Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2014-03-19

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

All times shown according to UTC.

Time Nick Message
00:49 edwardk joined #darcs
02:14 colDrMcBeardman joined #darcs
02:14 edwardk joined #darcs
02:34 colDrMcBeardman joined #darcs
02:57 colDrMcBeardman joined #darcs
03:11 mizu_no_oto joined #darcs
03:59 IbnFirnas_ joined #darcs
04:06 gh_ joined #darcs
04:12 mizu_no_oto joined #darcs
04:56 alexei_ joined #darcs
06:37 lelit joined #darcs
08:18 raichoo joined #darcs
09:57 exlevan joined #darcs
10:35 adnap joined #darcs
10:40 adnap_ joined #darcs
11:14 exlevan joined #darcs
11:42 mizu_no_oto joined #darcs
12:08 mizu_no_oto joined #darcs
12:31 edwardk joined #darcs
12:33 whaletechno joined #darcs
13:29 alexei_ joined #darcs
14:55 colDrMcBeardman joined #darcs
15:34 favonia joined #darcs
15:48 alexei_ joined #darcs
15:52 mizu_no_oto joined #darcs
16:07 alexei_ joined #darcs
16:12 mizu_no_oto joined #darcs
16:57 mizu_no_oto joined #darcs
18:07 edwardk joined #darcs
18:12 whaletechno joined #darcs
18:15 lelit joined #darcs
18:27 mizu_no_oto joined #darcs
18:36 dolio joined #darcs
18:41 gh_ hi
18:42 raichoo joined #darcs
18:42 gh_ Heffalump, http://irclog.perlgeek.de/darcs/2014-03-18#i_8455714  1) yes and 2) I haven't seen a lot of project proposals for haskell.org up to now, maybe we can ask for 3 slots
18:43 favonia joined #darcs
18:54 * gh_ hopes we'll see RFC's for GSoC projects on darcs-users no later than tonight
18:57 mornfall :-)
18:57 Heffalump how many are you hoping for?
18:59 gh_ there are 2 students interested here, I'm helping them with the proposals (on the first 1 projects, global cache and history reordering)
18:59 gh_ s/1/2
19:07 mornfall there's been a patch floating around implementing bucketed cache for many years btw
19:08 alexei_ joined #darcs
19:08 mornfall centering the project around "darcs undo" might be useful (although not particularly related to the global cache)
19:11 Heffalump did it manage backwards compatibility?
19:11 Heffalump actually changing it incompatibly seems trivial
19:15 gh_ yes, there's a matter of detecting and migrating unbucketed cache to bucketed
19:17 Heffalump I guess even that is simple if you don't care about having older clients running too
19:27 gh_ yes
20:13 adnap joined #darcs
20:14 gh_ mornfall, I feel that "darcs undelete" would be a more interesting "main point" of the project, since it has a lot of potential and uses the global cache
20:45 Heffalump I'm more convinced by darcs undo than darcs undelete as a useful feature
20:48 sm does darcs need more features ?
20:49 Heffalump undo would be cool because it would allow you to remove some confirmation prompts
20:49 sm I think it's more in need of polish and industrial-strengthening
20:50 sm not that they need be mutually exclusive, just sayin
20:50 n-dolio I've seen requests for some new features.
20:50 sm what about pull --rebase ? I still want that
20:50 n-dolio Like patch signing.
20:53 n-dolio I could imagine some new features that would be useful to have, too, probably.
20:53 n-dolio Like, working out some way of using nested repositories to work similar to branches.
20:53 n-dolio That's probably way too ambitious for a summer project, though.
20:55 n-dolio What is pull --rebase, by the way? Rebase your working set changes onto the patches you pull?
20:55 sm yes, just like git pull --rebase. An easy ui to the existing rebase feature
20:57 sm probably more than any of this, darcs needs a release, and a regular release cadence
20:57 n-dolio Oh. I've never used darcs rebase.
20:57 sm n-dolio: because it's in head
20:57 n-dolio No, I have head.
20:57 n-dolio I just never rebase.
20:57 sm ah, my mistake
20:57 n-dolio I don't use git rebase, either.
20:57 sm not something I do often either, but I would pull --rebase often
20:57 gh_ Heffalump, yes, "darcs undo" would be used more often
20:58 Heffalump sm: I told you it exists, right (rebase pull)?
20:59 Heffalump sm: agreed (re polish etc) but you can't really make that a gsoc project too easily
20:59 sm Heffalump: maybe you did ? Still, pull --rebase would be the easy and familiar ui right
20:59 Heffalump I guess so, you think it should be an alias?
20:59 sm yes alas, I wish we could get gsoc'ers to do the release !
21:00 sm yes I think that would be great.. does it just work (or just fail with a nice explanation), or is there more required ?
21:00 n-dolio 'Make darcs help text format according to terminal size.' :)
21:00 n-dolio That's probably not an entire summer.
21:02 Heffalump sm: do you have a definition of "just work"?
21:02 Heffalump (preferably not involving me finding out in detail how git pull --rebase works :-)
21:02 Heffalump s/finding out/finding out for myself/ # I don't mind if you tell me
21:04 sm ok :) sure, when I used rebase, there was new terminology and a number of manual steps required. A pull --rebase that just works has no extra steps - it does the rebase, or tells you why it can't, and you're done
21:05 sm my experience of git pull --rebase: when git pull tells me it can't merge because of my local unpushed commits, I do git pull --rebase and it rewrites the local commits, and the pull succeeds (or it can't merge automatically and tells me which files)
21:06 Heffalump ok, so you still have to unsuspend the patches with darcs rebase pull
21:06 Heffalump but, iI think it differs from git rebase in that there will invitably be conflicts if you are rebasing (otherwise why not just merge)
21:06 Heffalump so it couldn't just work
21:06 sm hmm
21:07 sm well how does it work at all ? I have forgotten. Does darcs rebase always require manual conflict resolution then ?
21:08 Heffalump well, you can use it even in caes where it doesn't, but then why bother?
21:08 sm ok
21:08 Heffalump oh, I take that back - deep amend-record is a case where you would want to
21:15 sm Heffalump: aside from what you can do with the existing rebase feature, isn't there a common case which we don't yet support with a single command, namely where you have recorded local patches which conflict with upstream changes, and you want to re-record them on top of latest upstream so you can push them as conflict-free patches ?
21:18 Heffalump I agree we don't support that with a single command, but I don't know how to do that sensibly given that you have to actually resolve one or more conflicts in the middle of the process
21:19 mizu_no_oto joined #darcs
21:20 sm I don't either, I wonder how git does it. I guess it doesn't, it's just in effect doing what we rely on darcs pull to do automatically (though probably also its automatic merge strategies are smarter than ours)
21:23 n-dolio They might also be dumber.
21:23 n-dolio So that they work when a smart strategy would figure out that it shouldn't. :)
22:01 vikraman Heffalump: emailed you
22:02 mizu_no_oto joined #darcs
22:21 raichoo joined #darcs
23:08 alexei_ joined #darcs

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