Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2014-04-04

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

All times shown according to UTC.

Time Nick Message
00:14 edwardk joined #darcs
02:46 strokov joined #darcs
04:09 ibespal joined #darcs
05:25 edwardk joined #darcs
06:14 lelit joined #darcs
07:06 raichoo joined #darcs
08:02 raichoo joined #darcs
08:28 ordnungswidrig joined #darcs
08:49 gal_bolle joined #darcs
09:02 notdan is it me, or the branches feature is slow/doeesnt work on darcshub?
09:02 notdan http://hub.darcs.net/co-dan/darcsden/patches <- this doesn't load for me
09:08 gal_bolle hub.darcs.net/galbolle/darcs-interactive-diff/patches does load for me
09:12 gal_bolle but your darcs den doesn't
10:12 edwardk joined #darcs
11:25 amgarchIn9 joined #darcs
12:36 donri joined #darcs
13:06 colDrMcBeardman joined #darcs
13:27 sm notdan: it can be slow and exceed the maximum allowed processing time when there are many/large/conflictful branches
13:29 sm I think fixing it requires optimising darcs, but it would be great to have someone investigate
13:29 sm bbl
13:54 gh_ joined #darcs
14:02 amgarchIn9 joined #darcs
14:31 mizu_no_oto joined #darcs
14:59 raichoo joined #darcs
15:03 edwardk joined #darcs
15:15 schlaftier joined #darcs
15:38 dolio joined #darcs
15:50 mdiaz joined #darcs
16:19 dolio joined #darcs
16:36 edwardk joined #darcs
16:54 gh_ hi
17:11 mizu_no_oto joined #darcs
17:14 rdesfo joined #darcs
17:21 mizu_no_oto joined #darcs
17:27 c74d joined #darcs
17:37 amgarchIn9 joined #darcs
17:45 rdesfo left #darcs
18:00 sm ho, what's this secret darcs amend-unrecord command I just heard about
18:14 edwardk joined #darcs
18:18 sm I'd love a doc for darcs (& others) that lists the commands with one line description of the situation in which you'd use them
18:19 sm I guess that's slightly more verbose than darcs help
18:50 mornfall sm: it's an alias, that's why it's secret ;-)
18:50 mornfall (but I didn't know about it either)
18:58 gh_ "amend-record" could be just called "amend"
18:58 mornfall that's what I call it
18:59 mornfall which reminds me, darcs chan no longer works (or any other contraction of changes)
19:00 gh_ indeed
19:00 mornfall (darcs rec does work though)
19:01 mornfall (and amend, and many others... that's a weird regression)
19:01 mornfall (and darcs c resolves to darcs convert :)
19:01 mornfall oh
19:01 mornfall changes is now an alias to log
19:02 gh_ it's because the new name is log
19:02 mornfall and aliases don't participate in shortenings (amend-un doesn't work either)
19:02 mornfall that's pretty silly though
19:03 mornfall my muscle memory is already broken by the last regrets thing
19:05 mornfall and well, having different command names from git is actually a feature for me :) I confuse them less often that way
19:11 xstill hm, another thing which I will have to get used to. I found just today that darcs rec --loo stopped working
19:12 mornfall :\
19:12 mornfall I use -l for that I think...
19:12 mornfall (but normally I don't, because I have a lot of junk everywhere...)
19:12 xstill nice, I did not know about that
19:13 xstill I usually preffer long commands because they are easier to remember
19:13 xstill and with shortening I have nice compromise between remembering and length
19:13 xstill every program should have command shortening :-)
19:18 siel joined #darcs
19:23 mizu_no_oto joined #darcs
19:25 mornfall hmm, and there's no prompt for short comment anymore
19:25 mornfall gah
19:29 * mornfall goes back to darcs 2.8
19:30 mornfall my life is miserable enough with random incompatible UI changes
19:30 mornfall without, too
19:31 mornfall (I need to fix up darcs-fastconvert anyway, as I kinda need incremental conversion)
19:39 xstill I actually like the way commit messages work now. But incremental conversion would be usefull :-)
19:48 nebnesknarf joined #darcs
19:49 mornfall xstill: well, the problem is that I have been typing messages inline for half a dozen years... hard to unlearn
19:52 mornfall for typing one line, the overhead of a real editor just doesn't pay (especially if it's stateful or runs in a different window or both)
19:52 xstill I see, I'm using it just for year and half and git (and svn) handle commit messages in same way as new darcs, so it is ok for me.
19:53 xstill I can see your point I'm using vim in console so it is not a problem
19:54 mornfall you still need to hit an extra i, don't you? (which is fine if it's mechanised, but if you automatically start typing, bad things happen (tm))
