Camelia, the Perl 6 bug

IRC log for #darcs, 2013-07-19

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

All times shown according to UTC.

Time Nick Message
00:05 mizu_no_oto joined #darcs
00:12 edwardk joined #darcs
02:28 mizu_no_oto joined #darcs
02:29 bfrank joined #darcs
03:04 amgarchIn9 joined #darcs
03:13 alexei joined #darcs
03:35 mizu_no_oto joined #darcs
04:23 mizu_no_oto joined #darcs
04:43 dolio joined #darcs
06:52 kmels joined #darcs
07:13 favonia joined #darcs
07:14 lelit joined #darcs
07:24 raichoo joined #darcs
08:05 gh_ joined #darcs
08:09 alexei joined #darcs
09:33 owst joined #darcs
09:38 whaletechno joined #darcs
09:53 jlneder joined #darcs
10:13 owst gh_: is windows really identified by a mingw prefix on the os string?!
10:15 gh_ owst, I haven't tested it myself but according to this: http://stackoverflow.com/questions/93​41611/what-are-possible-values-for-th​e-system-info-os-function-in-haskell
10:16 gh_ and according to the existing windows constant introduced by eric in Darcs.Util.Ssh
10:16 gh_ that should be it.
10:16 owst Aha, so it's to do with what compiled the Haskell compiler, not a runtime thing
10:16 owst cool
10:16 owst I was just concerned about people running darcs without mingw!
10:17 owst (Maybe that would be a good comment somewhere!)
10:17 gh_ good point
10:36 owst gh_: the patch looks good. I'm at work so don't have perms to push the patch, but feel free to do so
10:37 owst s/perms/keys/
10:38 gh_ owst, ok thanks
10:39 owst oh and if you do, could you push my patch from yesterday too? I forgot to do it last night
11:01 burp_ joined #darcs
11:03 donri joined #darcs
11:53 uniquenick joined #darcs
11:56 favonia joined #darcs
12:27 mizu_no_oto joined #darcs
12:51 bfrank__ joined #darcs
12:52 bfrank__ hi, can anyone describe to me or redirect me to a place that describes the benefits of darcs?
12:53 konundra joined #darcs
12:54 owst bfrank__: benefits relative to...?
12:54 owst (SVN, Git, hg, ...)?
12:54 bfrank__ yeah
12:55 bfrank__ like, I am still trying to wrap my head around it
12:55 bfrank__ I mean, it seems like it is easy to pull subsets of patches
12:55 bfrank__ I guess, I am trying to understand from someone more experienced with it, how common workflows work with it
12:56 owst It is easy to pull subsets of patches, yes.
12:56 owst Do you have experience with other VCS tools?
12:56 bfrank__ how exactly is that done? Some of the examples I've seen seem to show that you can pull patches with patterns?
12:56 bfrank__ Yes, HG, SVN
12:56 owst bfrank__: yes, that's one way
12:56 owst Darcs is quite interactive, so if you just say `darcs pull` it'll ask you which patches to pull on an individual (ish) basis
12:57 bfrank__ I guess, I am also a bit perplexed from the context of how to return to previous states of code
12:57 owst Yeah, that is something that isn't quite so easy with darcs
12:57 bfrank__ for instance, that is a common workflow in hg/svn
12:57 bfrank__ it seems like you can "tag" stuff
12:57 owst bfrank__: can you give an overview of the workflow you're talking about?
12:57 owst yeah, tags identify a set of patches
12:57 owst so you can pull tags, and their associated patches
12:58 owst For me, most things I do with darcs are very similar/the same as those with e.g. git
12:58 bfrank__ but, I mean, for instance in mercurial, or svn, I could say checkout the code as it existed as of a specific date
12:58 owst Right
12:58 owst that isn't so easy with darcs
12:58 owst Although it depends what you mean... time on your local machine, or global time?
