Camelia, the Perl 6 bug

IRC log for #darcs, 2012-12-15

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

All times shown according to UTC.

Time Nick Message
00:29 kmels joined #darcs
00:39 gpiero left #darcs
00:45 madjestic should not darcs be a part of standard library?
00:52 owst Eh? Which standard library? What do you mean?
00:59 kmels joined #darcs
01:40 kmels joined #darcs
01:40 mornfall joined #darcs
01:49 sm joined #darcs
01:59 xymox joined #darcs
01:59 dleverton_ joined #darcs
01:59 teratorn joined #darcs
02:01 mizu_no_oto joined #darcs
02:03 pierreee joined #darcs
02:04 kmels joined #darcs
02:04 mornfall joined #darcs
02:05 donri joined #darcs
02:05 mulander joined #darcs
02:22 hualet joined #darcs
02:29 schlaftier joined #darcs
02:39 kmels joined #darcs
02:43 mizu_no_oto joined #darcs
02:47 hualet` joined #darcs
02:52 intripoon joined #darcs
02:57 hualet joined #darcs
03:22 alexsuraci` joined #darcs
03:43 astor joined #darcs
04:14 astor left #darcs
07:52 kowey joined #darcs
07:52 kowey good morning!
07:52 * kowey tries his GitHub-pull-request => hub.darcs.net again
07:52 kowey this time using darcs-bridge apply (I think) on the darcs end
07:53 kowey and then pushing via bridge to git, effectively doing a git rebase
07:53 kowey blithely undid my first attempt on GitHub, (a) because not too many followers and (b) eh, they're Git users; they're used to randomness
07:59 kowey ... so, method
08:00 kowey 0. repos are: antfarm (darcs), antfarm_bridge/antfarm, antfarm_bridge/antfarm_git, antfarm_git
08:00 kowey 1. cd antfarm; darcs-fastconvert apply . foo.patch (available as email attachment in GitHub notification)
08:00 kowey 2. darcs push ../antfarm_bridge/antfarm
08:01 kowey 3. darcs-fastconvert sync ../antfarm-bridge git
08:01 kowey 4. cd ../antfarm_git (my local one)
08:02 kowey 5. git pull bridge master (bridge was created with something like git remote add bridge ../antfarm_bridge/antfarm_git)
08:02 kowey 6. git push (to GitHub)
08:02 kowey 7. (oops), cd ../antfarm; darcs push (to hub.darcs.net)
08:10 kowey (so whilst I can't offer any work on darcs, I think what I will do is just try and use darcs-bridge and experiment with darcs-centric GitHub usage, ie. keep using darcs, but try to meet people in GitHubia)
08:10 kowey and just be noisy like i am now
08:11 kowey (but not be too vocal about outside of #darcs as owst doesn't want publicity on this yet due to branchy issues)
08:24 kowey well, ok maybe a little vocal https://plus.google.com/u/0/114991​347543804898741/posts/8zfRKcqdPRK
09:41 drostie joined #darcs
10:06 owst joined #darcs
10:20 madjestic joined #darcs
10:37 schlaftier joined #darcs
12:01 amgarchIn9 joined #darcs
12:23 kmels joined #darcs
13:38 mizu_no_oto joined #darcs
13:42 drostie joined #darcs
14:47 kmels joined #darcs
15:15 kmels joined #darcs
15:32 owst joined #darcs
15:42 kmels joined #darcs
15:47 nomeata joined #darcs
16:03 alpounet hi guys, coming to bug you a bit with scoutess. so, we recently discovered 'darcs show index' - is that a correct/decent way of getting some kind of hash for the current state of the repo, when applied to the root dir of a darcs repo (or even a subdir of the repo) ?
16:09 kmels joined #darcs
16:10 mizu_no_oto joined #darcs
16:33 kmels joined #darcs
16:33 dons joined #darcs
17:11 Heffalump alpounet: off the top of my head, I'd guess not
17:11 Heffalump what kind of hash do you actually need? We've been mulling over the idea for ages, so I'd be willing to explore actually adding something as needed
17:12 alpounet Heffalump, i still have the trick you told us about a few months ago, somewhere in my logs. but one of the scoutess devs discovered 'darcs show index' and it may fulfill our needs, but we're not sure.
17:12 Heffalump darcs show index is affected by local changes, btw.
17:13 Heffalump also, I think it's only a hash of the tree, rather than the patch history
17:15 alpounet Heffalump, hmm, rather than re-explaining, let me just point you to our discussion in April: http://irclog.perlgeek.de/​darcs/2012-04-10#i_5424092
17:16 alpounet it's not really long
