»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 moritz on 22 December 2015.
grondilu m: use NativeCall; role Foo[Sub &native] { has &.native }; role Bar is repr('CPointer') {}; 00:00
camelia ( no output )
grondilu m: use NativeCall; role Foo[Sub &native] { has &.native }; role Bar is repr('CPointer') {}; class :: does Foo(sub {}) {}
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in typename; couldn't find final ')' ␤at <tmp>:1␤------> 3 repr('CPointer') {}; class :: does Foo(7⏏5sub {}) {}␤»
grondilu m: use NativeCall; role Foo[Sub &native] { has &.native }; role Bar is repr('CPointer') {}; class :: does Foo(my & = sub {}) {}
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in typename; couldn't find final ')' ␤at <tmp>:1␤------> 3 repr('CPointer') {}; class :: does Foo(7⏏5my & = sub {}) {}␤»
grondilu m: use NativeCall; role Foo[Sub &native] { has &.native }; role Bar is repr('CPointer') {}; class :: is Foo(sub {}) {}
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Cannot resolve caller trait_mod:<is>(<anon|78520784>, Foo, Sub); none of these signatures match:␤ (Mu:U $child, Mu:U $parent)␤ (Mu:U $child, :$DEPRECATED!)␤ (Mu:U $type, :$rw!)␤ (Mu:U $ty…»
grondilu wth
oh 00:01
m: use NativeCall; role Foo[Sub &native] { has &.native }; role Bar is repr('CPointer') {}; class :: is Foo[sub {}] {}
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Cannot resolve caller trait_mod:<is>(<anon|59022496>, Foo, Array); none of these signatures match:␤ (Mu:U $child, Mu:U $parent)␤ (Mu:U $child, :$DEPRECATED!)␤ (Mu:U $type, :$rw!)␤ (Mu:U $…»
grondilu m: use NativeCall; role Foo[Sub &native] { has &.native }; role Bar is repr('CPointer') {}; class :: does Foo[sub {}] {} 00:02
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤No appropriate parametric role variant available for 'Foo'␤at <tmp>:1␤»
grondilu damn it
back to basic:
m: role Foo[Int $] {}; class :: does Foo[13] {}
camelia ( no output )
grondilu ok 00:03
m: role Foo[Sub &] {}; class :: does Foo[sub {}] {}
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤No appropriate parametric role variant available for 'Foo'␤at <tmp>:1␤»
grondilu hum
m: say .WHAT given sub {}
camelia rakudo-moar d43ea0: OUTPUT«(Sub)␤»
grondilu Can't I use a Sub as a parameter in a parametric role?
00:16 mtj_ left, mtj_ joined
grondilu m: role Foo[$] {} 00:18
m: say "hi"
camelia ( no output )
rakudo-moar d43ea0: OUTPUT«hi␤»
grondilu wow that first one took its time
m: role Foo[$] {}
camelia ( no output )
grondilu actually not
m: role Foo[$] {}; class :: does Foo[my $ = sub {}] {} 00:19
camelia ( no output )
grondilu m: role Foo[$f] { my $.f = $f; method f { $!f() } }; class :: does Foo[my $ = sub { say "hi!" }] {}.f 00:20
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Package 'Foo' already has a method 'f' (did you mean to declare a multi-method?)␤at <tmp>:1␤»
grondilu m: role Foo[$f] { my $.f = $f; method talk { $!f() } }; class :: does Foo[my $ = sub { say "hi!" }] {}.talk
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Attribute $!f not declared in role Foo␤at <tmp>:1␤------> 3] { my $.f = $f; method talk { $!f() } }7⏏5; class :: does Foo[my $ = sub { say "hi␤ expecting any of:␤ horizontal whitespac…»
grondilu m: role Foo[$f] { has $.f = $f; method talk { $!f() } }; class :: does Foo[my $ = sub { say "hi!" }] {}.talk 00:21
camelia rakudo-moar d43ea0: OUTPUT«Cannot look up attributes in a type object␤ in method talk at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
grondilu m: role Foo[$f] { method talk { $f() } }; class :: does Foo[my $ = sub { say "hi!" }] {}.talk
camelia rakudo-moar d43ea0: OUTPUT«hi!␤»
grondilu m: role Foo[$f] { method fork { $f() } }; class :: does Foo[my $ = sub fork() is native {}] {}.fork; say "hi" 00:22
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Can't use unknown trait 'is native' in a sub declaration.␤at <tmp>:1␤ expecting any of:␤ rw raw hidden-from-backtrace hidden-from-USAGE␤ pure default DEPRECATED inlinable nodal␤ …»
grondilu m: role Foo[$f] { method fork { $f() } }; class :: does Foo[my $ = sub fork() is native() {}] {}.fork; say "hi"
camelia rakudo-moar d43ea0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Can't use unknown trait 'is native' in a sub declaration.␤at <tmp>:1␤ expecting any of:␤ rw raw hidden-from-backtrace hidden-from-USAGE␤ pure default DEPRECATED inlinable nodal␤ …»
grondilu m: use NativeCall; role Foo[$f] { method fork { $f() } }; class :: does Foo[my $ = sub fork() is native {}] {}.fork; say "hi"
camelia rakudo-moar d43ea0: OUTPUT«hi␤hi␤»
grondilu nice
00:23 mtj_ left, mtj_ joined
grondilu m: use NativeCall; role Foo[$f] { method fork { $f() } }; class :: does Foo[sub fork() is native {}] {}.fork; say "hi" 00:24
camelia rakudo-moar d43ea0: OUTPUT«hi␤hi␤»
grondilu m: use NativeCall; role Foo[$f] is repr('CPointer') { method fork { $f() } }; class :: does Foo[sub fork() is native {}] {}.fork; say "hi" 00:25
camelia rakudo-moar d43ea0: OUTPUT«hi␤hi␤»
grondilu that looks a lot like what I was considering doing. Cool.
b2gills ｢Sub &｣ declares that it requires a Callable that returns a Sub 00:29
tbrowder ugexe: thanks 00:31
.tell ugexe thanks 00:32
yoleaux tbrowder: I'll pass your message to ugexe.
00:33 mcmillhj joined 00:38 BenGoldberg joined, mcmillhj left 00:40 nbrown joined 00:41 nbrown left 00:43 nbrown joined
grondilu m: role Foo[Sub &] {}; class :: does Foo[sub (--> Sub) { sub {} }] {} 00:43
camelia ( no output )
grondilu b2gills: you're right
m: role Foo[&] {}; class :: does Foo[sub {}] {} 00:44
camelia ( no output )
00:53 pierre_ joined 00:55 araujo_ joined 00:58 mcmillhj joined 00:59 araujo left, zacts left, nbrown left 01:04 mcmillhj left 01:09 olinkl left 01:10 kipd left, mithaldu_ left, BuildTheRobots left, notbenh_ left, Lucas_One left 01:11 AngeloMichael left, SmokeMachine____ left 01:12 woodruffw left 01:13 zengargoyle left, Actualeyes joined 01:14 zengargoyle joined, mcmillhj joined 01:16 _slade_ left 01:17 woodruffw joined, woodruffw left, woodruffw joined 01:18 _slade_ joined 01:19 mcmillhj left, Lucas_One joined 01:20 SmokeMachine____ joined 01:23 gfldex left, olinkl joined 01:24 beatdown joined 01:25 kipd joined 01:26 mithaldu_ joined 01:27 notbenh_ joined, AngeloMichael joined, grondilu left 01:28 BuildTheRobots joined 01:31 BillSussman left 01:33 mcmillhj joined 01:37 mcmillhj left 01:38 zacts joined 01:42 TEttinger left 01:45 TEttinger joined, r_i_d joined 01:51 mcmillhj joined 01:56 mcmillhj left 02:05 mcmillhj joined 02:10 mcmillhj left 02:12 _slade_ left, Praise left, _slade_ joined 02:16 wamba joined 02:23 mcmillhj joined 02:28 mcmillhj left 02:33 n1lp7r left 02:37 mcmillhj joined 02:43 noganex_ joined, mcmillhj left, r_i_d left 02:45 noganex left 02:51 mcmillhj joined 02:54 zacts left 02:55 mcmillhj left 03:07 r_i_d joined
r_i_d is there a built-in way to check if a variable is an element in a given array? 03:09
right now i iterate over the array with a for-loop and trigger a duplicate flag accordingly -- is there a better way? 03:10
03:11 cuonglm joined
cuonglm jnthn: Is there any design doc about NFG implementation in MoarVM? 03:12
harmil r_i_d: you can use a Set, but if you are doing this just once, it's more efficient to just scan the list
cuonglm I read your article www.jnthn.net/papers/2015-spw-nfg.pdf 03:13
r_i_d okay, thanks
ugexe m: my @arr = <a b c d>; say "a" ~~ @arr.any; say "z" ~~ @arr.any
yoleaux 00:32Z <tbrowder> ugexe: thanks
camelia rakudo-moar d43ea0: OUTPUT«True␤False␤»
r_i_d oh, that's nice. 03:14
ugexe m: my @arr = <a b c d>; say @arr.contains("a"); say @arr.contains("z")
camelia rakudo-moar d43ea0: OUTPUT«True␤False␤»
r_i_d perfect. 03:16
harmil I didn't even know there was a contains. That's handy. I assume it just scans the array? 03:18
r_i_d Larry thought of everything, aparently.
geraud m: my @a = <a b c d e>; say "d" ∈ @a;
camelia rakudo-moar d43ea0: OUTPUT«True␤»
ugexe m: say "abc".contains("a"); say "abc".contains("z")
camelia rakudo-moar d43ea0: OUTPUT«True␤False␤»
03:19 mcmillhj joined
ugexe `cd rakudo && git grep "method contains"` will generally make it easy to find where and how something is implemented 03:21
03:28 mcmillhj left 03:32 mcmillhj joined, r_i_d left, wamba left 03:37 mcmillhj left 03:38 TEttinger left 03:41 TEttinger joined 03:44 zacts joined 03:46 araujo__ joined, cuonglm left 03:49 TEttinger left, araujo_ left 03:52 TEttinger joined 03:53 nester joined 03:54 mcmillhj joined 04:04 notostraca joined 04:07 TEttinger left 04:13 notostraca left 04:15 notostraca joined 04:26 mcmillhj left 04:28 khw left 04:34 notostraca left, cgfbee left 04:37 notostraca joined 04:39 mcmillhj joined 04:40 pierre_ left 04:41 skids left 04:44 mcmillhj left 04:46 notostraca left 04:47 pierre_ joined 04:48 notostraca joined, mohae_ joined 04:49 nester left 04:51 mohae left, pierre_ left 04:55 holyghost left 04:56 mcmillhj joined 04:58 zacts left 05:00 mcmillhj left 05:10 notostraca left 05:14 notostraca joined 05:20 mcmillhj joined 05:22 zacts joined 05:23 Actualeyes left 05:25 mcmillhj left 05:32 mcmillhj joined 05:34 araujo__ left 05:35 notostraca left 05:36 Actualeyes joined 05:37 mcmillhj left 05:38 notostraca joined 05:47 _slade_ left 05:48 notostraca left 05:50 notostraca joined, pierre_ joined 05:52 CIAvash joined 05:55 pierre_ left 06:03 pierre_ joined, pierre_ left 06:05 mcmillhj joined, robinsmidsrod left 06:06 robinsmidsrod joined, pierre_ joined
avuserow m: class Foo {has $!x; method x {return Proxy.new(FETCH => sub ($) {$!x}, STORE => sub ($, $v) {note "got here"; $!x = $v;})}}; my $f = Foo.new; $f.x = 1; say $f.x; my @attrs = Foo.^attributes(:local)[0].set_value($f, 2); say $f.x; note "pre-3"; $f.x = 3; say $f.x; 06:09
camelia rakudo-moar d43ea0: OUTPUT«got here␤1␤2␤pre-3␤got here␤Cannot assign to an immutable value␤ in sub at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
avuserow looks like set_value and Proxies interact poorly?
