Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2015-07-29

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

All times shown according to UTC.

Time Nick Message
01:47 ilbot3 joined #darcs
01:47 Topic for #darcs is now http://darcs.net/ | logs: http://irclog.perlgeek.de/darcs/ | darcs 2.10.0 is out http://darcs.net/Releases/2.10
02:44 mizu_no_oto joined #darcs
03:21 Riastradh joined #darcs
05:38 Heffalump notdan: shall we meet soon, either tonight or Friday maybe?
07:05 alexei joined #darcs
09:19 notdan Heffalump: yes, let's have a meeting today?
09:43 alexei_ joined #darcs
10:08 lambdabot joined #darcs
10:59 mizu_no_oto joined #darcs
11:02 sea-gull joined #darcs
13:12 Heffalump notdan: great, 8pm london time as usual?
13:20 byorgey joined #darcs
14:00 Riastradh joined #darcs
14:05 burp joined #darcs
14:15 lambdabot joined #darcs
14:39 notdan yep
15:17 Riastradh joined #darcs
16:19 alexei_ joined #darcs
16:46 alexei_ joined #darcs
17:37 mizu_no__ joined #darcs
18:45 MasseR joined #darcs
18:50 Heffalump I'm here from now
18:53 notdan Heffalump: hi
18:55 Heffalump hi :-)
18:56 Heffalump I've been using the local backend in anger a bit to review patches
18:57 notdan why in anger?
18:57 Heffalump that's an English idiom meaning "for something real", it doesn't imply any actual anger :-)
18:58 notdan Oooh
18:58 notdan TIL!
18:59 notdan Cool
18:59 notdan Are you using the latest version with local jqueryui and full patch hashes?
18:59 Heffalump no. I should pull.
19:00 Heffalump actually, what should I pull, the latest local-take2?
19:00 notdan http://hub.darcs.net/co-dan/darcsden-local-take2
19:00 notdan yeah
19:00 Heffalump ok, doing that now
19:00 Heffalump How's merging going?
19:01 notdan I am reformatting the history
19:01 notdan doing so I realized that it would be cool to be able to list suspended patches in darcsden
19:01 notdan at least for me
19:01 Heffalump it would indeed - rebase is another thing that would really benefit from a UI
19:02 notdan and extended contexts for hunk patches, too
19:03 notdan this is supported in darcs proper, right?
19:03 dolio That's why "Anger" is a great name for a programming language.
19:03 Heffalump you mean that darcs already has code that adds context to hunk patches? Yes, it does.
19:04 notdan Dammit, my ~/.darcs folder is gone. Today I almost rm'd my home directory, accidentally
19:06 notdan Anyway, so the patches that I worked on for a couple of days was that JSON unrecorded changes thinh that I mentioned to you already. The UI should also pick up the dependencies: when you select a patch, it also selects also its immediate dependencies.
19:06 Heffalump recursively, presumably?
19:07 notdan I also rewrote the javascript code. It was hella ugly: callback towers, etc. Now it is more functional and uses promises for synchronization and code readability purposes: https://blog.jcoglan.com/2013/03/30/callbacks-are-imperative-promises-are-functional-nodes-biggest-missed-opportunity/
19:08 notdan Heffalump: dependencies? Yeah, but I still haven't figured whether I should do this on client side or server side
19:08 notdan "this" = taking the transitive closure
19:08 Heffalump which javascript code did you rewrite, the original code in darcsden or your own first draft?
19:09 notdan oh only the code that I wrote for record
19:09 notdan the rest of the js code in darcsden is fairly straightfoward
19:10 Heffalump I like the selection UI
19:12 Heffalump record needs to pick up my details (name/email address) from the darcs defaults
19:13 notdan very true
19:14 Heffalump unselecting hunks doesn't unselect things that depend on them
19:17 Heffalump do you have a todo list somewhere, btw? Just checking you haven't forgotten the other stuff I mentioned as well - wrong slashes on Windows, config file location needs fixing
19:19 notdan yeah
19:20 notdan let me paste it actually
19:23 notdan https://privatepaste.com/d475c343fd
19:24 Heffalump great
19:24 Heffalump so, about the merge - hoow much is left?
