Experimental IRC log happs-2007-07-20

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.

00:57:45<shapr>jmelesky, kaol, msouth: the Roadmap document is nearing completion.
00:58:02<shapr>The branch repo should be out sometime next week at the latest.
00:58:11<msouth>great
00:58:33<shapr>And I've already written working examples for the new API, it's much nicer imho.
00:58:57<shapr>But I want feedback from people who haven't been developing API iterations because I could be wrong!
00:59:04<msouth>heh
00:59:05<shapr>er, s/I/we
00:59:46<msouth>well, I'm a haskell newbie, I don't know how much help I'll be, but I'll try anything that you think will be of use.
01:00:10<shapr>I'd like to hear about anything that doesn't "make sense" or leads to misdirection.
01:02:51<msouth>at our company, there's one guy that knows haskell, who just joined. I started learning it last week. We're at the beginning of switching our architecture to a service-oriented one for new work, and anything is possible...
01:03:14<msouth>however, there's a lot of momentum for Rails to be "the thing" that the services get developed in.
01:03:42<shapr>msouth: Oh, who just joined? Someone from #haskell?
01:03:53<msouth>Dino Morelli
01:03:59<shapr>er, hey .. I know that name.
01:04:06<shapr>ACTION thinks
01:04:21<msouth>yeah, he has at least mailed to haskell-cafe from time to time
01:04:26<msouth>he is very enthusiastic.
01:04:30<shapr>I'm almost positive I've read blog entries by him.
01:05:13<msouth>he has been learning it on his own--I think he asked one of you guys about his command line utilities, whether they would be of interest to the community.
01:05:27<msouth>he's been on here as DinoMI lately
01:05:57<shapr>He's not on #haskell or #happs though?
01:05:58<msouth>we work for McClatchy Interactive, which is the interactive wing of the 3rd biggest newspaper company in the us
01:06:04<shapr>Wow, nifty.
01:06:38<msouth>I don't know how much time he spends on irc
01:06:54<msouth>I told him about #happs, I think he came in yesterday for a while
01:07:00<shapr>cool
01:07:54<shapr>HAppS is positioned in the same spot as RoR for a lot of thinsg.
01:08:26<msouth>Dino's impression of RoR is "junk food" :)
01:08:28<shapr>haha
01:08:55<shapr>That's pretty funny.
01:09:10<shapr>And fitting too
01:09:19<msouth>I haven't really formed an opinion of RoR, but I'm concerned about performance/scaling.
01:10:31<shapr>My impression isn't detailed either, but it seems to me that RoR almost requires the use of Mac OS X, TextMate, and the whole RoR workflow.
01:10:43<shapr>I've heard that it's hard to do Rails on Linux.
01:12:15<msouth>I don't know. we have about a dozen guys all over it trying to figure out how to use it, how to get it into our systems, what perfomance and stuff is about. I figured that was enough people so I struck out on my own to see what haskell could do.
01:12:30<msouth>The CTO is hyped up about rails
01:12:48<msouth>well, informed hype
01:12:54<chessguy_>msouth, what newspaper
01:12:56<shapr>heh
01:13:10<msouth>partly for him it's the network effect of there being a whole community
01:13:15<shapr>Yeah
01:13:29<chessguy_>haskell has a great community
01:13:30<shapr>HAppS is limited to the #haskell community, but that's a very focussed and supportive communiy.
01:13:33<msouth>chessguy_: newsobserver.com sacbee.com miamiherald.com
01:13:38<msouth>those are the largest
01:14:12<msouth>yeah, i am totally impressed with #haskell
01:14:33<msouth>all these brilliant people willing to help complete newbies with their questions.
01:15:08<msouth>(and I understand the "promote" idea--but I'm just saying, it's nice to see very, very advanced people with a lot of patience)
01:15:12<msouth>contrast that with #perl
01:15:26<msouth>I love perl, but #perl is brutal
01:15:38<chessguy_>not to mention ##c++
01:15:51<msouth>haven't been there
01:15:58<chessguy>count yourself lucky
01:16:01<msouth>:)
01:16:23<msouth>are happs questions ok on #haskell or should they be in here only?
01:16:49<shapr>Probably better in here, but you can ask in #haskell as well.
01:17:58<msouth>so is the place you guys are committing the current development to available for public to check out?
01:19:27<msouth>chessguy: here is a list of the daily papers: http://www.mcclatchy.com/102/story/354.html there are also weekly and other smaller things.
01:19:29<lambdabot>Title: The McClatchy Company - Newspapers
01:24:42<shapr>msouth: As soon as the roadmap is completed, the repository will be public.
