Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
00:42 klapperl left 00:49 klapperl joined 05:07 vrurg left 06:22 robertle left 06:32 Kaiepi left, Kaiepi joined 06:40 epony joined 07:40 ufobat joined 07:46 robertle joined 07:56 committable6 left, undersightable6 left, shareable6 left, squashable6 left, unicodable6 left 07:57 benchable6 left, coverable6 left, statisfiable6 left, greppable6 left, notable6 left, nativecallable6 left, reportable6 left, evalable6 left, releasable6 left, bloatable6 left 07:58 quotable6 left, bisectable6 left 08:08 quotable6 joined, ChanServ sets mode: +v quotable6 08:09 releasable6 joined, statisfiable6 joined, ChanServ sets mode: +v statisfiable6, committable6 joined, ChanServ sets mode: +v committable6, greppable6 joined, ChanServ sets mode: +v greppable6 08:10 squashable6 joined, shareable6 joined, undersightable6 joined, coverable6 joined, ChanServ sets mode: +v coverable6, bisectable6 joined, evalable6 joined, ChanServ sets mode: +v evalable6, nativecallable6 joined, ChanServ sets mode: +v nativecallable6, bloatable6 joined, ChanServ sets mode: +v bloatable6, unicodable6 joined, ChanServ sets mode: +v unicodable6 08:11 notable6 joined, ChanServ sets mode: +v notable6, benchable6 joined, ChanServ sets mode: +v benchable6, reportable6 joined, ChanServ sets mode: +v reportable6 08:43 leont joined 10:22 pmurias joined 10:31 ExtraCrispy joined 10:42 ExtraCrispy left
lizmat Files=1266, Tests=88069, 410 wallclock secs (21.15 usr 6.55 sys + 2959.26 cusr 242.88 csys = 3229.84 CPU) 10:51
10:55 ExtraCrispy joined
pmurias is ashamed to admit he has to run roast in rakudo.js (bundling the js and running it in chrome) overnight :( 11:38
timotimo hmm 11:39
pmurias I'll have to experiment more with parceling/compiling and loading multiple test files at once 11:43
12:28 pmurias left
lizmat notable6: weekly 12:58
notable6 lizmat, 11 notes: gist.github.com/fad50d8156f1c0ab0a...cfe3414246
AlexDaniel also: 2019.03.1 released 12:59
lizmat reportable6: 2019-03-11T12:00:00Z 2019-03-18T00:00:00Z 13:28
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
lizmat, gist.github.com/e627d4ff823c9ba647...5439ab3f91 13:29
AlexDaniel that doesn't look right 13:30
oh, actually
all of these were labeled O_O
lucasb++
that probably deserves a mention in the weekly 13:31
lizmat AlexDaniel: what does ?
AlexDaniel lizmat: lucasb++ labeled more than 250 tickets 13:32
lizmat aaahh ok
AlexDaniel gist.github.com/Whateverable/e627d...ickets-294
13:36 leont left
lizmat hopes we'll see a Rakudo Star release before the P6W is published 13:42
13:59 vrurg joined
lizmat notable6: weekly reset 14:19
notable6 lizmat, Moved existing notes to “weekly_2019-03-18T14:19:10Z”
14:58 vrurg left 15:00 vrurg joined 15:38 pyrimidine left 15:48 gfldex left, gfldex joined 15:51 ggoebel joined 15:56 pyrimidine joined 16:06 robertle left
lizmat and another Perl 6 weekly hits the Net: p6weekly.wordpress.com/2019/03/18/...te-course/ 16:14
jnthn lizmat++ 16:20
yoleaux 06:24Z <nwc10> jnthn: will master merged in, more-pea gets all the way to Rakudo spectests (but does not pass all of them)
16:28 ufobat left
AlexDaniel I don't understand 16:43
why people submit grant proposals without coming here first and asking for feedback
timotimo not everybody is as irc-centric as we are 16:44
AlexDaniel send an email then 16:45
start a git repo
whatever
lizmat AlexDaniel: add a comment ? It's not too late :-) 16:47
AlexDaniel of course I will, except that I don't think it's going to change anything
if it's accepted, they will do whatever they want without really trying to integrate their work with the community 16:48
if it's not accepted, then we're losing the chance to get some of the proposed stuff done
lose-lose 16:49
lizmat well, it's that type of comment that the grant committee would like to know about
ugexe "they will do whatever they want without really trying to integrate their work with the community" -- can you back this up? 16:50
AlexDaniel they have a plan, and it is described in detail in the proposal. The plan is to create a separate resource that is independent from the docs, and to solve some problems (e.g. code execution) for that separate website 16:53
ugexe can the docs infrastructure support his plan?
AlexDaniel now of course maybe they'll rethink some of the parts, but it's going to change the proposal significantly, so I don't believe that can happen. I co 16:54
I can't prove it to you unless we actually wait and see what happens, obviously
no, but there is no infrastructure yet
timotimo proposals can be resubmitted in the next round
AlexDaniel instead of modifying the doc website the proposal is to create a new website from zero 16:55
ugexe i see him as trying to get the actual valuable work done and avoiding the bike-shedding stuff that can be handled later by whoever
AlexDaniel I see him not giving a fuck about everything and everyone else and doing everything his way 16:57
and when the bus comes, *we* will have to maintain that separately
ugexe if it needs to be integrated into the docs site at a later time then im sure they are not going to stop anyone. but thats lots of additional scope that isnt strictly needed for deliverable
AlexDaniel so, fund his work so that someone else can then later clean it up and integrate into what we've been working on? 16:59
this is making zero sense to me
ugexe its not often i hear the sentiment of 'we need more scope' 17:06
AlexDaniel it's not more scope, it's in fact about reducing the amount of work in total
we need code execution on the docs too, why not solve that also with one shot 17:07
ugexe well you are trivializing the complexities of these features
maybe you are right. i dont think you are.
"why not just ..." 17:08
17:08 jmerelo joined
ugexe if you could convince me its actually a reduction of scope, then i'm sure you could convince them as well. 17:09
AlexDaniel the scope is same, but all the details will be different 17:10
jmerelo Infrastructure maintenance is difficult enough as it is now. Throwing another thing would be the proverbial straw that broke Camelia's back. 17:11
AlexDaniel e.g. not markdown but pod6, not a separate domain but docs.perl6.org, etc.
and then you don't have to worry about user experience issues like users having to jump from docs to that website and back (like if they want to see other methods of some class like Str or whatever), everything is in the same place 17:12
jmerelo Plus being able to run whatever is in the doc sites would also mean we would be able to test the output of everything that's in there, and that would be just awesome. Wd don't do that now. From time to time, "output is not that" issues appear. 17:14
AlexDaniel btw I feel like using rakudo-js is a good idea for that thing. The plan is to have some introductory tutorial, and most basics I think work with rakudo-js
jmerelo AlexDaniel: The problem with rakudo-js is that right now it takes a while to load; it's rather heavy. If we found some way of lazy loading, that would be much better for mobile devices, for instance. It's great for desktop, however. 17:15
AlexDaniel jmerelo: yeah but if you want to go through the course, it's probably OK to wait a bit for it to load 17:24
17:40 |Tux| left
jmerelo AlexDaniel: right. 17:44
timotimo AlexDaniel: unless you say that "for every 0.5s a site takes longer to load, you lose 50% of your audience" or so 17:46
i believe there's a new API that lets you decouple the computations inside an iframe from the frame that contains it 17:47
17:47 |Tux| joined 17:48 robertle joined 18:35 lgtaube left 18:51 lgtaube joined
AlexDaniel well, I wrote something 19:57
there are however certain limits on how much time I'm willing to spend on justifying my position when the submitter spent zero time on doing the same besides saying that it will be “different from the documentation” 19:59
20:01 Kaiepi left 20:02 Kaiepi joined 20:48 hankache joined 21:04 jmerelo left 21:17 epony left
gfldex I'm getting: 21:17
Type check failed in binding to parameter '$s'; expected Str but got Int (100000)
the line in question is: 21:18
CATCH { .fail }; from-json post |c
where is that $s coming from?
jnthn: can your inlining improvements cause Rakudo to point to the wrong line? 21:19
21:20 lucasb joined
gfldex also that is clearly a regression 21:21
21:23 epony joined
jnthn gfldex: There's long been cases of optimizations throwing line numbers off a bit; you can always run without if you think that's to blame (just set MVM_SPESH_DISABLE=1 in the environment) 21:25
There's something a bit up with line numbers even without that, though, I think; line numbers around while loops, for example, seem to have something a bit iffy, because coverage data comes out showing a loop body statement being hit, but not the loop's opening line, which is trivially wrong. 21:27
The $s will come from some varible inside of from-json or post, I guess
gfldex jnthn: I will dig deeper. 21:28
21:30 hankache left
gfldex jnthn: it can't be post because the whole file doesn't contain a single $s. It is: github.com/perl6-community-modules...ib/WWW.pm6 21:32
cant't be from-json either 21:33
here is a complete stacktrace gist.github.com/74f72a5f25badd4ee3...77a880b1b9 21:35
timotimo it does have from-json in there 21:48
and it looks like it's trying to stringify a failure object perhaps?
but yeah, there doesn't seem to be a parameter or variable named $s anywhere in there, but perhaps it has something to do with the coercer? 21:50
jnthn Well, it could also be in the setting, which is, after all, just a bunch of Perl 6 code :) 22:13
gfldex jnthn: golfed: gist.github.com/c357af88e042d5af39...ccc7188dde 22:15
shall I rakudobug? 22:17
timotimo gfldex: can you golf the contents of the jpost sub a little? maybe cut out the http request completely? 22:25
oh, does it die before it can even do that?
gfldex looks like it 22:41
bisectable6: gist.github.com/c357af88e042d5af39...ccc7188dde 22:44
bisectable6 gfldex, It looks like a URL, but mime type is ‘text/html; charset=utf-8’ while I was expecting something with ‘text/plain’ or ‘perl’ in it. I can only understand raw links, sorry.
gfldex bisectable6: gist.githubusercontent.com/gfldex/...ntifier.p6
bisectable6 gfldex, Successfully fetched the code from the provided URL
gfldex, Bisecting by output (old=2015.12 new=6d8077c) because on both starting points the exit code is 1
gfldex, bisect log: gist.github.com/f800892956479ed784...8e7b6157a4 22:45
gfldex, (2015-12-25) github.com/rakudo/rakudo/commit/07...dc61f84053
gfldex, The result looks a bit unrealistic, doesn't it? Most probably the output is different on every commit (e.g. ｢bisect: say rand｣)
22:53 Kaiepi left, Kaiepi joined
vrurg When ACCEPTS is being called by a metaobject at compile time find_best_dispatchee only considers the two candidates from Mu ignoring one(s) from the invocant class. Is it normal? 23:00
jnthn Probably, yes. Multi-dispatch candidates are not incorporated until composition time. 23:01
And further, cannot be.
And I guess you're saying that the class hasn't been composed yet 23:02
By "compile time"
Nothing specific to ACCEPTS, just a genreal properly of multi-methods
AlexDaniel gfldex: yeah it doesn't have that module
vrurg jnthn: ACCEPTS is just what I've got on my hands. But generally, yes – it'd be any method. 23:03
Then, so far, it means there is no way to make subsets bahave similarly to definities.
*behave
jnthn It's not clear how that follows 23:05
Example?
vrurg SubsetHOW, method accepts_type checks against $!refinement by calling ACCEPTS on it. This is a universal way to handle any kind of refinees: blocks, regexes, etc. 23:06
jnthn Yes, following so far. 23:08
vrurg In actions default value is checked with check_default_value_type.
subset S of Any where { True }; my S $v; – fails to compile. 23:10
jnthn Why? Block is composed (it comes from CORE.setting) and S is composed by that point 23:11
m: subset S of Any where { True }; BEGIN say Any ~~ S 23:12
camelia True
vrurg This is what I'm trying to find out now. A smal but important note: current code avoids calling ACCEPTS on SubsetHOW because the $!default is set to the subset itself and accepts_type validates against subset itself. 23:14
So, this happens to be enough. I have removed the check against self making $!refinement.ACCEPTS mandatory.
23:51 MasterDuke joined 23:52 MasterDuke left, MasterDuke joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!