»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.
lelf smth more end-user-friendly? the problem is that ‘sub foo is native’ is too lazy with regard to dyn library bindings 00:01
timotimo if it's okay to be unix-specific, you can try binding dlopen and friends with nativecall 00:02
lookatme o| 00:40
TimToady
.oO(a grim cyclops?)
 00:41
lookatme o_O 00:42
TimToady
.oO(the top of a movie projector?)
 00:44
lookatme (°ー°〃) 00:48
TimToady
.oO(an upside-down skateboard flying out of a dustcloud?)
 00:51
gfldex
.oO(What is the hippie smoking…again?)
 00:52
.oO(And can I have some too?)
 00:53
lookatme (＠_＠;)
TimToady
.oO(a winking snail towing another snail out of a tunnel?)
 00:56
lookatme (⊙﹏⊙) 00:57
TimToady
.oO(hard to shoot at the target when I drank too much...)
 00:58
lookatme back to work 01:00
TimToady o|
Geth ecosystem: a02d987859 | (Zoffix Znet)++ (committed using GitHub Web editor) | META.list
Add CoreHacker::Q to ecosystem

  "Perl 6 QAST Visualizer": github.com/zoffixznet/q		 01:13
Zoffix .tell lizmat you had some interest in the module. The first baby version got released as CoreHackers::Q: github.com/zoffixznet/q 01:14
yoleaux Zoffix: I'll pass your message to lizmat.
Todd_ Ji All. I am stumped. I just installed rakudo-pkg on a Fedora 27 server (removed all the zef's and rakudo's first) and I can't get anything to zef properly. All packages have the same issue. " --force-test" is required on all of them to complete. When I go to "use xxx", you get an error messages as to it can;t find one of the long letter directories in the /opt path (it really isn't theWould someone mind looking at my vpaste and tel 03:34
" --force-test" is required on all of them to complete. If I force, when I go to "use xxx", you get an error messages as to it can't find one of the long letter directories in the /opt path (it really isn't there). Would someone mind looking at my vpaste and telling me what I am doing wrong?
vpaste.net/MnJEg
sorry for the weird copy and paste. 03:35
Zoffix Todd_: do you have ~/.zef or ~/.perl6 directory by any chance? 03:38
I may be mistaken, but it's possible zef is installing the modules user-local, but the path when you `use` them doesn't include the path where it's installing them or something.
Todd_ yes and yes
Zoffix Todd_: and when you `use` the module? Does the error message list ~/.perl6 in the list? 03:39
Todd_ # eraseme.pl6 ===SORRY!=== Could not find Terminal::ANSIColor at line 9 in: /home/linuxutil /root/.perl6 /opt/rakudo-pkg/share/perl6/site /opt/rakudo-pkg/share/perl6/vendor /opt/rakudo-pkg/share/perl6 CompUnit::Repository::AbsolutePath<54193536> CompUnit::Repository::NQP<30451248> CompUnit::Repository::Perl5<30451288> 03:40
Zoffix Todd_: are you root? 03:41
Todd_ yes and I just realized that and renamed /root/.zef and /root/.perl6 03:42
lookatme Is it lake some test tool ? or something else
like prove
Todd_ rerunning "zef install Terminal::ANSIColor" gives the same error
lookatme I faced same problem in fresh installation of Fedora 27 03:43
Zoffix Todd_: renamed to what?... 03:44
Todd_ mv ~/.zef ~/.zef.000 mv ~/.perl6 ~/.perl6.000
both the root and users 03:45
Zoffix No idea
Todd_ I am doing a "dnf --enablerepo=* upgrade" to see if that helps 03:46
lookatme Todd_, you can try it with --debug 03:49
Todd_ the zef command?
lookatme yeah, it print the debug message
Todd_ in progress 03:50
vpaste.net/21U9f 03:51
Is this a hint: 03:53
ListUtil.c: loadable library and perl binaries are mismatched
where are these loadable libraries?
I think I am on to something. Don't anybody help me until I get back 03:54
rats. renaming .precomp did not help 03:55
lookatme idk
no idea :)
geekosaur that message is a *perl 5* mismatch 03:56
geekosaur you have a perl 5 XS module that doesn't match the perl 5 executable (or shared library, depending on how perl was configured0 03:57
something tells me your F27 is much more hosed than just rakudo. or you have a perl 5 you copied over that needed to not be copied. 03:58
lookatme so some perl5 script will failed ?
Todd_ ahhhh. now it is starting to make sense. I will try wiping p5 and reinstalling it. I presume zef is using p5
lookatme And I think the zef maybe use prove(a perl5 script for test) to test something
geekosaur I think it uses Inline::Perl5 for some things
probably have to ask nine
Todd_ I have had the WORST luck with Inline::Perl5. Probpably a good thing as it forced me to dive into p6 harder 03:59
lookatme And I am wonder why you copy the Perl5 or Rakudo and other things ?
