»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend! 🦋
Set by Zoffix on 25 May 2018.
00:04 MasterDuke left 00:07 imcsk8 left, imcsk8 joined 00:08 MasterDuke joined
wbn tyil: i am not #1 expert at this but wouldn't you use a Grammar here? 00:08
to parse the module name into a class with like $!name, $!ver, $!api etc 00:09
tyil possibly
wbn then you could match based on class attributes
tyil preferably I'd do as little work as possible, of course
perhaps someone knows of a module, or other thing that will Just Work for me 00:10
I'd expect something would exist for this, for use in zef for instance
wbn yeah 00:11
just as a first go though you might find it fairly simple (as long as the parameters of the problem are known to you) 00:12
tyil all tags that work for module dependency entries must be supported 00:13
wbn yeah maybe not so simple
tyil I could match all :(key)<(value)> stuff and throw it into a hash, and then check if all the things provided in the query are correct in the hash 00:14
if nothing exists for it, it sounds like an interesting project for tomorrow :p
00:16 raynold joined 00:20 imcsk8 left, imcsk8 joined 00:23 epony left 00:28 noganex joined 00:31 noganex_ left 00:34 molaf left
lookatme tyil, some project need to parse the META6.json, such as zef, maybe you can have a look at their code, 00:34
wbn, some project need to parse the META6.json, such as zef, maybe you can have a look at their code, 00:35
sorry :)
00:35 aindilis left 00:36 aindilis joined 00:38 markoong left 00:41 Zoffix joined
Zoffix tyil: just eval the entries and extract it that way? 00:42
Like, "Config::Parser::toml:ver('1.0.1+')" is also a valid way to write that stuff 00:43
m: my $v := ｢Config::Parser::toml:ver('1.0.1+'):api<meows>｣; with 'class \qq[$v] {}'.EVAL { [.^ver, .^api].say } 00:45
camelia ( no output )
Zoffix m: my $v := ｢Config::Parser::toml:ver('1.0.1+'):api<meows>｣; given 'class \qq[$v] {}'.EVAL { [.^ver, .^api].say }
camelia [v1.0.1+ meows]
Zoffix m: my $v := ｢Config::Parser::toml:ver('1.0.1+'):api<meows>｣; given 'class \qq[$v] {}'.EVAL { [.^name, .^ver, .^api].say }
camelia [Config::Parser::toml v1.0.1+ meows]
lookatme Zoffix, great!
wbn oh wow 00:46
Zoffix Depending on what you're building it might be OK to use &EVAL for this (if it's not obvious, you're executing arbitrary code so someone can stick evil things into "depends" key to make this stuff run evil things)
00:47 lex_ joined
Zoffix And if you don't wanna use it, might look into stealing `token longname` and related machinery from github.com/rakudo/rakudo/blob/mast...rammar.nqp 00:47
The $*W you might see is the Perl6::World object: github.com/rakudo/rakudo/blob/mast.../World.nqp 00:48
00:55 stmuk_ joined 00:57 stmuk left, epony joined 00:59 jameslenz joined 01:00 lex_ left 01:01 lex joined, lex is now known as lindylex 01:04 jameslenz left 01:12 fake_space_whale joined 01:13 DarthGandalf left 01:15 DarthGandalf joined 01:27 Zoffix left 01:44 spycrab0 left 01:49 stmuk joined 01:50 subroot joined 01:51 stmuk_ left 01:57 aborazmeh joined, aborazmeh left, aborazmeh joined
benji__ What HTML parser do we recommend? 02:06
Gumbo, HTML::Parser, and HTML::MyHTML seem outdated 02:08
02:28 imcsk8 left, imcsk8 joined 02:29 lindylex left 02:38 ufobat___ joined 02:41 ufobat_ left
b2gills benji__: This is the the time when the fewest people are active. It may be best to ask again in 6+ hours. (I haven't had a need to parse HTML) 02:43
03:00 jameslenz joined 03:04 jameslenz left 03:14 psychoslave joined 03:20 MasterDuke left 03:31 salasrod joined 03:32 subroot left 03:34 epony left 03:42 benji__ left 04:05 BenGoldberg joined 04:16 epony joined 04:35 Aceeri_ left 04:47 ToddAndMargo joined
ToddAndMargo Hi All, How do I do an "or" inside a subsitute (I am converting from Perl 5)? p6 'my $x="1_2-3:4u5"; $x ~~ s:global/ "-|_|:|u" / "." /; say $x;' I want `1.2.3.4.5` 04:48
figured it out. It is all in the quotes: p6 'my $x="1_2-3:4u5"; $x ~~ s:global/"-"|"_"|":"|"u" /./; say $x;' 1.2.3.4.5 04:51
04:52 imcsk8 left, imcsk8 joined 04:57 aborazmeh left 04:58 curan joined
tobs ToddAndMargo: or use a character class with <[...]> 04:58
m: say S:g/<[-_:u]>/./ given $ = "1_2-3:4u5"
camelia 1.2.3.4.5
tobs (you can even forget about the "$ =" part) 04:59
04:59 imcsk8 left 05:00 imcsk8 joined, jameslenz joined 05:03 xtreak joined 05:05 jameslenz left 05:08 BenGoldberg left 05:10 fake_space_whale left 05:15 sauvin joined 05:23 troys left 05:24 fake_space_whale joined
ToddAndMargo $ p6 'my $x="1_2-3:4u5"; $x ~~ s:global/ <[-_:u]> /./; say $x;' 1.2.3.4.5 Thank you! 05:25
05:26 Sgeo__ left 05:27 ToddAndMargo left 05:34 kerframil left 05:38 Ven` joined 05:42 sno left, Ven` left 05:43 psychoslave left 05:46 fake_space_whale left 05:51 robertle joined 05:58 sno joined 06:01 wamba joined 06:03 xinming_ left 06:04 raynold left, xinming joined 06:06 x[LGWs4x4i]uG2N0 joined 06:39 konsolebox left 06:41 konsolebox joined 06:51 hami joined 06:54 psychoslave joined
lookatme m: my $x="1_2-3:4u5"; $x ~~ s:global/ <[-_:u]> /./; say $x; 06:58
camelia 1.2.3.4.5
07:00 domidumont joined, jameslenz joined 07:04 jameslenz left 07:06 domidumont left 07:07 domidumont joined 07:12 benjikun joined 07:39 gregf_ left 07:40 salasrod left, xtreak left 07:42 lizmat joined 07:43 xtreak joined 07:55 wamba left 08:08 casdr joined, hami left 08:18 dakkar joined 08:23 Technato1 joined 08:27 wamba joined 08:30 Ven` joined 08:34 ChoHag_ joined, ChoHag left 08:41 wamba left 08:43 xtreak left
Geth doc: 1ddd7c8858 | (JJ Merelo)++ | README.md
Adds notice for p6doc installation closes #2166
 08:46
08:46 spycrab0 joined 08:50 wamba joined 08:58 Technato1 left 08:59 wamba left 09:00 jameslenz joined 09:05 jameslenz left 09:07 wamba joined 09:16 lizmat left 09:18 sena_kun joined 09:19 xtreak joined, lizmat joined 09:22 robertle left 09:25 Yogesh joined 09:27 Yogesh left 09:30 nativecallable6 joined 09:38 lizmat_ joined 09:40 lizmat left 09:48 jmerelo joined
Ulti weekly: news.ycombinator.com/item?id=17496789 09:49
notable6 Ulti, Noted!
lizmat_ Ulti++
09:49 lizmat_ is now known as lizmat
Ulti little front page trend there thanks to brrt 09:49
jmerelo Ulti++ 09:54
Geth Pod-To-HTML: f121904216 | (JJ Merelo)++ | README.md
Clarifies minimum version

Refs to perl6/doc#2166, which reported the problem.		 10:03
synopsebot DOC#2166 [closed]: github.com/perl6/doc/issues/2166 [docs] zef install p6doc fails on tests
jmerelo "Routines are the smallest means of code reuse in Perl 6". Is that true? Aren't blocks "smaller"? (that's from the functions doc) 10:05
lizmat yes, Blocks are the smallest that make sense at the HLL level 10:07
jnthn WhateverCode is even smaller, since it doesn't have its own lexical scope :P 10:08
lizmat you could consider Code smaller still
jnthn You can't really write a Code explicitly though :)
lizmat that's what I mean, for "normal" usage I think we should consider "Block" the smallest, no? 10:09
jnthn Probably yes :)
Geth doc: deb3c72da7 | (Tom Browder)++ (committed using GitHub Web editor) | doc/Language/functions.pod6
“one” refers to a person
 10:12
synopsebot Link: doc.perl6.org/language/functions
jmerelo jnthn, lizmat: better then to change that in the documentation, right? 10:21
10:22 pmurias joined
jmerelo Maybe simply "Routines are a way to reuse code in Perl t" 10:23
Maybe simply "Routines are a way to reuse code in Perl 6"
Geth doc: 6f9840a739 | (Luca Ferrari)++ | doc/Language/functions.pod6
Fix ambigous statement about %*SUB-MAIN-OPTS.

