Camelia, the Perl 6 bug

IRC log for #darcs, 2010-08-16

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
02:29 zooko joined #darcs
02:32 exlevan joined #darcs
02:39 maurer_ joined #darcs
02:55 twb joined #darcs
03:13 sshc_ joined #darcs
03:58 JaffaCake joined #darcs
04:18 maurer_ joined #darcs
05:27 ManateeLazyCat joined #darcs
05:47 ManateeLazyCat left #darcs
06:13 maurer_ joined #darcs
06:14 maurer_ joined #darcs
06:18 maurer_ joined #darcs
06:55 maurer_ joined #darcs
07:02 darcscommitbot joined #darcs
07:03 darcswikibot joined #darcs
07:06 exlevan joined #darcs
07:15 sxs joined #darcs
07:20 devdos joined #darcs
07:20 devdos left #darcs
08:36 exlevan joined #darcs
09:52 kowey joined #darcs
09:52 kowey morning
09:57 * exlevan waves
10:09 kowey exlevan: you may want to think about sending what you have now so that mornfall has a chance to look at it
10:09 kowey (whilst you continue on with the hacking/debugging that you may want to do)
10:23 intripoon joined #darcs
11:30 [1]intripoon joined #darcs
11:32 [2]intripoon joined #darcs
12:38 [1]intripoon joined #darcs
13:41 zooko` joined #darcs
14:17 balor joined #darcs
14:34 gh_ joined #darcs
14:34 sxs i just watched a youtube about camp
14:34 sxs is camp the new darcs?
14:35 gh_ it is a parallel project, aimed at developing a correct patch theory
14:36 sxs and when its ready for use darcs will be substituded by camp?
14:36 gh_ but it also has nice ideas about "concrete stuff" also (repository versionning for instance)
14:36 gh_ basically it is what darcs 3 will be
14:36 gh_ i think the plan is to replace darcs' core by camp's
14:37 sxs what IDEs are you using for haskell?
14:37 sxs i wanna look the src
14:37 Igloo I wouldn't worry about that happening in the near future, though
14:37 gh_ so no full substitution, because OTOH darcs has all that is needed for command line arguments parsing, file system crawling, etc.
14:38 * Igloo uses vim
14:38 gh_ sxs, vim and ghci
14:38 sxs *rofl* i guessed that
14:38 sxs i'm a nano user :P
14:38 gh_ i have see leksah was nice but cabal-install failed me so i can't test it
14:39 gh_ sxs, never tried joe?
14:39 maurer_ joined #darcs
14:39 sxs uhhh, i guess i tried it long time ago. windows ide?
14:40 sxs http://joe-editor.sourceforge.net/ < this?
14:40 gh_ yes, this one
14:40 sxs i like netbeans. maybe theres a plugin for haskell
14:40 gh_ it's just that it seems almost as simple as nano but has vim and emacs mode if you want to try
14:41 gh_ I don't think so
14:41 sxs http://forums.netbeans.org/topic7641.html someone tried to do so
14:41 gh_ you can also try geany, i've see a ddc programmer blog about it, it's close to gedit
14:41 sxs but hasekll is not that wide spread language
14:42 gh_ http://www.mega-nerd.com/erikd/B​log/CodeHacking/Geany/index.html
14:42 gh_ (has a haskell color mode)
14:42 gh_ there is a eclipsefp plugin that is supposed to handle haskell, but I don't know if it is up to date now
14:43 sxs yeah, nicer. i'm a klickiklicki developer :)
14:44 kowey I wonder if Cream http://cream.sourceforge.net/ and Haskellmode would play nice together http://projects.haskell.org/haskellmode-vim/
14:44 kowey Cream is a configuration for vim for klickiklicki developers
14:46 gh_ cream looks interesting
14:47 sxs http://osum.sun.com/profiles/b​logs/netbeans-haskell-support
14:48 gh_ cream really hides vim :) there seems to be nothing in common
14:48 kowey one good thing is that the author is dedicated to keeping it as a config and not a fork
14:49 gh_ that's good
14:49 gh_ sxs, I suspect haskell projects in general are more succint than java programs, and maybe a full fledged IDE like netbeans is just not required
14:50 gh_ sxs, and so any combination of ghci + darcs + text editor should be fine, granted you know how to use them
14:50 gh_ but maybe I'm just too lazy to learn netbeans/eclipse/emacs
14:50 sxs the big thing is: im used to netbeans for some years now. so i need not to learn a new tool
14:51 sxs geany and its deps are stil compiuling. while that i test that nb plugin
14:52 gh_ gentoo user?
