Experimental IRC log happs-2007-09-11

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.

00:01:23<alexj>:-)
00:01:41<alexj>the only constraint is that it doesn't work for packages that require compilation of other languages.
00:03:25<tuukkah>also the source dir must be recursively gettable via http?
00:04:00<alexj>yes. I think curl also supports file URLs.
00:04:20<alexj>I think it is actually whatever URL format is supported by curl.
00:04:37<tuukkah>or perhaps not recursively, if we only download the hs files that some other file wants to import?
00:08:11<tuukkah>now regarding the cosmetic problems, i take it have wrong version of curl. curl --version says 7.15.5
00:08:56<alexj>I am using curl 7.13.1 and it works for me.
00:09:28<tuukkah>no warnings?
00:09:46<alexj>some warnings. but basically haskell warnings not curl warnings.
00:09:54<alexj>can you paste the what you were seeing somewhere?
00:10:47<tuukkah>ok
00:14:54<tuukkah>here's all the command prints: http://tuukka.iki.fi/tmp/SourcePath-curl-warnings.txt
00:18:06<alexj>did you delete .haskell-cache?
00:18:12<alexj>or is that after a second run?
00:18:55<tuukkah>that was in a clean darcs checkout of HAppS-Begin
00:19:05<alexj>hmm. can't reproduce.
00:19:34<tuukkah>could also be that i have a too new curl
00:19:44<alexj>so that the -z option doesn't work?
00:19:52<alexj>can you check whether the -z option works on your version of vurl?
00:19:54<alexj>curl
00:20:06<tuukkah>yeah, i'm trying to do that
00:23:28<tuukkah>so -z wants either a date expression or a filename
00:24:33<tuukkah>i'll install the man page the warning refers to
00:27:27<tuukkah>ok could the reason be that -z is given a filename that doesn't exist
00:29:02<tuukkah>yeah, it only complains about the ones that don't exist
00:29:26<alexj>interesting. so perhaps they updated curl to do that.
00:29:29<alexj>mildly annoying.
00:29:45<alexj>would you be up for patching searchpath to only use -z if the file exists?
00:30:00<tuukkah>sure
00:30:09<tuukkah>so why don't those files exist?
00:30:18<alexj>because they haven't been downloaded yet.
00:30:31<alexj>-z is basically, download the file if they have been modified since the date.
00:30:46<alexj>-z translates into an http if-modified-since header.
00:31:07<alexj>the relevant portion of the code is the very end... the getURI function.
00:31:10<tuukkah>but the files aren't there on later runs either
00:31:17<alexj>?
00:31:33<alexj>oh, perhaps curl refuses to download if it has an error?
00:32:17<alexj>maybe it doesn't download if -z is present with an illegal date format.
00:32:19<tuukkah>ok i'll try to figure it out :-)
00:32:28<alexj>cool. thanks.
00:39:05<tuukkah>yeah that's it: we try to download lhs and hs and it keeps complaining about the one that doesn't exist
01:00:18<tuukkah>btw, i'd say runhaskell haskell/Main.hs takes too much memory for many machines: VIRT 206m RES 178m
01:06:27<tuukkah>alexj, patch sent
03:00:05<tuukkah>sent a new bundle with two more patches
15:51:25<tuukkah>shapr, did you read the log of me installing HAppS?
15:51:46<shapr>No?
15:51:57<shapr>But I can look at it in a bit when I figure out this type error.
15:52:48<shapr>Or while I'm compiling :-)
15:52:48<shapr>ACTION looks
15:53:12<tuukkah>half of it got resolved already yesterday but there are some things left that could be documented better
15:53:20<shapr>What things?
15:53:42<tuukkah>the order the repos should be compiled
15:53:58<shapr>Ah
15:54:01<tuukkah>and the external dependencies, their locations and required versions
15:54:05<shapr>install order: Util, Data, IxSet, State, Server, Begin
15:54:07<shapr>Yeah
15:54:25<shapr>Yeah, I think we need to add more granularity to the versions.
15:54:31<shapr>0.9.1 has had several incompatible changes.
15:54:39<shapr>I'd like to switch to 0.9.1.$DATE
15:55:19<tuukkah>why not bump the version numbers liberally?
15:55:54<shapr>I don't know, alex doesn't seem so interested in that.
15:56:00<shapr>I'll ask him when he gets online today.
15:58:04<Igloo>0.9.1.$DATE would mean having to rebuild all the time
15:59:01<Igloo>And when you're working from a darcs repo the version is never going to uniquely define the code (unless every patch changes it, which would mean that every patch conflicts with every other patch)
16:00:19<tuukkah>the HAppS-Begin tutorial suggests using -http-port, should be --http-port. also, i'd add --log-level NOTICE to the example invocation
16:00:22<Saizan>i think you want to rebuild the package and its dependencies after you pull some patches, no?
16:03:00<Igloo>Yes, but normally only those files that have changed
16:04:50<Saizan>uhm, you lose that granularity when you rebuild the packages that depend on it, though, right?
16:05:27<shapr>tuukkah: Oh, I'll add that extra dash, thanks.
16:06:36<tuukkah>the point in log-level is that you'd actually see the compilation has finished and happs has started. would be nice to get a message like "accepting http requests on port 8000"
16:06:51<shapr>Ah, good point
16:07:33<Igloo>Saizan: That only matters if cross-module inlining has happened across package boundaries, and what has been inlined has changed. But yes, you do occasionally have to clean everything
16:08:08<Igloo>You can also avoid it by turning off optimisation

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