»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by masak on 28 November 2015.
joeschmoe2 m: my $Link = 'www.google.com'; say $Link /msg 00:00
camelia rakudo-moar c27a00: OUTPUT«5===SORRY!5=== Error while compiling /tmp/yA7ERKIWq6␤Undeclared routine:␤ msg used at line 1. Did you mean 'msb'?␤␤»
_nadim So long, and thanks for all the fish 00:01
Zoffix \o
00:01 donaldh left
joeschmoe2 i suck 00:02
diakopter m: ***=*=***=****=**=***=*=****=*=**
camelia ( no output )
joeschmoe2 +camelia m: my $Link = 'www.google.com'; say $Link 00:03
Zoffix joeschmoe2, /msg is an IRC command to send private messages. You use it by typing /msg NICK_TO_SEND_MESSAGE_TO MESSAGE ... and the message you'd send is the same one you're typing here in the channel
/msg camelia m: say 42
^ type that exactly and it'll work :) 00:04
diakopter, hax!
nine Something for the bikeshedding commitee: how should I call CompUnit::RepositoryRegistry's method that takes a string of the form "inst#/path/to/site" and gives you a CompUnit::Repository object?
xjrK I have a class in a .pm6 module... I want to access it in another file. use Class::Whatever should work, no?
marchelzo isn't it kind of inconsistent that 'for ((1, 2), (3, 4), (5, 6)) -> $a, $b { ... }' uses the pairs to populate $a and $b, but 'for (1, 2, 3, 4, 5, 6) -> $a, $b { ... }' uses two elements each time?
Zoffix m: for ((1, 2), (3, 4), (5, 6)) -> $a, $b { say "$a $b" } 00:05
camelia rakudo-moar c27a00: OUTPUT«1 2 3 4␤Too few positionals passed; expected 2 arguments but got 1␤ in block <unit> at /tmp/NP1kzGr44W:1␤␤»
diakopter nine: bikeshed->woodshed
Zoffix m: for |((1, 2), (3, 4), (5, 6)) -> $a, $b { say "$a $b" }
camelia rakudo-moar c27a00: OUTPUT«1 2 3 4␤Too few positionals passed; expected 2 arguments but got 1␤ in block <unit> at /tmp/ZmQcAKx0ew:1␤␤»
marchelzo Zoffix: huh
00:06 lucasb joined
Zoffix m: for (1, 2, 3, 4, 5, 6) -> $a, $b { say "$a $b" } 00:06
camelia rakudo-moar c27a00: OUTPUT«1 2␤3 4␤5 6␤»
Zoffix marchelzo, seems terribly consistent to me :)
joeschmoe2 I have a new friend her name is +camelia
marchelzo this perl6advent thing is wrong then
nine diakopter: are we meta-bikeshedding now? ;)
Zoffix perhaps, is it from last year?
marchelzo from 2009
Zoffix Ohh.
marchelzo, yeah, there likely have been a billion changes since then :|
nine, path-to-repo? :) 00:07
lucasb here, a bikeshed team member arrived :)
_nadim "you cannot create an instance of this type" ... got to love the error messages.
lucasb nine: doesn't it make sense to call it '.from-string' ?
marchelzo I was warned about that. Apparently the authors have been trying to update it for modern perl6 but there are still some things that are outdated.
lucasb *does it
Zoffix nine, or just .repo() :) Hard to say without knowing what other methods are :) 00:08
nine lucasb, Zoffix: I guess what bothers me is that I'm reluctant to have "repo" twice in the name, yet a plain .from-string doesn't really tell you what you get
Zoffix CompUnit::RepositoryRegistry.repo('inst#/path/to/site')
*shrug* 00:09
nine There will also be a method that gives you the 'site' or 'vendor' repo
marchelzo Zoffix: so how _would_ you loop over two lists in parrallel? flat $a Z $b?
Zoffix marchelzo, no idea. I barely know any Perl 6, sorry :) 00:10
marchelzo Zoffix: oh, ok. no worries.
lucasb timotimo: when you wake up, see if you've put this test in the wrong place: github.com/timo/json_fast/blob/mas...ast.pm#L81 00:12
I think where it is, it'll never fire and end the loop.
nine I think I will go with repository-for-spec (because it may just return a cached object) and repository-for-name
Zoffix lucasb, is that the probable cause of the freeze? 00:13
Zoffix is trying to debug that ATM
lucasb Zoffix: that's my hypothesis :) 00:14
00:15 TEttinger left
marchelzo How to interleave multiple lists together? 00:15
00:16 joeschmoe2 left 00:17 joeschmoe joined, n0tjack joined
nine marchelzo: maybe [>]? 00:17
marchelzo: err [Z]
marchelzo I want (1, 2, 3) and (4, 5, 6) to become (1, 4, 2, 5, 3, 6) 00:18
so flat after [Z] would work
but I was wondering if there was a shorter way
lucasb m: say <1 2 3> Z <4 5 6>
camelia rakudo-moar c27a00: OUTPUT«((1 4) (2 5) (3 6))␤»
marchelzo it seems like a common thing to want to do (e.g., for $list1 ??? $list2 -> $elem1, $elem2 { ... }) 00:19
joeschmoe sorry
nine marchelzo: for <1 2 3> Z <a b c> -> ($a, $b) { ... } 00:20
marchelzo so the parens around $a, $b are important?
00:20 Sgeo__ joined
Zoffix m: my @l1 = 1..4; my @l2 = 5..8; for @l1 Z @l2 -> ($a, $b) { say "[$a $b]" } 00:20
camelia rakudo-moar c27a00: OUTPUT«[1 5]␤[2 6]␤[3 7]␤[4 8]␤»
Zoffix m: my @l1 = 1..4; my @l2 = 5..8; for @l1 Z @l2 -> $a, $b { say "[$a $b]" }
camelia rakudo-moar c27a00: OUTPUT«[1 5 2 6]␤[3 7 4 8]␤»
nine yes, that makes it unpack
marchelzo nine: great, that makes sense. thanks 00:21
lucasb another variant is: <1 2 3> Z[&slip] <4 5 6>
but that looks ugly :)
joeschmoe quick question how do you do a regex search and replace as a oneliner
marchelzo m: <1 2 3> Z[&slip] <4 5 6>
camelia ( no output )
marchelzo m: say (<1 2 3> Z[&slip] <4 5 6>).fmt 00:22
camelia rakudo-moar c27a00: OUTPUT«1 4 2 5 3 6␤»
joeschmoe quick question how do you do a regex search and replace as a oneliner and print it to std out
marchelzo so.. is Z a meta-operator as well as a regular operator?
Zoffix joeschmoe, you mean in files?
joeschmoe sure or just all on the command line quoted 00:23
Zoffix no idea
Zoffix is as helpful as a brick :P
00:23 cbk joined, Axord joined
nine marchelzo: yes 00:25
marchelzo what does the <1 2 3> syntax mean exactly? is it the same as (1, 2 3)?
00:26 lucasb left
nine marchelzo: yes, <foo bar baz> is equal to ('foo', 'bar', 'baz') 00:26
Zoffix aw
lucasb left seconds before I found that their hypothesis was right :) 00:27
00:28 z8 joined, M-eternaleye left, M-eternaleye joined
_nadim can one reach non exported sub via the package they reside in? EG A::B::C::<&foo>() 00:30
00:30 Psyche^ joined
Zoffix .tell timotimo Sent you a PR to fix the freezing issue. lucasb++ first found the problematic line. github.com/timo/json_fast/pull/8 00:30
yoleaux Zoffix: I'll pass your message to timotimo.
00:31 skids joined 00:33 captain-adequate joined
dalek kudo/repository_registry: 10eb41b | (Stefan Seifert)++ | / (9 files):
Rename CompUnitRepo to CompUnit::RepositoryRegistry

Its job will be to manage the mapping of repository spec strings to repositories.		 00:34
kudo/repository_registry: 75ffcdc | (Stefan Seifert)++ | src/ (4 files):
Rename CompUnit::RepositoryRegistry.new to .repository-for-spec

