Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2014-03-13

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

All times shown according to UTC.

Time Nick Message
00:01 edwardk joined #darcs
00:31 siel_ joined #darcs
00:38 haasn joined #darcs
01:51 edwardk joined #darcs
03:28 mizu_no_oto joined #darcs
03:54 mizu_no_oto joined #darcs
04:06 amgarching joined #darcs
05:21 edwardk joined #darcs
06:13 balor joined #darcs
06:44 lelit joined #darcs
07:08 kowey joined #darcs
07:39 bishboria joined #darcs
07:47 alexei joined #darcs
08:15 raichoo joined #darcs
10:01 alexei_ joined #darcs
10:19 alexei___ joined #darcs
10:19 balor joined #darcs
12:17 mizu_no_oto joined #darcs
12:34 balor joined #darcs
13:34 balor joined #darcs
13:39 mizu_no_oto joined #darcs
14:03 alexei___ joined #darcs
14:13 colDrMcBeardman joined #darcs
14:42 mizu_no_oto joined #darcs
15:26 edwardk joined #darcs
15:27 edwardk joined #darcs
15:37 mizu_no_oto joined #darcs
16:54 raichoo joined #darcs
17:18 donri joined #darcs
17:35 briarrose joined #darcs
17:58 mizu_no_oto joined #darcs
18:00 favonia joined #darcs
18:22 edwardk joined #darcs
18:22 soyuzmjaso joined #darcs
18:37 colDrMcBeardman joined #darcs
18:44 alexei___ joined #darcs
18:54 favonia joined #darcs
19:17 favonia joined #darcs
19:21 mizu_no_oto joined #darcs
19:31 gh_ joined #darcs
20:20 favonia joined #darcs
20:32 mizu_no_oto joined #darcs
20:56 alexei___ joined #darcs
21:17 byorgey_ joined #darcs
21:17 byorgey_ joined #darcs
21:23 vikraman Heffalump: pong
21:26 vikraman Heffalump: I started looking at the Darcsden Issue and IssueChange types
21:30 vikraman Heffalump: when are you usually around on irc?
21:33 sm vikraman: Heffalump is in GMT time, and tends to be on at commute time
21:34 vikraman ok thanks, I already sent him an email
21:35 sm are you planning a darcsden gsoc project ?
21:35 vikraman distributed issue tracking using patch theory, that is the one I am interested in :)
21:35 sm oh right
21:36 vikraman I have been going through the darcsden source, it's awesome
21:37 sm have you seen a lot of haskell projects' code ?
21:37 vikraman are there any interesting issues that I could work on that you recommend?
21:38 vikraman well, I try to, I am actually a member of the gentoo haskell herd
21:38 sm nice. I agree, darcsden was a very impressive project
21:39 vikraman was?
21:40 sm well still is, I mean the initial project which was all the work of alexsuraci
21:40 sm if you mean darcsden/darcs issues to work on, there is no shortage
21:48 vikraman http://hub.darcs.net/simon/darcsden/issue/50
21:48 sm oops, network trouble
21:48 vikraman is this still relevant?
21:48 vikraman couchdb -> postgresql ?
21:49 sm yes but it's not a burning priority and the best course is not clear
21:49 sm Heffalump is abstracting the db storage
21:52 sm darcs' excessive memory use for certain operations is the biggest problem for darcs hub, though not necessarily for local darcsden use
21:56 * Heffalump appears
21:59 Heffalump I'm generally around in the evenings GMT - some parts of the time between 1800 and 1930, and again off and on after 2000. Also at some points between 0530-0800 GMT.
22:00 Heffalump the latest round of patches I sent to sm make using local storage for the darcsden data very feasible
22:00 sm yay
22:00 Heffalump http://hub.darcs.net/ganesh/darcsden-abstract-backend
22:01 Heffalump I've started building a locally hacked up "Ganesh" backend on top of them to prove I can :-)
22:04 sm it seems to me you can't disallow conflicts when pulling
22:04 Heffalump --skip-conflicts
22:04 Heffalump except there's a little hole in that that can occasionally leave you with conflicts anyway, but the reasons are a bit subtle
22:05 Heffalump actually --dont-allow-conflicts exists for darcs pull too
22:05 sm pull doesn't seem to have the conflicts flags. And IIRC setting it on apply in defaults didn't work
22:05 Heffalump ok, so setting it in apply won't help for pull
22:05 * sm is looking at 2.9.8 (+ 21 patches)
22:05 Heffalump apply is for pushes from remote repos
22:06 sm ok
22:06 Heffalump I see --dont-allow-conflicts in the help for pull in 2.9.8+34
22:07 sm odd, I shall build HEAD
22:07 sm by which I mean -screened
22:07 Heffalump I can't think of anything that could have changed that any time recently
22:07 Heffalump are you sure that e.g. darcs help pull | grep conflicts doesn't show you that flag?
22:08 sm woah, it does
22:08 sm urgh, I typed darcs push --help. Mystery solved. :)
22:10 sm and, I had put no-conflicts in defaults, should be no-allow-conflicts
22:13 vikraman I'm back, lost my connection there for a while
22:14 sm in -screened, it seems attoparsec needs to be added to the cabal file for  Convert.hs
22:15 Heffalump I see it there, >=0.11
22:15 Heffalump vikraman: is there anything you need to know to get started on a proposal?
22:15 Heffalump or indeed anything else worth discussing now
22:17 vikraman ok, what points do you want me to write about in the proposal?
22:18 Heffalump I don't have any strong feelings about exactly how this project should be done
22:18 vikraman objectives, deliverables, timeline?
22:18 Heffalump ah, I see - yes.
22:19 vikraman do you think it should be a cli tool like darcs that works independently of any vcs, or should it be tied to darcsden?
22:20 Heffalump I would lean towards darcsden for a GSoC project, as that's more likely to lead to something practically usable
22:20 sm aha.. I have a perhaps obsolete local patch "implement log --fast-export to output git-fast-import stream" with a conflictor for that hunk in darcs.cabal