12:59 bfrank__ so I guess, as a result of that not being so easy, I guess, what I am trying to figure out, is what darcs would give me, that would be worth sacrificing that ability
12:59 owst Interactiveness; sane hunk recording; cherry-picking for free; nice UI
12:59 bfrank__ hmm
12:59 owst Just try it for a small project, you'll see what it's all about
12:59 owst And ask questions here :-)
12:59 bfrank__ when you say cherry-picking, you mean, both the ability to roll forward with patches, and roll them back as well?
13:00 bfrank__ for instance, could I get my code back to what it was a week ago, simply by rolling back the patches committed in that week, if I were to track exactly which ones I did in the week?
13:00 owst I mean picking patches out-of-order compared to how they were recorded. Whether that is push/pull/obliterate or other operations
13:00 bfrank__ hmm
13:00 owst yes
13:01 bfrank__ does darcs track the dates patches were applied or created?
13:01 owst Commonly, you might tag at the start of the week. Then you could delete all patches since the tag
13:01 owst created, yes
13:01 owst applied, no
13:01 bfrank__ hm
13:01 bfrank__ I noticed as I was recording stuff, that it asks me for a patch name
13:01 owst I use darcs for all my coding projects, and it really is the most convenient VCS tool
13:01 bfrank__ what exactly is that?
13:02 owst It's a commit message
13:02 owst used as (part of) the identity of the patch
13:02 bfrank__ I guess, the concept of this fascinates me a bit, I am trying to wrap my head around it
13:02 bfrank__ I also notice it asks me if I want to save a long message
13:02 owst Yeah, it's a bit different, which is why it's worth trying
13:02 bfrank__ what is the difference between long message and patch name?
13:02 owst indeed it does, that's for changelogs >1 line
13:03 owst patch message/name is the "identity" of the change, whereas long message is more descriptive
13:03 owst just like hg
13:03 MasseR bfrank__: It's explicit in that regard whereas git (and hg?) it is implicit. You can think of it as the header and content of the commit message
13:03 bfrank__ so patch name is like a sha1?
13:03 owst bfrank__: sort of
13:03 owst the author,date,message and some other bits go towards the patch id (or SHA1, if you like)
13:03 MasseR owst: Really? You can have multiple commits with the same message
13:04 bfrank__ but they have to have a different name, right?
13:04 owst No, you can make as many patches with the same name as you like
13:04 bfrank__ hm
13:04 owst darcs internally gives them unique names
13:05 bfrank__ has anyone in here written or knows of a tutorial that guides people through workflows you'd typically do with git/hg/svn/bzr, but shows how they are done in the darcs way?
13:05 owst bfrank__: I do suggest just diving in, and trying it. Darcs tries to be helpful. If you get stuck, come here
13:05 bfrank__ like, if I do multiple records, how do I diff those records?
13:05 owst `darcs changes -v` will show the contents of each patch
13:05 bfrank__ ah
13:06 bfrank__ so that is like hg log -p ?
13:06 owst If hg log -p does what git log -p does, yes
13:06 bfrank__ hehehe
13:06 owst (except darcs is interactive by default)
13:06 bfrank__ I did notice that it was asking me a lot of stuff
13:06 bfrank__ is there a way to say yes to all, no to all, etc?
13:06 owst You can quiet most it down if you need/want
13:06 bfrank__ also wasn't sure what the W meant
13:06 owst yes
13:07 owst Hit '?' it'll tell you the choices
13:07 MasseR bfrank__: IIRC w is wait. It lets you skip the change and decide later
13:07 bfrank__ I think it was a capital W though
13:07 MasseR Probably wait on all
13:07 bfrank__ hm
13:07 bfrank__ let me download darcs on this machine real quick
13:08 owst No, captials show you what the default is, if you just hit return
13:08 bfrank__ ahhh
13:08 bfrank__ that makes sense
13:08 owst And that's actually silly, because darcs says "invalid response" if you enter an uppercase letter
13:08 bfrank__ well, because as I could tell from help, there were no capital letter options
13:08 * owst adds a bug report for that - darcs should notice you've tried that, and tell you what the uppercase letter means
13:09 bfrank__ haha
13:09 owst indeed
13:09 bfrank__ yes, that would be helpful
13:09 bfrank__ I guess, coming from "traditional" vcs, darcs is kinda scary
13:09 owst Why is it scary? :-)
13:09 owst It's friendly!