17:19 Heffalump thanks, i was about to ask for that but got distracted :-)
17:20 Heffalump I see, ok. So I think darcs show index is definitely inappropriate.
17:21 Heffalump did you get as far as trying out the tagging approach?
17:24 alpounet Heffalump, unfortunately, we have been busy with other things, and then had to solve bigger issues in scoutess. but i'm getting back at this little by litle.
17:25 Heffalump no problem
17:26 Heffalump dinner, biaw
17:31 alpounet so everytime we pull from a repo, we tag what we have. but then since we don't necessarily have control over the "distant" repository, can we compare the tagged, local repo and the distant repo's state?
17:50 madjestic joined #darcs
17:52 mizu_no_oto joined #darcs
17:54 madjestic hey guys, trying to get darcs working with cabal (after wiping .cabal and .ghc and getting 'cabal update' and 'cabal install cabal-install')
17:54 madjestic https://pastee.org/g84s7
17:54 madjestic this is what I get.  Basically darcs fails to install because of haskeline dependency
17:55 madjestic I do 'cabal unpack haskeline', edit the haskeline.cabal file, fix the dependency ranges, do 'cabal install'
17:56 madjestic I do 'ghc-pkg list' and can see haskeline package (listed in the pastee link above)
17:56 madjestic however when doing 'cabal install darcs', I still get the error, because of haskeline dependency (as seen in the pastee above)
17:57 madjestic why is that?  Is not haskeline get properly installed?  Or does darcs require it rebuilt with some extra dependencies which fail?
18:20 madjestic any ideas?
18:29 madjestic13 joined #darcs
18:30 * Heffalump reappears
18:31 Heffalump madjestic13: are you using GHC 7.6?
18:31 Heffalump alpounet: so you need a way to see if the distant repo has new stuff without actually pulling from it?
18:33 Heffalump madjestic13: you need to look in the output to see why hashed-storage and haskeline failed
18:55 alpounet Heffalump, we can pull in a temporary dir or w/e, I guess we can't really know without pulling? :)
19:10 madjestic hey Heffalump I am using ghc  7.4.1
19:10 madjestic should I upgrade to 7.6?
19:12 madjestic also, is it worth to wipe haskell platform and start from scratch?
19:15 iago joined #darcs
19:29 owst joined #darcs
19:32 gpiero joined #darcs
19:43 kmels joined #darcs
19:48 Heffalump madjestic: no, don't use 7.6, darcs doesn't support it yet
19:48 Heffalump and yes, it might be if you don't have anything else you care about
19:48 favonia joined #darcs
19:59 Heffalump alpounet: well, pull knows if there's anything to pull before it actually does i
19:59 Heffalump t
20:00 Heffalump what do you need to do if there's anything new in the distant repo?
20:45 alpounet Heffalump, pull and generate a "version" string from it
20:46 Heffalump what's the version string for, identifying what you tested?
20:56 raichoo joined #darcs
21:17 alpounet Heffalump, yes, uniquely identifying to be accurate
21:18 alpounet Heffalump, it's for scoutess, the build-bot kind of thing. we want to be able to test the build against HEAD versions of the dependencies, that requires us to assign some kind of unique identifier to the state of a darcs repo at a given time
21:19 Heffalump darcs itself uses last tag + count of patches to give a human readable, albeit not globally unique, identifier
21:22 alpounet hm
21:44 mizu_no_oto joined #darcs
22:22 schlaftier joined #darcs
22:23 nomeata joined #darcs
22:43 Heffalump alpounet: otherwise the tag idea should work fine, I think
22:43 Heffalump alpounet: again imagining that you're scripting darcs, you can run an interactive pull to find out whether there are any new patches at the remote end, and then use the tags to actually identify repository states
22:48 gpiero left #darcs
23:14 alpounet Heffalump, i'm going to talk with stepcut about this when we have a chance
23:15 alpounet we'll figure out what's best for our use case, and come bug you if we have trouble doing so
23:15 alpounet i guess we'll stick with the tag idea though
23:26 drostie joined #darcs

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