Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2015-05-21

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

All times shown according to UTC.

Time Nick Message
02:30 mizu_no_oto joined #darcs
03:42 amgarching joined #darcs
03:52 c74d joined #darcs
04:03 Riastradh joined #darcs
06:09 c74d joined #darcs
06:31 c74d joined #darcs
06:43 quarantenne joined #darcs
06:43 lelit joined #darcs
08:14 _loic_ joined #darcs
08:20 amgarching joined #darcs
08:38 gal_bolle joined #darcs
11:39 _loic_ joined #darcs
11:47 siel joined #darcs
12:16 c74d joined #darcs
14:57 karlpinc joined #darcs
14:58 amgarching joined #darcs
15:01 karlpinc What does this mean and what can I do to keep it from happening?  http://paste.debian.net/181644/  I've already run a "darcs check" on the source repo and it passes.  (version 2.8.3 on both ends)
15:01 karlpinc And in another repo (of the same stuff) I get this: $ darcs check
15:01 karlpinc The repository is consistent!
15:01 karlpinc Hash mismatch(es)!
15:01 karlpinc db/functions
15:01 karlpinc index: 9eca10c7db8adb5e1b6822683a8706adbdb50ccf56c39e5f1e3f83a01900c9c8
15:01 karlpinc working: eee6aed81080a56e8ed6a2173b9df5654088ab8ffe7477414fc504c3cfc7e3bd
15:01 karlpinc db
15:02 karlpinc index: 01ea3c1ef5863a7a2c81a58d587998c13f483089015f3dad1fdc95cd2a930efa
15:02 karlpinc working: 8c10f76a78f343cd869aa373c7766fdbba8e4463a50936aa8e4dbc131c000237
15:02 karlpinc chado/gff
15:02 karlpinc index: eebc7b94fc403ac671e8aa563850d7db351e9ed19bd6e968ee78db900738f9b0
15:02 karlpinc working: 70aea5fe1e4e2d8ee80186b092d068d4d91a51b949417303816a445ca5fb7b25
15:02 karlpinc chado
15:02 karlpinc index: 2c25f6a1d360f3d23d9704ef19434d660a89e67b6d976edd9e36b4f20b6633ec
15:02 karlpinc working: 22e0a321b125eb118c2296a78c373e01f59c5b647989fd452be04c6781804cbb
15:02 karlpinc Bad index.
15:02 karlpinc whooops.
15:02 karlpinc Er, ... : http://paste.debian.net/181645/
15:05 maerwald I'd say try to reproduce it with 2.10.0 first...
15:05 karlpinc In the latter paste all 3 are directories, and there's stuff in the directories that are not added to the repo (or recorded obviously)
15:07 karlpinc maerwald: I'm just nervous, having just converted the repos to darcs2.
15:08 maerwald karlpinc: I don't have a specific idea what might be the problem, but it's always better to debug/file bugs against the latest version
15:08 karlpinc maerwald: Understood.  I'm looking to see if I can upgrade with cabel now (on a different box.  Can't mess with the boxes where there's problems.)
15:09 sm karlpinc: it's not much help, but I can tell you I've had errors like that with no ill effects. Perhaps cloning a fresh copy helps
15:10 karlpinc sm: Should I erase my cache?
15:11 sm you could try, or  move it out of the way
15:12 maerwald karlpinc: I'd leave a copy of the repo, so for debugging purposes
15:23 karlpinc Humm.  Just doing a get --no-cache to make a fresh copy of the source repo does not help.
16:03 karlpinc Ok... So having done a "cabal install foo", how do I run "foo".  (I tried "cabal install cabal" since "cabal install darcs" told me I'd break something.)
16:09 maerwald karlpinc: cabal run foo
16:16 Riastradh joined #darcs
16:19 karlpinc maerwald: Thanks.  (I'm getting help getting the latest darcs installed on #haskell.)
16:20 maerwald karlpinc: just "cabal install" and then something like: export PATH="~/.cabal/bin:$PATH"
16:20 maerwald unless you use a sandbox
16:25 karlpinc maerwald: Not getting a lot of help on #haskell vis my darcs problem.  Whcih is "cabal install darcs" says that regex-posix-0.95.1 and regex-compat-0.95.1
16:25 karlpinc will likely break and to use --force-reinstalls.  Seems like a bad idea.
16:25 maerwald I think there is no way around
16:26 karlpinc maerwald: I can always remove ~/.ghc and ~/.cabal later.  So I guess I'll go ahead.
16:57 Riastradh joined #darcs
16:57 thorkilnaur_ joined #darcs
17:04 karlpinc Ok.... So, with 2.10.0 a "darcs get --no-cache" does not give me an error.  When I copy the result back to the 2.8.3 box and try get from that with 2.8.3 I have the same error as before.  Unless someone has another suggestion I think I'll forget about the problems and figure they'll probably go away when upgrade happens eventually.
17:09 karlpinc (I see 2.10.0 gives me color curses.  I _hate_ color.  I see that there's an env variable I can set.  FYI. "darcs check" has permanently altered my text color in the window I'm using.)
17:34 amgarching joined #darcs
18:00 maerwald "I hate color"... errrr? :D
18:00 maerwald life must be quite sad that way :P
18:44 karlpinc maerwald: Yup.  :)  It's a brain thing.  Can't read much when there's color.  Too distracting.
18:57 Heffalump karlpinc: you can delete _darcs/index to "reset" it
18:57 Heffalump are you doing anything with scripting darcs where things are being updated faster than once a second?
18:58 Heffalump karlpinc: with modern cabal, I'd recommend a sandbox to build darcs
18:59 Heffalump though I guess #haskell advised that too
19:02 maerwald karlpinc: I'd guess that there would be something to turn off colors system-wide, instead of per-application
19:03 karlpinc maerwald: Probably frobbing the global curses.  I settle for turning it off per application.
19:03 karlpinc Heffalump: So, getting rid of _darcs/index in my repos would "fix" the problem?
19:04 maerwald karlpinc: debatable whether it makes sense to implement a color-off switcher for all applications instead of once on desktop-level (e.g. in your desktop environment)
19:04 karlpinc problems (one on get and one on check)
19:04 maerwald I am pretty sure you can even do it globally + have exclude rules in e.g. KDE
19:05 karlpinc Heffalump: No.  Darcs should not be updated more than once a second.
19:12 amgarching joined #darcs
19:18 gh_ joined #darcs
19:29 favonia joined #darcs
19:54 favonia joined #darcs
19:58 Heffalump karlpinc: well, it should fix it until something breaks it again. I guess something must be.
20:13 favonia joined #darcs
20:27 favonia joined #darcs
21:02 karlpinc Heffalump: I did convert from darcs1, hashed to darcs2.  There were at least 2 lossy patches.  ?
21:09 Heffalump I guess that could mess up the index
21:09 Heffalump the index is just a cache of the files in your working directory (timestamps and maybe size) so darcs record/whatsnew etc can see what's changed fast
21:41 karlpinc Thanks.
22:24 mizu_no_oto joined #darcs

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