from happs into an external database? I mean, I know it has its own storage mechanism, but is it able to, say, put stuff into mysql?
18:59:20<msouth>or is that a "why use happs if you want your data in mysql" type question
18:59:33<shapr>Er, sort of yes to both :-)
18:59:45<shapr>You can put stuff in an external database.
19:01:23<kaol>and get it from there too
19:02:05<shapr>But as you also mentioned, one of the ideas behind HAppS is to get rid of the marshalling problem.
19:02:25<shapr>I have a HAppS user question...
19:03:10<shapr>Lots of big changes coming up in HAppS soon, do you want those changes incrementally released as they are completed...
19:03:14<shapr>or do you want them all in one big chunk?
19:04:39<msouth>personally I'm too new to know, sorry.
19:05:47<msouth>are there apis that let you talk to the data in happs from a language like perl? Or would you just do that by exposing REST or something?
19:05:55<kaol>I'm tempted to leave my attempts at trying to learn to use HAppS until later, since it still seems to be very much alive...
19:06:45<shapr>kaol: Ah, but why wait until it's dead? Don't you want a chance to influence the design? :-)
19:07:53<kaol>like I would have anything to contribute in that regard
19:08:48<shapr>Well, sure you would.
19:09:37<shapr>If you use the code, you'll see things you wish were different. Then you either fix it yourself, or tell us what bugs you.
19:10:04<kaol>and even if I got something to work, I suspect that it would cease doing so as soon as I did a darcs pull
19:10:31<shapr>That's exactly the problem I'm asking about...
19:10:53<shapr>Is it better to get lots of incremental changes, or is it better to get a few massive breakages?
19:11:26<msouth>another question--what kind of scaling capability is there? Can you throw another machine at it and have that machine start taking up load?
19:12:38<msouth>shapr: if I was a business using it right now, I think I would want the big bang, so I could plan for it once, convert what needed to be converted, and be done with it.
19:13:00<msouth>there's overhead involved in planning a release that I wouldn't want to repeat.
19:13:31<shapr>Yeah
19:13:59<tuukkah>shapr, can you publish incrementally in a branch?
19:18:09<shapr>That sounds like a good option
19:24:12<kaol>anyone want to comment on some noncompiling code? I'm afraid that I'm using the snapshot that Debian has... Hope that won't matter. http://hpaste.org/1715
19:24:53<Igloo>What's the error?
19:25:02<Igloo>Your indentation looks wrong in the paste
19:25:11<Igloo>Oh, no, I misread it
19:25:13<kaol>Couldn't match expected type `() -> m (Either Request (im Result))' against inferred type `Either request (IO Result)'
19:25:33<Igloo>Where?
19:25:51<kaol>In the expression: respond (...)
19:26:34<Igloo>Hmm, not sure OTTOMH
19:27:40<kaol>perhaps I should have still tried to swap things around to see if the types would click together...
19:28:19<kaol>I'm guessing that I'm not even making clear what I'm trying to do, either :-/
19:44:54<Saizan>you've forgotten a monad layer, so to speak
19:46:45<Saizan>oh, no
19:47:54<Saizan>where does the error occurs?
19:50:49<kaol>it's at respond, and from that until the toMessageM i
19:51:08<Saizan>ah! it's your use of val
19:51:33<Saizan>val :: (Monad m) => response -> () -> () -> m (Either request response)
19:52:40<kaol>it's basically a placeholder... I suppose I should replace them with some XML replies already
19:52:42<Saizan>val is supposed to be used directly as an handler, to it takes two () as the uri and request inputs, but you've yet "consumed" those in the declaration of respondInfo
19:53:09<Saizan>you could just use strings there,
19:53:20<kaol>would it be ok if they all were something like plain_xml $ toMessageM foo?
19:53:55<Saizan>toMessageM is the last thing you call, usually
19:54:06<kaol>I should go back to studying the types a bit closer...
19:55:02<Saizan>if you want XML responses just use types that are instances of ToElement and call toMessageM on them
20:00:04<Saizan>or if you have the xml as string just use toMessageM and add the right content-type header (i'll make a newtype and a new instance of ToMessage in this case)
20:00:14<Saizan>s/i'll/i'd/
20:02:15<Saizan>we should make some graphs to explain all these possible conversions, maybe
20:02:41<kaol>make them kaolproof :-P
20:02:53<kaol>(that's a tall order)
20:02:55<kaol>ACTION sighs
20:06:35<Saizan>well, if there's something that need to more clarification the best is to ask
20:13:49<kaol>ACTION goes back to read SimpleHTTP2.hs again instead of asking clueless questions
20:52:10<kaol>yay. thinking about types got me far enough to get it to compile, at least. (still at http://hpaste.org/1715#a1 if anyone cares)
21:01:44<kaol>it's regular haskell code that's giving me grief now
21:08:41<kaol>hmm... does HAppS already process the query parameters in a GET? (ie. the stuff after the ?)
21:14:17<tuukkah>it does
21:17:57<kaol>and how would I access them?
21:18:19<tuukkah>the tutorial shows it using FromMessage and POST
21:18:48<kaol>ok, the same way then
21:19:18<tuukkah>we like lookM though
21:20:02<tuukkah>directly on the implicit ?req =)
21:21:22<Saizan>you can also retrieve an association list or a Map through functions in MessageWrap
21:24:05<kaol>wait, what. I'd like to see the HTML form used to issue those POST query for the example in 3.3. You can't make a POST simply by entering an URL, after all...
21:24:58<kaol>s/query/queries/
21:25:39<Saizan>kaol: the look* functions work on both POST and GET
21:26:12<kaol>yeah, but I can't see what's used as input for the 3.3 example without seeing the <form method="POST">...
21:27:31<kaol>but fair enough, since I know what to do already
21:28:10<Saizan><form method="POST" action="/set"><input type="text" name="val"/></form>, i think
21:31:09<shapr>Does anyone know if #happs is archived?
21:32:02<tuukkah>i'm supposed to but i haven't been updating the logs lately
21:33:09<shapr>Ah, ok
21:33:19<shapr>Where are the logs?
21:34:10<tuukkah>http://tuukka.iki.fi/tmp/logindex
21:34:40<shapr>spiffy, thanks
21:35:41<tuukkah>now that the happs apps are moved to another server i should be able to make it realtime again
21:36:12<shapr>I bet hpaste could be hacked to be an irc logger.
21:36:32<tuukkah>indeed
21:45:34<kaol>yay. finally got a version that works as designed, too. Got my code that fetches a value from the DB, now. http://hpaste.org/1715#a2
21:45:38<kaol>ACTION smiles happily
21:45:53<kaol>thanks for the help
21:48:55<kaol>and I'm starting to feel that I understand how HAppS works, too...
21:49:17<kaol>though I'm looking at a two months old snapshot.
21:51:02<shapr>Even with all the exciting new code, you'll understand it.
21:54:17<tuukkah>fast-forwarding the logs now :-)
21:57:30<kaol>ok, next question. How do I set a response code in my respondInfo? Those NotFoundReplies should really go along with a 404 status.
21:58:20<kaol>hmm... by setting rsCode.
22:01:58<alexj>notFound should set the correct status code.
22:02:10<tuukkah>shapr, ok, back at realtime. of course it takes a while for google to catch up
22:02:43<shapr>yay
22:07:07<kaol>alexj: but I'm operating after a respond $. I don't think that I could stick a notFound in there anywhere.
22:10:08<kaol>ah. I could use "liftM (\res -> res {rsCode = 404}) $ toMessageM" instead of a plain toMessageM

Back to channel and daily index: content-negotiated html turtle