13:09 bfrank__ ha
13:10 bfrank__ I guess, is there a way to have darcs auto tag after each commit?
13:10 owst Yeah, you can create a script that is hooked into post-record
13:10 bfrank__ if I could do that, it might not seem as scary
13:10 bfrank__ of course some of this could just be a fear of not knowing how to do things the darcs way
13:11 owst I wouldn't be too scared. Unless you're using obliterate, or amend-record, there aren't many ways to truly irrevocably change data
13:11 owst s/many/any/ ??
13:12 bfrank__ I guess, the workflows I use vcs for, are to return to previous states, and that is where with darcs it feels like it would be harder to do
13:12 owst So darcs doesn't have in-repo branching, which makes that a bit more cumbersome, but it's not so bad
13:12 bfrank__ So is the way that this is done in darcs, to pull the changes you want to a new repo?
13:12 owst Exactly
13:13 bfrank__ is it possible to pull everything and exclude what you don't want?
13:13 bfrank__ or only to pull what you do want
13:13 owst darcs get repo repo_copy --to-patch 'everything to the patch with this name'
13:14 bfrank__ what exactly does that mean? Is it going sequentially up to that patch?
13:15 bfrank__ is it possible to retroactively add patches to a tag?
13:15 bfrank__ like, if I start with a repo with no tags, and have 1000 records, could I go back and add different patches to different tags?
13:15 owst it goes sequentially in the order the patches exist in the repository
13:16 bfrank__ and/or remove patches from a tag
13:16 gh_ owst, I screened your patch from yesterday
13:16 owst gh_: thanks
13:16 gh_ I did it yesterday actually
13:16 owst bfrank__: no, I don't think it is, without pulling into a clean repo and tagging there
13:17 owst if you did that, and then push the tags to the original repo, you can tell darcs to optimise itself around those tags
13:17 owst gh_: there was another small follow-up patch I think
13:17 MasseR owst: gh_ Are you screening each others patches? :)
13:17 owst MasseR: "being screened" means to be pushed into the darcs.net repository
13:18 bfrank__ btw, that capital letter doesn't seem to mean the default, because I am pressing enter, and it won't let me proceed
13:18 owst I am at work, and don't have the necessary SSH key to do so, so gh_ is helping me out
13:18 owst bfrank__: oh
13:18 owst Sorry, space
13:18 owst not enter
13:18 owst hitting '?' told me that
13:18 bfrank__ ah
13:18 gh_ owst, about my bundle, did you mean you were ok to screen or accept it?
13:18 owst accept
13:18 gh_ k
13:18 owst But I haven't tested on OSX
13:23 gh_ I'm screening it and will ask on darcs-users for testers.
13:23 bfrank__ is there a way when pulling to get more context in the diff?
13:25 owst bfrank__: I don't think so
13:27 mizu_no_oto joined #darcs
13:35 owst gh_: good point re: just accepting captial default choice
13:36 owst I did think that, but thought it might be confusing?
13:38 sm good morning
13:38 owst hey sm
13:39 gh_ owst, I don't see how
13:39 * sm wonders what good works owst, gh are doing
13:39 owst No, thinking about it, me either, nevermind :-)
13:39 owst We're talking about the fact that if you enter the default choice as it appears in the choice list (i.e. uppercase) darcs barfs
13:40 owst We just agreed that darcs should just accept it
13:40 sm oh yea, saw that - good
13:40 sm it sounded like you're working on patches too. I guess I should look at the repos on hub
13:41 owst Oh just some small things
13:41 owst I made annotate give you line numbers which had been bugging me for a while
13:41 sm noticed that - thanks!
13:41 owst gh_ did some stuff with caches in correct places
13:41 sm and that
13:41 owst *with putting
13:42 mal`` joined #darcs
13:42 gbeshers joined #darcs
13:43 sm I see markstos has a bunch of unscreened docs patches at http://hub.darcs.net/darcs/darcs-screened/patches
13:43 gh_ hmm I accepted a couple of his a few weeks ago
13:43 owst Ack, we've missed the patch queue on hub
13:43 owst Oh
13:44 sm and galbolle has a bunch of interactive diff patches at http://hub.darcs.net/darcs/darcs-reviewed/patches . I think he probably should have that repo forked off -screened
13:47 sm yay, new darcsden-gsoc patches
13:48 owst sm: ooh, where?