See <github.com/perl6/doc/commit/deb3c7...668834>		 10:26
synopsebot Link: doc.perl6.org/language/functions
Geth doc: 9761710770 | (Luca Ferrari)++ | doc/Language/functions.pod6
Change truthy to true.
doc: 64d8419049 | (Luca Ferrari)++ | doc/Language/functions.pod6
Small rephrasing of named-anywhere problem.
doc: 01e8a77f97 | (JJ Merelo)++ | doc/Language/functions.pod6
Rephrases routine def

Explanation in the #perl6 channel log colabti.org/irclogger/irclogger_log...07-11#l182		 10:28
doc: 6dec88921d | (Luca Ferrari)++ | doc/Language/functions.pod6
Change script to program.
 10:29
doc: 1d721a4202 | (Luca Ferrari)++ | doc/Language/functions.pod6
Merge branch 'master' of github.com:perl6/doc
synopsebot Link: doc.perl6.org/language/functions
10:42 jmerelo left 10:45 Some-body_ joined, DarthGandalf left
tbrowder_ i would like to use a unit scoped MAIN but can i somehow specify an option parameter? i have tried putting a ? after the param as well as a default value. shouldn’t that work (assuming i don’t mess up anything else) 10:45
10:46 Some-body_ is now known as DarthGandalf 10:47 Zoffix joined
tbrowder_ *optional parameter? 10:47
Zoffix eco: DOM::Tiny
buggable Zoffix, DOM::Tiny 'Very small, self-sufficient DOM parser manipulator': github.com/zostay/p6-DOM-Tiny 1 other matching results: modules.perl6.org/s/DOM%3A%3ATiny
Zoffix benjikun: I assume you are the benji___ ? ^ that's a good HTML parser
(might want to install with --/test as it got a ton of them and takes forever to install) 10:48
10:48 Zoffix left 10:50 Ven` left
tyil tbrowder_: got a link to the source you're using? 11:01
on the MAIN thing
11:01 jameslenz joined
tbrowder_ tyil: hi! give me a few minutes and i’ll gist it. goona be here long? 11:02
tyil I might have to go shower and to the store inbetween, but after that I'll be back for ~ 10 hours I think 11:03
when the shower and store are depends on samcv :>
samcv wow
tbrowder_ er, i won’t ask, thnx 11:04
samcv hah. yeah we need to go to the store
tyil preferably before 5, so you can get the things you want
I think that store closes at 5
11:04 psychoslave left 11:05 jameslenz left
samcv is there anything equivilant to python's range(0, 100, 2) in perl6 11:13
11:13 markoong joined
samcv which would go from 0 to 99 inclusive but add 2 every time 11:13
equivilant to loop (my $i = 0; $i < 100; $i += 2) in perl6 essentially. but is there some more ideomatic way to wright it 11:14
timotimo the ... operator comes closest
tyil m: 0, 2, 4 ... 100
camelia Potential difficulties:
Useless use of ... in sink context
at <tmp>:1
------> 030, 2, 4 ...7⏏5 100
tyil m: dd 0, 2, 4 ... 100
camelia (0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, 72, 74, 76, 78, 80, 82, 84, 86, 88, 90, 92, 94, 96, 98, 100).Seq
tyil this?
samcv yeah that seems like it
tbrowder_ tyil: i solved my problem—misunderstanding of docs, plus fumble fingers
samcv though how can i do that programmatically
tyil tbrowder_: alright :D
timotimo samcv: you can put any numbers in there that you want, it's Just Lists™ 11:15
samcv gist.github.com/samcv/9a65a586ea21...le-time-p6 so this
tyil samcv: my $step = 2; dd 0, $step ... 100;
ilmari m: dd 0, { * + 2 } ... 100
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed double closure; WhateverCode is already a closure without curlies, so either remove the curlies or use valid parameter syntax instead of *
at <tmp>:1
------> 3dd 0, { * + 2 }7⏏5 ... 1…
ilmari m: dd 0, * + 2 ... 100
camelia (0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, 72, 74, 76, 78, 80, 82, 84, 86, 88, 90, 92, 94, 96, 98, 100).Seq
timotimo m: sub myrange($start, $end, $step) { $start, $start + $step ... $end }; .say for myrange(0, 10, 2); .say for myrange(10, 20, 3)
camelia 0
2
4
6
8
10
10
13
16
19
tyil I think timotimo's solution is the cleanest for what you want to get to 11:16
but take your pick from the options :>
tbrowder_ yep, docs need some love on unit sub MAIN, issue filed
Geth ¦ doc: tbrowder self-assigned sub MAIN needs better examples github.com/perl6/doc/issues/2167 11:17
tyil samcv: feel free to ask that question on SO as well (on the range stuff) 11:20
samcv well that takes 2 minutes :)
compared to 27 seconds if using loop()
timotimo yeah, loop is basically completely inlined
tyil samcv: a nice topic to optimize next? :p 11:21
samcv 5 seconds for nqp, 27 secs for perl 6 with loop(), 2 mins with 0, $interval ... $end -1
python takes 0.5 sec
tyil hmm
samcv tyil: yeah
11:22 wamba left
samcv for 0..$x has code in the optimizer but the other one doesn't 11:22
Geth Pod-To-HTML: c20d3388cd | (Zoffix Znet)++ | t/01-basic.t
Skip non-break test on older rakudos

  github.com/perl6/Pod-To-HTML/issues/39
timotimo for 0..$x also only optimizes if the $x is compile-time-known
Geth Pod-To-HTML: a0635fb657 | (Zoffix Znet)++ | META6.json
Bump version
timotimo otherwise for at least turns into a while with explicit .pull-one calls
samcv timotimo: ah
timotimo so it at least doesn't have to go through map
samcv it can't turn it into a while loop? 11:23
make a new lexical just in case $x was changed somewhere else and do the while loop i'd assume
timotimo if it's not compile-time-known, we can't be sure it's not funky 11:25
even if the variable has an Int type constraint, it could still have mixins or so
samcv m: <1 str thing>.perl.say
camelia (IntStr.new(1, "1"), "str", "thing")
samcv is this expected to make an IntStr?
11:25 eliasr joined
timotimo yes 11:26
<1> doesn't but < 1 > does
11:26 psychoslave joined
lizmat
.oO( significant whitespace, we haz it )
 11:26
tyil I thought every element in < > was stringified
but apparently only if there's space
lizmat well, if you have more than 1 element, there's a space 11:27
the special casing is only for single element <foo>
tyil I also generally write it as < some list of items >
tbrowder_ turns out, at least for my purposes, i JUST discovered unit SUB main, and all its dressing, is EXACTLY what i’ve been looking for since i first dipped into p6. in mid-2015 it looked useful, but i didn’t get much help with ironing out details so i couldn’t see how to use my p5 style of code writing and arg handlng without an extra scope and multi subs, and, imo for a p6 noob, the docs were confusing, so i gave up and 11:30
kept rolling my own.
p6 is AWESOME 11:31
tyil heh
MAIN is my favourite thing about Perl 6
it takes all the pain out of writing a cli application for me
El_Che MAIN became useful once lizmat++ enable random order or parameters like gnu (and most useful UNIX commands) 11:32
tbrowder_ amen, bruddah, i can now agree (plus kebab case, etc.)
