Camelia, the Perl 6 bug

IRC log for #darcs, 2013-01-08

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

All times shown according to UTC.

Time Nick Message
00:00 jochu joined #darcs
00:53 tranma joined #darcs
01:03 * owst wonders where the name glump09 came from http://hub.darcs.net/darcs/darcs-reviewed​/browse/src/Darcs/Patch/V1/Commute.hs#396
01:13 owst Here's an idea: "hunk history". Show all the patches affecting a given line-range in a file, back to when the line-range was created.
01:14 owst I'm trying to work out what's so clever about cleverCommute; "hunk history" might give me clues, rather than annotate, which doesn't (in this case) give me much useful information
01:27 tranma joined #darcs
01:27 * owst was wondering wtf was up with the witnesses for Merger, then I saw the comment "Merger constructor of Patch[...] where the witnesses seem odd" :-)
01:36 iago joined #darcs
01:53 sm joined #darcs
01:56 owst wtf, actualMerge computes both sides of the commuting square http://hub.darcs.net/darcs/darcs-reviewed​/browse/src/Darcs/Patch/V1/Commute.hs#282, throws away one side, only for it to be recomputed in merge http://hub.darcs.net/darcs/darcs-reviewed​/browse/src/Darcs/Patch/V1/Commute.hs#153
01:56 owst Oh, that comment suggests there used to be some sort of sanity-checking comparision
01:57 * owst searches changes
02:00 tranma joined #darcs
02:03 owst No such luck, sigh
02:03 owst This code history is painful
02:03 favonia joined #darcs
02:05 owst Aha, glump09 is the version of glump that was originally for 0.9 mergers, whatever that means
02:11 owst WTF is unwinding all about?! :-s
02:12 owst trueUnwind just looks like magic.
02:20 mizu_no_oto joined #darcs
02:23 * owst wonders where all the .lhs commentry that used to be in the merger code went - it's pretty useful!
02:24 intripoon_ joined #darcs
02:28 owst eek, 862MB of resident for obliting back to 2005 darcs.net... slooooowly
02:29 tranma joined #darcs
02:32 owst 7.5 mins and a peak of ~930MB resident for darcs ob --from-patch 'Split Patch up more'
02:37 owst bingo: "The conflict resolution code (glump)"
03:13 mizu_no_oto joined #darcs
05:04 M30W joined #darcs
05:04 M30W Heh...
05:04 M30W Who's simon?
05:05 M30W HTML on hub.darcs.net kinda messed up. All the >'s are on the next line with the next tag.
05:32 Heffalump you want sm
05:56 jochu joined #darcs
06:11 M30W How can you use a dpatch? O.o
06:14 M30W apply -i :)
06:15 donri joined #darcs
06:20 favonia joined #darcs
06:22 Heffalump M30W: the >s being on the line with the next tag is the way many Haskell HTML/XML pretty-printing libraries work. I can't remember why
06:22 Heffalump I think it's valid, and it makes printing out attributes simpler, IIRC
06:23 * M30W is trying to get darcs to compile
06:24 M30W But it doesn't want to build with ghc 7.4 however; I'm trying to apply this patch: http://bugs.darcs.net/patch987 with.. No luck.
06:24 M30W Heffalump: ^
06:24 M30W 7.6 *
06:28 Heffalump it doesn't build with 7.6.* yet
06:29 Heffalump 7.4 should work
06:29 Heffalump if you really want to use 7.6, you can use the top of the 2.8 release branch, but the handling of old repositories with non-UTF8 encoded metadata will be broken
06:29 Heffalump s/top of the 2.8/tip of the 2.8/
06:29 Heffalump http://darcs.net/releases/branch-2.8/
06:30 Heffalump patch987 is for HEAD (what you get from http://darcs.net) and is already in the repo - but one of the patches got rolled back because of the character set brokenness
06:31 Heffalump you can unpull the one called "ROLLBACK: use bytestring-handle instead of haskeline for encoding" if you want 7.6 support on head (with the same broken handling of old repos)
06:34 favonia joined #darcs
06:36 M30W Heffalump: So get and convert?
06:36 M30W Or can't it even get?
06:38 tranma joined #darcs
07:08 M30W ./darcs push always leads to "darcs: fd:9: hClose: resource vanished (Broken pipe)" (always as in; no matter what I try) http://darcs.net/Binaries < Just got the GNU/Linux standalone binary.
07:09 M30W 32-bit binary.
07:11 tranma joined #darcs
07:45 tranma joined #darcs
08:18 tranma joined #darcs
08:51 tranma joined #darcs
09:25 tranma joined #darcs
10:39 raichoo joined #darcs
11:02 owst joined #darcs
13:17 tranma joined #darcs
13:21 * owst doesn't understand why you'd want to consider patches as forming an inverse semi-group, rather than just a group. We have an identity patch, and all patches have inverses, so what's the problem?