13:48 sm http://hub.darcs.net/simon/darcsden/patches
13:48 sm bsrk's compare repo feature
13:50 sm no ui for it yet I guess
13:50 * owst wonders what the function iff does
13:51 owst Looks like a nice feature though!
13:52 * sm wonders where bsrk's feature proposals are
13:53 sm aha, http://www.google-melange.com/gsoc/proposa​l/review/google/gsoc2013/bsrkaditya/32002
13:54 gh_ also darcs.net/GSoC/
13:55 * owst would love "search in this repo" on hub
13:55 sm nice. I'd better link his page there
13:55 sm owst: yeah, or click function name to jump to definition
13:56 sm (iff is in Darcsden.Validate)
13:56 owst sm: That sounds sweet ;-)
13:56 owst That's my favourite part of using agda-mode (and forcing myself through the pain of losing vim!)
13:57 * sm wonders how to speed up hub (and darcs) development
13:58 owst More time/people?
13:58 owst I have been very non-darcs-y for quite a while now
13:59 owst I've recently felt like finally getting back into it though, so we'll see
14:00 gh_ more people yes
14:09 bfrank__ is there a way to create a list of patches in a text file, that can be used to pull only those patches into a new get?
14:10 owst How do you want to select the patches?
14:11 owst And anyway, I'm not sure get could be persuaded to select only from a list of patches
14:11 bfrank__ I guess, I am picking a huge swath, and instead of manually having to try to match a pattern, I just want to be able to pass a file with a list of them to a command, and have it pull only those patches
14:11 bfrank__ I probably still have a lot to learn about darcs then
14:17 sm @tell bsrk hi, I added links to http://darcs.net/GSoC/2013-Darcsden , consider linking this page in each blog post
14:17 lambdabot Consider it noted.
14:17 sm more people yes.. what's the most cost-effective way to attract more people ?
14:18 sm I think it's probably "focus on shiny ui"
14:20 gh_ make students use it? I'm a teacher :-)
14:21 sm ha. But doesn't that make them turn against it at the first opportunity :)
14:21 sm how about making them use github
14:21 sm hmm no they'll probably stay there >:(
14:23 sm bfrank__: usually we are selecting patches up to or starting from a certain matched patch. Do you really need to select a large number of nonconsecutive patches ?
14:25 sm with --[to|from]-[m|p]atch
14:26 bfrank__ hmm, I guess, the problem is that darcs doesn't guarantee to preserve the order
14:26 sm or we tag a known set of patches so they can easily be pulled later
14:27 bfrank__ how freeform can tagging be, though
14:27 bfrank__ for instance, if I don't initially tag, but then realize, crap, I should have started a tag, and want to ammend a whole slew of patches to that tag
14:27 sm a tag depends on whatever patches are currently in the repo.. so you can arrange that in a branch, make the tag, then push the tag back
14:27 sm a little clunky, but conceptually simple (darcs' strength)
14:28 mizu_no_oto joined #darcs
14:30 owst sm: interest them in either: a) nice UI, b) nice underlying (incomplete) model, c) real world haskell!
14:31 owst I'm worried that those things are already there, and people just aren't interested though :-(
14:34 sm I think there's enough demand to keep growing darcs/hub. At least until something devastatingly better appears. Of course it's a tough call to decide how to spend one's time on an individual level
14:34 owst indeed
14:34 owst Personally, I'm interested in finding a conflict representation that is (provably) correct/complete
14:35 sm projects that thrive make that decision easy for people by being so damn compelling
14:35 owst My motivation for the bridge has (seriously) waned.
