Camelia, the Perl 6 bug

IRC log for #darcs, 2011-03-24

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

All times shown according to UTC.

Time Nick Message
00:15 copumpkin joined #darcs
00:26 iago left #darcs
00:42 gwern so I tried out darcs optimize --http today
00:42 gwern is it supposed to increase _darcs size so much?
00:47 gwern oh yes, I also gave up keeping the output of `find ~/` in darcs when I fed darcs 30GB of disk space for a 200MB repo and it still couldn't finish the record/conversion
00:47 gwern (that problem should probably be covered in the docs or fixed)
01:17 twb joined #darcs
01:54 lispy left #darcs
02:15 gwern no one?
02:24 lispy joined #darcs
02:35 Gwern-away joined #darcs
02:35 gwern left #darcs
02:35 Gwern-away is now known as gwern
03:15 secorp joined #darcs
03:20 gwern left #darcs
03:26 secorp left #darcs
03:33 intripoon joined #darcs
03:36 intripoon_ left #darcs
03:53 gwern joined #darcs
03:53 gwern left #darcs
03:53 gwern joined #darcs
04:23 owst left #darcs
04:32 kolmodin left #darcs
04:32 kolmodin joined #darcs
04:32 kolmodin left #darcs
04:32 kolmodin joined #darcs
04:49 secorp joined #darcs
04:59 sm left #darcs
06:07 Riastradh left #darcs
06:10 Riastradh joined #darcs
07:01 darcscommitbot left #darcs
07:02 darcswikibot left #darcs
07:02 darcscommitbot joined #darcs
07:32 jutaro joined #darcs
07:44 jutaro left #darcs
08:00 jeltsch joined #darcs
08:04 jeltsch left #darcs
08:04 jeltsch joined #darcs
08:05 lelit joined #darcs
08:08 jeltsch left #darcs
08:16 jutaro joined #darcs
08:16 jutaro left #darcs
08:22 jeltsch joined #darcs
08:25 jeltsch1 joined #darcs
08:25 jeltsch left #darcs
08:28 jeltsch1 left #darcs
08:30 jeltsch joined #darcs
08:42 jeltsch left #darcs
08:59 shenshei joined #darcs
09:03 jonkri joined #darcs
09:05 raichoo joined #darcs
09:06 jeltsch joined #darcs
09:39 Weltraumschaf joined #darcs
10:03 darcswikibot joined #darcs
10:03 twb left #darcs
10:12 jutaro joined #darcs
10:45 shenshei left #darcs
11:10 lelit` joined #darcs
11:10 preflex left #darcs
11:14 preflex joined #darcs
11:17 gh_ gwern, about optimize --http, did your repo double in size? or more?
11:18 jonkri left #darcs
12:05 owst joined #darcs
12:08 shenshei joined #darcs
12:12 shenshei left #darcs
12:32 shenshei joined #darcs
12:35 lelit` left #darcs
13:02 gwern gh_: roughly double, yeah, from around 70M to ~120M iirc
13:19 lispy left #darcs
13:19 lispy joined #darcs
13:33 Riastradh left #darcs
13:38 gh_ gwern, we shall take advantage of the sprint to see whether there is a problem with optimize --http then, or at least to know if there are cases where it's not efficient
13:45 Riastradh joined #darcs
13:50 raichoo left #darcs
13:57 copumpkin left #darcs
13:58 copumpkin joined #darcs
14:12 copumpkin left #darcs
14:36 copumpkin joined #darcs
14:46 kowey joined #darcs
14:51 sm joined #darcs
15:36 gbeshers joined #darcs
16:05 shenshei left #darcs
16:25 kowey http://twitter.com/#!/ddicki​son/status/50952334875955200
16:32 lispy kowey: interesting.  For me it was github
16:32 lispy I just can't ignore github
16:32 kowey force to be reckoned with
16:33 kowey I do hope for a "github? gerrit? cool git GUI? no problem!" future
16:35 kowey interoperability is the name of the game
16:35 raichoo joined #darcs
16:36 owst (kowey, did you see the talk here a while back with myself, Heffalump and others, coming to the conclusion that a git<->darcs interop would make a nice GSoC project?)
16:37 kowey no, but hooray!
16:37 kowey I hope there were concrete things and realisable ideas being thrown around
16:37 owst Heffalump made the observation that it would be a good candidate project, with attainable milestones.
16:38 owst We can thrash out the details at the sprint, but I'm certainly keen to work in that area.
16:39 owst Being a reborn darcs user, having gone to git and back ;)
16:39 kowey :-)
16:40 kowey we do need more people who know how to think in git
16:40 kowey and darcs simultaneously
16:40 owst I'm training my brain back to darcs, but I still do think in git most of the time. Learning the underlying system of git was fairly useful, it makes me realise that it's probably not the best way to do things!
