Camelia, the Perl 6 bug

IRC log for #darcs, 2013-09-04

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

All times shown according to UTC.

Time Nick Message
00:21 arpunk joined #darcs
01:38 intripoon_ joined #darcs
01:38 mizu_no_oto joined #darcs
01:51 srinup joined #darcs
04:30 whaletechno joined #darcs
05:22 arpunk joined #darcs
05:55 mulander joined #darcs
06:12 adnap_ joined #darcs
06:31 gbeshers joined #darcs
06:55 mlamari_ joined #darcs
07:00 raichoo joined #darcs
07:05 Igloo joined #darcs
07:18 stepcut joined #darcs
08:50 donri joined #darcs
09:24 mulander joined #darcs
09:47 donri joined #darcs
09:54 shafi4442 joined #darcs
09:54 shafi4442 hi
10:21 arpunk joined #darcs
10:45 idnar joined #darcs
11:24 mlamari joined #darcs
12:48 srinup joined #darcs
13:04 srinup_ joined #darcs
13:07 mizu_no_oto joined #darcs
13:28 srinup joined #darcs
14:14 arpunk joined #darcs
15:12 srinup joined #darcs
15:13 mizu_no_oto joined #darcs
15:21 arpunk joined #darcs
15:59 karlpinc joined #darcs
16:00 karlpinc How do I get a list of diffs of each recorded patch?  I want to scan through them to look for where I made a specific sort of change.
16:04 int-e darcs changes -v  lists the hunks with the patch descriptions.
16:05 karlpinc int-e: Thanks.  I don't know why I didn't see that.  I've done it before.
16:05 int-e well, "-v         --verbose               give verbose output" is not the most helpful of flag descriptions.
16:07 karlpinc Hey, but it's terse!  ;-)
16:09 sm with latest darcs builds, you'll probably want darcs log -v
16:10 sm is interactive-by-default changes going to stay, or be rolled back I wonder
16:10 int-e oh, so  darcs log  is no longer an alias for  darcs changes?
16:11 sm currently darcs log does what changes always did, and darcs changes does what changes -i did (interactive viewing)
16:32 adnap joined #darcs
17:31 favonia joined #darcs
17:35 stulli joined #darcs
17:44 burp joined #darcs
17:46 WarzoneCommand joined #darcs
17:46 WarzoneCommand is there propper documentation of hashed-storage somewhere?
17:47 WarzoneCommand I think I may want to use it for something. but it seems some of the functions don't work as advertised
17:47 WarzoneCommand or I am understanding something incorrectly
17:48 sm WarzoneCommand: probably just the source, or you could as mornfall
17:48 sm ask
17:50 Heffalump I've used it a fair bit (outside of darcs) and I also found the API somewhat idiosyncratic
17:52 WarzoneCommand basically I have an Index, which I've just read and/or created
17:52 WarzoneCommand and I'm trying to get an fully expanded tree
17:53 WarzoneCommand so something like: "do idx <- readIdx .... ; stub <- updateIndex idx  ; tree <- expand stub"  should do that right?
17:54 Heffalump that sounds plausible, but I haven't used the index myself
17:54 WarzoneCommand the tree I get by doing that is apparently empty
17:55 Heffalump whereas if you read directly you get a valid tree?
17:55 Heffalump (got to go, back in a while)
17:56 WarzoneCommand yes
17:57 WarzoneCommand my testing code: http://lpaste.net/92572
17:57 sm WarzoneCommand: you saw the note in the haddock about stubs, right
17:57 sm yes I guess you did
17:57 WarzoneCommand sm: that is why I'm calling expand
17:58 WarzoneCommand or at least, I would expect that expand fully expands the tree as advertised
18:01 sm expand: "This is the basic implementation, which may be overriden by some Tree instances (this is especially true of the Index case). "
18:01 sm sounds ominous
18:03 WarzoneCommand I fail to see how Index is able to do that. Nor do I see any evidence that it does so/tries to do so
18:06 WarzoneCommand by the way (the index-file itself is created on the filesystem, and looking at it it seems to contain something like a list of files in it
18:14 mizu_no_oto joined #darcs
18:14 Heffalump have you tried not calling expand?
18:14 Heffalump I can't see any cases in darcs where the index is read then expand is called
18:14 Heffalump (I agree that I would have also expected you would need to call it!)
18:15 WarzoneCommand well not calling expand would be identical to printing idxTree (see the Hpaste)