14:35 owst But I will get it done
14:35 sm tough to do if you're not into it
14:38 * gh_ obliterates spam patches on darcs-wiki
14:53 bfrank__ get --to-match isn't working right for me
14:53 raichoo joined #darcs
14:56 sm git / darcs discussion: http://www.reddit.com/r/haskell/comment​s/1iki8p/how_to_help_develop_hackage2/
14:58 gh_ "in-place branching" I like that phrase
14:59 gh_ recently I found myself using obliterate -O to do it
14:59 gh_ but it's only convenient with a few patches to remove from the current  repo
15:09 konundra joined #darcs
15:17 sm yes I do that quite a bit
15:17 sm or sometimes darcs get --lazy . branch1
15:21 gh_ but then you move to another directory, so your .o and .hi files are no longer there
15:27 whaletechno joined #darcs
15:28 raichoo joined #darcs
15:35 drostie joined #darcs
15:46 lelit joined #darcs
15:55 uniquenick joined #darcs
15:56 kmels joined #darcs
15:58 uniquenick joined #darcs
16:02 sm so. I should do my part and get this darcsden release/hub news post out
16:03 sm but there must also be something visibly different on the front page
16:03 sm otherwise it's.. *YAWN*
16:04 sm hello Heffalump, any eta on your db changes ?
16:28 uniquenick joined #darcs
16:30 edwardk joined #darcs
17:09 mizu_no_oto joined #darcs
17:42 drostie joined #darcs
18:30 konundra joined #darcs
18:33 MasseR joined #darcs
18:37 bfrank__ what does darcs: fd:9: hPutBuf: resource vanish (Broken pipe) mean?
18:37 mtp it means the far end of whatever you were doing hung up
18:38 bfrank__ I am pushing to a local repo
18:44 alexei joined #darcs
18:47 gbeshers joined #darcs
18:49 bfrank__ is it normal to have that kind of problem when trying to push to another repo?
18:49 bfrank__ pulling doesn't seem to have that problem
18:50 bfrank__ ahh, I see, pushing in general fails
18:50 bfrank__ what do I need to do to make a push work?
18:52 mtp pushing where?
18:53 mtp oh, to a local repo?
18:53 mtp yeah, i've never seen that go wrong
18:53 bfrank__ I am pushing from one repo, to a blank repo
18:53 bfrank__ is that a cause for problem?
18:54 bfrank__ I am not sure how it would be, though, that I can pull into a blank repo, but can't push into a blank repo
18:54 mtp oh
18:55 mtp i think pushing into a blank repo is possibly not kosher
18:55 mtp try using 'put' instead
18:55 mtp or 'darcs get . /path/to/other/repo' from the blank one
18:55 mtp manpage suggests the former
18:55 bfrank__ can I make get interactive, like a push/pull would be?
18:56 mtp not that i know of; you might want to use lazy and then pull to deepen
18:56 mtp shrug
19:04 uniquenick joined #darcs
19:32 mizu_no_oto joined #darcs
19:39 alexei joined #darcs
19:39 edwardk joined #darcs
19:56 uniquenick joined #darcs
20:56 uniquenick joined #darcs
21:02 bfrank___ joined #darcs
21:02 bfrank___ is get --to-match broken?
21:02 bfrank___ I can't seem to get it to work
21:23 konundra joined #darcs
21:31 sm bfrank___: what are you doing ?
21:31 bfrank___ just trying to get up to a specific patch
21:32 bfrank___ and it gives me an error
21:32 bfrank___ it says
21:32 bfrank___ darcs failed: invalid --match pattern 'remove word repetition in fileHelpAuthor string'.
21:32 bfrank___ "match" (line 1, column 1):
21:32 bfrank___ unexpected 'r'
21:32 sm what's your command ?