The correct way is reinstall them, not copy 04:00
geekosaur and/or ugexe 04:01
Todd_ I am using `# dnf remove perl` and the corresponding install. ther is a reinstall, but I wanted to hunt down any leftovers first 04:02
removed and reinstalled p5. Still get: loadable library and perl binaries are mismatched 04:04
Todd_ update: the perl 5 chat guys are helping me fix battling cpan and installed binaries, which is causing the issue 04:35
wish list: please update zef to use perl6 instead of perl5 to do its magic 04:38
okay: # perl -v bash: /usr/bin/perl: No such file or directory 04:43
Todd_ I useed find to hunt down any leftovers and now I will reinstall 04:46
# dnf install perl perl-interpreter this will be a while! 04:47
Todd_ Yippee!!!! This command show the offender! # perl -le 'for (@INC) { print $_ if -f "$_/DBI.pm" }' 05:09
`mv /home/todd/perl5 /home/todd/perl5.000` fix the issue
Todd_ "if" Fedora had our moduels loaded up, would this be the right way to translate their names into the repo names (word in Perl 5): 05:20
# dnf repoquery --whatprovides 'rakudo(DateTime::Format)'
Todd_ I am having trouble installing modules on Fedora 26. Would someone please look at my vpaste and see what I am doing wrong? vpaste.net/x04uk 06:29
All moduesl give the same error 06:30
El_Che Todd_: it looks like you're using system zef instead the one of the package 07:07
/opt/rakudo-pkg/bin/zef (as root) or /opt/rakudo-pkg/bin/install_zef_as_user if you want them in your $HOME
(the latter is the beter option for dev stuff) 07:08
Todd_: rakudo-pkg does not touch system path to no interfere with system rakudo (if present). You need to look at the PATH section in the doc to prevent this type of error 07:10
in short "export PATH=/opt/rakudo-pkg/bin:$PATH" will do (and adding ~/.perl6/bin if you install locally) 07:11
Todd_ I uninstalled the rakudo package and reinstalled Fedroa's and all is well. Go figure .... 07:14
El_Che it's not related 07:16
you are using /usr/bin/zef
while the package provides /opt/rakudo-pkg/bin/zef
I decided not to change the user $PATH when I created the packages, maybe I need to review that, but not interfearing with system rakudo is a good thing, imho 07:18
even when not adapting the path can lead to weird (but safe) results 07:19
Todd_ I uninstalled rakudo and rakudo-ref. Then did a global find and aced ANYTHING that had the letters "zef" in it. I then installed the rakudo-pkg rpm. That is when I got into trouble. I did this same thing an fe hours earlier on a Fedora 27 server and it went perfectly, excewpt for the Perl 5 module mismatches that blew zef's mind. 07:21
I have this Fedora 26 machien scheduels for upgrade to 27 tomorrow evening, so I may have to go through this again. 07:22
El_Che well, having to manually delete binaries is a bad thing is they were installed from packages 07:23
for the record, rakudo-okg is self contained in /opt/rakudo-pkg and modules installed locally are in ~/.perl6
nothing is written elsewhere 07:24
Todd_ It is going to be a fun night. Thank you all for the wonderful help! 07:48
Geth doc: 7a0bb28bdc | (Luca Ferrari)++ | doc/Type/Any.pod6
Improve &exit documentation.

Specify that exit status code has to be interpreted by calling process, and that LEAVE is prevented when using exit. Remark that &exit should be used to terminate abnormally only the main process, while exceptions should be used for other cases (e.g., methods).
This does not remove the TODO item attached, but I believe there's not a better place to document exit since it is already documented with the usage of MAIN and the phasers.		 08:45
synopsebot Link: doc.perl6.org/type/Any
doc: fed435e1b2 | (Luca Ferrari)++ | doc/Language/functions.pod6
Explain that 'return' in MAIN will produce always an exit code of zero.
synopsebot Link: doc.perl6.org/language/functions
torbjorn the token/rule names found in key names in the resulting Match object - is there magic involved in determining whta class they get? 09:13
ie token foo { [foo]* } # would that result in $/<foo> being an Array because of the * ? 09:14
moritz [foo] doesn't capture 09:18
(foo)* does, and yes, $<foo> would be an array
the only exception is the ? quantifier, which produces either a Match or a Nil capture
torbjorn ah yes, () of course 09:33
so token foo { foo } # woueld not be an array? 09:34
moritz that doesn't even capture
well, $<foo> in the caller caputres 09:35
lelf How do you do “:(Foo[::T] \a, Foo[T] \b)”? Akin to “:(::T \a, T \b)”. [Foo is defined] 09:36
moritz m: my token foo { foo }; 'foo' ~~ /<foo>/ and say $<foo>.^name
camelia Match
moritz m: my token foo { foo }; 'foo' ~~ /<foo>+/ and say $<foo>.^name
camelia Array
Geth doc: e3edec2d40 | (Luca Ferrari)++ | doc/Language/functions.pod6
Small rephrasing about coercion.

Also define the example class Bar at the beginning of the example, there is no point here to leave it abstract.		 09:45
synopsebot Link: doc.perl6.org/language/functions
dakkar hmm. re: coercion, I feel it mixes "target type" and "method used to get it". I can't say «sub x(IO::Path() $d) { ... }», even though «sub x(IO() $d) { ... }» will actually coerce to IO::Path 09:54
m: sub x(IO::Path()$d) { $d.^name.say}; x('a') 09:55
camelia No such method 'IO::Path' for invocant of type 'Str'
in sub x at <tmp> line 1
in block <unit> at <tmp> line 1
dakkar m: sub x(IO()$d) { $d.^name.say}; x('a')
camelia IO::Path
travis-ci Doc build errored. Luca Ferrari 'Small rephrasing about coercion. 10:00
travis-ci.org/perl6/doc/builds/327159440 github.com/perl6/doc/compare/fed43...edec2d409b
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 10:00
Zoffix dakkar: it's just a convention that a method named "Foo" converts an object to type "Foo". The coersion merely calls a method with .^name of target type. Str does not have "IO::Path" method so that coersion rightly fails. IO::Path does role IO tho and Str type has IO method that returns an IO object (IO::Path in this case). 10:59
But nothing's stopping you from defining a method with IO::Path as a name and using IO::Path as target type in coercer 11:00
m: use MONKEY; augment class Str { method IO::Path { self.IO } }; -> IO::Path(Str) $x { dd $x }("foo")
camelia IO::Path.new("foo", :SPEC(IO::Spec::Unix), :CWD("/home/camelia"))
Zoffix FWIW, there are also two paths in coersion: the slow path will try to see if there's a method and complain if there isn't: 11:04
m: try '-> IO::Path(Str) :foo(:bar(:$ber)) {}("x")'.EVAL; say [$!.^name, $!]