06:09 mcmillhj left
avuserow anyway I don't need to use a Proxy per se, I just want to update one attribute whenever another is set... not sure what good ways there are for that 06:10
nine harmil: do _not_ use .contains! That checks if the given value is a substring. On lists it will check if the value is a substring of the stringified list. So it'll find "1" in (10, 11, 12). 06:19
ugexe: ^^^
06:19 mcmillhj joined
nine lizmat++ noticed this and already suggested a fix, but for now that's unfortunately how it is. 06:20
06:21 labster joined 06:24 mcmillhj left 06:33 nadim joined 06:41 domidumont joined 06:45 domidumont left 06:46 domidumont joined 06:47 nadim_ joined, mcmillhj joined 06:49 movlex left 06:50 nadim left 06:53 mcmillhj left 07:02 mcmillhj joined, BenGoldberg left 07:06 mcmillhj left 07:10 notostraca left 07:11 cgfbee joined, notostraca joined 07:21 notostraca left 07:24 notostraca joined 07:28 espadrine joined 07:30 canopus left 07:33 RabidGravy joined 07:34 mcmillhj joined 07:38 canopus joined 07:40 mcmillhj left, ufobat joined 07:43 notostraca left 07:44 firstdayonthejob joined 07:45 mcmillhj joined, notostraca joined 07:47 ufobat left, ufobat joined 07:48 darutoko joined 07:50 mcmillhj left 07:51 notostraca left 07:54 pierrot joined 07:58 notostraca joined 07:59 mcmillhj joined 08:02 jonas2 joined 08:04 mcmillhj left 08:06 notostraca left, quester joined 08:07 notostraca joined 08:09 Jody_ joined, g4 joined 08:11 Jody_ left 08:16 firstdayonthejob left 08:19 leont joined 08:20 firstdayonthejob joined 08:22 araujo joined 08:24 notostraca is now known as TEttinger 08:28 TEttinger left 08:29 lizmat left 08:30 TEttinger joined 08:31 salva joined 08:37 mcmillhj joined 08:42 mcmillhj left 08:43 TEttinger left 08:44 TEttinger joined 08:45 mvorg joined 08:46 DrEeevil left 08:47 xiaomiao joined
jnthn .tell cuonglm No design doc, though the key data structures are described somewhat in the .h files; of note see github.com/MoarVM/MoarVM/blob/mast...ings/nfg.h and github.com/MoarVM/MoarVM/blob/mast...ormalize.h 08:49
yoleaux jnthn: I'll pass your message to cuonglm.
08:53 mcmillhj joined 08:55 TEttinger left 08:57 mcmillhj left 08:59 TEttinger joined, Actualeyes1 joined, Actualeyes left 09:03 pierre_ left 09:06 smls joined
TEttinger jnthn: is that grapheme normalization stuff? guessing by names 09:08
09:09 TEttinger left, TEttinger joined, mcmillhj joined
jnthn TEttinger: Yes :) 09:10
(Was answering a question from the backlog :))
09:12 user9 joined 09:15 mcmillhj left, quester left 09:16 notostraca joined 09:19 TEttinger left, notostraca is now known as TEttinger 09:26 mcmillhj joined 09:31 mcmillhj left 09:33 Actualeyes1 left 09:35 nadim_ left 09:43 TEttinger left 09:54 Actualeyes joined 09:57 mcmillhj joined 10:02 mcmillhj left
melezhik Good day gentlemen 10:16
how I can create a perl6 class with a "nested" name? Like class Foo::Bar::Baz 10:17
looks like simple having "class Foo::Bar::Baz {}" does not work
jnthn m: class Foo::Bar::Baz { method m { 'bacon' } }; say Foo::Bar::Baz.m 10:22
camelia rakudo-moar f2df2c: OUTPUT«bacon␤»
jnthn Seems to work fine, so I suspect you're running into some other problem :) 10:23
melezhik jnthn: will double check thanks! 10:24
ok, I see the root now. Had to add class Name / file path into META.info , so t/ succeed when run panda --force install . 10:26
10:27 kaare_ joined
melezhik anybody knows the reference to Perl6 Test ? all I find is perl6maven.com/how-to-test-perl6-modules but it is not enough for me ... 10:28
DrForr docs.perl6.org/language/testing 10:30
melezhik DrForr: thanks
looks good :)
10:30 leont left
melezhik how can have panda installer to show verbose test output when run my tests with `panda install .` 10:31
?
looks like --verbose don't work
10:31 mcmillhj joined 10:32 labster left
DrForr "Doesn't work" isn't as helpful as "I would have expected... but didn't see it." 10:33
melezhik ok. "panda --force --verbose install ." give me a panda help info, instead of running tests 10:34
"panda --force install ." runs my tests and makes install 10:35
10:36 mcmillhj left
DrForr I'll have to install locally to confirm. 10:36
nine melezhik: add --verbose to your ~/.proverc 10:37
melezhik DrForr: github.com/melezhik/outhentix
head ~/.proverc 10:38
--verbose 10:39
DrForr 'panda --force --verbose install .' appears to be failing correctly as '.' isn't a list of modules, but a directory. If you're used to that from Perl 5 tools, keep in mind this is perl 6 :)
melezhik but still have a none verbose output from "panda --force install ."
10:39 salva left
nine DrForr: but "panda install ." can actually be used to install an unpacked dist from a directory 10:40
melezhik DrForr: ok, let me say in other words. I have unit tests under t/ and I want to run them and see verbose output ...
how can I do this?
prove --exec perl6 -r t ? 10:41
nine melezhik: I usually run tests with prove -e 'perl6 -Ilib' -v t
melezhik yes, prove --exec perl6 -r t works
tbrowder melezhik: I'm having better luck with zef. 10:42
melezhik I just expected that pand would respect --verbose
ok. for now, "prove" is fine for me
10:42 salva joined 10:49 mcmillhj joined 10:54 mcmillhj left 10:56 leont joined
tbrowder ugexe: do you know how to get info or verbose output from Build.pm? I'm using zef and only get info messages using --debug, nothing with --info (default) or --verbose. 11:01
dalek osystem: cae96fc | (Sam Gillespie)++ | META.list:
Adding Numeric::Pack to ecosystem.
 11:07
osystem: 4f525ef | azawawi++ | META.list:
Merge pull request #245 from samgwise/patch-2

Adding Numeric::Pack to ecosystem.
11:10 rindolf joined
avar shorter oneliner for this in perl6: perl -le"print substr(hex(substr(shift,0,2)),-1)" $(hostname -f | md5sum) ? 11:10
11:11 bjz joined
avar for "Pick N% of hosts depending on their hostname" 11:11
11:15 leont left
avar best I can think of: 11:15
perl6 -e 'exit :16(@*ARGS[0].substr(0,2)) % 10 == 0 ?? 0 !! 1' $(hostname|md5sum)
11:18 xinming joined
avar substr not needed due to native bigint: 11:20
perl6 -e 'exit :16(@*ARGS[0]) % 10 == 0 ?? 0 !! 1' $(hostname|md5sum)
11:22 xinming_ left 11:26 mcmillhj joined 11:28 MasterDuke joined 11:29 bioduds left
MasterDuke avar: perl6 -e 'exit :16(@*ARGS[0]) %% 10 ?? 0 !! 1' $(hostname|md5sum) 11:29
11:29 bioduds joined
bioduds im having trouble installing perl6 11:29
DrForr What kind of trouble? 11:30
bioduds i have it installed but a very old version
11:30 bjz left
bioduds ===SORRY!=== Error while compiling hello.pl Can't use unknown trait 'is required' in an attribute declaration. 11:30
11:30 mcmillhj left
El_Che that reads like Moo/Moose 11:30
bioduds and i dont seem to be able to update it 11:31
no matter what I do
I can't get it to work
El_Che bioduds: have you tried rakudobrew? 11:33
bioduds yes
tried a couple of times
El_Che how did you install the version you have?
os package?
bioduds i installed by running sudo apt-get install rakudo 11:34
11:34 pierre_ joined
bioduds im on ubuntu 14.04 11:34
moritz try uninstalling it, and building from source instead
bioduds I did
El_Che sudo apt-get remove --purge rakudo
bioduds let me do it again 11:35
El_Che then just follow the rakudo instructions, that should work
DrForr Right, so you got the versionfrom the repository, which isn't up-to-date. Have you tried removing the package that you installed, and building via rakudobrew?
bioduds i actually only removed it not purged
El_Che (as there is no rakudo on the system)
(removed should be ok, purge is because I am a digitally clean type :) )
11:35 Actualeyes left
bioduds let me try 11:35
i would like to make a suggestion to the top cat p6 guys if i may and if they see this. The best installing way is Meteor 11:36
curl install.meteor.com/ | sh 11:37
perhaps this may be done with Perl6
DrForr Well-volunteered, grab the file, change it to download rakudobrew and post a link in the channel? 11:38
11:38 mcmillhj joined
bioduds ok, so I did sudo apt-get remove --purge rakudo 11:39
moritz you can also use the docker container: github.com/perl6/docker
bioduds now Im to follow which line?
this here? perl6.org/downloads/
11:41 MilkmanDan left
DrForr bioduds: Are you Eduardo Capanema? If so, welcome! 11:42
bioduds yep, I am. thanks DrForr 11:43
DrForr (and even if you're not, still, welcome!)
bioduds I have one successful installation on my local machine
ubuntu 16.04
11:43 bjz joined, mcmillhj left
bioduds but on my AWS server ubuntu14.04 it is not working 11:43
i believe 16.04 repo apt is correct and 14.04 is not 11:44
and once you put bad p6 version there is no way to upgrade it
11:46 mr-fooba_ left
DrForr Well, the version from Ubuntu 16.04 is certainly going to be more recent. Like any other project, you can either wait for the next version to become available as a package, or go out and install it yourself. 11:47
El_Che bioduds: read your post! 11:48
:)
bioduds: damn straight to the point :)
DrForr Installing from packages is more convenient, certainly, but means that you have to wait until someone else packages the latest version, or you can go out and get the latest yourself.
El_Che bioduds: I build often from source for my docker setup and it's pretty straightfwd (but it takes a long time to finish) 11:49
pkgs is something we indeed need to provide
bioduds im trying again
11:50 Actualeyes joined
El_Che ls 11:50
oops
bioduds my point is if perl6 is aimed at beginners, installation must be made reaaaaaly easy
i purged current p6 version 11:51
now i downloaded rakudobrew
moritz then help making it easy
bioduds and im running the script
DrForr No argument there. But doing that takes time and effort, effort that people have channeled into making the software actually work.
bioduds I am helping, I am trying to point out the easiest way I know 11:52
which is Meteor shell script after curl
DrForr Right. Feel free to rewrite it to use rakudobrew and submit it here.
perlbrew has almost exactly the same setup, you might want to look at that first. 11:53
moritz bioduds: thing is, throwing "hey everybody, please do what I want" rarely works for open source projects, even if you offer a valid perspective
bioduds moritz, i am not throwing do what i want 11:54
not at all
I am suggesting
very different
moritz bioduds: the difference is just in wording, not in action
bioduds Im not demanding a thing moritz
it is merely a suggestion 11:55
11:55 smls_ joined
El_Che ok, let's get past this. bioduds, is rakudobrew running? 11:55
moritz bioduds: and it's received. And I'm merely suggesting that you actually implement your suggestion, because that has the best chances of success
bioduds no not running
11:55 smls left
bioduds Command failed (status 512): make Failed running /usr/bin/perl Configure.pl --backends=moar --gen-moar --git-reference="/home/ubuntu/.rakudobrew/bin/../git_reference" --make-install at /home/ubuntu/.rakudobrew/bin/rakudobrew line 58. main::run('/usr/bin/perl Configure.pl --backends=moar --gen-moar --git-r...') called at /home/ubuntu/.rakudobrew/bin/rakudobrew line 386 main::build_impl('moar', undef, '') called at /home/ubuntu/.raku 11:55
DrForr And your suggestion is noted. Feel free to modify install.perlbrew.pl's script and submit it.
bioduds: That's cut off. Please post your full install log as a github gist or something like that, where we can see it. 11:56
El_Che bioduds:
How to get Rakudo Perl 6
The recommended way to use Rakudo is by downloading Rakudo Star – a useful and usable production distribution of Perl 6. Rakudo Star includes both the Rakudo compiler and some commonly used libraries.
OS_Windows_8 .msi Rakudo Star
2016.07
11:56 El_Che left, pierre_ left, El_Che joined
bioduds how do i do that? 11:57
tbrowder jnthn: where is specification for the META6 file?
El_Che sorry for the flooding. Damns copy paste. I hope my leaving the channel stopped the flooding
bioduds post as a github gist?
