Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2015-06-12

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

All times shown according to UTC.

Time Nick Message
00:04 codygman joined #darcs
00:04 codygman Does the name darcs stand for anything?
00:15 sm I think originally david's advanced revision control system
01:14 mizu_no_oto joined #darcs
02:26 mizu_no_oto joined #darcs
02:35 lf94 joined #darcs
03:58 thorkilnaur_ joined #darcs
05:37 alexei___ joined #darcs
05:37 peb` joined #darcs
07:05 dolio joined #darcs
07:13 haasn joined #darcs
08:52 maerwald joined #darcs
09:26 notdan sm: aaah, I was always under the impression that a stands for "awesome" :P
09:26 notdan gh_: uh that sucks :(
09:27 notdan maybe we can resurrect it
09:31 alexei___ joined #darcs
11:07 mizu_no_oto joined #darcs
11:53 alexei___ joined #darcs
13:50 alexei___ joined #darcs
14:58 gh_ joined #darcs
15:03 f-a joined #darcs
15:19 luoke joined #darcs
16:57 Riastradh joined #darcs
17:17 gh_ Heffalump, so far the following patches go into 2.10.1: bump attoparsec dependency, support QuickCheck 2.8, Generalize Darcs.UI.Commands.Dist.doFastZip, Fix the previous doFastZip patch, fix warning, support zlib 0.6
17:17 maerwald Heffalump: what do you think of integrating issue tracking into darcs? I mean darcs, not darcsden. We have everything terminal-compatible except issue tracking. Or do I miss something and that's already possible somehow?
17:17 maerwald similar to what fossil does: https://www.fossil-scm.org
17:21 gh_ maerwald, what do you mean by "We have everything terminal-compatible" ?
17:21 f-a time to spam my barebones issue tracker http://hackage.haskell.org/package/lentil
17:22 f-a maerwald: have you used fossil in the wild?
17:22 f-a I am quite curious
17:22 maerwald gh_: I cannot do issue tracking without a browser unless I write a plugin that interacts with darcsden, no?
17:22 maerwald gh_: I think it would be interesting to have issue tracking integrated in the vcs...
17:22 gh_ maerwald, ok I see
17:22 f-a maerwald: how would people report issues?
17:22 maerwald so darcsden would just use that API as well
17:22 f-a i.e. non developers
17:23 gh_ maerwald, someone proposed a summer of code project about this last year but it didn't get accepted.
17:23 maerwald f-a: via darcsden ofc
17:23 maerwald but darcsden needs to be hacked as well for that
17:23 maerwald non-trivially
17:23 f-a I see
17:24 f-a so, if I follow you correctly, you are envisioning a way for darcs to, mh, be plugged into darcsden (or whatever else) more easily, regarding issue tracking?
17:24 maerwald f-a: not really, no... I want bug tracking to be independant of darcsden in darcs
17:24 maerwald darcsden would just connect to the API then and visualize stuff etc
17:25 maerwald but that's really a major change... I'm just brainstorming
17:25 maerwald it would allow for way better tooling if you think long term
17:26 f-a ok, maerwald I will spam my stuff once again http://www.ariis.it/static/articles/decentralised-lentil/page.html xD
17:26 maerwald e.g. one could write a gtk gui for darcs which shows the history as well as the issues
17:26 f-a tl;dr, "issue tracking" is really many different activities
17:30 dolio Going in kind of the opposite direction is interesting, too, I think.
17:30 dolio Launchpad, for instance.
17:31 maerwald worst approach ever :D
17:31 dolio Issue tracking isn't associated with repositories/projects at all.
17:31 f-a projects like xmonad use a mailing list
17:31 f-a other project use github
17:31 dolio So when a bug gets reported against one project, but it's actually due to another, the bug can correctly be characterized as being against the second project.
17:32 maerwald but with darcsden, you have issue trackers per repository as well
17:32 dolio Yes. That's bad. :)
17:32 maerwald if you want something more complex, you probably end up using redmine or whatever anyway
17:33 dolio It's probably not really that bad, because hub will never be big enough to be like launchpad.
17:33 dolio So the amount of cross-project tracking that could ever go on is limited.
17:36 maerwald dolio: another approach is just having a cli-tool to interact with darcsden. I know there is something for github that does that
17:37 dolio On a tangent, does darcsden e-mail about issue creation and whatnot?
17:37 dolio I kind of feel like it doesn't.
17:37 maerwald dolio: not really no
17:37 maerwald TODO :P
17:37 dolio Okay.