El_Che how a little change tranformed a feature from meh to superb
tyil kebab case results in nicer syntax imo
tbrowder_ thnx to ++lizmat
and kebab case is so useful for non-touch typists! 11:33
tyil also for touch typists
El_Che docs.perl6.org/routine/MAIN#%*SUB-MAIN-OPTS <-- maybe %*SUB-MAIN-OPTS<named-anywhere> = 1 should be the default in 6.d?
tyil no need to hold shift
tbrowder_ yepper! 11:34
El_Che better link: docs.perl6.org/routine/MAIN#named-anywhere
tyil El_Che: it would be nice to have that as a default
11:35 robertle joined
tyil I asume -- is also interpreted as the end of all named opts, so calling script.pl --foo -- --bar would result in $foo = True, and --bar passed as a positional string 11:36
it does 11:37
I should add that to the docs for clarity 11:38
lizmat wouldn't mind seeing that as default for 6.d 11:40
11:42 rindolf joined 11:43 Ven` joined
AlexDaniel .seen zostay 11:48
yoleaux I saw zostay 30 Jun 2018 21:43Z in #perl6: <zostay> Yes, I was using it for the first time just this week and it just worked for what I needed.
AlexDaniel fwiw for some MAIN is still not good enough 11:50
github.com/rakudo/rakudo/pull/688
↑ that's just one example, I'd say
zostay: ping :) 11:53
zostay: I was trying to figure out why array-hash module tests are failing on HEAD, but I'm a bit confused 11:55
zostay: first of all, these two test sets seem to be identical? github.com/zostay/perl6-ArrayHash/....t#L33-L46 11:56
masak m: my $x; role R { method foo { say $x } }; class C does R {}; $x = "OH HAI"; C.new.foo
camelia OH HAI
masak strangelyconsistent.org/blog/lexpad...eed-fixups (in case anyone was wondering) 11:57
(as I re-read that post, I'm a bit surprised that I never use the word "frame", only "lexpad") 11:58
moritz does the fame contain the lexpad? 12:02
or is it the same?
lizmat moritz: I seem to recall things have changed since then 12:03
timotimo frames also have the locals
lizmat e.g., frames used to be refcounted :-)
masak moritz: yes, a frame has a lexpad 12:04
locals (as far as I understand them) are essentially an optimization where we don't have to do look up to outer frames because the variable is *right here* in ours
timotimo there's still a little difference between locals and registers 12:05
masak (and it makes sense to separate out that case for other reasons too, such as escape analysis)
timotimo every local has a register, but there's more registers than locals usually
though registers are allowed to be re-used between locals based on their lifetimes
moritz frames exists in the language (can be accessed by callframe() or whatever we call it), registers are purely a VM construct 12:06
masak hehe. I'm guessing "there's more registers than locals" comes from the fact that we're talking about a VM, not a physical M
timotimo right :)
masak frames are the noun, and "call this routine/block" is the verb :) 12:07
tyil AlexDaniel: on that issue, his `which perl6 -a` actually doesn't work like `which -a perl6` :p 12:11
but subcommands are technically already possible 12:12
by making a multi sub MAIN ("subcommand", *@args)
just replace the subcommand string to any other subcommands you want to have
AlexDaniel tyil: yeah, seems to be the wrong example 12:13
in fact makes me wonder which tools actually allow that 12:15
tyil ls for instance allows it
ls . -a is similar to ls -a .
but things like echo are even more of a pain to deal with, as echo -- -n "hi" will show "-- -n hi" 12:16
timotimo d'oh 12:17
tyil but echo -n -- "hi" shows "-- hi" on most shells
El_Che "Torture the implementers for the sake of the users.” kind of thing. Make the CLI work whatever users throw at it is my idea of a good cli
12:18 xtreak left
AlexDaniel well, echo is stupid :) 12:18
El_Che is stupid 12:19
pid
id
:)
tyil I personally dislike the use of echo for any use other that interactive and without any args
printf is much more sane, and also POSIX
AlexDaniel I mean, you actually have to do `printf "%s\n" "$foo"` if you want to print arbitrary stuff
because -- is not supported for whatever random reason
timotimo "echo injection" 12:21
12:26 itaipu joined
Altreus I heard that unix sockets aren't supported on moarvm and therefore there's no p6 interface to them. Is that still current? 12:31
timotimo you can use NativeCall for unix domain sockets as a work-around 12:32
Altreus has anyone written a thing to abstract this so I don't have to learn stuff? 12:33
Which is to say, I'm already learning P6 so I dunno if I can cope with learning sockets too
timotimo hm, you can open the socket and get a fileno for it, then you'll use it just like a regular pipe 12:34
a file that can't be rewound
12:34 sarna joined
Altreus rw? 12:36
timotimo no, seek
Altreus sorry, I meant would it be a rw pipe or would I have to open one for reading and another for writing?
guy at work is inventing problems :P
12:36 itaipu left
timotimo ah, no, sockets are actually duplex 12:36
Altreus I'll have to TIAS when I get around to trying it 12:37
timotimo on top of that, you don't have to open a domain socket as a stream socket, it can be a datagram socket, too
Altreus And there's the bounds of my knowledge :D
12:37 itaipu joined
Altreus datagram makes me think TCP 12:38
timotimo nah, tcp is stream, and udp is datagram 12:40
actually, i'm not so sure about unix domain sockets; do they have a "listen socket" with "accept" as well?
Altreus I'm just going to smile until someone else answers - :)
(this is how I remain employed) 12:41
timotimo yup, smile and nod
El_Che www.userfriendly.org/cartoons/smili...ingman.gif
12:42 wamba joined
Altreus looks like either Northernlion or Frontalot 12:43
tbrowder_ ref MAIN: is there any way to get the gnu way of handling an arg like “-m someval” and some-val is taken as m’s value? 12:44
lizmat tbrowder_: you mean, have one argument serve as the default for another ? 12:45
or do you mean: support named parameters
Altreus Presumably named parameters, except with single-letter short options as aliases 12:46
El_Che I don't get it either
timotimo there's a problem with that and multis; imagine one candidate takes a Str as the argument to :$m, and one candidate takes :$m as a bool, and a positional 12:47
it'd be ambiguous if -m foo is m => True, "foo" or m => "foo"
Altreus In P5, you have one option spec and a manual dispatch to decide what to do. P6 is nice because you can have a multi MAIN, but suffers therefore in that a single script has multiple option specs 12:48
tbrowder_ hm, i think Altreus is right, single letter dhort option. instead of “—long-name=value” one can use “-m value”
Altreus I would say that if you make a single script in which -m foo can mean different things, that's the developer's fault 12:49
furthermore
I discovered that it only supports -m=foo
or --m=moo
12:50 psychoslave left
Altreus basically: it doesn't do it GNU style 12:50
masak it's interesting to read back and realize that the entire promise of strangelyconsistent.org/blog/macros...ak-shaving is now known to be wrong
(that Q::Unquote is a Qnode)
Altreus you can do -moo=foo or --moo=foo and it means the same
tbrowder_ i can definitely live with = for assigning value, just curious
Altreus Whereas, GNU style, -moo is -m -o -o=foo
Given that I only ever want to write GNU-style scripts myself I don't think i'd ever use the named parameters version of MAIN anyway, which puts us back almost in the perl5 camp 12:51
tbrowder_ the single letter (not grouped) option would be nice though
12:52 wamba left 12:53 dj_goku joined, scimon joined 12:58 psychoslave joined 13:00 p6noob left 13:01 lizmat left, jameslenz joined 13:02 wamba joined 13:04 lizmat joined 13:06 jameslenz left 13:10 wamba left 13:13 raschipi joined 13:39 psychoslave left 13:40 warriors joined 13:44 bdmatatu joined
tyil .tell AlexDaniel you said Matrix gets much better performance when using a homeserver that is not matrix.org, do you know of a list of different public home servers so I can try that out? 14:00
yoleaux tyil: I'll pass your message to AlexDaniel.
