Camelia, the Perl 6 bug

IRC log for #darcs, 2013-02-20

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

All times shown according to UTC.

Time Nick Message
00:25 dixie joined #darcs
01:22 favonia joined #darcs
01:27 intripoon joined #darcs
02:18 intripoon joined #darcs
02:38 intripoon_ joined #darcs
03:36 mizu_no_oto joined #darcs
04:45 preflex_ joined #darcs
04:51 mizu_no_oto joined #darcs
05:32 donri joined #darcs
05:55 jyyou joined #darcs
06:15 Thaalos joined #darcs
06:35 carter joined #darcs
06:53 lelit joined #darcs
07:07 Igloo joined #darcs
07:24 Igloo joined #darcs
07:34 amgarchIn9 joined #darcs
07:52 kmels_ joined #darcs
08:09 raichoo joined #darcs
08:10 amgarchIn9 joined #darcs
08:16 amgarchIn9 joined #darcs
08:34 gh_ joined #darcs
08:36 Igloo joined #darcs
09:33 florent_ joined #darcs
09:50 owst joined #darcs
09:55 jnaimard hi all
09:56 jnaimard hi owst, branch on hub.darcs.net coming soon(-ish)
09:56 owst cool
10:55 amgarching Hi! I do "darcs cha file" and it complains that "Can't find changes prior to:" followed by a patch name. What is going on?
10:57 jnaimard is your repository lazy?
10:58 jnaimard did you use get --lazy or hit Ctrl-C durinng darcs get?
10:58 iago joined #darcs
11:05 iago did somebody build darcs HEAD in Debian?
11:05 owst iago: when?
11:05 iago recently
11:05 owst jnaimard: might be on debian?
11:06 owst I'm on Arch, so no help.
11:06 iago I'm having problems with libiconv, not sure if it is Debian specific
11:06 owst Oh, install curldev or something like that
11:06 owst this came up recently
11:07 owst iago: http://irclog.perlgeek.de/​darcs/2013-02-19#i_6470109
11:11 gh_ I just found this (RCS class that involves git, darcs and others) http://www.cs.tufts.edu/comp/150GIT/
11:11 gh_ from fall 2012
11:12 owst That sounds very cool
11:12 gh_ looks like students were asked to read iago's report from 2011 :-)
11:13 owst " What, if anything, do we know about Darcs merge?
11:13 owst heh
11:13 iago I should have polished the report a little more... :-P
11:13 gh_ a group of students implemented a RCS in haskell as a project for this class https://github.com/jmont/nor
11:16 jnaimard jnaimard: i'm on debian indeed
11:17 iago «We present our model of a simplistic VCS and show how the chosen representation
11:17 iago of patches enables a simple method for conflict detection and resolution when
11:17 iago combining lines of development.
11:17 iago »
11:18 owst merge -- difficult problem
11:18 owst git rebase + merge -- WAT
11:18 owst :-)
11:18 gh_ yeah their claims deserve some examination
11:19 jnaimard I think I did build HEAD yesterday without trouble
11:21 owst They say they can easily convert from parallel patches to sequential patches... "easily"?
11:22 gh_ their git repo has some activity from last weekend
11:23 jnaimard I don't think they care about consistency (that's ok, that's what anagram(ioaleus) does)
11:23 owst hmm
11:24 iago I don't have too much time to read this in depth right now, but it seems that they are not 100% patch-based
11:25 owst No
11:25 owst It's not
11:25 jnaimard they're doing a better merge for git, not darcs (at least, that's what i'm ready to accept for now)
11:25 jnaimard s/not/not something comparable to/
11:26 iago yep, it looks like that
11:27 iago btw, thanks owst, installing libcurl-dev solved the problem
11:27 owst cool
11:27 jnaimard still, the idea that you want to commute as many patches as possible to some common origin before you look at conflicts might have some merit
11:28 jnaimard but they could have opened a book about graphs for ten minutes
11:28 owst jnaimard: ?
11:30 jnaimard their section about maximum conflict set could use words such as bipartite and coupling to be a bit more readable
11:31 owst Oh, there's a paper?
11:31 * owst missed that
11:31 owst Oh yeah, got it
11:32 owst Right, so they're doing git merges, via a darcs-like merge using patches
11:33 * owst was trying to figure it out from the code
11:40 jnaimard page 12, they magically commute all patches to the least common ancestor (<- just say lower bound, go read some math)
11:40 jnaimard I'm not sure how that's possible