17:52 gh_ Heffalump, pushed "support network 2.5 and 2.6" to 2.10
17:52 sm f-a: wow, that's quite good timing (lentil)
17:53 f-a :) thanks sm
17:54 f-a I have gotten some encouraging responses, I hope someone will find it useful
17:56 sm gives a nice summary of code todos in hledger
17:56 f-a :)
17:57 f-a if I may sm, http://ariis.it/static/articles/lentil-manual/page.html check the man out! if you here and there add some tags, it would make stuff even more browable and useful
17:57 sm I am, thanks
17:57 f-a ( http://ariis.it/link/repos/lentil/issues.txt ) look at my output
17:57 sm f-a you have shipped useful software that Just Works - my very favourite kind. nice!
17:57 * f-a blushes
17:58 f-a thanks for the encouraging feedback. Please write to me if you need anything or have suggestions or anything.
17:58 sm it also gives a nice report on the darcs source, but there are a few parse errors
17:58 sm on ./Hlint/impossible.h
17:58 f-a argh
17:58 * f-a downloads
18:05 f-a thanks for the bug report sm , I will check it now and report back when fixed
18:07 f-a (fun fact, it's an empty file xD)
18:09 sm aha
18:17 sm one valuable thing that conventional issue trackers provide is issue numbers
18:18 sm also an easy way to link to the discussion/history of the issue
18:22 f-a those two will provide some interesting design challenges. Thanks sharing, noted!
18:24 sm what if lentil could assign numbers, where missing, by adding tags
18:24 f-a I could even make a handy syntax
18:24 sm and optionally display associated notes from a numbered file in a known location
18:25 f-a -- TODO something something [tag] [poo] #number
18:25 f-a that's a great idea!
18:29 sm perhaps generate those files too. Maybe even update them if the comment in the source changes
18:29 f-a just to see if I understood, something like this? http://ariis.it/link/t/sm.txt
18:30 sm and an issue file might have one, many, or no related source comments
18:31 sm I was imagining something like ./.issues/101
18:31 sm or it could be in _darcs, but really why would you make it require darcs
18:33 Heffalump maerwald: what I'd like is for darcsden issues to be expressed as darcs patches
18:36 f-a sm: I guess in a file like ./.issues/101 I would find a markdownish description of the issue + other useful info (url to discussion, etc.), right?
18:37 sm yes, or plain text, or anything they want
18:39 * sm tries not to go all armchair-design-crazy on you
18:39 f-a keep doing it!
18:39 Heffalump I saw the lentil announcement, it looked nice, was thinking of using it on darcs
18:39 sm those files could be web-served when they want something linkable. Perhaps with source code links auto-added
18:40 f-a I wrote lentil for my own needs and now that I decided to release it I am grateful if someone shares their own needs
18:41 f-a Heffalump: that's nice to hear, in general I think there is a great field, yet to be occupied, on, what to call it, decentralised or distributed issue tracking
18:42 sm lentil has a very nice low barrier to adoption - most projects already have lentil issues, and it needs no configuration
18:42 f-a one folk in the thread (Henning Thielemann) said that he was planning to use gitit/darcsit
18:42 sm but it seems like it could scale up quite smoothly, for programmer/code issues and perhaps beyond
18:43 f-a so whatever you choose Heffalump be sure to make a blog post, it's nice to see what other people are using in their development
18:44 f-a sm: I received encouraging feeback which was nice and surprising (for such a simple tool)
18:47 f-a I see hledger issue tracker quite busy! Nice to see
18:48 sm I suppose..
18:48 sm stupid bugs.. >:|
18:48 f-a project with no issues are dead project :P
18:49 f-a (ok, dead projects or TeX)
18:49 f-a I noted down what you guys said, it will be nice for a while to work on lentil not at the keyboard but with pen and paper, playing with ideas
18:50 * sm wonders what's going on in the darcs tracker and if it should announce here
18:53 Heffalump I haven't had time to digest all this conversation yet, but my thoughts about darcs patches were issue tracking were that there are two broad options:
18:53 Heffalump (1) = text files, i.e. restructure the concept of issue tracking to fit well with the filesystem
18:53 Heffalump (2) = custom patch types, i.e. do more work but perhaps get something that's got nicer merge/commute behaviour
18:54 Heffalump I think vikraman and I had some discussions about it on here sometime maybe February 2014ish
18:54 f-a Heffalump: on irc on mailing list?
18:54 sm what would be an example of how 2 is helpful ?
18:56 dolio I would think the most problematic aspect of either of those is that you want people to be able to create/comment on issues, publicly, even if you don't want them to be able to push to the code parts of your repository.