14:00 psychoslave joined
ilmari tyil: www.hello-matrix.net/public_servers.php 14:02
tyil ilmari: thanks :>
I couldn't find anything on matrix.org about it
ilmari has finally got around to getting himself a bytemark VM and is planning to run his own homeserver
tyil: first ddg hit for "matrix public homeserver" 14:03
;)
raschipi Is it really a homeserver then?
tyil I searched for "matrix alternativ home server"
:(
ilmari raschipi: "homeserver" is where users live
and room aliases (but not rooms themselves, they are distributed) 14:04
raschipi I see, thanks ilmari.
tyil rooms have some relation to the homeserver, don't they
for instance, you have #matrix:matrix.org
ilmari tyil: that's just the alias. the room itself is distributed between all the servers that have users participating in it 14:07
tyil ah
ilmari the internal id also has the homeserver name in it, but that's just for namespacing
tyil I'm regging at matrixim.cc to see if that's a bit more responsive
matrix.perl.org when tho 14:10
14:17 molaf joined 14:21 tyil[m]1 joined, tyil[m]1 is now known as tyil[m]2
tyil hmm 14:22
failed to join room, CORS request rejected
when I try to join #perl6 from matrixim.cc
14:24 raschipi left 14:25 curan left 14:26 itaipu left, kerframil joined 14:27 itaipu joined 14:29 tyil[m]2 is now known as tyil[m4, itaipu left 14:33 itaipu joined
Geth doc: 4e62a0f976 | (Will "Coke" Coleda)++ | Makefile
The script that did this was removed when adding xt/examples*
 14:33
14:34 itaipu left 14:37 itaipu joined
warriors guys, this is a general question .. how do you deal with focus and attention issues, i mean for me for example, i keep jumping from learning one language to the other, i end up knowing a little about a lot 14:40
i am afraid, me now trying to spend time on perl6, is just my bad habbits, jumping to perl6 instead of ... learning more about one the languages i already know a bit 14:41
lizmat as long as you learn from what you spend time on, it's not a waste, is it ? 14:43
moritz warriors: I typically spend more time with a technology/language when I use it to solve a problem I have
I tried to learn golang for example, but didn't have a use case. Gave up after 1.5 days
14:44 itaipu left
warriors i see your point, but .. my use cases dont always fit the thing i want to learn, and i feel i have weird psychological barriers, my work in BI (business intelligence) 14:45
14:45 itaipu joined
warriors so it make more sense to me, to learn python, R maybe even julia 14:45
but , man, those R and Python, are so dry 14:46
14:46 itaipu left
warriors anyway .. i know the problem and i know the solution ... its always will power 14:47
moritz are you learning for work, or in your spare time?
warriors well, i try to do both
again fragmenting my time
moritz I only manage to learn stuff for work in my spare time if I'm really into the topic
scimon I learnt Golang. Spent the entire time going "Perl6 does this bit better" (string, arrays, threading...) 14:49
masak performance... :P
masak , troll
but yes, I think I agree on the others 14:50
14:50 itaipu joined
masak though with threading, I think Go at least has some neat ideas 14:50
scimon Ah yes. Though speed of development wasn't as fast.
masak largely I think Go looks like a very polished packaging of some fairly old ideas 14:51
scimon It does, but I think Perl6 has done them better. Fire and forget async functions where you have to wire in your own communications aren't as fully features.
masak I can see it being a step up from C++ for Google
warriors Go is GC lang, cant really compete with C++, where C++ make sense
El_Che I don't think perl6 does a lot of things better than Go
masak especially with things like modularity and build times
warriors Go is step between Python and C++ 14:52
El_Che it just does a lot of things go doesn't do
warriors Rust is the true C++ replacement
masak warriors: I do think a garbage-collected language can be constructively compared to C++
lizmat El_Che: which things does Perl 6 do not as well as Go ? Just OOC :-)
masak warriors: as far as I understand, a lot of the features that were prioritized in Go were pain points from C++
14:52 itaipu left 14:53 pukku joined
masak at least, I don't think that just because Go is a GC'd language, any comparison to C++ just goes out the window. 14:53
El_Che lizmat: eg the array/list methods
lizmat: you need to write a lot of loops to do simple things like contains 14:54
lizmat El_Che: specifically ?
El_Che one loop for each type
warriors sure not C++ have been used in so many domain, from application programming to systems programming
lizmat El_Che: example ?
El_Che lizmat: "is this element part of this array?"
write a for loop
warriors i am just saying, C++ really have an edge when you need to be GC free
El_Che let me see I I can show you code
warriors go wont replace C++ in those domains
lizmat m: my @a = ^100; say 42 (elem) @a 14:55
camelia True
lizmat m: my @a = ^100; say 142 (elem) @a
camelia False
El_Che lizmat: github.com/nxadm/ldifdiff/blob/mas...ff.go#L138
timotimo i think you got liz' question the wrong way around, El_Che
warriors i learnt a little C++ at some point :)
14:55 itaipu joined
El_Che lizmat: the other way around. Perl 6 has built-in methods, Go has 1 for loop 14:55
warriors never touched Go though
lizmat "I don't think perl6 does a lot of things better than Go" # El_Che 14:56
El_Che yes
lizmat so you're saying Go is better or not ?
El_Che quote also the second line
no
perl 6 is not better for the same features
but it has more features
lizmat ok, then maybe what do you mean by "better" ?
warriors for me, Go doesnt offer any novelty .. Go is .. the Python of compiled languages
timotimo a friend told me that go is more pythonic than python is, iirc 14:57
El_Che lizmat: I mean go is nog lacking is the quality of it's implementation or regular features, like concurrency
lizmat: but it misses a lot of fatures that you have to implement yourself 14:58
like that small loop for trivial things
lizmat I see, I think :-)
El_Che it's the mini-maxi discussion :)
not better-worse
my brain can handle both :)
14:59 itaipu left
pmurias warriors: building something using a language is a good way to learn more about it 14:59
timotimo building a compiler for the language is also a good way to learn more about it :P 15:00
15:00 itaipu joined, epony left, fake_space_whale joined
Geth doc: 52194a8e97 | (Tom Browder)++ (committed using GitHub Web editor) | Makefile
add Gnu standard distclean target
 15:01
15:01 jameslenz joined
El_Che lizmat: Cro is moving perl6 into golang's space by making it easy to create webservices. 15:02
pmurias timotimo: you do learn a lot of silly details that way 15:03
warriors pmurias, i completely agree, i think the way i always learned was fundamentally problemetic ... i spend too much learning, without doing ... i need to learn to use a language before i know all its quirks 15:04
scimon (Sorry if I put the cat among the pigeons there. Go being like Python makes sense to me. They are both dull ;) )
El_Che lizmat: I did some POC code with cro for a rest service I needed to write, but I end up writing it in go because I needed oauth2 support (web and db looked ok)
15:05 jameslenz left
warriors Go is a very safe bet, it is popular, simple have good ide, tons of code on github 15:05
El_Che scimon: depends what your workspace is. If you're thinking SRE, devops and cloud, golang is pretty much the nr 1 option (and python non-existent)
scimon While failing to write slides I have started on a new project : github.com/Scimon/p6-Range-SetOps Updates the Set Operations to work on Ranges as continous Set (so 3 (elem) (1.5..9.7) would work) 15:06
scovit I think nothing around will replace C++
jnthn Nor C :)
warriors rust will, eventually
15:06 gabiruh left
warriors rust i a must learn in my opinion, it is really the future, super solid community and interesting language 15:06
15:06 gabiruh joined
moritz maybe a language a generation later that picks up ideas from rust, but makes the concepts easier accessible 15:07
El_Che golang was supposed to replace C/C++, but ended killing big parts of the ecosystems held by python and ruby (they moved, C/C++ people didn't)
15:07 epony joined
El_Che jnthn: every time more impressed by Cro. Great job! 15:07
warriors i also think the c/c++ combination is unfair to C, i think C is a lot harder to replace than C++ ... this is mainly because C is more niche compared to C++ which is more general 15:08
El_Che warriors: yet C++ seems to be very entranched in some niches 15:09
I think more people will move from C to Rust, than from C++ to Rust
(my very subjective prediction)
jnthn :)
15:10 domidumont left
warriors maybe :) ... but regardless i think rust is very worthy of learning 15:10
moritz replacing C only really helps you if you also manage to replace the calling conventions that C brought about
and the lack of type information in the ABI
15:10 Khisanth left
El_Che moritz: one ABI compatible piece of Rust code at the time, they say :) 15:10
warriors C is like some people describe it, the lowest common denominator 15:11
scovit with C you almost know exactly what the code run by the processor will look like 15:12
hobbs because everything else comes with some baggage that makes it harder to interoperate with than C
except Fortran :)
scovit C++ adds the possibility to abstract algorithms over complex data structures, if needed 15:13
15:13 robertle left
moritz scovit: except when you manage to write undefined behavior and have a clever compiler 15:13
15:14 diakopter joined, itaipu left, troys joined 15:15 itaipu joined, salasrod joined
scovit moritz at least it will be easy to debug 15:16
Ulti scovit: the one exception is C doesn't really have nice abstractions that fit caching and branch prediction in modern CPUs which are sort of outside of what code is emitted 15:19
though are there any languages that kind of make those problems obvious? 15:20
scovit Ulti: this is right, and could be that the future will be heterogeneous computer architectures that can only be optimally programmed by much more sophisticated runtimes. But is it going to ever happen? 15:22
my impression is that it could be that there will always be a spot for something with minimal abstraction (that requires lot of work to program with) 15:23
15:30 Khisanth joined 15:31 lizmat left 15:33 lizmat joined
b2gills warriors: If you learn *all* of Perl 6's quirks, you would probably be the first. (backlogging) 15:38
15:38 robertle joined
SmokeMachine m: sub bla { KEEP { .say }; 42 }; bla 15:41
camelia (Any)
SmokeMachine shouldn't it print 42?