El_Che bioduds: sudo apt-get instal build-essential git
jnthn tbrowder: No idea, sorry; I don't really work on that side of things. :) 11:58
El_Che bioduds: maybe you're missing a depency for building rakudo
jnthn tbrowder: S11 or S22 at a guess though
DrForr Whatever pastebin you prefer. The point is that we could only see part of the log. pastebin.ca, snit.ch, any pastebin will do.
bioduds well, happened twice
once in a digital ocean ubuntu 1404
and now on AWS ubuntu 14.04
DrForr bioduds: Yes, and if you don't install the dependency it'll keep breaking.
tbrowder thanks, but how did you develop yr meta6 test? 11:59
bioduds sorry?
not following you tbrowder
DrForr tbrowder: He's talking to someone else.
bioduds oh, sorry 12:00
tbrowder sorry, question was for jnthn
12:00 smls_ is now known as smls
bioduds any ideas how should i proceed? 12:00
jnthn tbrowder: Um...I don't think I did, are you confusing me with then other Jonathan? :)
(Who goes as RabidGravy here :)) 12:01
DrForr bioduds: Yes, put your full install log somewhere that we can see it.
tbrowder probably, sorry!
jnthn :)
DrForr bioduds: Otherwise we're going to have to guess at what's going on, and that takes more time and patience than most of us have.
jnthn You wouldn't be the first one ;)
bioduds ok
12:01 pierre_ joined
bioduds i installed build-essential git 12:02
can you tell exactly how to put log on?
please
im doing rakudobrew build moar 12:03
El_Che bioduds: paste.ubuntu.com/ <-- copy the code there, paste the url here
by code I mean error
DrForr bioduds: Go to the site El_Che mentioned, paste the log there, share the link with us here.
That way we can see what failed and can try to help.
bioduds doing it 12:04
12:05 eliasr joined
bioduds posted 12:06
paste.ubuntu.com/23107097/
tbrowder ref META6 test: i did find the reference right where it should be expected: Jonathan Stowe has the reference to S22 toward the end of the README--i just hadn't read far enough: "RTFM, tbrowder!" 12:08
12:08 SpaceMario joined
moritz bioduds: looks like too little RAM. How much do you have available? 12:08
SpaceMario m: gist.github.com/zoffixznet/dc09594...29756bb11c
camelia rakudo-moar f2df2c: OUTPUT«False␤»
bioduds not much
DrForr Stage parse : MVM_platform_alloc_pages failed: 12 # I'm guessing you're RAM-limited.
bioduds its a free tier in amazon
SpaceMario Are <ws> tokens not available for use in regular regexes?
bioduds and same error occurred in digital ocean minimal vps slice 12:09
moritz bioduds: you need a bit more than 1GB
maybe 1.5
bioduds but my local machine is much better and worked
SpaceMario I'd say like 2
bioduds: get more swap. I successfully built on 1GB boxes with tons of swap 12:10
moritz m: say so 'a b' ~~ /a <ws> b/
camelia rakudo-moar f2df2c: OUTPUT«True␤»
moritz SpaceMario: ^^ seems to work
bioduds im not sure i can manage swap in vps
SpaceMario moritz: Hm, I wonder why my gist fails. I seem to have all the right bits in place gist.github.com/zoffixznet/dc09594...29756bb11c
moritz bioduds: or build on your local machine, and copy over the resulting files (but you need the same paths)
nine DrForr: PLEASE stop recommending rakudobrew to normal users. There's probably only 5 people in rakudobrew's actual target audience and I don't think any one of us is actually using it. 12:11
DrForr Okay, done, over with. 12:12
smls nine: Explain?
I'm using rakudobrew and have recommended it to some people.
SpaceMario m: gist.github.com/zoffixznet/1e0f8db...1072d98f6a
camelia rakudo-moar f2df2c: OUTPUT«True␤»
DrForr Then you get the privilege of being chided as well :) 12:13
SpaceMario (was doing .lc on input, but regex still had uppercase letters)
bioduds so, just for me to know, I need at least 2 Gb Ram to safely install perl6?
nine smls: rakudobrew is for when people need to manage _multiple_ different rakudo installations on the same machine. It's not a good way to safe a user from typing in 3 commands on the shell. It causes much more problems than it solves.
DrForr Or 1GB and sufficient swap, as somene else says. 12:14
SpaceMario bioduds: to build it, yeah, somewhere around that. And not RAM but memory, so swap will do in a pinch.
bioduds ok
nine smls: there have been lots and lots of reports of issues with module installation reported here and I actually stop reading once I see rakudobrew mentioned.
bioduds got it
DrForr Note to self: remove from slides :) 12:15
smls nine: Interesting. Why would modules care where $PATH and stuff is set to?
12:15 buggable left
SpaceMario uses rakudobrew 12:16
nine smls: the problems usually come from having a mix of outdated and current rakudoversions installed and are undebuggable as rakudobrew users usually don't even know where the stuff gets installed to
12:16 buggable joined
SpaceMario I just nuke everything when I upgrade :P 12:16
12:16 MilkmanDan joined
smls OK, so what is the best alternative for normal users? (Assuming their distro doesn't have a Rakudo package, or only an ancient one.) 12:16
nine SpaceMario: exactly! That's completely unnecessary. Perl 6 is designed so you never have to re-install modules when upgrading Perl 6.
SpaceMario :) 12:17
12:17 SpaceMario left
nine git clone [email@hidden.address] && cd rakudo && perl Configure.pl --gen-moar --make-install 12:18
smls ok
nine That ^^^ is all you need to do to install rakudo. Put that in a script, put that onto perl6.org and point people at curl rakudo.org/install-script | curl
| bash of course 12:19
12:19 Juerd joined
nine Upgrading that is just git pull && bash config.status && make install 12:19
bioduds yep, or better yet install.rakudo.org 12:20
DrForr bioduds: ^^^ There you go, first contribution!
bioduds im happy with that :)
I learned to program in Perl!
In 2006
you dont see many 'I began in Perl' guys out there and let me tell you one thing. It made me look at programming languages veeeery differently as I went out learning c, c++, java and stuff 12:22
and this is the reason Im very excited with Perl6
MasterDuke i have a perl6 directory in my home, i clone moar, nqp, and rakudo into there and then point them all at the same --prefix
it works well for testing changes to those three individual components 12:23
but having to clone a ton of modules and run everything with a giant -I list is kind of annoying
12:23 buggable left
MasterDuke but when i first used panda several months ago it didn't work very well (and i apologize, but i don't remember why) 12:24
12:24 buggable joined
nine MasterDuke: if panda or zef give you trouble, please report that. Also until that is solved, you can install the cloned modules by: cd whatever-module ; perl6 ~/rakudo/tools/install-dist.pl . 12:25
MasterDuke nine: ah ha! i had no idea about install-dist.pl 12:26
12:29 mcmillhj joined
nine It is fairly unknown. But I do have plans to rename it to raccoon and have it installed alongside perl6 12:33
12:34 mcmillhj left
RabidGravy boo! 12:34
12:39 rindolf left
Woodi is there some way to clone module code and install it "manually", without panda or zef ? 12:41
nine Woodi: what I just told MasterDuke?
Woodi nine: oops, -dist part hijacked my understanding of that :) thanx 12:43
RabidGravy nine, +1 on actually installing it, have no feelings about the name 12:46
there should be a caveat with the above that if the module in question requires some build action (such as compiling so helper library,) it probably won't work 12:47
El_Che nine: I am kind of surprised. I get that rakudobrew does too much. But the only alternative according to the doc (rakudo.org/how-to-get-rakudo/#Insta...g-Rakudo). Isn't rakudo a better solution than directly compiling the src for a regular use (who will install systemwide and won't be able to clean it up afterwards)? 12:49
nine RabidGravy: I actually hope we are gonna make some progress on the build front soon. I do want raccoon to be able to execute build steps. Primary use case being creating distro packages for Perl 6 modules. 12:52
RabidGravy totally up for that :) 12:53
nine El_Che: the steps listed as manual installation will _not_ install rakudo system wide. It will install into rakudo/install which can be removed easily. 12:54
El_Che: that's even easier than some hidden .rakudobrew directory.
12:56 Averna joined 12:57 nadim joined
nadim hi, quick question, eqv is still not implemented for matches, right? 13:01
13:02 bjz_ joined, bjz left 13:06 rmmm left 13:08 mcmillhj joined 13:14 cdg joined 13:17 rmmm joined
Woodi possibly it's problem with current *nix practicess of sharp distinction between system wide and ~user private only... 13:20
timotimo Woodi: yeah, we should teach our platforms (like rakudo or v8 or python) to look into every user's home directories for libraries when they aren'n in the system's repository 13:21
"huh, today 'import sys' generated 95 terabyte of traffic over samba and nfs shares."
i meant to say 'import syss' 13:22
typos are hard when you're trying to make them intentonially
Woodi timotimo: :) i will write some story when system update will finish :) 13:23
[Coke] wonders how quickly irclog adds sends to its own search index. 13:27
13:27 zakharyas joined
[Coke] stmuk_: you end with $COLON because you're also avoiding having a literal : in a file on disk, which isn't pleasant on some platforms (and we support people running the app out of a checkout) 13:29
[Coke] is sure there's a prettier way to solve this eventually, but this works.
13:31 wamba joined, MilkmanDan left, ufobat left
[Coke] nine: I'm using rakudobrew for my local development. 13:33
if we're not supposed to be recommending it, why do we even have it?
13:33 MilkmanDan joined
[Coke] .... and I caught up, seeing nine's objections later in backlog. 13:33
I don't think recommending "install from source" is a better option for most users. 13:34
13:34 mr-foobar joined, bjz_ left
nine [Coke]: the best option would probably we a small script that does pretty much those manual steps. 13:34
[Coke]: btw. even tadzik said that rakudobrew was actually more for developers :) 13:35
mst [Coke]: rakudobrew is for developers, just like using perlbrew isn't something you should ever suggest to newbies
[Coke] I respectfully disagree with him, building from source is for developers. :)
mst [Coke]: that's ... not the point
13:35 domidumont left
mst the point is rakudobrew has lots of extra moving parts designed to accommodate multiple rakduo installs 13:36
[Coke] in my mind, it would be (devs) -> source -> rakudobrew -> R * (non devs)
mst and also is rather inflexible and weird
similar to perlbrew
whereas if you just want *A* perl, then you use perl-build
what we need basically is a rakudo-build equivalent
[Coke] As a user of perl 5, I use perlbrew all the time. YMMV.
13:36 nebuchadnezzar left
mst if you're develop[ing against multiple perl5 installs it's useful 13:36
13:37 nebuchadnezzar joined, pmurias joined
mst for newbies, it's a bunch of complicated fragile stuff they won't understand 13:37
[Coke] has never heard of perl-build, and a google search is not enlightening.
mst p3rl.org/Perl::Build
13:37 domidumont joined
mst honestly, once I get Alien::Rakudo shipped I would argue that that's the simplest installer for newbies we have available currently 13:38
[Coke] mst: that page doesn't really help explain what that is to someone who doesn't know, btw.
mst though, equally, I might consider obsoleting myself again later
[Coke] but I trust your collective judgement, that'sf ine.
pmurias for perl5 programmars 'cpanm Alien::Rakudo' seems like the easiest option
mst right, *if* you already have cpanm+local::lib I think it's a no-brainer 13:39
if you don't, well, I'll see about providing something for that case as well later 13:40
pmurias works for cpanm+perlbrew too ;)
mst I am still a little bit smug about how short twitter.com/shadowcat_mst/status/7...2365946883 is 13:41
MasterDuke arch linux has a relatively recent rakudo in the AUR, but the rakudo available to apt in kubuntu 16.04 is 2015.11 13:43
13:43 Sgeo left
tadzik nine, [Coke]: well, I am to blame for advertising it as an easiest way to install rakudo at some point :3 13:44
(because it is/was) 13:45
pmurias mst: in addition to having Alien::Rakudo it would also be nice to have something that installs zef/panda
mst yes. but I'm unhappy with Alien::Rakudo::Inline::Perl5 currently
nine tadzik: it may as well still be the easiest way to get up and running. But unfortunately also an easy way to run into trouble later on. 13:46
mst tadzik: it was. and I am both very glad you wrote it, and very glad I've now got something else I can use 13:47
pmurias mst: what's wrong with it? 13:48
mst 14:36 < mst> the point is rakudobrew has lots of extra moving parts designed to accommodate multiple rakduo installs
14:37 < mst> for newbies, it's a bunch of complicated fragile stuff they won't understand
I mean, I was never able to entirely work out what rakduobrew was *doing* 13:49
eventually I straced the fucker and grepped for exec() calls
but it's hardly the most transparent of systems
as a result of which, that nine said
*what
timotimo what's wrong with Alien::Rakudo::Inline::Perl5 except the long name?
mst it doesn't write a .packlist correctly yet because there's no way to get file lists out of the CUR, and I'm not sure there necessarily should be, so I'm not sure how I want to implement that 13:50
nine mst: is there anything in the target CUR besides Inline::Perl5?