CompUnit::RepositoryRegistry is uninstantiable and the method is used only in a few places. No need to shorten the name.
00:35 TEttinger joined, Zoffix left
joeschmoe im doing this wrong i just want to search and replace and then pring -e 's:g/ut/foo/'; say $_ 00:39
00:39 n0tjack left 00:40 n0tjack joined
[Coke] advent2013-day14.t is hanging. :| 00:43
flussence [Coke]: that's happened to me a few times, seems to happen at random 00:51
dalek rl6-roast-data: 71b6ae5 | coke++ | / (9 files):
today (automated commit)
joeschmoe m: 'hello world' ~~ s/'hello world'/bye/; .say 00:54
camelia rakudo-moar c27a00: OUTPUT«Cannot modify an immutable Str␤ in block <unit> at /tmp/Ac_YQSYh1Q:1␤␤»
[Coke] it shouldn't be ble to hang that, though ^^
joeschmoe m: 'hello world' ~~ s/hello world/bye/; .say
camelia rakudo-moar c27a00: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at /tmp/dpsKMLEkGX:1␤ ------> 3'hello world' ~~ s/hello7…»
joeschmoe m: 'hello world' ~~ 's/hello world/bye/'; .say 00:55
camelia rakudo-moar c27a00: OUTPUT«(Any)␤»
joeschmoe m: 'hello world' ~~ :s/hello world/bye/; .say
camelia rakudo-moar c27a00: OUTPUT«5===SORRY!5=== Error while compiling /tmp/gtPAxuuZtH␤Missing required term after infix␤at /tmp/gtPAxuuZtH:1␤------> 3'hello world' ~~ :s/hello world/bye/7⏏5; .say␤ expecting any of:␤ prefix␤ term␤»
joeschmoe what am i doing wrong in the search and replace 00:56
00:56 joeschmoe left 00:58 leont left 01:00 zakharyas left
skids m: my $a = "hello world"; $a ~~ s:s/hello world/bye/; $a.say; 01:03
camelia rakudo-moar c27a00: OUTPUT«bye␤»
skids m: $_ = "hello world"; s:s/hello world/bye/; .say
camelia rakudo-moar c27a00: OUTPUT«bye␤» 01:04
skids The first s is "run a substitution" and the second one is an adverb that means spaces are not ignored in the pattern. 01:05
01:10 yeahnoob joined 01:13 cdg left 01:15 cdg joined 01:20 cdg left 01:21 Zoffix joined 01:32 km3 joined 01:34 rurban left 01:36 herby_ joined
herby_ Evening, everyone! 01:36
marchelzo hi herby_
herby_ m: if "abcde" ~~ /e/ {say "Found it"} 01:37
camelia rakudo-moar c27a00: OUTPUT«Found it␤»
herby_ I'm still having a problem wrapping my head around Grammars
if I do: grammar Test { token TOP { c } }, then look for a match, it doesn't find anything 01:38
Zoffix herby_, have you seen today's Advent article?
herby_ I have :)
great article
Zoffix :)
k
herby_ everyone is talking about how great grammars are, i just can't seem to grasp them yet
Test.parse("abcde") would not find a match
Zoffix m: grammar Test { token TOP { <c> }; token c { \d+ }; }; say Test.parse("42"); 01:39
camelia rakudo-moar c27a00: OUTPUT«｢42｣␤ c => ｢42｣␤»
herby_ what if you did token c { 4 }
so looking for a specific digit, instead of a digit 01:40
Zoffix m: grammar Test { token TOP { .+? 'c' .+ } }; say Test.parse("abcde");
camelia rakudo-moar c27a00: OUTPUT«｢abcde｣␤»
Zoffix m: grammar Test { token TOP { <c> .+ }; token c { 4 }; }; say Test.parse("42");
camelia rakudo-moar c27a00: OUTPUT«｢42｣␤ c => ｢4｣␤»
herby_ hmmm
what the world
Zoffix IIUC, the grammar needs to match the whole parse string
herby_ ahhh
Zoffix m: grammar Test { token TOP { <c> }; token c { 4 }; }; say Test.parse("42");
camelia rakudo-moar c27a00: OUTPUT«Nil␤»
herby_ thats what I discovered
the fact that I need to know the structure of the whole string 01:41
Zoffix so this fails, because there's no match rule for the "2"
herby_ ok, i just wanted to make sure I wasn't missing a concept
flussence wait, does that do an implicit ^ and $ ? 01:42
01:42 julien_ joined, julien_ left
herby_ m: grammar Test { token TOP { <c> }; token c { 4 } ; }; say Test.parse("24"); 01:43
camelia rakudo-moar c27a00: OUTPUT«Nil␤»
01:44 cdg joined
Zoffix Don't know if it's an implicit ^ and $, but, I'd think "parse" would mean to "interpret the string"... and if part of it doesn't make sense to the grammar, it's fair for it to fail 01:44
marchelzo Zoffix: you mean there is a perl6 advent for this year as well?
Zoffix marchelzo, yeah: perl6advent.wordpress.com/
marchelzo Zoffix: ah, I'm looking at the 2009 one right now. Have they done one every year since then? 01:45
Zoffix *shrug* probably
marchelzo I've got a lot of reading to do then
Zoffix :)
m: grammar Test { token TOP { <c>+ }; token c { \d } ; }; say Test.parse("24"); 01:46
camelia rakudo-moar c27a00: OUTPUT«｢24｣␤ c => ｢2｣␤ c => ｢4｣␤»
01:46 lustlife joined
Zoffix Grammars are so nice. 01:46
herby_ :)
I'm trying to rewrite a simple server log parser (currently written in Perl 5) using Grammars
but the format of each line might change slightly, depending what activity occured
so it seems like with Grammars, I don't want to use them unless I know exactly what the string will look like? 01:47
or am I oversimplifying it
I'm thinking I'm oversimplifying it 01:48
01:48 cdg left
geekosaur depends on whether you can characterize what you are looking at. log lines tend to have a well defined prefix and a reasonably well defined end (timestamp, newline respectively are common) 01:49
you might capture the rest and pass it to another grammar which tries to match various kinds of log messages
herby_ yeah. there are mainly two types of string entries for this. one entry shows the user logging into the server with workstation info etc 01:50
and the other type of line will be user activity
I'll try and plan for both 01:51
geekosaur but remember that at the bottom is just regexes so it can handle quite a lot; you do need to be able to discern one of those types from the other, but if you can't do that then you've got a bigger problem than grammars anyway, I suspect
Zoffix m: grammar Test { token TOP { <log-line>+ }; token log-line { .+?\n } ; }; say Test.parse("foo\nbar\nbaz");
camelia rakudo-moar c27a00: OUTPUT«Nil␤»
Zoffix m: grammar Test { token TOP { <log-line>+ }; token log-line { .+?\n } ; }; say Test.parse("foo\nbar\nbaz\n");
camelia rakudo-moar c27a00: OUTPUT«｢foo␤bar␤baz␤｣␤ log-line => ｢foo␤｣␤ log-line => ｢bar␤｣␤ log-line => ｢baz␤｣␤»
01:52 sufrostico left
Zoffix m: grammar Test { token TOP { <log-line>+ }; token log-line { .+?\n? } ; }; say Test.parse("foo\nbar\nbaz"); 01:52
camelia rakudo-moar c27a00: OUTPUT«｢foo␤bar␤baz｣␤ log-line => ｢f｣␤ log-line => ｢o｣␤ log-line => ｢o␤｣␤ log-line => ｢b｣␤ log-line => ｢a｣␤ log-line => ｢r␤｣␤ log-line => ｢b｣␤ log-line => ｢a｣␤ log-line => ｢z｣␤»
Zoffix m: grammar Test { token TOP { <log-line>+ }; token log-line { .+\n? } ; }; say Test.parse("foo\nbar\nbaz");
camelia rakudo-moar c27a00: OUTPUT«｢foo␤bar␤baz｣␤ log-line => ｢foo␤bar␤baz｣␤»
Zoffix m: grammar Test { token TOP { <log-line>+ }; token log-line { .+?[\n|\z] } ; }; say Test.parse("foo\nbar\nbaz"); 01:53
camelia rakudo-moar c27a00: OUTPUT«5===SORRY!5=== Error while compiling /tmp/NV7alei5Md␤Unsupported use of \z as end-of-string matcher; in Perl 6 please use $␤at /tmp/NV7alei5Md:1␤------> 3log-line>+ }; token log-line { .+?[\n|\z7⏏5] } ; }; say Test.parse("foo\nbar\nbaz")␤»
Zoffix aw
m: grammar Test { token TOP { <log-line>+ }; token log-line { .+?[\n|$$] } ; }; say Test.parse("foo\nbar\nbaz");
camelia rakudo-moar c27a00: OUTPUT«｢foo␤bar␤baz｣␤ log-line => ｢foo␤｣␤ log-line => ｢bar␤｣␤ log-line => ｢baz｣␤»
Zoffix weee
geekosaur don't you really want log lines separated by \n ?
hoelzro I just noticed while trying to debug further panda-build problems that `perl6 --ll-exception $(which panda-build)` doesn't pass --ll-exception to the new perl6 process. Is there a way to find out the command line arguments given between perl6 and the name of the script? 01:54
Zoffix herby_, BTW, there's modules.perl6.org/repo/Grammar::Debugger (which also includes Grammar::Tracer) you may find useful when writing grammars 01:55
01:55 lustlife` joined
geekosaur m: grammar Test { token TOP { <log-line>+ % '\n' }; token log-line { .+} ; }; say Test.parse("foo\nbar\nbaz"); 01:55
camelia rakudo-moar c27a00: OUTPUT«｢foo␤bar␤baz｣␤ log-line => ｢foo␤bar␤baz｣␤»
geekosaur no, did that wrong
m: grammar Test { token TOP { <log-line>+ % '\n' }; token log-line { .+? } ; }; say Test.parse("foo\nbar\nbaz"); 01:56
herby_ thanks Zoffix, I saw a youtube video where he was using a perl6-debugger but I couldn't get it to work
camelia rakudo-moar c27a00: OUTPUT«Nil␤»
herby_ I'll check those out
what does the % do? 01:57
<log-line>+ %
<log-line>+ % '\n' 01:58
01:59 lustlife left
Zoffix m: grammar Test { token TOP { <log-line>+ % '\n' }; token log-line { .+? } ; }; say Test.parse("foo\nbar\nbaz"); 02:02
camelia rakudo-moar c27a00: OUTPUT«Nil␤»
Zoffix no idae
m: grammar Test { token TOP { [ <log-line> \n? ]+ }; token log-line { <-[\n]> } ; }; say Test.parse("foo\nbar\nbaz");
camelia rakudo-moar c27a00: OUTPUT«｢foo␤bar␤baz｣␤ log-line => ｢f｣␤ log-line => ｢o｣␤ log-line => ｢o｣␤ log-line => ｢b｣␤ log-line => ｢a｣␤ log-line => ｢r｣␤ log-line => ｢b｣␤ log-line => ｢a｣␤ log-line => ｢z｣␤»
Zoffix m: grammar Test { token TOP { [ <log-line> \n? ]+ }; token log-line { <-[\n]>+ } ; }; say Test.parse("foo\nbar\nbaz");
camelia rakudo-moar c27a00: OUTPUT«｢foo␤bar␤baz｣␤ log-line => ｢foo｣␤ log-line => ｢bar｣␤ log-line => ｢baz｣␤»
lucs ack'ing for "EXECUTABLE" in the design docs (github.com/perl6/specs) matches nothing -- looking for P5's $^X equivalent. 02:03
Zoffix m: say $*PROGRAM 02:04
camelia rakudo-moar c27a00: OUTPUT«"/tmp/zKU4Shd0tn".IO␤»
herby_ <log-line>+ % '\n'
Zoffix m: say $*EXECUTABLE-NAME
camelia rakudo-moar c27a00: OUTPUT«perl6-m␤»
Zoffix lucs, check out doc.perl6.org/language/5to6-perlvar
herby_ whoops, ignore that
geekosaur herby_, the % means parse a list of something separated by something else 02:05
lucs Zoffix: Yeah, that's where I saw EXECUTABLE(-NAME, yes), but I was looking for more details.
geekosaur so <number> % ',' would be a comma-separated list of numbers
er
<number>+ % ',' 02:06
saying at least one number, separated by commas
herby_ ah ok, thats pretty neat
lucs (details like, can I get its full path from inside the program?) 02:07
geekosaur that isn't always available 02:10
lucs Hmm...
herby_ m: grammar Test { token TOP { <c>+ % ',' }; token c { \d }; }; say Test.parse("1,2,3,4,5");
camelia rakudo-moar c27a00: OUTPUT«｢1,2,3,4,5｣␤ c => ｢1｣␤ c => ｢2｣␤ c => ｢3｣␤ c => ｢4｣␤ c => ｢5｣␤»
marchelzo is there an alternative way to write f($_) without explicitly mentioning $_? f is not a method, it's a block. 02:11
zengargoyle m: say $*EXECUTABLE.abspath
camelia rakudo-moar c27a00: OUTPUT«./rakudo-m-inst/bin/perl6-m␤»
zengargoyle m: say $*EXECUTABLE.^methods 02:12
camelia rakudo-moar c27a00: OUTPUT«(BUILD new-from-absolute-path abspath is-absolute is-relative parts volume dirname basename extension Numeric Bridge Int succ pred IO open pipe watch absolute relative cleanup resolve parent child chdir rename copy move chmod unlink symlink link mkdir rmdi…»
zengargoyle you might be able to find something in there...
lucs geekosaur: I happen to want to launch perl6 from within the program, and I'd like not to have to trust the PATH and make sure I'm invoking the same instance. 02:13
flussence m: sub f { $^a * 2 }; given 5 { say .&f } 02:14
camelia rakudo-moar c27a00: OUTPUT«10␤»
lucs geekosaur: Why do you say it's not always available?
geekosaur zengargoyle, there turn out to be ways to defeat that...
the program is usually passed a basename as $0 if it was found on $PATH. one sting in the tail is that the shell may have a different $PATH it's using than it has exported 02:15
marchelzo flussence: thanks
geekosaur (this is not common but is possible)
I should say as argv[0] since I am talking about C here 02:16
or if it's not started by a shell, it could do pretty much anything
(not to mention edge cases like traditional login prepending a - to argv[0] to indicate that it's a login shell)
lucs geekosaur: Ah, I see what you mean. 02:17
zengargoyle m: say (slurp "/proc/self/cmdline").trans( "\x00"=>' ');
camelia rakudo-moar c27a00: OUTPUT«/home/camelia/rakudo-m-inst-2/bin/moar --execname=./rakudo-m-inst/bin/perl6-m --libpath=/home/camelia/rakudo-m-inst-2/share/nqp/lib --libpath=/home/camelia/rakudo-m-inst-2/share/perl6/lib --libpath=/home/camelia/rakudo-m-inst-2/share/perl6/runtime /home/ca…»
geekosaur also IIRC windows cannot pass a path
herby_ Thanks for the help and tips on understanding Grammars. Now I have some ideas for my log parser
lucs geekosaur: I guess I'll have to trust the PATH :)
02:17 herby_ left
geekosaur so, you can get something that *usually* works but not that *always* works, except by going to platform dependent stuff like Linux /proc/self/exe 02:17
02:19 jdong left
zengargoyle wonders if Perl 6 has a way to set program name ala $0 = 'foobar' 02:19
tho even if it's platform dependent, it should be settable by that platform dependent method during startup... 02:21
then again, the actual executable is 'moar' with a bunch of options... :) 02:24
02:24 cdg joined 02:27 Sqirrel left 02:28 astrofyziky joined, astrofyziky left
zengargoyle heh: $ /lib/x86_64-linux-gnu/ld-linux-x86-64.so.2 /opt/rakudobrew/moar-nom/install/bin/moar --execname=/opt/rakudobrew/bin/../moar-nom/install/bin/perl6 --libpath=/opt/rakudobrew/moar-nom/install/share/nqp/lib --libpath=/opt/rakudobrew/moar-nom/install/share/perl6/lib --libpath=/opt/rakudobrew/moar-nom/install/share/perl6/runtime /opt/rakudobrew/moar-nom/install/share/perl6/runtime/perl6.moarvm -e 'say 02:29
$*EXECUTABLE-NAME'
output: perl6
the LIES!
02:29 cdg left
hahainternet day 8: "with an impressive collection of utility to get most of our common", utility should be plural i think 02:31
"Tools that take the repetitive and automatable parts of our work the way" should be "away" i believe
"Just as classes are first class citizen that we can introspect" citizen should also be plural
hope people don't mind me giving these corrections, it's all i can do to help D: 02:32
lucs (introspecting citizens is an unusual metaphor, eh)
hahainternet classes is plural, and so citizen should be too i think 02:33
that's how the rule goes iirc
so says my grade C in GCSE English :D
"Each expression is a list of terms, or possibly and alternative of them" and should be 'an' 02:34
02:35 secwang joined
marchelzo is it possible to load some code into the repl and use things that are defined in it? 02:39
zengargoyle lucs: looks like the executable name is set by the perl6 wrapper script itself. passed in as an option to the moar vm instance running the code. 02:40
dj_goku marchelzo: like print them or inspect them?
02:40 telex joined
lucs zengargoyle: Do we happen to have access to it? 02:40
zengargoyle my .../bin/perl6: exec /opt/rakudobrew/moar-nom/install/bin/moar --execname="$0" 02:41
marchelzo dj_goku: like I made a grammar, and I want to try parsing some stuff in the repl.
zengargoyle m: say $*EXECUTABLE.path 02:42
camelia rakudo-moar c27a00: OUTPUT«./rakudo-m-inst/bin/perl6-m␤»
zengargoyle so it depends in this case on how you start it... 02:43
lucs zengargoyle: Cool, thanks :)
zengargoyle m: say $*EXECUTABLE.absolute
camelia rakudo-moar c27a00: OUTPUT«./rakudo-m-inst/bin/perl6-m␤»
02:43 BenGoldberg joined, vendethiel joined
zengargoyle m: say $*EXECUTABLE.resolve.absolute 02:44
camelia rakudo-moar c27a00: OUTPUT«/rakudo-m-inst/bin/perl6-m␤»
02:45 z8 left
zengargoyle wonder if camelia is in a chroot... 02:45
lucs It expands correctly for me here.
marchelzo does this look reasonable? I don't know if I need to use regex for TOP or if I could use token: sprunge.us/CfWK 02:46
zengargoyle or probably started up funny to handle rebuilds and such. like: cd /home/camelia; ./rakudo-m-inst/bin/perl6-m; probably wrapped up in an easily kickable script. 02:47
02:47 AlexDaniel joined
AlexDaniel m: my $foo = 5; say “x: $_” for $foo 02:48
camelia rakudo-moar c27a00: OUTPUT«x: 5␤»
zengargoyle it doesn't get PATH expanded since it's started with an absolute path
AlexDaniel m: my @foo = 5, 4; say “x: $_” for @foo
camelia rakudo-moar c27a00: OUTPUT«x: 5␤x: 4␤»
AlexDaniel m: my $foo = [5, 4]; say “x: $_” for $foo
camelia rakudo-moar c27a00: OUTPUT«x: 5 4␤»
AlexDaniel m: my $foo = [5, 4]; say “x: $_” for $foo; say $foo.WHAT
camelia rakudo-moar c27a00: OUTPUT«x: 5 4␤(Array)␤»
AlexDaniel hmm 02:49
m: my $foo = [5, 4]; say “x: $_” for |$foo
camelia rakudo-moar c27a00: OUTPUT«x: 5␤x: 4␤»
AlexDaniel m: my $foo = 5; say “x: $_” for |$foo
camelia rakudo-moar c27a00: OUTPUT«x: 5␤»
AlexDaniel It's just interesting how sometimes it doesn't matter whether it is an array or not 02:50
zengargoyle marchelzo: TOP can be anything. i think i've even seen it be a method that does some stuff and then calls another rule explicitly.
hahainternet AlexDaniel: where doesn't it matter? it seems that everything there is quite consistent
AlexDaniel hahainternet: yes it is consistent, yet it is pretty useful 02:51
zengargoyle it's just the name TOP that gets called by default. (you can also start at another rule/regex/token/method by passing an argument to parse)
hahainternet oh right, i thought you were complaining about something
zengargoyle AlexDaniel: that's the single thing iterates rule. 02:52
hahainternet i think retaining the clear distinction between scalar, array and hash is fantastic, it's something i find frustrating in python
marchelzo say I've got a Match object, and the pattern was <foo> | <bar> <baz>. How do I tell which one matched?
hoelzro zengargoyle: there's a MoarVM ticket open to address that, iirc
hahainternet i think the one thing i'm unsure of in 6 is how to turn a positional into an array member
marchelzo if $<foo>?
hoelzro marchelzo: you could check which of $<foo> or $<bar> is defined
flussence m: use Test:<1.*>; 02:53
camelia rakudo-moar c27a00: OUTPUT«===SORRY!===␤Could not find Test:<1.*>:ver<True>:auth<True>:api<True> in:␤ /home/camelia/.perl6/2015.11-435-gc27a00c␤ /home/camelia/rakudo-m-inst-2/share/perl6/site␤ /home/camelia/rakudo-m-inst-2/share/perl6/vendor␤ /home/camelia/raku…»
marchelzo hoelzro: how do you check if something is defined?
flussence if I'm reading the docs correctly, that 1.* should end up in :ver, right?
hoelzro $<foo>.defined, or defined($<foo>)
but if $<foo> should do the trick
AlexDaniel hahainternet: sometimes you don't want it though. e.g. there could be some config value that is typically singular, yet you can make it so that if you shove an array ref into it will magically work :) 02:54
hoelzro since all defined Matches are truthy
marchelzo hoelzro: ok, thanks
zengargoyle m: use Test:ver<1.*>;
camelia rakudo-moar c27a00: OUTPUT«===SORRY!===␤Could not find Test:ver<1.*>:auth<True>:api<True> in:␤ /home/camelia/.perl6/2015.11-435-gc27a00c␤ /home/camelia/rakudo-m-inst-2/share/perl6/site␤ /home/camelia/rakudo-m-inst-2/share/perl6/vendor␤ /home/camelia/rakudo-m-in…»
zengargoyle m: use Test:ver<*>;
camelia rakudo-moar c27a00: OUTPUT«===SORRY!===␤Could not find Test:ver<*>:auth<True>:api<True> in:␤ /home/camelia/.perl6/2015.11-435-gc27a00c␤ /home/camelia/rakudo-m-inst-2/share/perl6/site␤ /home/camelia/rakudo-m-inst-2/share/perl6/vendor␤ /home/camelia/rakudo-m-inst…»
zengargoyle m: use Test:ver<True>;
camelia rakudo-moar c27a00: OUTPUT«===SORRY!===␤Could not find Test:ver<True>:auth<True>:api<True> in:␤ /home/camelia/.perl6/2015.11-435-gc27a00c␤ /home/camelia/rakudo-m-inst-2/share/perl6/site␤ /home/camelia/rakudo-m-inst-2/share/perl6/vendor␤ /home/camelia/rakudo-m-i…»
zengargoyle heh
AlexDaniel hahainternet: otherwise you'd have to pass [$val] all the time… which is probably not too bad 02:55
hahainternet AlexDaniel: well in python, it's frustrating that if i pass a string to a function and then say 'for x in y', it will iterate the damn string
AlexDaniel hahainternet: hahaha
hahainternet wheras what i really want to say is "this positional is always an array, even if a string is passed, make it [$x]"
it doesn't matter as much in p6 because of better design
but the closest i know is slurpy args, it's just one thing on my list :D
marchelzo how can you do if-then-else as an expression? Like $cond ? $val : $val2 or something. 02:56
zengargoyle ?? !!
AlexDaniel marchelzo: $cond ?? $val !! $val2
hahainternet you can also use 'if' as the rhs can't you?
AlexDaniel hahainternet: can't use else though :) 02:57
hahainternet oh really? i thought there was a nice idiom for that
zengargoyle orelse?
hahainternet i don't have enough time in my day to write nice perl6 :(
zengargoyle m: say do if True { "bar" } else { "foo" }; 02:58
camelia rakudo-moar c27a00: OUTPUT«bar␤»
zengargoyle m: say do if False { "bar" } else { "foo" };
camelia rakudo-moar c27a00: OUTPUT«foo␤»
hahainternet else works fine yeah
i just checked locally lol
AlexDaniel huh
hahainternet expression -> statement or what have you 02:59
AlexDaniel why would you use it instead of ?? !! ?
lucs zengargoyle: Not the most elegant, but if you must... :)
hahainternet it's very explicit, and you might have a more complex conditional etc
lucs Yeah, the structure is useful. 03:00
hahainternet i've written quite a bit of golang recently
and going back to python, dear god i miss a lot of the explicit nature
AlexDaniel hahainternet: well, Perl 6 is also not the most explicit language 03:01
hahainternet AlexDaniel: on the contrary, to me it appears to be as explicit as you like
and a lot of the implicit assumptions are consistent and predictable
i'm very happy with how p6 is turning out, i really look forward to finding work in it
AlexDaniel hahainternet: well, if you write a couple of things explicitly, then perhaps 03:02
03:02 noganex_ joined
AlexDaniel hahainternet: but most Perl 6 code that you will stumble upon is probably not like this 03:02
hahainternet AlexDaniel: too early to say imho
AlexDaniel though we don't know right now, yes
03:02 Herby_ joined
hahainternet the current thing i'm most intrigued with is the ability to produce a 'proper' ORM 03:04
AlexDaniel m: say "test" if True or say 42 and False
camelia rakudo-moar c27a00: OUTPUT«test␤»
AlexDaniel m: say "test" if False or say 42 and False
camelia rakudo-moar c27a00: OUTPUT«42␤»
marchelzo wow I am so impressed with perl6
hahainternet i've been using a number of different ORMs of late, and they're all kinda bleh
AlexDaniel look at this… I think that I wrote a bit too much of prolog…
hahainternet AlexDaniel: what's with 'and False'? 03:05
hoelzro marchelzo: glad to hear that =)
AlexDaniel hahainternet: it has to fail…
hahainternet: otherwise you'd get both
03:05 vendethiel left
AlexDaniel hahainternet: so it is an “else” emulator… 03:05
marchelzo hoelzro: does this look okay style-wise? I think maybe I should have used if ... {...} else {...} instead of the ?? !! syntax. sprunge.us/KgGd 03:06
hahainternet interesting, and 'xor' doesn't seem to work locally
marchelzo grammars are super nice though
hahainternet i don't quite see how that works :/
marchelzo: fwiw i think your use of ?? and !! is pretty clear
AlexDaniel marchelzo: usually you'd use ?? !! when you want to assign 03:07
hoelzro looks alright to me; a few pointers:
AlexDaniel marchelzo: so I don't recommend doing it this way, although you might
hoelzro 1) you don't need to do $/.make; just make is fine
2) you can also use regexes with different long names instead of |, if you prefer
AlexDaniel marchelzo: e.g. $val = $cond ?? 42 !! 69; # here is where it totally makes sense
hoelzro so regex factor:times { <number> '*' <factor> } ; regex factor:number { <number> } 03:08
AlexDaniel marchelzo: everywhere else, well, not so much, unless you want to return something
hoelzro and then have method factor:times($/) and method factor:number($/) in your actions class
that's more a personal taste thing for me, though
AlexDaniel marchelzo: Oh, you are returning something, aren't you?
03:09 cdg joined
marchelzo hoelzro: oh, that's exactly what I was wishing I could do 03:09
incredible
what exactly does the make method do? I don't fully understand its significance. I get what .made does, but not .make. 03:10
hahainternet AlexDaniel: oh i just looked at it again and i see the precedence, i'm just dumb :D 03:11
marchelzo AlexDaniel: yea
hoelzro marchelzo: make is kind of like return for action methods
I don't know if I even fully get it =/
it basically sets .made/.ast for $/
marchelzo I tried replacing $/.make: with just make: but I got an error. 03:12
03:12 khw left
marchelzo "Redeclaration of symbol make" 03:12
hoelzro huh, that's odd. 03:13
oh, drop the :
03:13 cdg left
hoelzro just "make $value" 03:13
marchelzo oh, ok 03:14
so action methods are not regular methods? 03:15
hoelzro they are
I just don't know why make is used and not return
marchelzo I see 03:17
the fact that you can write make instead of $/.make:... is that a special thing just for grammars?
hoelzro I think that make just compiles to $/.make 03:19
hoelzro looks
yes
it just desugars to $/.make 03:20
03:21 kid51 left
AlexDaniel m: sub infix:<relse>($cond, $b) is looser(&infix:<or>) { $cond ?? True !! ($b() and False) }; say “just look” if 42 < 69 relse { say “at this insanity” }; 03:25
camelia rakudo-moar c27a00: OUTPUT«just look␤»
AlexDaniel m: sub infix:<relse>($cond, $b) is looser(&infix:<or>) { $cond ?? True !! ($b() and False) }; say “just look” if 42 > 69 relse { say “at this insanity” }; 03:26
camelia rakudo-moar c27a00: OUTPUT«at this insanity␤»
hahainternet nice
AlexDaniel I wonder if there is any way to get rid of { } 03:27
I don't think so, unless you modify the grammar more crazily
hahainternet shrugs 03:28
AlexDaniel hahainternet: this reminds me of some Python feature…
hahainternet: the one that does !! ?? 03:29
｢a if test else b｣
hahainternet yeah, the 'ternary' type operator added in 2.6 or so?
AlexDaniel m: sub infix:<python>($cond, $b) is looser(&infix:<or>) { $cond ?? True !! ($b() and False) }; say “just look” if 42 > 69 python { say “at this insanity” };
camelia rakudo-moar c27a00: OUTPUT«at this insanity␤»
hahainternet i wonder how those quotes are done on my keyboard 03:30
AlexDaniel ｢doSomething if cond else pass｣ – RHS if in python? 03:31
hahainternet it's only for assignment iirc but it is 3:30am so i'm not exactly at 100%
AlexDaniel “SyntaxError: invalid syntax” 03:32
hahainternet var = "foo" if True else "bar" 03:33
is the correct syntax
i had to actually test it lol
i'm going to get some tea :D
AlexDaniel hahainternet: sure, I know
just hoped that I can trick it :) 03:34
hm
actually, you can use it! 03:35
foo() if 2 < 10 else 1
and it works just fine in void context
so there you have it, rhs if in python… just if this ｢else 1｣ does not bother you too much 03:37
hahainternet AlexDaniel: so how do you type your quote characters? 03:38
i should start customising my compose key / third level keys
AlexDaniel m: say 0_0
camelia rakudo-moar c27a00: OUTPUT«0␤»
AlexDaniel hahainternet: my keyboard layout files.progarm.org/2015-12-09-05392..._scrot.png 03:39
hahainternet: some of these arrows are typable arrows, others are normal arrows :) 03:40
MadcapJake my program for adventofcode day 6 takes like 2 hours to run :_(
AlexDaniel hahainternet: designed for japanese keyboard
hahainternet: the ones with short spacebar and more thumb buttons 03:41
hahainternet AlexDaniel: interesting, how did you produce that png?
AlexDaniel hahainternet: keys with ♿ are sticky
hahainternet: that's just a table in my private wiki. I took a screenshot
hahainternet ah ok 03:42
03:42 Herby_ left
lucs In P5, arbitrary information can be passed to a module's import() function by doing something like use Foo ('bar') . 03:43
How is something similar done in P6 (import() function not being the same notwithstanding)? 03:44
"Similar" meaning: pass some information to the module so it can set itself up as desired. 03:45
AlexDaniel hahainternet: this layout turned up being quite interesting. Well, if you are interested in keyboard layout :) 03:46
hahainternet: for example, AltGr+x is delete. And guess what x does in vim!
hahainternet AlexDaniel: i'm trying to find out how to modify my layout in gnome 3 :)
deletes the character under the cursor?
AlexDaniel yeah
03:46 Ben_Goldberg joined
AlexDaniel hahainternet: I think that you can go straight to creating your own layout 03:47
hahainternet: which is done with xkb
hahainternet: /usr/share/X11/xkb/symbols/us is a file with the layout description 03:48
hahainternet AlexDaniel: not in gnome it seems, it uses some other source for them
AlexDaniel hahainternet: honestly, I don't buy it. I think that almost all GNU/Linux distros are using xkb 03:49
03:49 BenGoldberg left, BenGoldberg_ joined
geekosaur sadly 03:49
yes, they use xkb. but gnome apps use ibus which is informed by but not configured by xkb; they have additional configuration
AlexDaniel okay, but let's say you create your own xkb layout, is it going to be available? 03:50
hahainternet looks like git.gnome.org/browse/gtk+/tree/gtk...mpleseqs.h is the source
geekosaur it will be available. but ibus will intercept some keys before xkb can process them
03:51 Ben_Goldberg left
geekosaur and hat imcontext thing, yeh, which can intercept keys before either ibus or xkb can see them 03:51
zengargoyle ibus is just an input method. xkb is the event generation. 03:52
03:52 orbus joined
zengargoyle you can easily remove ibus or replace it with uim or scim for instance. 03:52
geekosaur enjoy figuring out which layer is messing with you\
zengargoyle you just have to load appropriate modules for other ime's to work with gnome/gtk/qt 03:53
AlexDaniel well, the question is what is going to happen if you use something like “key <AD04> { [ p, P, End, End ] };”
zengargoyle has a love/hate relationship with ibus...
hahainternet i just want to have fancy japanese quotes ;) 03:54
AlexDaniel is it going to get in your way somehow? e.g. swallow End button event or block your third level completely
zengargoyle yeah, that's why i keep ibus around, for anthy. :P
AlexDaniel anyway, I recommend everybody to experiment with their keyboard layouts. Fancy quotes are just part of it, what is way more useful is to have arrow keys on your home row. 03:58
hahainternet i just discovered Ω which is nice :D
didn't know how to type that the other day
03:59 noganex joined
AlexDaniel what I've just discovered is that there's <dead_greek> key 04:00
zengargoyle but which key is it?
AlexDaniel zengargoyle: you have to assign it 04:01
zengargoyle that's the part i never figured out...
plus, laptop with few actually free keys
04:02 noganex_ left
AlexDaniel AltGr gives you a whole set of keys which are free 04:02
twice
zengargoyle no altgr key on my keyboard...
04:03 kaare_ joined
AlexDaniel reassign Caps Lock to AltGr then 04:03
zengargoyle caps is compose :)
AlexDaniel set compose to Caps + something
like Caps+l or whatever 04:04
unless you really want to have a dedicated key for compose :)
zengargoyle i think i'm just going to use that XCompose file that was mentioned in the advent. 04:05
AlexDaniel but in this case you can use right ctrl or whatever. What keyboard is that without AltGr?
lucs & # ZZ
zengargoyle has tons of mappings. greek are Multi * O (for omega)
AlexDaniel what is multi? 04:06
zengargoyle laptop keyboards are often short a few keys, no altgr, only one 'windows' key,
Multi_key is just another name for Compose
AlexDaniel ah right 04:07
zengargoyle: well, right, that's one way to do it. Personally I think that pressing Shift + some key is more comfortable than pressing Compose + some key 04:08
though there are guys who make their shifts sticky
and AltGr is just another kind of a shift :) 04:09
zengargoyle well compose as capslock is just to the left of A on my keyboard, quite easy to hit. and it's sticky so it compose and then other characters, you don't have to hold it down or anything.
AlexDaniel I know, but that's kinda the difference 04:10
by the way, if you don't like the number of keys on your laptop you can always try to find japanese keyboard
I mean, for your laptop
zengargoyle yeah, in the end, i guess i find it easier to remember 'compose + these two sorta mnemonic characters' vs this magic key + this other certain key, or some combination of shift+alt+key :) 04:12
AlexDaniel compose is great for entering rare characters. For common stuff, um, not so much!
e.g. you don't need mnemonics for common stuff :) 04:13
zengargoyle but it's the uncommon stuff that's the most fun. :) 💡 04:14
marchelzo how do you do .pick with replacement?
zengargoyle what's pick with replacement?
AlexDaniel “”‘’｢｣…– «»􏿽xE2􏿽x8C􏿽xA9􏿽xE2􏿽x8C􏿽xAA
^ I'm not sure that these things are uncommon 04:15
marchelzo random selection with replacement
AlexDaniel marchelzo: use splice
zengargoyle i don't get the replacement part...
TimToady that's called "roll", as in dice
zengargoyle do you mean roll?
heh
marchelzo yes. roll. outside of programming this is called random selection with replacement. 04:16
AlexDaniel aaaaaah
zengargoyle m: say (1..6).roll xx 12
camelia rakudo-moar c27a00: OUTPUT«(4 2 4 6 6 4 5 2 6 4 2 5)␤»
TimToady m: say ('⚀' .. '⚅').roll(5)
camelia rakudo-moar c27a00: OUTPUT«(⚁ ⚂ ⚅ ⚀ ⚀)␤»
marchelzo like if you were randomly pulling 5 things out of a bag, .roll would be like replacing each think before pulling out the next one.
each thing*
TimToady hmm
AlexDaniel m: say ⚀
camelia rakudo-moar c27a00: OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at /tmp/gb9C0Qp6jX:1␤------> 3say7⏏5 ⚀␤Bogus postfix␤at /tmp/gb9C0Qp6jX:1␤------> 3say 7⏏5⚀␤ expecting any of:␤ infix␤ infix stopper␤ pos…»
AlexDaniel awww
TimToady m: say ⚅ 04:17
camelia rakudo-moar c27a00: OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at /tmp/engd6PCjA_:1␤------> 3say7⏏5 ⚅␤Bogus postfix␤at /tmp/engd6PCjA_:1␤------> 3say 7⏏5⚅␤ expecting any of:␤ infix␤ infix stopper␤ pos…»
TimToady aww
zengargoyle cards aren't numeric either. :(
AlexDaniel m: say '⚅' - '⚀' # I think that you can do this… 04:18
camelia rakudo-moar c27a00: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5⚅' (indicated by ⏏)␤ in block <unit> at /tmp/0Y2RJbNn8n:1␤␤Actually thrown at:␤ in block <unit> at /tmp/0Y2RJbNn8n:1␤␤»
AlexDaniel noooo
ah
m: say '⚅'.org - '⚀'.ord # I think that you can do this…
camelia rakudo-moar c27a00: OUTPUT«Method 'org' not found for invocant of class 'Str'␤ in block <unit> at /tmp/isVC8tQzEX:1␤␤»
AlexDaniel m: say '⚅'.ord - '⚀'.ord # I think that you can do this…
camelia rakudo-moar c27a00: OUTPUT«5␤»
AlexDaniel m: my \term:<⚅> = 6; say ⚅; 04:22
camelia rakudo-moar c27a00: OUTPUT«6␤»
dalek osystem: 97800c8 | (Anthony Parsons)++ | META.list:
Add Minecraft-Tools to ecosystem