16:41 kowey hopefully have an appreciation for all the enthusiasm people have for it
16:41 owst for example, want to pull with dirty working dir in git, nope, can't do that.
16:41 kowey I really hate that
16:41 owst Yes, of course.
16:42 kowey I wonder if it's  "just" a UI issue
16:42 kowey or if it's something fundamental/deep about git philosophy
16:42 kowey or if it's just that people have gotten so used to pushing/popping things on/off stash that they don't even realise it could be otherwise
16:43 secorp left #darcs
16:43 owst I can't think why it should be difficult, why does it not do an implicit stash/apply?
16:43 owst maybe there are fundamental reasons, and I can't think of them right now.
16:44 kowey well, I suppose there's the problem of working directory conflicts
16:45 kowey compounded with actual conflicts
16:45 owst Yeah, perhaps they take the view that it is worse to (potentially) break your current work
16:45 kowey how do you make a git UI for that
16:46 kowey the twitterverse on darcs (when it tweets about darcs) seems to be a mixture of hostility for paper-cut problems, struggling with conflicts, and affection mixed with regret
16:46 kowey http://twitter.com/#!/ivey​/status/50737215113932800
16:47 kowey (and slow get, which is perceived as just darcs is slow)... not saying that we should care, just randomly reporting
16:47 owst "had won" it still can win!
16:48 kowey one thing that used to come up a lot is complaints from mac users about darcs taking forever to install (via MacPorts)
16:48 kowey I suspect this is occurring less frequently because fewer people are trying to install darcs ;-)
16:48 kowey it's not the complaints you need to worry about; it's the silence
16:49 Riastradh I hate having to do git stash and git pop and so on.
16:49 owst I suppose it's easier to treat a wound you know is there, so to speak.
16:49 owst I've had a feeling that darcs annotate is slower than it should be, but nothing more constructive than that.
16:49 kowey yep, but also at the same time to be careful to identify the sorts of problems that people just swallow silently
16:50 owst kowey: did you see the talk about darcs diff giving broken patch files?
16:50 kowey mornfall fixed that, but we didn't have our act together to merge it all in
16:50 owst why does it not default to a unified diff?
16:50 kowey I saw a glimpse of that yesterday
16:50 kowey I doubt there's a good reason
16:50 Riastradh I'm using Git mainly these days because (a) all the cool kids are using Git, (b) Git runs everywhere I care about and Darcs doesn't (e.g., NetBSD/macppc), and (c) after about three years of Gitting, I have finally begun to come to terms with understanding what Git is doing for me, and I'm a little more comfortable with having consistent names for complete states of the repository.
16:51 kowey and I'll note that in darcs-land, we're actually quite happy to change defaults
16:51 owst Open question: does Darcs not consider a set of patches to have a "name" that could be used to identify a given state?
16:51 kowey even if this breaks past-consistency for users because (A) not that many people use darcs anyway -- avoid success and (B) better to suffer now than limp along with crappy UI
16:51 owst Well, I'd push for a change to unified-by-default.
16:52 owst normal diff only makes sense if you touch one file.
16:52 owst kowey: should I make a bug for the "change-to-unified-default" issue, or talk to the mailing list?
16:53 kowey owst: first the bug, then lead the discussion
16:53 owst Since the current behaviour is broken, but it seems to me that people may have an opinion on the changing default resolution.
16:53 kowey the bug is just there for tracking
16:53 owst You may have to "hold my hand" if I do something wrong :p
16:53 kowey :-)
16:54 mornfall $  cat ~/.darcs/defaults
16:54 mornfall diff diff-opts -up
16:54 kowey Riastradh has kinda covered quite a few of our struggles in one swoop, (a) network effects, (b) well sort of (a) too, (c) people really like that sense of solidity
16:56 kowey Designing for the Digital Age says "don't make people configure something you can choose a sensible default for", which sounds right
