»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.
alphah_ samcv: it worked for me, I merged the pull request. 00:23
haxmeister hello perl6 00:29
sammers hello 00:52
haxmeister hello 01:01
how can I get line 35 to be concatenated onto line 34? pastebin.com/eZtqasPY 01:11
MIKE998 Ahoy! 01:17
What's the good tutorial for sub MAIN?
haxmeister finding good tutorials has been a struggle for me..lol
MIKE998 *crickets* 01:19
Perl is dead, eh?
haxmeister nope
I'm loving this shit
but they are more talkative about 4hrs prior to now 01:20
MasterDuke_ MIKE998: a lot of people have come here from perl6intro.com/, have you looked at it yet? 01:39
MIKE998 MasterDuke_: don't see MAIN in there 01:43
MasterDuke_ i see one or two mentions here learnxinyminutes.com/docs/perl6/, but i'm not aware of anything that goes into great depth on MAIN 01:45
MIKE998 I'm almost positive I saw a post about it. Maybe on learningperl6.org but I don't see it there now 01:47
.com 01:48
MasterDuke_ MIKE998: here's something linuxtot.com/parsing-command-line-a...in-perl-6/ 01:50
MIKE998 Yey! 01:51
MasterDuke_++ thanks
MasterDuke_ np 01:52
MIKE998 I'm gonna learn all the stuff now! 01:53
colomon The article is a little sloppy and incomplete, but does seem to have the basics pretty well. 01:54
For instance, he left out that the Perl 6 script is run like normal BEFORE calling MAIN. So if you have code which isn’t encapsulated in a sub or whatever, that will be executed just like MAIN wasn’t there. Then once that is done, MAIN is called. 01:56
raschipi Yep, MAIN at the end. 01:57
colomon and MAIN can be a multi sub
colomon wonders what other tricks he’s forgetting 01:58
raschipi m: say 'a'; sub MAIN { say 'b' }; say 'c';
camelia a
c
b
colomon there’s a way to specify alternate spellings of options, but it always seemed really fiddly to me 02:00
colomon (like -v versus —verbose) 02:01
raschipi At that point, I think it's better to get getopt out. 02:02
haxmeister \o/ 02:03
made my first simple class and object in perl6 today :-D
colomon haxmeister++ 02:04
MasterDuke_ colomon: you mean for generic named arguments, like `sub ( :color(:colour($c)) ) { } `, or something specific to MAIN?
colomon MasterDuke_: that looks right to me? 02:05
MasterDuke_ heh, i copied it from the docs
colomon I just got a headache every time I tried to do that, and eventurally I gave up.
haxmeister looks like greek to me.. and I did lisp and scheme fairly well a year go 02:06
so what is the difference between a module and a package? 02:07
raschipi m: sub f (:arg(:argu($c))) {say $c}; f(:arg('a')); f(:argu('b'))
camelia a
b
haxmeister oh thank you thank you thank you for making that syntax possible 02:09
Foo::Bar::<$baz>
haxmeister loves the little things 02:10
raschipi haxmeister: a module is one kind of package. 02:11
haxmeister trudging through docs.perl6.org/language/packages
raschipi haxmeister: I think that's has too much detail to start. 02:18
A package is something that says where a variable declared with 'our' will be created. 02:19
'our' creates a variable in the current package. 02:20
So, a package is one kind of namespace.
haxmeister I need to just make one.. maybe make one in a script I already have.. but even better, one in a separate file 02:24
but this page I'm reading doesn't tell me how.. it's just a reference page really
raschipi Other types of namespaces are the lexical scope, the dynamic scope and method resolution. 02:25
You want to make a module? 02:26
haxmeister: Here's a tool that automates it for you: github.com/gfldex/perl6-meta6-bin 02:29
haxmeister sorry raschipi I need to hand code one so it will stick 02:47
raschipi Have you seen docs.perl6.org/language/modules.html ? 02:50
haxmeister can't get the page up.. I guess my router is glitching or something 02:57
but I can get this one: perl6intro.com/#_perl_6_modules 03:01
only thing is I can't use zef 03:02
well that was a let down 03:04
we need more tuts guys
raschipi That one is about using them, not what you want.