  github.com/flussence/Minecraft-Tools
Just for fun, this contains two modules whose IDs differ only in version number. Let's see how that goes...		 04:23
04:25 aoniao joined
AlexDaniel 20s startup for my thing using bailador… uh 04:25
flussence that might just be precomp
should be faster the second time 04:26
AlexDaniel slow every time
flussence oh... :(
04:27 aoniao left
zengargoyle has been pleasantly surprised how much second time is faster. well usually... 04:27
AlexDaniel actually, I haven't noticed any speedup when precomp arrived…
any way to force no precomp?
so that I can see if it is much slower
zengargoyle 0m12.424s vs 0m2.460s for testing Algorithm::Trie::libdatrie 04:33
04:34 BenGoldberg_ left 04:35 secwang left
zengargoyle that's 4 test files so 4x that it doesn't have to compile from scratch... 04:35
04:36 bpmedley left
dj_goku yay just connected to connection using IO::Socket::Async to a real system (gearmand)! 04:38
blah that didn't come out right. I made my first real connection to gearmand using IO::Socket::Async! 04:39
04:39 marchelzo left 04:42 kaare_ left
skids .tell masak gist.github.com/skids/97378d26a684591f1b89 # I only felt like bikeshedding tonight 04:47
yoleaux skids: I'll pass your message to masak.
zengargoyle wonders if NativeCall needs to be a depends in META 04:48
04:49 AlexDaniel left 04:50 n0tjack left 05:00 bpmedley joined, secwang joined
hoelzro zengargoyle: I don't think so 05:05
it's bundled with Rakudo, right?
05:13 secwang left 05:21 mattp__ left 05:32 xinming joined 05:44 xinming left 05:46 skids left 05:47 lustlife` left, lustlife` joined 05:48 zengargoyle left 05:50 secwang joined 05:51 xinming joined 05:53 Sqirrel joined 05:55 xinming left 05:56 xinming joined 05:57 glaukommatos left 06:02 zengargoyle joined 06:03 adu joined
adu \o 06:04
06:06 xinming left 06:08 xinming joined 06:17 xinming left 06:26 xinming joined 06:30 xinming left, xinming joined 06:32 Actualeyes joined 06:35 xinming left
MadcapJake how would you do bitwise negation that *doesn't* give a negative result? 06:37
wait, nevermind, problem is elsewhere :P 06:38
no that is it, how do you do one's complement? 06:42
MadcapJake loves bitwise math :P
adventofcode.com/day/7 has bitwise NOT of 123 equals 65412 ?? that doesn't make sense to me 06:44
oh i think i need to have an unsigned int 06:46
06:46 xinming joined 06:48 Tux__ joined 06:54 glaukommatos joined, kjs_ joined, kjs_ left 06:56 geraud left 07:00 glaukommatos left 07:01 CIAvash joined 07:08 sno left 07:09 adu left 07:10 ambs joined 07:18 rurban joined 07:23 M-eternaleye is now known as eternaleye
[Tux] test 50000 22.975 22.864 07:38
test-t 50000 16.793 16.683
csv-parser 50000 25.370 25.259
07:40 eyck joined
grondilu adventofcode? I failed at day 3, no idea why. Seemed easy enough. 07:45
07:48 sno joined 07:51 CIAvash left
_nadim Good morning everyone. 07:51
grondilu say 1 + unique [\+] map { when '^' {+i}; when 'v' {-i}; when '<' {-1}; when '>' {+1} }, slurp.comb; # that was a solution I thought was smart. 07:53
07:54 RabidGravy joined
grondilu oh wait that does not work with the examples 07:56
m: say unique 0, 0+0i 07:57
camelia rakudo-moar c27a00: OUTPUT«(0 0+0i)␤»
07:57 glaukommatos joined
grondilu yeah I remember now 07:57
that's how I found the (0+0i).narrow bug 07:58
moritz grondilu: map *.narrow, ...
grondilu m: say (0+0i).narrow
camelia rakudo-moar c27a00: OUTPUT«Attempt to divide by zero using /␤ in block <unit> at /tmp/vlOiRxr9BK:1␤␤Actually thrown at:␤ in block <unit> at /tmp/vlOiRxr9BK:1␤␤»
moritz eeks
grondilu I got annowed and stopped searching.
moritz grondilu: looks like an easy fix 07:59
grondilu I did submit a PR though: github.com/rakudo/rakudo/pull/620
dalek kudo/nom: b2ae6dc | grondilu++ | src/core/Complex.pm:
handling nul real part in narrow method

you don't want to divide by $!re unless you've dealt with the case $!re == 0e0		 08:01
kudo/nom: 49d8728 | grondilu++ | src/core/Complex.pm:
deal with nul case

if self is actually null, the nul integer should be returned.
kudo/nom: 40fe92d | moritz++ | src/core/Complex.pm:
Merge pull request #620 from grondilu/patch-1

handling nul real part in narrow method
08:02 glaukommatos left
grondilu thanks 08:02
still won't solve day 3 for me I'm afraid
say +unique [\+] 0i, |map { when '^' {+i}; when 'v' {-i}; when '<' {-1}; when '>' {+1} }, slurp.comb;
moritz grondilu: still needs a .narrow somewhere 08:03
_nadim How can one access a non exported sub?
grondilu no, as I put 0i instead of 0
moritz grondilu: an
_nadim: if it's an "our" sub, through the package name
_nadim: ThePackage::thesub()
_nadim: if it's neither "our" nor exported: not at all
BooK grondilu: advent of code looks funny 08:04
are people publishing their answers?
grondilu surely, somewhere
on reddit for instance
BooK I'd have thought github 08:05
grondilu probably as well
_nadim thnks, while I am at it, can I alias then with something like MY::<&the_alias> = ThePackage::<&thesib>
BooK adventofgolf would be an interesting twist
moritz _nadim: my &the_alias = &ThePackage::sub # iirc 08:06
08:07 Ch0c0late joined
_nadim moritz: thank you 08:07
dalek ast: 66d8f6c | moritz++ | S32-num/narrow.t:
Fix, unfudge and expand RT #126828 tests
08:08 mattp__ joined
grondilu m: say +unique [\+] 0i, |{ '^' => i, v => -i, '<' => -1, '>' => 1 }{.comb} for qw{ < ^>v< ^v^v^v^v^v}; 08:08
camelia rakudo-moar c27a00: OUTPUT«2␤4␤2␤»
_nadim I did try to find a documentaion about exporting and such but was not lucky enough to find the right one, any links?
08:09 Timbus left
grondilu should be S12, shoudn't it? 08:09
moritz _nadim: there's doc.perl6.org/routine/is%20export though it's not very much
grondilu sorry meant S11
_nadim moritz: that one I found :) 08:10
moritz _nadim: then what are you now looking for
MadcapJake grondilu: unfortunately i wrote over my part 1, but the only change was alternating between a robot and santa www.reddit.com/r/adventofcode/comm...ns/cxnwg43 08:11
hahainternet here's a weird one, the doc site for me when i 'first' load it has code with screwed up kerning 08:12
an f5 fixes it
08:12 darutoko joined
MadcapJake m: sub prefix:<u+^>(Int $i) {:2($i.base(2).split('').map({$_ ~~ '0' ?? '1' !! '0'})).base(10).Int}; say u+^123; #unsigned bitwise negation 08:15
camelia rakudo-moar 40fe92: OUTPUT«This call only converts base-2 strings to numbers; value ("0", "0", "0", "0", "0", "1", "0", "0", "0").Seq is of type Seq, so cannot be converted!␤(If you really wanted to convert ("0", "0", "0", "0", "0", "1", "0", "0", "0").Seq to a base-2 string, use …»
08:22 xinming left
MadcapJake hmm base returns a string so why doesn't that split work? 08:23
_nadim moritz: a single place whisth all the details, but I understand that I may have to put it together myself ;)
MadcapJake m: sub prefix:<u+^>(Int $i) {my $b = $i.base(2); my @c = $b.split("").map: {$_ ~~ "0" ?? "1" !! "0"}; :2(@c.join).base(10).Int}; say u+^123; 08:24
camelia rakudo-moar 40fe92: OUTPUT«8␤»
MadcapJake I think that's the wrong math :( anyone know how I could do unsigned bitwise negation?
moritz _nadim: ah, maybe doc.perl6.org/language/modules#Expo..._Importing also interests you 08:25
08:28 xinming joined
_nadim indeed :) 08:29
08:33 xinming left 08:34 xinming joined 08:36 Ven joined 08:38 xinming left, abraxxa joined
pdcawley Any chance the "we are not most men..." bit of p6advent day 6 be changed to 'we are not most people'? Doesn't affect the sense; does affect the message. 08:38
Sorry, day 8.
moritz pdcawley: afaict it's an obscure reference to some movie or another 08:39
08:39 xinming joined
DrForr And I didn't write it, I must be slipping. 08:39
pdcawley moritz: So? 08:40
dalek c: 9020737 | moritz++ | doc/Type/ (2 files):
Link to more detailed documentation on "is export" and importing
moritz pdcawley: so the reference will probably be not recognizable afterwards. Just sayin'
in the end, I'll leave it to tadzik, who wrote it.
08:42 koo8 joined 08:44 xinming left, leont joined
pdcawley moritz: I'm not sure it's that recognizable now; Google search for "we are not most men" is remarkably short list of results. 08:45
08:49 koo8 left, g4 joined, g4 left, g4 joined 08:50 leont left 08:51 CIAvash joined
_nadim Who is the author/maintainer of Test? diag does not ouput anything before a test is done. 08:51
moritz _nadim: Test is maintained alongside rakudo
_nadim: and are you sure that's not a feature of the test harness? 08:52
_nadim I hope it is not :)
which takes me to the next question? bugs are reported in github, right? 08:53
moritz _nadim: rakudo bugs are reported via email to [email@hidden.address] (which makes them show up at rt.perl.org) 08:54
m: use Test; diag "foo"; sleep 100; # should run into a timeout
camelia rakudo-moar 40fe92: OUTPUT«(timeout)# foo␤» 08:55
moritz _nadim: in my local tests, the "# foo" shows up pretty much immediately, not at test exit
_nadim: so I suspect it's the harness after all
_nadim it may very well be, I'll double check just to make sure I did not miss it then I'll report a bug and wait for someone to tell me I am completely wrong ;) 08:56
08:57 llfourn_ joined
rurban _nadim: github is good for nqp and moarvm, not rakudo 08:57
08:58 glaukommatos joined
_nadim ok 08:59
RabidGravy _nadim, it outputs to stderr immediately it is called (I just looked at the code)
nine .tell AlexDaniel no precompilation; in your module and it will compile your module every time. Its dependencies however will still be precompiled. 09:01
yoleaux nine: I'll pass your message to AlexDaniel.
09:03 glaukommatos left 09:04 xinming joined, Timbus joined 09:06 n0tjack joined
_nadim RabidGravy: I just reported 2 errors. the code says something, my eyes something else, it's probably my eyes 09:08
09:11 n0tjack left 09:12 leont joined
masak good antenoon, #perl6 09:18
yoleaux 04:47Z <skids> masak: gist.github.com/skids/97378d26a684591f1b89 # I only felt like bikeshedding tonight
09:19 matt_ joined, matt_ left, matt_ joined, matt_ left, matt_ joined
_nadim masak: good morning 09:21
masak :)
09:26 dakkar joined
RabidGravy good foo all 09:27
moritz RabidGravy: and a happy bar to you 09:30
BooK first time in my life I'm using named captures in Perl 5
guess I'm gearing up for perl 6
masak \o/ 09:31
El_Che good morning #perl6
masak BooK: early signs of an onset of Perl 6 are subtle, such as sudden usage of named captures in Perl 5
RabidGravy :)
BooK not writing regexp, though, generating them rather
masak as you do
El_Che Channels got me hooked
BooK I want to write happy code, so I let it write part of itself
El_Che Book: That's how Skynet happened 09:32
BooK El_Che: you'll never guess the name of the (work) project
moritz "Perl 6: paving the road for Skynet"
masak BooK: there's a scene in Girl Genius where they make a point of Agatha being so much of a spark that even the little machines she invents have the spark
BooK Skijnet # it's a Dutch company
masak BooK: O.O 09:33
moritz wouldn't that be a good title for an advent post :-)
masak BooK: I'm genuinely terrified now.
El_Che Book: SchijtNet for the detractors?
RabidGravy If I wanted to get the "package" of the call site of a method, is there any way to get that from callframe/backtrace/something else ?
moritz RabidGravy: once you have the code object, and it's a Routine, you can call .package on it 09:34
RabidGravy: and I think that Backtrace exposes the code object
maybe callframe too, dunno
RabidGravy ah, okay, will test in a bit, cheers. First have to go to the quack to have my non-ideal lifestyle condemned in order to get a repeat prescription 09:35
09:35 koo8 joined 09:36 domidumont joined
BooK masak: we're building a system that will first read network configs and make reports for humans to fix them, according to our world view, and eventually, as we get more confident with the system, focus on the "model" and let it write the configs and manage the devices 09:37
so yes, the name was obvious
09:38 CQ joined
BooK my job will be done when the network people ask for a "JFDI" button :-) 09:41
I'll commit that change with -m"Judgement Day" 09:42
nine callframe exposes the code object 09:43
El_Che semi random paste from the code working now: masak: we're building a system that will first read network configs and make reports for humans to fix them, according to our world view, and eventually, as we get more confident with the system, focus on the "model" and
let it write the configs and manage the devices
10:37 < BooK> so yes, the name was obvious
damn
nine callframe(1).code
El_Che I meant this: has Str $.separator = "(ノಠ益ಠ)ノ彡┻━┻"; # Yes we can utf-8!
BooK wrong paste buffer is as embarrasing as wrong window 09:44
El_Che is the modules.perl6.org repo also meant for applications?
BooK shouldn't that be applications.perl6.org ?
DrForr I've put some examples/ directories up, but not as part of any official installs. 09:45
El_Che BooK: you don't know my pain. I am a linux user working on a linux vm in a windows machine (@work)
masak BooK: "when armageddon finally arrived, it arrived because people were fully aware and decided to pick a funny name" :P 09:47
m: class A { has @.x; method m($op) { my class B { method y { $op } }; @.x.push(B.new); say .y for @.x; } }; my $s = A.new; $s.m(1); say "--"; $s.m(2)
camelia rakudo-moar 40fe92: OUTPUT«1␤--␤2␤2␤»
masak I have a question about the above eval.
I expected OUTPUT«1␤--␤1␤2␤» -- are my expectations out of whack?
and, assuming they are, how can I do something similar but which has that output? 09:48
jnthn masak: classes are not closures 09:49
So yes, you should has $.op in B and do .new(:$op) or so 09:50
masak ah, yes
arnsholt So that code would explode horribly if the B.y method was called in an outside scope?
jnthn Or, to spend lots more memory doing the same thing, you can go with a parametric role :P
masak jnthn++
jnthn masak: Given it's marked "my", it can't easily be ;)
uh, arnsholt ^^
arnsholt You could return an object, though, couldn't you? 09:51
jnthn Sure, and then you create the usual "package scoped inside lexically scoped" situation, where you'll probably see Any or so
masak *nod* 09:52
arnsholt Right, right
masak I'm actually completely fine with that answer
jnthn OK, I probably should not time at the airport to "not one tiny thing can possibly go wrong" levels :)
*cut time
Even if I am in Norway :)
bbl &
arnsholt Usually a good idea =)
Peaceful travels, jnthn 09:53
mrf Are any people here attending LPW. Or more importantly pre LPW social on Friday?
vytas mrf, I'll be at LPW and possibly in post LPW social but not pre 09:57
mrf vytas: I should likely be at the post social as well :D 09:58
10:01 glaukommatos joined 10:02 TEttinger left 10:05 glaukommatos left
DrForr I'll be there, but not for the social :( 10:06
10:07 yeahnoob left 10:08 leont left
mrf DrForr: No Beer!!! 10:10
DrForr With my leg it's probably not a good idea for me to go stumbling around :) 10:11
mrf thats true 10:13
DrForr It's pretty decent these last few days (I'm off my cane, though I still bring it with me) but it still needs to be strengthened.
10:14 Ch0c0late left 10:15 rurban left
mrf At least its getting better 10:22
10:23 koo8 left
tadzik pdcawley: yeah, I think I'll go and change that. For the record, I did check if I'm correct in assuming that "men" is for gender-neutral and "man" is for male, but I must've been wrong either way 10:25
masak nope, "men" is just the plural of "man". 10:26
moritz "men" is just the plural of "man", and both can either mean "human" or "male human", depending on context
masak there's a slow, gradual cultural shift from "of course man/men includes women" to "of course it doesn't"
and at any given time, you will have people arguing either side
Ven o/ #perl6 10:28
BooK words are hard, let's hug!
vytas there is always singular they :) 10:30
nine or just "people" 10:37
BooK or simply * 10:39
10:46 espadrine joined 10:48 Ven left
lizmat $ perl6 --ll-exception t/spec/S03-metaops/reduce.rakudo.moar 10:53
Cannot find method 'orig'
at gen/moar/stage2/QAST.nqp:5461 (/Users/liz/Github/rakudo.moar/install/share/nqp/lib/QAST.moarvm:compile_all_the_stmts:182)
this new spectest breakage goes deep into the rabbit hole
10:54 km3 left
pdcawley tadzik: Thanks. 10:56
10:59 lucs joined
nine lizmat: looks scary 11:00
sergot for vscode users: marketplace.visualstudio.com/items...l6Snippets 11:01
11:02 glaukommatos joined
masak yep, 007 is completely busted on Rakudo HEAD 11:02
lizmat m: [\orelse] Any, 0, 1 # scare breakage golfed 11:03
camelia rakudo-moar 40fe92: OUTPUT«===SORRY!===␤Cannot find method 'orig'␤»
11:03 xinming_ joined
masak lizmat: I wonder if we're looking at the same error or not. 11:04
I'm seeing "Cannot reference undeclared local 'flattening__1'"
masak bisects 11:05
I'd say whatever this is, it should be reverted on sight 11:06
11:06 brrt joined
RabidGravy sergot, that reminds me I wanted to make a Perl 6 template thingy for DevAssistant 11:06
lizmat masak: pretty sure it's TimToady's last commit
11:06 xinming left
lizmat verifying now 11:06
11:07 glaukommatos left
masak it's interesting with Travis CI nowadays, I'm much more inclined to stay on top of Rakudo changes :) 11:07
lizmat the commit before is ok 11:09
now double checking TimToady's
aka c27a00ca36b8baaa6d6
11:09 Zoffix left 11:10 znpy joined
lizmat yup, that's the first one that fails 11:10
11:11 Zoffix joined 11:16 secwang left, n0tjack joined
masak I'm not quite there yet, but the set still contains that possibility 11:19
11:20 rurban joined 11:21 n0tjack left
lucs In Perl 5, arbitrary data could be passed to a module's import() function by doing "use Foo ('some', 'data')". 11:21
How is something similar (passing arbitrary data to a module when 'use'ing it) done in Perl 6?
El_Che doe close channels get garbage collected even when the class they where created and consumes is still in scope?
nine lucs: we have EXPORT functions now 11:22
lizmat El_Che: as soon as objects cannot be reached anymore, they're up for GC at some point
lucs nine: grepping, thanks. 11:23
lizmat El_Che: not really an answer, but that's the essence: how it applies to your specific situation, I don't know
11:24 kid51 joined
El_Che lizmat: thx. If I open the channels in a method and pass the around, they won't be part of the instance state and hence I hope them to be GC'ed once the method exits (that's the rationale) 11:25
masak ok, c27a00ca36b8baaa6d668c025d76aa3c6b8bd711 confirmed as Bad. 11:26
people have about 2 minutes to raise objections before I push a revert.
Zoffix Think of the children! 11:27
CQ it's xmas season, don't refuse gifts? :)
lizmat El_Che: fwiw, if you consume the channels fast enough, I don't think there's much memory footprint to begin with 11:28
masak just confirming that the revert makes things work in 007 again.
lizmat masakL if it had been only that spectest, I would have said: no 11:29
but if it blocks you working with 007, I'm ++ on it: a commit should never block development
El_Che lizmat: I am consuming LDAP db's of 500+MB each. One channel read it as records, a second channel sort the record and change small stuff, a third write it to a new file. From the look of it, this 4gv-b VM will kill moar very soon 11:30
masak lizmat: yeah, something is clearly wrong.
lizmat: if I were motivated, I guess I could golf exactly what breaks.
lizmat: right now I only know that the 007 test suite hangs, with a Rakudo-internal error message. 11:31
indeed, the revert causes the test suite to work again.
pushing.
lizmat El_Che: well, sorting implies keeping it in memory, I guess
El_Che lizmat: only one record at the time, though 11:32
lizmat ah, so internally, sorting inside the record you mean?
El_Che yes
dalek kudo/nom: c5f81ae | (Carl Masak)++ | src/ (5 files):
Revert "get reductions thunking for left/list assoc"

This reverts commit c27a00ca36b8baaa6d668c025d76aa3c6b8bd711, which breaks, among other things, the 007 test suite with the following message:
   $ prove -r --exec=perl6 t/
   t/features/begin.t .......................... ===SORRY!===
   Cannot reference undeclared local 'flattening__1'
  (And then it hangs.)
It also breaks other things, like the spectest suite, and this:
   <lizmat> m: [\orelse] Any, 0, 1 # scare breakage golfed
   <camelia> rakudo-moar 40fe92: OUTPUT«===SORRY!===␤Cannot find method 'orig'␤»
El_Che the records are orderen, their attributes aren't, so I need to sort those before I can compare the data consistency
masak TimToady: if you're interested, I can golf the break in 007. 11:33
sergot RabidGravy++ 11:35
11:41 n0tjack joined 11:46 n0tjack left
brrt completely offtopic, but this in a way the funniest thing i've read all week 11:48
www.linkedin.com/pulse/mongodb-32-...hn-de-goes
tl;dr mongodb 'supports analytics' by letting itself be wrapped by postgresql, which is bad because....? 11:49
(because the author builds a competing database, that is why) 11:50
ilmari s/database/analytics tool/ 11:51
brrt right
11:59 snarkyboojum joined, pdcawley_ joined 12:03 glaukommatos joined 12:04 znpy left
lucs EXPORT question: gist.github.com/lucs/d2fba5bec39f8f677f4f 12:04
12:04 rindolf joined 12:08 glaukommatos left
RabidGravy lucs, afaik the EXPORT sub must be outside the package 12:09
lucs Not sure where I'd put it then :/ 12:10
Before the "unit module Foo;" line? 12:11
12:11 cygx joined
RabidGravy module Foo { .... }; sub EXPORT() { } 12:12
cygx lucs: if you want to use an explicit EXPORT sub, the easiest thing to do is not using a unit declaration
RabidGravy there has to be an advantage to using the block form 12:13
;-)
lucs RabidGravy: Having it outside like that, I'm not sure how "use Foo..." will know how to call the right sub EXPORT, but, trying ... 12:14
RabidGravy it's per "compilation unit" 12:15
lucs I'll need to understand those more clearly :) 12:16
(now, not sure where to scope my $val) 12:17
cygx lucs: workaround: 12:19
my &EXPORT; unit module Foo; &EXPORT = sub { say 42; {} }; 12:20
RabidGravy lucs, it is not all quite implemented as per design.perl6.org/S11.html
lucs cygx: The point is missing: pass the 42 with the 'use': use Foo 42; 12:21
RabidGravy: Yes, I've been reading that, but I'm not understanding as well as I'd like.
cygx my &EXPORT; unit module Foo; &EXPORT = sub ($i) { say $i; {} }; 12:22
or did you mean something else?
lucs cygx: I meant that last thing, thanks.
(I think, trying...)
12:24 kid51 left
_nadim is it possible, while running tests, to make a module not available? IE, I want to test for the non existance of Terminal::ANSIColors 12:25
12:26 Ven joined
RabidGravy ouch travis-ci.org/sergot/http-useragen...s/95370459 12:28
I was just about to say that Travis seemed to be working again 12:30
ilmari _nadim: Test::Without::Module, Devel::Hide 12:31
lizmat ilmari: that appears to be a perl 5 solution? 12:33
ilmari lizmat: sorry, I didn't notice which channel this was 12:34
12:34 rurban left
ilmari shoud go for lunch 12:34
lizmat _nadim: I'm afraid there is no solution for that yet
_nadim thank you for the information :) 12:37
12:41 rurban joined
lucs cygx: What does the {} represent in &EXPORT = sub ($i) { say $i; {} }; ? 12:42
RabidGravy the export function has to return a Map
see what happens if you don't put it there 12:43
lucs I saw that in an error message, but nowhere in the docs :/
12:43 snarkyboojum left
lucs (Well, not in S11) 12:43
And what should that map hold? 12:44
12:44 sufrostico joined
lucs cygx's example works, but I can't manage to make the Foo::val function visible in foo.pl6. 12:45
12:47 koo8 joined
RabidGravy { '&val' => sub () { "<$val>" }} 12:48
would be what you want
lucs Aha, I tried many variations around that, but not that one, trying ... 12:49
RabidGravy it's a map of the names of the things to be exported to the thing
lucs Hmm... I keep getting Could not find symbol '&val' 12:52
RabidGravy single quotes on the LHS 12:53
lucs Yep.
vytas what's the easiest way to wait for new input and eval in the same scope ?
lucs RabidGravy: Oh, I was invoking with Foo::val (fails), but with just val , it works. 12:55
RabidGravy yeah, that's what export does 12:56
lizmat vytas: loop { EVAL prompt }
vytas lizmat, sorry what i meant was that declarations and scope would stay between evals 12:57
lizmat vytas: that's tricky, perhaps something like this will do? 13:01
my $program = ''; loop { EVAL $string ~= prompt }
my $program = ''; loop { EVAL $program ~= prompt('') } # better 13:03
13:04 Skarsnik joined
vytas lizmat, wow ++ 13:04
lucs RabidGravy: Do you just happen to know all this stuff, and/or am I just reading the wrong documentation? 13:05
RabidGravy I think I just "know" it, probably just long term osmosis from the channel 13:06
lucs (or, reading the documentation wrong)
RabidGravy: Okay, thanks :)
dalek c: 480db57 | (Claudio Ramirez)++ | doc/Language/5to6-perlvar.pod:
update doc for nl-in and nl-out
 13:07
c: eef2490 | (Claudio Ramirez)++ | doc/ (3 files):
update doc for nl-in and nl-out
c: 4498203 | (Claudio Ramirez)++ | doc/ (4 files):
Merge branch 'master' of github.com:perl6/doc
c: f99e92b | RabidGravy++ | doc/ (3 files):
Merge pull request #231 from nxadm/master