13:23 owst I suppose it depends whether p;p^-1 is always the same as the identity patch. I can't think of an example where it isn't
13:31 M30W joined #darcs
13:41 M30W joined #darcs
13:42 mizu_no_oto joined #darcs
13:50 tranma joined #darcs
14:08 Guest46714 joined #darcs
14:19 favonia joined #darcs
14:23 tranma joined #darcs
14:33 alpounet owst, he answered
14:42 mizu_no_oto joined #darcs
14:56 tranma joined #darcs
14:58 owst yup saw that, I guess I didn't think about the fact that you can't restrict composition to those patches that are able to be composed
15:30 tranma joined #darcs
15:34 ronny joined #darcs
15:37 ronny hi
15:46 mizu_no_oto joined #darcs
15:52 owst hi ronny
15:56 ronny 2 things i wonder about 1. state of patch theory v2, 2. documented interfaces to interface with darcs from programs
15:58 alpounet for 2/, darcs can be used as a library
16:00 ronny alpounet, im not using haskell as the interfacing language
16:00 alpounet oh.
16:03 ronny oh, and another thing, good ways to communicate with git/hg?
16:03 tranma joined #darcs
16:07 alpounet there's darcs-bridge for this, but it has a few issues
16:12 ronny hmm
16:12 ronny does darcs have some kind of "meta-patch" which allows to attach methdata + some wished order to patches?
16:12 donri ronny: darcs can output xml from most commands, might help interfacing outside haskell
16:14 ronny hmm, might be some help
16:16 owst the bridge is mostly useless as a bridge, until you hear otherwise
16:26 ronny hmk
16:27 ronny owst, ah, i was your soc project?
16:31 M30W_ joined #darcs
16:36 tranma joined #darcs
16:42 owst it was - turns out, it was tricky!
16:43 owst I am very very slowly coming to understand how to fix things, but it's not something I find much time to work on, unfortunately
17:10 tranma joined #darcs
17:43 tranma joined #darcs
17:58 sm hi M30W
17:59 sm yeah that's normal, some frameworks emit html that way so that unintended whitespace between tags doesn't affect page layout
18:00 mizu_no_oto joined #darcs
18:12 M30W sm: Hello
18:12 M30W Ha
18:12 M30W I had a feeling it was by that reason
18:13 M30W sm: The whole site is JUST darcsdon correct?
18:13 M30W Exc the database and darcs folders ofc
18:17 tranma joined #darcs
18:17 ronny2 joined #darcs
18:22 sm__ joined #darcs
18:25 M30W sm: ping
18:26 sm M30W: eh.. what ? yes hub.darcs.net is darcsden
18:26 M30W sm: Great code ^_^
18:27 M30W Oh and I noticed a little issue. Forgot to report it; On comit changes (the page) there is a whitespace issue on the first line for the changes
18:28 gienah joined #darcs
18:28 sm yeah that's in the issue tracker
18:28 M30W Okay. :)
18:28 sm alexsuraci wrote it, nifty eh ?
18:29 M30W Very
18:29 M30W Oh (I'm new with darcs) how can you add modified files? I expected just darcs add file but it doesn't take it?
18:31 M30W alexsuraci: Love your code man ^_^
18:31 sm M30W: darcs add (now you'll see it in darcs whatsnew -s) then darcs record (now you'll see it in darcs changes)
18:32 M30W Ahh
18:32 M30W ofc
18:32 M30W So it adds all files that are already added but adds their modifications :D
18:32 M30W Thanks
18:33 M30W sm: Hmm
18:33 M30W M ./readme +1
18:34 M30W Ah.
18:34 M30W Works :) I looked at it as m not M
18:34 M30W Needs colours haha
18:34 * sm off to work, bbl
18:34 M30W Cya sm :)
18:39 gienah joined #darcs
18:39 gh___ joined #darcs
18:40 gh___ hi
18:40 lambdabot gh___: You have 1 new message. '/msg lambdabot @messages' to read it.
18:51 tranma joined #darcs
18:57 schlaftier joined #darcs
19:24 markstos joined #darcs
19:24 sm__ joined #darcs
19:25 tranma joined #darcs
19:47 owst joined #darcs
19:58 tranma joined #darcs
20:13 M30W joined #darcs
20:13 M30W joined #darcs
20:32 tranma joined #darcs
20:53 gh___ owst, about http://irclog.perlgeek.de/​darcs/2013-01-08#i_6308467 , I think these comments went into a chapter of the latex manual that used to live in the darcs repository
20:53 gh___ owst, and that got kicked from it ~6 months ago
20:54 gh___ owst, it was a chapter called "patch theory" IIRC, that was not build with the manual, hence you won't find it in the bigpage.html that is currently on darcs.net
20:56 mizu_no_oto joined #darcs
21:06 tranma joined #darcs
21:25 favonia joined #darcs
21:40 tranma joined #darcs
22:15 tranma joined #darcs
22:49 tranma joined #darcs
23:00 M30W joined #darcs
23:00 mizu_no_oto joined #darcs
23:04 jochu joined #darcs
23:23 tranma joined #darcs
23:57 tranma joined #darcs

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