camelia [X::AdHoc Unable to coerce value for '$ber' from Str to IO::Path; no coercion method defined
in block <unit> at EVAL_0 line 1
in block <unit> at <tmp> line 1
]
Zoffix But the fast path will go right away to try to call it and will just throw method not found error as a result: 11:05
m: try '-> IO::Path(Str) {}("x")'.EVAL; say [$!.^name, $!]
camelia [X::Method::NotFound No such method 'IO::Path' for invocant of type 'Str'
in block <unit> at EVAL_0 line 1
in block <unit> at <tmp> line 1
]
Zoffix (that's an implementation detail, but perhaps the docs should be more vague about what gets thrown when coercion fails) 11:07
Zoffix m: -> ::T \a where Array[T.of], \b where Array[T.of] {}(Array[Int], Array[Int]) 11:13
camelia ( no output )
Zoffix m: -> ::T \a where Array[T.of], \b where Array[T.of] {}(Array[Int], Array[Num])
camelia Constraint type check failed in binding to parameter 'b'; expected anonymous constraint to be met but got Array[Num] (Array[Num])
in block <unit> at <tmp> line 1
Zoffix lelf: doesn't look like that's implemented (filed as R#1383). That ^ looks like a slower workaround you could use 11:15
synopsebot R#1383 [open]: github.com/rakudo/rakudo/issues/1383 [NYI] Type capture in parametarization NYI
lelf ha! alright 11:22
m: (my Int @ = [1..3]).WHAT # ← is there a shortcut for this? 11:24
camelia ( no output )
lelf specifiying the *.of type I mean… this is also too wordy: 11:30
m: Array[Int]([1..3]).of
camelia ( no output )
jnthn Could be (my Int @ = 1..3) or Array[Int](1..3) I'd think 11:36
Saves creating a throw-away Array
moritz m: dd Array[Int].new(1, 2, 3) 11:49
camelia Array[Int].new(1, 2, 3)
dakkar re: type coercions, I understand everything that Zoffix said, and I know why it's that way. My issue is that the *documentation* says that in «Foo() $x», Foo is the parameter type that things get coerced to. 11:53
but that's not correct. It's both a type constraint, and the name of the conversion method
if I want a IO::Path, but not a IO::Special, and allow my callers to pass me a string… I can't 11:54
(and no, sub x(IO() $x where IO::Path) is a terrible work-around) 11:55
could we add a «sub IO::Path» alongside «sub IO», and document more clearly what the coercion actually does?
I may try doing that this weekend…
scimon Generally if I want a string that should be a valid IO::Path I make a subset but that's me. subset DirPath of Str where *.IO.d 12:01
That sort of thing.
scimon I dunno. I'm not feeling 100% at the moment so am probably not understanding. 12:03
andrzejku hi
can you see my messages?
DrForr andrzejku: Yes. 12:05
andrzejku thnks :)
andrzejku I decided to use sublime for irc and it looks it is working 12:05
DrForr irssi+tmux=shiny. 12:06
ilmari tmux, schmux 12:15
screen 4lyf
DrForr I started out with screen, but split-display sold me. 12:22
Altreus irssi ftl 12:26
weechat 4 lyf
moritz wtf 12:28
DrForr Shrug. 12:29
lelf I'm prob missing something trivial: how do you “X fff Y” getting list of lists (one for each matching sequence)? 12:55
moritz lelf: example?
lelf <begin 1 2 end begin 3 end X begin …> ⇒ [[1,2], [3], …] 12:57
masak lelf: not sure `fff` is the best tool for that, but it's definitely doable with (say) a recursive function or a loop/given construction 12:59
lelf: the problem you're bound to run into with `fff` is that its internal state is per *site* in the code -- which doesn't square well with needing one state per [] list 13:01
Geth doc: 0cfb87fb9e | (Luca Ferrari)++ | doc/Language/exceptions.pod6
Explain die uses $*ERR.

See issue #1722.
synopsebot Link: doc.perl6.org/language/exceptions
Geth doc: 28d8512437 | (Luca Ferrari)++ | doc/Language/exceptions.pod6
Add a $*ERR.say on a simple catch block as example of how to use STDERR.