13:51 Matthew[m] left, M-Illandan left, tadzik left, ilmari[m] left
mst nine: currently, yes, it has to install into the Alien::Rakudo site CUR so that Inline::Perl6 finds it 13:51
I'm tempted to give it its own CUR instead and tweak Inline::Perl6 to know to look for it
but basically then we're looking at "how does one handle module installation for Alien::Rakudo in general" 13:52
and that's a bit more complicated
ugexe which files do you want to get from CUR? I'm pretty sure they should all be available in the meta
mst ok, so, during ::EU nine and I looked at it and it wasn't immediately obvious to me how to convert that to a .packlist 13:53
nine The way baking rpm packages will work is: install-dist.pl --to /tmp/some_empty_temporary_build_directory --for vendor /path/to/module
mst also really I don't just want to do that
I want to stage via blib/
nine The files in /tmp/some_empty_temporary_build_directory can then be packaged and extracted over / on installation 13:54
13:54 g4 left
mst nine: ah, and --to is your snowflake name for DESTDIR? 13:54
nine yes
Because of course it was amateur hour :) 13:55
13:55 tadzik joined
tadzik nine: yeah, true 13:55
nine --for vendor or --for site is so the paths in the precomp files will be relative to those repositories.
13:56 brrt joined
mst nine++ # bahahahahahaha 13:56
14:06 r_i_d joined 14:07 ilmari[m] joined 14:09 Averna left 14:10 telex left 14:12 telex joined 14:13 zakharyas left
r_i_d when writing a test, how do I include the file to be tested? 14:13
dalek ateverable: b919be5 | MasterDuke17++ | Bisectable.p6:
Convert Bisectable to use get-short-commit
 14:14
ugexe use lib 'lib'; but really you should be passing it via the command line: `prove -e 'perl6 -Ilib' t/` (the -Ilib)
14:15 zakharyas joined
r_i_d when using 'lib', do I use the whole filename like "../filename.pm6"? 14:16
ugexe if you have a module "Foo::Bar" in lib, then: `use lib 'lib'; use Foo::Bar;` 14:17
moritz r_i_d: use lib is for directories containing modules, not for file names
r_i_d oh, okay, i don't have my module set up right. 14:18
ugexe you can map the module name to file name differently in your META6.json `provides` field but its generally a good idea to follow the s{::}{/} naming convention
r_i_d how would i do it if I had in my project directory "filename.pm6" and "t/test-filename.t"? 14:19
timotimo if filename.pm6 is supposed to be an executable script, it'd go into bin/ and most of its code into lib/App/YourMagicalApp.pm 14:20
ugexe if your shell is in the project root: -I. otherwise `use lib <..>`
timotimo then the whole thing would be called App::YourMagicalApp
14:21 skids joined
timotimo when panda, zef, and friends install your lib, they'll have the "lib/" folder set up appropriately for modules in it to be found with just "use TheModuleName" 14:21
ugexe ~/my_project $ prove -v -e 'perl6 -I.' t/test-filename.t
14:22 pmurias left
timotimo in general, though, it's quite uncommon to have scripts or modules in the root directory of a project 14:22
r_i_d yeah, that makes sense. If i have "bin/filename.pm6" and "t/test-filename.t", how do I install it using panda? 14:23
14:23 domidumont left
ugexe Build.pm 14:23
14:23 pmurias joined
ugexe pm6 is a perl module extension, it probably doesn't belong in bin/ 14:24
moritz uhm, adding them to the META6 file should be enough, no?
ugexe r_i_d: just look at other perl6 modules to get a feel for the boilerplate/layout 14:27
tbrowder i'm still experimenting but i'm not sure the META6 file can handle a complex installation. Does the presence of a Build.pm override the META6? 14:28
ugexe no
nine @all: please push people into putting their modules into a lib/ dir more strongly. Otherwise they will run into problems, e.g. when trying %?RESOURCES 14:29
14:30 jonas2 left, benchable6 left, domidumont joined
timotimo ugexe: um, are you sure Build.pm6 is needed for that at all? 14:31
ugexe timotimo: is needed for what? 14:32
14:32 benchable6 joined
timotimo ugexe: having a bin/filename.pm6 14:33
literal avar: perl6 -e 'exit +!(:16(@*ARGS[0]) %% 10)' $(hostname|md5sum)
ugexe no, i was pointing out Build.pm is a module that is typically found in the project root 14:34
timotimo oh
OK, that's fair. i wouldn't expect it to be installed along with my module, though
that's potentially far beside the point
ugexe it doesnt get installed
timotimo of course not :)
so is the rule "modules directly inside the project root don't get installed"? 14:35
ugexe its more like only files inside bin/ get automatically installed. everything else must be declared explicitly in the META6
14:35 mcmillhj left
ugexe so technically you could install modules from the root 14:36
"provides" : { "Build" : "Build.pm" }
timotimo huh. for some reason i have completely forgotten about the existence of provides
clearly, you shouldn't come to me for thoughts/advice on modules at all 14:37
14:37 benchable6 left
ugexe but really `Build.pm` needs to be replaced with something more sane 14:38
nine oh yes, please! 14:39
harmil Sometimes it's impossible for me to express how much I love working in this language. From my recent sequences commit (Pascal's Triangle): "𝕀.triangle.map: -> ($n,$k) { $n choose $k }"
14:39 mcmillhj joined, ufobat joined
timotimo ugexe: a set of common functionalities as directives for the meta6.json? 14:39
nine ugexe: I thought about just starting by giving raccoon everything it needs to build Inline::Perl5
driven by META6.json of course
ugexe well Inline::Perl5 has dependencies, so that requires finding all that stuff as well 14:40
14:40 acrussell joined 14:41 committable6 left
mst does it? 14:41
Inline::Perl5 was the only thing I had to add to a vanilla rakudo
to have sstuff work
nine ugexe: I made them optional last week to simplify mst's life a bit
mst ah
ugexe the META.info still lists them as deps 14:42
nine LibraryMake didn't help all that much anyway and File::Temp was only used in tests.
ugexe timotimo: essentially. or hooks design.perl6.org/S22.html#hooks
nine ugexe: Build.pm still uses LibraryMake and having File::Temp for tests is nice, too. And panda and zef have no problem installing it. 14:43
14:43 committable6 joined, benchable6 joined
nine ugexe: installation via perl6 configure.pl6 && make && make install works without any dependencies 14:43
14:45 khw joined
ugexe nine: why not have the Build.pm just be `class Build { method build($dir) { run "perl6", "configure.pl6", :cwd($dir); } }` 14:45
nine That.....is quite clever 14:46
14:48 mcmillhj left
r_i_d where are modules stored so I can look at the layout for comparison? 14:50
14:50 rindolf joined 14:51 n1lp7r joined, kaare_ left
harmil r_i_d: That depends on how you installed Perl 6. If you used rakudobrew, then they're in your ~/.rakudobrew if you used Star, then they're under Star's prefix. If you used a Windows installer, then there's going to be a c:\rakudo I think 14:54
r_i_d I used rakudobrew under ARM linux -- should I put my modules with the rest?
tbrowder ugexe: ref zef: for option '--verbose' how does one get a message in Build.pm visible to the user?
ugexe tbrowder: set `ZEF_BUILDPM_DEBUG=1` 14:56
mst ugexe: amusingly, I *almost* ended up making Alien::Rakudo::Inline::Perl5's Makefile.PL that ;) 14:57
harmil r_i_d: If you're writing a module, see docs.perl6.org/language/modules#Di...ng_Modules Once you've written your META6.json, you should be able to just "panda install ."
14:58 cdg left
r_i_d ah, I have to WRITE a META6, got it. 14:58
15:00 notbenh left
melezhik Hi! Anybody could help me with '~' in perl6 regexps? 15:01
let's I have perl6 -e "say 'OK jjj OK' ~~ /OK ~ OK [ jjj ]/;" 15:02
it says Nil
15:02 notbenh joined
melezhik I could not find an proper explanation of ~ in perl6 regexp, but probably ( I might be wrong ) it is something for looking inside some sub area ? 15:03
15:03 zakharyas left
melezhik delimited by some regexps ( in my trivial case this is OK ... OK ) 15:04
ugexe mst: you can also execute the Build.pm itself ala `perl6 -Ilib -I. -e '::("Build").new.build($*CWD.absolute);'` (this is how zef executes the Build process)
er add a -MBuild after -I.
15:04 Ven joined
mst aha 15:04
mst was about to ask
Ven o/ 15:05
15:05 mcmillhj joined
melezhik another example for this construction is github.com/moritz/json/blob/master...mar.pm#L26 15:06
I understand this, and how this work but can't use the same thing at my regexps constructions 15:07
tbrowder ugexe: that doesn't work for me; the question is, what print-type statements inside Build.pm will show on stderr or stdout when using option --verbose? so far i can see say msgs with --debug but then i also get extra msgs from zef, verbose seems to do nothing
15:10 MasterDuke left
ugexe ZEF_BUILDPM_DEBUG=1 will pass the stdout and stderr outwards to zef. zef filters messages based on --verbose/--debug etc 15:11
15:12 n1lp7r left
pmurias Ven: \o 15:13
15:13 kaare_ joined
ugexe if you are just testing your build then use `ZEF_BUILDPM_DEBUG=1 zef -v build .`. This will also output the exact command used to execute the Build.pm so you can do it without zef 15:14
15:15 M-Illandan joined, Matthew[m] joined 15:19 kurahaupo joined 15:21 zacts left
Ven awww.. do we not have a grep-not of sorts? mmh 15:21
[Coke] melezhik: you're looking for "OK" delimited by "jjj"? 15:22
melezhik no, jjjj b/w 'OK' delimiters 15:23
[Coke] I think you want the %% operator 15:24
not sure where you got ~
Ven well, ~ is useful for this kind of things as well 15:25
melezhik ok , please take a look at github.com/moritz/json/blob/master...mmar.pm#L5 or github.com/moritz/json/blob/master...mar.pm#L26 15:26
I know that this one works. It validates something inside '[' .. ']' or something inside '"' 15:27
now I want to have the similar regexps working on my code but I can't 15:28
as far as I guess ( I might be wrong ) a construction kinda foo ~ bar is for looking something between foo and bar
moritz melezhik: note that ~ doesn't work too well with backtracking 15:29
melezhik but I can reuse this approach at my regexps
mortiz: I don't need a backtracking at least now ...
moritz melezhik: so, give a small example of what's not working for you? 15:30
melezhik sure, just a second
for example this one works fine "say '{ OK 9 OK }' ~~ /OK+ % \d /;"
｢OK｣ 15:31
and I want to changes this
[Coke] melezhik: btw:
m: say '{ OK 9 OK }' ~~ /OK+ % \d /
camelia rakudo-moar f2df2c: OUTPUT«｢OK｣␤»
melezhik to this one - say '{ OK 9 OK }' ~~ /'{' ~'}' [ OK+ % \d ]/
and have - Nil 15:32
not sure that this chunk - [ OK+ % \d ] is correct
moritz melezhik: nothing matches the whitespace in there
melezhik: you also see that "OK 9 OK" ~~ /OK+ % \d/ matches only one OK, not the whole string 15:33
m: say "OK 9 OK" ~~ /:s OK+ % \d/
camelia rakudo-moar f2df2c: OUTPUT«｢OK ｣␤»
moritz m: say "OK 9 OK" ~~ /:s OK + % \d/
camelia rakudo-moar f2df2c: OUTPUT«｢OK ｣␤»
moritz m: say "OK 9 OK" ~~ /:s OK + % \d /
camelia rakudo-moar f2df2c: OUTPUT«｢OK ｣␤»
moritz grml
melezhik this one "say '{OK9OK}' ~~ /'{' ~'}' [ OK+ % \d ]/;"
still return Nil
moritz ok
oh
smls 'OK' +
moritz OK
right
what smls said 15:34
OK+ matches OK, OKK, OKKK etc.
melezhik sure I know
moritz m: say "OK 9 OK" ~~ /:s 'OK' + % \d /
camelia rakudo-moar f2df2c: OUTPUT«｢OK 9 OK｣␤»
moritz m: say '{ OK 9 OK }' ~~ /:s '{ 'OK' + % \d /
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in single quotes; couldn't find final "'" ␤at <tmp>:1␤------> 3ay '{ OK 9 OK }' ~~ /:s '{ 'OK' + % \d /7⏏5<EOL>␤ expecting any of:␤ single quotes␤»
moritz m: say '{ OK 9 OK }' ~~ /:s '{' 'OK' + % \d /
camelia rakudo-moar f2df2c: OUTPUT«｢{ OK 9 OK ｣␤»
moritz m: say '{ OK 9 OK }' ~~ /:s '{' ~ '}' [ 'OK' + % \d ] /
camelia rakudo-moar f2df2c: OUTPUT«｢{ OK 9 OK }｣␤»
moritz ... and it works if you do it right :-) 15:35
m: say '{ OK 9 OK }' ~~ /:s '{' ~ '}' 'OK' + % \d /
camelia rakudo-moar f2df2c: OUTPUT«｢{ OK 9 OK }｣␤»
moritz the brackets aren't necessary, it seems
15:35 n1lp7r joined
avuserow m: class Foo {has Int $.bar is rw;}; my $f = Foo.new(:bar(1)); say $f; $f.^attributes(:local).first(*.name eq q{$!bar})[0].set_value($f, 2); say $f; $f.bar = 3; # Anyone know why `bar` has become immutable here? 15:36
camelia rakudo-moar f2df2c: OUTPUT«Foo.new(bar => 1)␤Foo.new(bar => 2)␤Cannot modify an immutable Int␤ in block <unit> at <tmp> line 1␤␤»
15:36 kaare_ left
jnthn Because set_value binds rather than assigns, I'd guess 15:36
moritz avuserow: because set_value binds
melezhik yes, but this one still return Nil - say '{ OK 9 OK }' ~~ /:s '{' ~ '}' [ OK + % \d ] / 15:37
jnthn Do get_value and assign to the returned container
moritz avuserow: try .set_value($f, (my $ = 2))
melezhik if I have OK without quotes
moritz melezhik: that's the problem with OK+ again
melezhik strange
moritz melezhik: that matches OK, OKK, OKKK etc.
melezhik: the + only applies to the last token, which is the K
(so here, it matches OK, OK 9 K, OK 9 K 5 K etc.) 15:38
you can observe this by matching without the ~ part
avuserow moritz, thanks, that works.
moritz m: say say '{ OK 9 OK }' ~~ /:s [ OK + % \d ] /
camelia rakudo-moar f2df2c: OUTPUT«｢OK ｣␤True␤»
melezhik yes, but can't still understand why + prevent in succeed , this one is at least one K which is true ...
