Experimental IRC log happs-2007-08-05

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.

19:16:36<tuukkah>put IO in a Handle filter using addSideEffect and it's ran after the ModResp filter
19:35:22<Saizan>yeah, and only if there's no server error
19:35:32<Saizan>e.g. fail ""
19:35:56<tuukkah>if there's a server error does anything happen?
19:36:22<tuukkah>i mean, doesn't the whole transaction roll back?
19:37:00<Saizan>i'm not sure
19:37:34<tuukkah>i was already thinking about that since the example of counting the number of requests and responses doesn't make much sense if it's like that
19:43:57<tuukkah>for each server error, both the request and response counts increase by *two*
19:44:28<Saizan>2?
19:45:23<tuukkah>also the request gets printed twice and response naturally not at all
19:46:47<Saizan>really? i didn't get the request printed last time i've used this
19:48:45<tuukkah>let me paste the code...
19:51:49<tuukkah>http://hpaste.org/1900#a4
19:55:08<tuukkah>although if i request /error the first thing when i start the server, the request gets printed only once
20:00:20<Saizan>i don't experience the doubling
20:00:48<Saizan>but when i call error i don't get the server error page, it loads indefinitely
20:01:32<tuukkah>i see wget retrying since it gets no data
20:02:06<tuukkah>firefox live http headers doesn't say it would try twice but perhaps it still does
20:02:33<Saizan>ah, but you respond $ fail ""
20:02:50<Saizan>i was talking about a fail in the Ev monad, not IO
20:03:17<tuukkah>yeah, i was trying both
20:04:21<tuukkah>oh, in Ev the roll back works
20:05:13<Saizan>well the IO Result cannot inspected inside the transaction, so it makes more sense that it doens't rewind
20:05:30<tuukkah>indeed
20:05:50<tuukkah>perhaps it should still catch the exception
20:07:30<Saizan>definitely
20:08:09<tuukkah>in another thread, i've been wondering what to put in the tutorial about smtp retries
20:08:57<tuukkah>"Sending email is straightforward." is getting less and less true in my eyes
20:10:58<Saizan>now i get weird errors like:
20:11:01<Saizan>Fail: <stdout>: hPutStr: illegal operation (handle is finalized)
20:11:11<Saizan>just by starting the server
20:12:01<tuukkah>perhaps those are because of the uncaught exceptions
20:12:29<Saizan>yeah, but caused by what?
20:13:11<Saizan>the user code can't cause exceptions before a request arrives
20:17:50<Saizan>ah, inconsistent state..
20:18:09<Saizan>C-c is not the best way to shut down the server :)
20:20:20<sorear>Saizan: If you set the entirely undocumented -DUNIX, SIGINT becomes a clean shutdown
20:21:02<tuukkah>sorear, not entirely, it was some patch names ;-)
20:21:27<tuukkah>you could add a section to the deployment part of the tutorial
20:21:49<Saizan>sorear: i told you so :) i'm just lazy, but i'll put it in the .cabal for the next happs' build
20:21:53<sorear>No, I expect the documentation to be written for me. :)
20:22:53<tuukkah>=)
21:41:09<tuukkah>since when can the arrow in do-notation point to the right?
21:41:49<Igloo>Since a GHC bug was introduced, and up until the point it was removed :-)
21:42:55<tuukkah>such a bug sounds weird
21:46:09<tuukkah>now the next question should be answerable by darcs, as in since when my code contains such notation
21:48:53<xerox>They can?!
21:49:26<tuukkah>http://hackage.haskell.org/trac/ghc/ticket/1555
21:51:31<xerox>How cute.
22:11:39<tuukkah>not answerable by darcs because of some darcs bug or something
22:14:01<chessguy>that sounds like an April 1 joke :)

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