Camelia, the Perl 6 bug

IRC log for #darcs, 2013-07-23

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

All times shown according to UTC.

Time Nick Message
00:51 mizu_no_oto joined #darcs
01:34 sm aha, finally we see the real darcs users! http://hub.darcs.net:8901/
01:35 sm 144 of 319 user accounts actually have some repos. Ignoring private repos, which aren't counted yet
01:46 sm 125 private repos. Hmm, that's more than expected
01:47 bfrank ha
01:47 bfrank I created a repo earlier today
01:48 bfrank took a long time, with record --pipe
01:48 bfrank but got it done
01:51 sm woah, why did you need that
01:55 bfrank converted svn repo
01:59 sm bfrank: next time you can maybe get it into fastimport format and use http://hackage.haskell.org​/package/darcs-fastconvert
01:59 sm but congrats
02:06 sm and good night
03:06 dolio Time to fork all my repos to get to the top of the list.
03:26 idnar joined #darcs
03:34 preflex_ joined #darcs
03:39 mizu_no_oto joined #darcs
03:41 kofno joined #darcs
03:52 edwardk joined #darcs
04:33 mizu_no_oto joined #darcs
06:05 lelit joined #darcs
06:14 edwardk joined #darcs
07:09 amgarchIn9 joined #darcs
08:10 raichoo joined #darcs
08:47 owst Hmm, I'm having some trouble getting www.darcs.net: (Couldn't fetch `0000017073-ac60cee205c110af2f54efcc0ca​27ce2b2e04cc045d03f06bc89d4d75ddf7594'
08:47 owst in subdir pristine.hashed
08:47 owst Anyone else able to get www.darcs.net?
08:49 owst Err, there's something fooey going on, that hashname is the "Starting With" inventory of http://darcs.net/_darcs/hashed_inventory
09:07 owst Hmm, I'm not sure what's going on, darcs retrieves that hashed file from inventories, but then later tries to get it from pristine.hashed too. I can't see why that would be happening
09:07 owst The inventory file now exists locally
09:35 gh__ joined #darcs
09:36 gh__ hi
09:36 owst hi gh__
09:37 owst gh__: Can you try and get www.darcs.net?
09:37 owst I'm getting a weird failure (see the scrollback for details)
09:40 gh__ yes, I just tried get with --no-cache --lazy, and it worked fine
09:40 owst Well that is weird then.
09:43 gh__ can you reproduce it? what about passing the --debug-verbose flag?
09:43 gh__ could it be related to packs?
09:45 owst Weird, with --lazy it's fine
09:45 owst Otherwise `darcs get http://www.darcs.net` is enough to reproduce it here
09:45 owst I'm trying to build a debug version, so I can get a stacktrace where the error occurse
10:26 whaletechno joined #darcs
10:51 nomeata joined #darcs
11:10 owst alexsuraci_: had a chance to look at darcsden.com's SSH server?
11:26 kofno joined #darcs
11:45 pointfree-w530 joined #darcs
12:41 gal_bolle joined #darcs
13:01 konundra joined #darcs
13:02 owst Just an FYI, I've started reviewing Jose's patch, I'll try and finish it this evening.
13:02 owst (the monster patch, adding patience diff in most places)
13:08 gh__ owst, wait, I'm on it
13:11 amgarchIn9 joined #darcs
13:12 amgarchIn9 joined #darcs
13:13 amgarchIn9 joined #darcs
13:17 mizu_no_oto joined #darcs
13:29 owst gh__: Oh ok, damn, too late!
13:29 owst I don't think the Rebase changes are necessary btw
13:48 bfrank__ joined #darcs
13:49 bfrank__ Is tailor the tool for converting to darcs?
13:50 sm bfrank__: no, darcs-fastconvert is the best. tailor is one of several outdated tools
13:51 bfrank__ where is darcs-fastconvert?
13:51 sm http://hackage.haskell.org​/package/darcs-fastconvert , I mentioned it yesterday
13:52 bfrank__ ahh
13:52 bfrank__ can it do svn?
13:52 sm afk
13:54 donri joined #darcs
14:09 amgarchIn9 joined #darcs
14:45 owst bfrank__: depends if svn can export in the (git) fast-import format
14:47 owst gh__: nice review.
14:47 gh__ owst, I'm not sure yet about your observation on rebase
14:47 Heffalump 'lo
14:47 owst I think my only concern was the extra complexity encountered when passing through the diff algorithm to Rebase, I'm not sure the user will ever care to change it, and it should probably be fixed
14:47 gh__ owst, you can comment about it if you want, or wait for the next iteration of the patch
14:47 owst Hey Heffalump
14:47 gh__ hi Heffalump
14:47 owst I was going to defer to Heffalump's opinion anyway
14:48 owst s/fixed/a fixed choice/
14:48 Heffalump Jose did email me to ask if patience diff should have any practical impact on rebase, and I didn't think it did
14:48 edwardk joined #darcs
14:49 jlneder joined #darcs
14:50 owst I suppose it could affect the pushing of patches "in", if one diff gave you hunks that commuted and the other didn't, but I'm not convinced by that (I haven't thought about an example)