haxmeister yeah
I can't get that page you linked 03:05
in fact docs.perl6.org is down
ah man.. I only got an hour left tops 03:06
perl6.org is down completely 03:08
haxmeister perl6maven.com is up though 03:16
raschipi Works for me.
haxmeister still no bones 03:18
modules section at perl6maven is incomplete 03:19
skids forgets if there is a single-item ".map" method, like .first for .grep 03:46
TreyHarris skids: map in lazy context should work fine: 03:56
m: my $counter = 0; my @f = 1..20; say (@f.map({$counter++; $_ * 2})).min; say $counter; $counter = 0; say (@f.map({$counter++; $_ * 2}))[0]; say $counter
camelia 2
20
2
1
TreyHarris skids: that's what you need, right? 03:57
skids I'm wondering if there's a way to not involve a Seq when I need to append generated code the the end of a value to process it. 03:59
TreyHarris skids: example? first involves a Seq too 04:00
skids so "chunk of code" ~ '.map({' ~ $other_chunk_of_code ~ '})' is what I'm using, but I really just want to call the second chunk of code. 04:01
Ah. 04:02
m: 4.&{$_ + 1}.say
camelia 5
skids That does it.
TreyHarris skids: oh, on a scalar. I thought, like first, you meant on a sequence of some sort. 04:03
m: {$_ + 1}(4).say 04:04
camelia 5
TreyHarris skids: ^^ nicer syntax I think.
skids Right but I can only append. 04:05
TreyHarris though if it's in a chain it might not be possible
skids Which was the problem.
TreyHarris oh, literally append generated code, like... eval'ing?
not even Callables?
skids Yeah, generating Perl6 code. Eventually I'll refactor the crap and eliminate that. 04:06
TreyHarris oh. then yeah. I assumed you meant "append" in a hand-wavy AST sort of manner.
sorry
I would've given you .&{} right off the bat in that case 04:07
there is a slightly conciser way with whatever though:
m: 4.&(* + 1).say 04:08
camelia 5
TreyHarris (parens, not braces, because it's a Whatever)
skids: follow?
skids Yeah the code could get complex enough to bury the ownership of the Whatever star.
TreyHarris docs.perl6.org is down for me 04:12
skids here too
TreyHarris I'd argue that if star is confusing, you should probably skip $_ and just use named placeholders to give you even better control and debuggability. 04:13
(Not knowing your use case, but imagining it, I mean. Like if you have an iterator of some sort you could use ($^it-1x, $^it-1y), ($^it-2x, $^it-2y), etc. 04:14
TreyHarris erm.... $^it1-x, etc. I meant. You can't have a \d to the right of a hyphen in an identifier. 04:15
TreyHarris (er, actually, you *can*, but if you do, you take the risk of arithmetic rules taking precedence and seeing a subtraction, so I think itt's best to just pretend you can't.) 04:19
skids Never been much of a fan of the hyphenization trend. Especially for allcaps. 04:20
Anway, placeholders do offer some deconfliction beyond $_, but it's a margnal gain compared to how easy it is to confuse *. 04:21
TreyHarris skids: if you're writing a translator, style is te change underscores to hyphens, because underscores are supposed to indicate "beware, this may be somewhat magical"
s/change/change Perl 5/ 04:22
skids Yeah I know. It's something I more tolerate about Perl6 than like, though.
TreyHarris not saying you have to follow that. just be aware that connotation is out there. 04:23
Perl 6 identifiers are actually subject to precedence-busting in quite a few edge cases, particularly if you play with sigilless variables or with quoting constructs. 04:24
m: my $isn't = False; say "It is $isn't." 04:25
camelia It is False.