See discussion in issue #1722.
masak (more exactly, there's one internal state per clone of the surrounding function. but that helps little here. I guess you *could* clone the function once per new list -- say, by re-entering the surrounding function -- but at this point I'd argue that we're far into X/Y land, and should use another tool.) 13:02
lelf: on a high level, what you appear to be doing is turning SAX into DOM :)
lelf: how do you plan to handle mis-nested data such as <end begin> or <begin end end begin> ? 13:03
(another thing `fff` won't do for you)
lelf no, no nesting 13:15
masak no nesting? um, but `[1, 2]` is nested inside `[[1, 2], [3], ...]` 13:21
lelf there can't be begin-end inside begin-end. so… say, one-level nesting 13:23
masak ok, now I see what you're saying 13:24
travis-ci Doc build passed. Luca Ferrari 'Add a $*ERR.say on a simple catch block as example of how to use STDERR. 13:26
travis-ci.org/perl6/doc/builds/327219218 github.com/perl6/doc/compare/e3ede...d8512437a1
masak m: my @result; my @part; for <begin 1 2 end begin 3 end> { when "begin" { @part = () }; when "end" { @result.push(@part) }; default { @part.push($_) } }; say @result.perl 13:26
camelia [[IntStr.new(3, "3")], [IntStr.new(3, "3")]]
masak hm :)
m: my @result; my @part; for <begin 1 2 end begin 3 end> { when "begin" { @part := [] }; when "end" { @result.push(@part) }; default { @part.push(+$_) } }; say @result.perl 13:27
camelia [[1, 2], [3]]
lelf I'm itching for a Haskell’s groupBy function (group elements by comparing adjacent ones). So like squish, but groupping, not uniquing. 13:28
jnthn lelf: See categorize and classify 13:30
Oh, but that doesn't require adjacency
masak yeah, that's where the `fff` angle isn't all bad 13:40
it just fails in other ways
nothing stopping you from writing your own grouping function, though 13:41
[Coke] waves good morning over his bacon & coffee. 13:44
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin') Z @i.grep(:k, 'end') 14:09
camelia ((0 3) (4 6))
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1) 14:10
camelia ((1 2) (5 5))
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1).map({@i[&infix<..>(|$_)}) 14:11
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in subscript; couldn't find final ']' (corresponding starter was at line 1)
at <tmp>:1
------> 3 'end').map(*-1).map({@i[&infix<..>(|$_)7⏏5})
expecting any of:…
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1).map({@i[&infix<..>(|$_)]}) 14:12
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
infix used at line 1. Did you mean 'indir', 'index'?
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1).map({@i[&infix:<..>(|$_)]})
camelia Too few positionals passed; expected 2 arguments but got 1
in block <unit> at <tmp> line 1
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1).map({@i[&infix:<..>(|$_.list)]})
camelia Too few positionals passed; expected 2 arguments but got 1
in block <unit> at <tmp> line 1
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1).map({ say $_.perl })
camelia 2
5
((1 True) (5 True))
moritz m: my @i = <begin 1 2 end begin 3 end>; say @i.grep(:k, 'begin').map(*+1) Z=> @i.grep(:k, 'end').map(*-1).map({ say $_.perl })
camelia 2
5
(1 => True 5 => True)
jkramer p6: '11325992321982432123259' ~~ m:ex/[ (\d ** 2..*) \d* ] ** 2..*/ 14:13
camelia MoarVM panic: Memory allocation failed; could not allocate 100208 bytes
moritz m: my @i = <begin 1 2 end begin 3 end>; (say @i.grep(:k, 'begin').map(*+1) Z=> @i.grep(:k, 'end').map(*-1)).map({@i[&infix:<..>(|$_.list)]})
camelia (1 => 2 5 => 5)
Too few positionals passed; expected 2 arguments but got 1
in block <unit> at <tmp> line 1
jkramer Does this work for anyone? It uses 100% CPU for me and hangs forever
moritz m: my @i = <begin 1 2 end begin 3 end>; (say @i.grep(:k, 'begin').map(*+1) Z=> @i.grep(:k, 'end').map(*-1)).map({say .perl }) 14:13
camelia (1 => 2 5 => 5)
Bool::True
jkramer Ah I guess it doesn't :) 14:14
moritz jkramer: try it with fewer characters first?