21:32 bfrank___ expecting match rule or end of input
21:33 bfrank___ darcs get --to-match="remove word repetition" repo1 repo2
21:33 bfrank___ repetition in fileHelpAuthor string" repo1 repo2, I mean
21:33 sm you should use --to-patch for that. darcs help patterns has some examples of match syntax
21:34 bfrank___ oh
21:34 mizu_no_oto joined #darcs
21:35 sm also re blank repo.. you can push to an empty repo (in which you have done darcs init) but not to an ordinary directory - darcs put does that
21:36 bfrank___ I did do a darcs init
21:37 bfrank___ but it wouldn't let me push
21:37 bfrank___ I tried pushing just the first record, and that failed, tried pushing all records, that also failed
21:37 sm that's odd
21:38 bfrank___ yea
21:38 sm if something fails, I usually add --debug-verbose and try again
21:38 sm it certainly works here
21:39 bfrank___ it spews a lot of stuff, but nothing that specifically explains why it is failing
21:39 sm nothing relevant at the end ?
21:39 bfrank___ just the message it said before
21:40 bfrank___ darcs: fd:9 hPutBuf: resource vanished (Broken pipe)
21:42 bfrank___ ahhh
21:42 bfrank___ apparently push requires that darcs be in my path
21:42 dolio joined #darcs
21:42 bfrank___ that certainly was never explains anywhere
21:42 bfrank___ explained
21:43 bfrank___ I wonder why it would need to be in my path to push, but not to pull
21:43 sm urgh.. sometimes darcs' error messages are its achilles heel
21:44 sm push requires it to run a separate darcs apply process I guess
21:50 bfrank___ Can someone put that on the troubleshooting page?
21:51 sm bfrank___: please do! that would be a fine contribution :)
21:52 bfrank___ I am not sure how to
21:52 sm what's the page url ?
21:52 bfrank___ darcs.net/Troubleshooting
21:52 sm just click login / register at top right
21:53 sm it's a wiki
21:53 bfrank___ Does stuff get reviewed?
21:53 bfrank___ to confirm that what I right is valid?
21:53 bfrank___ write
21:53 sm I'll review it as soon as you're done :)
21:53 bfrank___ ha
21:53 bfrank___ well, have you been able to confirm that you get the same error when you remove darcs from your path?
21:53 sm seriously, newcomers are often the best documenters of stuff
21:54 sm I haven't dont that but I believe you
21:54 sm since you've tested it
21:54 sm and it sounds plausible. push and pull work a little differently
21:54 sm ok fine I'll try it
21:55 bfrank___ ha
21:56 sm easy enough: PATH= /usr/local/bin/darcs push newdir -> darcs: fd:9: hPutBuf: resource vanished (Broken pipe)
21:56 sm that'd be an excellent bug report, bugs.darcs.net
21:56 bfrank___ you confirmed it, you mean?
21:56 sm yes
21:56 bfrank___ so should it be a bug report, or a wiki troubleshooting entry?
21:57 sm a bug report first, wiki entry if you feel like it
21:59 bfrank___ where are bugs reported
21:59 sm bugs.darcs.net -> register -> new issue
22:02 bfrank___ I got a permission denied trying to update the wiki
22:02 sm bfrank___: so, you're logged in ok ?
22:03 bfrank___ yep
22:03 bfrank___ damnit
22:03 bfrank___ it lost my text too
22:04 bfrank___ oh, wait, I got it
22:04 bfrank___ yeah
22:04 bfrank___ it says
22:04 bfrank___ Server error: UnknownError: Could not darcs record Troubleshotting.page
22:04 bfrank___ Troubleshooting.page
22:05 bfrank___ darcs: _darcs/index: opening of '_darcs/index' failed: permission denied (Permission denied)
22:05 bfrank___ I can email you the text, it is really quite simple
22:05 sm I see the same thing
22:05 bfrank___ oh
22:05 sm the wiki is broken. Add to bugs.darcs.net :)
22:05 bfrank___ ha
22:05 bfrank___ wtf
22:06 bfrank___ grumble grumble
22:06 sm I don't think any admins are on right now. Thanks for the reports.
22:10 * Heffalump appears
22:10 bfrank___ ok, well I submitted it
22:14 Heffalump I've fixed the wiki editing, I hope.
22:15 sm yay Heffalump
22:17 sm I'm cleaning up http://hub.darcs.net/ front page and about to do a darcsden release
22:20 Heffalump cool
23:14 sm needs a logo
23:17 mizu_no_oto joined #darcs
23:31 gh_ joined #darcs

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