16:56 JaffaCake left #darcs
16:56 owst Also, is there a historic reason for the annotate output being verbose to the extreme? It multiplies the output (compared to a straight `cat`) by 5 times in my instance!
16:56 gwern "Every configuration option in a program is a place where the program is too stupid to figure out for itself what the user really wants, and should be considered a failure of both the program and the programmer who implemented it."  http://web.archive.org/web/20080603011809/ht​tp://www.fishshell.org/user_doc/design.html
16:56 owst kowey: I agree.
16:56 Riastradh I still hate Git's UI, but after three years I have managed to isolate a small set of commands that I understand well enough to get things done with Git.  (With Darcs, it was closer to three minutes, back seven years ago when I started using Darcs.  But now that's moot for me.)
16:56 mornfall kowey: I am just pointing out that I agree with making -u default, there.
16:57 kowey oh! yeah :-)
16:57 JaffaCake joined #darcs
16:57 kowey I still enjoy the speculation that part of the enthusiasm for Git is the sense of accomplishment people have when they finally master it
16:58 Riastradh Ha.  Yeah, that could be.
16:58 owst kowey: I'd say that's somewhat true. It's nice to be the go-to expert ;)
16:58 Riastradh My reaction was not enthusiasm but `holy cthulhu what a bloody waste of time that was'...
16:58 jeltsch left #darcs
16:59 mornfall Riastradh: How agreeable. :)
17:01 Riastradh Also, I still haven't figured out how to do *distributed* development with Git...
17:03 kowey I still find the github pull-request model a bit tricky
17:03 kowey at least, in the sense of "what happens when your patch is rejected or modified by the author?"
17:04 kowey I guess you just move the main branch pointer back to a previous commit, if such a thing is possible
17:05 * kowey would love to work with a design team to learn how to apply Goal Directed Design to version control
17:11 owst First bug! http://bugs.darcs.net/issue2052 What a momentous occasion. :-p
17:12 * owst can't find any persuasive reasoning for why git can't pull with a dirty working dir.
17:14 kowey could be worth venturing into #git
17:14 owst Already lurking there. I'll ask.
17:21 Weltraumschaf left #darcs
17:21 shenshei joined #darcs
17:26 shenshei left #darcs
17:28 shenshei joined #darcs
17:28 owst Nothing concrete.
17:28 owst A few "doing stuff behind my back" comments, but nothing fundamental.
17:29 jutaro left #darcs
17:30 jutaro joined #darcs
17:32 shenshei left #darcs
17:33 shenshei joined #darcs
17:37 pumpkin joined #darcs
17:39 copumpkin left #darcs
17:51 darcscommitbot left #darcs
17:51 darcswikibot left #darcs
17:54 owst left #darcs
17:55 owst joined #darcs
17:57 gwern left #darcs
17:58 owst left #darcs
18:00 jutaro left #darcs
18:35 lispy owst: it's because git pull == git fetch; git merge.  Git can fetch with a dirty working dir, but it doesn't let you merge with a dirty working dir.
18:36 Riastradh left #darcs
18:36 Riastradh joined #darcs
18:37 lispy Merging in darcs with a dirty working directory is something I always avoid as a user
18:37 ezyang Merging in darcs is something I always avoid as a user ;-)
18:37 lispy It's just really, really unpleasant when the merge has conflicts
18:37 lispy ezyang: I bet you avoid conflicts no merges :)
18:37 lispy Everything in darcs is a merge (just about)
18:38 ezyang Oh, yes, that's right :-)
18:38 Riastradh Git refuses to merge even if there wouldn't be conflicts.
18:39 lispy Yeah, but they made that not-so-bad by adding the stash command.
18:41 lispy In darcs you do the same thing as stash by recording a patch and then unrecording it
19:15 copumpkin joined #darcs
19:19 pumpkin left #darcs
19:38 raichoo left #darcs
20:04 secorp joined #darcs
20:18 owst joined #darcs
20:33 Igloo left #darcs
20:39 Igloo joined #darcs
21:20 pumpkin joined #darcs
21:21 exlevan joined #darcs
21:24 owst left #darcs
21:24 copumpkin left #darcs
21:50 gh_ left #darcs
22:30 iago joined #darcs
23:19 pumpkin left #darcs
23:22 iago left #darcs
23:58 copumpkin joined #darcs

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