Update documentation for .nl => .nl-in and .nl-out
El_Che according to the 5to6-perlvar is the autoflush var ($|) not implemented in perl6. Is there an other way to force the flush on each write? 13:08
brrt my suspicion is that there is a method for that....
El_Che I am looking ;)
vytas of course evaluating previous commands are annoying and troublesome 13:09
Skarsnik El_Che, .flush ?
$*OUT.flush ?
El_Che Skarsnik: no flush in the doc. I'll try it quickly 13:10
Skarsnik Wait I was thinking of some Qt code I was looking at. Yes IO::Handle seen to not have something to flush 13:11
El_Che it works though :) 13:12
Skarsnik m: say IO::Handle.^can("flush"); 13:13
camelia rakudo-moar c5f81a: OUTPUT«()␤»
Skarsnik m: say IO.^can("flush");
camelia rakudo-moar c5f81a: OUTPUT«Too many positionals passed; expected 2 arguments but got 3␤ in block <unit> at /tmp/gT0dC3hCG0:1␤␤»
El_Che I'll add it to the doc, if no one opposes (being something to be removed or the like) 13:14
Skarsnik m: say IO::Handle.^can("flush").elems;
camelia rakudo-moar c5f81a: OUTPUT«0␤»
El_Che it's certainly there and not marked as private
github.com/rakudo/rakudo/blob/d3e8.../Handle.pm 13:15
Skarsnik why ^can does not find it Oo 13:16
psch Skarsnik: because the sig specifies IO::Handle:D as invocant 13:17
m: class A { method foo(A:D $:) { } }; A.can('foo').say; A.new.can('foo').say
camelia rakudo-moar c5f81a: OUTPUT«(foo)␤(foo)␤»
psch ...
no
:(
Skarsnik lock/unlock not documented too x) 13:18
psch m: say IO::Handle.^mro 13:19
camelia rakudo-moar c5f81a: OUTPUT«IO::Handle is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in method gist at src/RESTRICTED.setting:33␤ in block <unit> at /tmp/xDai_dvqwg:1␤␤»
Skarsnik m: say IO::Handle.can("flush").elems;
camelia rakudo-moar c5f81a: OUTPUT«0␤»
psch *that* is why
'cause IO::Handle in camelia is an empty class that has RESTRICTED-CLASS and Mu as mro
dalek c: f6be6c0 | (Claudio Ramirez)++ | doc/Type/IO/Handle.pod:
Add method flush
c: b9f6fb0 | nxadm++ | doc/Type/IO/Handle.pod:
Merge pull request #232 from nxadm/master

Add method flush
psch Skarsnik: cf. src/RESTRICTED.setting 13:20
El_Che what a man has to do to get the memory usage down :)
Skarsnik It's sad rakudo take so much ram to build. I can't build it in cloud9 :( 13:21
El_Che build it locally and upload it? 13:22
Skarsnik my virtualbox decided that it could not see my intel virtualisation so I can't run 64 bits guess anymore x) But I will look at using a ssh workspace 13:25
grondilu wonders why we don't have an Int.ordinal method 13:26
m: say 'first', 'second'... *; 13:27
camelia rakudo-moar c5f81a: OUTPUT«(...)␤»
grondilu m: say ('first', 'second'... *)[^10];
camelia rakudo-moar c5f81a: OUTPUT«(first second secone seconf secong seconh seconi seconj seconk seconl)␤»
grondilu lol
Skarsnik fun
El_Che :)
secong would be a nice name for a language. C-Kong 13:28
Skarsnik use MONKEY-SLANG; 13:30
13:30 lustlife` left 13:34 znpy joined
dalek kudo/nom: fadbb54 | peschwa++ | src/RESTRICTED.setting:
Remove RESTRICTED override for not-anymore-existing class IOU.
 13:34
13:35 fooflare joined 13:36 n0tjack joined
MadcapJake m: sub prefix:<u+^>(Int $i) {my @bin = $i.base(2).split('')[1..*]; my @zer = @bin.unshift: |('0' xx 16 - @bin.elems); my @com = @zer.map: { $_ ~~ '0' ?? '1' !! '0' }; return :2(@com.join).base(10).Int}; say u+^123; 13:38
camelia rakudo-moar c5f81a: OUTPUT«65288␤»
MadcapJake why does that output 65288 but the adventcode question says it should output 65412??
13:38 fooflare left
[Coke] should we make Configure.pl default to moar-backend only? 13:39
Skarsnik it build parrot too? 13:40
13:40 n0tjack left
MadcapJake my intent is to make a bitwise complement that doesn't return a negative result (which to the adventcode puzzle gives you `NOT 123 -> 65412`, but I can't seem to arrive at that result) 13:41
Skarsnik Should perl6 pod be called pod6? for github.com/perl6/doc/issues/167 , well it's just an example but there is nothing to tell if a .pod is perl6 or perl5
psch m: say 123.base(2)
camelia rakudo-moar c5f81a: OUTPUT«1111011␤»
psch m: sub prefix:<u+^>(Int $i) {my @bin = $i.base(2).comb; my @zer = @bin.unshift: |('0' xx 16 - @bin.elems); my @com = @zer.map: { $_ ~~ '0' ?? '1' !! '0' }; return :2(@com.join).base(10).Int}; say u+^123; 13:43
13:43 ChristopherBotto joined
camelia rakudo-moar c5f81a: OUTPUT«65412␤» 13:43
psch MadcapJake: i'm guessing you still had a '' at the end from split('')
13:43 ZoffixW joined
ZoffixW m: subset Foo of Str where any <foo bar>; sub foo (Foo $x) { say $x }; foo 'foo'; foo 'ber'; 13:43
camelia rakudo-moar c5f81a: OUTPUT«foo␤Constraint type check failed for parameter '$x'␤ in sub foo at /tmp/qMmdyOI6DI:1␤ in block <unit> at /tmp/qMmdyOI6DI:1␤␤»
psch MadcapJake: .comb is better suited for that use-case, in any case
ZoffixW Didn't that used to mention the name of the subset in the error message?
m: subset Foo of Str where any <foo bar>; sub foo (Int $x) { say $x }; foo 'foo'; foo 'ber'; 13:44
MadcapJake psch: cool thanks!
camelia rakudo-moar c5f81a: OUTPUT«5===SORRY!5===␤Calling foo(Str) will never work with declared signature (Int $x)␤at /tmp/51QTaIggrp:1␤------> 3<foo bar>; sub foo (Int $x) { say $x }; 7⏏5foo 'foo'; foo 'ber';␤Calling foo(Str) will never work with declared signature (Int …»
_nadim please tell me thaat someone has a module that, when used instead for Test, makes the framework "tranparent". If not, and if there is more than me who wants it, I'll write it. I have always hated this in P5
ZoffixW huh. I'm getting a different error locally: Type check failed in binding $set; expected Int but got Str
[Coke] (parrot), no, it includes the jvm.
"tranparent" ? 13:45
ZoffixW _nadim, what do you mean "makes transparent"?
psch [Coke]: i'd agree with "perl Configure.pl" being equivalent to "perl Configure.pl --backends=moar --prefix=./install"
[Coke]: i usually also use --make-install, but people might not want that vOv
_nadim diag works like say ... say works normally ... is defined ... 13:46
Skarsnik ?
ZoffixW _nadim, still not following.
psch _nadim: what would &is do in that transparent framework?
[Coke] your transparent sounds like "this test framework is broken" to me. 13:47
13:47 PotatoGim joined
_nadim psch: the same thing or nothing if one would prefer 13:47
when a test fails, I want to be able to very quickly run it without the framework 13:48
ZoffixW _nadim, but how would you test things?
without is 13:49
Skarsnik perl6 -I lib t/mytest.t ?
_nadim ZoffixW: often, I need to see output, dmp data structures to fix the tests
psch _nadim: well, if it seems useful to you, go for it. i don't quite understand the purpose
ZoffixW neither
Skarsnik run the test directly, it will display everything you output 13:50
ZoffixW _nadim, I think you may simply be asking for the -v flag to prove :)
Skarsnik only prove hide it
_nadim I will probably write something, it will be easier to describe after I have written it :)
ZoffixW Sure
Write all the things! :)
psch isolating a breaking test case isn't annoying enough for me to want a module that i replace Test with instead of running the test alone
_nadim ZoffixW: not -v, I actually run prove with -v all the time
psch like, < copy & paste the breaking test-case into an -e or camelia > is usually fine with me :) 13:51
Skarsnik I run the test directly when I want to test a test x)
13:51 cygx left
ZoffixW _nadim, then I'm not at all sure what issue you're trying to solve. What you describe is hinting me at my P5 version of metacpan.org/pod/Test::Mojo::Role::Debug 13:51
_nadim psch: you are starting to get what I want, except that I don't want to copy paste
ZoffixW _nadim, so I'd say write the module and make it available :)
psch _nadim: well, as i said, if it seems useful to you go for it 13:52
_nadim that's the plan ;)
psch me personally i think copying and pasting the test case is cleaner than altering a (potentially upstream) test-file vOv
don't wanna start commit broken test files again :P
especially not to roast >_> 13:53
ZoffixW never had to copy-paste tests :S
ISAGN for DIE_ON_FIRST_FAIL type of thing. If there's a failure, stop further testing.
(I think I already mentioned that)
Skarsnik There is a DIE on FIRST FAIL? 13:54
ZoffixW Is that a question or...?
Skarsnik Yes, sorry
ZoffixW There is Test::Most in P5 that does that sort of thing. Makes doing TDD much easier
Skarsnik It was annoying when testing stuff in DBIish x) 13:55
"Oh test 10 has failed, wait there is 40 tests left"
psch m: my \blk = (-> --> Int { Str }); say blk.returns; say blk() 13:56
camelia rakudo-moar fadbb5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/VqmQ1Eq46l␤Variable '&blk' is not declared␤at /tmp/VqmQ1Eq46l:1␤------> 3 --> Int { Str }); say blk.returns; say 7⏏5blk()␤»
psch eh
13:56 prammer joined
psch m: my &blk = (-> --> Int { Str }); say blk.returns; say blk() 13:56
camelia rakudo-moar fadbb5: OUTPUT«Method 'returns' not found for invocant of class 'Str'␤ in block <unit> at /tmp/7KsFZ1L7D8:1␤␤»
psch m: my &blk = (-> --> Int { Str }); say &blk.returns; say blk()
camelia rakudo-moar fadbb5: OUTPUT«(Int)␤(Str)␤»
psch i am looking at about where that could be corrected
but i don't get it :/
ZoffixW m: class Meow { subset Foo of Str where any <foo bar>; method foo (Foo $x) { say $x }; }; Meow.new.foo: 'foo'; Meow.new.foo: 'ber'; 13:57
camelia rakudo-moar fadbb5: OUTPUT«foo␤Constraint type check failed for parameter '$x'␤ in method foo at /tmp/hc0UhYrCM0:1␤ in block <unit> at /tmp/hc0UhYrCM0:1␤␤»
ZoffixW ^^ that error is LTA. It should mention that Foo type was expected
As in this case:
m: class Meow { subset Foo of Str where any <foo bar>; has Foo $!bar = 'meow'; method foo (Foo $x) { say $x }; }; Meow.new.foo: 'foo'; Meow.new.foo: 'ber';
camelia rakudo-moar fadbb5: OUTPUT«Type check failed in assignment to $!bar; expected Meow::Foo but got Str␤ in block <unit> at /tmp/QM2m1KlW7Q:1␤␤»
RabidGravy WTF is "Serialization Error: Unimplemented case of read_ref" all about 13:58
Skarsnik In what?
RabidGravy well, in this case the t/082-exceptions.t in H::UA 13:59
psch m: my subset Foo of Str where any <foo bar>; sub f(Foo $x) { }; &f.signature.say
camelia rakudo-moar fadbb5: OUTPUT«(Str $x where { ... })␤»
psch ZoffixW: i don't know if we can still know that. as per above, the subset type constraint gets translated into a Str type with a where clause 14:00
ZoffixW: whereas the container for a variable (in the assignment case) probably knows it comes from a subset
ZoffixW Too bad. 14:02
My brag-tweet would've been more awesome with included useful error messages mentioning the subtype :) twitter.com/zoffix/status/674589605325180928
psch ZoffixW: well, feel free to look closer at the source. it starts around src/Metamodel/BOOTSTRAP.nqp:475-505
ZoffixW I don't know how to run that stuff :)
Skarsnik RabidGravy, I wil blame a weird error in sockets? 14:04
ZoffixW Like, I see the README mentions how to build the entire nqp, but do you guys rebuild the whole thing every time you make a change and just want to see what the effect was? github.com/perl6/nqp/
lizmat ZoffixW: I usually do "make install" which usually does the right thing
ZoffixW: except in a case of a NQP version bump
ZoffixW lizmat, does that take forever to run? 14:05
14:05 glaukommatos joined
psch realizes how spoiled all those "i only use r-m" people actually are... :P 14:06
ZoffixW r-m?
moritz ZoffixW: rakudo-moarvm
psch ZoffixW: fwiw, on the machine i hack on rakudo a complete rakudo-moar build - with building moar, nqp and rakudo - takes 3 minutes
ZoffixW: contrast with iirc almost 15 minutes r-j :P 14:07
lizmat my settings build cycle using make install is at about 80 seconds
this is excluding any Moar or NQP building
psch ZoffixW: if you only change stuff below $rakudo-git-path/src you only need to rebuild rakudo itself too, which as lizmat++ points out might be less than two minutes 14:08
lizmat psch: once upon a time, building with parrot to >5 minutes
*took
arnsholt Just parsing the setting took at least two minutes on this machine with Parrot, IIRC 14:09
ZoffixW k, I'll try to play around with the guts on the weekend. 14:10
arnsholt I remember being a bit envious of jnthn's machine that did the parse stage in about 100 seconds =)
ZoffixW rakudobrew build moar parse stage took 59seconds on my home box last night :)
14:11 glaukommatos left 14:12 prammer left
psch lizmat: on hack i have real 6m28.179s for r-j with existing nqp-j and an RTE during install-core-dist.pl 14:12
lizmat many modules installed maybe ? 14:13
psch and stage parse is 5 sec shy of 2 minutes
lizmat: no, Configure.pl
lizmat ah...:-) 14:14
14:14 ZoffixW left
RabidGravy "Stage parse : 182.712" on a 1.58GHz celeron :-\ 14:15
psch hack is a pretty beefy machine, i think 14:16
model name : Intel Core i7 9xx (Nehalem Class Core i7) cpu MHz : 2399.996
yup :P
14:17 yaewa joined 14:18 prammer joined 14:19 marchelzo joined 14:20 moei joined
marchelzo how can you get everything but the head of a list? there must be a better way than $mylist.tail($mylist.elems - 1) 14:20
RabidGravy why would I be getting "P6M Merging GLOBAL symbols failed: duplicate definition of symbol Response" in a test on travis but not locally? 14:21
psch m: my @a = ^10; say @a[1..*]
camelia rakudo-moar fadbb5: OUTPUT«(1 2 3 4 5 6 7 8 9)␤»
psch marchelzo: ^^^ is another way 14:22
marchelzo the whatever star is so mystifying 14:23
14:23 yaewa left
moritz m: my @a = ^10; say @a.tail 14:24
camelia rakudo-moar fadbb5: OUTPUT«(9)␤»
_nadim marchelzo: my ($ignore, @rest) = @list
marchelzo _nadim: I like that way 14:25
but if you don't want to assign the tail to anything then it's weird
Skarsnik ($head, @list) = @array; 14:26
probably work x)
_nadim m: my @list = ^5; my ($ignore, @rest) = @list ; @rest.say
camelia rakudo-moar fadbb5: OUTPUT«[1 2 3 4]␤»
grondilu m: say q{\\} ~~ /'\\'/
camelia rakudo-moar fadbb5: OUTPUT«｢\｣␤»
psch m: my @a = ^10; say [*] (($, @) = @a)[1] #
camelia rakudo-moar fadbb5: OUTPUT«362880␤»
grondilu was expecting \
gfldex m: my @a = (1..10).list; my ($,@rest) = @a; dd @rest;
camelia rakudo-moar fadbb5: OUTPUT«Array $var = $[2, 3, 4, 5, 6, 7, 8, 9, 10]␤»
_nadim m: my @list = ^5; my ($ignore, @rest) = @list ; @list.say ; @rest.say
camelia rakudo-moar fadbb5: OUTPUT«[0 1 2 3 4]␤[1 2 3 4]␤»
marchelzo psch: ok, what the fuck 14:27
psch marchelzo: well, ($, @) is a List with two elements
marchelzo right
_nadim marchelzo: what do you mean with "not want to assign the tail to anything"?
lizmat m: my @list = ^5; my ($, @rest) = @list ; @list.say ; @rest.say # no need to name the scalar to be ignnored 14:28
camelia rakudo-moar fadbb5: OUTPUT«[0 1 2 3 4]␤[1 2 3 4]␤»
psch marchelzo: and (($, @) = @a) puts the first elem of @a into the anon Scalar, and the rest into the anon Array
marchelzo I didn't know ($, @) = @a was a thing.
psch marchelzo: of that two element List we take the second element and multiply all elements of *that* afterwards
_nadim lizmat: tanks, I did try Nil but it complained heavily ;)
marchelzo I didn't know about anonymous Arrays / Scalars.
grondilu wth how do I quote a backslash?
m: say q{\}
camelia rakudo-moar fadbb5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/30d3hzcYzD␤Couldn't find terminator } (corresponding { was at line 1)␤at /tmp/30d3hzcYzD:1␤------> 3say q{\}7⏏5<EOL>␤ expecting any of:␤ }␤»
marchelzo _nadim: that I don't want to use the = operator. 14:29
psch grondilu: q[] still does some backslash interpolation
m: say Q[\]
camelia rakudo-moar fadbb5: OUTPUT«\␤»
psch grondilu: you want Q{}
grondilu ok, and in a regex?
psch m: say q[\qq<\c[SNOWMAN]>] 14:30
camelia rakudo-moar fadbb5: OUTPUT«☃␤»
masak "{ Q{} }" should work in the worst case
_nadim marchelzo: doc.perl6.org/routine/splice#class_List but why not use the above 14:32
grondilu m: say q{\\} ~~ /\x5C ** 2/
camelia rakudo-moar fadbb5: OUTPUT«Nil␤»
grondilu m: say Q{\\} ~~ /\x5C ** 2/ 14:33
camelia rakudo-moar fadbb5: OUTPUT«｢\\｣␤»
14:33 kaare_ joined
marchelzo _nadim: I will in cases where I want to assign the tail to something 14:33
grondilu m: say Q{\\} ~~ /'\\'/
camelia rakudo-moar fadbb5: OUTPUT«｢\｣␤»
grondilu ^what am I missing here?
marchelzo it's good to know multiple ways I guess
moritz grondilu: that \ escapes, even in single-quoted strings 14:34
grondilu so what the equivalent of Q{} inside a regex? 14:35
_nadim then put something into the construct. ($, @the_thing_you_want_to_assigne_to) = @source 14:36
marchelzo: ^^
marchelzo right
grondilu m: say Q{\\} ~~ /'\\\\'/
camelia rakudo-moar fadbb5: OUTPUT«｢\\｣␤»
marchelzo ok thanks
grondilu m: say Q{\\} ~~ /\\\\/ 14:37
camelia rakudo-moar fadbb5: OUTPUT«｢\\｣␤»
grondilu ^that will do I guess
14:39 jdrab joined 14:45 sunnavy joined 14:52 zakharyas joined 14:55 Ven left 14:57 skids joined
moritz eek 14:57
found a very nasty test code bug at work
lizmat eek!
moritz which would, under some not-too-obscure conditions, swallow all exceptions 14:58
and since test data setup is very fiddly, test writes tend to avoid doing it, and stuff their test into existing files with existing test data
colomon found a bug a few days ago for $work which under obscure-ish conditions would cause his code to process entirely incorrect data.
hahainternet just fixed a bug that resulted in 'nonsense' data production, as it mistook one field for another in a cache 14:59
python unfortunately
moritz so now we have a few a test files with huge amounts of tests in each, and some of them are downright bogus, and nobody ever noticed
hahainternet: I had a similar-ish bug recently, where a cache would be written by a wrong key, causing people to be mis-identified in our ticket system 15:00
not at login time, but at "send this person a notification" time
15:00 loren joined 15:01 rurban left
loren evening, perl6, long time no see.. 15:01
15:01 rurban joined
hahainternet moritz: i was naive and thought i could use a postgres array with a key field first, and then the rest of the fields deterministically ordered 15:01
later switching to the key field just being part of the array order, but didn't switch from using array[0] as the search key 15:02
whoops :D
this is why i am so interested in strongly type restricted ORMs in p6
with a bit of annotation i could easily have had PrimaryKey and FilterKey types, and at least noticed there was something wrong :D 15:03
also hi @ loren
15:03 AlexDaniel joined
moritz hahainternet: an assumption later, silently invalidated. Yes, that's a problem 15:04
15:04 glaukommatos joined
moritz hahainternet: in my case (test bug) it was reckless use of "goto" 15:04
Skarsnik feel free to write a orm that use attr.WHY x)
ChoHag Does :exists work on Positionals?
loren I give up build perl6 under msys2 , there are many path problem. Though i success make it once, i wonder it's a accident.... 15:05
hahainternet, hi ..
hahainternet Skarsnik: i learned about .WHY only yesterday, looking forward to finding a use 15:06
loren ChoHag, may be not, i think ..
lizmat m: my @a; @a[3] = Any; say @a[*]:exists 15:07
camelia rakudo-moar fadbb5: OUTPUT«(False False False True)␤»
lizmat ChoHag: ^^^
moritz lizmat++ # data points for RT #126842 15:08
15:08 dalek left
jnthn wtf :/ 15:08
dalek--
lizmat moritz: fwiw, the run with nqp::force_gc is still running at 22K iterations
Skarsnik hahainternet, could probably be used to say has $.id; #db=PrimaryKey and have an ORM class mapping that x)
15:08 glaukommatos left 15:09 dalek joined, ChanServ sets mode: +v dalek
hahainternet wb dalek 15:09
15:09 llfourn_ left
loren lizmat, will be work like this? sub my(Str $str?) { if $str:exists { } } 15:09
lizmat loren: that's not a Positional 15:10
ChoHag Nifty.
jnthn Anyway, I just pusehd a few commits, one of which adds a `use experimental`, and is cached is now needing `use experimental :cached`.
RabidGravy $str.defined
jnthn Also fixed that the newline pragma wasn't installed by make install.
loren eh, lizmat
jnthn loren: Just use if $str.defined 15:11
lizmat loren: something more Perl 6ish: 15:12
m: my $a; with $a { .say }
camelia ( no output )
lizmat m: my $a = 42; with $a { .say }
camelia rakudo-moar fadbb5: OUTPUT«42␤»
jnthn yes, or with :)
psch jnthn: do you have an idea re: RT #126232? i had thought adding a p6typecheckrv to the block could work, but it doesn't 15:13
15:13 jdrab left, Ven joined
psch m: (-> --> Int { Str })() # for reference 15:13
camelia ( no output )
loren I just think isn't it $str:eixsts cool than $str.defined, haha 15:14
jnthn, lizmat em, got it
masak loren: no, :exists is an adverb on hash access
lizmat masak: or array access 15:15
masak or what lizmat said
m: my %h; say %h<foo> :exists
camelia rakudo-moar fadbb5: OUTPUT«False␤»
ChoHag m: my %x = a => 42; say %x<b>:!exists; say !(%x<b>:exists); say !%x<b>:exists
camelia rakudo-moar fadbb5: OUTPUT«True␤True␤Unexpected named parameter 'exists' passed␤ in block <unit> at /tmp/0JAhK8EoFz:1␤␤»
masak loren: the above should be read as `<foo> :exists` -- that's a "unit"
loren I know that .. , masak
ChoHag Why is the third say statement an error?
masak loren: so with `$str :exists`, the adverb is not "sitting on" some operator. and it must do that.
ChoHag Is ! doing something I don't know about? 15:16
15:16 rurban left
moritz ChoHag: it's a precedence problem 15:16
loren If we can use like that, every variable can use :exists, may be it can be a postfix ? 15:17
moritz ChoHag: I guess the adverb binds to the ! instead of the <>
m: my %x = a => 42; say !(%x<b>:exists)
camelia rakudo-moar fadbb5: OUTPUT«True␤»
moritz m: my %x = a => 42; say not %x<b>:exists
camelia rakudo-moar fadbb5: OUTPUT«True␤»
gfldex you use ! unless you use unless 15:19
gfldex chuckles
CQ you could double negate it with !unless! ...?
loren m: my %x = a => 42; say !(%x<b>:exists); if !(%x<b>:exists) { say 'True' }; unless (%x<b>:exists) { say 'True' } 15:20
ChoHag m: my %x = a => 42; say not:exists %x<b>
camelia rakudo-moar fadbb5: OUTPUT«True␤True␤True␤»
rakudo-moar fadbb5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/1y0MB7M9ET␤Undeclared routine:␤ x used at line 1␤␤»
ChoHag m: my %x = a => 42; say !:exists %x<b>
camelia rakudo-moar fadbb5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/dhbNgODBLX␤Undeclared routine:␤ x used at line 1␤␤»
masak ChoHag: in the case of `!`, it's "above" `<b>` in the AST, and so `:exists` binds to it
ChoHag: an adverb basically looks back in the source code for something to bind to 15:21
ChoHag: but (and this is key) it doesn't take the textually closest thing it finds, but the AST-ly topmost thing
moritz sounds like worth of an FAQ entry 15:22
ChoHag Adverbs are going to needs a lot of clear explaining. They're a very weird thing. 15:23
dalek kudo/nom: 2f91998 | jnthn++ | / (3 files):
Make macro/quasi experimental.
ast: 4039698 | jnthn++ | / (17 files):
Add `use experimental :macros` as needed.
 15:24
15:25 zakharyas left 15:26 mdaq left 15:33 khw joined
zengargoyle good * #perl6 15:39
loren zengargoyle, o/ 15:40
15:40 molaf joined
AlexDaniel ｢0..($wolf_quantitiy-1)｣ hmmm… ｢0..^$wolf_quantity｣ ! 15:40
yoleaux 09:01Z <nine> AlexDaniel: no precompilation; in your module and it will compile your module every time. Its dependencies however will still be precompiled.
15:41 hankache joined
hankache hello #perl6 15:41
psch hm, i did figure something out to make < (-> --> Int { Str })() > throw, but that complains about not being able to serialize a Block during install-core-dist.pl6 :/
AlexDaniel: probably just < ^$wolf_quantity > instead of < 0..^$wolf_quantity >? 15:42
AlexDaniel psch: correct, that's the next step :)
loren m: say ^10;
camelia rakudo-moar 2f9199: OUTPUT«^10␤»
psch i mean, they are functionally identical still..
hankache m: say (^10);
camelia rakudo-moar 2f9199: OUTPUT«^10␤»
psch m: say @ = ^10; say eager ^10 15:43
camelia rakudo-moar 2f9199: OUTPUT«[0 1 2 3 4 5 6 7 8 9]␤(0 1 2 3 4 5 6 7 8 9)␤»
loren m: say [^10];
camelia rakudo-moar 2f9199: OUTPUT«[0 1 2 3 4 5 6 7 8 9]␤»
hankache m: say |^10
camelia rakudo-moar 2f9199: OUTPUT«0123456789␤»
RabidGravy resorts to adding "diag HTTP::Resonse.^mro>>.perl" to try and understand what is going on in travis
hankache m: say ^10.WHAT 15:45
camelia rakudo-moar 2f9199: OUTPUT«Potential difficulties:␤ Precedence of ^ is looser than method call; please parenthesize␤ at /tmp/OPaSTs7mBW:1␤ ------> 3say ^107⏏5.WHAT␤Use of uninitialized value of type Int in numeric context in block <unit> at /tmp/OPaSTs7mBW:1…»
AlexDaniel m: say ^10 .WHAT
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
hankache m: say (^10).WHAT
camelia rakudo-moar 2f9199: OUTPUT«(Range)␤»
hankache voila! 15:46
AlexDaniel m: say (|^10) .WHAT
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
AlexDaniel m: say (|^10).WHAT
camelia rakudo-moar 2f9199: OUTPUT«(Slip)␤»
hankache oh!! i thought that would return a list ^^
psch m: say ^10 .list.WHAT 15:47
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
AlexDaniel what's wrong with ｢ .WHAT｣ ?
psch m: say (^10) .list.WHAT
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
psch m: say (^10).list.WHAT
camelia rakudo-moar 2f9199: OUTPUT«(List)␤»
AlexDaniel m: say 5 .WHAT
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
psch huh, when did that break
AlexDaniel camelia: help
camelia AlexDaniel: Usage:
..<(nqp-js|star-j|rakudo-MOAR|niecza|nqp-parrot|rakudo-moar|p5-to-p6|debug-cat|pugs|nqp-jvm|nqp-moarvm|star-m|prof-m|std|rakudo-jvm|rPn|nPr|r-jvm|rj|rn|nrP|r-j|nqp|perl6|rakudo|Prn|star|n|nqp-m|nr|sm|nom|rnP|nqp-mvm|P|M|p56|Pnr|m|r|rm|sj|p6|nqp-q|r-m|nqp-p|nqp-j|j)(?^::\s(?!OUTPUT)) $perl6_program>
hankache m: 5.WHAT.say 15:48
camelia rakudo-moar 2f9199: OUTPUT«(Int)␤»
hankache m: say 5.WHAT
camelia rakudo-moar 2f9199: OUTPUT«(Int)␤»
AlexDaniel hmm, no way to tell camelia to run it on 2015.11?
hankache AlexDaniel you have a space after 5
AlexDaniel hankache: so what
psch star-m: say 5 .WHAT 15:49
camelia star-m 2015.09: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YDmUQYVYQD␤Two terms in a row␤at /tmp/YDmUQYVYQD:1␤------> 3say 57⏏5 .WHAT␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement m…»
psch ah, too old for that
AlexDaniel hankache: I bet that it was working
dalek c: 606d556 | (Wenzel P. P. Peppmeyer)++ | doc/Language/traps.pod:
add trap of adverb precedence
c: 6f59d83 | (Wenzel P. P. Peppmeyer)++ | doc/Language/traps.pod:
Merge pull request #233 from gfldex/master

