Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2016-09-01

| 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 Darcs/Pijul Sprint, Helsinki, 6th-8th May http://darcs.net/Sprints/2016-05 | http://darcs.net/ | logs: http://irclog.perlgeek.de/darcs/ | darcs 2.10.3 is out http://darcs.net/Releases/2.10
02:44 castlelore joined #darcs
02:52 mizu_no_oto joined #darcs
03:15 Riastradh joined #darcs
03:34 lispy joined #darcs
03:37 lispy_ joined #darcs
03:51 lispy joined #darcs
03:54 lisp_ joined #darcs
04:15 vikraman joined #darcs
04:18 lispy joined #darcs
04:25 lispy joined #darcs
04:28 lispy joined #darcs
04:40 aristid joined #darcs
05:08 ggherdov`_ joined #darcs
05:14 Big_G joined #darcs
06:56 castlelore joined #darcs
06:56 castlelore joined #darcs
08:07 Weltraumschaf joined #darcs
08:33 Weltraumschaf joined #darcs
09:03 thorkilnaur joined #darcs
12:34 Weltraumschaf_ joined #darcs
12:44 BitPuffin joined #darcs
13:35 dolio Thought: certain whitespace-only changes would be better off not affecting annotate output (because, e.g. I want to know who originally wrote the line, not who changed a separate line and re-aligned things to look nice). Is there a good way to accomplish this?
14:21 gh_ joined #darcs
14:22 gh_ hi
14:30 gh_ so, I'm going to send patches to drop support for GHC's older than 7.10 in HEAD and the 2.12 branch
14:36 Riastradh joined #darcs
14:51 gh_ dolio, sounds like an interesting feature ("darcs annotate --ignore-spaces"). But I'm pretty sure the fastest way to have it is to implement it into darcs' code.
15:03 dolio gh_: Oh. I mean, I know implementing it would get it. :) I meant more, thoughts on what the right approach would be.
15:03 dolio For instance, some whitespace-only changes would be semantically meaningful, and you might want to take that into account.
15:03 dolio Could, even.
15:04 dolio I wasn't asking for a way that I could somehow get darcs to accomplish it as-is.
15:12 gh_ ah alright :)
15:13 gh_ I guess there is no universal right approach because semantics of whitespaces depend on the programming language (on one hand python, haskell, etc. and on the other and C, Java, etc.)
15:35 gh_ patch sent to http://bugs.darcs.net/patch1504 , will screen it soon
16:09 Heffalump gh_: sounds good. You're very right about our slow responsiveness, I'd even completely forgotten about that sandi 0.4 patch from before
16:09 Heffalump There's this config fix we should get into the next 2.12 release too, I hope to have that ready within the next couple of days (issue2498)
16:53 sm gh_: great, thanks for that. darcs-2.12 + your patch builds easily with ghc 8 and stackage nightly-2016-09-01
16:57 sm once it's on hackage I'll get it re-added to stackage
17:35 gh_ Heffalump, sm, ok thanks for the feedback :) I'm going to self-accept the bundle then
17:58 mal`` joined #darcs
18:39 Riastradh joined #darcs
18:43 Weltraumschaf joined #darcs
19:09 pointfree dolio: yeah, wouldn't it be nice if "indentation" or "comment style" were stored in separate patches so I could pull in someone's changes but not their comment or indentation style?
19:10 pointfree or... perhaps what we need is a generalized darcs replace: https://twitter.com/lowfatcomputing/status/719749900519354369
19:10 dolio I'd just like to be able to fix formatting without taking ownership of the code.
19:11 dolio I don't want to have to choose between code that looks nice and good metadata from tools.
19:11 Heffalump I've been thinking about indentation patch types off and on, I think they're doable in theory
19:12 Heffalump but then there are two further issues to figure out: how to make use of them on existing repos where the changes weren't originally recorded as indentations, and how to make annotate etc "ignore" those patches
19:13 pointfree We could leverage the various syntax highlighting plugins that exist already for every language that exists.
19:16 dolio Of course, one solution is to not store text anymore. Only store syntax trees, which get formatted in some custom editor however you want to see it.
19:16 dolio But I'm not expecting that any time soon.
19:17 dolio I'm not sure I really even want it.
19:17 Heffalump yeah, it seems nice on the surface, but the details would be pretty painful
19:18 Heffalump I worked on a project that did that once, but the editor itself was a structure editor rather than working on text
19:18 pointfree not everything I put under source control is code.
19:22 pointfree I've been entertaining the idea of a dvcs that uses SDRs/semantic-fingerprints/context-vectors instead of hashes. SDRs would expose semantic relationships instead of hiding them.
19:22 Heffalump SDRs?
19:24 pointfree https://arxiv.org/abs/1503.07469 http://numenta.com/blog/wait-the-brain-is-a-bloom-filter.html https://github.com/numenta/nupic/wiki/Sparse-Distributed-Representations
19:24 Heffalump what would that mean in practice for a DVCS?
19:28 pointfree I've put the idea on ice for now because as with all of these machine learning oriented data structures there is a small fuzz factor which I don't like for version control.
19:28 pointfree Heffalump: I'm trying to eliminate positional logic from version control.
19:32 pointfree Victor Grishchenko described why he has the same goal with his ctre project here: https://github.com/gritzko/ctre/blob/master/doc/ws10.pdf
19:38 pointfree http://odroid.0xffffffff.in/~deploy/Rachkovskij%20Kussul%20Baidyk%20Building%20a%20world%20model%20with%20structure-sensitive%20sparse%20binary%20distributed%20representations%20BICA%202013%203%2064-86.pdf
19:39 pointfree ^ structure and function composition in SDRs
19:43 pointfree I'm trying to do something like semantic merge without maintaining grammars of various programming languages.
19:51 pointfree but don't hold your breath
20:01 Weltraumschaf joined #darcs
20:35 byorgey joined #darcs
20:41 byorgey joined #darcs
20:41 byorgey joined #darcs
23:25 Heffalump gh_: I sent in the patches for issue2498
23:26 Heffalump we should bump the QuickCheck dependency too
23:27 f-a joined #darcs
23:41 * Heffalump does so

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