m: my @i = <begin 1 2 end begin 3 end>; (say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1)).map({say .perl })
camelia ((1 2) (5 5))
Bool::True
jkramer p6: '112112' ~~ m:ex/[ (\d ** 2..*) \d* ] ** 2..*/
camelia ( no output )
moritz m: my @i = <begin 1 2 end begin 3 end>; (say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1)).map({@i[&infix:<..>(|$_.list)]})
camelia ((1 2) (5 5))
Too few positionals passed; expected 2 arguments but got 1
in block <unit> at <tmp> line 1
jkramer Oh wait that regex is broken anyway :D 14:15
timotimo "memory allocation failed" is often the result of infinite recursion of some kind
moritz m: my @i = <begin 1 2 end begin 3 end>; (say @i.grep(:k, 'begin').map(*+1) Z @i.grep(:k, 'end').map(*-1)).map({@i[.[0] .. .[1]]})
camelia ((1 2) (5 5))
Index out of range. Is: 1, should be in 0..0
in block <unit> at <tmp> line 1
moritz :(
timotimo moritz: why not Z.. actually? and then index into @i directly with the result? 14:17
m: my @i = <begin 1 2 end begin 3 end>; (say @i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({@i[@$_]})
camelia (1..2 5..5)
timotimo m: my @i = <begin 1 2 end begin 3 end>; say (say @i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({@i[@$_]}) 14:18
camelia (1..2 5..5)
((1))
timotimo m: my @i = <begin 1 2 end begin 3 end>; say (say @i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({say $_; @i[@$_]})
camelia (1..2 5..5)
True
((1))
timotimo oh, haha
m: my @i = <begin 1 2 end begin 3 end>; say (@i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({say $_; @i[@$_]})
camelia 1..2
5..5
((1 2) (3))
buggable New CPAN upload: List-Util-0.0.1.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.1.tar.gz
timotimo m: my @i = <begin 1 2 end begin 3 end>; say (@i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({@i[@$_]})
camelia ((1 2) (3))
jkramer p6: '112112' ~~ m:ex/(\d ** 2..*) [\d* ($0)]+/ 14:20
camelia ( no output )
timotimo m: my @i = <begin 1 2 end begin 3 end>; @i.grep('begin'|'end') Zeq |('begin', 'end') xx * ?? (@i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({@i[@$_]}) !! fail "malformed begin/end pairins" 14:21
camelia Potential difficulties:
Useless use of Zeq in sink context
at <tmp>:1
------> 3nd begin 3 end>; @i.grep('begin'|'end') 7⏏5Zeq |('begin', 'end') xx * ?? (@i.grep(:
timotimo m: my @i = <begin 1 2 end begin 3 end>; ([&&] @i.grep('begin'|'end') Zeq |('begin', 'end') xx *) ?? (@i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({@i[@$_]}) !! fail "malformed begin/end pairins"
camelia ( no output )
timotimo m: my @i = <begin 1 2 end begin 3 end>; say ([&&] @i.grep('begin'|'end') Zeq |('begin', 'end') xx *) ?? (@i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({@i[@$_]}) !! fail "malformed begin/end pairins"
camelia ((1 2) (3))
timotimo m: my @i = <begin 1 2 begin 3 end end>; say ([&&] @i.grep('begin'|'end') Zeq |('begin', 'end') xx *) ?? (@i.grep(:k, 'begin').map(*+1) Z.. @i.grep(:k, 'end').map(*-1)).map({@i[@$_]}) !! fail "malformed begin/end pairins"
camelia malformed begin/end pairins
in block <unit> at <tmp> line 1
jkramer Any ideas why this doesn't match?
Shouldn't it return both 11 and 112 twice each? 14:22
timotimo you want :ex and :ov maybe? 14:23
p6: '112112' ~~ m:ex:ov/(\d ** 2..*) [\d* ($0)]+/
camelia ( no output )
timotimo p6: say '112112' ~~ m:ex:ov/(\d ** 2..*) [\d* ($0)]+/
camelia ()
timotimo p6: say '112112' ~~ m:ex/(\d ** 2..*) [\d* ($0)]+/
camelia ()
jkramer Doesn't ex imply ov? 14:24
> To find all possible matches of a regex – including overlapping ones – and several ones that start at the same position, use the :exhaustive (short :ex) adverb.
timotimo ah 14:25
p6: say '112112' ~~ rx/(\d ** 2..*) [\d* ($0)]+/
camelia Nil
timotimo p6: say '112112' ~~ rx/(\d ** 2..*)/ 14:26
camelia ｢112112｣
0 => ｢112112｣
timotimo p6: say '11112' ~~ rx/(\d ** 2..*) [($0)]+/
camelia Nil
timotimo p6: say '11112' ~~ rx/(\d ** 2..*) [($0)]/
camelia Nil
timotimo hm?
p6: say '11112' ~~ rx/(\d **? 2..*) [\d*? ($0)]/ 14:27
camelia Nil
jkramer p6: '112112' ~~ m:ex/(\d ** 2..*) \d* $0/
camelia ( no output )
jkramer Hmm weird the versionwith $0 without parens works for me locally, just not with ($0)
timotimo you're not outputting anything
p6: say '112112' ~~ m:ex/(\d ** 2..*) \d* $0/
camelia (｢112112｣
0 => ｢112｣ ｢11211｣
0 => ｢11｣ ｢12112｣
0 => ｢12｣)
jkramer Oh 14:28
p6: say '112112' ~~ m:ex/(\d ** 2..*) \d* ($0)/
camelia ()
timotimo oh
jkramer Why doesn't $0 work with parens?
timotimo because nested parens, i bet
m: say 'aabbcde' ~~ m/(\w) $0 ((\w) $0)/
camelia ｢aabb｣
0 => ｢a｣
1 => ｢bb｣
0 => ｢b｣
timotimo i should get waked up or something 14:29
m: say 'aabbaaba' ~~ m/(\w) $0 [(\w) $0]/
camelia ｢aaba｣
0 => ｢a｣
1 => ｢b｣
timotimo there you can see the difference more clearly; but most clearly if it's printed with proper newlines
p6: say '112112' ~~ m:P5:ex/(\d{2,}) \d* ($1)/ 14:33
camelia ()
timotimo p6: say '112112' ~~ m:P5:ex/(\d{2,}) \d* ($0)/
camelia ()
timotimo i wonder how we compile that
p6: say '112112' ~~ m:P5:ex/(\d{2,}) \d* $0/
camelia ()
timotimo jkramer: does that help? 14:36
not the last three lines, of course
those are just noise
jkramer timotimo: Yeah thanks, I found some kind of workaround using another parens around the whole thing: '112112' ~~ m:ex/((\d ** 2..*) \d* $0)+/ 14:38
moritz if you're already using :ex, why use the + inside the regex at all? 14:40
timotimo p6: say '112112' ~~ m:P5:ex/(\d{2,}) \d* \1/ 14:41
camelia ()
timotimo how does p5 regex do backreferences to numbered captured groups again ...