19:56 xstill that is prety automatic, I would actually need to do almost all text editing it vim to improve my efficiency (like writing mails I mean).
19:56 mornfall I write emails in emacs, yes. :)
19:57 nebnesknarf my 2 cents: I very much like the one-line prompt and I have configured it to be the default for me
19:57 xstill (vim does not come with integrated everything)
19:58 mornfall nebnesknarf: oh, that's configurable? :)
19:59 mornfall I didn't find a switch for the last regrets prompt, so I didn't try looking for stuff to revert other UI changes...
20:00 nebnesknarf are we talking about 2.9.8 (unreleased)?
20:00 nebnesknarf I mean the current screened repo
20:00 mornfall whatever is in screened
20:01 nebnesknarf ok, the option is --prompt-long-comment and you can set it just fine in ~/.darcs/defaults
20:03 nebnesknarf I find some of the late changes in darcs UI questionable, but I can live with them if I can configure them away
20:04 nebnesknarf what cannot be configured is that 'darcs changes' is now a hidden command; doh
20:04 nebnesknarf it means command completion does not work for it
20:07 nebnesknarf wrt last regrets: I found out just yesterday that it gives you more than just [yn], you can also say [l] to see the selection you made. that's pretty nice
20:08 mornfall nebnesknarf: can the extra prompt be disabled too? (I didn't find it)
20:08 nebnesknarf no i don't think so
20:08 edwardk joined #darcs
20:09 mornfall I'll stay with the familiar then. :)
20:15 nebnesknarf mornfall: does that mean you'll never upgrade to 2.10 or later? ;-)
20:16 mornfall nebnesknarf: possibly
20:17 mornfall nebnesknarf: if there's a compelling improvement, I'll consider re-training, or patching it up
20:17 nebnesknarf you'll miss out on a large number of very useful features (rebase among them)
20:18 mornfall nebnesknarf: I no longer have a project that'd benefit from rebase ...
20:18 mornfall I used it a long while ago (before it was part of the upstream repo)
20:20 nebnesknarf I'd hope a patch implementing a --no-last-regrets option has a chance of being accepted
20:21 nebnesknarf since we are using darcs at work, our team collaborates on a number of repos, so I have lots of use for rebase
20:22 nebnesknarf grmph, this came out bad
20:22 mornfall :-)
20:23 mornfall well, rebase is mainly useful as a workaround for problems with conflicts
20:23 nebnesknarf hope it is clear what I wanted to say
20:24 mornfall well, that it's useful to you, I guess :)
20:24 mornfall which is legit, I used it for conflicty repos too
20:24 mornfall we tried a long-running branch a while back, and had to rebase at the end, because it wouldn't merge otherwise
20:25 nebnesknarf yes, darcs still has performance problems when a merge encounters lots of conflicts
20:25 mornfall correctness problems, too
20:26 nebnesknarf personally, i try to avoid conflicts, I like a clean (or rather: not too dirty) history
20:26 nebnesknarf yes
20:26 mornfall that's what we do now, too :)
20:27 mornfall branches are short and merged often, and usually concurrent work is on different parts of the project
20:28 nebnesknarf yes, exactly, if your project allows this kind of workflow darcs is just great
20:29 nebnesknarf another reason to avoid long-running branches is that I find it very hard to manually resolve the conflicts
20:30 nebnesknarf I mean, even if darcs has no technical problems
20:39 mornfall hmm
20:39 mornfall it's somewhat annoying in git, but it's not very hard (obviously depends on how nasty the conflicts are)
20:41 nebnesknarf mornfall: my thesis is that resolving conflicts does not scale nicely: it gets disproportionally harder the more conflicts you have
20:42 nebnesknarf obviously I have no proof for that ;-)
20:44 gh_ joined #darcs
20:44 nebnesknarf wrt recent UI changes: a configuration file that says (overrides) which commands are hidden and which not would be cool
20:44 mornfall nebnesknarf: seems a little on the complex side :-)
20:45 nebnesknarf hmm
20:46 nebnesknarf or maybe a slight extension to the defaults format? like: [unhide] command option
20:47 edwardk joined #darcs
20:47 mizu_no_oto joined #darcs
21:03 edwardk joined #darcs
21:10 edwardk joined #darcs
22:05 edwardk joined #darcs
22:11 edwardk_ joined #darcs
22:15 edwardk joined #darcs
22:16 edwardk joined #darcs
22:42 mizu_no_oto joined #darcs
22:55 dolio joined #darcs
23:18 schlaftier joined #darcs
23:36 mizu_no_oto joined #darcs

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