TreyHarris m: my $isn't = False; say q'It is $isn't.'
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in single quotes; couldn't find final "'"
at <tmp>:1
------> 3my $isn't = False; say q'It is $isn't.'7⏏5<EOL>
expecting any of:
dotty method or postfix…
TreyHarris m: my $isn't = False; say q'It is {$isn't}.'
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in single quotes; couldn't find final "'"
at <tmp>:1
------> 3y $isn't = False; say q'It is {$isn't}.'7⏏5<EOL>
expecting any of:
dotty method or postfi…
skids recalls something with grammar rule names not liking them.
TreyHarris (I'm not actually sure there's any way to use an apostrophe-containing identifier inside a single-quoted interpolating construct.) 04:26
skids Oh maye it was `
erm '
TreyHarris grammar rules not liking "them" == "identifiers with non alphaunderscorenumerics"? 04:27
TreyHarris I haven't seen that 04:27
skids may be fixed. 04:28
I remember working around it in Grammar::ABNF. Maybe should go un-work-around it. 04:29
TreyHarris it's just that quoting and arithmetic operators--which can cause ambiguity with identifiers--are higher in precedence.
TreyHarris I've learned a lot by running perl6 --target=ast. 04:32
skids Hrm maybe Grammar::ABNF wasn't where I ran into that. Just confusing with a similar problem. 04:33
TreyHarris chuckles 04:34
Guido would be very, very unhappy with this lexical ambiguity business :-)
skids Obviously need to go to bed. Just spent 5 minutes figuring out I had a test file with plan 63 and no tests, that actually had no tests, and wasn't crashing silently before running them. 04:35
TreyHarris skids: nighty night
samcv o/ all 04:40
lookatme o/ 04:44
samcv hey lookatme 04:45
how is your grammar going?
TreyHarris looks at samcv
samcv hehe 04:46
TreyHarris what is the 'xt' test dir for? 04:51
samcv xt? 04:53
TreyHarris oh, author tests.
samcv design.perl6.org/S22.html site is down 04:56
xt folder isn't in the design specs 04:57
sammers I think modules.perl6.org is down... 04:59
samcv everybody panic! 05:00
samcv runs around screaming
TreyHarris *.perl6.org has been sporadic for a few hours now. 05:08
ugexe TreyHarris: i dont understand the second half of your issue 05:09
TreyHarris ugexe: what issue?
ugexe zef 05:10
TreyHarris oh, with zef? Probably an editing issue, as I only had one. Let me look
ugexe you were talking about a url library, and the -u flag defaulting to true
TreyHarris oh... `homebrew browse` opens the browser onto the URL. `homebrew browse -u` just prints the URL the browse would have been opened to
I don't see a portable browser-opening lib in ecosystem, so I was suggesting to default to the -u behavior for now. 05:11
ugexe oh not open source, now i understand
TreyHarris ah-hah. yes. I'll edit. 05:12
done.
ugexe the long way is `zef info zef|grep bugtracker` 05:13
TreyHarris are the synopses on GitHub, since design.perl6.org is down? 05:16
I thought they were in doc, but apparently not 05:17
ah. "specs" 05:18
TreyHarris ugexe: oh, interesting. S22 doesn't provide a homepage attribute (I thought I saw one), just source and bugtracker, both of which can be used to derive a GitHub homepage, but won't work if there's a dedicated project domain. Is that intentional? 05:20
ugexe "possible" names are: ... 05:21
its just listing things that make sense 05:22
TreyHarris ugexe: oh, I see. I read that like "possible states of an electrical charge are positive and negative" 05:23
i.e., things not listed there are not possile
ugexe yeah, thats probably what it meant, but im saying theres a loophole
realistically there is no reason for support to be defined in s22 05:24
ugexe other than as a suggestion 05:25
TreyHarris since many/most projects won't fill-in "phone", I think it's understood as not being a list of required fields. but it does read as an exhaustive list of fields to me. I'd say either add a "project homepage", or explicitly specify, "s/possible names are/possible names may include/" 05:27
TreyHarris and since I think homepage, like source, are likely to be programmatically useful (as I've given an example), I'd prefer it be specified. 05:28
but it needn't be mandatory any more than phone.
ugexe not just phone! A fully qualified phone number (with potential cost indication) 05:31
TreyHarris :-)
samcv TreyHarris, what's the windows api? 05:35
on linux it's xdg-open or whatever
not sure on mac
TreyHarris samcv: dunno. on mac it's just "open". you can type a URL into Run: and it opens a browser, so it may just be cmd.exe, but I'd have to test and my Win machine's currently updating 05:36
if I have a class Foo { has Tup2 $.x... } where Tup2 is a 2-ple, can I do "handles" so that .fst delegates to $!x[0] and .snd delegates to $![1], or do I need explicit methods for that? 05:39
ugexe powershell -Command Start-Process google.com 05:40
lookatme samcv, oh, fine. Current I make a terminal table module, it will be available soon . 05:42
samcv ok :)
lookatme I took a afternoon nap just now . 05:47
TreyHarris Ah, in this case, I can do handles <fst first> and handles <snd tail> because there are unary methods for the first and last elements of a tuple, but what if it were a 3-tuple and I wanted <fst snd thd>?
TreyHarris did y'all just get a rash of racist privmsgs, too? 05:54
TreyHarris okay, time for bed. & 06:00
lookatme :) 06:02
moritz just seen on twitter: "I hate fractals. They are so full of themselves." :-) 06:41
Geth perl6.org: 5026b1922c | (Steve Mynott)++ | source/downloads/index.html
How to build 2017.04 R* on Raspbian
 07:32
moritz FYI I've rebooted www.p6c.org (which also runs perl6.org) because it hung up during the night 07:42
samcv cool 08:11
Geth whateverable: 0ff5de677d | (Aleks-Daniel Jakimenko-Aleksejev)++ | Evalable.p6
Don't search for camelia if no space after m:

Camelia does not answer on “m:say 42” anyway, no need to check if camelia is present.		 10:35
whateverable: 31bc4c8515 | (Aleks-Daniel Jakimenko-Aleksejev)++ | Evalable.p6
Detect messages starting with “say” and answer sometimes

This change has been there for a while, I simply forgot to commit it. It seems that there was at least one time it answered unnecessarily, so perhaps the regex can be improved.
But generally, the idea is really nice. If it becomes annoyng for some reason, I can always make the regex more strict.
masak harsh lesson of the day: do not expect Perl 5 subs to take a lexical closure of a surrounding while loop iteration... :/ 11:50
the little things you start taking for granted...
(hi, #perl6)
timotimo hi masak
masak $ perl -Mstrict -wE'my $c = 2; while ($c) { my $d = $c--; sub foo { say $d }; foo() }' 11:52
2
there we have it, exhibited
m: my $c = 2; while $c { my $d = $c--; sub foo { say $d }; foo() } 11:53
camelia 2
1
masak sobs happily
jnthn masak: What if you put an explicit my on the sub? 11:58
Juerd 1;0 juerd@cxien:~$ perl -Mexperimental=lexical_subs -Mstrict -wE'my $c = 2; while ($c) { my $d = $c--; my sub foo { say $d }; foo() }' 12:02
2
1
Juerd I'm amazed there's no warning for putting a non-lexical sub in a while loop 12:03
moritz I'm sure somebody on p5p suggested doing that, and somebody else cited a semi-valid use case that prevented the warning from getting in 12:05
masak jnthn: that would work from 5.18 and onwards, but the code I'm writing (for reasons outside of my control) needs to be 5.14-compatible. 12:14
timotimo hah, ouch
jnthn masak: Ah, fair enough
timotimo 5.14 isn't ridiculously old, is it?
colomon 2011 12:22
timotimo not that bad 12:26
nine LOL I just discovered that there's an actual /usr/bin/[ command 13:14
Juerd Syntactic sugar is bad for you :) 13:15
nine Btw. from the backlog. The actual Perl 5ish way is: nine@sphinx:~> perl -Mstrict -wE'my $c = 2; while ($c) { my $d = $c--; my $foo = sub { say $d }; $foo->() }' 13:16
2
1
Juerd nine: Or "use experimental qw(lexical_subs);" :) 13:37
ugexe timotimo: didn't you disable telemeh by default? rakudo appveyor tests have kept failing with telemeh 14:29
timotimo i did, yes 14:34
i don't think rakudo appveyor pulls in moar-HEAD 14:35
samcv src\profiler\telemeh.c(23) : warning C4391: 'unsigned int __rdtscp(unsigned int *)' : incorrect return type for intrinsic function, expected 'unsigned __int64' 14:36
hmm
intrisic function cannot be defined it says 14:37
timotimo i think i fixed that
samcv after that
msdn.microsoft.com/en-us/library/s85a3w4b.aspx
ah
timotimo, can you get jnthn to enable appveyor for moarvm? 14:39
timotimo do you know how we could put it under moarvm/moarvm on appveyor?
rather than timo/moarvm or samcv/moarvm?
samcv i do not
but any appveyor better than none
Ulti so if I wanted to read a file line by line what's the way thats going to be the best P6 can pull off right now? 14:52
perf wise
jnthn Ulti: Quite possibly the obvious one (for $fh.lines { }) 14:53
Ulti: In so far as it's been optimized a bunch
Ulti okedoke, currently comparing that to the P5 near equivalent while($line = <$fh>) {}
so far its not the best story 14:54
jnthn You may find doing a while loop that calls .get is faster, but it might equally be slower because lines can avoid various checks
Ulti yeah thats what I was wondering is it the iteration and the overhead of all the fancy nice things in P6 that is slow vs the raw IO 14:54
guess I can profile this :S 14:55
timotimo you should also try if slurp + lines is any faster
Ulti this is a 9GB SQL file
jnthn Also, decoding/NFG ain't free
timotimo oh, not slurp, then :)
Ulti yeah exactly this is more an actual real world test
because Im noticing most things are fast enough but whenever I hit files of an order I work on from minute to minute its just not really possible yet 14:56
jnthn: yeah maybe buf would be better
so like a binary version of lines
.records
jnthn Well, you can :env<ascii> on opening if you know that's what you have
Ulti *shrug*
jnthn uh, enc
or latin-1
Ulti I have bytes jnthn :3
its obviously bytes I can see them 14:57
but yeah most SQL dumps are about a million encodings at once in my experience
jnthn hah :)
timotimo if you go with :enc<latin1> you'll get almost what you had as bytes, but we do turn \r\n into a single grapheme 14:58
Ulti P5 0m28.867s vs P6 7m16.588s
I'll go with the encoding specified 14:59
14x isnt the worst
given this is on the bigger end of the scale 15:00
it was defo like an order of magnitude or more worse a couple of years ago
MasterDuke_ i just ran another perf record of `for $fh.lines {}' on a 1m line file, 20% in MVM_string_utf8_decodestream, 11% in find_separator.isra.6 15:01
Ulti I mean thats 14x on Perl 5 which has super optimal IO 15:02
MasterDuke_ if i do `for $fh.lines(:enc<latin1>) {}' it's 15% in find_separator.isra.6 and 11% in MVM_string_latin1_decodestream 15:04
Ulti 4m53.429s with latin1 set 15:06
lichtkind m: 10 % -3 15:13
camelia WARNINGS for <tmp>:
Useless use of "%" in expression "10 % -" in sink context (line 1)
lichtkind m: say 10 % -3
camelia -2
lichtkind why?