add trap of adverb precedence
AlexDaniel m: 5 .say
camelia rakudo-moar 2f9199: OUTPUT«5␤»
AlexDaniel m: 5 .WHAT
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
AlexDaniel m: 5 .WHICH 15:50
camelia ( no output )
AlexDaniel m: 5 .sin.say
camelia rakudo-moar 2f9199: OUTPUT«-0.958924274663138␤»
hankache AlexDaniel: dunno just saying
ilmari m: 5 .WHAT .say
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
AlexDaniel hankache: look ↑ it works
psch m: say 5 .++
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/TvwB0m7vKs␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/TvwB0m7vKs:1␤------> 3say 5 .+7⏏5+␤»
AlexDaniel but not with .WHAT
gfldex m: my %x = a => 42; say !%x<b>:exists; CATCH { default { .WHAT.say } } 15:51
camelia rakudo-moar 2f9199: OUTPUT«(AdHoc)␤»
hankache m: say (5) .WHAT
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
gfldex not sure i like that to be AdHoc
hankache ok i give up
no spaces for WHAT
ilmari m: pi.sin.say
camelia rakudo-moar 2f9199: OUTPUT«1.22464679914735e-16␤»
hankache AlexDaniel: is this a bug? 15:52
AlexDaniel m: π . sin . say
camelia rakudo-moar 2f9199: OUTPUT«1.22464679914735e-16␤»
AlexDaniel hankache: yes, probably
15:52 prammer left
hankache m: say "helllo" .WHAT 15:52
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
psch m: 5 .Real.WHAT.say # this is the curious bit imo
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
AlexDaniel if not, then at least it is LTA because it lacks the line number… 15:53
AlexDaniel submits rakudobug…
ilmari m: say sin pi/2
camelia rakudo-moar 2f9199: OUTPUT«1␤»
hankache m: say 5 . Real
camelia rakudo-moar 2f9199: OUTPUT«5␤»
AlexDaniel j: π . WHAT
hankache weird
camelia rakudo-jvm 6c0f93: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
AlexDaniel hankache: one reason why you might want to use it is in ｢^10 .say｣ case 15:54
hankache: otherwise you need parens
m: ^10.say
camelia rakudo-moar 2f9199: OUTPUT«Potential difficulties:␤ Precedence of ^ is looser than method call; please parenthesize␤ at /tmp/yznGid3Yyh:1␤ ------> 3^107⏏5.say␤WARNINGS:␤Useless use of "^" in expression "^10.say" in sink context (line 1)␤10␤»
AlexDaniel m: ^10 .say
camelia rakudo-moar 2f9199: OUTPUT«^10␤»
AlexDaniel m: (^10).say
camelia rakudo-moar 2f9199: OUTPUT«^10␤» 15:55
hankache AlexDaniel aha ok
psch cb25b2f475f31335d77d018474482aceec2c74de is the original "spaces around infix:<.>" commit, checking if that actually worked for .WHAT before
i'm pretty sure it did, but it's a good start for a bisect anyway :P
15:55 abraxxa left
hankache I am a big fan of putting parentheses everywhere 15:56
AlexDaniel psch: oh wow, you are bisecting it. I'll wait then
RabidGravy okay I put that diag in and now the test passes, this actually sucks
hankache i always get precedence messed up
AlexDaniel hates parentheses… 15:57
psch AlexDaniel: well, the RT fix i was working on is running into CURLI-related problems, which i don
't quite feel up to look into
hankache should learn more about precedence 15:58
AlexDaniel m: ^10 .^methods.say
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/IVXMYbL_BG␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/IVXMYbL_BG:1␤------> 3^10 .^7⏏5methods.say␤»
AlexDaniel hmm…
hankache . to concatenate!! 15:59
zengargoyle AlexDaniel: ^methods is a macro or something...
hankache Devil get out of camelia!!
psch m: Nil .?say
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/UQCoK1BjKy␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/UQCoK1BjKy:1␤------> 3Nil .?7⏏5say␤»
psch only the normal method dot works with space
all the mutators for it don't
hankache starts incantation
psch because we look for an ident directly after the dot (minus whitespace) 16:00
AlexDaniel I'm not sure if it should not work…
but maybe that's the price for this awesome error message
which in this case is less than awesome
psch well, .WHAT anywhere in a space-y method call chain didn't work with the original commit either 16:01
so it's at least not a regression
AlexDaniel okay, interesting
psch probably 'cause .WHAT doesn't go the normal QAST::Op(:op<callmethod>,...) route
moritz right, it's macro-Y 16:02
psch which fits the error message, 'cause we apparently build such an Op but don't populate the children correctly, because .WHAT isn't a method that can be called with callmethod
hankache what is a slip? 16:04
16:04 znpy left
hankache the result of flattening? 16:04
m: say (|^10).WHAT 16:05
camelia rakudo-moar 2f9199: OUTPUT«(Slip)␤»
AlexDaniel m: say 5 . WHICH
camelia rakudo-moar 2f9199: OUTPUT«Int|5␤»
hankache and what's the difference between slip and list
?
pyrimidine I'm seeing an error pop up that appears to be from precomp: gist.github.com/cjfields/ebe5d55f76041b34e6eb
The plugin in this case is loaded dynamically here: github.com/cjfields/bioperl6/blob/...IO.pm6#L36 16:07
Based on the timing, I think it's in the plugin, not in SeqIO. any way to debug this? 16:09
Skarsnik No idea, there is this bug with other method in similar code
nine pyrimidine: perl6 --ll-exception might give us further information 16:10
pyrimidine nine: will give that a try
marchelzo is lines()[0] the best way to read one line? 16:11
ugexe .get 16:12
marchelzo ugexe: thanks
16:13 FROGGS joined
AlexDaniel by the way, what is ｢.?｣ ? 16:14
m: Nil.?say
camelia rakudo-moar 2f9199: OUTPUT«Nil␤»
ugexe call method say if it exists
AlexDaniel right
thanks
16:15 brrt left
pyrimidine nine: bingo, worked, see second gist here: gist.github.com/cjfields/ebe5d55f7...-exception 16:15
nine: I suppose I could also submit as a bug. Strangely enough it appears to work, but I do see some oddities with the class in the plugin which I need to diagnose a bit more. Could try turning off precomp for that module, see if it fixes the issue 16:19
16:20 dwarring left
psch m: say ({ a => 1, b => 2 }).WHAT 16:23
camelia rakudo-moar 2f9199: OUTPUT«(Hash)␤»
psch hm
that's really weird 16:24
16:24 marchelzo left
psch like, i've added on line to Perl6::Actions.pblock, and that makes that bit of code say (Block) instead 16:25
[Coke] ORMS: just say no.
16:26 zengargoyle left, joeschmoe joined, hankache left
psch huh, maybe not... 16:26
[Coke] jnthn++ #experimental!
16:26 zengargoyle joined
joeschmoe m: 'oldstring' ~~ 's/oldstring/newstring/'; .say 16:28
camelia rakudo-moar 2f9199: OUTPUT«(Any)␤»
joeschmoe I want it to say newstring what am i doing wrong 16:29
ilmari m: $_ = 'oldstring'; s/oldstring/newstring/; .say 16:30
camelia rakudo-moar 2f9199: OUTPUT«newstring␤»
joeschmoe thanks i was missing $_ =
the default variable
ilmari also, you were just smart-matching one string against another 16:31
FROGGS joeschmoe: don't wrap your regexes in quotes
m: 'oldstring' ~~ s/oldstring/newstring/; .say 16:32
camelia rakudo-moar 2f9199: OUTPUT«Cannot modify an immutable Str␤ in block <unit> at /tmp/ZQt37OfOCc:1␤␤»
joeschmoe m: 'oldstring' ~~ s/oldstring/newstring/; .say
camelia rakudo-moar 2f9199: OUTPUT«Cannot modify an immutable Str␤ in block <unit> at /tmp/4W8daPWGUs:1␤␤»
Hotkeys hello people of perland
joeschmoe m: $_ = 'oldstring' ~~ s/oldstring/newstring/; .say
camelia rakudo-moar 2f9199: OUTPUT«Cannot modify an immutable Str␤ in block <unit> at /tmp/p2yDVz2SiK:1␤␤»
FROGGS joeschmoe: use the .subst method if you want to operate on constant strings
psch hrm, how *does* pblock know if it's a Block or a Hash..? :/
FROGGS psch: it guesses
joeschmoe m: $_ = 'oldstring';s/oldstring/newstring/; .say 16:33
camelia rakudo-moar 2f9199: OUTPUT«newstring␤»
16:33 ab6tract joined
joeschmoe yeah i did it 16:33
ab6tract o/
FROGGS psch: grep for 'is_hash' in the actions
psch FROGGS: ahh, that helps. although i don't know if that's too late to attach the p6typecheckrv... 16:34
16:35 rhr joined
psch but that does explain why everything turned into Blocks for me with adding the p6typecheckrv in pblock 16:35
joeschmoe m: $_ = 'oldstring';d/old/; .say 16:36
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/AlXlGaNaWQ␤Missing required term after infix␤at /tmp/AlXlGaNaWQ:1␤------> 3$_ = 'oldstring';d/old/7⏏5; .say␤ expecting any of:␤ prefix␤ term␤»
joeschmoe what is the delete letter substitute is s/../ what is delete 16:37
16:38 marchelzo joined
marchelzo how reverse a Str? 16:38
I'm a native English speaker, I swear.
FROGGS m: say 'hello'.flip
camelia rakudo-moar 2f9199: OUTPUT«olleh␤»
FROGGS marchelzo: ^^ 16:39
marchelzo ah, so _that's_ what flip does
FROGGS :D
psch m: $_ = "foo"; tr:d/o//; .say
camelia rakudo-moar 2f9199: OUTPUT«f␤»
marchelzo is flip only for strings?
psch m: $_ = "foo"; tr:d/o/u/; .say
camelia rakudo-moar 2f9199: OUTPUT«fuu␤»
psch hm
marchelzo I was thinking .comb.reverse.join
psch i thought i had implemented tr:d :o
16:39 lucasb joined
FROGGS marchelzo: I think so... .reverse is for lists 16:39
joeschmoe so i need tr: 16:40
_nadim Hi, there a class that is either a Str or an Int?
ilmari m: 'hello'.comb.reverse.join.say
camelia rakudo-moar 2f9199: OUTPUT«olleh␤»
mrf _nadim: Cool
IIRC
FROGGS _nadim: can you rephrase that?
_nadim thanks
FROGGS m: say Str.^mro; 16:41
ilmari m: <42>.WHAT.say
camelia rakudo-moar 2f9199: OUTPUT«((Str) (Cool) (Any) (Mu))␤»
rakudo-moar 2f9199: OUTPUT«(IntStr)␤»
_nadim is there a class that can contain a Str or an Int
FROGGS m: say Int.^mro;
camelia rakudo-moar 2f9199: OUTPUT«((Int) (Cool) (Any) (Mu))␤»
psch m: my $x = <5>; say $x.isa(Str); say $x.isa(Int)
camelia rakudo-moar 2f9199: OUTPUT«True␤True␤»
ilmari m: say IntStr.^mro
camelia rakudo-moar 2f9199: OUTPUT«((IntStr) (Int) (Str) (Cool) (Any) (Mu))␤»
FROGGS _nadim: an IntStr is basically a string that look like an Int
mrf I thought IntStr was for quoted numbers. rather than Int || Str. 16:42
joeschmoe m: $_ = "your dumb smart"; tr:d/d*b//; .say
camelia rakudo-moar 2f9199: OUTPUT«your um smart␤»
joeschmoe yeah i did it cool
what is the "tr:" thing called 16:43
FROGGS translate
joeschmoe got it
thanks Froggs and psch 16:44
m: $_ = "your dumb smart"; tr:d/r*s//; .say 16:45
camelia rakudo-moar 2f9199: OUTPUT«you dumb mat␤»
joeschmoe m: $_ = "your dumb smart"; tr:d/d..b//; .say
camelia rakudo-moar 2f9199: OUTPUT«your dumb smart␤»
RabidGravy okay let's try this all again 16:46
joeschmoe m: $_ = "your dumb smart"; tr:d/dumb//; .say
camelia rakudo-moar 2f9199: OUTPUT«yor sart␤»
AlexDaniel you need substitute I guess
joeschmoe m: $_ = "your dumb smart"; tr:d/"dumb"//; .say
camelia rakudo-moar 2f9199: OUTPUT«yor sart␤»
AlexDaniel m: $_ = "your dumb smart"; s/dumb\s+//; .say
camelia rakudo-moar 2f9199: OUTPUT«your smart␤»
FROGGS you're* 16:47
(btw)
joeschmoe Your're smart FROGGS
16:47 n0tjack joined
FROGGS :P 16:47
AlexDaniel FROGGS: well, unless you copy-paste it 16:48
16:48 znpy joined
joeschmoe So my ultimate goal is to get rid of all the tracking junk in google links via a regex so i can copy and paste it into instapaper 16:49
AlexDaniel m: $_ = "your dumb smart"; s/dumb\s+//; "$_ dog".say
camelia rakudo-moar 2f9199: OUTPUT«your smart dog␤»
marchelzo Why doesn't this work? (*.sort)([1,3,1])
AlexDaniel marchelzo: what is it supposed to do?
marchelzo AlexDaniel: be equivalent to [1,3,1].sort 16:50
AlexDaniel m: sort([1,3,1]).say
camelia rakudo-moar 2f9199: OUTPUT«(1 1 3)␤»
joeschmoe That's good Alex
marchelzo AlexDaniel: I'm not woried about how to actually sort the list, I'm just trying to learn how the Whatever star works.
psch m: say (*.sort)([1,3,1]) # ...? 16:51
camelia rakudo-moar 2f9199: OUTPUT«(1 1 3)␤»
marchelzo when I tried it in the repl it said: cannot stringify this
I didn't use say, though.
psch m: say (*.sort)([1,3,1]).WHAT # ...? 16:52
camelia rakudo-moar 2f9199: OUTPUT«()␤»
marchelzo Why does that make a difference?
psch m: say ((*.sort)([1,3,1])).WHAT # ...?
camelia rakudo-moar 2f9199: OUTPUT«()␤»
16:52 ZoffixW joined
AlexDaniel what is () ? 16:52
ZoffixW m: say ٦1٥٠3 + ٤६੬៩ - ৭۹੧
camelia rakudo-moar 2f9199: OUTPUT«65381␤»
ZoffixW grins
psch marchelzo: no idea, the REPL has been in varying states of weird for pretty much ever
16:52 n0tjack left
marchelzo psch: ok, well I'm glad that my intuition was right 16:53
16:55 ZoffixW left 16:56 cognominal left 16:57 prammer joined
TimToady m: say Q{\\} ~~ /｢\\｣/ 16:58
camelia rakudo-moar 2f9199: OUTPUT«｢\\｣␤»
TimToady grondilu: ^^
m: say ｢\\｣ ~~ /｢\\｣/ 16:59
camelia rakudo-moar 2f9199: OUTPUT«｢\\｣␤»
AlexDaniel m: say ｢hello world｣ 17:00
camelia rakudo-moar 2f9199: OUTPUT«hello world␤»
AlexDaniel what's the difference between ｢｣ and ‘’?
TimToady single quotes still recognize a few backslashes 17:01
in particluar, the trailing quote, and backslash itself
17:01 prammer left
AlexDaniel m: say 'hello \\ world' 17:01
camelia rakudo-moar 2f9199: OUTPUT«hello \ world␤»
AlexDaniel m: say ｢hello \\ world｣
camelia rakudo-moar 2f9199: OUTPUT«hello \\ world␤»
psch and \qq 17:02
AlexDaniel right
TimToady ｢\\｣ is just the non-Texas version of Q[\\]
psch m: say 'a \qq[\c[SNOWMAN]]'
camelia rakudo-moar 2f9199: OUTPUT«a ☃␤»
TimToady m: say 'a pair of \Q[\\]' # aww 17:04
camelia rakudo-moar 2f9199: OUTPUT«a pair of \Q[\]␤»
marchelzo Why is +(123.List) == 1? 17:05
AlexDaniel marchelzo: only one element in the list
marchelzo >.<
that's obvious
sorry
ilmari marchelzo: because a list in numeric context is the number of elements
AlexDaniel m: say +[8, 25, 42]
camelia rakudo-moar 2f9199: OUTPUT«3␤»
RabidGravy I'm glad panda-test is working again
AlexDaniel ilmari: sounds very p5-ish :) 17:06
ilmari m: say ~[3,5,7]
camelia rakudo-moar 2f9199: OUTPUT«3 5 7␤»
AlexDaniel m: say ?[3,5,7]
camelia rakudo-moar 2f9199: OUTPUT«True␤»
lizmat m: say [+] 3,4,7
camelia rakudo-moar 2f9199: OUTPUT«14␤»
17:06 Ven left
lizmat m: say [**] 3,4,7 17:07
camelia rakudo-moar 2f9199: OUTPUT«1427701207893295829254572822096315603914668438510753143971001074259359396579389699820627167240291515779836466627227091534232812571357130969315818188029243277766289210859772322572719318511462543138040955367493331114541581266774920391602594047652970259262924…»
lizmat afk for a bit&
AlexDaniel m: say [R**] 2,32
camelia rakudo-moar 2f9199: OUTPUT«1024␤»
dj_goku .messages
17:09 Actualeyes left
AlexDaniel m: say [infix:<+>] 2,4,8 17:10
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/pA5leA_W3x␤Two terms in a row␤at /tmp/pA5leA_W3x:1␤------> 3say [infix:<+>]7⏏5 2,4,8␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ …»
17:10 Ven joined
AlexDaniel is there any way to make it work? 17:10
TimToady m: say [[&infix:<+>]] 2,4,8
camelia rakudo-moar 2f9199: OUTPUT«14␤»
AlexDaniel why double [[]]?
psch m: say 2 [&infix:<+>] 2
camelia rakudo-moar 2f9199: OUTPUT«4␤»
TimToady outer one is reduce, inner one turns function into infix 17:11
17:11 domidumont left
TimToady only works with & though 17:11
17:12 rurban joined
lucasb m: say List.() 17:13
camelia rakudo-moar 2f9199: OUTPUT«()␤»
lucasb m: say List.new.()
camelia rakudo-moar 2f9199: OUTPUT«Invocant requires a type object of type List, but an object instance was passed. Did you forget a 'multi'?␤ in block <unit> at /tmp/UwnVEiNbnS:1␤␤»
TimToady m: say [[&[+]]] 2,4,8
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/sL27fn93sc␤Unable to parse expression in bracketed infix; couldn't find final ']' ␤at /tmp/sL27fn93sc:1␤------> 3say [[&7⏏5[+]]] 2,4,8␤»
lucasb List needs to answer to 'call me' for some reason?
CIAvash RabidGravy: when I tried to install HTTP::UserAgent, I was getting the same error as travis, now after your change I'm getting this: # Block Code Any Mu
Default constructor for 'Block' only takes named arguments
in block <unit> at t/030-cookies.t:172
17:13 joeschmoe left
_nadim can't call method 'perl' on a null object. what's a null object? 17:14
dj_goku m: say [+] [1, 2, 3]
camelia rakudo-moar 2f9199: OUTPUT«6␤»
RabidGravy which is odd, which is quite what I expected to get but it passes now
AlexDaniel _nadim: what is the code? :)
dj_goku m: say [+] 1,2,3
camelia rakudo-moar 2f9199: OUTPUT«6␤»
RabidGravy CIAvash, I would suggest getting a new rakudo as it isn't doing it now 17:15
TimToady m: say 2 [&[+]] 2 17:16
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/w6sedakjos␤Unable to parse expression in bracketed infix; couldn't find final ']' ␤at /tmp/w6sedakjos:1␤------> 3say 2 [&7⏏5[+]] 2␤»
TimToady hmm, I wonder why that doesn't work
AlexDaniel how can a postfix take two parameters?
_nadim AlexDaniel: nopaste.linux-dev.org/?878914 just figuring out things
dj_goku .tell Zoffix just wanted to say thanks for the IRC Client advent post. I took the plung on prototyping a interacting with gearmand in perl6. I was able to get connected last night!
yoleaux dj_goku: I'll pass your message to Zoffix.
RabidGravy there's something righteously messed up here "Type check failed for return value; expected NativeCall::Types::CArray[uint8] but got NativeCall::Types::CArray[uint8].new"
AlexDaniel _nadim: can you golf it down? 17:17
marchelzo How can I do custom formatting with a list of say 2 things? Like ('Foo', '32').fmt('%s -> %d') or something.
The above doesn't work, but maybe you can see what I mean. 17:18
17:18 egor joined
TimToady .tell masak I wouldn't mind a golf of the failure, or better yet, a spectest 17:18
yoleaux TimToady: I'll pass your message to masak.
_nadim AlexDaniel: yes but I have something else right now. but the my( [@a, @b, @c]) = ... is the place that bombs it I think 17:19
CIAvash RabidGravy: I tried after updating rakudo. Do I need to uninstall and reinstall rakudo?
psch m: printf '%s -> %d', |('Foo', '32')
camelia rakudo-moar 2f9199: OUTPUT«Foo -> 32»
17:19 BuildTheRobots joined, egor left
AlexDaniel m: say sprintf("%s → %s", [42, 69]) 17:19
camelia rakudo-moar 2f9199: OUTPUT«42 → 69␤»
marchelzo hmm 17:20
AlexDaniel m: say sprintf("%s → %s", [42, 69, 20])
camelia rakudo-moar 2f9199: OUTPUT«Directives specify 2 arguments, but 3 arguments were supplied␤␤»
_nadim sub xxx { my @a, @b, @c ... return @a,@b,@c) .... my (@a, @b, @c) = xxx. @a gobbles it all, how does one write that?
AlexDaniel return [@a, @b, @c] perhaps? 17:21
17:21 zakharyas joined
_nadim Nope, I tried that first 17:21
RabidGravy CIAvash, I don't know, I couldn't replicate locally which is why I put that diag in
lucasb ^^ I found strange that sprintf flattened the list second argument 17:22
[Coke] /home/coke/sandbox/perl6-roast-data/rakudo.moar-jit/perl6.moarvm t/spec/integration/advent2013-day14.t hanging again in daily test run
AlexDaniel _nadim: Ok I'm looking into it 17:23
dalek rl6-roast-data: d4dc4cb | coke++ | / (9 files):
today (automated commit)
17:23 Ven left
tadzik pdcawley_ moritz timotimo: I just learned, from a fellow anthropologist, that they often specifically mark each use of "men" in articles with a note saying "I'm using it as «people» rather than «male specimen»", or so 17:23
quite interesting
17:23 xtreak joined
psch m: sub f { [1,2,3],[4,5,6],[7,8,9] }; my ($a, $b, $c) = f; say $a 17:23
camelia rakudo-moar 2f9199: OUTPUT«[1 2 3]␤»
marchelzo Is there any way to get more detailed error messages? I get an error and it doesn't even tell me the line number. 17:24
lucasb I think just changing "men" to "dudes" would sound more inclusive :D
marchelzo "Directives specify 2 arguments, but 1 argument was supplied" is all it says.
[Coke] marchelzo: with printf*? known issue.
_nadim AlexDaniel: doin ($a $b, $c) = xxxx() works but that is not what I want to do 17:25
AlexDaniel [Coke]: is it rakudobugged?
_nadim: see psch answer ↑
marchelzo [Coke]: why does it only affect printf*?
AlexDaniel _nadim: aaah
[Coke] AlexDaniel: rt.perl.org/Ticket/Display.html?id=126063
AlexDaniel _nadim: that is not what you want… OK I have to learn to read :)
17:26 domidumont joined
[Coke] marchelzo: implementation detail 17:26
17:26 koo8 left
psch _nadim: what do you want to do? 17:26
17:26 dakkar left
marchelzo Is printf implemented in perl6 or in C? 17:26
psch marchelzo: in NQP, almost completely
[Coke] marchelzo: NQP
psch marchelzo: so, of those two options, Perl 6 is the correcter one
17:27 molaf_ joined
[Coke] jnthn: we need to find one of your old slide decks that shows the structure of rakudo (inc. moar, nqp, etc.) and throw that on perl6.org somewhere. or rakudo.org 17:27
mainly for the ongoing Endless Christmas. 17:28
AlexDaniel m: my ([@a]) = 1; say @a.perl; # here is the “null object” thing
camelia rakudo-moar 2f9199: OUTPUT«Cannot call method 'perl' on a null object␤ in block <unit> at /tmp/4f49kOW4_t:1␤␤»
marchelzo does all perl6 get compiled to NQP as an IR before becoming moar bytecode? Is it meant to be written by humans?
17:28 molaf left
[Coke] doesn't get compiled to nqp, no. 17:29
AlexDaniel HOORAY
SEGMENTATION FAULT!
m: my ([@a]) = 1; say @a.WHAT;
camelia rakudo-moar 2f9199: OUTPUT«(signal SEGV)»
TimToady jnthn: I do think we should consider renaming 'experimental' to 'MONKEY-AROUND', or maybe just 'MONKEY', on the theory that a quick audit of a project for dangerous code can just grep for MONKEY
marchelzo [Coke]: So where does nqp come from?
psch [Coke]: rakudo/docs/ has architecture.html and *.svg, maybe those are a start 17:30
AlexDaniel _nadim: so your null object finding is a bit more serious that it seemed :)
_nadim: I don't think that it is critical, but it should not segfault anyway 17:31
TimToady if we then declare EVAL to be dangerous, we can also have MONKEY-SEE-NO-EVAL :)
geekosaur rakudo perl6 is written in nqp
zengargoyle lol
[Coke] marchelzo: read www.jnthn.net/papers/2013-apw-lessons.pdf
AlexDaniel geekosaur: again… not true!
geekosaur: it is NQP and Perl 6
[Coke] marchelzo: crap, that's not right. one sec. 17:32
lucasb you mean these? edumentab.github.io/rakudo-and-nqp-...ls-course/ 17:33
[Coke] jnthn++ has so many slide decks, I'm trying to scan for the one that has the diagram that i think will help explain.
zengargoyle nqp == perl 6 for suitably augmented versions of nqp. :P
psch thinks github.com/rakudo/rakudo/blob/nom/...ecture.svg is pretty good
AlexDaniel
.oO( can we declare “shell” to be dangerous and have MONKEY-SHELL ? )
17:34 znpy left
[Coke] gives up trying to find the slides in particular he was looking for, ah well. 17:37
marchelzo [Coke]: heh I went through that entire set of slides trying to find NQP
TimToady we should gather up all the known monkey idioms in a list somewhere... 17:38
marchelzo psch: does (NQP) mean the component is implemented in NQP?
17:38 domidumont left
marchelzo or does it mean the code is represented as NQP during that phase 17:39
TimToady monkey-shine, monkey-business...
17:39 domidumont joined
TimToady
.oO(use FLYING-BUTT-MONKEY)
 17:40
psch marchelzo: no to the latter, yes to the former. the language in parens is what code that handles that stage is written in 17:41
[Coke] jnthn: is there a version of www.jnthn.net/papers/2013-yapceu-jvm.pdf but for moarvm?
psch marchelzo: as in, the parser and actions are written in NQP, the CORE.setting is written in Per l6
Hotkeys when should one use parens for function calls? "foo($x)" vs "foo $x"
TimToady when one wishes to guard against a future keyword "foo" 17:42
17:42 leont joined
marchelzo psch: I see 17:42
TimToady when one wishes to extend the expression with another .bar
AlexDaniel
.oO( oh! MONKEY-HELL for shell! )
 17:43
TimToady though of course there's also the option (foo $x).bar in that case
shades of Lisp
AlexDaniel
.oO( might also make sense to rename shell to hell )
Hotkeys mmkay 17:44
17:45 domidumont left, lucasb left
AlexDaniel
.oO( hell-out-of(“ls”) )
 17:45
pdcawley_ afk calling it a day, but I'll be in tomorrow morning to make up some time after my cold.
TimToady o/ 17:46
[Coke] TimToady: S03-metaops/reverse.t has a passing test that might be a result of your work.
RabidGravy oh deary, deary me
perl6 -Ilib t/030-cookies.t
===SORRY!===
Cannot find method 'run_alt'
17:47 cdg joined, cdg left
TimToady hmm, I have a bunch of new tests there that I forgot to check in 17:47
17:47 cdg joined
RabidGravy I think that rakudo is completely messed up right now 17:48
AlexDaniel RabidGravy: you mean, after curli?
RabidGravy nope in the last hour or so 17:49
17:49 xtreak left
RabidGravy This is rakudo version 2015.11-443-gc769890 built on MoarVM version 2015.11-34-gc3eea17 implementing Perl v6.b. 17:50
[jonathan@cannibal http-useragent]$ perl6 -Ilib t/030-cookies.t
1..29
ok 1 - new 1/3
This is rakudo version 2015.11-444-g2f91998 built on MoarVM version 2015.11-34-gc3eea17 implementing Perl v6.b. 17:51
moritz m: say <a b c>.roll(*).elems
camelia rakudo-moar 2f9199: OUTPUT«Cannot .elems a lazy list␤ in block <unit> at /tmp/uQ_R9jGw_Q:1␤␤Actually thrown at:␤ in block <unit> at /tmp/uQ_R9jGw_Q:1␤␤»
RabidGravy # Block Code Any Mu 17:52
Default constructor for 'Block' only takes named arguments
in block <unit> at t/030-cookies.t:172
17:52 ZoffixW joined
moritz m: say <a b c>.roll(*).rotor(5)[^10] 17:52
camelia rakudo-moar 2f9199: OUTPUT«((b a c a a) (b b b b b) (c c a c b) (c a a c a) (a c c a a) (c a b b a) (b c c b c) (c b b a a) (c a c c b) (b a b c a))␤»
ZoffixW RabidGravy, if that run_alt resurfaces, you may wanna comment on this ticket: rt.perl.org/Ticket/Display.html?id=126832 I had the exact same issue when installing a module with panda, though cloning a repo and installing from current dir worked
AlexDaniel _nadim: rt.perl.org/Public/Bug/Display.html?id=126857
RabidGravy okay it's something to do with the presense of blib 17:53
marchelzo how do you append two lists? 17:54
ZoffixW marchelzo, .append ?
marchelzo there's no infix operator?
AlexDaniel m: say append([1, 2, 3], [10, 11, 12]) 17:55
camelia rakudo-moar 2f9199: OUTPUT«[1 2 3 10 11 12]␤»
ZoffixW m: say [1, 2, 3] .append: [10, 11, 12] 17:56
camelia rakudo-moar 2f9199: OUTPUT«[1 2 3 10 11 12]␤»
ZoffixW look! infix! :D
AlexDaniel ZoffixW: hahahaha
marchelzo lol
ZoffixW m: sub infix:<»«> ($l1, $l2) { [|$l1, |$l2] }; say [1, 2, 3] »« [10, 11, 12] 17:58
camelia rakudo-moar 2f9199: OUTPUT«[1 2 3 10 11 12]␤»
AlexDaniel ZoffixW: why not call it A :) 17:59
ZoffixW m: sub infix:<¯\_(ツ)_/¯> ($l1, $l2) { [|$l1, |$l2] }; say [1, 2, 3] ¯\_(ツ)_/¯ [10, 11, 12]
camelia rakudo-moar 2f9199: OUTPUT«[1 2 3 10 11 12]␤»
moritz autarch: is your advent post for day 11 ready? because if yes, I'd like to post it tomorrow already (swap days 10 and 11) 18:00
autarch moritz: one person reviewed it, I wouldn't mind another reviewed, but otherwise yes
gfldex m: my @a = 1,2,3; @a[*-1] = |(11,22,33); dd @a; 18:01
camelia rakudo-moar 2f9199: OUTPUT«Array $var = $[1, 2, (11, 22, 33)]␤»
gfldex why does that refuse to slip?
m: my @a = 1,2,3; @a.push(|(11,22,33)); dd @a;
camelia rakudo-moar 2f9199: OUTPUT«Array $var = $[1, 2, 3, 11, 22, 33]␤»
marchelzo How can I do something like this? for lines() { /(<digit>+) (<alpha>+)/; say $1; }
TimToady note that .append is mutating the left arg
geekosaur I would suspect that you are telling it to assign to an element
marchelzo I keep getting Nil for $1.
AlexDaniel m: sub infix:<¯\_(ツ)_/¯> ($a, $b) { [$a, $b].pick }; say 42 ¯\_(ツ)_/¯ 69
camelia rakudo-moar 2f9199: OUTPUT«69␤» 18:02
AlexDaniel m: sub infix:<¯\_(ツ)_/¯> ($a, $b) { [$a, $b].pick }; say 42 ¯\_(ツ)_/¯ 69
camelia rakudo-moar 2f9199: OUTPUT«42␤»
ZoffixW marchelzo, first match is $0
marchelzo ZoffixW: I know, but $1 should still be defined shouldn't it?
ZoffixW m: for 'nope' { /(<digit>+) (<alpha>+)/; say $1; } 18:03
camelia rakudo-moar 2f9199: OUTPUT«Nil␤»
ZoffixW m: for '42nope' { /(<digit>+) (<alpha>+)/; say $1; }
camelia rakudo-moar 2f9199: OUTPUT«｢nope｣␤ alpha => ｢n｣␤ alpha => ｢o｣␤ alpha => ｢p｣␤ alpha => ｢e｣␤»
marchelzo oh, I know the problem
ZoffixW \o/
marchelzo my input didn't actually match the regex
ZoffixW :)
moritz autarch: ok, I'll review and swap dates 18:04
autarch moritz: thanks
18:04 ZoffixW left
RabidGravy ah, the random "Use of uninitialized value %ENV of type Any in string context" is coming out of qqx/ .../ here 18:07
dalek : 81d2e76 | moritz++ | misc/perl6advent-2015/schedule:
Perl 6 advent: Swap days 10 and 11
 18:08
: 92dc957 | moritz++ | misc/perl6advent-2015/schedule:
Perl 6 advent: Update post title for day 10
TimToady m: say WHAT ((*.sort)([1,3,1])) 18:09
camelia rakudo-moar 2f9199: OUTPUT«(List)␤»
gfldex m: my @a = |(1,2,3); my @b; @b[0] = |(1,2,3); dd @a, @b;
camelia rakudo-moar 2f9199: OUTPUT«Array $var = $[1, 2, 3]␤Array $var = $[(1, 2, 3),]␤»
masak TimToady: ok. I have some time tonight; I'll get on it. 18:10
yoleaux 17:18Z <TimToady> masak: I wouldn't mind a golf of the failure, or better yet, a spectest
TimToady thanks 18:11
AlexDaniel j: my ([$a]); $a.WHAT
camelia rakudo-jvm 6c0f93: OUTPUT«java.lang.NullPointerException␤ in block <unit> at /tmp/y_AIDyesNw:1␤␤»
RabidGravy it's possible that "use NativeCall; say CArray[uint8].new ~~ CArray[uint8]" may not work when precompiled
moritz autarch: I've read your advent post, and liked it. I changed one link from http to https, that's all :-) 18:12
marchelzo anyone know a nice way to turn (a, b, c, d) into ((a, b), (b, c) (c, d))?
TimToady m: my $f = (*.sort)([1,3,1]); say $f.WHAT 18:13
camelia rakudo-moar 2f9199: OUTPUT«()␤»
TimToady m: my $f = (*.sort)([1,3,1]); say $f
moritz m: say <a b c d>.rotor(2, -1).perl
camelia rakudo-moar 2f9199: OUTPUT«Cannot find method 'gist': no method cache and no .^find_method␤ in block <unit> at /tmp/wfsqiIAVuD:1␤␤»
rakudo-moar 2f9199: OUTPUT«Cannot have elems < 1, did you mean to specify a Pair with => -1?␤ in block <unit> at /tmp/00fEQdi9bR:1␤␤»
moritz m: say <a b c d>.rotor(2 => -1).perl
camelia rakudo-moar 2f9199: OUTPUT«(("a", "b"), ("b", "c"), ("c", "d")).Seq␤»
moritz marchelzo: ^^
marchelzo amazing 18:14
gfldex m: my @b; @b = @b.list, |(1,2,3); dd @b; say @b;
camelia rakudo-moar 2f9199: OUTPUT«Array $var = (my \Array_74891312 = $[Array_74891312, 1, 2, 3])␤(\Array_74891312 = [Array_74891312 1 2 3])␤»
gfldex is that intentional?
m: my @b; @b = @b, |(1,2,3); dd @b; say @b;
camelia rakudo-moar 2f9199: OUTPUT«Array $var = (my \Array_80543712 = $[Array_80543712, 1, 2, 3])␤(\Array_80543712 = [Array_80543712 1 2 3])␤»
gfldex m: my @b; @b = @b, |(1,2,3); dd @b; put @b; 18:15
TimToady m: my $f = (*.sort)([1,3,1]); say $f.^name
camelia rakudo-moar 2f9199: OUTPUT«Array $var = (my \Array_57139168 = $[Array_57139168, 1, 2, 3])␤Memory allocation failed; could not allocate 29248 bytes␤»
rakudo-moar 2f9199: OUTPUT«BOOTArray␤»
gfldex m: my @b; @b = @b, 1; put @b; 18:16
camelia rakudo-moar 2f9199: OUTPUT«Memory allocation failed; could not allocate 8192 bytes␤»
gfldex is that rakudos way of saying: "infinite recursion" ? 18:17
TimToady likely
gfldex i shall rakudobug
TimToady m: my $f = (*.sort)([1,3,1]); say $f.^name 18:18
camelia rakudo-moar 2f9199: OUTPUT«BOOTArray␤»
TimToady so shall I
ugexe are there any changes planed for Version, like restrictions/transformations on what is passed to .new? (i.e. should version comparison try to be smart such that ?(Version.new("v1.0") ~~ Version.new("1.0")))
i remember there was a discussion on it but i dont remember if anything came of it :( 18:19
autarch moritz: cool, thanks for reviewing it
I note that the script I refer to at the end in my repo still fails with gist.github.com/autarch/1405c19f1ee561cb6fbd 18:20
18:20 ab6tract left
gfldex m: my @b; @b[0] = |(1,2,3); dd @b; # my question if that's a bug still stands because i think it is 18:20
camelia rakudo-moar 2f9199: OUTPUT«Array $var = $[(1, 2, 3),]␤»
autarch would be nice to fix that
AlexDaniel m: my @b; @b = @b, 1; put @b;
autarch the script in question is at github.com/autarch/perl6-advent-20...ewalker.p6
camelia rakudo-moar 2f9199: OUTPUT«Memory allocation failed; could not allocate 8192 bytes␤» 18:21
AlexDaniel oh come on! 8192 bytes!
b2gills In Perl 5 a list in scalar context will give the either the last element, or the number of elements; depending on if it is an immediate list or the result of a function call. Perl 6 returns the list itself, which in coercing numeric context always returns the number of elements.
autarch I am of course going afk - if someone figures out my issue please have a bot tell me ;)
b2gills That means it could not allocate an additional 8192 bytes 18:22
18:22 loren left
dalek kudo/repository_registry: f729c17 | (Stefan Seifert)++ | src/core/CompUnit/RepositoryRegistry.pm:
Remove obsolete method ctxsave from CompUnit::RepositoryRegistry
 18:23
kudo/repository_registry: 90da3cf | (Stefan Seifert)++ | / (3 files):
Move repository setup into CompUnit::RepositoryRegistry