b2gills m: sub bla { $_ = 255; KEEP { .say }; 42 }; bla 15:44
camelia 255
15:44 psychoslave left
b2gills m: sub bla { $_ = 255; KEEP { .say }; 42 }; say bla 15:44
camelia 255
42
SmokeMachine docs.perl6.org/language/phasers#KEEP 15:45
15:45 psychoslave joined, p6noob joined 15:49 psychoslave left 15:59 psychoslave joined
vrurg Empirical observation: <?after <rule>> doesn't work despite the fact is not mentioned in the docs. Am I right? Similarly, <?before <rule>> works. Example: f.perl.bot/p/mbmiu4 16:02
mst please don't send #perl6 pastes to #perl, it's confusing 16:03
vrurg mst: f.perl.bot supports rakudo. What's wrong with it? 16:04
timotimo the bot also puts the link in the channel you select 16:05
and #perl denizens aren't happy about getting perl6 unexpectedly
mst vrurg: yes? because perlbot is in here. so you select #perl6 to send the paste to.
vrurg: it would also be rude if we sent random perl5 pastes to here.
hobbs vrurg: you just want to set the "where" to Freenode #perl6 if it's relevant to here, not to #perl :) 16:06
vrurg Sorry guys, missed the 'where" field.
hobbs np 16:07
vrurg Any way, is there a workaround for <?after>? I can use same regex in a token and a lookbehind – but rule would me logically more the right way to do it. 16:08
tyil[m4 o/ 16:09
b2gills vrurg: I think you may just have to remove the `?`
zostay AlexDaniel: They are identical. I clearly intended for them to be different, but I'm not able to recall at the moment what my intention was. 16:10
AlexDaniel: I've been intending to check on those failing tests myself, but I haven't made time for it yet.
vrurg b2gills: Nah, makes no difference. Basically, <?after and <after should be the same. 16:11
b2gills Exactly why I thought that removing the `?` might work. (Just because they should be the same doesn't mean they work the same) 16:12
16:13 sjn left
vrurg b2gills: It seems that in this particular case 'should be' ~~ 'are' ;) 16:13
16:16 psychoslave left
b2gills vrurg: Actually maybe it is the `&&` as I don't see a reason for them being there 16:16
m: say 'abcd' ~~ m:g/ <alpha> && c / 16:17
camelia (｢c｣
alpha => ｢c｣)
vrurg b2gills: Ah, sorry. && shouldn't be there. I was just playing around and forgot to remove it. Though it makes no difference. 16:18
b2gills vrurg: I'm having a little difficulty understanding why it isn't just `token paragraph { <line>+ }` 16:19
16:20 sjn joined
vrurg b2gills: Strict approach. Paragraph is what comes after an empty line of the beginning of text. Only. 16:20
Basically, it's more about studying grammars for now. So, mostly experimenting. 16:21
b2gills vrurg: Well I am absolutely sure that everything after `<line>*` can just be removed as token is ratcheting. (doesn't backtrack) 16:22
Actually I wonder if the reason `<after>` isn't working is because it is a `token` and not a `regex` 16:23
16:23 Ven` left
vrurg Basically, for this demo <?before> isn't needed, you're right. But I extracted it from a bit more complex grammar where paragraph definition is slightly different. 16:24
16:25 epony left
vrurg I tried a regex too. Fails the same way. It only works with <?after [^^ \h* \n]> 16:26
b2gills token doc { ^$ | <emptyLine> ␤ <paragraph>+ %% <emptyLine> } 16:27
vrurg b2gills: %%.. Hm, nice solution! Thanks! 16:29
16:29 fake_space_whale left
b2gills vrurg: You may have stumbled onto a bug, as the people who understand grammars rarely use `<after>` 16:31
16:31 scimon left
b2gills Basically when I see a `<after>` that means to me that the grammar may need to be restructured. 16:31
vrurg Though still might fail in more complex setup. Basically, I'm training myself of markdown-like format. A paragraph might be followed by a list/quote/whatever. But then instead of <paragraph> I might use a compound rule... Just thinking aloud. ;) 16:32
16:32 pukku left
vrurg suspected that there must be a simpler solution. 16:32
b2gills vrurg: You will want to look into proto tokens. For an example see github.com/moritz/json/blob/master...Grammar.pm 16:33
vrurg: Perl 6 has a lot of really useful features. The biggest problem is finding out that they exist. 16:34
vrurg I already use a proto in the bigger one. Though don't see how it applies to this example. 16:35
b2gills It's one of the features that gets often overlooked. So just making sure you knew about it in case it proved useful 16:36
vrurg Aha, I agree. Wonder how to manage to fit both perl5 and perl6 together in the limited space of my skull. :) Though perl6 might overtake some areas of my work already.
b2gills I remember one person who wrote a bunch of code to make sure that a Hash doesn't overwrite other values that hash to the same value. At which point I told him that Perl 6 already does that. 16:37
vrurg: I've forgotten how to do symbolic references in Perl 5. 16:38
vrurg protos and multi are definitely not something /me would miss because it's the only feature I was missing since departing from C++ almost two decades ago. ;)
Ulti < b2gills> warriors: If you learn *all* of Perl􏿽xA06's quirks, you would probably be the first. (backlogging) <--- you'd also be a human that can pass the test suite and thus an actual implementation of Perl 6 16:41
16:42 epony joined, dakkar left 16:45 kerframil left
mst vrurg: btw, f.perl.bot/#perl6 as a bookmark should I think send it to the right place to start with 16:45
b2gills: unless you're using that to mean something other than what it usually means ... good. you shouldn't be doing those in the first place :D 16:46
vrurg mst: thanks.
AlexDaniel tyil: I don't know how bridging with IRC works, but I guess it doesn't matter that much if it's #freenode_#perl6:matrix.org or #freenode_#perl6:perl.org 16:49
yoleaux 14:00Z <tyil> AlexDaniel: you said Matrix gets much better performance when using a homeserver that is not matrix.org, do you know of a list of different public home servers so I can try that out?