15:38 gfldex joined 15:39 shicheng joined
Ven moritz: what do you say say?:) 15:39
moritz it doesn't match the whole string inside the curlies
15:39 Zoffix3 joined
moritz Ven: whatever I want, in every way way! 15:39
Zoffix3 Are goroutines like perl 6 promises? I'm preparing a talk and wanna know what to answer if someone asks... 15:40
15:40 shicheng left
gfldex Zoffix3: no, there are more like processes in erlang (IIRC) 15:40
melezhik mortiz: thanks for help, need to rethink all of this ))) 15:41
Zoffix3 hm /me doesn't know earlang.
gfldex you don't know go either, so it's all good :->
moritz Zoffix3, gfldex: I think they share some similiarities; they are both threads managed by a thread pool 15:42
Zoffix3 Well, I'm reading about them and I see they're concurrent and not parallel, but if you increase GOMAXPROCS then they will use more cores and so I see the difference blur between P6 promises
gfldex m: enum Collision(<Str Int>); my %c{Collision}; %c{Str}++; dd %c; my Str $s; 15:43
camelia rakudo-moar f2df2c: OUTPUT«===SORRY!===␤This type cannot box a native string: P6opaque, Collision␤»
gfldex should that ^^^ work or is it just LTA?
moritz gfldex: this roughly the same as 15:44
m: my class Str { }; say "foo"
camelia rakudo-moar f2df2c: OUTPUT«===SORRY!===␤This type cannot box a native string: P6opaque, Str␤»
15:45 elohmrow joined
moritz LTA, well-known, and kinda "shoot your own foot" 15:45
Zoffix3 Is it? The hash is typed, so it shouldn't be stringifying things.
m: enum Collision(<Str Int>); my %c{Collision}; %c{Str}++;
camelia ( no output )
dj_goku is currently learning elixir at work.
Zoffix3 Ah, nevemind
moritz Zoffix3: the enum Collision locally overrides the Int and Str types
gfldex m: enum Collision(<Collision::Str Collision::Int>); my %c{Collision}; dd %c; 15:46
camelia rakudo-moar f2df2c: OUTPUT«Hash[Any,Collision] %c = (my Any %{Collision} = )␤»
gfldex m: enum Collision(<Collision::Str Collision::Int>); my %c{Collision}; dd %c.enums.keys;
camelia rakudo-moar f2df2c: OUTPUT«Method 'enums' not found for invocant of class 'Hash[Any,Collision]'␤ in block <unit> at <tmp> line 1␤␤»
moritz so the compiler tries to look them up to box stuff into, and gets something that cannot (un)box
gfldex m: enum Collision(<Collision::Str Collision::Int>); my %c{Collision}; dd %c.keyof.enums.keys;
camelia rakudo-moar f2df2c: OUTPUT«("Collision::Str", "Collision::Int").Seq␤»
gfldex m: enum Collision(<Collision::Str Collision::Int>); my %c{Collision}; %c{Collision::Str}++; dd %c;
camelia rakudo-moar f2df2c: OUTPUT«Could not find symbol '&Str'␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
gfldex should quallified enum keys work?
moritz uhm, not really :-) 15:47
tbrowder ugexe: here is a simple Build.pm file: 15:48
elohmrow The README for rakudo-star-2016.07 says: "Please note that this release of Rakudo Star is not fully functional with the JVM backend from the Rakudo compiler. Please use the MoarVM backend only." <-- what is the most recent version I can rely on, if I super-really want to use the jvm backend, please? 15:49
tbrowder www.irccloud.com/pastebin/7EHwoo17/Build.pm
using that file, the only way a msg gets to the user is by using --debug. -verbose doesn't work. suggestions? 15:50
Zoffix3 elohmrow: 2016.08.1 but it's not really fully functional there either. 15:51
stmuk_ elohmrow: that release probably does basically work in terms of simple code examples
Zoffix3 elohmrow: rakudo.org/downloads/rakudo/
15:51 kaare_ joined
ugexe 1: you dont need `use Panda::Builder` or `is Panda::Builder`. secondly i guess you could print to $*ERR 15:51
tbrowder but what is the intent of the verbosity options then? 15:52
Zoffix3 elohmrow: note that you can use our bot camelia with j trigger to run code on JVM: /msg camelia j: say 'hello'
15:53 pierre_ left
elohmrow Zoffix3 stmuk_: perhaps I have an env problem then. I can build on Moar, but on JVM fails every time, on every machine I tried. 15:53
stmuk_ I think I could only get it to work with OpenJDK 1.6
ugexe Build.pm is not specced and will go away, so its output is not a major concern 15:54
15:54 obfusk left, ufobat left
stmuk_ I successfully built a week or two back using that version from debian stable 15:54
15:54 obfusk joined
ugexe the verbosity flags are used in many other places 15:54
elohmrow I have tried J7 and J8 on OSX 10.10.5 (2 different machines), CentOS 7, and Ubuntu 16.04
15:55 zacts joined
stmuk_ I couldn't get OpenJDK 1.7 (7) to work 15:55
Zoffix3 elohmrow: what's the error output that you get?
tbrowder um, printing to $*ERR doesn't work, either. 15:56
elohmrow Zoffix3: well, if i do get an error, it ends up being about lack of RAM, and an error file. so if i fix up the jvm opts to give lots-o-ram, then i get a blank error 2 with no output file
dj_goku Just at first glance a lot of the Perl6 concurrency stuff feels very similar in erlang/elixir. But I have very little knowledge of both.
Zoffix3 no idea
dj_goku Zoffix3: what kind of talk are you doing? 15:57
ugexe dont count on Build.pm for delivering messages to the user
Zoffix3 dj_goku: it's a short lightning talk: "Perl 6: What Programming In The Future Is Like?"
tbrowder ugexe: then what will be the future, any ideas? the "run ..." thing you showed earlier? is that the way to go? or use the META6 hook attribute?
Zoffix3 dj_goku: where I convert a 7-line password hashing program that runs in 63 seconds, to one that runs in 3.1 seconds, to one that's just one line of code.
ugexe i dont know. long ago zef had a hook implementation that simply ran a script ala `hooks/$pre-or-post_$phase.pl6` 15:58
stmuk_ elohmrow: I didn't see that with 1.7 .. I can't remember the errors I saw .. but it does work with 1.6 (or did a week or so ago when I tried it)
dj_goku Zoffix3: nice! any new modules coming out from the talk?
ugexe whatever the solution is should allow the user to run it directly 15:59
elohmrow so, stmuk_: you got 08.1 or 07.1 working with java 6? what OS? <-- debian, you say? v?
Zoffix3 dj_goku: nope, it's all raw, core Perl 6.
tbrowder ugexe: thanks, i'll keep experimenting then
dj_goku Zoffix3: cool!
16:00 cdg joined
stmuk_ elohmrow: I got a version close to 08.1 to work .. one sec I will see if I have more details 16:00
elohmrow stmuk_: thanks! 16:01
dj_goku Zoffix3: where is the lightning talk at?
stmuk_ elohmrow: I don't have access to the laptop :( but I'm sure it was OpenJDK 1.6 (6) 16:02
Zoffix3 dj_goku: Toronto. There also should be a live online feed and (later) a YouTube vid: www.meetup.com/Toronto-Perl-Mongers...233588645/
dj_goku cool. are you in toronto?
Zoffix3 Brampton, about a 40-minute drive from Toronto.
dj_goku ahh 16:03
elohmrow stmuk_: no problem, you've been a big help with this info alone. do you recall roughly the debian version?
16:03 eliasr left 16:04 Xliff_zzzzz left 16:08 bob777 left, grondilu joined
grondilu hello #perl6 16:08
stmuk_ elohmrow: I think I was wrong and I used 1.7 (I recall I had to downgrade a version for it to work) 16:10
elohmrow: github.com/rakudo/rakudo/blob/nom/README.md
elohmrow: the instructions there refer to JDK 1.7 and also have a possible fix for your out of memory error
elohmrow stmuk_: the memory error i'm not very concerned about, as i can make it go away by changing the -X options 16:11
stmuk_ probably was openjdk-7-jdk:amd64 7u111-2.6.7-1~deb8u1 16:12
I can confirm in a day or two
16:12 Zoffix3 left, zacts left
elohmrow stmuk_: but, i will admit i had not read that part of the README :( <-- i only read the Configure line and then ran make, make test. but then make install always died 16:12
stmuk_: armed with the info from your experience, i will try once more ... thanks again! 16:13
stmuk_ I'm not even sure I ran "make install" I probably ran some code "in place"
also see
irclog.perlgeek.de/perl6-dev/2016-...i_12688822
"rakudo-j compiles but installing fails due to a precomp related error" 16:14
elohmrow stmuk_: ^^ very useful - maybe an older version of rakudo is better then ... 2016.04 seems like the one before the ones i tried? 16:17
so i downgrade java to 7, fiddle with some X-opts through the code, and use a slightly older rakudo, might be the trick to getting something that works right now 16:18
stmuk_ elohmrow: I doubt that helps .. the JVM backend has had problems for a while
16:20 Ven left
stmuk_ I mean I don't know if older versions are better .. the rest of your plan is sound 16:20
16:20 domidumont left
elohmrow stmuk_: yeah, and yest i already tried for ex: install/bin/nqp-j <-- mucking around in there for memory stuff. but then i get sad --> note slides 16-19, and thene specially 20, here: www.slideshare.net/tokuhirom/perl6-meets-jvm <-- seems to have a very different experience than i have ;) 16:25
geekosaur mostly you're trading off different bugs; the jvm backend has never been stable
elohmrow slide 20 is a vicious, vicious lie
[Coke] elohmrow: there is no fully functional JVM version. 16:26
not since (and including) the Christmas release in 2015.12 16:27
elohmrow [Coke]: but can i at least install with ?
i am all for using the version i used last year (august-ish) ... which is the last time i tried p6 :(
i'm just trying to get it to work. nothing fancy yet
[Coke] elohmrow: you can install any version you want from source. 16:28
you'll have to experiment yourself to find the one that works for you, though.
I would recommend using the latest version and opening tickets for any issues you find.
but our release process doesn't test to make sure it even compiles on the JVM at this point. 16:29
elohmrow [Coke] only reason i had not done this yet is everywhere i looked, it seemed easy to do, so i began to suspect a deficiency between KAC
stmuk_ elohmrow: one of the probs is that the JVM version didn't fully survive a radical list refactoring (GLR) so even an old functioning version would be different to modern perl 6
[Coke] KAC?
elohmrow kayboard and chair, sorry
[Coke] ah, PEBKAC. :)
no, it's not you. It's us. :0
elohmrow yeah, but i left out the peanut butter ;-P
[Coke] er, :)
I will fire up a build right now and see how HEAD is doing. 16:30
16:30 nadim_ joined
elohmrow ok well, this is all very good info. i will give one more good try, then i will give up for now. 16:30
16:31 nadim left
[Coke] ok. if you're will to be a canary, bug reports on what's not working make it easier to get it running again. 16:31
elohmrow thanks again all, especially stmuk_ Zoffix3 [Coke]
will do. i have pretty good notes on what i have tried, and on different machines.