All code dealing with include specs is now in CompUnit::RepositoryRegistry. This way we no longer need to make SHORT-ID2CLASS, INCLUDE-SPEC2CUR and PARSE-INCLUDE-SPEC publicly visible.
  %*CUSTOM_LIB<site> is now replaced by
  %CompUnit::RepositoryRegistry.repository-for-name('site')
zengargoyle hrm, can you use self in a has? 18:24
nine TimToady: your commit d3c278d34e6da5b8b39b26bab32d6fcdf3e9a969 causes a regression in Inline::Perl5 where in a method self is a completely different object (higher up in the call stack) than the one the method was called on.
Skarsnik zengargoyle, probably
zengargoyle m: my class x { has $foo = self; }; 18:25
camelia ( no output )
zengargoyle yeah...
marchelzo Can someone help me understand why this isn't working? sprunge.us/jQPJ
AlexDaniel m: my @b; @b = @b, 1; say @b; 18:26
camelia rakudo-moar 2f9199: OUTPUT«(\Array_75579232 = [Array_75579232 1])␤»
marchelzo It says "WhateverCode object coerced to string".
AlexDaniel that's pretty cool though
Skarsnik m: class A { has $.self; method foo() {say self.self = "foo"}; my A $a = A.new; $a.foo();
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/cqxYXHzwMZ␤Missing block␤at /tmp/cqxYXHzwMZ:1␤------> 3elf = "foo"}; my A $a = A.new; $a.foo();7⏏5<EOL>␤»
Skarsnik m: class A { has $.self; method foo() {say self.self = "foo"}}; my A $a = A.new; $a.foo();
camelia rakudo-moar 2f9199: OUTPUT«Cannot modify an immutable Any␤ in method foo at /tmp/Ru2txYp6tp:1␤ in block <unit> at /tmp/Ru2txYp6tp:1␤␤»
ugexe *.say
psch m: *.say
camelia ( no output )
psch m: { *.say }()
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/hg0Ckumkjy␤Malformed double closure; WhateverCode is already a closure without curlies, so either remove the curlies or use valid parameter syntax instead of *␤at /tmp/hg0Ckumkjy:1␤------> 3{ *.say }7…»
ugexe m: *.join('-')
camelia ( no output )
ilmari m: class :: { method bar { 42 } has baz = self.bar + 3 }.new.baz
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/vqnYBhNO6l␤Strange text after block (missing semicolon or comma?)␤at /tmp/vqnYBhNO6l:1␤------> 3class :: { method bar { 42 }7⏏5 has baz = self.bar + 3 }.new.baz␤ expecting any of:␤ inf…»
Skarsnik m: class A { has $.self is rw; method foo() {say self.self = "foo"}}; my A $a = A.new; $a.foo();
camelia rakudo-moar 2f9199: OUTPUT«foo␤»
nine TimToady: With method call_context() { note self.^name }; I get the wrong self, while with method call_context(\SELF:) { note SELF.^name } it's correct! 18:27
ilmari m: class :: { method bar { 42 }; has $.baz = self.bar + 3 }.new.baz
camelia ( no output )
ilmari m: class :: { method bar { 42 }; has $.baz = self.bar + 3 }.new.baz.say
camelia rakudo-moar 2f9199: OUTPUT«45␤»
ugexe m: my %x; %x{*.join('-')}
camelia rakudo-moar 2f9199: OUTPUT«WhateverCode object coerced to string (please use .gist or .perl to do that) in block <unit> at /tmp/MUnHeYtLjD:1␤»
TimToady nine: is it in the lexical scope of any \self declaration?
marchelzo ugexe: ok, I see what the problem is 18:28
ugexe marchelzo: use a map block and $_ instead of *
marchelzo ugexe: yea
nine TimToady: only \self is in github.com/niner/Inline-Perl5/blob...5.pm6#L969 and line 984 while the method in question is created by an EVAL in github.com/niner/Inline-Perl5/blob...5.pm6#L520 18:30
marchelzo ugexe: hmm. now it says: "Use of uninitialized value of type Any in numeric context in block" (line 12)
nine TimToady: note that the wrong self is the Perl6Callbacks object...the one where the EVAL that creates the method is run in 18:31
18:31 prammer joined
nine TimToady: Perl6Callbacks.create runs an EVAL that creates a class and the method of this class gets the Perl6Callbacks object as self instead of its own 18:31
TimToady ah, so there is already a self defined when it sees the EVAL's self, hmm, so it thinks someone said my \self = ... 18:32
nine yes, that sounds like it
TimToady maybe we'll need to back out the self part of that patch, and make self 'more reserved'
marchelzo Is there something like .sort for strings?
nine Better safe than sorry... 18:33
That one was quite a challenge to nail down :)
ugexe marchelzo: might try to name your outer map parameters in case you are stepping on $_ `.map(-> $arg { $arg ... })`
psch m: say <a c e b d>.sort
camelia rakudo-moar 2f9199: OUTPUT«(a b c d e)␤»
Skarsnik m: say "acebd".split('').sort;
camelia rakudo-moar 2f9199: OUTPUT«( a b c d e)␤»
gfldex m: 'bca'.comb.sort.say;
camelia rakudo-moar 2f9199: OUTPUT«(a b c)␤»
gfldex m: 'bca'.comb.sort.join.say; 18:34
camelia rakudo-moar 2f9199: OUTPUT«abc␤»
AlexDaniel m: my @b; my @b[0] = @b
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5===␤Cannot use variable @b in declaration to initialize itself␤at /tmp/aNwB2NF9OU:1␤------> 3my @b; my @b[0] = @7⏏5b␤ expecting any of:␤ term␤Other potential difficulties:␤ Redeclaration of symbol @b␤ at…»
gfldex i do agree that there should be Str::sort tho
psch m: say Str.^can('sort') 18:35
camelia rakudo-moar 2f9199: OUTPUT«(Method+{<anon|77418096>}.new)␤»
AlexDaniel m: my @b; my @a; my @a[0] = @b; my @b[0] = @a; say @b
camelia rakudo-moar 2f9199: OUTPUT«Potential difficulties:␤ Redeclaration of symbol @a␤ at /tmp/DVpk9qT3U5:1␤ ------> 3my @b; my @a; my @a[0]7⏏5 = @b; my @b[0] = @a; say @b␤ Redeclaration of symbol @b␤ at /tmp/DVpk9qT3U5:1␤ ------> 3my @b; my @a; my @a[0…»
AlexDaniel my favorite “Index 0 for dimension 1 out of range (must be 0..-1)” :)
18:35 zakharyas left 18:36 CIAvash left
ugexe m: my @b; my @a; @a[0] = @b; @b[0] = @a; say @b 18:36
camelia rakudo-moar 2f9199: OUTPUT«(\Array_74018864 = [[Array_74018864]])␤»
dalek kudo/nom: 67749a7 | TimToady++ | src/Perl6/Grammar.nqp:
make self non-overridable

