Camelia, the Perl 6 bug

IRC log for #darcs, 2013-02-07

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

All times shown according to UTC.

Time Nick Message
00:11 gbeshers joined #darcs
01:42 mizu_no_oto joined #darcs
01:58 glguy_ joined #darcs
02:32 schlaftier joined #darcs
03:23 glguy joined #darcs
04:08 glguy joined #darcs
04:09 mizu_no_oto joined #darcs
05:00 preflex_ joined #darcs
05:06 mizu_no_oto joined #darcs
06:00 kowey joined #darcs
06:12 carter_ joined #darcs
08:13 raichoo joined #darcs
08:32 kmels__ joined #darcs
09:39 iago joined #darcs
10:59 xymox joined #darcs
11:19 owst joined #darcs
11:33 xymox joined #darcs
11:37 xymox joined #darcs
11:44 xymox joined #darcs
11:52 donri joined #darcs
11:56 xymox joined #darcs
13:04 mizu_no_oto joined #darcs
13:05 kowey joined #darcs
13:57 owst joined #darcs
13:59 mizu_no_oto joined #darcs
14:22 iago joined #darcs
15:03 mizu_no_oto joined #darcs
16:04 owst Can we somehow ban someone from the bts?
16:05 owst http://bugs.darcs.net/issue2298
16:34 kmels joined #darcs
17:16 raichoo joined #darcs
17:46 mizu_no_oto joined #darcs
18:20 nomeata joined #darcs
19:10 stepcut joined #darcs
19:49 amgarchIn9 joined #darcs
20:28 owst joined #darcs
21:15 Heffalump 'lo
21:15 owst Evenin'
21:15 owst Gah, how do I do unrecord in git?
21:15 owst I should know this
21:16 owst Some sort of reset I suppose
21:16 Heffalump I use gitk
21:16 Heffalump and yes, some sort of reset. --soft I think
21:16 owst Cool
21:16 Heffalump I thought you were the git expert around here? :-)
21:16 owst Me too!
21:16 owst I've forgotten it, and got used to how nice darcs is :-)
21:17 * Heffalump is just releasing 2.8.4
21:17 donri what's new?
21:18 * owst is trying to finish the paper/code last 3 night's bedtimes have been 2am,5am,3am
21:18 Heffalump I'll send the email shortly (not a lot)
21:18 Heffalump well, GHC 7.6 support
21:18 Heffalump need to work on 2.10, that'll be the exciting release
21:19 byorgey what's slated for 2.10?
21:19 Heffalump rebase, patch index
21:19 owst Rebase?
21:19 byorgey ooooh, nifty
21:19 Heffalump last regrets
21:19 Heffalump maybe packs if we can figure out what went wrong last time
21:19 owst I'm so used to last regrets (and their annoyances!)
21:19 owst packs would be sweet
21:20 byorgey last regrets?
21:20 owst byorgey: darcs will say "do you really want to FOO those N patches?"
21:20 byorgey ah
21:20 owst Quite handy, but the UI is sometimes a bit annoying
21:21 byorgey as
21:21 byorgey as long as I can turn them off
21:21 byorgey =)
21:21 byorgey but perhaps only certain ones... yeah, the UI sounds tricky
21:22 owst Heffalump: the answer was `git reset HEAD~1` to reset the index/HEAD-pointer, but leave the working dir alone
21:22 owst --soft would have left the changes in the index
21:23 Heffalump clearly
21:23 Heffalump we need to make sure we don't actually make pending like the index.
21:24 Heffalump s/actually/accidentally/
21:24 byorgey --soft < no option < --hard  in terms of destructiveness.  very weird.
21:24 owst Yeah
21:24 byorgey they should add an option like --firm
21:24 owst byorgey: yeah, it's because --mixed is the default
21:25 byorgey oh, I didn't even know about --mixed
21:25 byorgey I mean I didn't know it had a name
21:25 owst I knew the name, I didn't know it was default :-) I have only really used --hard before
21:25 owst --hard = obliterate, essentially
21:25 byorgey right
21:26 owst Heffalump: in what sense? I don't recall, but there are some situations where you end up with hunks in pending?
21:26 owst s/there are/are there/
21:27 Heffalump I'm not actually sure :-)