14:52 sxs mac
14:52 gh_ ok
14:52 gh_ kowey, we still don't know why darcs is slower under mac?
14:53 sxs is it?
14:53 kowey gh_: nope, but we have darcs get http://code.haskell.org/~kowey/cutdown-darcs to play with
14:53 kowey I've been slowly trying to purge things from it, like patch application
14:53 gh_ sxs, we are looking for people with macs to investigate weird slowness of darcs
14:53 sxs maybe i can help you
14:53 kowey it'd be good if somebody could test that locally getting a hashed GHC repo with cutdown-darcs is still slower than on Linux
14:54 kowey the goal is to one day have like a 100 line haskell script we can give to haskell-cafe and say "uh, why is this slow?"
14:55 sxs ok. i understand 20 % of what you said. i never coded haskell.
14:55 kowey sxs: http://lists.osuosl.org/pipermail/d​arcs-users/2010-August/024675.html
14:55 kowey sxs: oh I didn't use much Haskell in that; I was just taking the darcs source code and simplifying it so that we have a minimal test case for Darcs being slow
14:56 kowey without all the Darcs stuff in it
14:56 sxs GHC is the haskell compiler, right?
14:56 kowey the hope is that by doing this, we can involve more people from the Haskell community, show them a simplified version of our problem
14:56 kowey yeah, so we have a test repository which coincidentally happens to the source code for GHC
14:57 kowey "locally getting" just means doing a darcs get but from a local copy (because I removed the network code)
14:57 kowey and "hashed" is just saying that the repo needs to be in hashed format (because I removed support for non-hashed repos)
14:57 kowey but nowadays if you darcs get http://darcs.haskell.org/ghc, you'll get a hashed repo
14:58 sxs hashed repo is a feature of darcs 2?
14:58 sxs i think i read s.th. bout that
14:58 kowey that's right, http://wiki.darcs.net/RepositoryFormats
15:00 sxs i think thats a good chance for me to test the new instruments tools from Apple
15:00 kowey that's actually something we were hoping somebody would do (in that mailing list thread)
15:00 sxs where are (most of) you guys from? Europe or USA? I ask because the time shift
15:01 kowey 5/6 of active darcs developers are in Europe
15:01 kowey http://wiki.darcs.net/DarcsTeam
15:02 kowey oh, sorry was just counting the review team
15:02 sxs hehe
15:02 kowey 8/10 if you include everybody on that page that's said something in the past month (or something vague like that)
15:05 sxs cool. i'm also in the #hudson channel. wanna make a darcs plugin for hudson ci. but most of them are sleeping, when i'm awake :/
15:09 gh_ hudson is like buildbot?
15:10 sxs didnt knwo buildbot. ist a continious integration like cruise control. but much more easier to install and maintain.
15:13 kowey you can see darcs' use of buildbot here: http://buildbot.darcs.net/waterfall
15:13 kowey and here: http://buildbot.darcs.net/console
15:15 sxs i think ist like hudson
15:17 sxs haskell has an interesting syntax. like erlang
15:24 gh_ never seen that language. i know alexsuraci programmed in it (he reverse-engineered a ragnarok server)
15:25 sxs ist also functional and uses the -> to show, err, i didnt know it in english.
