Experimental IRC log happs-2008-03-29

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.

08:51:02<porrifolius>Is HAppS.State.Util.getRandom intended for use (H.S.U isn't exported), or should we have a pure random generator in our state?
10:25:19<Lemmih>porrifolius: It is exported by HAppS.State
10:25:44<Lemmih>porrifolius: You might be better off passing in a StdGen from the outside world.
19:16:58<darrint>Anyone know where to get system.log? It appears the darcs repo is gone.
19:19:36<Saizan>darrint: it's in the hslogger package right? there's a tarball on hackage
19:20:23<darrint>Saizan: Can I edit .searchpath.default.map and make searchpath install it?
19:24:28<Saizan>darrint: i think so if you unpack the tarball somewhere, but i'm not familiar with searchpath
19:43:15<kaol>HAppS 0.9.2.1 uploaded for Debian. Tell me if it doesn't work. At least I got mightybyte's example to work with my package, so I hope it's ok.
19:44:18<kaol>I ended up combining all the HAppS-*-0.9.2.1.tar.gz tarballs into one big HAppS-0.9.2.1.tar.gz, adding a common cabal file for them all
19:44:56<Lemmih>kaol: A single cabal file for all the packages?
19:45:37<kaol>yup. combined all the dependencies and exposed modules and dropped any HAppS-foo dependencies and named it as HAppS
19:45:48<Lemmih>kaol: That will not work.
19:47:11<Lemmih>All non-toy applications specify HAppS-Server, HAppS-State, etc.
19:47:56<kaol>hmm... there's a definite mismatch with the package names that people expect to see
19:48:01<Lemmih>kaol: How about a single tarball that contains all the necessary libraries?
19:49:59<Lemmih>I don't see why you wanna merge all the packages.
19:50:09<kaol>sounds like trouble, since they have dependencies on each other and need to be installed at build time. eg. HAppS-State depends on HAppS-Util
19:50:17<Lemmih>(A separate dep for each library would be even better, of course)
19:51:29<Lemmih>kaol: Can't you install them one by one? (I'm not familiar with the internals of debian packages)
19:51:39<kaol>well, I can do another upload after this one... 5 tarballs means 5 times the work. eh.
19:52:20<kaol>5 source packages instead of one. it's a pain.
19:52:55<Lemmih>I still don't see why you can't install all the libraries in a single go.
19:53:46<kaol>what'll happen if HAppS-Util and HAppS-State are built at the same time? won't HAppS-State need to have HAppS-Util installed in order to satisfy its dependencies?
19:54:09<Lemmih>kaol: Specify the build order manually?
19:54:36<Lemmih>I assume there is some way of sequencing IO actions.
19:54:51<kaol>I'm talking about HAppS-State.cabal. It has Build-Depends: HAppS-Util>=0.9.2 in it
19:55:02<Igloo>He means that they don't get registered in the package.conf during the build process
19:55:08<kaol>I can't satisfy that without building and installing HAppS-Util first
19:55:14<Igloo>build and install are separate for Debian packages
19:55:25<Igloo>What's the problem with doing them in separate packages, kaol?
19:55:46<kaol>ok, sure. I'm a regular packaging droid. All work and no play.
19:56:19<Igloo>The alternative would be to register them in-place. I'm not sure how much effort that would require
19:56:20<Lemmih>Can't he just use the --package-db flag?
19:56:44<Igloo>but someone will have to make it work at some point
19:57:39<kaol>at least what I uploaded now is less broken than what was there before.
19:58:25<kaol>I'll package the individual packages tomorrow...
20:00:07<kaol>with going through NEW, they'll hit the archive in 2 weeks. With luck.
20:00:56<Lemmih>Super, thanks.
20:01:21<kaol>and sorry for my grumbling. Been having a headache all day.
20:24:33<MarcWeber>kaol: Not sure but: Have you talked to cabal people? Maybe there is some automatically create deb with dependecies script ? Something like this does exist for gentoo..
20:24:47<MarcWeber>Hi Igloo_ :)
20:27:40<MarcWeber>Do we have kind of connection close event?
21:00:47<fxr>re
21:04:41<Lemmih>MarcWeber: What connection?
21:12:37<MarcWeber>Lemmih: The http-connection (keep-alive).. The idea is to get notified if the user closes the browser window.. But this can only be done using time out I guess
21:13:07<Lemmih>MarcWeber: The web-server isn't connected to the event system.
21:13:57<Lemmih>The http server doesn't raise any kind of signal when a connection is closed.
21:14:39<MarcWeber>Lemmih: I don't know the http protocol and tcp/ip stuff well enough.. But do you know wthere the server could do it?
21:15:21<MarcWeber>This would be incredible. You can free resources, in life systems you can tell the "friends" in realtime when they go offline etc.
21:17:55<Lemmih>MarcWeber: It wouldn't be very reliable. Connections close all the time.
21:20:16<kaol>you can use AJAX to get to know when the browser window is closed, too. Though that's not a surefire method either

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