01:25:08<msouth>ok
02:37:53<jmelesky>shapr: I'll happily try and develop something against the new API when it comes out.
02:51:08<blackdog>wow. hello #happs, you've woken up... :)
02:51:12<shapr>jmelesky: yay!
02:51:19<shapr>blackdog: Ok, going to sleep now ;-)
02:51:36<blackdog>man, every time...
02:51:44<blackdog>i'm beginning to think i'm a soporific
02:52:34<blackdog>so i'm still wondering - if i run a couple of HAppS servers, how do they talk to each other?
02:52:48<shapr>Roadmap will be out soon!
02:52:50<blackdog>the standard webapp method would be to communicate through a DB, and obviously that isn't the HAppS way...
02:53:03<blackdog>yeah, i'm impatient :)
02:53:29<jmelesky>do we know how soon soon is?
02:53:45<jmelesky>i'm guessing it will wait at least until this weekend's contest is past
02:53:48<blackdog>shapr: incidentally, RoR with a terminal and Emacs works just fine
02:54:01<blackdog>heh, was going to ask how many people are entering
02:54:12<blackdog>i'm seriously thinking of giving up a party and a date to do it...
02:54:39<jmelesky>keep perspective. programming contests are probably not worth missing a date
02:54:56<jmelesky>however, you could participate before and after
02:57:17<blackdog>before, maybe. :P
02:57:40<jmelesky>alternately, there's this: http://code.google.com/p/reddit-icfp-lounge/wiki/Welcome?updated=Welcome&ts=1184897122
02:57:42<lambdabot>Title: Welcome - reddit-icfp-lounge - Google Code, http://tinyurl.com/2hsmlx
02:57:52<jmelesky>thanks, lambdabot
02:58:46<blackdog>nah, if you're gonna do it you gotta get into the ring
02:58:57<blackdog>i'm tentatively on a UNSW team
02:59:38<jmelesky>i'm tentatively on a london.pm team, despite being in Chicago, US
03:00:38<jmelesky>i need to get some sleep between now and the start, though
03:01:21<jmelesky>good luck in the contest
03:41:16<msouth>blackdog: how much have you worked on happs?
03:44:22<msouth>I know the answer to this might be "wait for the roadmap", but I was wondering, with all the data on disk (it is on disk, right) if one server had it mounted read/write, and several others mounted it read-only, if read-only ops could maintain consistent views by seeing changes on the disk.
03:44:40<msouth>I guess, though, that's not the way it works now, changes in memory then go to disk
03:49:40<shapr>yarr!
03:50:18<shapr>msouth: The new codebase supports multimaster.
03:50:50<msouth>I knew if I worked hard enough I could force some roadmap leakage.
03:50:55<msouth>:)
03:51:02<msouth>that's awesome
03:51:19<shapr>heh
03:51:26<msouth>I figured with your reference to the amazon s3 and ec2 something like that had to be in the works.
03:51:53<shapr>you got it :-)
03:52:54<msouth>you aren't going to be at oscon by any chance are you?
03:53:04<shapr>I don't think so, when/where?
03:53:13<msouth>starting monday in portland
03:53:24<msouth>portland , oregon, us
03:53:29<shapr>Ah, no.
03:53:40<shapr>Simon Peyton-Jones will be giving some Haskell talks there, I think.
03:53:45<shapr>Are you going to be there?
03:54:33<msouth>yes. I am signed up for SPJ's taste of haskell, i'm very excited about that. And I'll be going to all the other things he's doing.
03:54:56<dino->msouth: oo, I didn't know that, ALL other things
03:55:20<msouth>btw shapr , dino- is the same dino we were talking about before
03:55:28<dino->You, sir, have totally flipped for the big H
03:55:32<msouth>I just didn't know it was the same person behind the nick
03:55:36<shapr>hi dino-!
03:55:54<dino->shapr: Hello. You guys have been helping me for the past 14 months with Haskell.
03:55:57<shapr>Yay!
03:58:18<dino->mouth has probably said: trying to infiltrate the day-job place with a Haskell REST service built with happs.
03:59:09<msouth>yeah, I told him about your RoR = junk food theory
03:59:20<shapr>That's pretty funny :-)
03:59:35<msouth>careful what you say at work dino- , it gets on teh intarwebs
04:00:03<dino->Kind of said some crazy things lately to our people in charge there.
04:00:24<msouth>basically, instead of a RoR cloud of services we could have a happs cloud. I'm expecting that exact sentence to be at the bottom of the roadmap
04:01:16<msouth>dino-: I think the stuff you sent to Clint is why he was supportive when I went in and told him what I was doing.