AlexDaniel tyil: but maybe I misunderstood your question 16:50
b2gills mst: I've also forgotten how to do it without using symbolic refs (%::MAIN or whatever it is)
Geth ecosystem: ec6ee28869 | (Zoffix Znet)++ (committed using GitHub Web editor) | META.list
Add App::Lang::French::VerbTrainer to ecosystem

  "Training program for learning tenses and conjugations of French verbs":
  github.com/zoffixznet/perl6-App-La...erbTrainer		 16:52
vrurg b2gills: Tracing down stashes in Perl5 is really a headache. Must easier and faster with the refs. 16:55
vrurg wrote a special sub for fetching symbols from a namespace... 16:56
17:02 jameslenz joined 17:06 jameslenz left 17:14 sarna left 17:22 koto joined 17:24 koto left, sena_kun left 17:25 domidumont joined 17:27 diakopter left 17:29 bwisti joined
SmokeMachine m: sub bla { KEEP { .say }; 42 }; bla # shouldnt it print 42? 17:30
camelia (Any)
timotimo m: sub bla { KEEP .say; 42 }; bla 17:31
camelia (Any)
timotimo i think it should
m: sub bla { POST { .say }; 42 }; bla
camelia 42
SmokeMachine is that a bug? 17:32
timotimo possibly an NYI? 17:33
17:33 psychoslave joined
SmokeMachine I thought I already used that... I think I'm wrong... 17:34
bisectable6: sub bla { KEEP { .say }; 42 }; bla
bisectable6 SmokeMachine, On both starting points (old=2015.12 new=88e913b) the exit code is 0 and the output is identical as well
SmokeMachine, Output on both points: «(Any)␤»
17:37 pecastro joined
SmokeMachine m: sub bla { LEAVE { .say }; 42 }; bla 17:38
camelia (Any)
17:40 psychoslave left 17:42 domidumont left 17:52 molaf left 17:56 itaipu left, itaipu joined 17:58 pecastro left 18:05 itaipu left 18:07 pecastro joined 18:16 w17t joined, w17t left 18:19 w17t joined, w17t left 18:22 wamba joined, itaipu joined 18:23 Schepeers left 18:24 sauvin left
TimToady in general, LEAVE-style phasers are there for the purpose of maintaining invariants that are not related to the return value, so there's little point in binding them to $_ (and in so doing shadowing any $_ in the outer scope) 18:36
we have POST if you want to do the other thing
18:37 gabiruh left, psychoslave joined
TimToady though arguably KEEP/UNDO are related to the return value insofar as they select one of two sets of invariants based on the return value 18:37
18:38 gabiruh joined
TimToady but LEAVE/KEEP/UNDO have to run even if there is no return value (such as when unrolling the stack for an exception) 18:38
18:38 konsolebox left 18:41 vrurg left, konsolebox joined
tyil AlexDaniel: I've asked a freenode oper more on bridging, and asked in the matrix channel about it as well 18:41
it seems freenode irc bridges will most always be slow on matrix, since they mostly still use the matrix.org bridge
which is overloaded 18:42
it is possible to setup a home server with our own bridge to freenode
which would then probably not be as overloaded, and thus faster
SmokeMachine TimToady: So the docs are what's wrong?
tyil however, we would have to contact the iline team about it as well, so we not get banned
18:43 molaf joined
tyil the bridging part itself seems to be part of matrix appservices 18:43
TimToady SmokeMachine: well, perhaps the docs copied the speculations, which still say: "For phasers such as C<KEEP> and C<POST> that are run when exiting a scope normally, the return value (if any) from that scope is available as the current topic within the phaser." 18:47
but rakudo has never done that, and I think the fact that nobody has noticed till now is indicative of how useful a feature that would be :) 18:48
18:49 pmurias left
TimToady also working against this is the fact that in a "will leave { .close }" the topic is actually whatever the object is that the trait was applied to, not the return value 18:49
(I think) 18:50
m: { my $x will leave { .say } = 42; }
camelia 42
TimToady yep 18:51
SmokeMachine makes sense...
TimToady so probably, if we want to make the return value (if any) available to exit phasers, we should put the onus on that code to dig it out, not just blindly bind to $_ 18:52
dunno if there's an implementation-independent way to dig it out currently...
TimToady guesses not
unless callframe(0).retval or some such works 18:53
SmokeMachine m: sub bla { $_ = 42; KEEP { .say } }; bla
camelia ( no output )
TimToady you returned Nil
18:53 Ven`` joined
SmokeMachine m: sub bla { return $_ = 42; KEEP { .say } }; bla 18:53
camelia 42
SmokeMachine :)
m: sub bla { return $_ = 42; KEEP { .say } }; say bla 18:54
camelia 42
42
TimToady and since we might have code that currently depends on not shadowing $_, we probably won't change it
19:00 Zoffix joined
SmokeMachine m: sub bla { POST { .say }; 42 }; bla 19:00
camelia 42
SmokeMachine m: sub bla { POST { .say }; throw "bla" }; bla
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
throw used at line 1. Did you mean 'THROW'?
SmokeMachine m: sub bla { POST { .say }; die "bla" }; bla 19:01
camelia (Mu)
bla
in sub bla at <tmp> line 1
in block <unit> at <tmp> line 1
Zoffix It was noticed before (because the docs claimed it should work): R#1290 and looks like there was a trial implementation that did indeed break a few modules
synopsebot R#1290 [open]: github.com/rakudo/rakudo/issues/1290 KEEP does not set $_ to the return value
TimToady Zoffix++ 19:02
19:02 jameslenz joined
SmokeMachine Zoffix++ 19:02
m: sub bla { POST { .?say }; die "bla" }; bla
camelia (Mu)
bla
in sub bla at <tmp> line 1
in block <unit> at <tmp> line 1
TimToady but I think the "will leave" argument is pretty powerful here
19:03 stephane_ left
[Coke] travis-ci.org/perl6/doc/jobs/402694645 - do we need to update some zef installs somewhere? 19:03
TimToady we should, however, make some way of getting at the return value for the highly motivated :)
SmokeMachine: the .say is succeeding, or the POST would fail 19:05
m: sub bla { POST { .say; False }; die "bla" }; bla
camelia (Mu)
Postcondition '{ .say; False }' failed
in sub bla at <tmp> line 1
in block <unit> at <tmp> line 1
19:06 jameslenz left
SmokeMachine TimToady: thanks 19:07
Zoffix SmokeMachine: where in the docs does it say that it sets $_ to return value? 19:08
19:09 dogbert17 left
SmokeMachine Zoffix: docs.perl6.org/language/phasers#KEEP 19:09
geekosaur "For phasers such as KEEP and POST that are run when exiting a scope normally, the return value (if any) from that scope is available as the current topic within the phaser."
at that link
Zoffix [Coke]: the only avaiable version is 0.3.18, not 0.3.17. Previous versions aren't available for p6c ecosystem ATM, so it can't find it. The version prereq should be removed from docs prereqs IMO 19:10
Thanks.
Geth doc: d6d75b0397 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/phasers.pod6
Remove speculative $_ setting

It's not part of the spec (just a speculative fudge exists) and might end up not being specced the way it's described:
  github.com/rakudo/rakudo/issues/12...-404277299
  colabti.org/irclogger/irclogger_log...07-11#l776		 19:12
synopsebot Link: doc.perl6.org/language/phasers
doc: 225b7fc212 | (Zoffix Znet)++ (committed using GitHub Web editor) | META6.json
Allow higher versions of HTML::To::Pod in prereqs
 19:13
19:15 psychoslave left 19:17 lizmat left
[Coke] Zoffix++ 19:17
Zoffix Thinking of making "Your™" that does the very minimum to comply to the Perl 6 Specification. Other than having a laugh at how useless it will be, despite passing every test, it's main purpose will be to test our entire chain to ensure we fully support multiple compilers, so that when someone wants to join the ranks, they can easily do it. The places where "Your™ Compiler"'s description would appear would 19:21
be a plug to encourage people to write more Perl 6 compilers.