lichtkind cheers khw 15:13
why has perl 6 sam modulo bug as in perl 5? 15:14
cosimo -win 18
timotimo lichtkind: and python 15:15
lichtkind this shouldnt be an argument :)
perl 5 took the shitty oo from python 2 and regretted it since too 15:16
timotimo i didn't know they did that
lichtkind in old interview larry said at least so
not sure it was a pun
timotimo OK
lichtkind so i have to define my own? 15:17
lichtkind if i understeand you correctly it was very intentional? 15:18
lichtkind even mathematically wrong 15:18
Zoffix lichtkind: it's not a bug. There's more than one right answer and many langs do different things 15:19
araraloren Yeah, it not only one answer
lichtkind it long socks style of math
its
Zoffix points to the table on the right on en.wikipedia.org/wiki/Modulo_opera..._operation 15:20
haxmeister I always loved perl5.. it was my first love.. perl6 is a step above all other interpreted languages.. very good work guys 15:21
timotimo lichtkind: i think you mean "longstocking" 15:22
lichtkind we all love perl, except when we dont :)
timotimo you got it
Zoffix haxmeister: not sure if you got an answer but: you can do it as `say "Surface footage at 12224 RPM = {$mill.calc-sfm :rpm(12224)}";` or `say "Surface footage at 12224 RPM = $mill.calc-sfm(:rpm(12224))"` or `say "Surface footage at 12224 RPM = $mill.calc-sfm(:12224rpm)";` 15:23
TIL $o.meth :named(arg) is valid syntax :D
s/:D/:S/;
timotimo of course it is ;( 15:24
because we don't want to force commas between colonpair named arguments
haxmeister yes Zoffix I figured it out shortly after asking.. but I managed it like this: say "Surface footage at 12224 RPM = " ~ ($mill.calc-sfm :rpm(12224));
Zoffix I always wrote $o.meth: :foo(42)
timotimo oh, does that actually attach the named argument to the method call?
Zoffix timotimo: but I meant there's no () or : after the method.
timotimo i'm afraid you might fall on your nose in some circumstances 15:25
Zoffix m: class Foo { meth (:$bar) { dd $bar } }.meth :42bar
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$bar' is not declared
at <tmp>:1
------> 3class Foo { meth (:7⏏5$bar) { dd $bar } }.meth :42bar
Zoffix wat
timotimo "meth" :)
Zoffix ah
haxmeister thank you for an answer though, I do have irssi running in screen from a server and I am able to see messages from when I was away
Zoffix m: class Foo { method meth (:$bar) { dd $bar } }.meth :42bar
camelia Int $bar = 42
Zoffix m: class Foo { method meth (:$bar, :$meow) { dd [ $bar, $meow ] } }.meth :42bar :72meow
camelia [42, 72]
Zoffix meat
timotimo m: class Foo { method meth (:$bar) { dd $bar } }; say Foo.meth :42bar, Foo.meth :99bar 15:26
camelia Int $bar = 42
Int $bar = 99
Unexpected named argument 'bar' passed
in block <unit> at <tmp> line 1
Zoffix *neat
It went into `say`
timotimo okay that's just a bug
no, it went into Foo.meth *and* into `say`
Zoffix m: class Foo { method meth (:$bar) { dd $bar } }; say Foo.meth: :42bar, Foo.meth: :99bar
camelia Int $bar = 99
Too many positionals passed; expected 1 argument but got 2
in method meth at <tmp> line 1
in block <unit> at <tmp> line 1
jnthn And in today's edition of "surprising upshots of adverbs being parsed as infixes and then we lie about it"... :)
Zoffix m: class Foo { method meth (:$bar) { dd $bar } }; say Foo.meth(:42bar), Foo.meth: :99bar
camelia Int $bar = 42
Int $bar = 99
NilNil
Zoffix :) 15:27
Ulti: what P6 version are you on? 15:29
Ulti Rakudo version 2017.04.3-172-gf2af3db16 built on MoarVM version 2017.04-56-g8ad18b84 so HEAD or close to it 15:31
Zoffix ok; that'll have all the recent IO perf improvements :/ 15:32
Ulti :/
I mean there will have been a lot of memory pressure given my laptop is complete cack at work for everything I have to do 15:33
but its still like for like with respect to P5
Zoffix P6 IO::Path.lines with a for loop is 25x slower than P5 while (readline) on my box... 15:34
Zoffix tries bare nqp
Basically just 8x slower 15:40
For a 5-meg file. `time perl6 -e 'use nqp; my $fh := nqp::open("foo", "r"); nqp::while(($_ := nqp::readlinechompfh($fh)), nqp::null)` vs `time perl -wlE 'open my $fh, "<", "foo"; while (<$fh>) {}'`
jnthn Zoffix: And if you get Perl 5 to do utf-8 decoding too? 15:42
araraloren perl6 25x slower, nqp 8x slower . 15:43
Zoffix Oh right. 15:44
jnthn: then bare-nqp version is 4.4x slower than perl -wlE 'open my $fh, "<:encoding(UTF-8)", "foo"; while (<$fh>) {}' 15:45
Pretty good!