18:56 sm and, who knows the state of the art for 1 ? I think there must be a bunch of attempts at this already
18:57 sm that's true
18:58 sm well in design 1, your .issues/ can be a separate repo, maybe a gitit wiki
18:58 sm and.. hmm.
18:58 dolio With public push rights?
18:59 sm yes, with a different access policy
18:59 dolio I guess that might be where the commutation comes in.
18:59 sm but with a separate repo you lose the integration of code fix and issue update
18:59 dolio Comment patches shouldn't conflict.
18:59 f-a would a ML (or anything other mean of communication) do? you can then reference the thread on your issue tracker without giving access to the user
19:00 f-a (who wouldn't like to learn a dvcs to report a bug in the first place, maybe)
19:00 f-a a wiki would be apt too
19:00 f-a maybe even omre than a ML
19:02 Heffalump f-a: IRC
19:02 * f-a checks if the channel has logs
19:02 * sm remembers how gigantic this rabbit hole is
19:02 Heffalump it does
19:02 f-a \o/
19:02 f-a this will be the doom of us all xD
19:02 Heffalump (sorry for my disjointed contributions to the conversation, I'm very interested in the topic but am a bit busy right this minute)
19:02 dolio Yes, a mailing list would work, but then it's up to the developers to add issues to the tracker whenever there's a mailing list thread.
19:02 dolio Unlike most project hosting sites.
19:03 f-a dolio: it would be automatedish, but in my opinion manual add gives a sweet impulse to immediate triaging, which, at least with me, is required
19:03 f-a I agree automation is super convenient
19:05 f-a has anyone tried gitit or darcsit?
19:05 sm yes, it's great
19:05 sm darcs.net might be the biggest user of it
19:06 f-a Heffalump: http://irclog.perlgeek.de/darcs/2014-02-04 mmmmaybe here
19:06 sm huh, being maintained too
19:07 sm how does john mcfarlane get any professoring done
19:07 Heffalump f-a: don't think so, that's about imports in the code
19:08 f-a argh right
19:09 Heffalump http://irclog.perlgeek.de/darcs/2014-02-27 pointing to http://irclog.perlgeek.de/darcs/2013-08-15#i_7455609
19:11 Heffalump it might be worth starting a darcs.net wiki page about this topic if it doesn't already exist (it might well do, I could easily have forgotten)
19:13 f-a http://darcs.net/Ideas/BugTracker Heffalump I found this
19:16 sm http://docs.bugseverywhere.org/master/tutorial.html
19:20 alexei___ joined #darcs
20:01 f-a I've fixed that bug sm (0.1.1.2)
20:03 f-a will now enjoy eating. Thanks for your ideas folks, it has been really an interesting chat
20:03 f-a_ joined #darcs
20:04 notdan Hey everyone
20:05 notdan I forgot, how do I generate the manual?
20:05 notdan darcs help markdown, and then what?
20:07 f-a notdan: man manual,  you mean?
20:07 f-a darcs help manpage | man -l -
20:07 f-a from http://darcs.net/HintsAndTips#view-darcs-help-in-markdown-and-manpage-format
20:07 notdan html manual
20:07 notdan as in http://darcs.net/manual/bigpage.html
20:08 maerwald joined #darcs
20:08 dleverton joined #darcs
20:10 f-a notdan: pandoc?
20:12 sm good question..
20:13 notdan f-a: yeah, but I want to actually generate the full manual
20:13 notdan i.e. with headers and stuff
20:15 sm looks like gh nuked it from orbit: darcs log -s -h d20e6633853cfa5da354b643587922335c412d3f
20:16 sm it's a pity those hashes are not used in darcs hub urls
20:16 notdan is this the reason the manual on the web site is outdated?
20:17 sm I would guess so. Not very good eh
20:17 sm I don't remember what the new plan was
20:17 notdan I am not in a position to keep complaining, but there's been a darcs 2.10 release, yet there is no binaries and no documentation
20:20 sm it deserves complaint
20:22 notdan it's kinda hard to sell darcs to other people in this very moment
20:22 notdan well OK, the man page is actually generated by Setup.lhs
20:22 sm and discussion, hopefully stimulating improvements
20:23 notdan from `darcs help manpage`
20:26 alexei___ joined #darcs
21:47 notdan OK, I've updated my OSX binaries and Homebrew formula: http://darcs.covariant.me/
21:47 notdan `brew install http://darcs.covariant.me/darcs.rb` should just work(tm)
22:32 mizu_no_oto joined #darcs
22:59 mizu_no_oto joined #darcs

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