19:24 lizmat joined
warriors hi in the repl "a".say.WHAT returns a , i was extpecting (Bool) which is the return type of .say 19:24
what am i missing
Zoffix + info on internals courses, etc, if people want to use the NQP toolchain 19:25
warriors: the autoprinting of the last value is done only if you did not produce any output
warriors: and since you do, it assumes you don't want it to print bool
m: "a".say.^name.asy
camelia a
No such method 'asy' for invocant of type 'Str'. Did you mean 'any'?
in block <unit> at <tmp> line 1
Zoffix m: "a".say.^name.say
camelia a
Bool
warriors i see, so its a repl thing
i think ocaml repl does it better, it prints the return value, then what went to STDOUT :) 19:26
Zoffix (.WHAT isn't a good thing to find what a thing is. It's for getting the actual typeobjects, which don't stringify well and .gistify only to a shortname. It's also a fake method so you can't use it in some places. The .^name metamethod is better and returns a Str object)
m: say IO::Handle.WHAT
camelia IO::Handle is disallowed in restricted setting
in sub restricted at src/RESTRICTED.setting line 1
in method gist at src/RESTRICTED.setting line 33
in block <unit> at <tmp> line 1
Zoffix e: say IO::Handle.WHAT 19:27
evalable6 (Handle)
Zoffix e: say IO::Handle.^name
evalable6 IO::Handle
Zoffix e: say "The object is " ~ IO::Handle.WHAT
evalable6 Use of uninitialized value of type IO::Handle in string context.
Methods .^name, .perl, .gist, or .say can be used to stringify it to something meaningful.
The object is
in block <unit> at /tmp/kazvK_t0De line 1
Zoffix e: say "The object is " ~ IO::Handle.^name
evalable6 The object is IO::Handle
Zoffix e: say "The objects are " ~ ($*OUT, $*ERR, $*IN).map: *.WHAT
evalable6 (exit code 1) Cannot resolve caller map(List: Whatever); none of these signatures match:
($: Hash \h, *%_)
(\SELF: &block;; :$label, :$item, *%_)
in block <unit> at /tmp/M63_hQP7oF line 1
Zoffix e: say "The objects are " ~ ($*OUT, $*ERR, $*IN).map: *.^name 19:28
evalable6 The objects are IO::Handle IO::Handle IO::Handle
geekosaur warriors, the REPL is as yet something of a hack
warriors :)
Zoffix m: sub new-thing-maker { $^v.WHAT.new }; my $o := class Meows {}.new; say [$o.WHICH, $o.&new-thing-maker.WHICH] 19:29
camelia [Meows|67665248 Meows|67665280]
warriors the REPL needs autocompleton then it will be berfect
:)
Zoffix I think CommaIDE has autocompletion now. Use that instead :D 19:30
hobbs warriors, come out to play?
19:30 cognominal joined
warriors it cost almost 200 canadian 19:31
i will wait for free comma
Zoffix paid the 200 canadian...
warriors :(
Zoffix warriors: where in Canada are you at? 19:32
warriors montreal
Zoffix C'est á Brampton :)
warriors qui?
Zoffix *à
Damn it :)
warriors c'est quoi ca qui est a Brampton
Zoffix Not c'est. Je suis :) 19:33
warriors je suis un immigrant, ca ne fait pas longtemp que je suis ici
:)
a toi, tu es a Brampton
enchante
:)
Zoffix I've been to Montreal like a month ago and started learning French 'cause I had trouble getting coffee there :)
warriors montreal is very bi langue 19:34
Zoffix enchante aussi
warriors u can get by, with english only
sepecially in down town
Zoffix Yeah, but it just feels like a dick move to speak in English when everything is in French :)
warriors well ... i didnt at first, it took me a while getting used to speak in french 19:35
:) .. french is cool though
Zoffix :)
warriors and its enchanté, i forget i have a french keybaord :) 19:36
Zoffix :)
[Coke] (perl 6 compiles) - btw, did anyone see that someone make niecza compile with a recent mono? 19:39
tbrowder_ an aside: anyone using a windows host (running win 10 pro) and using hyperV for a Linux VM? 19:44
Zoffix tbrowder_: I do, except using VirtualBox for Linux VM
tbrowder_ i hear hyperV should be faster since its a level 1 vs level 2 for virtual box, thoughts? 19:46
Zoffix vOv... VirtualBox is fast enough for me 19:47
moritz tbrowder_: I don't think these theories applies anymore
tbrowder_ i’m condidering getting a new laptop and, for the first time in a long time, not using duel boot 19:48
dual
has anyone tried hyperV?
19:53 itaipu left, itaipu joined 19:55 itaipu left, Zoffix left
tbrowder_ well, i have used vbox a lot, so i guess i could start with that... 19:56
19:57 benjikun left 19:58 itaipu joined 20:06 itaipu left 20:07 itaipu joined 20:09 cognominal left
Kaiepi m: BEGIN $*LANG.refine_slang: 'MAIN', role Lolcode::Grammar { token scope_declarator:sym('I HAZ A') { <sym> <scoped('my')> }; token initializer:sym<ITZ> { <sym> <.ws> <EXPR('e=')> } }, role Lolcode::Actions { method scope_declarator:sym('I HAZ A')($/) { make $<scoped>.ast }; method initializer:sym<ITZ>($/) { make $<EXPR>.ast } }; I HAZ A $CHEEZBURGER ITZ 'EPIC'; say $CHEEZBURGER 20:16
camelia ===SORRY!===
Type check failed in binding to parameter '$/'; expected Any but got Perl6::Grammar+{Lolcode::Grammar} (Perl6::Grammar+{Lolcode::Grammar}.new() #`[14001853426...)
Kaiepi is there something i'm doing wrong here?
apart from reviving dead memes
20:16 cognominal joined 20:18 p6noob left 20:25 Zoffix joined
Zoffix Kaiepi: based on the error, you're attempting to stick a non-Any object into Any param 20:25
stick a Mu type constraint on it 20:26
tobs Kaiepi: the $/ parameter is missing a Mu and you can't $<foo> somehow. See hastebin.com/mowepuyuko.rb
I don't know why but I saw that in the sources of Slang::SQL
TimToady the Any constraint is coming from nqp's .refine_slang
Kaiepi ah perfect, that works 20:28
thanks
20:29 psychoslave joined
TimToady ah yes, ignore what I said 20:30
the mention of $/ should'a clued me
20:31 warriors left
TimToady not to mention I probably need a geezer nap... 20:32
20:34 psychoslave left
Kaiepi what does lk stand for? 20:36
tobs Kaiepi: sLang cargo Kulting, maybe? :) github.com/tony-o/perl6-slang-sql/...QL.pm6#L41 20:37
or "lookup key"
20:40 MasterDuke joined 20:42 benjikun joined
Kaiepi i'm gonna make a port of c's version of lolcode to get to learn how Perl6::Grammar and Perl6::Actions work 20:44
SmokeMachine m: multi bla() { say "not" }; multi bla(|c (:$bla!)) { say c }; bla :42bla
camelia Unexpected named argument 'bla' passed
in sub bla at <tmp> line 1
in block <unit> at <tmp> line 1
SmokeMachine m: multi bla(|c (:$bla!)) { say c }; bla :42bla
camelia \(:bla(42))
SmokeMachine m: class :: {multi method a() { say "not" }; multi method a(|c (:$bla!)) { say c } }.a: :42bla 20:46
camelia not
SmokeMachine are those bugs? 20:47
TimToady for subs, prolly, but methods have the *% thing going as well 20:49
timotimo DrForr: the tensorflow repository has your perl6::parser's readme (but you haven't touched it in a year anyway, so maybe not useful any more? 20:50
Zoffix SmokeMachine: no, named args do not participate in dispatch
SmokeMachine: only for tie breaking
20:51 p6noob joined
Zoffix m: multi bla(| ()) { say "not" }; multi bla(|c (:$bla!)) { say c }; bla :42bla 20:51
camelia \(:bla(42))
Zoffix Here now both candidates are tied, so it used the named args to tie break
SmokeMachine m: class :: {multi method a(|()) { say "not" }; multi method a(|c (:$bla!)) { say c } }.a: :42bla 20:54
camelia 5===SORRY!5=== Error while compiling <tmp>
Obsolete use of | or \ with sigil on param ()
at <tmp>:1
------> 3class :: {multi method a(|()7⏏5) { say "not" }; multi method a(|c (:$bl
SmokeMachine m: class :: {multi method a(| ()) { say "not" }; multi method a(|c (:$bla!)) { say c } }.a: :42bla
camelia \(:bla(42))
SmokeMachine Zoffix: thanks!