21:27 Heffalump I guess that's the main mistake of git?
21:28 owst Last regrets annoyance: "darcs unrec --last 1" says "Shall I unrecord this patch?" then "Do you want to unrecord these patches?"
21:28 Heffalump I suspect it might have to happen if a replace depended on some hunk changes
21:28 owst I don't think the index is seen as a mistake
21:28 owst it's quite nice when merging: all your cleanly merged changes go in the index, and any conflicts go in working, to mark them as resolved, you add them to the index
21:29 Heffalump it feels like a UI mistake most of the time
21:29 owst Which admittedly is annoying in that you can't merge with a dirty working dir
21:30 owst Hmm, sometimes. I guess you can work in small chunks, adding to the to-be-commited index as you go
21:30 owst not sure what the point of that is though :-)
21:30 Heffalump I think darcs uses amend-record for some of that workflow
21:30 owst also: `git checkout -p` is  `darcs revert`
21:30 Heffalump anyway, my point is that it's quite tempting to do something similar.
21:30 owst Yeah
21:31 Heffalump but the UI problems/confusion it causes are something we should keep in mind
21:31 owst What would it do for us?
21:31 Heffalump it's tempting for a reason; as you say having an explicit intermediate state can be handy.
21:31 owst right
21:31 Heffalump I'm not sure - I can just see that cleaning up or doing any more stuff with pending might lead to that kind of thing.
21:32 donri git doesn't really do obliterate unless you gc (or wait some weeks for git to gc)
21:32 owst yeah, sure
21:32 owst but darcs could also do that
21:32 owst :-)
21:32 donri agreed, i wish it did
21:32 owst well, it does in a sense - the cache probably saves you
21:32 owst (not sure when patches/inventories get put in the cache)
21:32 donri i thought obliterate deleted caches?
21:33 * owst isn't sure
21:33 Heffalump I'm never quite clear on how gc works even inside a darcs repo
21:33 Heffalump I suspect they just grow out of control. I certainly found loads of junk in mine when I was hacking on the internals.
21:33 owst Yeah, I think you're probably right
21:34 owst wasn't that the problem with packs? You always accumulate them, and can't delete them, since someone might be getting the repo
21:34 Heffalump isn't that also true for non-packs?
21:34 owst Yeah, I think so
21:41 gh_ joined #darcs
22:07 schlaftier joined #darcs
22:55 Lemmih joined #darcs
23:05 Topic for #darcs is now can't talk? see -> | http://darcs.net/IRC | http://darcs.net/ | darcs 2.8.4 is out http://darcs.net/Releases/2.8
23:08 mizu_no_oto joined #darcs
23:11 owst What is the point of the "Can't talk" bit?
23:13 Heffalump I can't remember :-)
23:13 Heffalump some technical issue with channel modes
23:13 * Heffalump thinks removing it might be sensible
23:14 owst I think so too
23:15 Topic for #darcs is now http://darcs.net/ | darcs 2.8.4 is out http://darcs.net/Releases/2.8
23:15 owst Nice
23:15 Heffalump do you have channel op thingummies?
23:15 owst Umm, I don't think so, how do I know?
23:16 owst I'm an IRC n00b.
23:17 Heffalump try /msg chanserv op #darcs
23:17 Heffalump (I just added you)
23:17 owst Cool
23:23 donri linking the irc logs would be useful though
23:24 * owst attempts this
23:25 Topic for #darcs is now http://darcs.net/ | darcs 2.8.4 is out http://darcs.net/Releases/2.8
23:26 owst Damn
23:26 Topic for #darcs is now http://darcs.net/ | logs: http://irclog.perlgeek.de/darcs/ | darcs 2.8.4 is out http://darcs.net/Releases/2.8
23:28 donri good job!
23:41 mizu_no_oto joined #darcs
23:53 mizu_no_oto joined #darcs

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