15:25 alexsuraci heh, erlang's syntax is wacky.
15:25 gh_ http://darcsden.com/alex/aliter
15:25 gh_ hi alex :)
15:25 alexsuraci hey
15:25 sxs the mat "picture" of a mathematic function
15:26 gh_ alexsuraci, I didn't have time to continue trying to make darcsden work on my machine, but when I do I'll let you know (maybe not before september)
15:26 alexsuraci ah, alright
15:27 sxs what is ragnarok?
15:27 alexsuraci a korean MMORPG
15:28 sxs ah, then i found the right thing in google
15:31 alexsuraci gah, my haskell setup is borked pretty bad
15:31 gh_ alexsuraci, is it safe to point people to darcsden.com for hosting?
15:31 alexsuraci should be. i'd say it's pretty stable now.
15:31 alexsuraci i use it every day
15:31 alexsuraci the new ssh server rocks :P
15:32 kowey time for it to go up on hackage?
15:34 alexsuraci once i split it out from darcsden, sure
15:34 gh_ the current version of darcsden.com has this new ssh server?
15:34 alexsuraci yea
15:34 gh_ nice
15:37 kowey I think Thomas blogged about wanting to put into patch-tag
15:37 kowey http://blog.patch-tag.com/2010/07/21/patch-ta​g-lessons-learned-and-announcing-open-source/
15:38 alexsuraci yeah he's contacted me, I haven't had the time though
15:38 gh_ alexsuraci, will darcsden keep on using public ssh keys in the future, or a generated password, ala google code hosting?
15:38 gh_ I thought you mentioned passwords once,
15:38 gh_ maybe I'm mistaken
15:38 sxs have to go. cu all
15:39 alexsuraci i don't plan on having anything other than pubkeys
15:39 gh_ ok fine
15:39 gh_ don't know where I got that idea from
15:41 gh_ when pushing to darcsden with darcs HEAD, there is a line saying "this server only accepts exec requests"
15:41 gh_ was that written by darcs or the ssh server?
15:41 gh_ ok, it's darcsden
15:42 alexsuraci odd
15:42 gh_ so darcs tried to do something wrong?
15:42 alexsuraci mind pushing with --debug-verbose?
15:42 gh_ hmm ok (got to create a patch)
15:45 * alexsuraci curses happy under his breath
15:48 gh_ alexsuraci, the message did not show up again
15:48 alexsuraci dun dun dun
15:48 gh_ it is a private repository
15:49 alexsuraci looks like it's pushing to patch-tag there?
15:49 gh_ and the 1st push I made (with the message) was pushing 6 patches against an empty repository
15:49 gh_ wut
15:49 gh_ err
15:49 gh_ ok, trying again :)
15:49 alexsuraci haha
15:50 gh_ (caught by new flag behaviour) (which I support in theory)
15:52 mornfall joined #darcs
15:52 mornfall 3 hours to deadline and no code from exlevan?
15:52 lambdabot mornfall: You have 3 new messages. '/msg lambdabot @messages' to read them.
15:52 gh_ alexsuraci, this time the push failed (trying once more without --debug-verbose)
15:53 gh_ Oo
15:53 gh_ this time without --debug-verbose, the push was successful
15:53 gh_ with the same "this server only accepts exec requests" message
15:53 alexsuraci to darcsden? :P
15:54 gh_ yes
15:55 alexsuraci actually i don't expect it to work with --debug-verbose anyway
15:55 alexsuraci since that adds --debug to the command, which darcsden doesn't handle
15:55 alexsuraci http://darcsden.com/alex/d​arcsden/browse/Main.hs#L89
15:55 gh_ I supposed so
15:57 gh_ but that's bad for debugging :)
15:58 alexsuraci no idea why it's saying "this server only accepts exec requests", that should be pretty foolproof
15:59 alexsuraci as in, it must be sending a non-exec request
15:59 mornfall Like sftp?
16:00 alexsuraci would it do that?