14:50 owst And I'm probably forgetting how that Rebase code works, anyway
14:52 Heffalump skimming email, I don't think removing the double canonize calls is appropriate - I remember they were needed for something
14:58 jlneder hi, for what i see and minimally test, like i say in my last post, the diff algorithm is used only to get a patch from the unrecorded changes from the current repo.
15:03 jlneder About the canonize calls, for what i see, before the patch of the double canonize calls is was called in this way "sortCoalesceFL . concatFL . mapFL_FL (canonize diffa)" and now i put in this other way "concatFL . mapFL_FL (canonize diffa) . sortCoalesceFL". Correct me if i'm wrong but from what i understand of the comments in the code and the code itself sortCoalesceFL must be first for the canonize to succeed.
15:03 Heffalump got to go out for a bit, I'll be back in a while
15:04 Heffalump jlneder: it'd be useful to have a brief summary of why the diff algorithm needs to be passed around in Rebase now
15:04 Heffalump it's not immediately obvious where the need originates, but I can well imagine that the somewhat tangled darcs code might well require it
15:05 jlneder After i get this theory i tested with issue525.sh and it is running ok, so i assumed i was right about it.
15:06 jlneder the diff algorithm make hunks about the unrecorded changes to check for conflicts and make merges.
15:08 owst Heffalump: it's needed since there's an eventual call to canonize (which now takes the diff algorithm)( in (IIRC pushIn, or whatever it's really called)
15:09 Heffalump oh, I see
15:09 Heffalump ok, so it is actualy used by rebase (when merging two fixups), I hadn't thought of that
15:09 Heffalump well, if the tests pass, that's fine
15:09 owst And pushing things past a rebase too, no?
15:09 Heffalump not sure
15:09 owst sorry, past a fixup
15:09 Heffalump (really gone)
15:09 owst ok
15:12 burp_ joined #darcs
15:19 owst joined #darcs
15:19 mizu_no_oto joined #darcs
15:36 amgarchIn9 joined #darcs
16:03 Heffalump pushing things past a fixup and merging two fixups is kind of the same thing :-)
16:05 owst That's alright then :-)
16:07 Heffalump when pushing A past B, I think it tries to merge A and B, then to commute A and B, then it gives up and leaves A before B
16:12 owst What is an example of not Merge(A,B) but Commute(A,B) ?
16:19 konundra joined #darcs
16:37 gal_bolle owst: these don't have the same type: Merge : Patch a b -> Patch a c -> Whatever while Commute : Patch a b -> Patch b c -> Whatever
17:07 amgarchIn9 joined #darcs
17:20 mizu_no_oto joined #darcs
17:26 javier_rooster joined #darcs
17:48 pr joined #darcs
18:13 alexsuraci_ darcsden ssh server should be back up; God smote it
18:15 mtp chunk that shit into runsv :)
18:15 bfrank__ how do I build darcs-screened?
18:25 lelit joined #darcs
18:40 Heffalump bfrank__: are you familiar with cabal?
18:41 bfrank__ no, but I have it installed
18:46 Heffalump if you want to build darcs in the local directory: "cabal configure" then "cabal build"
18:46 Heffalump if you want to build it then install it into ~/.cabal/bin , "cabal install"
18:53 bfrank__ hm
18:53 bfrank__ let me try that
18:55 bfrank__ telling me some dependencies are missing
18:55 bfrank__ base16-bytestring
18:55 bfrank__ cryptohash
18:55 bfrank__ data-ordlist
18:57 owst joined #darcs
18:57 owst bfrank: `cabal install --only-dependencies`
18:58 bfrank__ apparently I need to run cabal update or something to update my package list?
18:59 amgarchIn9 joined #darcs
18:59 bfrank__ now, I think it might be installing dependencies
19:02 bfrank__ I think it might be compiling darcs now
19:06 bfrank__1 joined #darcs
19:25 mizu_no_oto joined #darcs
19:37 bfrank__1 how do you do log -p in darcs again?
19:42 Heffalump what does it do?
19:43 Heffalump changes, perhaps?
19:43 mtp darcs changes -v
19:52 favonia joined #darcs
19:54 bfrank__1 hmm
19:56 bfrank__1 I think that was what I was looking for
19:57 bfrank__1 thanks
19:57 owst joined #darcs
20:01 bfrank__1 that would have been nice to see on the rosetta stone page
20:07 owst Isn't that on a (the darcs.net) wiki?
20:07 owst If so, add it!
20:07 bfrank__1 yeah, rosetta stone is on the darcs.net wiki
20:08 bfrank__1 also, why does darcs record have a --unified option, but darcs pull doesn't?
20:15 bfrank__1 added it to the wiki
20:15 owst_ joined #darcs
20:40 sm g'day all
20:47 bfrank__1 hey
21:27 mizu_no_oto joined #darcs
22:09 amgarchIn9 joined #darcs
22:14 konundra joined #darcs
23:40 mizu_no_oto joined #darcs

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