otherwise an EVAL creating a method within a method gets the wrong self
TimToady nine: ^^ should fix it
gfldex m: "bca".sort.say 18:37
camelia rakudo-moar 2f9199: OUTPUT«(bca)␤»
18:37 Ven joined
nine checking... 18:38
Hotkeys m: "bca".split.sort.join.say 18:39
camelia rakudo-moar 2f9199: OUTPUT«Cannot call split(Str: ); none of these signatures match:␤ (Cool $: Regex:D $pat, $limit = { ... };; :$all, *%_)␤ (Cool $: Cool:D $pat, $limit = { ... };; :$all, *%_)␤ (Str:D $: Regex:D $pat, $parts = { ... };; :$v is copy, :$k, :$kv, :$p, :…»
Hotkeys m: "bca".split('').sort.join.say 18:40
camelia rakudo-moar 2f9199: OUTPUT«abc␤»
dalek ast: 200ec05 | TimToady++ | S03-metaops/reverse.t:
test thunking of reversed logical ops
ast: 6c2c99c | TimToady++ | S03-metaops/reverse.t:
fudge temporarily failing test
18:41 prammer left, leont left
AlexDaniel m: my @a; @a[0] = @b; @b[0] = @a; put @b 18:41
camelia rakudo-moar 2f9199: OUTPUT«5===SORRY!5=== Error while compiling /tmp/aT5vfdqkeN␤Variable '@b' is not declared␤at /tmp/aT5vfdqkeN:1␤------> 3my @a; @a[0] = 7⏏5@b; @b[0] = @a; put @b␤»
AlexDaniel m: my @b; my @a; @a[0] = @b; @b[0] = @a; put @b 18:42
camelia rakudo-moar 2f9199: OUTPUT«Memory allocation failed; could not allocate 49120 bytes␤»
nine \o/
TimToady++ # Inline::Perl5 works again :) 18:43
.tell [Tux] TimToady++ just fixed the rakudo regression that broke Inline::Perl5's t/v6.t
yoleaux nine: I'll pass your message to [Tux].
marchelzo Is anybody here doing the adventofcode.com challenges in perl6? 18:44
grondilu marchelzo: I do 18:46
TimToady is there a way to add [BUG] to an RT ticket I forgot to mark?
grondilu well, not everything works in Perl 6 though
18:47 Ven left
grondilu I had to do day 4 in Perl 5, for instance. 18:48
and day 6
zengargoyle grondilu: now you can go back and use Inline::Perl5 :)
grondilu I've never used this. How different is it from v5? 18:49
m: use v5;
camelia rakudo-moar 2f9199: OUTPUT«===SORRY!===␤Could not find Perl5:ver<True>:auth<True>:api<True> in:␤ /home/camelia/.perl6/2015.11-444-g2f91998␤ /home/camelia/rakudo-m-inst-2/share/perl6/site␤ /home/camelia/rakudo-m-inst-2/share/perl6/vendor␤ /home/camelia/rakudo-m-…»
zengargoyle no clue, i was just being snarky.
18:49 Ven joined
marchelzo grondilu: why? 18:49
grondilu Perl 6 was too slow
marchelzo yeah that's my problem right now 18:50
brute force travelling salesman problem in perl6 isn't working out so well
nine grondilu: largest difference to v5 is that you can use all of CPAN with Inline::Perl5 18:51
TimToady: "Basics" in the menu, there you can edit the subject 18:52
grondilu marchelzo: that's day9 right? I brute-forced it with in Perl 6. Not very fast but reasonable.
18:52 xfix joined
grondilu on a Raspberry Pi even 18:52
18:53 prammer joined
grondilu marchelzo: make sure you have a recent rakudo, though. 18:53
18:54 Ven left
marchelzo grondilu: how long did it take to execute for you? 18:54
mine eventually finished but took several minutes 18:55
18:55 Ven joined
grondilu it did take a few minutes. Just consider displaying progress. When you see it not making any you can try what the program currently has as a minimum. Worked for me. 18:55
marchelzo hmm 18:56
I don't think my code is very good, but I'm tempted to port it to pypy3 just to see the speed difference.
18:56 koo8 joined
grondilu port it to Perl 5, you'll see enough of a difference. 18:57
RabidGravy jnthn, just added "use experimental to OO::Monitors" ;-)
marchelzo grondilu: I don't know Perl 5
masak m: say [&&] |0 18:58
camelia rakudo-moar 67749a: OUTPUT«0␤»
masak TimToady: ^
TimToady: that's the golf.
marchelzo at least not very well
masak TimToady: not sure where to add a spectest for that -- but I can, no problem :)
let's see if I can make it fail on the commit in question, but not elsewhere.
TimToady that's okay, I can put in a test as I fix it 19:00
masak TimToady: stacktrace: gist.github.com/masak/8fcd83e1054fc1e74e8c
TimToady: ok, then I leave it to you from this point on.
TimToady: I already found that at least `try EVAL` catches the bad mood and is testable.
TimToady I figured it probably had somethign to do with a slip
masak (see gist)
TimToady since the patch wants to take comma-args literally
at minimum I probably need to exempt slip arguments from thunking 19:01
masak TimToady: if you want to re-do your patch with the bug fixed, if you push it to a branch, I can test 007 on it before we merge to nom. 19:02
TimToady or just give up entirely on thunking lists if there's a slip anywhere
okay
masak I can show you the original un-golfed line of code 19:03
github.com/masak/007/blob/master/l...#L121-L122
so, essentially an "all of these values have to be truthy" construct 19:04
ISTR I added the slip after GLR
[Tux] runs «make again»
yoleaux 18:43Z <nine> [Tux]: TimToady++ just fixed the rakudo regression that broke Inline::Perl5's t/v6.t
19:05 Ven left
marchelzo grondilu: oh wow 19:07
it's almost instant in python3
that's sort of disappointing
ChristopherBotto marchelzo: Rakudo Perl 6 hasn't been optimized yet. 19:09
marchelzo ChristopherBotto: when will it be optimized?
ChristopherBotto That will be a large part of the work in 2016. 19:10
marchelzo: That will be a large part of the work in 2016.
[Tux] nope, still fail 19:11
must run. feedback later
marchelzo I really hope it ends up being fast
masak I think it's a little bit unfair to say that it "hadn't been optimized" 19:13
but surely we are not at the limit point yet. there is more optimization to be had
ChristopherBotto masak: Sorry, it hasn't been "fully" optimized.
masak it's perhaps comparable to early Ruby or something like that 19:14
psch .oO( s/fully/seriously/ )
RabidGravy certain things have become faster
ChristopherBotto But its design will make it more optimizable than Perl 5 ever could be, right? 19:15
TimToady that's what we think, but we've thunk wrong before :) 19:16
19:16 leont joined 19:17 Ven joined 19:19 Ven left
masak :P 19:20
ChristopherBotto: there are many opportunities for optimization in Perl 6 that couldn't be optimized in Perl 5, even in theory. 19:21
ChristopherBotto: of course *saying* that, and *knowing* that, is different from being as fast or faster than Perl 5 ;)
MadcapJake marchelzo: I've done AdventOfCode days 3, 5, and 6. I'm in the midst of working on day 7 19:23
_nadim AlexDaniel: Cool, in a way, first P6 bug I find 19:24
marchelzo MadcapJake: nice
ChristopherBotto masak: It's already fast enough for many things. I use Perl 6 as my language of choice when starting a new project. On the rare occasion that I need raw speed, I'll prototype in Perl 6 before coding it in Perl 5.
_nadim AlexDaniel: actually the second one, I segaulted with something else yesterday, I kept the code for later ;) 19:25
MadcapJake day 6, I only did part 1 because compilation/execution time was over 2 hours and I would much rather solve other problems than wait that long xD
[Coke] TimToady: note that adding [BUG] to the title of a bug doesn't magically set the BUG flag. 19:26
MadcapJake originally I tried doing day 6 with shaped arrays but that was even slower (current solution is nested arrays, a double for loop, and dispatch function of sorts)
[Coke] ... note that no one actually seems to care about the bug flag.
19:27 ilbelkyr joined
MadcapJake in Day 7 I am doing another grammar and actions object! Love these! Debugging them is a bit hard though because action method sare called in post-order. 19:29
_nadim I'll ask again, sub X { return @a, @b) ... I'd like to do my(@x, @y) = X() or something like that. I know my ($x, $y) = X() work fine. but I'd like to know how to do it with arrays and not have the first one gobble everything. 19:30
Skarsnik I am not what you can return to make that work
return [@a], [@b]; maybe? 19:31
moritz m: sub f { my @x = 1, 2; return @x, @x }; my (@a, @b) := f(); say @b.perl 19:32
camelia rakudo-moar 67749a: OUTPUT«[1, 2]␤»
moritz just use binding :-)
19:32 ilbelkyr left
_nadim ah! thanks moritz 19:33
masak ChristopherBotto: sounds nice. I think there's a large potential user base who could strike the same deal with Perl 6 speed/whipuptitude. 19:35
Skarsnik *love mi6, added a build-readme cmd to mi6* 19:36
My main complain about speed for now is starting time (october rakudo) x)
19:37 khw left 19:40 ilbelkyr joined
skids nine: Re: "That one was quite a challenge to nail down" unfortunately duplicated effort, yes that was a hard one to find RT#126754 19:43
timotimo too much backlog :( 19:45
yoleaux 00:30Z <Zoffix> timotimo: Sent you a PR to fix the freezing issue. lucasb++ first found the problematic line. github.com/timo/json_fast/pull/8
grondilu m: sub f { return "first" unless $++; return rand < $ = rand; }; say f; say f 19:46
camelia rakudo-moar 67749a: OUTPUT«first␤Use of uninitialized value of type Any in numeric context in sub f at /tmp/Z5yFjg80Hh:1␤Cannot modify an immutable Bool␤ in sub f at /tmp/Z5yFjg80Hh:1␤ in block <unit> at /tmp/Z5yFjg80Hh:1␤␤»
psch hrm, this ContextRef thing is nigh-impenetrable from where i'm standing :/ 19:47
grondilu not sure but I think it's a bug
psch and apparently something is mismatched with Makefile generation between moar and jvm, which 0ec5ad56 fixed but went broken again somewhen later... 19:48
maybe i have to actually read and understand all this CURLI stuff /o\
timotimo ShimmerFairy: your post has dominated all other posts since the day it came out :) 19:49
moritz wonders if this was the Perl Weekly newsletter hitting the blog 19:51
19:51 joeshcmoe joined 19:52 demayl joined
grondilu still trying to improve permutations: gist.github.com/grondilu/61a28aca4bcb7ea423e4 This one does not seem faster though. 19:55
ah nevermind, does not even quite work 20:01
20:03 raoulvdberge joined 20:04 prammer left 20:05 espadrine left 20:06 Ven joined, adhoc joined 20:11 spider-mario joined, Ven left 20:12 joeshcmoe left
[Coke] how can I tell which parent class or role is providing a particular method? 20:12
20:12 yqt joined
Skarsnik hm sounds tricky 20:13
A.^roles.grep: $_.^can("method") ?
moritz m: say Int.^can('sqrt')[0].package 20:15
camelia rakudo-moar 67749a: OUTPUT«(Int)␤»
20:15 Ven joined
moritz m: say Int.^can('list')[0].package 20:15
camelia rakudo-moar 67749a: OUTPUT«(Any)␤»
moritz m: say Int.^can('roots')[0].package 20:16
camelia rakudo-moar 67749a: OUTPUT«(Numeric)␤»
moritz [Coke]: that seems to work
[Coke] TimToady: RT #126860 is an implicit request that iterating over a string give you the characters of a string. I'm assuming the answer is no and we should reject the ticket?
m: say Str.^can("sort")[0].package; # should be Any
camelia rakudo-moar 67749a: OUTPUT«(Any)␤»
[Coke] moritz++ 20:17
20:18 darutoko left 20:19 ChristopherBotto left
timotimo hey folks, guess what 20:19
grondilu what?
timotimo grondilu: 15.5% of time spent in a simple permutations benchmark is spent in "find_best_dispatchee"
which means somewhere we're failing to go through the dispatch cache 20:20
MadcapJake If anyone has a moment to look at my adventofcode day 7 parser, I am stuck! It solves the example given, I've fixed the case where a 1 is used instead of a wire, and I've sorted assignments at the beginning. Still I am getting zero for wire A! gist.github.com/MadcapJake/558eff70f13d2ff2c223
timotimo ah
it might be the .= for .=reverse
20:21 Ven left
grondilu let me try without 20:21
timotimo please do
using .=reverse for two elements in the array can very well be a SUPER CRAZY overhead
dj_goku m: say $*CWD 20:22
camelia rakudo-moar 67749a: OUTPUT«"/home/camelia".IO␤»
jdv79 15% is SUPER CRAZY to you?
150% may qualify to me 20:23
timotimo grondilu: i'll microbenchmark a bit of stuff for you
jdv79: if the script spends more than 100% of its time in just one function, that'd be crazy to me :)
[Coke] jdv79: ... we can't spend 150% of the time in a given function.
RabidGravy I just discovered that the use Foo:ver<v0.0.1> works now 20:24
El_Che [Coke]: MAIN?
:)
grondilu nope , doesn't seem faster. Actually it looks slower without the .reverse.
s/.reverse/.=reverse/ 20:25
jdv79 right. i meant 1.5x slower. still 15% of runtime doesn't seem SUPER CRAZY.
RabidGravy: in what way?
20:25 FROGGS left
jdv79 can one actually install multiple version of the same package and use them solely or concurrently? 20:25
RabidGravy i.e. if you have "use Foo:ver<v0.0.2>" and you only have v0.0.3 installed then it craps out 20:26
dj_goku thanks for whoever wrote camelia and thought of this: Proc::Async is disallowed in restricted setting
RabidGravy but yes you can have multiple versions installed 20:27
timotimo grondilu: that's because you're now slicing twice
grondilu yes, would make sense.
but then what is the dispatch thing causing slowliness? 20:28
timotimo the difference between @foo[4,5] = @foo[4,5].reverse and @foo[4,5] = @foo[5,4] is very small
grondilu I mean there are very few function calls in this if any.
jdv79 ah
timotimo oh, damn 20:29
it wasn't .=
.= is at 1.3% exclusive time and 9.2% inclusive time
grondilu notice that we could use a faster algorithm but it would require a rewrite of the tests. See github.com/perl6/roast/issues/79 20:31
timotimo there's some range objects that are getting created that are surprisingly expensive to be created
that also goes through the slow path binder, but not the "find best dispatchee" thing 20:32
grondilu which range object? There's only ^$!n
$k+1 ... * not being a range but a sequence. 20:33
timotimo it could be that POSITIONS is what spends all that damn time in the dispatch find thingie
POSITIONS is at 30% inclusive time 20:34
POSITIONS itself is cheap enough
grondilu oh it seems $k+1..@!a.end is quite faster than $k+1..*
timotimo aye, POSITIONS goes through the slow path binder 20:35
grondilu was wrong in thinking $k+1...* was used. It's $k+1..*
timotimo quite a bit, eh?
m: my @a = (^20).list; for ^1000 { @a[5..*] }; say now - INIT now 20:36
camelia rakudo-moar 67749a: OUTPUT«0.63605866␤»
timotimo m: my @a = (^20).list; for ^1000 { @a[5..@a.end] }; say now - INIT now
camelia rakudo-moar 67749a: OUTPUT«0.2863622␤»
timotimo m: my @a = (^20).list; for ^1000 { @a[5..^+@a] }; say now - INIT now
camelia rakudo-moar 67749a: OUTPUT«0.29446261␤»
timotimo it's noticable at least
grondilu it's a bit silly IMHO. Shouldn't the optimizer make it as fast in all cases? 20:37
Skarsnik yay another error in NC/Rakudo determining the size of a C structure
timotimo pfft!
the optimizer can't know that stuff
in the range, the * doesn't get whatever-curried, you know?
m: my @a = (^20).list; for ^1000 { @a[{ 5..$_ }] }; say now - INIT now
camelia rakudo-moar 67749a: OUTPUT«0.3378164␤»
timotimo ^- that's 2x faster than the range with whatever as an end, but slower than putting @a.end in directly
20:38 Ven joined
grondilu well, I'll submit a PR I guess 20:38
20:38 Ven left
Skarsnik damn I hate when my code is right: Your representation is smaller than the cstruct, but total size of fields match. Did you forget a field? 20:38
timotimo :)
your module seems kinda cool, Skarsnik
Skarsnik I am running it on DBIish 20:39
timotimo grondilu: did you run a little benchmark and have an overall speed difference?
20:40 Ven joined
grondilu yeah. 20:40
20:40 Ven left
timotimo does it make any difference? :< 20:40
grondilu yes.
I ran it with permutations(5). 3.9s instead of 4.5s or something. 20:41
quite significant anyway.
20:41 zacts left
grondilu it's not much of a benchmark though, just a manual test. 20:41
timotimo right, fair enough 20:42
thank you for looking into it :)
with the new code, i'll do some more benchmarking and profiving
profiling
Skarsnik Thx mysql doc to not show the last field on their documentation. I really wish I could compile clang-smokegen :( 20:43
moritz that you have to care about the fields of the mysql structs is already a WTF
sqlite3 and postgres manage just fine with opaque data structures that you just pass along 20:44
timotimo grondilu: exchanging the two values via a variable and only doing single-value slices is a TON faster than .=reverse
grondilu one thing that clearly takes time is .List. I wonder if there's a quicker way to do that.
really? 20:45
20:45 Ven joined
timotimo yes. 20:45
Skarsnik moritz, to get the type of a field
timotimo i can either do 10_000 .= reverse or i can do 450_000 via a variable
your choice!
grondilu let me try 20:46
20:46 pdcawley_ left
timotimo do you think you can come up with something faster for the [$k+1 .. *] .= reverse? 20:47
grondilu apart from using @!a.end you mean?
leont META.info doesn't work until installed? 20:48
El_Che ideone.com/ supports perl6
grondilu wasn't there a trick to substitute two variables without using a temp var?
timotimo grondilu: yes
leont is having multiple packages in one file, and it doesn't DWIM when loading them using "use"
timotimo with xor
Skarsnik leont, what do you mean?
grondilu timotimo: can you show me? my brain is tired :P 20:49
TimToady [Coke]: yes, we can reject #126860
grondilu timotimo: nevermind, I got it: en.wikipedia.org/wiki/XOR_swap_algorithm
leont has a file with at least two packages, the unit has the name of one of them, but I also want to export from the other one, I can't get this to work
20:50 demayl left, Ven left, zacts joined
moritz only works for integers, no? 20:51
timotimo grondilu: please note that you're currenly working with an Any array for the numbers
whether or not using a native int array is better right now, i can't tell.
if you're up for making it a tiny bit faster still, $i++ on a native is considerably more expensive than $i = $i + 1 is
in this case, it's not $i, it's $l and ök
$k 20:52
and it's not ++, it's --
but it still applies
20:53 marchelzo left
timotimo that won't be a noticable part of the whole. not yet anyway. 20:53
dalek kudo/nom: 5ed63cd | grondilu++ | src/core/native_array.pm:
small permutations improvement

use explicit .end upper slice limit instead of Whatever. As suggested by timotimo.		 20:54
kudo/nom: 1fe560d | timo++ | src/core/native_array.pm:
Merge pull request #627 from grondilu/patch-5