[Coke] I can tell you that back when it was closer to working, it was missing most of the unicode work. pretty sure concurrency didn't land there. 16:33
elohmrow wanders off for a bit to try once more 16:34
16:52 lizmat joined 17:00 mcmillhj left 17:03 mcmillhj joined
tbrowder ugexe: rooting around I see the zef client (the Build class in Build.pm I believe) is supposed to have a logger. can you, or anyone say how to use it? maybe that's the tie to verbosity 17:06
stmuk_ Stage jast : 72.799 17:07
17:08 mvorg left
[Coke] Stage jast : 45.793 17:14
(having just got to the same spot.) 17:15
stmuk_ you will probably overtake! 17:16
[Coke] make test has a lot of failures (mostly nativecall) 17:21
one or two non nativecall.
REPL is busted.
timotimo uh oh :( 17:22
but that's just java?
er, that came out wrong
[Coke] that's what I'm testing, yes.
timotimo only on the java backend, yes?
[Coke] m: / "asdfadsf" / ~~ 'asdf'
camelia rakudo-moar f2df2c: OUTPUT«Regex object coerced to string (please use .gist or .perl to do that) in block <unit> at <tmp> line 1␤»
[Coke] m: (/ "asdfadsf" / ~~ 'asdf').perl
camelia rakudo-moar f2df2c: OUTPUT«Regex object coerced to string (please use .gist or .perl to do that) in block <unit> at <tmp> line 1␤»
[Coke] hee 17:23
m: 'adsfasdf' ~~ / 'asdf' /
camelia ( no output )
[Coke] so, jvm's issue there is that it can't give you the nice error message when you do something dumb. :)
grondilu a very basic example of NativeCall use with a typedef-ed structure: gist.github.com/grondilu/dc60a8f75...015505302a
I thought it could be a useful example to add for instance in github.com/jnthn/zavolaj/tree/master/examples 17:25
17:25 tomboy65 left
timotimo we still don't have an easy way to put an array of things with a given length in-line into a CStruct 17:25
grondilu you mean that for instance with something like: struct vector_t { double coeff[3]; } ? 17:27
timotimo yup
grondilu that's regrettable.
timotimo and we also don't yet have arrays of in-lined CStructs
that's also qiute regrettable
stmuk_ make install does work on JVM so that's been fixed in the last 2 mon 17:29
grondilu on the other hand, isn't that only a problem if we need to access the members of the CStruct from the Perl 6 code? Usually only the C code does it, it's often opaque to libraries.
timotimo hmm. 17:30
of course you can create an array of 8bit ints that's the right size
stmuk_ I get an impressive traceback trying to install panda 17:36
17:37 domidumont joined 17:38 webstrand left
bartolin fwiw, I'm still running my daily spectest for rakudo-j. apart from one failing test file (which runs fine standalone) there are no failures. (but there are quite some fudged tests) 17:40
github.com/usev6/perl6-roast-data/...ummary.out
elohmrow doesn't work for me yet, stmuk_: == Installing modules for JVM cd modules/panda && sh -c "PATH=/home/ghost/rakudo-star-2016.07/install/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games /home/ghost/rakudo-star-2016.07/install/bin/perl6-j bootstrap.pl" ^@==> Bootstrapping Panda ^@^@^@^@Cannot unbox a type object in result (gen/jvm/stage2/QAST.nqp) ... 17:42
bartolin I've seen some strange failures recently (when all test files starting with integration/advent2014-something started to fail), but those seemed to be due to an OOM for the eval-server
[Coke] bartolin: are you not running 'make test' ?
elohmrow I did not this time, because it takes approximately as much time as the universe has existed 17:43
bartolin [Coke]: no, I don't. I'm really only running the spectest to keep roast up to date wrt rakudo-j
(in order to make it easier to start working on the JVM backend)
elohmrow stmuk_: <-- same-ish as what i pasted? 17:44
stmuk_ elohmrow: yes I got "Cannot unbox.." too
[Coke] theoretically, 'make test' is more core than 'make spectest' 17:45
elohmrow stmuk_: then i get at the end, in (gen/jvm/main.nqp) The spawned process exited unsuccessfully (exit code: 1) in sub MAIN at bootstrap.pl line 68 in block <unit> at bootstrap.pl line 9 Makefile:60: recipe for target 'modules-install-j' failed make: *** [modules-install-j] Error 1 17:46
17:46 kurahaupo left
elohmrow at least it is consistent - i've gotten exactly this under java 7, 8 and 4 different machines on 3 OSs 17:46
harmil_wk Is there a module of common mathematical constants in the works? I know we define a few globally, but there's a pretty big list including those at en.wikipedia.org/wiki/Mathematical..._constants that I could throw into a module... 17:52
17:53 pierre_ joined 17:55 Wiertek joined 17:58 pierre_ left, MasterDuke joined 17:59 Piotr joined 18:03 Wiertek left, MasterDuke left 18:05 Piotr left
timotimo s: Array, permutations 18:05
SourceBaby timotimo, Something's wrong: ␤ERR: Too few positionals passed; expected 1 argument but got 0␤ in block <unit> at -e line 6␤␤
timotimo will i ever learn this? 18:06
s: Array, &permutations
SourceBaby timotimo, Something's wrong: ␤ERR: Cannot resolve caller sourcery(Array, Sub); none of these signatures match:␤ ($thing, Str:D $method, Capture $c)␤ ($thing, Str:D $method)␤ (&code)␤ (&code, Capture $c)␤ in block <unit> at -e line 6␤␤
timotimo %)
s: permutations
SourceBaby timotimo, Something's wrong: ␤ERR: Too few positionals passed; expected 1 argument but got 0␤ in block <unit> at -e line 6␤␤
timotimo s: &permutations
SourceBaby timotimo, Sauce is at github.com/rakudo/rakudo/blob/f2df...rs.pm#L663
[Coke] bartolin++ running the java stuff. 18:07
18:08 cyphase left
masak today's autopun spotting (actually 2 of them): twitter.com/methode/status/768337146780475392 twitter.com/RuthanneReid/status/77...8764889092 18:09
stmuk_ zef errors with "===SORRY!=== Serialization Error: could not locate static code ref for closure" on JVM 18:10
ufo can be hacked up to work though :)
18:13 cyphase joined, n1lp7r left
grondilu a more elaborate example than above: paste.debian.net/794867/ 18:14
interestingly enough, we can actually just define C<class CStruct is repr('CStruct') {}> and use it for any opaque type defined as a pointer to a structure. 18:15
(or rather as Pointer[CStruct]) 18:16
timotimo i find it a bit muddy to have CStruct be a pointer, but Pointer[CStruct] also .. or is that a pointe-to-pointer?
grondilu no CStruct is a struct, not a pointer to a struct. I may have expressed it poorly above. 18:17
timotimo ah 18:18
grondilu see lines 48 and 55 in above code.
timotimo the only difference it makes if you define the structure properly or not is when you .new it
oh i didn't notice the link
did you know about Inline::C?
it'll automate the compiling step for you
and you can then have the C code in-line in your perl6 code as sub bodies 18:19
grondilu I knew it existed, but is it really working?
I mean that would look too good to be true.
oh wait, I misunderstood.
timotimo it worked at one point
haven't tested in a logn time 18:20
grondilu it creates the Perl 6 bindings?
timotimo my sub a_plus_b( Int $a, Int $b ) is inline('C') returns Int {' 18:21
DLLEXPORT int a_plus_b (int a, int b) {
this is how you use it. that should explain everything, hopefully
grondilu I'm more interested in something that takes a header file and creates the Perl 6 bindings. 18:22
timotimo that's gptrixie
find it on modules.perl6.org and give it a shot
18:23 CptnK joined
CptnK Is it possible from within sub EXPORT {} to get the file of where I'm exporting to? I'm tring to die() with a helpful error 18:24
And I've tried callframe().file with a bunch of values to callframe, but I can either get the file where the sub EXPORT is or Rakudo's guts...
timotimo right, the importation happens at compile time, so all you'll find on the stack is the compiler 18:25
CptnK Ahh... ok
elohmrow stmuk_: correction - i did not pay enough attention. we do get a successful make install; what fails is the bootstrapping panda step. i may or not need panda right now. so i *do* have perl6 on jvm right now. oops. sorry for the false alarm. 18:27
18:28 setty1 joined
stmuk_ elohmrow: I guessed that! in short the JVM "mostly" works (with fudged tests) but not functional enough yet to install either panda or zef 18:29
elohmrow exactly. 18:30
nine $*W.current_file? 18:33
CptnK: ^^^
CptnK: note that I'm not sure if this is actually stable API
timotimo definitely not stable API 18:35
CptnK nine++ hah, yeah it worked :)
gfldex m: class Str {}; say CORE::Str; 18:36
camelia rakudo-moar f2df2c: OUTPUT«===SORRY!===␤This type cannot box a native string: P6opaque, Str␤»
gfldex m: class Str {}; print CORE::Str;
camelia rakudo-moar f2df2c: OUTPUT«===SORRY!===␤This type cannot box a native string: P6opaque, Str␤»
gfldex m: class Str {}; print SETTING::Str;
camelia rakudo-moar f2df2c: OUTPUT«===SORRY!===␤This type cannot box a native string: P6opaque, Str␤»
elohmrow stmuk_: and for completeness, same here for zef: ../rakudo-star-2016.07/perl6 -Ilib bin/zef install . ^@^@^@^@^@===SORRY!=== Serialization Error: could not locate static code ref for closure 18:41
so we are all in the same place, i guess ;) 18:42
[Coke] bartolin: don't need to add [MOARVM] to tickets. tagging the VM is fine. 18:44
bartolin [Coke]: thanks, will (try to) remember :-) 18:45
18:45 Praise joined
[Coke] also, moarvm queue is mainly for stuff that is already identified as internal - anything that presents in the rakudo layer, it's fine to keep as an RT. 18:46
18:47 darutoko left, cdg left
grondilu discovered the `tabular` vim plugin. github.com/godlygeek/tabular Neat. 18:50
18:52 TEttinger joined
nine nadim_: irclog.perlgeek.de/perl6-toolchain/...i_13112051 18:54
CptnK m: say $*SPEC.splitdir: 'foo\bar\ber'; 19:00
camelia rakudo-moar f2df2c: OUTPUT«(foo\bar\ber)␤»
CptnK That kinda sucks
19:01 CIAvash left
timotimo you can just use the spec that has \es 19:01
CptnK timotimo: what do you mean?
timotimo m: say IO::Spec::Win32.splitdir: 'foo\bar\ber' 19:02
camelia rakudo-moar f2df2c: OUTPUT«(foo bar ber)␤»
timotimo m: say IO::Spec::Win32.splitdir: 'C:\foo\bar\ber'
camelia rakudo-moar f2df2c: OUTPUT«(C: foo bar ber)␤»
CptnK m: say IO::Spec::Win32.splitdir: 'foo/bar/ber'
camelia rakudo-moar f2df2c: OUTPUT«(foo bar ber)␤»
timotimo hm, not sure how handling drive letters works, actually
19:03 espadrine left
dalek osystem: 2c47dd0 | (Zoffix Znet)++ | META.list:
Add SPEC::Func to ecosystem

Import $*SPEC methods as functions: github.com/zoffixznet/perl6-SPEC-Func		 19:13
19:16 girafe joined
gfldex m: IO::Path::Win32.new('C:\a\b').volume.say 19:16
camelia rakudo-moar f2df2c: OUTPUT«C:␤»
19:17 ufobat joined
timotimo ah 19:19
thanks, gfldex :)
19:27 nadim_ left 19:29 nadim_ joined 19:33 domidumont left, canopus left 19:42 canopus joined 19:43 nadim_ left 19:45 labster joined 19:50 CptnK left 19:52 zakharyas joined
moritz observer.com/2016/08/not-a-drill-se...eep-space/ wohoo! 19:55
19:59 cdg joined
mspo moritz: all of those screen savers paid off? 20:02
grondilu inserts "alien" meme
lizmat starts on this week's Perl 6 Weekly 20:03
vcv lizmat++ love that these are on a monday, an otherwise dreadful day. 20:04
timotimo derp space! 20:05
grondilu: "alien meme"?