16:00 exlevan mornfall: I'm somewhat stuck with cgi stuff... Hopefully will make something working in a couple of hours
16:00 lispy joined #darcs
16:00 mornfall Well I sure hope you will.
16:04 alexsuraci gh_: I'll have it dump out the invalid request it received at least
16:05 mornfall alexsuraci: darcs is using sftp yes
16:05 mornfall Well, at least it tries (with a scp fallback).
16:08 alexsuraci is that a new thing? what about transfer-mode?
16:15 kowey exlevan: I'd still suggest sending your work in progress, could make your life easier
16:15 mornfall Well, it first tries transfer-mode, I guess.
16:15 mornfall kowey: I don't think it's going to help anymore.
16:15 mornfall I mean, now or in 2 hours doesn't make a difference.
16:15 sm go exlevan, go
16:16 * sm sends sandwiches
16:20 lispy joined #darcs
16:28 arjanb joined #darcs
16:37 gh_ bye
17:31 lispy joined #darcs
17:33 balor joined #darcs
17:34 balor joined #darcs
17:40 lispy joined #darcs
17:41 FunctorSalad joined #darcs
17:45 mornfall 1:15 left
17:57 lispy and 0:45 right?
17:57 lispy > Left "1:15"
17:57 lambdabot Left "1:15"
17:59 balor joined #darcs
18:04 balor joined #darcs
18:21 zooko joined #darcs
18:31 gwern joined #darcs
18:35 kowey joined #darcs
18:56 mornfall Well... Пнд Авг 16 18:56:27 UTC 2010
18:57 mornfall And no mail.
19:01 exlevan Successfully sent patch bundle to: patches@darcs.net.
19:01 mornfall You certainly are playing on the edge.
19:02 pierreee joined #darcs
19:05 exlevan just seen a horrible mistake 30sec before deadline... Fixed it to be less horrible
19:06 exlevan i'll send description now
19:08 mornfall We'll see.
19:11 balor joined #darcs
19:31 sshc joined #darcs
19:45 Gracenotes joined #darcs
19:46 iago joined #darcs
19:53 balor joined #darcs
20:14 sshc joined #darcs
20:14 balor joined #darcs
20:55 mornfall I'll have a closer look tomorrow, I have to sleep. Goodnight.
20:58 balor kowey: FB chat sucks
20:58 balor kowey: Why don't mortals use IRC>
20:59 kowey night, mornfall
20:59 * kowey wanders off too
21:00 balor kowey: Have a good sleep
21:00 kowey balor: there's always jabber
21:51 FunctorSalad_ joined #darcs
21:56 Heffalump grmph. Darcs repository handling is not very compositional.
22:02 Igloo IWW?
22:04 * lispy can't think of what "IWW" translates too
22:04 lispy oh
22:04 lispy In what way
22:05 lispy Heffalump: I agree with Igloo
22:06 Heffalump I'm trying to run fragments of two commands, both of which add/delete patches and change working, in succession. I'm ending up with the second one claiming there's a conflict in working which there isn't (AFAIK)
22:07 lispy hmm
22:08 lispy Heffalump: hmm
22:08 Igloo Curious
22:08 Igloo Has getCurrentWorkingDirectory changed after the first command?
22:09 lispy Heffalump: Sorry, that second "hmm" was because I started to type something and deleted without relaizing I had already said that :)
22:09 Heffalump NAFAIK, and if it had it'd be completely broken, wouldn't it?
22:09 Igloo I don't know
22:10 Igloo But that and atexit are the only things that come to mind, and I hope nothing relevant is atexit'ed
22:10 Heffalump I think it has to do with what darcs thinks the working dir contains.
22:12 lispy Heffalump: doesn't it report the reason for the conflict?
22:12 Heffalump well, yes. It claims there's a conflict in the state of working with the patch I'm just applying. But I already undid the state of working that would be in conflict.
23:35 sm joined #darcs

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary