»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
Zoffix What's the difference between +@ and *@ slurpries? 00:37
m: sub a(+@a){ dd @a }; sub b(*@a){ dd @a }; a 1, 2, 3; b 1, 2, 3; a 1; b 1; my @a = ^2; a @a; b @a;
camelia rakudo-moar 8be36b: OUTPUT«[1, 2, 3]␤[1, 2, 3]␤[1]␤[1]␤[0, 1]␤[0, 1]␤»
Zoffix No idea what this sentence is supposed to mean: "The single argument rule allows to treat arguments to subroutines, for-loops and list constructors based on context." docs.perl6.org/type/Signature#Sing...ule_Slurpy 00:38
Zoffix m: sub a(+@a){ dd @a }; sub b(*@a){ dd @a }; a [1, 2], 3; b [1, 2], 3; 00:41
camelia rakudo-moar 8be36b: OUTPUT«[[1, 2], 3]␤[1, 2, 3]␤»
skids Gah the error disappears under valgrind. :-( 00:44
skids m: sub a(+@a){ dd @a }; sub b(*@a){ dd @a }; a((1,2),3); b((1,2),3); # Zoffix 00:52
camelia rakudo-moar 8be36b: OUTPUT«[(1, 2), 3]␤[1, 2, 3]␤»
skids see also docs.perl6.org/language/functions#...onventions 00:53
TimToady + says a semantic list is expected without caring whether that list is made by syntactic commas, while * and ** mean a syntactic comma list is expected; ** hides commas inside parens so you can make a list of lists, while * makes a flat list by hoisting parenthesized lists (and interpolated lists) up to the top level
skids m: sub a(+@a){ dd @a }; sub b(*@a){ dd @a }; sub c(**@a){ dd @a }; a(($_ for 1,2,3)); b((1,2),3); c((1,2),3); a(($_ for 1,2,3)); b(($_ for 1,2,3)); c(($_ for 1,2,3)); 00:57
camelia rakudo-moar 8be36b: OUTPUT«[1, 2, 3]␤[1, 2, 3]␤[(1, 2), 3]␤[1, 2, 3]␤[1, 2, 3]␤[(1, 2, 3),]␤»
MasteDuke m: say "⒗".uniprop; say "١".uniprop 02:04
camelia rakudo-moar 8be36b: OUTPUT«No␤Nd␤»
MasteDuke m: say "⒗".unival 02:08
camelia rakudo-moar 8be36b: OUTPUT«16␤»
MasteDuke .tell harmil_wk ⒗ won't work by design for a bunch of stuff. see github.com/rakudo/rakudo/commit/e2...6d25995a90 for some discussion. it has a .uniprop of 'No' instead of 'Nd' 02:14
yoleaux MasteDuke: I'll pass your message to harmil_wk.
sprocket hello #perl6! 02:48
i just updated to the latest version of rakudo on my raspberry pi, and wanted to say thank to whomever is responsible for the markedly faster startup times :) 02:49
skids o/ They aren't awake probably but I'm sure they'll backlog. 02:50
AlexDaniel sprocket: huh, did you compile it yourself right on raspberry pi? :) 03:02
sprocket using rakudobrew, yes 03:02
AlexDaniel sprocket: how long did it take? 03:03
sprocket the last version i’d had was 2016.04 i believe, and it was about a 20 second startup when loading a perl6 program
i think compilation of rakudo took about a hour? i dunno, i was doing other things :)
but 2016.09 perl6 binary starts up in a couple seconds 03:04
AlexDaniel interesting. I've always pulled it from debian unstable because I am impatient
sprocket makes testing a lot nicer
AlexDaniel at this moment it is slightly outdated :( 2016.07.1
skids m: class A { my $e = 42 }; import A; say A::("\$e") # I wonder what that message is trying to say 03:23
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Combination of indirect name lookup and call not supported␤at <tmp>:1␤------> 3onder what that message is trying to say7⏏5<EOL>␤ expecting any of:␤ argument list␤»
skids Also trying to figure out how to acess symbols from another module without importing into own namespace. 03:24
m:
m: class A { $e = 42 }; import A; $A::e.perl.say 03:25
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '$e' is not declared␤at <tmp>:1␤------> 3class A { 7⏏5$e = 42 }; import A; $A::e.perl.say␤»
skids m: class A { my $e = 42 }; import A; $A::e.perl.say
camelia rakudo-moar 8be36b: OUTPUT«Any␤»
skids This is confusing.
(not the first one, that the symbol exists but is Any in the second one.) 03:26
geekosaur I think I could argue that the bug there is that it is visible at all; a "my" should not be accessible, at least without using namespace magic 03:39
m: class A { our $e = 42 }; $A::e.perl.say 03:40
camelia rakudo-moar 8be36b: OUTPUT«42␤»
skids "our" only works in the same compunit. 03:41
Hrm. Grr. No it seems to work, in a simple test module. I guess it must be something to do with the fact that my modules share a base namespace. 03:43
TimToady m: class A { my $e = 42 }; import A; $a::nonesuch.perl.say 03:45
camelia rakudo-moar 8be36b: OUTPUT«Any␤»
TimToady that's what's really going on
m: class A { my $e = 42 }; import A; $A::nonesuch.perl.say
camelia rakudo-moar 8be36b: OUTPUT«Any␤»
TimToady $A::nonesuch is assumed to be a package variable that might or might not exist yet
geekosaur heh. that first one is actually even more informative 03:46
TimToady otherwise escape analysis would be completely impossible
skids Ahah maye figured it out. 03:49
One of the modules isn't defining a name that matches its compunit's path. 03:51
yeah that did it. I wonder if that counts as a bug in the file-based repo code. 03:57
Oh nm. I'm obviously just up too late. 04:02
nine DrForr: rt.perl.org/Public/Bug/Display.htm...xn-1410204 07:23
El_Che build 2016.09 packages as we speak 08:41
building 2016.09 packages as we speak
raydiak in a grammar, is there a better way to entirely ignore things like comments without having a separate preprocessing grammar or putting <comment> between every atom of every rule? 08:43
timotimo you can make it part of your own "ws" rule 08:44
and then use sigspace in the appropriate places
raydiak thought about that...guess I could...but then I can't use sigspace as intended elsewhere in the grammar, because I even want it to ignore them in the middle of sequences which don't allow whitespace 08:46
timotimo ah
arnsholt Depending on how the grammar is factored, you might hack around it with dynvars 08:47
I did that to have two different <ws>es in Snake, depending on whether we're inside parens or not
raydiak I see 08:48
arnsholt So if the toggle is called $*WS-MODE, you can then do method ws { $*WS-MODE ?? self.ws1 !! self.ws2 } 08:50
raydiak was just looking at github.com/arnsholt/snake/blob/mas...ar.nqp#L74
arnsholt And ":$*WS-MODE = 1;" in a rule to change it
Yup
That's it
Of course, that's NQP, but it should be pretty much identical in Perl 6 08:51
Except you can use assignment instead of binding =)
timotimo how does ":$*WS-MODE = 1" behave in regards to backtracking? 08:52
wouldn't it be wiser to have a "my" to go with it first?
arnsholt Oh, derp
It's :my $*WS-MODE = 1; yeah
raydiak would make things a bit messy, having to have spaces literally everywhere and changing a dynvar instead of just writing "rule". any reason that'd be better than a preprocessing grammar? I have doubts about it performing better. 08:54
timotimo right. difficult. 08:56
jnthn You'd still write rule, the mode would just tweak what ws considers itself allowed to match 08:57
raydiak well yes but I'd be writing rule everywhere because token wouldn't do what I want, its no longer a useful choice 08:59
arnsholt What are you parsing, OOC?
raydiak markdown in this case 09:00
arnsholt Oh, horror! =)
IIRC MarkDown is not super-amenable to CFG parsing
raydiak guess that explains why there's not an actual Grammar for it in our ecosystem yet 09:01
arnsholt ISTR it originally being implemented by a cascade of regex search-replace
And it turns out cascades of search-replace let you do context *sensitive* things 09:02
raydiak incidentally, I don't even know if markdown usually allows comments or not in the places I'm being asked to allow them 09:03
timotimo when people see "markdown supported", they probably expect commonmark instead of the original markdown anyway, no? 09:04
arnsholt I have no idea
jnthn reads the first answer to stackoverflow.com/questions/4823468...n-markdown and is mildly horrified :) 09:05
raydiak actually it's a vague mix of multimarkdown and some other variants and some of our own proprietary additions that I'm supposed to parse :) 09:06
timotimo yay
raydiak its coming along well so far though, near as I can tell
if I just do "pile of regexes", I run into issues differentiating between things like bulleted lists and bolded chunks of text with linebreaks in the middle 09:10
brrt wonders if org-mode syntax is more parseable... 09:11
given as it has multiple-language support
arnsholt I'm currently leaning towards RST 09:13
Haven't looked at org-mode though, being a non-emacser =) 09:14
raydiak yeah that SO answer is pretty horrifying. we had thought about adding another comment syntax, but I'm hoping to avoid adding much of our own because I know nobody here is a real language designer and I fear we'll end up in impossible situations which are difficult to forsee
arnsholt RST syntax is a bit clunkier than MarkDown, but it has the definite advantage of being much more clearly defined, and has well-defined ways of extending it
timotimo raydiak: if you build your own comment syntax, you can potentially pre-process the whole document to strip out the comments and continue with a regular grammar :P
arnsholt raydiak: Are you gree-fielding, or implementing to support something that's already extant? 09:15
*green-fielding
raydiak preprocessing was what I waas wondering if there was a way to avoid int he first place. I can preprocess the html comments out just as easily
timotimo since grammars can't do streaming parsing yet anyway, you'll not lose much from pre-processing 09:16
arnsholt Oh, that's the other horrifying thing about MarkDown: Allowing HTML
timotimo a little bit, yeah
on the other hand: yay, graphics embedded into markdown by way of SVG!
raydiak there is a large amount of existing data in this...vaguely-defined pile of PHP loops and PCREs that I'm replacing 09:18
brrt org-mode is awesome
raydiak but not so much with the mixing in of HTML, if my understanding is correct. I only have a small amount of test data
brrt in both senses of that word
arnsholt raydiak: Yeah, that's why I asked. If you were making something new, I'd have urged you to consider something not MarkDown, but when you have an existing corpus to work with, you can't 09:19
raydiak it's good advice, wish I could take it :)
arnsholt Anyways, if you have an existing implementation and a corpus of source/rendered pairs, the first thing I'd do is extract all the source/rendered pairs into unit-tests 09:20
And then copy the existing implementation, horrors and all, as closely as possible
raydiak also there are a number of tools people are already using here that need to support what we're doing, even if they don't recognize our additions and modifications
arnsholt And then try to replace the horrors once you have something that works 09:21
raydiak I'll think about that with at least a subset of the data. it's a lot
arnsholt Yeah, if it's a massive corpus, extract a random sample
But I think I'd extract a really large sample as well as a kind of stress test 09:22
Actually, it might be attractive to do low-level hackery on the Cursor API here
The token/regex/rule stuff lets you do CFG-stuff, but if you do things with methods and manipulate the underlying cursor directly, you might be able to use declarative stuff for most of it, and then wire the loops and horid stuff together using lower-level primitives 09:23
arnsholt I'm not very familiar with the Cursor API though, and it might not be very well documented, so you might have to do some non-trivial source-diving to figure out exactly how to do it 09:24
raydiak don't know that I have the time to get that deep into it, this is only a small part of a much larger project with a way-too-soon deadline 09:25
just gonna preprocess for now, but you all helped me figure that out :) 09:26
arnsholt Good that we could help =)
raydiak might even end up with more than two grammar stages if it gets hairy, to somewhat mimic the regex pile...but making tests is a smart idea 09:28
and that cursor api thought is really interesting, I've got that filed away in the back of my mind for another time :)
raydiak one of my saving graces is that we don't just have proprietary additions, but also a few subtractions and simplifications...really calling it "markdown" is a tad bit of a stretch in places 09:34
anyway, thanks timotimo++ arnsholt++ brrt++ jnthn++ 09:35
nadim Good morning all
raydiak good morning nadim
raydiak thinks about future needs and considers trying again to fix Grammar::Generative 09:42
but who knows how far gone that is now that it's been rotting since...2013...so much has changed since then, not just the GLR and method names 09:43
llfourn_ "Method 'foo' must be resolved by class Bar because it exists in multiple roles" 10:26
how does one "resolve" this?
jnthn Write a method foo in the class that decides what to do
llfourn_ jnthn: kk thanks 10:27
lizmat m: role A { method a {"A"} }; role B { method a {"B"} }; class C does A does B { method a { self.A::a } }; dd C.new.a # llfourn_ 10:44
camelia rakudo-moar 8be36b: OUTPUT«"A"␤»
llfourn_ lizmat: thanks :) 10:49
maybe the error message could be clearer... 10:54
"Bar must have a method 'foo' because it exists in multiple roles"
although I get the reason why "resolve" is used now it does immediately suggest the solution 10:55
doesn't*
jnthn It'd be nice if the error said which roles also 10:56
llfourn_ jnthn: it does 10:57
I just cut it short :)
jnthn Oh! 10:58
So, it' snot entirely LTA then :)
eww, snot
*it's not
llfourn_ no, it's pretty good. Just might need to point more directly at the solution. 10:59
lizmat clickbaits p6weekly.wordpress.com/2016/09/19/...he-robots/ 11:14
masak lizmat++ 11:16
Ven` raydiak: that'd be really interesting :-) 11:22
moritz lizmat++ 12:05
[Coke] caffeine by committe this morning: should I drink get coke zero, coffee, or tea to start? 12:30
*ee 12:31
s/drink// - 'cause I need something, it seems.
brrt tea
tadzik is "zero" in "coke zero" the amount of teeth that you'll have in your 80s? :P 12:40
tadzik I'd usually go for tea on normal days and coffee if I really desperately need to wake up after too little sleep 12:41
masak [Coke]: tea, but good tea 12:43
or, failing that, coffee but good coffee
spoil yourself a little with quality stuff!
Ven` ;q 12:44
oh well.
masak Ven`: why is the little q crying? :/ 12:45
or is it just trying to exit vim?
moritz don't cry for me little q, the truth is, I never left you!
masak .oO( come on, colon, I only semi-left you ) 12:46
masak .oO( exit half of vim ) 12:47
jkramer m: perl6 -e 'class Foo {}; my $x = Foo; { my $x.=new; say $x.WHAT }'
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Two terms in a row␤at <tmp>:1␤------> 3perl6 -e7⏏5 'class Foo {}; my $x = Foo; { my $x.=ne␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end…»
jkramer Oops
m: class Foo {}; my $x = Foo; { my $x.=new; say $x.WHAT }
camelia rakudo-moar 8be36b: OUTPUT«(Any)␤»
masak .oO( ؛q -- exit the other half of vim )
jkramer m: class Foo {}; my $x = Foo; { my $y = $x.new; say $y.WHAT }
camelia rakudo-moar 8be36b: OUTPUT«(Foo)␤»
jkramer Why is this?
masak jkramer: because instances also have a .new method 12:48
jkramer: does that answer your question?
it's occasionally useful -- I believe the 007 code base uses or used the fact that you can call .new on a non-type object
jkramer masak: Nope, I'm wondering why the result of the first one is Any
masak jkramer: because in that case `$x` starts out as `Any`, and then you call `.new` on that 12:49
jkramer Also:
m: class Foo {}; my $x = Foo; { say $x.new.WHAT }
camelia rakudo-moar 8be36b: OUTPUT«(Foo)␤»
jkramer here it works again
masak jkramer: ah, I see you *think* that you're using the *outer* `$x` in the first program
jkramer Just the assignment not
masak jkramer: that's not the case.
jkramer Oh
Ven` masak: yes. I have `(define-key state-map (kbd ";") 'evil-ex)` in my config :)
masak jkramer: once you introduce a new `$x` in the innter scope,
that's what you get
lizmat m: class Foo {}; my Foo $x; { $x.=new.WHAT } 12:50
camelia ( no output )
lizmat m: class Foo {}; my Foo $x; { say $x.=new.WHAT } 12:50
camelia rakudo-moar 8be36b: OUTPUT«(Foo)␤»
Ven` (yes, I'm running evil, vim emulation inside emacs. And I'm connected to IRC via said emacs...)
jkramer Ok, I see :)
masak jkramer: I think this is a difference from Perl 5.
Ven` masak: I'm curious as to how you type that "reversed" semi, however :) 12:51
.oO( Is the semi-colon half-full or half-empty? )
masak .u ؛ 12:51
yoleaux U+061B ARABIC SEMICOLON [Po] (؛)
masak I didn't type it, I used au++'s App::Uni 12:51
I use that one quite a lot
[Coke] masak: quality is not an option this morning. 12:52
tea it is.
Ven` yoleaux++
I must say, I'm pleasantly surprised with ERC. But I need to find a way to enable vim emulation here.
masak [Coke]: I curse you with the task of seeking out quality ASAP, if not this morning.
Ven`: I was on ERC for a long time. I liked it a lot.
[Coke] lipton bag tea, with splenda. 12:56
masak .oO( I am altering the curse. pray I do not alter it any further. ) 13:00
[Coke] wow, this is not good. 13:01
masak whyy, committee, whyyy?!
[Coke] next up, I need to find a comittee to tell me what to work on today! 13:12
... and more letters to use.
jkramer m: role R { has Str @x }; class C does R { @x = <foo bar baz> } 13:28
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '@x' is not declared␤at <tmp>:1␤------> 3role R { has Str @x }; class C does R { 7⏏5@x = <foo bar baz> }␤» 13:28
jkramer How can I make this work?
Declaring @x in C as well also throws an error
masak use @!x 13:28
OO attributes want a twigil
and you may want to initialize @!x inside a BUILD submethod in the class 13:29
jkramer m: role R { has Str @!x }; class C does R { @!x = <foo bar baz> }
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable @!x used where no 'self' is available␤at <tmp>:1␤------> 3 R { has Str @!x }; class C does R { @!x7⏏5 = <foo bar baz> }␤»
masak m: role R { has @.x }; class C does R { submethod BUILD { @!x = <foo bar baz> } }; say C.new.x
camelia rakudo-moar 8be36b: OUTPUT«[foo bar baz]␤»
jkramer Hmm, is there no easier way to override the value? 13:30
masak m: role R { method x { () } }; class C does R { method x { <foo bar baz> } }; say C.new.x
camelia rakudo-moar 8be36b: OUTPUT«(foo bar baz)␤»
jkramer Can role parameters be named? I.e. role R[Str :@x] or something? 13:32
masak sure
they're full-fledged parameters
jkramer \o/
masak I could've answered "try it and see" on that one :)
jkramer Arrrgh, I presume a role parameter can't have the role as type? :D 13:36
m: role R[R :$x] {}
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Invalid typename 'R' in parameter declaration.␤at <tmp>:1␤------> 3role R[R7⏏5 :$x] {}␤»
masak I think that runs up against R not being declared yet inside the `[ ]` brackets 13:37
jnthn, correct me if I'm wrong
jnthn You'd be right
I guess because the signature is considered part of the role's long name
masak right
m: role R {}; role R[R :$x] {}
camelia ( no output )
masak that works :) 13:37
jnthn We might be able to fudge it though.
masak m: role R { ... }; role R[R :$x] {} 13:38
camelia ( no output )
jnthn But yeah, that workaround works fine enough 13:38
masak yeah
I don't think we should over-promise by making names available before they're even declared :P
jkramer m: sub x(--> Array[Int]) { return [] }; x 13:47
camelia rakudo-moar 8be36b: OUTPUT«Type check failed for return value; expected Array[Int] but got Array ($[])␤ in sub x at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
jkramer How can I make this work? :)
Ah got it
m: sub x(--> Array[Int]) { return Array[Int].new }; x
camelia ( no output )
masak m: sub x(--> Array[Int]) { my Int @x = 1, 2, 3; return @x }; say x 13:50
camelia rakudo-moar 8be36b: OUTPUT«[1 2 3]␤»
masak typing of containers is nominal in Perl 6, not structural 13:51
masak so it's not enough for an array to contain only the desired type of values; it must also be declared to do so 13:51
jkramer Alright 13:54
masak m: sub x(--> Array[Real]) { my Int @x = 1, 2, 3; return @x }; say x 13:57
camelia rakudo-moar 8be36b: OUTPUT«Type check failed for return value; expected Array[Real] but got Array[Int] (Array[Int].new(1, 2, 3))␤ in sub x at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
masak also, arrays are not covariant, it seems. probably a good thing.
SmokeMachine____ moritz: thanks! 14:06
Ven` masak: I'd agree with that :-)
dalek c: 724e5fa | coke++ | doc/Language/ (2 files):
fix whitespace
 14:12
MetaZoffix Human relations question: is naming an IRC bot responsible for reporting test coverage results "CoverGirl" sexist or undesirable? 14:14
moritz I like the name, but I can see the potential for ill will 14:16
MetaZoffix k. I'll think of something else :) 14:17
ilmari covertly? 14:17
MetaZoffix That's good :)
AlexDaniel … coverable? ;D 14:18
moritz lid
because the lid covers the eye
(idea courtesy by masak++)
optikalmouse goooood morning
mst MetaZoffix: given it's mostly there to upset people about their coverage being insufficient 14:19
MetaZoffix: I vote for coverrage
COVERRAGE SMASH 20% REPORT
[Coke] undercover(ed) 14:20
mst [Coke]++
DrForr 'Undercover Brother' is as bad in the wrong direction :) 14:21
MetaZoffix goes with undercover, as it's an available nick. [Coke]++
vcv spin off of [Coke]'s idea: uncover(ed) 14:21
MetaZoffix Thanks, all! 14:22
timotimo is there a reason we don't allow #`» ... « for comments? 14:24
m: say » hi there «
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Missing infix inside hyper␤at <tmp>:1␤------> 3say »7⏏5 hi there «␤ expecting any of:␤ infix␤ infix stopper␤»
timotimo well ...
for code i understand it
moritz because the quotes pair the other way around 14:25
timotimo not if you're french
masak was gonna say -- think it's language-dependent
timotimo we allow smart quotes for all kinds of stuff, but not for this 14:26
masak I for one wouldn't grieve too much if the above never worked
but YMMV
timotimo m: say foo ~~ / foo / 14:27
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Expected a term, but found either infix ~~ or redundant prefix ~␤ (to suppress this message, please use a space like ~ ~)␤at <tmp>:1␤------> 3say foo ~~7⏏5 / foo /␤»
timotimo m: say foo ~~ rx/ foo / 14:27
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Expected a term, but found either infix ~~ or redundant prefix ~␤ (to suppress this message, please use a space like ~ ~)␤at <tmp>:1␤------> 3say foo ~~7⏏5 rx/ foo /␤»
timotimo m: say "foo" ~~ rx/ foo /
camelia rakudo-moar 8be36b: OUTPUT«｢foo｣␤»
jkramer Is there a nicer way for saying: <foo 123 baz>.map: -> $x { val($x) } 14:29
jkramer .map: val(*) doesn't seem to work and >>.val doesn't work because it's not a method 14:30
timotimo m: say < foo 123 baz >.perl
camelia rakudo-moar 8be36b: OUTPUT«("foo", IntStr.new(123, "123"), "baz")␤»
timotimo m: say <foo 123 baz>.perl
camelia rakudo-moar 8be36b: OUTPUT«("foo", IntStr.new(123, "123"), "baz")␤»
timotimo it already goes through val
jkramer Wut
timotimo m: say qqA test A
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared name:␤ A used at line 1␤Undeclared routines:␤ qqA used at line 1␤ test used at line 1␤␤»
timotimo m: say qq☺ test ☺
camelia rakudo-moar 8be36b: OUTPUT« test ␤»
timotimo m: say qq☺test☺ 14:31
camelia rakudo-moar 8be36b: OUTPUT«test␤»
AlexDaniel m: say qq» test »
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Use of a closing delimiter for an opener is reserved␤at <tmp>:1␤------> 3say qq7⏏5» test »␤»
jkramer Is this specific to <...> or lists of stuff in general?
masak specific to <...>
well, and <<...>>
jkramer Ah, then ignore that and assume it's an Array of Str :)
jkramer I just used <> for demonstrating 14:31
DrForr m: say qq AtestA 14:32
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Alphanumeric character is not allowed as a delimiter␤at <tmp>:1␤------> 3say qq7⏏5 AtestA␤»
masak m: say ("foo", "123", "baz")>>.&val
camelia rakudo-moar 8be36b: OUTPUT«(foo 123 baz)␤»
masak m: say ("foo", "123", "baz")>>.&val.perl
camelia rakudo-moar 8be36b: OUTPUT«("foo", IntStr.new(123, "123"), "baz")␤»
DrForr Huh, thought I tested for that.
jkramer So >>.& works for any routine? 14:33
timotimo that's right
just like .&foo will also work
jkramer Sweet, thanks! :)
masak .oO( we're lazy, so we just made it work consistently for everything )
jkramer Ha, didn't know that either. I should read through the infinite list of operators some time :)
masak >>. is technically a dotty, not an operator :P 14:34
and & is a sigil
AlexDaniel m: say ("foo", "123", "baz")>>.&{“Blah! $_”}.perl
camelia rakudo-moar 8be36b: OUTPUT«("Blah! foo", "Blah! 123", "Blah! baz")␤»
AlexDaniel ↑ and here's how you can do almost anything
jkramer OMG :D 14:35
masak our aim is to leave language purists with their mouths agape, unable to utter a word in protest
jkramer That's genius
timotimo twitter.com/AFresh1/status/778054292804546561 - does my answer seem good?
Juerd .{ ... } works? Wow!
jkramer So what's the point of .map then? :) I'll never use it again 14:36
AlexDaniel jkramer: um, they are completely different
moritz jkramer: map works with multi-arg blocks, for example 14:37
also, map is lazy, >> is very eager
AlexDaniel and » is potentially parallelizable
vcv m: say <a 1 b 2 c 3>.map: * ~ "=" ~ * 14:38
camelia rakudo-moar 8be36b: OUTPUT«(a=1 b=2 c=3)␤»
Juerd Ah, I forgot the &. But still, .&{ ... } is amazing.
timotimo &.foo is also a thing, but a very different thing :P 14:39
gfldex m: dd <foo 123 baz>.map: { .&val }
camelia rakudo-moar 8be36b: OUTPUT«("foo", IntStr.new(123, "123"), "baz").Seq␤»
timotimo and & .foo is yet another very different thing 14:40
mst well, that's neat
one can't use * as a method name
but one can define a postfix .*
Juerd mst: What are you going to use that for? Call a random method? :P 14:40
timotimo fwiw, .*blah already exists 14:41
mst timotimo: which is?
masak mst:
m: class C { method foo { say "C" } }; class D is C { method foo { say "D" } }; D.new.*foo
camelia rakudo-moar 8be36b: OUTPUT«D␤C␤»
timotimo yes, that
masak "call all methods in the mro"
timotimo in that vain, .+ also exists
masak vein* :)
timotimo m: say "heyo".+vain
camelia rakudo-moar 8be36b: OUTPUT«No such method 'vain' for invocant of type 'Str'␤ in block <unit> at <tmp> line 1␤␤»
timotimo m: say "heyo".*vain
camelia rakudo-moar 8be36b: OUTPUT«()␤»
timotimo m: say "heyo".?vain
camelia rakudo-moar 8be36b: OUTPUT«Nil␤»
mst yep, but that's not what I care about
perlpilot timotimo: for completion you should mention that .?foo also exists
timotimo perlpilot: i just did :)
masak .oO( it is vain to correct someone's spelling ) :P
mst Juerd: nope
perlpilot yeah, I'm too slow typing this morning :) 14:42
mst and that means
timotimo perlpilot: don't worry, i'm just very fast
mst sub postfix:<.foo> ($foo) { sub ($bar) { $bar ~ ' ' ~ $foo.f() } }
$f.foo(3)
works
which means
I can potentially do C# style extension methods
EXCELLENT
lexical augmentation of available methods on an object, transparently
timotimo they won't show up in things like .^methods or other introspection, though 14:42
masak mst: you kind of already get that through .&foo, though 14:43
mst I can live with that
masak: the & OFFENDS ME
masak a-ha
timotimo OK
Juerd Where's .& documented?
masak Juerd: it's an emergent feature :P
it falls out of being able to use variables in place of a literal method name 14:44
Juerd Still :)
mst masak: also, AFAICS, with appropriate matching, I can, for things that already have a .foo, fall through to that
timotimo watch out if you spell it like that, masak
mst: indeed
mst and that's really interesting
since IIRC '$x but Foo' copies $x whereas this would work in-place on $x 14:45
timotimo yeah, and "does" will mutate the original object 14:45
mst right 14:46
and I want augment-if-necessary-within-my-current-scope
basically
there's uses for that
dalek c: c4a4116 | gfldex++ | doc/Language/operators.pod6:
remove superstitiuos parentheses from example
 14:47
vcv Semi-related, is there a way to call variants of a multi that match instead of just the "best" match?
vcv call all variants* 14:47
timotimo that's what .* does 14:48
moritz .* (for method calls)
timotimo it'll still only call those that match the passed arguments
vcv right
I was just trying.. I must have been doing it wrong then :(
timotimo oh
if you have just the multi, you can use something like .can or .cando or something 14:49
i never needed that
masak m: class C { multi method foo($x) { say "x" }; multi method foo($y) { say "y" } }; C.new.*foo(42) 14:52
camelia rakudo-moar 8be36b: OUTPUT«Ambiguous call to 'foo'; these signatures all match:␤:(C $: $x, *%_)␤:(C $: $y, *%_)␤ in block <unit> at <tmp> line 1␤␤»
jnthn .* is only about the inheritance tree, not about multis
masak ...what jnthn said :)
jnthn .cando can be used to find all matching multis
masak I don't remember if this changed somewhere along the line, or if it was always just the mro chain 14:53
jnthn It changes years and years ago, when proto methods started to exist :)
*changed
jnthn So ~2010 14:54
Maybe 2011
masak ah.
jnthn It was debated a few times, but the final ruling in the run-up to 6.c was that it stays the way it is. 14:55
jnthn .cando can be used to get all multi candidates matching a Capture 14:57
vcv fffffff~./~. 15:03
right~.
timotimo die, ssh connection, die!
moritz RETURN TILDA DOT 15:04
poohman Hello all, dont we have Str.len or Str.length?
timotimo m: say "foobar".length 15:04
camelia rakudo-moar 8be36b: OUTPUT«Method 'length' not found for invocant of class 'Str'␤ in block <unit> at <tmp> line 1␤␤»
moritz poohman: nope; you must ask for a specific unit
timotimo i thought we used to have a nice error message for that
moritz poohman: like .chars or .codes (or .bytes on buffers) 15:05
poohman oh ok
poohman perfect thanks 15:06
mst poohman: means you're never confused about "which length" 15:07
timotimo but how do i figure out how long a string would be if i asked javascript about it? :( 15:08
mst roll a d6, assume that's as accurate as anything else
timotimo i guess i'd just have to go through the string codepoint-per-codepoint and if the codepoint is on the astral plane (or something?) count it as two characters 15:09
jkramer Is there a nice and safe way to generate a bunch of methods in a role or class? Not necessarily monkey-patching at runtime, I'm just looking to safe myself from some typing 15:14
perlpilot EVAL? :>
jkramer I.e. curryently I have "multi method x(*@) { all-do-the-same-thing() }" ~10 times with different names instead of x 15:15
perlpilot docs.perl6.org/routine/add_method 15:15
moritz jkramer: look for add_method and compose in the docs
jkramer And I'd like to do "for <lots of method names> -> $name { make multi method $name(*@) { ... } }
perlpilot see also docs.perl6.org/language/mop 15:16
moritz for @names ->name { ::?CLASS.^add_method($name, method (*@) { ...}) };
pmurias timotimo: wouldn't you need to go through codepoint-per-codepoint (if it's on the basic plane i++ and if on the astral/supplementary plane i += 2) 15:17
timotimo pmurias: yeah, that's what i meant?
jkramer moritz: That looks like what I want, but doesn't seem to work with multi methods 15:18
timotimo i believe you can add_multi_method or something? 15:19
jkramer timotimo: I think it's because it's an anonymous method in the call to add_method 15:20
timotimo multi methods absolutely require you to compose something
pmurias timotimo: ok, I have looked at generated javascripted too much today ;)
jkramer "An anonymous method may not take a multi declarator"
timotimo yeah 15:21
jkramer Ah no it works :)
timotimo we don't have a way to associate multiple multi candidates into the same "thingie" if they don't have a name
jkramer Ok, got it :) add_multi_method but with "method $name" instead of "multi method $name"
pmurias timotimo: str.replace(/([\uD800-\uDBFF][\uDC00-\uDFFFF]|[\S\s])/g, "_").length 15:23
timotimo _ rather than __? 15:24
perlawhirl timotimo: shell(q{node -e 'console.log("ñoñó".length)'}, :out).out.lines 15:26
yoleaux 19 Sep 2016 07:57Z <nine> perlawhirl: note that just installing an older version when a newer one was already installed, won't make rakudo use that old version. If no :ver is specified in the use statement, we pick the newest version.
perlawhirl ducks
timotimo :)
skaji Hi, 15:29
IO::Socket::Async used to read data all at once, if data is small. 15:30
But now IO::Socket::Async always does not read data all at once.
Is this expected?
Please look at gist.github.com/skaji/664077e99b80...d1a54a0910
timotimo if you read non-binary, you'll have to wait for a control character to get the very last character 15:31
because the next bytes you receive could introduce a composing character
and if you already got the data via string, it'll be too late for perl6 to give you the right data, it might already have "escaped" 15:32
bioduds hey guys, I installed MongoDB but it is not working
how can I check if install was ok?
timotimo bioduds: what are the symptoms?
bioduds when I run the program
it says
😀 bem vinda Já conhece o trabalho do BUKKAKEROS?
sorry
bioduds wrong copy paste 15:32
here 15:33
===SORRY!=== Could not find MongoDB::Client at line 7 in:
could not find?
but I installed
im confused
:)
timotimo it could be the module doesn't have a provides section?
nine ^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H/win 13
timotimo did the module run its tests successfully?
bioduds I dont know
how can I check?
timotimo the module does have a provides section that looks correct 15:34
just try to install it again, maybe put --force after "panda" or "zef"
skaji timotimo: Oh, composing character, I see.
bioduds ok
timotimo skaji: we're in the process of exposing encoding and such better to the user 15:35
skaji: until then you might want to just set latin1 as the encoding or use binary blobs and do your own decoding
nine Considering that --force only helps if the very same version is already installed, I would only suggest it really sparingly and only if I knew it would actually solve something.
timotimo right 15:36
bioduds re-installing
timotimo that makes sense
nine Also the precomp repository now puts more trust in the immutability of distributions and I'm not sure we actually outdate precomp files when a distro gets overwritten with a modified version. 15:37
Well modified code but same version to be precise.
skaji timotimo: OK, I will use Supply(:bin). Thanks! 15:38
timotimo good luck! 15:39
skaji: in the future we'll be properly able to change the encoding in the middle of a connection, which will be good for HTTP
mst m: sub roll { for %_.kv -> $k, $v { my ($d) = $k.comb(rx{<[0..9]>+}); say &CORE::roll(0+$v, 1..$d).join(", ") } }; roll :6d6 15:43
camelia rakudo-moar 8be36b: OUTPUT«1, 6, 4, 1, 4, 2␤»
mst m: sub roll { for %_.kv -> $k, $v { my ($d) = $k.comb(rx{<[0..9]>+}); say &CORE::roll(0+$v, 1..$d).join(", ") } }; roll :d12
camelia rakudo-moar 8be36b: OUTPUT«11␤»
mst :D
timotimo that's cute 15:44
does it fit into a tweet? looks a tad long for that 15:45
mst wc says 142
so
and that includes the 'perl6 -e' bit
timotimo it'd want some description of what it does, too 15:47
m: say "foobar" ~~ /<digit>+/ 15:48
camelia rakudo-moar 8be36b: OUTPUT«Nil␤»
timotimo m: say "9d10" ~~ /<digit>+/
camelia rakudo-moar 8be36b: OUTPUT«｢9｣␤ digit => ｢9｣␤»
timotimo m: say "<[0..9]>".chars; say "<digit>".chars
camelia rakudo-moar 8be36b: OUTPUT«8␤7␤»
timotimo but <digit> may include some weird unicode ones, too, if you don't like that
.u die face 15:50
yoleaux U+2680 DIE FACE-1 [So] (⚀)
U+2681 DIE FACE-2 [So] (⚁)
U+2682 DIE FACE-3 [So] (⚂)
timotimo m: say (:⚀d⚂)
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Bogus statement␤at <tmp>:1␤------> 3say (:7⏏5⚀d⚂)␤ expecting any of:␤ colon pair␤»
timotimo OK, So doesn't go there
m: say "⚂".&uniprop 15:51
camelia rakudo-moar 8be36b: OUTPUT«So␤»
timotimo committable6: releases say "⚂".&uniprop
committable6 timotimo, ¦«2015.10,2015.11,2015.12,2016.02,2016.03,2016.04,2016.05,2016.06,2016.07.1,2016.08.1,HEAD»: So
timotimo oh, i thought it might have something numeric about it 15:52
have had*
mst timotimo: yes, that's exactly why I didn't use <digit> 15:53
timotimo understood
m: my $k = "d20"; my ($d) = $k ~~ /<[0..9]>+/; say $d 15:54
camelia rakudo-moar 8be36b: OUTPUT«｢20｣␤»
timotimo this ought to be shorter than using .comb and the parens
BBL
pmurias [Coke]: I fixed t/docs/opcodes.t 16:00
[Coke] Thanks. Hopefully you didn't have to mangle your code. 16:15
jkramer Is the precedence of multi methods with the same signature defined (and documented somewhere)? 16:29
perlpilot "multi methods with the same signature"? 16:30
timotimo if you have two multi methods with the same signature, you should get a compile-time error
jkramer They're not 100% the same, but technically they both match the same stuff. One is (*@) and one is (*@foo) 16:31
timotimo that's the same signature and ought to give you an error 16:32
jkramer It seems like (*@foo) is preferred (which is what I want) but I'd feel better if that was defined behavior :)
Well it doesn't
timotimo are you using add_method or add_multi_method?
jkramer Yup :) How did you know ;)
timotimo it's probably the grammar (or rather the actions) that cause that particular error 16:33
jkramer Ah 16:33
timotimo are you composing the class after you're done adding stuff?
jkramer The add_multi_method happens in a role which is then applied to a class. The other confliciting method is defined in yet another role which is applied to the same class, if that matters 16:34
However the other method isn't added with add_multi_method
timotimo are you adding a method with add_method and then a method with the same name with add_multi_method?
jkramer No, only with add_multi_method and in the other role the other method is defined as usual with "multi method foo..." 16:35
timotimo mhm, mhm
i'd do some prodding with .^methods and .candidates on the methods to see what's really going on
jkramer Here's a dump of the ^methods and their .candidates dpaste.com/3VY1DVY 16:41
The method I'm concerned about is "get"
post, put, patch, head and delete are the other ones added with add_multi_method
So the candidates for get() are one proper method and one anon 16:42
timotimo ah 16:44
well, you've got different signatures there anyway :)
so it shouldn't give you an error
and the narrowness is clear for that 16:45
jkramer Because of the Mu vs Image?
timotimo yes
jkramer Hmm, interesting. So the Image methods have precendence because the others are in Mu which is higher up in the inheritance thing? 16:46
timotimo the part before the : is the type constraint on the "self" (aka the invocant) 16:47
a narrower type constraint gets preferred
jkramer And the Mu comes from the add_multi_method thing I guess? So if I define it as multi method get(*@) in that role it should crash? 16:49
Oh yeah it does.
Nice, so the add_multi_method is actually helping me quite a bit here :)
vcv What's the shortest way to print a new line to stdout? say ''; anything shorter? 17:53
mst sub nl () { say '' }
nl;
TimToady m: say() 17:54
camelia rakudo-moar 8be36b: OUTPUT«␤»
vcv ah. i wrongly assumed say() was the same as just say 17:55
harmil_wk Should FIRST now be defined inside a loop? I get:
yoleaux 02:14Z <MasteDuke> harmil_wk: ⒗ won't work by design for a bunch of stuff. see github.com/rakudo/rakudo/commit/e2...6d25995a90 for some discussion. it has a .uniprop of 'No' instead of 'Nd'
harmil_wk m: sub sd ($d) { $d.DateTime.local }; for ^3 { say "Time: {sd BEGIN now}, {sd INIT now}, {sd FIRST now}, {sd ENTER now}, {sd now}"; } 17:56
camelia rakudo-moar 8be36b: OUTPUT«Use of Nil in string context␤ in block at <tmp> line 1␤Time: 2016-09-20T19:56:11.317111+02:00, 2016-09-20T19:56:11.358320+02:00, , 2016-09-20T19:56:11.373095+02:00, 2016-09-20T19:56:11.375848+02:00␤Use of Nil in string context␤ in block at <tmp…»
psch harmil_wk: it's 'sub FIRST $val' vs 'FIRST sub $val' what gives you the warning 17:58
m: for ^1 { FIRST say "hi"; };
camelia rakudo-moar 8be36b: OUTPUT«hi␤»
psch m: for ^2 { FIRST say "hi"; };
camelia rakudo-moar 8be36b: OUTPUT«hi␤»
psch m: for ^2 { say FIRST "hi"; };
camelia rakudo-moar 8be36b: OUTPUT«WARNINGS for <tmp>:␤Useless use of constant string "hi" in sink context (line 1)␤Nil␤Nil␤»
psch harmil_wk: ...plus something about FIRST apparently not returning a value..?
harmil_wk psch: That can't be. It still happened before I added sd
psch (note the second example has two Nils)
harmil_wk: right, that's what i just noticed myself. looks wrongish from here, but i'm a bit out of practice. better wait for someone else to chime in :) 17:59
TimToady according to S04:1391 FIRST is supposed to return a value, but I believe it is NYI 18:02
synopsebot6 Link: design.perl6.org/S04.html#1391_FIRS...rn_a_value
TimToady say wot?
psch bad heuristics, probably 18:03
timotimo whoops :)
psch i remember having fiddled with that once, probably waterbedded something :S
timotimo it's very easy to waterbed this particular piece of software 18:04
TimToady
.oO(waterbedding as a form of torture...)
harmil_wk Note that the example above was based on seeing INIT in the example at rosettacode.org/wiki/Handle_a_signal#Perl_6 18:22
I assumed it was generalizable to all phasers. 18:23
TimToady notice the asterisks in the table at S04:1380 18:25
synopsebot6 Link: design.perl6.org/S04.html#line_1380
poohman m: say "Hello testing" 18:26
camelia rakudo-moar 8be36b: OUTPUT«Hello testing␤»
TimToady phasers without an asterisk cannot be used for their value because that would involve time travel 18:27
which is one of the few things we haven't solved in Perl 6...
but note that has a speculated asterisk, indicating it should work 18:28
s/that/that FIRST/
speaking of time travel...
AlexDaniel synopsebot6: source 18:30
meh 18:31
SourceBaby: source
SourceBaby AlexDaniel, See: github.com/zoffixznet/perl6-sourceable
AlexDaniel SourceBaby: good bot
SmokeMachine____ Hi! can anyone help me? I am having a problem but I couldn't find how to simplify it to post here... :( but thats it: the error is intermittent: if I change a class to add a new attribute, it gives a error on another class... 18:34
www.irccloud.com/pastebin/bwRXBjJO/Error 18:35
the original code is in: github.com/FCO/ProblemSolver 18:36
psch would guess lingering precomp files 18:37
SmokeMachine____ psych: if I remove the .precomp dir, it gives error from the first time... www.irccloud.com/pastebin/iqGbQi4Y/ 18:38
psch SmokeMachine____: alright, i probably guessed wrong then, sorry :) 18:39
TimToady how is add-heuristic declared? 18:40
SmokeMachine____ psch: thanks for trying! :)
TimToady: its handled by a array attribute of Problem
TimToady I mean, what's the signature? 18:41
TimToady too lazy to download it all to grep...
SmokeMachine____ TimToady: no, sorry... its a normal method!
TimToady with what signature?
SmokeMachine____ method add-heuristic($var, &heu) {
and: has Array of Callable%!heuristics; 18:42
TimToady okay, so something inside that method
SmokeMachine____ the only line of the method: %!heuristics{$var}.push: &heu
TimToady huh, it's like it's expecting Array[Callable]:U or so 18:43
SmokeMachine____ TimToady: 👆
TimToady well, it's apparently not autoviving for you somehow, seems 18:44
SmokeMachine____ this is the method: github.com/FCO/ProblemSolver/blob/...em.pm6#L11
TimToady try making sure the element exists first?
timotimo oh, autovivification sometimes just doesn't work if you declare a constraint
timotimo hasn't investigated yet
SmokeMachine____ but it works if it does not exists a Array of Callable on another class... :( 18:45
TimToady might be a () vs some undef value thing 18:46
the attribute is probably forcing a .new at some point on the value that is complaining it's a .new, but that doesn't explain why the assignment wouldn't work 18:47
anyway, it does look a little buggy to me
SmokeMachine____ the problem occours when I add this line: github.com/FCO/ProblemSolver/blob/...ate.pm6#L7 18:48
TimToady yes, which is an array of callables, which is probably getting Array[Callable].new somewhere
which is not assigning right inside the push 18:49
SmokeMachine____ but its in another class...
and no-one uses that array...
TimToady what happens if, before the push, you say %!heuristics{$var} //= Array[Callable].new;
SmokeMachine____ let me try it... 18:50
TimToady this will rule out an autoviv problem
CIAvash the error message reminds me of this bug: rt.perl.org/Public/Bug/Display.html?id=127309 18:51
SmokeMachine____ TimToad: the same www.irccloud.com/pastebin/LSSo6eIY/ 18:52
TimToady wow CIAvash++
psch #127309 looks like some kind of parameterization caching problem?
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=127309
psch anyway, yeah, spooky :) 18:53
TimToady so praps a noan bugg
SmokeMachine____ This is Rakudo version 2016.09-19-g8be36b1 built on MoarVM version 2016.09 implementing Perl 6.c.
psch (wow that phonetic transcription took me a while to pull apart, TimToady++) 18:54
TimToady spellin' is my speshialitee 18:55
lunch & 18:56
broquaint If I want to run perl6 on the JVM is my best option to build from source or is it part of Rakudo Star (or other)? 20:28
konobi broquaint: ask mst 20:29
konobi ducks
timotimo huh?
what does mst have to do with rakudo-jvm? 20:30
broquaint: at the moment, rakudo-j has a few problems
geekosaur build from source, but note that jvm is known broken
broquaint Aha, thanks for the heads up :) 20:31
geekosaur there's ongoing work but it is prone to crash on things rakudo-m works on still 20:32
stmuk rakudo-j works well enough for simple short code examples but panda etc doesnt work 20:37
tailgate how do I check if an object has an attribute? 20:40
dogbert17 o/ #perl6
jnthn m: class C { has $.x }; say so C.^lookup('x'); say so C.^lookup('y') 20:43
camelia rakudo-moar 8be36b: OUTPUT«True␤False␤»
dogbert17 m: my @a = [1,[2,3],4]; dd @a.duckmap({ $_ ~~ Int ?? $_ !! Any }) # this works well
camelia rakudo-moar 8be36b: OUTPUT«(1, (2, 3), 4)␤»
dogbert17 m: my @a = [1,[2,3],'a']; dd @a.duckmap({ $_ ~~ Int ?? $_ !! Any }) # this not so much
camelia rakudo-moar 8be36b: OUTPUT«Memory allocation failed; could not allocate 6296072 bytes␤» 20:44
timotimo that somehow infinitely recurses
dogbert17 timotimo: yeah, it should work I guess 20:45
timotimo yeah, i'd say it should
dogbert17 guess I'll rt then 20:46
dogbert17 done, rt.perl.org/Public/Bug/Display.html?id=129321 20:54
tushar I have a private method that I am calling in each public method. Is there any way I can call private method after the object creation automatically? 20:55
timotimo you can do that in BUILD for example 20:56
moritz or BUILDALL
docs.perl6.org/language/objects#Ob...nstruction
tushar timotimo, I am not writing any custom BUILD method rather using the default BUILD method 20:57
tushar timotimo, moritz, thanks. I will explore it. 20:59
tushar what is "|" mean in BUILDALL(|)? Does it indicate Junction operator? 21:02
vcv | is a capture which gobbles up all remaining positional and named arguments 21:06
if those arguments arent going to be used, you dont have to name it, thus | instead of something like |c
it also means BUILDALL(|) can accept any number of parameters of any type 21:07
tushar vcv: thanks
gfldex tushar: see docs.perl6.org/type/Signature 21:11
tushar :gfldex thanks. 21:13
vcv Yeah, the Capture Parameters section. I don't see anything about nameless parameters though.. maybe it's there under a different name/term 21:14
Woodi so, here are some stats for Rakudo releases: gist.github.com/slunski/0c79c21557...70381e8bed 21:59
Woodi all that experiment tought me that Perl6 OO is quite stable since few years :) 22:00
El_Che no regressions lately :) 22:12
timotimo no ragrets 22:15
wiki.gnome.org/Outreachy/2016/Dece...anizations - it seems like Perl is already participating in outreachy; that's very cool 22:17
could applicants potentially work on perl6 stuff?
oh, the description reads TPF will sponsor an applicant to work on bugzilla 22:18
dylanwh That is because that's what I can reasonably make time to mentor for. 22:31
timotimo ah 22:32
thank you for mentoring, in any case! :)
dylanwh hopefully brings a new person to perl, and maybe a new person to perl6 afterwards. :-) 22:34
timotimo that'd be cool 22:36
dylanwh btw, I countered the person at mozilla that thought the whatever-star operator / fibonacci example was "line noise" with the CSS 2.1 parser, and the person was actually impressed by that grammar.
dylanwh so a small victory in mind share, maybe. 22:37
timotimo neat :) 22:38
the ... operator is fantastic 22:40
i used it to great effect to follow a node's "parent" pointer from the target to the source in a graph that had a dijkstra path finding algorithm run over it
dylanwh I think there is literally something in people's mental model of programming that either makes them love symbols or hate symbols, with no inbetween. 22:41
timotimo and the result of the ... was the list of nodes along the path
dylanwh it might have to do with exposure to maths notation
timotimo a friend of mine was astounded to see me switch from python to perl6
dataangel My exposure to maths notation made me hate maths notation
timotimo what with python having a dearth of symbols and perl6 having the opposite :)
dataangel In math pretty much no symbol has a consistent meaning 22:42
Instead of an easy-to-read name you get a Greek letter that tells you nothing without going back and referring to an earlier part of the paper
geekosaur and even when there are consistent names, they're only consistent within one branch 22:45
timotimo yeah 22:47
at least in perl6 symbols tend to keep their meaning and different things look different 22:49
dataangel Yeah math notation never gets executed so they never discover that they're being inconsistent
timotimo ipython notebooks (and by extension julia notebooks and some others) have made a nice dent in that issue 22:51
at one point there was a talk or article about "you can basically click one button to fire up an AWS instance that'll reproduce our results in a couple of hours or days"
lambd0x Hi everyone! 22:59
timotimo hello lambd0x 23:05
lambd0x timotimo: :D
lambd0x timotimo: do you know what does @array.grep uses internally for the search? I've just done some testing of finding an element on an array of ordered elems. sizing from 1,000 up to 15,000 using grep and compared it against a simple binary search and the second was significant better in time. tinyurl.com/zgc2lcq (codes), tinyurl.com/hrlx4v7 (results) 23:18
timotimo well, grep will definitely not use binary search, because it's not guaranteed to have a sorted list to search through 23:20
lambd0x timotimo: could we optimize grep by narrowing it down to cases? Such as for ascending, descending, etc...? 23:23
vcv dylanwh: do you have a link to the CSS parser? 23:26
timotimo definitely not 23:27
grep just goes through the list. if you want to figure out if the list is sorted, you also have to go through the whole list
so finding out if the list is sorted or not is about as expensive as doing the grep itself
lambd0x timotimo: I see. thanks for explaining :). 23:29
vcv You could, in theory, implement a SortedList object and write your own grep method for that
timotimo also, sorting a list is eager, but perl6 gives you a lazy grep 23:33
lambd0x vcv: I was thinking about that, but what timotimo said is true. Normally u wouldn't know if the array is or not already sorted and so, how to tell it otherwise than just checking it which will take its share of the time alonside with grep. 23:34
timotimo if you already know it's sorted coming in, or you've done the work up front, use a module from the ecosystem that gives you binary search or something similar :)
vcv i suppose binary search and grep dont fit together at all anyway -- maybe first, if given a known sorted list 23:36
actually that might not make sense either 23:37
lambd0x timotimo: the difference between eager and lazy lists is that the second just resolves the element when accessed?
timotimo that's right 23:38
lambd0x vcv: exactly :/
timotimo: thanks :)
timotimo grep gives you a list of items that match a predicate
if you have a predicate like "is exactly 9", you can easily binary-search, because it gives you "greater than" or "less than" results
but if your predicate is like "is a prime number" or "this field has a stone placed on it" or things like that ... not so much 23:39
lambd0x timotimo: that's interesting 23:40
lambd0x oh another thing, if I wanted to use hashes with a different hashing function would that be let's say using OOP, or not even with such? 23:43
timotimo i can't parse that sentence 23:44
lambd0x sorry. Is possible in perl6 to use a different hashing function for hashes?
timotimo the hashing function used for built-in hashes is an implementation detail 23:45
but you can implement your own hashes using arrays and your own hashing function
it'll also be kinda slow :)
vcv s: help
SourceBaby vcv, Something's wrong: ␤ERR: ===SORRY!=== Error while compiling -e␤Undeclared routine:␤ help used at line 6␤␤
lambd0x I'm literally wanting to reinvent the road, yes I know ahahha
vcv s help
lambd0x timotimo: just for learning I think its worth :) 23:46
thanks.
vcv s: Hash, 'keys'
timotimo i'm going to bed, gnite!
SourceBaby vcv, Something's wrong: ␤ERR: Type check failed in binding to &code; expected Callable but got Method+{<anon|74133312>} (Method+{<anon|7413331...)␤ in sub do-sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 42␤ in sub sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 33␤ in block <unit> at -e line 6␤
lambd0x timotimo: o/ 23:47
vcv s: Hash
SourceBaby vcv, Something's wrong: ␤ERR: Cannot resolve caller sourcery(Hash); none of these signatures match:␤ ($thing, Str:D $method, Capture $c)␤ ($thing, Str:D $method)␤ (&code)␤ (&code, Capture $c)␤ in block <unit> at -e line 6␤␤
vcv s: 'abs'
SourceBaby vcv, Something's wrong: ␤ERR: Cannot resolve caller sourcery(Str); none of these signatures match:␤ ($thing, Str:D $method, Capture $c)␤ ($thing, Str:D $method)␤ (&code)␤ (&code, Capture $c)␤ in block <unit> at -e line 6␤␤
timotimo s: &abs
SourceBaby timotimo, Sauce is at github.com/rakudo/rakudo/blob/8be3...ric.pm#L46
vcv thanks, i can never remember how to operate this bot
vcv lambd0x: github.com/rakudo/rakudo/blob/8be3...re/Hash.pm Perl 6 Hash object, but its heavily using underlying nqp ops which makes it much faster than you could with pure Perl 6 23:48
lambd0x vcv: thanks u very much \o/ 23:51

Powered by Cro and the Raku® Programming Language.

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

Thank you!