00:01 molaf joined 00:31 moon-child joined
moon-child is there any special interaction between Inline::Python and threading|networking|async on the python side? I'm trying to use chatexchange, but it's not routing message properly 00:57
01:12 Altai-man_ joined 01:14 sena_kun left 01:33 tejr left 01:35 vike left 01:42 tejr joined 01:51 molaf left 01:57 vike joined 01:59 OpenZen left 02:02 Manifest0 left 02:03 Manifest0 joined, molaf joined 02:09 Sgeo left 02:11 Sgeo joined
brass Hey, I'm trying to understand how the rakudo/moarvm compiler toolchain works 02:13
How does NQP fit in between rakudo and MoarVM?
Is rakudo written in NQP and NQP compiles it and emits MoarVM bytecode? Or does NQP run directly on MoarVM which then loads rakudo? 02:16
02:16 Noisytoot left
brass Sorry if the questions are a bit silly 02:16
Is there a document that explains how it works?
tobs brass: these ones have a lot more stuff but probably also the answer you're looking for edumentab.github.io/rakudo-and-nqp...ls-course/ 02:18
brass Oh my God that's MASSIVE 02:20
Is there a lecture video that goes along with them?
tobs I know of none 02:21
brass Also thank you tobs that looks like a great resource! 02:22
tobs Slide 78 of day 1 starts a subsection called The compilation pipeline. Maybe start looking there, after checking the first ten or so slides. 02:23
brass On Day 1? 02:24
tobs yes
brass Ok thanks, I'll go over them tomorrow when I have more time to read all the slides properly 02:25
02:30 TimToady left 02:31 TimToady joined 03:12 sena_kun joined 03:13 rock6loose joined 03:14 Altai-man_ left 03:18 rockxloose left 03:30 rock6loose is now known as rockxloose 04:06 xinming left 04:07 xinming joined, lucasb left 04:08 nige joined 04:10 nige left 04:11 nige joined 04:12 guifa2 joined 04:49 nige left 05:03 ensamvarg3 joined 05:33 Sgeo left 05:35 Sgeo joined 05:42 ensamvarg3 left 05:44 andrzejku joined 05:51 xinming left 05:52 xinming joined 06:06 molaf left 06:12 Altai-man_ joined 06:14 sena_kun left 06:17 rindolf joined 06:36 aluaces joined 06:39 guifa2 left 07:14 JJMerelo joined 07:15 gnufr33dom left, skids left 08:13 sena_kun joined 08:14 Altai-man_ left 08:16 molaf joined 08:23 xinming left, leont joined 08:24 xinming joined 08:41 cpan-raku left 08:42 cpan-raku joined, cpan-raku left, cpan-raku joined 08:45 Noisytoot joined 08:52 xinming left
Geth doc: 18b1b74a7d | (JJ Merelo)++ | doc/Type/IO/Notification/Change.pod6
Starting to document IO::Notifcation::Change spinning it off IO::Notification refs #475

For the time being, it's mostly copy/paste from the original; next commit will work with `IO::Notification`
doc: 0913cf7275 | (JJ Merelo)++ | doc/Type/IO/Notification.pod6
Modifying IO::Notification, closes #475

Eliminates indexing and text that has been moved elsewhere.
linkable6 Link: docs.raku.org/type/IO::Notification::Change
08:52 xinming joined
linkable6 DOC#475 [closed]: github.com/Raku/doc/issues/475 [docs][search][update] IO::Notification::Change is not documented
Link: docs.raku.org/type/IO::Notification
09:06 Sgeo left 09:08 JJMerelo left 09:23 Kaeipi joined, Kaiepi left 09:25 Kaiepi joined, Kaeipi left 09:55 JJMerelo joined 09:59 xelxebar_ joined 10:00 xelxebar left 10:12 Altai-man_ joined
tbrowder anyone still awake? 10:12
Geth doc: fc65fbebc4 | (JJ Merelo)++ | type-graph.txt
Adds IO::Notification::Chane to typegraph refs #475
doc: 7615186c3b | (JJ Merelo)++ | doc/Type/IO/Notification.pod6
Added an example and some details. Closes #2828
doc: b0b8f0270c | (JJ Merelo)++ | doc/Type/IO/Notification/Change.pod6
Adds reference to .watch, basically the same, #475
linkable6 DOC#475 [closed]: github.com/Raku/doc/issues/475 [docs][search][update] IO::Notification::Change is not documented
linkable6 Link: docs.raku.org/type/IO::Notification
DOC#2828 [closed]: github.com/Raku/doc/issues/2828 [RFE] Improve IO::Notification
Link: docs.raku.org/type/IO::Notification::Change 10:13
tbrowder i have a question on testing and subtests
10:14 sena_kun left
tbrowder how, using our testing setup for rakudo spec testing (with roast), does one see individual test msgs in subtests? 10:15
i wrapped a set of 10+ tests in a subtest and get one msg about test 1 ok. 10:16
i've tried using diag and note but nothing. 10:17
10:18 wamba joined
tbrowder i'm using rakudo's test harness like this "make t/spec/S26*/12-sometest.t" 10:18
Geth doc: 2d1615fdcb | (JJ Merelo)++ | doc/Type/IO/Notification.pod6
Tags correctly text
doc: dffa311e91 | (JJ Merelo)++ | doc/Type/IO/Notification.pod6
Tags _really_ correctly text
linkable6 Link: docs.raku.org/type/IO::Notification
JJMerelo tbrowder I think you need to use --verbose; they will be hidden if you don't 10:21
tellable6 2020-08-02T09:17:54Z #raku-dev <tobs> JJMerelo How did you invite me to the advent calendar? I didn't get anything. Feel free to /msg when you're around.
10:23 molaf left
tbrowder yeah, i think that's not available if i don't use prove directly. just want to "prove" my roast changes work for vrurg. 10:23
yuk yuk 10:24
jj, is the 20th advent for this Christmas or for earlier? 10:25
JJMerelo: ^^^
JJMerelo tbrowder it's for right now 10:27
We've already published a couple of posts 10:28
tbrowder and we'd love to have your collaboration...
tbrowder you can also just run with raku -Ilib 10:29
tbrowder i really can't help right now, sorry
JJMerelo: thanks for the "raku -I lib" that's what i needed. so, how long would i have to get an article together? 10:34
JJMerelo tbrowder Well, they're gonna be published until the 20th. So, 17 days tops. Of course, if they're ready earlier than that, that would be awesome 10:35
tbrowder erg...can't promise but i'll try to take a look today 10:36
10:40 JJMerelo left 11:40 linkable6 left, evalable6 left 11:41 linkable6 joined, evalable6 joined 12:13 sena_kun joined 12:14 Altai-man_ left 12:25 MasterDuke left 12:33 andrzejku left 12:39 xinming left 12:40 xinming joined 12:56 _jrjsmrtn joined 12:58 __jrjsmrtn__ left 13:30 xinming left 13:31 xinming joined 13:35 MasterDuke joined 13:38 cpan-raku left, xinming left 13:39 cpan-raku joined, cpan-raku left, cpan-raku joined 13:40 xinming joined 13:41 phogg left 14:02 gnufr33dom joined 14:08 phogg joined, phogg left, phogg joined, xinming left 14:09 Manifest0 left 14:12 Altai-man_ joined, xinming joined, Manifest0 joined 14:14 sena_kun left 14:21 xinming left 14:22 xinming joined 14:31 demostanis joined
demostanis Hello everyone! Why is Zef so slow? It literally took 50 seconds to run command `zef install UNIX::Privileges` 14:32
14:37 NickSeagull joined
Altai-man_ demostanis, hi. There are number of reasons, primary is that the current language implementation is not optimized enough, also the code itself is not optimized enough. 14:45
tobs demostanis: when you install something else now, is it faster? I remember such slowdowns after compiler updates which trigger re-precompilation, but I don't know if that's still a practice. 14:53
Altai-man_ Just installed it in `21,64 secs`. 14:59
15:04 xinming left 15:05 xinming joined
demostanis tobs: I tried installing NCurses to see, it takes 1 minutes and 47 seconds! 15:05
Altai-man_: Is optimizing the code planned in the future? 15:06
Altai-man_ demostanis, sure, there are ongoing projects on optimizing moarvm, as well as compiler-level things getting improvements, you can grep rakudo changelog for `fast` at github.com/rakudo/rakudo/blob/mast.../ChangeLog 15:07
More hands and funding would speed up development, but now we have what we have. 15:08
Altai-man_ tries to install NCurses
timotimo sees lots of cursing in Altai-man_'s near future 15:09
Altai-man_ Dammit
timotimo fiddlesticks!
Altai-man_ forgot to add `time` before `zef install NCurses`
Executed in 75,08 secs 15:10
demostanis That's faster than me 15:11
Altai-man_ That's kinda long, though I also have boinc attacking all the cores constantly.
timotimo annoyingly, you can currently not --profile a program that uses exit, and zef uses exit in many places
Altai-man_ demostanis, also, it is not zef which is hella slow. Zef just runs tests of the modules and some of them have not lots of tests, while others have plenty. For example, Cro::HTTP has a looot of tests. 15:12
More tests to execute takes longer.
demostanis Installing UNIX::Privileges without tests takes 43 seconds instead of 47. 15:14
rypervenche You'll really see the difference with modules that have a lot of (intensive) tests. Cro will show a bigger difference. 15:18
15:26 kensanata joined 15:44 JJMerelo joined 15:46 guifa2 joined 15:48 sena_kun joined 15:49 lucasb joined 15:50 andrzejku joined
raku-bridge <Fernando Santagata> I noticed that modules that use NativeCall are slow to install, even when the module contains just the C-language function declarations (sub xxx() is native is export { * }) Is the compilation process slow because it looks for the function starting location in the C library? (never looked inside Rakudo's guts) 15:51
15:53 Altai-man_ left, Altai-man_ joined
timotimo giving --stagestats to the rakudo commandline that precompiles these modules can give a tiny hint usually 15:59
guifa2 is there anyway to modify a signature at a runtime?
guifa2 isn't sure if .wrap is supposed to allow new args or not. 16:01
I feel like it should allow for its own signatures to override the original but it appears not to be the case in current Rakudo. But not sure if I should file an issue or if that's the intended design
16:03 sena_kun left, parsifal joined, parsifal left 16:10 Kaeipi joined 16:11 Kaiepi left 16:13 sena_kun joined 16:14 Altai-man_ left
JJMerelo .tell tbrowder thanks 16:18
tellable6 JJMerelo, I'll pass your message to tbrowder
16:20 Kaeipi left
timotimo i imagine a wrapped routine would want to return the wrapper's signature when introspected 16:20
guifa2 timotimo: actually I think I figured out 16:21
it seems to be more of a problem with wrapping multis
I think
I got what I needed to do to work but I'll investigate more to make sure behavior is as expected or could be improved 16:22
16:23 Kaiepi joined
timotimo without having looked closely, i think calling .signature on a multi ought to give you the signature of the proto, and if the multi is wrapped (which really means the proto got wrapped, unless you're really wrapping an individual candidate directly) ... 16:23
then you'd get the wrapper's signature
though i imagine you can't get at the candidates that way any more 16:24
which also makes sense, since you're not guaranteed to have the same dispatch semantics any more if the proto got wrapped
i.e. the wrapper could switch argument order or change types or whatever
then the individual signatures are not useful information any more
guifa2 Okay I think I've finally got this. Nwo just a question on naming 16:31
16:32 cpan-raku left
guifa2 There's two different timezone IDs. There's the finer geographical (America/New_York, Europe/Madrid) and then there's the broader political (Central European, Atlantic, etc) 16:32
16:32 xinming left
guifa2 I'm thinking .olson[-id] for the former, and .tz[-abbr] for the latter. .timezone is already taken up by the gmt-offset value 16:33
16:35 xinming joined 16:40 mandobot joined, mandobot left 16:49 cpan-raku joined, cpan-raku left, cpan-raku joined
JJMerelo timotimo we want _you_ to write an article for the 20th anniversary calendar 16:52
JJMerelo goes AFK and CTTS (close to the surf) while expecting lots of candidates for 20th anniversary articles.
guifa2 JJMerelo: so sorry I'm behind getting mine done 16:54
Finishing it up now
16:57 JJMerelo left, mandobot joined, mandobot left 17:01 andrzejku left
rypervenche I'm looking over this article: perl6.party/post/IRC-Client-Perl-6...IRC-Module and I was curious. Why is $ used in the signature, like in... multi method irc-to-me ($ where /bash/) {... 17:44
guifa2 it basically says "if the first argument contains bash" 17:45
but the method itself doesn't actually need the contents of the argument
so no symbol is installed in the scope for it 17:46
rypervenche Ah, I see. Interesting.