04:01:32<msouth>I was surprised he was so in favor of me working on it.
04:01:44<shapr>haha
04:01:47<msouth>but then when I saw what you had mailed him, it made more sense.
04:01:55<dino->:D I told the lovely Mrs. Dino. I'm spamming bosses with crazy stuff, I'm going to get promoted or fired soon, one or the other.
04:01:56<shapr>I think there are direct references to RoR in the HAppS roadmap :-)
04:02:03<shapr>dino-: haha!
04:02:10<dino->Something extreme is going to happen.
04:03:46<dino->msouth: I'm glad that people are thinking about it. Honestly, I never get any ack of any kind from the messages from those guys.
04:03:50<msouth>shapr: there's no reason that happs would be limited to POST and GET calls, is there? Because we're going to be going with the full REST kool-aid and using PUT and DELETE from the http spec
04:04:12<shapr>I think those are already supported.
04:04:12<msouth>dino-: they are the two busiest people in the company right now.
04:04:34<msouth>cool--I thought they probably were, but just wanted to check because almost no one ever uses them.
04:04:57<msouth>I don't think even REST people really do, they just POST something that says delete.
04:04:58<msouth>:)
13:58:22<shapr>yarr!
14:06:23<msouth>shapr: is there an example in the distribution or docs that shows how to get key/val pairs from POST or from query string?
14:09:00<Saizan>msouth: bodyToList :: Request -> [(String, String)]
14:11:36<Saizan>it'll give you both
14:15:56<msouth>thx Saizan
14:57:09<dinoMI>dinoMI == dino- at work
17:32:55<dinoMI>I see that there is a --default-port switch for a happs application. I'm having trouble finding out what other switches it supports.
17:33:27<dinoMI>I vaguely recall seeing something about data dir in the past, but can't find it now.
17:35:16<dinoMI>--help only shows me two, is that all there are? (--concurrent-threads, --http-port)
17:35:42<dinoMI>Well, --help make it fail and show me usage.
17:37:53<fxr>dinoMI: there are log-level and log-target defined in stdMain
17:38:38<fxr>and more, just fgrep with "Option" keyword
17:50:48<dinoMI>fxr: Ok, thanks.
19:06:17<kaol>hmm... my happs web server seems to have a 5% CPU load even though it's completely idle. What's it doing?
19:12:27<kaol>is this normal for happs?
19:35:19<Saizan>kaol: do you have CronEvents
19:35:21<Saizan>?
20:03:48<kaol>none (first time I've encountered that term, too)
20:04:37<Saizan>(they are sued to manage the expiring of sessions for example)
20:11:35<kaol>it's nothing I've consciously invoked, at least
20:11:53<kaol>ACTION would rather see %CPU stay at 0
20:43:44<msouth>I am trying to just get the server to answer a POST with one variable that was posted to it
20:43:58<msouth>using frankensteined code from hpaste
20:44:17<msouth>stitched horribly into the example from HAppS source
20:44:20<msouth>http://hpaste.org/1799
20:44:41<msouth>the function is showBody, but I don't know how it should be declared.
20:49:48<msouth>brb
20:53:49<dino->I'm working with msouth on that, and am logging this. So if anybody can help, please blurt something for us here.
22:08:03<msouth>so I went to a lot of trouble to connect from a screen session so I wouldn't miss anything while I was driving home
22:08:27<msouth>first i forgot to run screen, so I had to start over
22:08:46<msouth>when I started over, I connected, identified, then forgot to /join #happs again
22:08:50<msouth><- dork
22:09:29<msouth>so i apologize for asking and leaving
22:58:47<Saizan>msouth: what's the problem on that paste? and why showBody takes 3 arguments?
23:07:27<Saizan>also, if you don't use the "ok" handler you're supposed to return an IO Result
23:07:28<msouth>i was just guessing
23:08:07<msouth>i just started learning haskell this week, and i was having a hard time figuring out what got passed to the handler
23:08:16<Saizan>ah ok :)
23:08:19<msouth>that stuff I pasted doesn't even compile
23:12:23<Saizan>http://hpaste.org/1799#a1 <-- this is how i'd do it, feel free to ask about the modifications
23:18:16<Saizan>when you've type errors and such try asking on #haskell as well, there are more users and so it's more reponsive :)
23:25:06<msouth>oh
23:25:08<msouth>heh
23:25:29<msouth>I just apologized there for accidentally having my last paste announced accidentally
23:25:31<msouth>sigh
23:25:47<Saizan>heh, no problem :)
23:26:17<Saizan>however you've to add the noState handler to the list, so the type is no more ambiguous
23:27:24<Saizan>but.. aren't you using a state in val and set?