jkramer \1 :)
I think 14:42
timotimo right. $0 turned into an INTERPOLATE call in the :P5 regex i had up there 14:42
jkramer moritz: 'cause I want to count the repetitions of the number, but I haven't quite figured out how to do this with the Match object I'm getting 14:43
moritz jkramer: what problem are you trying to solve? 14:45
jkramer moritz: I want to find all numbers of 2+ digits in a string made of digits and the number of their occurances :) 14:47
It's not a real problem, just a random programming practice I found somewhere to get used to Perl 6 :)
moritz m: say '11211' ~~ m:ex/\d ** 2..* / 14:48
camelia (｢11211｣ ｢1121｣ ｢112｣ ｢11｣ ｢1211｣ ｢121｣ ｢12｣ ｢211｣ ｢21｣ ｢11｣)
moritz m: say ('11211' ~~ m:ex/\d ** 2..* /).map(*.Str).Bag 14:49
camelia Bag(11(2), 112, 1121, 11211, 12, 121, 1211, 21, 211)
timotimo m: say '11211'.comb.rotor(2=>-1).>>join("").Bag
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing dot on method call
at <tmp>:1
------> 3say '11211'.comb.rotor(2=>-1).>>7⏏5join("").Bag
expecting any of:
postfix
timotimo m: say '11211'.comb.rotor(2=>-1)>>.join("").Bag
camelia Bag(11(2), 12, 21)
timotimo jkramer: how's that?
moritz timotimo: that's only 2 digit numbers, not 2+ digit numbers 14:50
timotimo oh, of course!
timotimo m: say ('11211' ~~ rx:ex/\d ** 2..*/).Bag 14:50
camelia 5===SORRY!5=== Error while compiling <tmp>
Adverb ex not allowed on rx
at <tmp>:1
------> 3say ('11211' ~~ rx:ex/\d ** 2..*/7⏏5).Bag
timotimo m: say ('11211' ~~ m:ex/\d ** 2..*/).Bag
camelia Bag(｢11211｣, ｢1121｣, ｢112｣, ｢11｣, ｢11｣, ｢1211｣, ｢121｣, ｢12｣, ｢211｣, ｢21｣)
timotimo m: say ('11211' ~~ m:ex/\d ** 2..*/)>>.Str.Bag 14:51
camelia Bag(11(2), 112, 1121, 11211, 12, 121, 1211, 21, 211)
timotimo with the original regex implementaiton you had you also have the problem that the second time some two-digit thing appears, it'll be the start of another attempt again
so you'll count everything multiple times
moritz m: say bag '11211' ~~ m:ex/\d\d+/ # golfing! 14:52
camelia Bag(｢11211｣, ｢1121｣, ｢112｣, ｢11｣, ｢11｣, ｢1211｣, ｢121｣, ｢12｣, ｢211｣, ｢21｣)
moritz m: say bag ~<< '11211' ~~ m:ex/\d\d+/ # golfing!
camelia Bag(False)
moritz m: say bag ~<<('11211' ~~ m:ex/\d\d+/)
camelia Bag(11(2), 112, 1121, 11211, 12, 121, 1211, 21, 211)
timotimo that looks nicer 14:53
jdoege Good (whatever time of day it is wherever you may be). I have a question and i am hoping someone here may have an answer. I have a large grammar that is finally working really well. The last task I have for it is to ignore comments. My plan is to redefine the ws token to include a comment regex. However, to get started, I began by simply redefining the ws token to what I believe it already is: token ws { \s+ | <|w> } 14:58
When I test my parser with this redefined ws token, it quickly fails on test cases it previously passed on when I would expect (nearly) identical behavior. 15:00
Did I redefine ws incorrectly?
raschipi Did you use the non-capturing form? <.ws> 15:01
jdoege Also, I notice that the parser now explicitly parses ws (as seen in Grammar::Trace output) whereas previously it was implicit and invisible. Is that expected?
I never reference ws directly. This is just implicit rule usage. 15:02
perlpilot that last bit is interesting
raschipi So, after declaring it you lost the magic, is that it? 15:03
jdoege Yup.
jdoege Interesting. Despite the documentation saying that ws matches \s+ or <|w>, there is an example of redefining ws to not match vertical whitespace that looks like this: token ws { <!ww> \h* } When I change my definition to token ws { <!ww> \s* } my parser resumes working. A bit of documentation clarification/correction may be in order. I am not sure I understand the distinction between these two definitions. 15:18
jkramer timotimo: Nice, I didn't think of that, I focused too much on doing it all with a regex. :) I ended up with this one-liner (some formatting of the result included): 15:51
.match(/(..+)/, :ex)>>.Str.Bag.pairs.grep(*.value > 1).map(*.kv.join(':')).join(' ').say for $*IN.lines;
lizmat jkramer: I think you can leave out the .pairs there 15:52
jkramer lizmat: True, thanks :) 15:53
Saved a few more bytes by replacing .join(' ') with .Str :) 15:55
timotimo m: my $b = bag <hi hi hi bye bye lo>; say $b >>->> 1
camelia Cannot modify an immutable Bag (Bag())
in block <unit> at <tmp> line 1
timotimo m: my $b = bag <hi hi hi bye bye lo>
camelia ( no output )
timotimo m: my $b = <hi hi hi bye bye lo>.BagHash; say $b; say $b >>->> 1; say $b 15:56
camelia BagHash(bye(2), hi(3), lo)
BagHash(bye, hi(2))
BagHash(bye(2), hi(3), lo)
timotimo m: my $b = <hi hi hi bye bye lo>.BagHash; say $b; say ($b >>->> 1) >>+>> 1; say $b
camelia BagHash(bye(2), hi(3), lo)
BagHash(bye(2), hi(3))
BagHash(bye(2), hi(3), lo)
timotimo a nice, round-about way to filter out things that only have one entry in the bag
lizmat sometimes I think a dedicated method for Bags would help 15:57
timotimo dedicated what method? 16:01
lizmat for grepping out items by frequency 16:04
the .grep is general, so it goes out and creates Pairs, and take the value, etc. etc.