19:25 * Heffalump looks at http://hub.darcs.net/simon/darcsden/compare/co-dan/darcsden-local-take2
19:26 notdan Hm I think Simon hasn't pushed the changes that he merged already
19:27 notdan but it's not much unfortunately :(
19:28 sm notdan: I unmerged those 3/4 changes pending your cleanup.. nothing has been officially merged to trunk
19:28 sm (recently)
19:28 alexei_ joined #darcs
19:28 sm isn't rebasing your test patches to be first pretty easy ?
19:28 notdan Yeah that's reasonable, because I need to rewrite history
19:29 sm yup
19:29 notdan sm: yeah the problem is that some of them depened on non-test patches
19:29 notdan Let me push what I've got so far
19:29 Heffalump notdan: I think you should focus on getting the bulk of the merge out of the way before you do more stuff on record - as you said yourself, it'll make future work much simpler
19:32 notdan Heffalump: you are right :< I am just being.. less responsible than I should
19:32 notdan sm: so I manage to push the test patches down the history http://hub.darcs.net/co-dan/darcsden-local-take2/changes/page/2
19:32 sm there's always the nuclear option. Unrecord all, and record a completely new set of patches
19:33 notdan sm: but there are still a bit of non-test related patches inbetween
19:34 sm I know some sequence of cleanish patches is possible, whether it's tests or something else isn't critical
19:34 Heffalump is 'Add basic "download .zip" functionality' mergeable now?
19:34 notdan Yeah, it is completely separate from the rest of the changes
19:35 sm ok, merging that
19:35 notdan I also made this <http://hub.darcs.net/co-dan/darcsden-local-take2/patch/20150729191925-21ae5> Use full patch hash as patch ID. I don't know if you want to merge it or not (it breaks old urls) but I use it on local instance, because otherwise different patches get same URL
19:36 Heffalump presumably them getting the same URL is quite common after a rebase if you unsuspend lots of things at once?
19:36 sm download .zip patch is merged to trunk and can be removed from discussion :)
19:36 sm let me know when to pull more
19:36 notdan Heffalump: exactly
19:36 Heffalump right, so what makes sense to merge next? You were talking about tests?
19:37 Heffalump btw re rebase and dates, you can use --keep-date when unsuspending if you want to keep the original record date. Can lead to some weird "temporal anomalies" though
19:39 mizu_no__ joined #darcs
19:51 notdan sorry, I am trying to amend one particular test patch, but I screwed up my sandbox
19:51 sm "stack build" :)
19:51 * sm been there
20:20 * Heffalump wonders if notdan will ever emerge :-)
20:22 sm cabal hell! we left a good man in there 8-( !
20:24 sm and I'm serious, "stack build" is a fast elevator out of there
20:31 notdan sorry guys
20:31 notdan Ok I did some more rebasing
20:31 notdan http://hub.darcs.net/co-dan/darcsden-local-take2/changes/page/2
20:32 notdan sm: you could also merge "fix warnings in Dist.hs" which is again not realted to the -local code at all
20:33 notdan so could you also review the patches up to "Implement HTTP Basic auth"?
20:33 mizu_no__ joined #darcs
20:34 sm notdan: great taking a quick look
20:36 sm notdan: the hsp dependency still needs to be dealt with
20:37 sm you're still requiring 0.10.0.1 right
20:37 notdan there should be a patch that relaxes it
20:37 notdan well it is still technically required for tests to pass
20:37 notdan unfortunately, I haven't figured a workaround yet :(
20:37 sm you'll have to disable those tests I guess
20:40 notdan hm
20:40 sm how do things work after "read parent from _darcs/defaultrepo", eg with existing repo relationships on darcs hub ?
20:41 notdan can we both disable the tests and keep the code? or is that a bad idea?