11:41 jnaimard ouch page 13
11:41 jnaimard take an effing graph course
11:42 owst jnaimard is fired up today :-)
11:44 * owst goes back to agda.
11:45 jnaimard when you come back, you'll be as fired up as i am (ok, only judging from experience with coq rather than agda)
11:46 owst :-)
11:46 owst I'm trying to do something exceedingly simple, no doubt, but let's see where I get.
11:48 owst (for any nat n, I want to be able to give a definition of BooleanAlgebra (OrdinalSet n) such that the operations are checked at compile-time - e.g. for n = 3, {0,1} || {2} shouldn't be allowed to be {4})
11:48 owst I think it might probably all be defined somewhere already, I suppose it's a dependent-type learning exercise for me :-)
11:52 jnaimard my experience with agda has been limited to my inability to get it to talk with my emacs
11:52 owst I don't use emacs, which might be a bad thing
11:53 jnaimard after reading the paper and looking at the code, i'm really perplex about nor
11:53 owst What about?
11:53 owst (specifically, I mean)
11:53 jnaimard they don't seem to have anything about commuting patches, which seems odd given that they strongly need to do that before merging
11:54 jnaimard since they want all patches in the same context
11:54 jnaimard one of me and them is missing something obvious
11:55 jnaimard aha it might be in rebaseContinue
11:59 jnaimard and teach yourself bibtex
11:59 jnaimard for Knuth's sake
12:22 jnaimard $nor rebase 0d7a6b5493fc28298b9a42b9bf11c155e75bc1cf
12:22 jnaimard nor: Can't Happen
12:22 jnaimard that's what I thought
12:28 jnaimard that said, if/when they implement commutations, I'd be curious to see how well that idea performs
13:45 mizu_no_oto joined #darcs
13:47 favonia joined #darcs
14:19 raichoo joined #darcs
14:54 donri joined #darcs
15:00 nomeata joined #darcs
15:31 mizu_no_oto joined #darcs
16:17 raichoo joined #darcs
16:37 gh_ joined #darcs
17:05 Heffalump jnaimard: I think least common ancestor is a common term used in VCS talk, even if not suitably mathematical :-)
17:06 stepkut joined #darcs
17:06 jnaimard i guess so
17:10 sm g'day all
17:10 sm a small puzzle. I have a cron job that currently looks like this: cd /home/darcsden/users/darcs/darcs-screened; darcs pull -a http://darcs.net/screened >/dev/null # 2>&1
17:11 sm when I test it manually there's no output. When running automatically, some times it generates output like Identifying repository http://darcs.net/screened format
17:11 sm or: Defining set of chosen patches 1 done, 2 queued. DARCS_PATCHES
17:11 sm Writing inventory
17:13 sm I haven't been able to reproduce it yet but I suppose maybe it's writing some stderr output only when there are new patches to fetch
17:13 sm I need it to be silent.. unless there's an error
17:15 * sm tries darcs pull -qa http://darcs.net/screened
17:17 owst sm: you could obliterate a bunch of patches to test it
17:17 owst (i.e. re-pull some old patches)
17:18 sm owst: I couldn't reproduce it that way unfortunately
17:20 owst :-(
17:34 sm next puzzle.. why are certain patches in certain, inactive, repos repeatedly getting their file ownership changed
17:34 sm time to get to the bottom of this
17:36 Heffalump /me appears
17:44 iago Heffalump,   suspend        Select patches to move into *into* a suspended patch at the end of the repo. (darcs help rebase) --- I don't think filling a bug report is worthwhile
17:44 iago (btw, I'm using darcs-reviewed)
17:52 Igloo joined #darcs
17:56 Heffalump iago: ta
18:04 amgarchIn9 joined #darcs
18:08 stepkut joined #darcs
18:24 favonia joined #darcs
18:41 Igloo joined #darcs
18:44 mizu_no_oto joined #darcs
18:56 amgarchIn9 joined #darcs
19:25 mizu_no_oto joined #darcs
19:39 iago joined #darcs
20:08 favonia joined #darcs
21:03 sm confessions of a git skeptic is pretty good: http://news.ycombinator.com/item?id=5251359
21:08 gh_ joined #darcs
21:53 Igloo joined #darcs
22:08 carter joined #darcs
22:28 Igloo joined #darcs
22:57 iago joined #darcs
22:59 jnaimard joined #darcs
23:27 favonia joined #darcs
23:43 schlaftier joined #darcs

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