i've seen a project where you'd name branches "timo/foobar" and when a branch is named that, "timo" would be allowed to force-push 20:06
but i find code ownership slightly problematic
grondilu timotimo: rather "aliens" 20:08
20:08 MasterDuke joined 20:10 sufrostico joined
El_Che lwp only has timeslots of 60, 120 and 240 minutes (or do I read it wrong)? act.yapc.eu/lpw2016/ 20:12
timotimo the aliens meme is alien to me .. oh never mind, i've seen it! it has that bad-ass female protagonist in it! 20:13
20:16 wamba left 20:18 cdg left 20:19 cdg joined 20:32 ufobat left, elohmrow left
grondilu enum (Foo => 1, Bar => Foo); can't be done apparently. Yet it exists in C. 20:37
timotimo :D :D 20:38
could be done with a bit of parsing trickery, but ... bleeehhh :)
if you can't do it with the meatmodel, there's always the potatomodel! 20:39
20:40 leont joined 20:51 canopus left, r_i_d left, r_i_d joined
grondilu currently I'm looking at the Vulkan API, I saw this kind of enum there. 20:51
I think if it's possible and apparently done in C, it should also be done in Perl 6. 20:52
konobi vulkan? 20:54
grondilu a new 3D graphics API
20:54 zacts joined 20:55 skids left
grondilu it's weird because I vaguely recall it was possible to do something like that in Perl 6, but not with enums, rather with an other data structure. Can't quite remember which. 20:55
20:56 canopus joined
konobi well, in C, doesn't Foo just become a pointer, so therefore an address which is just an Int at the end of the day? 20:56
grondilu has no idea
konobi yeah... i think that basically how it works out 20:57
20:57 r_i_d left, r_i_d joined
arnsholt An enum in C is an int, IIRC 20:58
grondilu Is it not quite the same as writing this: 20:59
m: my ($a = 1, $b = $a);
camelia ( no output )
grondilu m: my ($a = 1, $b = $a); say $b
camelia rakudo-moar f2df2c: OUTPUT«(Any)␤»
grondilu hum
grondilu did not expect that
konobi so more like enum (Foo => 1, Bar => \Foo)
grondilu m: enum (Foo => 1, Bar => \Foo); say Bar 21:00
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared name:␤ Foo used at line 1␤␤»
21:00 r_i_d left
konobi (in concept, rather than as syntax ^_^) 21:00
grondilu it's killing me that I can't remember where I've seen something like that. 21:01
konobi \self.Foo might be more appropriate
grondilu maybe it was in constant definition. Is it possible to define several constants in a single statement?
oh wait, it might have been in Perl 5 21:02
21:02 r_i_d joined
grondilu like: use constant { Foo => 1, Bar => Foo } 21:03
^that's apparently valid Perl 5 code.
ahh no, it does not bode with 'use strict;' 21:04
timotimo so at least we don't regress in that regard :)
konobi is there a way to access the current enum from it's expression?
(it's own expression) 21:05
=0)
grondilu doubts it 21:06
konobi timotimo? 21:07
21:08 mcmillhj left
konobi well, expressions should be able to use reflection on themselves, iirc 21:08
21:08 sufrostico left 21:10 acrussell left
grondilu they do in signatures, don't they? 21:11
m: say :($a, $b where $a == $b)
camelia rakudo-moar f2df2c: OUTPUT«($a, $b where { ... })␤»
grondilu thought that's definitely not the case I had in mind
konobi well, nested expression =0) 21:13
m: say enum (Foo => 1, Bar => :Foo) 21:14
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Type error in enum. Got 'Pair' Expected: 'Int'␤at <tmp>:1␤------> 3say enum (Foo => 1, Bar => :Foo)7⏏5<EOL>␤»
grondilu I'm sure it was a bare word. I was using one in a statement while it was defined earlier in the same statement.
maybe it was an enum and there was a regression
21:15 sufrostico joined
grondilu looks in some old code of his 21:17
grondilu finds an example and it was indeed with constants in Perl 5 21:18
21:18 kyclark_ joined, kyclark_ left 21:19 espadrine joined
konobi m: say enum (Foo => 1, Bar => $:Foo) 21:19
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Cannot use placeholder parameter $:Foo in the mainline␤at <tmp>:1␤------> 3say enum (Foo => 1, Bar => $:Foo7⏏5)␤»
konobi huh, i'd have thought that should have worked
m: say enum (Foo => 1, Bar => \$:Foo) 21:20
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5===␤Cannot use placeholder parameter $:Foo in the mainline␤at <tmp>:1␤------> 3say enum (Foo => 1, Bar => \$:Foo7⏏5)␤Other potential difficulties:␤ To pass an array, hash or sub to a function in Perl 6, just pass it as is.…»
timotimo i expect the actual enum object is only composed when we reach the end of the parsing
grondilu I had no idea this code I had written was not "strict" compliant.
timotimo i'd assume it won't even be stubbed after we've reached the name of the enum
konobi but twigils should setting that to be somewhat lazily evaluated, no? 21:21
21:21 kyclark joined, zacts left, zacts_pi joined
pmurias konobi: hi 21:21
kyclark Given a list of pairs/lists, how can I convert to a hash?
konobi m: say enum (Foo => 1, Bar => $*Foo) 21:22
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Type error in enum. Got 'Failure' Expected: 'Int'␤at <tmp>:1␤------> 3say enum (Foo => 1, Bar => $*Foo)7⏏5<EOL>␤»
kyclark my @list = (("name", "George"), ("age", 23))
pmurias konobi: I'm working on packaging up nqp-js into proper npm-style packages
21:22 kaare_ left
konobi pmurias: oh, nice 21:22
kyclark I thought about something like ".map((=>))" but that doesn't work
konobi kyclark: check the list class
pmurias konobi: my current plan is it have seperate nqp-js-runtime and nqp-js-compiled-to-js packages 21:23
21:23 Actualeyes left
timotimo konobi: no, twigils will not make that happen for you 21:23
konobi yeah... makes sense
timotimo: ah
21:24 mcmillhj joined, cdg left
konobi m: say enum (Foo => 1, Bar => $.Foo) 21:24
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable $.Foo used where no 'self' is available␤at <tmp>:1␤------> 3say enum (Foo => 1, Bar => $.Foo7⏏5)␤ expecting any of:␤ term␤»
konobi =0)
yup 21:25
timotimo kyclark: mapping &[=>] onto a list of pairs like that is problematic because => expects two arguments, but it'll get a single list as its argument
kyclark: what you can do, however, is flatten the list first to get its inner lists as a long list of pairs 21:26
konobi pmurias: yeah, having the js-runtime separately helps focus an improving the JS implementation by itself outside of work that changes from NQP bring
timotimo otherwise i think we do have some method that does what you want. i could be wrong, of course
21:26 rindolf left
timotimo m: my @list = (("name", "George"), ("age", 23)); say @list>>.&[=>]; # i wonder 21:26
camelia rakudo-moar f2df2c: OUTPUT«Too few positionals passed; expected 2 arguments but got 1␤ in block <unit> at <tmp> line 1␤␤»
timotimo m: my @list = (("name", "George"), ("age", 23)); say @list.map(*.[0] => *.[1]).perl 21:27
camelia rakudo-moar f2df2c: OUTPUT«(:name(23),).Seq␤»
timotimo huh?
hah
m: my @list = (("name", "George"), ("age", 23)); say @list.perl; say @list.map(*.[0] => *.[1]).perl
camelia rakudo-moar f2df2c: OUTPUT«[("name", "George"), ("age", 23)]␤(:name(23),).Seq␤»
timotimo the star is a new argument each time, of course 21:28
m: my @list = (("name", "George"), ("age", 23)); say @list.map({ $^a.[0] => $^a.[1] }).perl
camelia rakudo-moar f2df2c: OUTPUT«(:name("George"), :age(23)).Seq␤»
timotimo there we go.
21:28 Actualeyes joined
timotimo m: my @list = (("name", "George"), ("age", 23)); say @%list 21:28
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '%list' is not declared. Did you mean any of these?␤ &list␤ @list␤ &last␤ List␤␤at <tmp>:1␤------> 3(("name", "George"), ("age", 23)); say @7⏏5%list␤»
timotimo m: my @list = (("name", "George"), ("age", 23)); say %@list # typod
camelia rakudo-moar f2df2c: OUTPUT«{name George => (age 23)}␤»
timotimo that's definitely not it :D
m: my @list = (("name", "George"), ("age", 23)); say @list.pairup 21:29
camelia rakudo-moar f2df2c: OUTPUT«((name George) => (age 23))␤»
timotimo also not it
kyclark Well, that will do for now. Seems like it should be more elegant. Thanks!
21:29 zacts_pi left 21:30 zakharyas left
konobi m: my @list = (("name", "George"), ("age", 23)); say %@list.antiparis 21:31
camelia rakudo-moar f2df2c: OUTPUT«Method 'antiparis' not found for invocant of class 'Hash'␤ in block <unit> at <tmp> line 1␤␤»
konobi m: my @list = (("name", "George"), ("age", 23)); say %@list.antipairs
camelia rakudo-moar f2df2c: OUTPUT«((age 23) => name George)␤»
konobi m: my @list = (("name", "George"), ("age", 23)); say %@list.kv
camelia rakudo-moar f2df2c: OUTPUT«(name George (age 23))␤»
21:31 MasterDuke left
konobi m: my @list = (("name", "George"), ("age", 23)); say @list.kv 21:31
camelia rakudo-moar f2df2c: OUTPUT«(0 (name George) 1 (age 23))␤»
konobi heh
kyclark gotta jet, but will reconnect at home. let me know if you find something cooler. 21:32
21:32 kyclark left
konobi m: my @list = (("name", "George"), ("age", 23)); say @list.flat.kv 21:32
camelia rakudo-moar f2df2c: OUTPUT«(0 (name George) 1 (age 23))␤»
konobi m: my @list = (("name", "George"), ("age", 23)); say %@list.flat
camelia rakudo-moar f2df2c: OUTPUT«(name George => (age 23))␤»
21:33 zacts joined
konobi k... time to recompile rakudo it seesm 21:33
avuserow m: my @list = (("name", "George"), ("age", 23)); say @list.>>pairup
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Missing dot on method call␤at <tmp>:1␤------> 3", "George"), ("age", 23)); say @list.>>7⏏5pairup␤ expecting any of:␤ postfix␤»
avuserow m: my @list = (("name", "George"), ("age", 23)); say @list>>.pairup
camelia rakudo-moar f2df2c: OUTPUT«((name => George) (age => 23))␤»
21:37 pmurias left
timotimo avuserow: hah! short and sweet. very good 21:37
m: my @list = (("name", "George"), ("age", 23)); say @list.flat.pairup
camelia rakudo-moar f2df2c: OUTPUT«((name George) => (age 23))␤»
21:37 cdg joined
timotimo m: my @list = (("name", "George"), ("age", 23)); say @list>>.Slip.pairup 21:37
camelia rakudo-moar f2df2c: OUTPUT«((name George) => (age 23))␤»
timotimo hmm.
m: my @list = (("name", "George"), ("age", 23)); say @list>>.list.pairup
camelia rakudo-moar f2df2c: OUTPUT«((name George) => (age 23))␤»
avuserow I find myself using >> a fair bit in p6 when processing collections like this. it's great to have. :) 21:38
21:38 mvorg joined
timotimo yup, it really, really, really is 21:38
gfldex m: my @list = (("name", "George"), ("age", 23)); my @pairs = do for @list -> [$k, $v] { $k => $v }; dd @pairs;
camelia rakudo-moar f2df2c: OUTPUT«Array @pairs = [:name("George"), :age(23)]␤»
timotimo to the point where i want every language to have that now
gfldex ». gets rid of lots of simple loops 21:39
tbrowder ugexe: I see the change you made, and what is needed to get the value into zef's guts. I just tried a few changes and think, at least for build, we can test for the actual value of $DEBUG (maybe change its name) instead of True or not. The ENV value tests as a Str and regex/match statements work well. For example I'll show some code fragments in a gist in 21:41
a few minutes.
21:41 leont left
lizmat and a new Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2016/08/29/...om-zagreb/ 21:41
21:42 leont joined 21:43 mvorg is now known as margeas
tbrowder ugexe: here are the changed lines in this gist "gist.github.com/tbrowder/9ea0d5acd...781939fdb" 21:47
timotimo lizmat: one of the links says "how to parse CSS" but the post is actually about RSS
imagine my confusion when people were recommending using libxml2
i thought "um, that won't work" :D 21:48
21:48 canopus left
lizmat timotimo++ # fixed 21:48
timotimo thanks :)
tbrowder ugexe: notice the value of $DEBUG is tested two different ways, as needed: truth or value
ugexe tbrowder: those aren't neccesary. it will output with either --verbose or --debug, because INFO/VERBOSE/DEBUG/etc is part of the VERBOSITY enum 21:49
tbrowder I know, but I don't want all the info from DEBUG for VERBOSE type items from user space, thus I have the option to test for values instead of truth. 21:50
Maybe I don't understand, but the ZEF_BUILDPM_DEBUG env var is treated as a string (which can be '') so it all just works in my short test. 21:51
21:52 leont left
ugexe m: %*ENV<foo> = True; %*ENV<bar> = False; say %*ENV<foo>; say %*ENV<bar> 21:53
camelia rakudo-moar f2df2c: OUTPUT«True␤False␤»
timotimo mst: Opan GangnaMSTyle!