20:41 notdan sm: oh, that patch only touches the local backedn
20:41 notdan *backend
20:41 notdan CouchDB stills stores this stuff explicitly
20:52 Heffalump I should head off soon. Anything else for us to discuss?
20:59 notdan Hm, no
20:59 notdan I don't think so
20:59 notdan I was also writing a blog post on directed types in darcs
21:00 notdan but it all got changed in screened
21:00 sm back
21:00 Heffalump you can still write a post about those :-)
21:00 notdan well not all, but some of my explanations dont work
21:02 Heffalump I guess you could also stick to the version in 2.10 and just add a note saying things have changed a bit in the latest code
21:02 mizu_no__ joined #darcs
21:03 notdan yeah
21:03 sm I have to go too.. notdan, I've merged up as far as the HTTP auth tests
21:03 sm will look at the rest later
21:03 notdan ok, thanks sm!
21:03 Heffalump yay!
21:03 sm thank you!
21:03 notdan yeah i have to fix the rest too
21:03 notdan meet again tomororw?
21:03 Heffalump me, or sm?
21:03 Heffalump I'm out tomorrow evening though will probably be online at some point in the evening
21:03 sm moving day tomorrow, I may not be allowed much screen time, but we'll see :)
21:04 sm later o/
21:05 notdan ooh important day
21:05 Heffalump notdan: does the local backend do something special to only bind to 127.0.0.1 ?
21:06 Heffalump hmm, no, it's bound to 0.0.0.0
21:06 Heffalump (was just trying to use it remotely for my own purposes and not succeeding)
21:08 notdan maybe you can use the hostname setting?
21:08 Heffalump I think it's windows firewalling screwing with me
21:13 notdan Tbh after those changes in screened I came to think that FL = directed lists associated to the right and RL = directed lists associated to the left
21:13 Heffalump that's basically what the changes amount to
21:14 Heffalump from the darcs point of view it's more logical to always think of patch sequences as going in one direction, and just to change the associativity to get the most efficient representation
21:15 notdan yeah
21:15 notdan two datatypes representing the same concept
22:52 lf94 joined #darcs
23:03 fr33domlover hello! I got this when trying to install darcsden from the 1.1 branch: http://paste.rel4tion.org/40
23:03 fr33domlover I realize there are dependency issues there - I just don't know how to resolve them :P
23:04 fr33domlover (I tried the master darcsden repo too, got some conflicts as well)
23:05 notdan fr33domlover: hm, can you try to manually edit darcsden.cabal and bumb the pandoc dependency a bit
23:05 notdan ?
23:06 fr33domlover notdan, trying :)
23:11 sm or use stack
23:16 fr33domlover sm, would stack solve the problem?
23:17 fr33domlover btw I bumped pandoc to essentially have no upper limit (as of now) and blaze-html upper bound is 0.8 instead of 0.7
23:17 fr33domlover that got resolution to succeed
23:17 fr33domlover (but it doesn't mean the code will build etc.)
23:21 sm fr33domlover: yes, it will likely solve the problem directly
23:22 sm slightly less reliable, but should also work right now: use a cabal sandbox
23:24 sm using just cabal without a project sandbox is only for experts with some time to fool around IMHO
23:26 fr33domlover sm, I'm using a sandbox indeed
23:26 fr33domlover still building :P
23:54 fr33domlover wow, still building :/
23:54 sm heh
23:54 sm I always install --dry to see what I'm in for
23:54 fr33domlover it's my first time installing something so big
23:54 fr33domlover Haskell isn't environment friendly, it seems :/
23:55 fr33domlover (unless you have solar power maybe)
23:56 sm it's true, there is a lot of building - not yet much binary packaging
23:57 sm with stack, most of your deps will be shared among projects which helps a lot
23:57 sm among projects which use the same snapshot, anyway
23:59 fr33domlover does stack work with the entire hackage?
23:59 fr33domlover i.e. even stuff not in stackage
23:59 sm yes, if you tell it to

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