22:20 Heffalump i.e. it should probably be hooked up to the existing darcsden issue tracking frontend code
22:21 Heffalump I think behind that frontend you would have a lot of flexibility in terms of the representation of the issues and how you used the darcs library code to manage that
22:22 vikraman that sounds good
22:22 vikraman also, I had earlier mentioned conflictors, how do we avoid conflicts here?
22:23 Heffalump if you look at the darcsden branch I mentioned above I think that provides enough of an API for a new backend for the issue tracking that you can plug into (and will no doubt need to tweak further, but it's heading in the right direction)
22:23 Heffalump good question - I lack an immediate answer to that.
22:23 vikraman ie, we need to talk about commutation rules for various patches
22:24 Heffalump the decision before that is I think about how to model issues. One option is just to make them be files in a normal filesystem that the traditional darcs patch types manage.
22:24 Heffalump Another is to define your own patch types for issues.
22:25 vikraman I have noticed that DarcsDen has a IssueChange type
22:26 Heffalump yeah, that does seem like an interesting thing for defining commute rules on (potentially)
22:26 Heffalump but it may also be appropriate to change it completely.
22:28 vikraman http://www.bugzilla.org/docs/4.0/en/html/lifecycle.html
22:28 vikraman bugzilla has a graph of the lifecycle of a bug
22:29 vikraman each edge in this is potentially a patch
22:29 vikraman then there are comments
22:29 vikraman also changes in any field etc.
22:29 Heffalump yeah - perhaps status is just a field
22:30 Heffalump (i.e. the edges in the graph are just patches to the status field)
22:31 vikraman yes, but perhaps this also has repercussions on merging
22:33 Heffalump agreed
22:34 Heffalump I'm not sure how important avoiding conflicts is - it may e simpler to just accept that they will happen and that people can resolve them if so, even though darcs doesn't handle them brilliantly at present
22:36 vikraman is the exponential blowup for nested conflictors still a problem?
22:37 Heffalump it's quite rare, but it is a possibility
22:37 vikraman I don't have a good understanding of the darcs conflictor
22:37 Heffalump sadly I don't think anyone does
22:37 vikraman oh wow
22:38 Heffalump well, I know roughly how they work. And Igloo probably knew about them very well at one point, though I don't know if he still does.
22:41 Heffalump you can kind of thing of them as a black box that manages conflicts for you, albeit badly.
22:41 Heffalump for the most part a simple two-way conflict works fine though
22:46 * Heffalump will be going to bed soonish
22:54 sm Heffalump: in HubPages.hs, import DarcsDen.Settings.Production hiding (frontPageContent) does not bring baseUrl and homeDir into scope, and I'm not sure how to fix it
22:55 sm I added them to the export list in Production and got an unbound implicit parameter error, but perhaps I'm on the right track
22:56 sm aha, I see the type signature I need to mimic
22:56 Heffalump I'll take a look, just building with -fhu should do it right?
22:56 Heffalump oh, ok
22:57 sm so does any function using something from settings need forall bp . (?settings :: Settings) => prepended to its signature ?
22:58 Heffalump the forall bp would be optional
22:58 Heffalump the ?settings :: Settings is the key part
22:58 Heffalump forall bp isn't exactly optional, it's just required for other reasons in some cases - largely orthogonal to the Settings thing
22:59 sm ok. And how do you select which settings you want ? I see some modules import Settings, others Settings.Production
22:59 Heffalump import either or both of those as needed
22:59 Heffalump just don't import Settings.Test
22:59 Heffalump and don't let the tests import Settings.Production
23:00 Heffalump I think possibly only the exes need Settings.Production
23:01 Heffalump I probably screwed up HubPages because I only did anything to it when a darcs replace hit it or something
23:01 Heffalump ah, I just didn't touch it at all
23:04 sm if you get any time to look at it, that would get me unblocked quicker. I'm currently getting src/DarcsDen/Pages/HubPages.hs:29:5: Couldn't match type `DarcsDen.State.Session.Session -> HSP XML' with `XMLGenT (HSPT' IO) XML'
23:09 Heffalump looking now
23:09 sm ah! I see it
23:09 sm never fails
23:10 Heffalump :-)
23:10 sm thank you :)
23:10 sm well, I moved forward by an inch. I'm really rusty on this stuff
23:11 Heffalump are you nearly there? I think there are a few changes that need making e.g. for the lenses
23:11 sm I think I'll have to leave it for today
23:11 sm the day job calls
23:12 sm so yeah your help would be most welcome
23:13 Heffalump I've done HubPages
23:13 Heffalump just hit a more annoying one where the getAllRepositiories call in Handlers is conditional on HUB
23:14 Heffalump so I'd stopped exporting it
23:14 Heffalump and in fact deleted it I think
23:20 adnap joined #darcs
23:27 Heffalump sm: I pushed a patch that should fix it
23:27 * Heffalump heads to bed
23:27 * Heffalump makes mental note for the future aout -fhub
23:28 Heffalump s/aout/about/
23:28 sm Heffalump: thanks, good night. Will try after I sort out cabal libs
23:28 gh_ joined #darcs
23:38 sm darcsden sure builds a lot of stuff
23:39 sm 156 or so dependencies
23:39 amgarchIn9 joined #darcs
23:45 IbnFirnas__ joined #darcs
23:45 IbnFirnas__ joined #darcs
23:54 mizu_no_oto joined #darcs

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