while a dedicated method wouldn't need to do all of that
Bag.frequency(1) would give all elements with a frequence of 1 16:05
Bag.frequency(* > 1) would give all elements with a frequence > 1
*frequency
Ulti by elements do you mean just the keys or the pairs? 16:06
this is a nice idea for if you're using bag models for stats 16:07
lizmat just the keys 16:07
[Coke] had to rediscover Test::More in p5 today, whoops. 16:28
[Coke] "why aren't I getting test descriptions!?" /me's brain full 16:29
moritz prove -v ? 16:34
[Coke] moritz: no, because I was using Perl 5's "Test", not "Test::More" 16:35
moritz aaaarg 16:38
[Coke] Wonder if it'd be worth trying to convert this to Inline::Perl5 (need WWW::Mechanize). 16:40
lizmat notes WWW::Mechanize is really needed 16:50
El_Che lizmat: or something minimalistic in the style of github.com/headzoo/surf 17:05
bdmatatu Hi folks, is there a p6 equivalent to Perl 5's 'do FILE'? 18:51
raschipi_ run? 18:53
DrForr You probably want EVALFILE, according to the 5to6 docs. 18:54
bdmatatu I think that runs it in another process...I'm thinking like 'use' but for something that's not *.pm or *.pm6
oh, thanks DrForr
lizmat bdmatatu: EVAL slurp $file 18:59
bdmatatu Great, that works too 19:01
lizmat FWIW, EVALFILE is better, and doesn't run in another process 19:02
lizmat in fact, it is basically implemented as EVAL slurp $fule 19:02
$file
bdmatatu I was kind of hoping it might also be possible to do that in the current lexical scope (e.g. any lexical vars in FILE would still be visible) 19:04
lizmat they are, if you use EVALFILE 19:07
bdmatatu oh!
El_Che EVALFILE will be the title of the first security trolling talk about Perl 6 ;)
lizmat bdmatatu: if you put a "say $a" in a file "foo", and then do "my $a = 42; EVALFILE 'foo', it *will* say "42" 19:08
bdmatatu Okay, I was thinking about declaring 'my $a' in file foo and then trying to use it after doing an EVALFILE 19:09
I could see how that would be bad securitywise
lizmat you cannot change a pad once it is executing 19:10
El_Che bdmatatu: isn't a yaml or json file an alternative for declaring variables? 19:11
[Coke] lizmat: I had someone actually give me a hard time about using perl 5 on this (even though I am the only one who writes one offs in perl here. He was suprised to see me use 5 instead of 6) 19:12
lizmat well, yes
[Coke] (he was just teasing, but I still felt like I cheated. :)
lizmat looks sternly at [Coke]
DrForr bdmatatu: Look at one of the config modules, or even (like others say) a .json or .yaml file - Letting configuration files run arbitrary code can get ugly. 19:13
bdmatatu Yeah I know it's ugly, but can be really handy for debugging and experimenting
timotimo or use a programming language that's made for config files: Lua
El_Che bdmatatu: It's not much effort to put the contents of a yaml or json file in an hash 19:14
lizmat m: my $a = 42 but role { method foo { "foo" } }; say $a; say $a.foo # perhaps another way to add debugging info to values ?
camelia 42
foo
bdmatatu The idea is that I want to load a script from a repl session and look at all the variables that were declared 19:15
lizmat m: role foo { has $.foo }; my $a = 42 but foo("bar"); say $a; say $a.foo # another way
camelia 42
bar
El_Che lizmat, AlexDaniel` , jnthn: 2017.12 rakudo linux packages are being built. Thx for fixing the bug 19:17
lizmat El_Che++ # perseverance! 19:17
El_Che lizmat: people call other names referring to that :) 19:18
TEttinger El_Che: tenacity?