small permutations improvement
timotimo grondilu: i was expecting you'd also put in the exchange thing for the two slots. i guess i'll do it myself :)
Skarsnik they get freaking lazy in mysqlclient "Here you row, everything is char *" even if we can give you info on the field type
grondilu timotimo: no I'm doing it 20:55
timotimo OK :) 20:56
grondilu give me a sec
timotimo for me the timing of permutations is very noisy
dalek kudo/nom: 294fca4 | timotimo++ | src/core/Exception.pm:
undefined types are often mistaken for subs and gobble blocks
 20:57
timotimo had that patch locally and forgot to push it 20:58
timotimo builds rakudo 20:59
hoelzro .oO( are gobble blocks blocks that look like turkeys? )
timotimo i never understood why turkeys supposedly say "gobble gobble"
20:59 Ven joined 21:00 geraud joined
timotimo stage parse ... 95 seconds!? 21:00
must be overheating or something
grondilu me wonders if he can write $a = $b +^ $a as $a [R+^]= $b
hoelzro ¯_(ツ)_/¯ 21:02
moritz wonders if any clarity is gained from doing so
timotimo should be able. it could be very expensive, though
grondilu I thought it would be faster
but I guess not
[Coke] (gobbled) btw, "flatten" is a better description of what's going on with that return, I think. 21:03
21:03 ZoffixMobile joined
ZoffixMobile ehehe. have you guys seen this? TimToady on a shirt; 72 hours left: www.kickstarter.com/projects/14228...arry-wall/ 21:04
Skarsnik This look so weird 21:05
21:05 n0tjack joined
ZoffixMobile the guy on the shirt on the pic looks like Gerge Lucas.... 21:06
Skarsnik My only perl T-shirt is the one I get a YAPC in Paris x)
ZoffixMobile "in common sizes".... well, not me then X) 21:07
The only computer shirt I ever owned was the one with xkcd's sudo make me a sandwich
Skarsnik me a sandwich not a valid target for make. 21:11
?
awwaiid re-discovers the "note" sub when reading some old code
ZoffixMobile Skarsnik: xkcd.com/149/ 21:13
timotimo grondilu: whaaaat. you actually used the xor swap trick?! is it any faster at all? 21:14
RabidGravy timotimo, would that commit explain the "HTTP::Response.new()" in this test somehow thinking it's Block.new?
timotimo RabidGravy: what?
what commit, what what?
RabidGravy "undefined types are often mistaken for subs and gobble blocks"
Skarsnik ZoffixMobile, Oh yeah xD I was just wondering what kind of error make will say x)
RabidGravy ^
timotimo no, i don't think so 21:15
it's only about an exception message
m: given ThisIsActuallyAType { say "yay" }
camelia rakudo-moar 294fca: OUTPUT«5===SORRY!5===␤ (or perhaps it's a class that's not declared or available in this scope?)␤at /tmp/nsMvko2NCb:1␤------> 3given ThisIsActuallyAType { say "yay" }7⏏5<EOL>␤Missing block (apparently claimed by 'ThisIsActuallyAType')␤at /tmp/…»
timotimo crap!
the message got garbled!
timotimo has a fix 21:16
grondilu timotimo: it is faster. Also I can actually use it for the @!a[$k+1..@!a.end].=reverse
RabidGravy I'm seeing "Default constructor for 'Block' only takes named arguments in block <unit> at t/030-cookies.t:172" which is HTTP::Response.new(200);
grondilu waait until I update to merge
21:16 sno left, domidumont joined
RabidGravy doing my nut 21:16
Juerd gives grondilu some free whitespace
timotimo grondilu: you ... really? that's kinda cool! 21:17
21:19 kaare_ left
dalek kudo/nom: 4313b30 | timotimo++ | src/core/Exception.pm:
fix precedence in previous commit
 21:19
21:20 molaf_ left
Skarsnik RabidGravy, maybe it's because of the ? arg 21:20
grondilu it's an additional loop but it appears to be faster
21:21 ZoffixMobile left
grondilu Juerd: did I not expand correctly? 21:22
timotimo grondilu: a little loop may be a lot cheaper still than a function call that calls something else on top
grondilu: if you use postfix-while or for, you can even prevent an extra lexical scope from forming, which may be cheaper, too
21:23 domidumont left
grondilu the XOR trick requires three statement, not sure I can put them all in one. 21:23
RabidGravy Skarsnik, I can't reproduce except if I do panda install 21:24
Juerd grondilu: I don't understand your question.
Skarsnik RabidGravy, err, not fun :(- 21:25
timotimo grondilu: you can, by using STATEMENT_LIST( ... )
i used that same evil trick in JSON::Fast
grondilu I can indeed but that's kind of ugly 21:26
timotimo this is internal core setting code. it's meant to be fast and allowed to look kidna ugly
kinda*
thing is: someone on twitter recently complained about permutations performance, and i'm sure they're right to.
grondilu looks like that:
(@!a[$k] +^= @!a[$l]), (@!a[$l] = @!a[$k] +^ @!a[$l]), (@!a[$k] +^= @!a[$l]), $l-- until ++$k >= $l; 21:27
hoelzro I remember observing some bad performance with permutations, but I couldn't find an easy way to fix it =/
MadcapJake rakudo dev question: how do you start work on something new with your fork? Do you delete it and just make a new one? Or do you rebase the fork to upstream HEAD? 21:28
timotimo hoelzro: you should have grabbed me! :)
Juerd IMHO anything is allowed to look ugly if it's accompanied by a comment that explains why it's ugly, and what it does.
hoelzro MadcapJake: delete the repo? I would just rebase
timotimo MadcapJake: with one bigger thing i've done i re-based onto master (that was in moarvm) a few times and gave the branch a number at the end to point out the rebasedness 21:29
21:29 Ven left
hoelzro timotimo: =( missed opportunity! 21:29
timotimo: it's still in my "look at" list if you want to collaborate on it
timotimo grondilu: it could even be that our static optimizer inlines that block for us, so it might not be worth it to use STATEMENT_LIST 21:30
hoelzro: dude, i'm working on permutations with grondilu right now :D
hoelzro =P
grondilu timotimo: that would be preferable indeed. 21:31
timotimo grondilu: has to be measured :)
how's your rakudo build times at the moment?
grondilu I may possibly golf the line above into:
(@!a[$k] +^= @!a[$l]), @!a[$k] +^= (@!a[$l] = @!a[$k] +^ @!a[$l]), $l-- until ++$k >= $l;
21:31 n0tjack left 21:32 colomon joined
grondilu or even: 21:32
(@!a[$k] +^= @!a[$l]), @!a[$k] +^= (@!a[$l] [R+^]= @!a[$k]), $l-- until ++$k >= $l;
but htat's kind of silly 21:33
hoelzro MadcapJake: btw, do you have outstanding work that makes a rebase necessary? could you just fast forward your nom?
timotimo m: my @a = (^15).list; my int $k = 3; my int $l = 10; (@a[$k] +^= @a[$l]), @a[$k] +^= (@a[$l] = @a[$k] +^ @a[$l]), $l-- until ++$k >= $l; say @a
camelia rakudo-moar 4313b3: OUTPUT«[0 1 2 3 10 9 8 7 6 5 4 11 12 13 14]␤»
timotimo that was fast
m: my @a = (^15).list; my int $k = 3; my int $l = 11; (@a[$k] +^= @a[$l]), @a[$k] +^= (@a[$l] = @a[$k] +^ @a[$l]), $l-- until ++$k >= $l; say @a 21:34
camelia rakudo-moar 4313b3: OUTPUT«[0 1 2 3 11 10 9 8 7 6 5 4 12 13 14]␤»
timotimo m: my @a = (^15).list; my int $k = 3; my int $l = 11; STATEMENT_LIST((@a[$k] +^= @a[$l]), @a[$k] +^= (@a[$l] = @a[$k] +^ @a[$l]), $l--) until ++$k >= $l; say @a 21:35
camelia rakudo-moar 4313b3: OUTPUT«[0 1 2 3 11 10 9 8 7 6 5 4 12 13 14]␤»
grondilu if you start from $l = @a.elems you can put the $l-- at the end I think
timotimo can you check locally if either variant of these two is slower/faster?
i meant with vs without STATEMENT_LIST
grondilu with STATEMENT_LIST will always be faster, won't it? 21:36
timotimo no clue. hopefully.
grondilu because no lexical scope and stuf
timotimo the one you had up there with the commas doesn't have the lexical scope either
but it may create a list
grondilu I think this is too weird, we should stick with the version with a proper block. 21:37
timotimo sure can do.
grondilu it's in my PR if I'm not mistaken.
timotimo let me see
ah, the one from before, still 21:38
after that i'm going to have a look-see if changing the ++ and -- to assignments makes a noticable difference 21:39
21:40 sufrostico left, sufrostico joined
jdv79 looks like larry's t-shirt might not make the cut 21:42
timotimo damn, these 2 minutes of rakudo compile time ;_;
WAITWHAT 21:43
jdv79 is there some theoretical way to cleave the setting so compiling could be shorter?
timotimo just with the latest patch you put into the PR, my dear grondilu, time taken went from 3s to 0.7s
jdv79 some sort of partial/merged compunit or something
Skarsnik lol this time cut
grondilu XOR trick FTW 21:45
timotimo well, i would have said it could have been done without the xor trick
okay, sorted by exclusive time, we now have the pull-one method from our permutations iterator at the #1 spot with 22.95% exclusive time 21:46
and iterator's push-exactly at 20.12% exclusive time 21:47
then comes postcircumfix:<[ ]> at 9.99% and then elems at 4.21%
jdv79 sounds like other random profiles i've done
MadcapJake hoelzro: I don't, what would be the command to just fast-forward? I don't want to add a merge commit.
jdv79 the new list stuff pops up often
hoelzro MadcapJake: git merge --ff-only 21:48
lucs Whoops, I can't use => as an autoquoting fat comma anymore eh,
timotimo yo grondilu we can use $n - 1 from the outer scope instead of @!a.end, no?
MadcapJake oh sweet, never knew that
timotimo lucs: yeah you can
grondilu yes
hoelzro I have that aliased to gff, I use it so often =)
timotimo wait ...
lucs Hmm...
timotimo not if you mean it to give you two positional things
lucs timotimo: I did.
timotimo yeah, no, that won't fly
grondilu but in fact you can start from $!n and start with decrementing $l
lucs And that's the error that I'm getting :)
grondilu I did it in the PR
timotimo ah, cool, let me pull that change 21:49
grondilu you should wait until travis confirms it's ok though
well, wait before merging I mean
lucs timotimo: What's actually happening with the => ? (pointer to docs please)
MadcapJake hoelzro: I have the zsh git plugin but i'm surprised there isn't an alias for that command
21:49 TEttinger joined
timotimo lucs: if you're in a parameter list, it gets turned into a named parameter, otherwise it creates a Pair object 21:49
hoelzro MadcapJake: plugin? do you use oh-my-zsh?
MadcapJake yeah 21:50
hoelzro ah ha
timotimo doc.perl6.org/type/Pair - doc about pair
lucs timotimo: Thanks
hoelzro if you're curious, here are the aliases I use: github.com/hoelzro/zsh-config/blob...liases.zsh
grondilu we only have two statement in the block now, maybe it's worth putting the until at the end.
MadcapJake there's gm (git merge), gmom (git merge origin/master), gmum (git merge upstream/master)
grondilu and do the statement-list thing
timotimo doc.perl6.org/type/Capture - this has a bunch on parameters
Skarsnik I bet it's already a alias for a git oscure-cmd -Xkkl256-21~~A50 insertlinusranthere
awwaiid lucs: also doc.perl6.org/language/syntax#Pair_literals has some info 21:51
timotimo grondilu: i can try that; how long does rakudo take to compile on your machine?
oh!
lucs Always more to read, and loving it!
timotimo i think i have optimization turned off in my moarvm
for debugging purposes
awwaiid (I need to figure out how to tag that so it comes up when you search for =>)
grondilu rakudo takes almost ten minutes to compile on my machine
RabidGravy what does "Serialization Error: Unimplemented case of read_ref" mean?
MadcapJake i almost feel like I should do gdad for git merge --ff-only ;) 21:52
lucs is learning Perl 6 from the inside out, so to speak, so hits different level snags at different moments.
timotimo grondilu: how about only the core setting?
MadcapJake grondilu: iirc TEST_JOBS might help 21:53
not sure if that's just for tests though
leont It's just for the tests
make -j <number of cores> is helpful during building though 21:54
timotimo grondilu: there's a typo in your code
there's only an "a" where a @!a should be
grondilu is it?
timotimo in that loop
grondilu oh yeah
RabidGravy "Missing serialize REPR function for REPR VMException" what is going on here?
timotimo RabidGravy: an exception (perhaps via a Failure) ends up in the precompiled output of some module 21:55
grondilu I wonder if @!a[$k] +^= @!a[$l] = @!a[$k] +^ @!a[$l] requires parenthesis. I think not.
RabidGravy oh bizzare
MadcapJake leont: forgot about -j, that's good to know 21:56
jdv79 moritz: i think the bot is missing stuff lately
timotimo copy out the whole permutations class into a simple script and try it out :)
grondilu that's what I've been doing from the beginning :)
timotimo oh, damn it. now i can't compare my numbers against the old numbers any more because i turned optimization off in moar for the old numbers and on for the new ones >:(
good!
i've been an idiot and editing inside the core setting
i need a bigger list to test this stuff with 21:57
it's getting too fast
grondilu been there, done that
21:57 rindolf left
[Coke] leont: "make -j" is sufficient, I think. 21:57
timotimo oh damn. added two numbers, the time ballooned up from 0.31s to 22.46s :D 21:58
21:58 rindolf joined
MadcapJake are command line options stored in %adverbs (in Compiler.nqp) 21:58
hoelzro MadcapJake: yes 21:59
timotimo m: STATEMENT_LIST(say "hi"; say "goodbye")
camelia rakudo-moar 4313b3: OUTPUT«hi␤goodbye␤» 22:00
timotimo i think i used a , in JSON::Fast
grondilu: using "statement_list" doesn't actually make it any better. 22:03
grondilu ok
masak m: say 2 + 2; sub infix:<+>($l, $r) { "!!!" } 22:04
camelia rakudo-moar 4313b3: OUTPUT«!!!␤»
masak overridden operators can be post-declared. freaky.
22:04 regreg joined
grondilu I've just realised @!a is not defined as a native array. Not sure it matters. 22:04
22:05 xfix left
xjrK never trust these hoes; 22:05
timotimo i told you before; i also said i don't know if a native int array would be helpful or not right now 22:06
flussence "make -j" without a limit is usually a terrible idea
22:06 skids left
MadcapJake how can I get environment info in nqp (perl6 version, vm, distro, time), none of the dynamic perl6 variables seem to be workingt 22:07
flussence if you don't want to hardcode a number, at least use "make -j $(nproc)"
timotimo flussence: that's how i build moarvm, tbh
flussence timotimo: most of my systems would blow up around the 30-40 process mark 22:08
timotimo :)
i don't actually know how many processes make fans that out to
Zoffix .botsnack
yoleaux :D
17:16Z <dj_goku> Zoffix: just wanted to say thanks for the IRC Client advent post. I took the plung on prototyping a interacting with gearmand in perl6. I was able to get connected last night!
Zoffix dj_goku, cool. Glad it was useful. 22:09
timotimo ja!
grondilu: replacing -- and ++ with explicit assignments got me to 2.27 from 2.4 22:11
hm, more like to 2.27 from 2.35 22:12
well, 2.37 would be more fitting
22:12 colomon left
grondilu it's a bit LTA that ++ and -- are not efficient. 22:12
timotimo yes.
we know, we'll work on it.
22:13 regreg left
timotimo grondilu: are you interested in fleshing out the permutations spec tests? 22:14
22:14 regreg joined
timotimo because damn, those are *weak* 22:14
grondilu: i expect there's more to be gained from turning postcircumfix:<[ ]> into .AT-POS
but that'll make things very ugly indeed 22:15
grondilu I've made suggestions already: github.com/perl6/roast/issues/79
it's definitely LTA that .AT-POS is more efficient than postcircumfix:<[ ]> 22:16
timotimo yes.
grondilu I suggest we keep [ ] until it's fixed.
and same for ++ and --
22:16 Zoffix left
timotimo can you compare the algorithm suggested in that ticket against the current implementation now that we've micro-optimized it? 22:17
grondilu it will still be faster I think
but ok I can check 22:18
though the other algorithm could also use some XOR magic 22:19
so there's not much point to compare.
(until the otehr algorithm is re-written, that is)
timotimo sure
grondilu and we can't use it until the test are re-written anyway 22:20
so let's just keep it in cold storage
timotimo fair enough 22:21
i would accept the suggested change to the permutations test, btw
to sort it
flussence «Stage parse : 372.765» -- my 32-bit, 1.6GHz netbook is only 4x as slow as my desktop at this. Thought it'd be much worse... 22:22
timotimo wow
grondilu: interestingly, the .List method isn't getting jitted. perhaps there's a tiny win to be had there, too. 22:25
grondilu: right now, though, Iterator's general "push-exactly" is where we spend 25% of run time and only 20% run time is spent in pull-one of the permutation iterator 22:26
quite frankly, that surprises me a little bit
22:27 ChoHag left
grondilu I don't know what "push-exactly" is. I'll have a look. 22:27
timotimo it basically iterates over .pull-one until the target has as much as it wants 22:28
lol. oh crap. it gets called exactly as many times as pull-one minus 2 22:29
so it seems like it's calling push-exactly with "please give me one entry" each time
22:30 raoulvdberge left
grondilu isn't that normal? 22:31
timotimo why doesn't it call push-one or pull-one instead?
perhaps it'd be worth it to give the general push-exactly a "fast path" for the argument "gimme one"
regreg does perl6 have any GUI library? 22:32
jdv79 wasnt that part of the design? all one needs to impl is pull-one and the rest builds on that (probably suboptimally)
grondilu I don't know. Those Seq and Iterator classes are complicated. 22:33
Skarsnik regreg, there is a part of Gtk that has a binding (search gtk-simple on the ecosystem)
:eco gtk
x)
timotimo tries it out
grondilu FYI, I took Str.comb as a model to write permutations.
22:33 joeschom joined
timotimo yeah, that was a good idea 22:35
weird. my fast path doesn't get executed 22:37
22:38 PotatoGim left 22:39 cognominal joined
timotimo please push exactly 2147483647 22:39
so why the F does that method get called a fuckton of times? 22:40
Skarsnik to piss you off?
timotimo probably. 22:41
hum. 22:43
it does push 8 at a time
i really don't get it :) 22:44
jdv79 you need to become one with the GLR
grondilu yeah that's definitely GLR-specific stuff 22:45
timotimo oh! 22:46
there's another pull-one that gets called 362951 times
40323 calls of push-exactly and 40321 calls of the "expensive" pull-one 22:47
imagine my surprise :)
22:49 sno joined 22:50 sno left 22:59 joeschom left
lizmat timotimo: are you sure it's not calling the "default" push-exactly ? 23:02
timotimo lizmat: it is! 23:03
lizmat ah, but every time only for 1 elem, right ?
timotimo nah, many thousands :)
lizmat I recall seeing that and wondered why it wouldn't be calling pull-one directly
timotimo it's not as bad as i thought it'd be 23:04
grondilu: i'm about to merge your PR. are you ok with that?
lizmat: i wonder if we can do better than what List.permutations does now 23:07
lizmat probably :-)
23:07 RabidGravy left
timotimo like, the normal permutations function currently does a full copy of @!a, but the permutations method of list doesn't modify that list, so it *could* return the original. but it would have to get that passed in somehow 23:08
masak 'night, #perl6
timotimo ginte masak!
the overhead of pulling the values from that list is about 6 seconds 23:09
that was less than half when we began optimizing, but now it's almost 3x as much :)
nine MadcapJake: there's nqp::getenvhash 23:10
dalek kudo/nom: 7e9610c | grondilu++ | src/core/native_array.pm:
small permutations improvement

use XOR swap trick, as suggested by timotimo		 23:12
23:12 dalek left 23:13 dalek joined, ChanServ sets mode: +v dalek, loren joined 23:15 loren left
nine Why oh why does the require Inline::Perl5; in src/core/control.pm's EVAL sub not find CompUnit::RepositoryRegistry at runtime ($*W.symbol_lookup(['CompUnit', 'RepositoryRegistry'], $/)) while every other require statement can? 23:15
lizmat nine: because you twiddle with the context ? 23:16
23:16 Skarsnik left
nine lizmat: err...what does that mean? 23:16
23:17 regreg left
lizmat so, you're not passing a :context ? 23:17
nine lizmat: not that I'm aware of, no 23:18
Could not find symbol 'RepositoryRegistry'
23:18 sufrostico left 23:19 kid51 joined
timotimo what does your block stack look like at that point? 23:19
because all it does is look at all the Block objects it's entered so far and try to find CompUnit::RepositoryRegistry in there
lizmat nine: maybe you need to stub it ?
nine Is symbol_lookup actually the right tool for this job? This used to be find_symbol but I degraded it to a runtime lookup (or at least try to) so I could get rid of CompUnit::RepositoryRegistry in the BOOTSTRAP 23:21
lizmat nine: not sure...
nine That's the full backtrace: gist.github.com/niner/67d4f2170d7e37db3858 23:22
What I find odd about it is the FALLBACK in there
Makes me think that maybe I'm looking at the wrong place. But there's not that many places which refer to CompUnit::RepositoryRegistry (formerly CompUnitRepo) directly. And only one where the name is actually split into those components. 23:23
timotimo nine: comes from the Failure class 23:24
nine Ok there's 2 places: the other is method comp_unit in Perl6::Actions
timotimo hoelzro: did you see? permutations($n) improved by almost 6x 23:27
hoelzro \o/ 23:28
timotimo++ grondilu++
way to go!
timotimo i think the permutations method on ranges should check if the range is from 0 to some value and just call the permutations sub directly instead of self.flat.permutations
nine Oooh...it _is_ actually method comp_unit if the debug print is to be believed 23:31
23:32 Zoffix joined
nine And there I still use find_symbol which will not find it because it's no longer in BOOTSTRAP 23:33
timotimo oh! 23:34
23:38 lucasb joined
lucasb I read in the backlog that ++/-- are slower than manually increment/decrement? 23:40
23:41 Zoffix left, PotatoGim joined
timotimo lucasb: on native integers 23:41
lizmat at the moment 23:43
lucasb ah, I just tested. thanks, timotimo
nine Now I'm back to "Cannot call method 'AT-KEY' on a null object" when loading a precomp file :/
lucasb strange that with "my $i", $i++ is faster than "$i=$i+1", but with "my int $i", it's the other way around
flussence is Str.encode('utf16') big or little endian?
23:44 Zoffix joined
Zoffix Hm, rakudobrew build-panda still freezes for me on JSON::Fast :( 23:45
lucasb panda wasn't updated
Zoffix oh, it includes it with it :/
lucasb you can call update-subtree.pl ext/JSON__Fast in panda root 23:46
timotimo oh, sorry, i forgot to do that
Zoffix Where's panda root?
23:47 cdg left
timotimo there we go. 23:47
Zoffix looks like ~/.rakudobrew/moar-nom/panda
lizmat lucasb: that's not strange if you realize that for a native int to be ++, it basically needs to upgraded to an Int, and then downgraded again
timotimo lizmat: more or less 23:48
Zoffix timotimo++ that fixed it thanks.
lucasb lizmat: hm, I didn't know that, thanks for the info
23:48 Zoffix left
timotimo lucasb: the main source of performance trouble is that our inliner doesn't yet understand ++ on native ints, so it ends up doing full function calls 23:48
lucasb so, IOW, boxing and unboxing just to increment it :(
timotimo not quite boxing/unboxing. taking a native reference in this case 23:49
plus, on something as simple and cheap as ++, making the reference and calling the function is 99% overhead at least
at some point, spesh will understand native references completely and when the ++ sub gets inlined then, it'll not have to generate the reference; it'll just do the increment/decrement in place on the register and *then* it'll be crazy cheap 23:50
the thing about Int, on the other hand, is that it's a full object with the potential to become infinitely big
23:51 colomon joined
timotimo lucasb: does that explain things? 23:51
lucasb timotimo++, explain very well! :) 23:52
timotimo i'm deep into the whole performance topic. sadly, not good enough at it to make a dent in either of those problems :| 23:56
either meaning a) the inliner recognizing NativeRef properly and b) spesh not turning NativeRef into a direct usage of the given register/lexical

Powered by Cro and the Raku® Programming Language.

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

Thank you!