Kaiepi m: BEGIN $*LANG.refine_slang: 'MAIN', role Lolcode::Grammar { token multi_declarator:sym<LOTZA> { :my $*LINE-NO := HLL::Compiler.lineof(self.orig(), self.from(), :cache(1)); <sym><.kok> :my $*MULTINESS := 'multi'; [ <?before '('> { $*W.throw: $/, 'X::Anon::Multi', multiness => $*MULTINESS } ] [ <declarator> || <routine_def('sub')> || <.malformed('multi')> ] } }, role Lolcode::Actions { sub lk(Mu \h, \k) { use nqp; nqp::atkey(nqp::findmethod(h, 'hash')(h), 20:55
k) }; method multi_declarator:sym<LOTZA>(Mu $/) { my \declarator = lk($/, 'declarator'); $/.make: declarator ?? declarator.ast !! lk($/, 'routine_def').ast; } }; LOTZA sub HAI { say 'HAI' }; HAI
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in argument list; couldn't find final ')' (corresponding starter was at line 1)
at <tmp>:1
------> 3qp::atkey(nqp::findmethod(h, 'hash')(h),7⏏5<EOL>
expecting …
Kaiepi whoops too long
when i try running that i get Undeclared name: LOTZA used at line 49
Zoffix Kaiepi: what if you stick that code into its own block? 20:56
Some things are so fundamental any slang changes get picked up only in a new scope after slangification
Kaiepi like this? hastebin.com/zekohohuda.rb 20:57
Zoffix Kaiepi: no, that just added a block into the begin. Not a block around the code that's meant to pick up slanged changes 20:58
Kaiepi what do you mean then?
Zoffix nothing, never mind 20:59
Kaiepi there's a line in src/Perl6/Grammar.nqp that's like token term:sym<multi_declarator> { <?before 'multi'|'proto'|'only'> <multi_declarator> } 21:00
maybe that's what's causing the problem?
21:00 bdmatatu left
Zoffix yeah 21:00
Kaiepi lemme compile and test again after removing the <?before> token 21:01
21:01 pmurias joined
Zoffix Just stick token term:sym<multi_declarator> { <?before 'multi'|'proto'|'only'|'LOTZA'> <multi_declarator> } into your grammar 21:01
21:02 jameslenz joined
Zoffix hm 21:03
it comes back after you fix the HLL::Compiler error
oh works 21:05
Ah, you missed the `?` on the `<before` thin in LOTZA
Kaiepi ohh
Zoffix m: gist.github.com/zoffixznet/b5197de...e482bbe45d 21:06
camelia HELLA FRICKIN EPIC
HALLO
21:06 jameslenz left 21:13 Zoffix left 21:22 molaf left 21:30 itaipu left, itaipu joined 21:32 itaipu left
benjikun What are the ways to use map asynchronously again 21:33
timotimo are you refering to asynchronous as in supplies, or do you want to map over a list using multiple cores at the same time? 21:34
21:35 itaipu joined
timotimo supplies have a map method, and for mapping over a list in parallel you can use race or hyper (hyper will keep the same order, race won't necessarily) 21:35
don't forget to tune :batch() and maybe :degree() to fit your workload better
Kaiepi what's the difference between *@args and **@args in parameters? 21:36
El_Che a pointer to a pointer? 21:37
El_Che ducks
21:37 itaipu left
timotimo Kaiepi: changes how nested structures are handled 21:38
Kaiepi ah
yeah i found the documentation for it
The single asterisk form flattens passed arguments.
The double asterisk form does not flatten arguments.
The plus form flattens according to the single argument rule.
timotimo m: sub onestar(*@args) { say @args.perl }; onestar(1, 2, (3, 4), (5, (6, 7), 8))
camelia [1, 2, 3, 4, 5, 6, 7, 8]
timotimo m: sub twostar(**@args) { say @args.perl }; twostar(1, 2, (3, 4), (5, (6, 7), 8))
camelia [1, 2, (3, 4), (5, (6, 7), 8)]
tyil[m] benjikun: are you looking for race/hyper?
benjikun timotimo: Map over a list 21:39
tyil[m]: Yeah, I'm aware of race/hyper, how would I use that with .map though?
timotimo you just @mylist.hyper.map({ })
tyil[m] benjikun: @list.race.map();
or hyper 21:40
yeah
21:40 itaipu joined 21:41 MasterDuke left
benjikun ah I see, didn't realize it would be that simple 21:41
tyil[m] benjikun: have you seen 6guts.wordpress.com/2017/03/16/con...semantics/ 21:42
21:42 itaipu left 21:44 MasterDuke joined 21:45 itaipu joined 21:47 itaipu left 21:48 Ven`` left 21:50 itaipu joined 21:52 itaipu left 21:55 itaipu joined 21:56 itaipu left 21:58 Ven`` joined 21:59 Ven`` left 22:00 itaipu joined 22:10 devmikey joined, Schepeers joined, itaipu left 22:11 itaipu joined 22:12 pmurias left 22:13 itaipu left 22:14 pmurias joined 22:16 itaipu joined 22:18 warriors joined
warriors hi 22:18
i noticed that you must include a space after a class name else, errors so for example class point{ ...} causes error but class point {...} is fine 22:19
timotimo greetings 22:20
yeah, { } after something is also syntax for hash access
warriors is this documented somewhere
timotimo m: sub class { }; my \point = %(a => 1); class point{"a"}.say
camelia ===SORRY!===
No compile-time value for point
timotimo oops :)
warriors hmmm 22:21
22:22 pmurias left 22:24 itaipu_ joined
warriors the error message was so criptic 22:24
is kiwi irc a safe client 22:26
22:26 itaipu left
warriors i mean the web client kiwiirc 22:27
timotimo i'd expect it to be reasonably good. why? 22:29
warriors just asking
timotimo if you're using freenode.net's own kiwiirc, it shouldn't be a problem, since your chat messages go through their network anyway 22:30
22:35 warriors left 22:36 benjikun2 joined, spycrab0 left 22:39 benjikun left 22:40 itaipu_ left 22:42 Zoffix joined 22:43 Zoffix left, h44gsyefw6fw joined, h44gsyefw6fw left, wamba left 22:44 sgsdfgsdfg joined, pecastro left 22:52 DarthGandalf left 22:53 sgsdfgsdfg left 22:57 vrurg joined 22:59 DarthGandalf joined 23:00 Khisanth left 23:02 jameslenz joined 23:04 Kaiepi left 23:05 Kaiepi joined 23:07 jameslenz left
Geth doc: f7c37c213a | 陈梓立++ (committed using GitHub Web editor) | README.zh.md
sync
 23:10
23:13 Khisanth joined
Geth doc: ace72dfa08 | 陈梓立++ (committed using GitHub Web editor) | README.zh.md
Update README.zh.md

though jp version is WIP, but `README.md` show a link to it. thus sync also add link to Dutch ver.		 23:15
23:29 rindolf left 23:30 kerframil joined 23:37 devmikey left
benjikun2 If I have some array @a where I'm mapping each to a start block asynchronously, how do I await on each of the items in the array to finish? 23:49
timotimo you can either just await @array-of-promises, or you can await Promise.allof(@array-of-promises) 23:51
the first version will throw on any broken promises
the one for allof is only for synchronization
if you use allof, you'll potentially silently lose exceptions 23:52
jnthn Yeah, the list form of `await` means you an just do my @results = await @promises; 23:54
If that's what you're wanting :)
If you're doing that, though, you often want .race or .hyper instead :) 23:55
timotimo changes the documentation
"If a list of promises is provided a list of promises is returned.
because that's wrong
jnthn indeed 23:56
timotimo oh, it doesn't say anything about a list of mixed things, like two promises and two channels 23:57

Powered by Cro and the Raku® Programming Language.

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

Thank you!