23:28:24<msouth>Saizan: ah, sorry--i commented that part out because it was complaining about the next handler
23:28:48<msouth>made me wonder if maybe i had copied the original wrong, and I just decided to cut out the question
23:28:58<msouth>should have put up my whole modification
23:29:14<Saizan>ok
23:29:40<Saizan>so without val and set you've to add noState, like you've debugFilter
23:30:33<Saizan>what was the problem with val and set? have you defined a corresponding datatype for that state in your code?
23:30:44<msouth>hang on a sec
23:30:49<Saizan>k
23:33:24<dino->But suppose we don't even want state at all. Just to unpack the post data, do something with it and get out of there.
23:33:36<msouth>woohoo!
23:33:50<msouth>sorry Saizan it compiled so I was excited to see if it would work
23:33:50<dino->Guessing we'd still need a data type to instance into FromMessage..?
23:33:54<msouth>and it did!
23:34:22<msouth>lemme paste it up and then we can look at the other questions.
23:34:26<msouth>i'm so excited !
23:34:40<Saizan>dino-: FromMessage it's not directly related to your state, you need a FromMessage instance for it only if you expect HAppS to extract a value of that type from the message
23:35:39<dino->msouth: I'll have to get your code, I left mine inaccessible at work.
23:36:04<Saizan>dino-: the only problem with having no handler that use the state it's that the compiler doesn't know which type to use for it, since stdHTTP is parametrized over it
23:36:07<dino->Saizan: But is there a simple way to grab the items from the POST, perhaps dumped into a Map.
23:36:24<Saizan>dino-: so you've to fix the type by using noState that has () as the state type
23:36:52<Saizan>dino-: there's bodyToMap :: Request -> Map String String
23:37:02<dino->Saizan: That function isn't exposed from that module.
23:37:15<Saizan>MessageWrap?
23:37:38<dino->Well, no API doc generates for it, but I can see it in the source.
23:38:20<Saizan>well, the documentation might be still lacking, however i've it in scope if i import HAppS
23:38:45<Saizan>we're talking about the darcs version, right?
23:38:45<dino->I don't have the source here yet, getting it now.
23:38:54<dino->But I thought it was more like it was "private"
23:38:59<dino->Not for public use.
23:39:10<Saizan>no, it's public, as bodyToList
23:39:38<Saizan>it wouldn't be exported otherwise
23:39:51<dino->I thought even something that has no doc comments should get plain vanilla treatment when haddock does its thing.
23:40:04<dino->I did not try to use the function though.
23:40:18<dino->ok
23:42:42<dino->I think I had gotten the tar.gz if that made a different. msouth probably as well
23:43:30<dino->Huh, 200 patches alone from that tag. You guys have been busy.
23:43:33<dino->220
23:43:46<Saizan>heh, and a lot of API changes :)
23:43:58<Saizan>(however i'm not one of the developers)
23:45:51<Saizan>mmh yeah the haddock is generated for bodyToMap here, so maybe it was introduced later
23:46:08<dino->0.8.8 or latest-latest ?
23:46:37<Saizan>there has been quite a few changes in MessageWrap when support for multipart has been added
23:47:08<Saizan>latest-latest
23:47:27<Saizan>what you get with darcs pull -a :)
23:48:50<dino->oic, 220 patches total but then the 0.8.8 tag rolled back half of them. Still super busy for since like a week ago.
23:50:04<Saizan>and next week we have a new big chunk of patches from a non-public branch
23:53:26<Saizan>msouth: does it work or just compiles?:)
23:53:32<dino->Hm. Any idea where I can get that generic-xml ?
23:53:58<Saizan>happs.org/generic-xml
23:54:04<Saizan>as the other
23:54:41<Saizan>http://happs.org/HAppS/generic-xml
23:54:48<Saizan>forgot a level ^^
23:55:04<dino->Bah, my browser says "No directory indexing!"
23:56:29<Saizan>err, you've to darcs get it
23:56:45<dino->ah, I see
23:56:49<dino->sorry
23:57:02<Saizan>no, i should have said that ^^
23:57:38<Saizan>you'll need to darcs get http://happs.org/HAppS/syb-with-class too
23:57:56<msouth>Saizan: works
23:58:18<Saizan>msouth: good :)
23:59:05<msouth>yeah, I'm very stoked. thanks for your help

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