So it's kind of a throwaway (not really a) variable then. 17:47
guifa2 exactly 17:48
That's probably the most common use for them, although you can use them in loops and stuff
m: for ^10 { say $++ } 17:49
camelia 0
rypervenche Nice. Knowledge++ 17:50
guifa2 Ack and another naming issue 17:54
.is-dst or .is-summer or .is-daylight or ? to find out if lcocks are currently shifted 17:55
rypervenche I like the first. .is-summer feels confusing. 18:00
guifa2 in Europe it's called summer time though and not daylight savings time 18:01
the super neutral would be .is-standard but that just feels.....too generic 18:02
18:02 tomaw left 18:03 tomaw joined
Geth doc: f3bd28f499 | Coke++ | xt/code.pws
track new word in code
doc: 7dc41501c7 | Coke++ | doc/Language/intro.pod6
doc: 429a319e29 | Coke++ | xt/code.pws
new variable names used in code
linkable6 Link: docs.raku.org/language/intro
18:06 NickSeagull left
tbrowder guifa2: +1 for is-dst 18:12
tellable6 2020-08-02T16:18:33Z #raku <JJMerelo> tbrowder thanks
18:12 Altai-man_ joined
guifa2 actually, crap. I'm going to need to have an extra named argument for .new to decide whether to use standard or daylight time when interpretting the time 18:13
*sigh* Working with timezones is annoying. I can see why core Raku just leaves it at an offset and nothing more :-) 18:14
18:14 sena_kun left 18:22 molaf joined
demostanis Any way I can ignore circular module loading errors? I have this code: use B; class A { method dosomething(B $b) { ... } } in B.rakumod and: use A; class B { method dosomethingelse(A $a) { ... } } 18:32
in A.rakumod
But it fails... 18:33
guifa2 use A; class B { method dosomethingelse(::('B') $a) { ... } 18:37
::('A') obviously
if possible, just put them all in one file though and stub one of them 18:38
18:52 OpenZen joined
tbrowder .ask JJmerelo i want add a little blurb on testing pod in the docs. my guess it should be on the Testing page, but let me know if you (or anyone interested) have a better place for it. 18:57
.ask JJmerelo ^^^ 18:58
tellable6 tbrowder, I'll pass your message to JJMerelo
tbrowder stupid leading space should be IGNORED 18:59
Geth advent: aea0e616f4 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 20th/README.md
Linkify RFC's to be blogged about
19:17 Altai-man_ left 19:19 Sgeo joined 19:20 Altai-man_ joined 19:26 kensanata left 19:28 phogg left 19:32 grumble left 19:39 grumble joined
Geth advent: 41f10f9f9f | (Vadim Belman)++ | 20th/README.md
Claim RFC 137
19:53 phogg joined, phogg left, phogg joined, rindolf left 20:10 camelCaser left 20:12 camelCaser joined
Geth advent: 3c2b785715 | Altai-man++ (committed using GitHub Web editor) | 20th/README.md
Claim RFC 188
20:12 sena_kun joined 20:14 Altai-man_ left
AlexDaniel` committable: uptime 20:33
committable6 AlexDaniel`, 3 weeks, 6 days, 4 hours, 14 minutes, and 23 seconds, 970.675781MiB maxrss. This is Rakudo version 2020.02.1-395-g2c9168b9a built on MoarVM version 2020.02.1-162-g3d8ff61e2 implementing Raku 6.d.
AlexDaniel` Dave: ↑
Nasrudin ah 20:41
20:45 wamba left 20:57 molaf left 21:16 phogg left
jdv79 it seems to me the docs on TWEAK are a bit misleading. they say "The TWEAK submethod allows you to check things or modify attributes after object construction:" but in a tweak if you try to use an accessor you'll still get something like "Virtual method call $.user may not be used on partially constructed object (maybe you mean $!user for direct attribute access here?)" 21:30
to me, a partially constructed instance is not possible "after" construction, no ?
21:32 demostanis left
jdv79 s/after/at the end of/ is probably better 21:35
guifa2 that's odd, I haven't had problems with that 21:41
m: class A { has $.foo; method TWEAK { say $.foo } }; A.new(:2foo);
camelia 2
guifa2 m: class A { has $.foo; method BUILD { say $.foo } }; A.new(:2foo); 21:44
camelia (Any)
jdv79 "submethod" 21:49
guifa2 ah whoops 21:51
21:57 gordonfish left 21:58 gordonfish joined 22:12 Altai-man_ joined 22:14 sena_kun left 22:35 leont left
raku-bridge <stu002> Posted this on the Raku discord, apologies if you're seeing it twice: 22:53
<stu002> Can any MacOS rakudo star users advise on this message please: "I ran into a problem while trying to set up Readline: Could not instantiate role 'ReadlineBehavior': Cannot locate native library '(null)': dlopen(libreadline.7.dylib, 10): image not found" Do I need to setup LD_LIBRARY_PATH for the libreadline library? This has been installed with Homebrew and with a working readline.
vrurg stu002: I don't have much time now, can only advise trying to symlink libreadline into ~/lib 23:18
tellable6 2020-08-02T07:15:06Z #raku-dev <JJMerelo> vrurg a short article will do :-) Pick your RFC!
2020-08-02T10:50:23Z #raku-dev <tbrowder> vrurg i just pushed my new changes to roast_657. looks pretty good to me thanks to your clever pod hackery! gonna put that in my tool box. it needs to be in the docs somewhere.
23:20 phogg joined, phogg left, phogg joined
tbrowder can anyone please point me to a howto for the new test hook for githup workflows used by mi6? thnx 23:27
is it preferred now over travis-ci? 23:28
holyghost The only thing I wanted to remark on the RFCs and milestones pages, a battleplan file 23:30