18:15 WarzoneCommand which also yields an empty tree
18:18 Heffalump do the paths you can see in the index-file make sense as paths inside /Users/frank/tmp/synced ?
18:20 WarzoneCommand well its a binary file, but if I cat it it does seem to contain the two filenames of the files in ~/tmp/synced
18:23 WarzoneCommand maybe the mmapping goes wrong somehow
18:25 WarzoneCommand readIndex and updateIdx seem to return and EmptyIndex if mmapIdex fails
19:13 markstos joined #darcs
19:14 markstos I suspect I'm running into an issue with the new "patience diff" code which has this symptom: I "record" all the "hunks" in a file, and later do a "darcs whatsnew", only to find that that are some "whatsnew" bits, such as empty lines. Anyone else running into behavior like that with the patience diff code?
19:15 markstos Sorry, no reproducible test case at this point, just a hunch.
19:22 markstos Ok, here's a specific case that is suspect: https://gist.github.com/mar​kstos/ac0ecfad05bf25a95a3f
19:22 markstos Perhaps I did something unexpected myself that I didn't realize, but it appears to be odd patience-diff behavior to me.
19:49 Heffalump are you sure your editor didn't mess around with the newlines at the end of the file?
19:49 Heffalump though the diff still looks slightly erroneous to me, at worst I'd have expected it to be in two chunks (with the TODO line itself left out)
19:50 markstos Heffalump: the end of the file would be much later.
19:50 markstos But no, I'm certain the problem is not on my end.
19:50 markstos I mean /not certain/
19:50 markstos the problem could be me.
19:51 markstos But in the course of series recording several patch hunks, strange bits like this came up several times.
19:51 markstos I'll keep an eye on it, and see if I can capture a reproducible case.
20:21 arpunk joined #darcs
21:42 mornfall hmm
21:43 mornfall WarzoneCommand: Which version do you use?
21:44 mornfall WarzoneCommand: when the unexpanded tree is empty, so must be the expanded one
21:44 mornfall WarzoneCommand: expansion will only replace existing stubs
21:50 mornfall WarzoneCommand: the main problem with hashed-storage is that it's not really maintained; darcs decided it wants to stick with 0.5 which is deprecated and there's no driving force to continue work on the 0.6/fslib codebase...
21:51 mornfall WarzoneCommand: all I can offer is http://repos.mornfall.net/fslib and http://repos.mornfall.net/hsbackup
21:52 mornfall WarzoneCommand: neither is quite finished though
21:52 javier_rooster joined #darcs
21:53 mornfall also there's http://repos.mornfall.net/gorsvet/ which uses fslib too
21:54 mornfall apparently, I have a number of unrecorded changes in my fslib working copy from two years ago
21:54 mornfall WarzoneCommand: I guess your best bet is not to use hashed-storage
21:56 mornfall maybe if darcs wakes up from its slumber one day
21:58 mornfall which reminds me, did anyone ever get to fix that MIME encoding bug I pointed out in April?
21:59 javier_rooster joined #darcs
22:01 mornfall Cf. http://irclog.perlgeek.de/darcs/2013-04-19
22:01 mornfall the bug *is* annoying, I have to hand-edit almost every bundle I get
22:04 javier_rooster joined #darcs
22:09 javier_rooster joined #darcs
22:15 javier_rooster joined #darcs
22:15 sm hi mornfall
22:16 sm I suspect the slumbering darcs have moved to latest hashed-storage by now if it was compelling.. what prevented that ?
22:17 sm would have
22:20 javier_rooster joined #darcs
22:25 javier_rooster joined #darcs
22:28 mizu_no_oto joined #darcs
22:31 javier_rooster joined #darcs
22:36 javier_rooster joined #darcs
22:41 javier_rooster joined #darcs
22:47 javier_rooster joined #darcs
22:52 javier_rooster joined #darcs
22:58 javier_rooster joined #darcs
23:03 javier_rooster joined #darcs
23:08 javier_rooster joined #darcs
23:14 javier_rooster joined #darcs
23:19 javier_rooster joined #darcs
23:25 javier_rooster joined #darcs
23:30 javier_rooster joined #darcs
23:39 javier_rooster joined #darcs
23:44 javier_rooster joined #darcs
23:50 javier_rooster joined #darcs
23:55 javier_rooster joined #darcs

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