21:54 mcmillhj left
tbrowder I know I'm not explaining this clearly, but the $verbosity takes on a value of the enum, and when it is placed in an environment variable, it can take on many values, not just true or false. The original statement assigned truth where my change assigned a value to $verbosity. 21:55
21:56 canopus joined
ugexe the ::Service namespace doesn't get to work with the logger verbosity levels, that the job of the adapter. the fact $DEBUG is even there is just an artifact from when it was in Zef::Client 21:56
i.e. verbosity output should be handled in Zef::Build, not Zef::Service::Shell::Build 21:58
Zef::Build will have access to the actual verbosity enum/types 21:59
22:01 canopus left
tbrowder I couldn't see how to get the verbosity value in Buid.pm other than the DEBUG env var, but that would be the proper way to do it. 22:02
ugexe you need to clarify between Zef::Build and Zef::Service::Shell::Build 22:03
because your gist says its for Zef::Build but its for Zef::SErvice::Shell::Build
what i'm saying is the $DEBUG bit can probably be moved into Zef::Build from Zef::Service::Shell::Build. Zef::Build will have access to the actual verbosity type, no need to stringify anything 22:04
tbrowder Oh, I was a bit sloppy ref the name--let me look closer.
22:05 mcmillhj joined, sufrostico left 22:06 canopus joined
tbrowder The actual file is "Zef::Service:🐚:Build.pm", I've corrected the gist. 22:09
22:09 leont joined 22:10 mcmillhj left
tbrowder Argh, this unicode stuff is getting out of hand! Anyone else see a shell picture instead of "::Shel"? 22:10
s/Shel/Shell/
22:12 kyclark joined, cyphase left 22:14 Sgeo joined
ugexe yes, i knew thats what you meant. but what needs to happen is to remove $DEBUG from ::Service::Build all together and filter the output in Zef::Build based on the verbosity level 22:16
22:16 zacts left 22:17 cyphase joined
ugexe ::Service::* are all dumb and essentially stand alone. Zef::Build/Test/whatever is the adapter for the various services, and what Zef::Client actually uses 22:17
22:18 skids joined
ugexe in fact it might Just Work if the $DEBUG conditions are all removed / assumed true 22:20
22:21 canopus left 22:24 firstdayonthejob left, margeas left 22:27 buggable left
tbrowder ugexe: so how can i help you to scratch my itch? 22:27
22:27 smls left, buggable joined 22:29 leont left
tbrowder Any Canadian p6 meetings during the period 16-22 Sep? 22:30
22:30 _slade_ joined 22:32 canopus joined, zjlin joined
tbrowder Specifically from Quebec City, Prince Edward Island, Sydney, and Halifax. Yes, a cruise with wife and friends--p6 meeting would be better than being dragged on a shopping tour! 22:33
El_Che I saw zoffix posting about the next toronto meeting, but later: 29 sept 22:35
kyclark What would be the most efficient way to find all overlapping sequences from something like this? 22:36
my @pos = [9..11], [1..5], [3..6], [7..10]
dalek c: 71df199 | gfldex++ | doc/Language/typesystem.pod6:
show how to get hold of type object for enum keys
 22:37
grondilu there is a known algorithm.
kyclark: it'd be tough to figure it out yourself. Look it up.
oh you want *all* overlapping sequences? Might even be harder. 22:38
kyclark It’s an n-squared thing, isn’t it?
22:38 canopus left
grondilu the algorithm I was talking about is for the longuest overlapping sequence. 22:38
IIRC
en.wikipedia.org/wiki/Longest_comm...ce_problem 22:39
kyclark for 0..@list.elems -> $i { for 1..@list.elems -> $j { if $i in $j { … }}
22:44 zjlin left 22:47 mcmillhj joined 22:48 canopus joined 22:51 bjz joined
timotimo we don't have an "in" operator :) 22:51
22:51 NeuralAnomaly joined 22:52 mcmillhj left, NeuralAnomaly left
timotimo and i think it's actually n^3? 22:52
hmm
i'm recognizing my brain is getting tireder than recommended for thinking 22:53
konobi grondilu: you can use a set operation with a sort for that 22:54
22:56 NeuralAnomaly joined
timotimo it looks like we only have an iterated version of permutations, at least the sub version 22:57
grondilu timotimo: iterated as opposed to what? 22:59
I mean how else would you implement it? 23:00
timotimo all-at-once
grondilu is confused
timotimo i.e. an implementation for the push-all method in the iterator
23:00 leont joined
ugexe tbrowder: it should do what you were looking for now 23:01
23:02 mcmillhj joined
gfldex Hash with fixed set of keys with the caveat of polluting the local scope: gist.github.com/gfldex/af2a861fc21...a944146c62 23:06
23:06 bjz left, canopus left 23:10 cdg left
kyclark m: my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred”)}]; put %things<dogs>.elems; for %things<dogs> -> $thing { put $thing<name> } 23:11
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '$thing' is not declared. Did you mean '%things'?␤at <tmp>:1␤------> 3hings<dogs>.elems; for %things<dogs> -> 7⏏5$thing { put $thing<name> }␤»
23:13 canopus joined
kyclark Hmm, I’m not able to show my problem. 23:13
Given a hash that has an array for values, how do I iterate over the values?
my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred")}]
for %things<dogs> -> $thing { put $thing<name> }
That gives me:
Type Array does not support associative indexing.
timotimo m: my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred”)}]; put %things<dogs>.elems; for (%things<dogs>) -> $thing { put $thing<name> } 23:15
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '$thing' is not declared. Did you mean '%things'?␤at <tmp>:1␤------> 3ngs<dogs>.elems; for (%things<dogs>) -> 7⏏5$thing { put $thing<name> }␤»
timotimo looks for non-closed double-quotes
m: my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred”)}]; put %things<dogs>.elems; for (%things<dogs>) { put $_<name> } 23:16
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in double quotes; couldn't find final '"' ␤at <tmp>:1␤------> 3ms; for (%things<dogs>) { put $_<name> }7⏏5<EOL>␤»
timotimo haha
do you see that?
m: ':name("Fred”)'.substr(*-3).comb.>>.uniname.say 23:17
camelia rakudo-moar f2df2c: OUTPUT«(LATIN SMALL LETTER D RIGHT DOUBLE QUOTATION MARK RIGHT PARENTHESIS)␤»
avuserow m: my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred")}]; put %things<dogs>.elems; for %things<dogs>.list -> $thing { put $thing<name> }
camelia rakudo-moar f2df2c: OUTPUT«2␤Bowzer␤Patch␤»
timotimo perl6 does a lot of different combos of smart quotes, but not a regular ("dumb") one and a right "smart" one
avuserow not sure if there's a more graceful way of doing that
23:17 kyclark left
timotimo avuserow: @ in front is short for ".list" at the end 23:17
avuserow m: my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred")}]; put %things<dogs>.elems; for @%things<dogs> -> $thing { put $thing<name> } 23:18
camelia rakudo-moar f2df2c: OUTPUT«2␤Type List does not support associative indexing.␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
avuserow m: my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred")}]; put %things<dogs>.elems; for @(%things<dogs>) -> $thing { put $thing<name> }
camelia rakudo-moar f2df2c: OUTPUT«2␤Bowzer␤Patch␤»
timotimo well, i'll be damned.
m: my %things = dogs => ({:name("Bowzer")}, {:name("Patch")}), people => ({:name("George")}, {:name("Fred")}); put %things<dogs>.elems; for (%things<dogs>) { put $_<name> }
camelia rakudo-moar f2df2c: OUTPUT«2␤Type List does not support associative indexing.␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
timotimo m: my %things = dogs => ({:name("Bowzer")}, {:name("Patch")}), people => ({:name("George")}, {:name("Fred")}); put %things<dogs>.elems; for @(%things<dogs>) { put $_<name> }
camelia rakudo-moar f2df2c: OUTPUT«2␤Bowzer␤Patch␤»
timotimo that doesn't change much, it seems
avuserow I prefer a postfix method call to circumfix here... 23:19
timotimo that's fair
23:20 canopus left 23:21 mcmillhj left
avuserow m: my %things = dogs => [{:name("Bowzer")}, {:name("Patch")}], people => [{:name("George")}, {:name("Fred")}]; put %things<dogs>.elems; my @thing-container = %things<dogs>; for @thing-container -> $thing { put $thing<name> } 23:22
camelia rakudo-moar f2df2c: OUTPUT«2␤Type Array does not support associative indexing.␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
23:23 adu joined
timotimo [Coke]: got any cool war stories to tell about angular 2 yet? :) 23:24
gfldex m: my %things = (dogs => (:name<Bowzer>, :name<Patch>)), (people => (:name<George>, :name<Fred>)); put %things<dogs>.elems; for %things<dogs>.flat -> %h { dd %h<name> }
camelia rakudo-moar f2df2c: OUTPUT«2␤"Bowzer"␤"Patch"␤»
23:24 setty1 left
gfldex m: my %things = (dogs => (:name<Bowzer>, :name<Patch>)), (people => (:name<George>, :name<Fred>)); put %things<dogs>.elems; for %things<dogs> -> \i { dd i } 23:25
camelia rakudo-moar f2df2c: OUTPUT«2␤List %things = $(:name("Bowzer"), :name("Patch"))␤»
gfldex m: my %things = (dogs => (:name<Bowzer>, :name<Patch>)), (people => (:name<George>, :name<Fred>)); put %things<dogs>.elems; for %things<dogs> -> \i { dd i.flat }
camelia rakudo-moar f2df2c: OUTPUT«2␤(:name("Bowzer"), :name("Patch")).Seq␤»
23:27 canopus joined 23:28 mcmillhj joined
gfldex m: my %things = dogs => [:name<Bowzer>, :name<Patch>], people => [:name<George>, :name<Fred>]; dd %things; for %things<dogs>.Seq -> %h { dd %h } 23:30
camelia rakudo-moar f2df2c: OUTPUT«Hash %things = {:dogs($[:name("Bowzer"), :name("Patch")]), :people($[:name("George"), :name("Fred")])}␤:name("Bowzer")␤:name("Patch")␤»
gfldex m: my %things = dogs => [:name<Bowzer>, :name<Patch>], people => [:name<George>, :name<Fred>]; dd %things; for %things<dogs> -> [%h] { dd %h } 23:31
camelia rakudo-moar f2df2c: OUTPUT«Hash %things = {:dogs($[:name("Bowzer"), :name("Patch")]), :people($[:name("George"), :name("Fred")])}␤Too few positionals passed; expected 1 argument but got 0 in sub-signature␤ in block <unit> at <tmp> line 1␤␤»
gfldex i'm confused that the destructuring doesn't work
23:32 xinming left
timotimo because there isn't a single hash in that list 23:32
23:32 xinming joined
timotimo wait ... what ... 23:32
can we have this with a shorter literal please? my eyes cross over when trying to read all that :)
oh, you just have a list of Pair objects in your hash there 23:33
that surely won't destructure with [%h]
skids m: my %things = dogs => [:name<Bowzer>, :name<Patch>], people => [:name<George>, :name<Fred>]; dd %things; for %things<dogs> -> [*%h] { dd %h } # :/ 23:37
camelia rakudo-moar f2df2c: OUTPUT«Hash %things = {:dogs($[:name("Bowzer"), :name("Patch")]), :people($[:name("George"), :name("Fred")])}␤{:name("Patch")}␤»
skids m: my %things = dogs => [:name<Bowzer>, :name2<Patch>], people => [:name<George>, :name<Fred>]; dd %things; for %things<dogs> -> [*%h] { dd %h } # :/ 23:38
camelia rakudo-moar f2df2c: OUTPUT«Hash %things = {:dogs($[:name("Bowzer"), :name2("Patch")]), :people($[:name("George"), :name("Fred")])}␤{:name("Bowzer"), :name2("Patch")}␤»
skids Ah. Yes that does work.
23:40 girafe left 23:47 canopus left 23:53 zoosha_ left, zoosha joined 23:54 canopus joined 23:58 mcmillhj left

Powered by Cro and the Raku® Programming Language.

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

Thank you!