El_Che pain in the ass
:)
TEttinger force of will? Nietzche's usage of a similar term in German? 19:19
El_Che don't go there :) 19:20
TEttinger persistence 19:22
:)
tobs I'd like to react whenever $*IO.lines but that code seems to want to slurp .lines. Iterating with 'for' works but I like Supply as a conceptual wrapper and also need multiple taps. 19:36
tobs Can someone help? A Supply of a lazy Seq seems like something natural 19:36
jnthn Try react whenever Supply.from-list($*IO.lines) { } 19:38
tobs That isn't interactive either. It waits until I Ctrl+D (Rakudo 2017-07 still, if that makes a difference) 19:39
tobs s/$*IO/$*IN/ specifically 19:41
jnthn Hm, curious, if I do perl6 -e 'react whenever Supply.from-list($*IN.lines) { .say }' then type foo and press enter, it immediately outputs foo 19:46
tobs Not here, with the same line. Updating Rakudo can't hurt (I hope) 19:47
lizmat STDIN isn't buffered, is it ?
jnthn It's either fixed in a more recent version or it's platform sensitive; handy I have 2017.12 on Linux (works) and 2017.07 on Windows (doesn't)
tobs 2017.07 on Linux here 19:48
jnthn OK, then it's most probably a fix since then
El_Che 2017.12 rpms/devs/apk should be out in some minutes
jnthn oh, also maybe try: react whenever Supply.from-list($*IN.lines, :scheduler($*SCHEDULER)) { .say }
tobs jnthn: that works 19:50
jnthn OK, then I can guess what change(s) fixed it :) 19:51
tobs (yes, 2017.12 works here too) 20:09
El_Che github.com/nxadm/rakudo-pkg/releas...g/v2017.12 20:27
2017.12 packages released 20:28
moritz El_Che++ 20:30
Zoffix m: &postcircumfix:<[ ]>() 20:30
camelia WARNINGS for <tmp>:
Useless use of constant value [ ] in sink context (lines 1, 1)
Cannot resolve caller postcircumfix:<[ ]>(...); none of these signatures match:
(\SELF, Any:U $type, |c is raw)
(\SELF, int $pos)
(\SELF, int $pos…
Zoffix m: &postcircumfix:<[ ]>(1, 0)
camelia WARNINGS for <tmp>:
Useless use of constant value [ ] in sink context (lines 1, 1)
Zoffix I'm debugging that warning.... Does this bit of QAST look like it should be there? gist.github.com/zoffixznet/549227c...xt-L53-L57 20:31
Nothing is there for `infix:<+>` and I tried removing that bit, but now the code doesn't find the right routine name. Wondering if that bit is supposed to be there of if I'm removing it too soon
moritz m: &circumfix:<[ ]>(1, 0) 20:33
camelia WARNINGS for <tmp>:
Useless use of constant value [ ] in sink context (lines 1, 1)
Cannot resolve caller circumfix:<[ ]>(Int, Int); none of these signatures match:
()
(Iterable:D \iterable)
(Mu \x)
in block <unit> at <tmp> li…
moritz Zoffix: ^^ do you mean that code?
'cause I only found a circumfix, not a postcircumfix, in that QAST tree 20:34
or is that the point?
comborico1611 Looking for an emacs tip. CHARACTER-HERE CURSOR-HERE. How do I drop character-here to a new line and preserve its indentation?
Zoffix moritz: yeah, that code, with &circumfix
comborico1611 and preserve its*
Zoffix moritz: no, the point is what appears two be two Stmts; one empty and one is just a `call &infix:<,>`, which I believe is causing the warning. They don't seem to be used for anything, are they? 20:35
moritz Zoffix: I wonder if the two elements of the infix:<,> call correspond to the [ and the ] from the routine name 20:36
Zoffix moritz: yeah. And it's made here: github.com/rakudo/rakudo/blob/mast....nqp#L2532
Well, s/yeah/I think so/; 20:37
Zoffix recompiles and dumps the values 20:38
damn, forgot .ast call; /me does recompiles again 20:43
Zoffix moritz: yeah, it's the circumfix: gist.github.com/zoffixznet/c4b9070...bdfeb5336a 20:47
the [ and ] chars for it
Zoffix My next best guess is something rips out ASTs for the colonpair with WVal(Str) for `infix:<+>`, but it forgets to handle the more complex colonpair for circimfix. Just need to find that rip-out bity 20:48
s/y$//
Zoffix relocates but will read replies, if any 20:49
timotimo DrForr: you may want to delete that comment on your blog, the one that advertises a lot of liposuction service related addresses? 20:54
actually, every one of those links points at the same url
masak .oO( those ads suck ) 21:03
timotimo oooooh 22:00
timotimo hmm. turning a for @plan { } with "when" statements inside into a while @plan with a $_ = @plan.shift in front makes when target the block outside of that 22:32
timotimo not surprising, but annoying in this case 22:32
buggable New CPAN upload: List-Util-0.0.2.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.2.tar.gz 22:48
Zoffix is disappointed Perl 5's Util vs Utils confusion made it into Perl 6 22:49
huh... 22:50
We have both List::Util and List::Utils? :S
Zoffix assumes the singular is the fruit of Project Camelia 22:52
lizmat Zoffix: List::Util is a more or less straight port of Perl 5's List::Util 22:55
colomon's List::Utils is a nice collection, but has no bearing on the Perl 5 version
lizmat goes to bed: enough porting for today 22:56
buggable New CPAN upload: Sub-Name-0.0.4.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.4.tar.gz 22:58
timotimo m: my $foo; my $p1 = a => $foo; my $p2 = a => $foo => 1; $foo = 99; say $p1; say $p2; 23:53
something wrong with camelia?
c: HEAD my $foo; my $p1 = a => $foo; my $p2 = a => $foo => 1; $foo = 99; say $p1; say $p2;
evalable6: say "hi"
what's happening
camelia a => 99
a => (Any) => 1
committable6 timotimo, ¦HEAD(0be9d55): «a => 99␤a => (Any) => 1␤»
evalable6 hi
timotimo OK
so, is this by design?

Powered by Cro and the Raku® Programming Language.

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

Thank you!