jnthn Some room for improvement still, though. :-) 15:46
Zoffix And perl6-only version is 13x slower 15:46
timotimo that's almost 2x faster than the 25x slower version 15:47
Zoffix timotimo: well yeah, perl5 with unicode decoding is like 2x slower 15:48
timotimo the thing is that even when we do latin1 decoding, we'll generate synthetic graphemes for \n and \r\n (and maybe also \r?)
Zoffix And actually if you go to IO::Handle.lines directly, it's only 9.25x slower
timotimo or ... only for \r\n?
Zoffix Which is weird :S 15:49
s: "foo".IO, 'lines', \()
Right. no robot. But IIRC it's just delegates to IO::Handle
Ah. It's IO::Handle.lines vs IO::Handle.lines(:close) 15:50
My close-iterator sucks, I guess :|
araraloren Why Perl6 slower than nqp . The parser ?
Zoffix araraloren: it does more things 15:51
araraloren: you get a proper Seq + phasers + IO::Path + choosing mode etc etdc
araraloren The gap is bigger than I thought 15:54
Zoffix araraloren: there's still a ton of work to be done to make the gap smaller :) 15:57
araraloren Em, thanks all the developer. 15:58
Perl 6 will be faster and more faster in the future. 15:59
alphah Hi, Is there difference between thiese two definitions for this token: 16:01
token star-core{ '-' <([ 'x86_64' | 'i386' ])> } 16:02
and:
alphah roto token star-core { * } 16:02
proto token star-core { * }
token star-core:sym<x86_64>{ '-' <(<sym>)> }
token star-core:sym<i386>{ '-' <(<sym>)> }
Zoffix alphah: don't think so. Why? 16:03
alphah any difference of any kind that can make first one match and second one thrwos NQPMu error, when parseed against '-x8_64' ?
araraloren no, I think
Zoffix alphah: what's your Perl 6 version?
alphah Zoffix: This is Rakudo version 2017.04.3-47-gf0414c468 built on MoarVM version 2017.04-44-gf0db8822 16:04
implementing Perl 6.c.
Zoffix m: grammar { proto token star-core { * }; token star-core:sym<x86_64> { '-' <(<sym>)> }; token star-core:sym<i386> { '-' <(<sym>)> } }.parse: :rule<star-core>, '-x8_64'
camelia ( no output )
Zoffix alphah: what's the NQP error?
m: say grammar { proto token star-core { * }; token star-core:sym<x86_64> { '-' <(<sym>)> }; token star-core:sym<i386> { '-' <(<sym>)> } }.parse: :rule<star-core>, '-x8_64'
camelia Nil 16:05
alphah No such method 'perl' for invocant of type 'NQPMu'
in block <unit> at ./bin/galaxy.p6 line 7
so, without this commit, I get the NQPMu error: github.com/alphah77/galaxy/commit/...801486489b 16:06
Zoffix alphah: you don't have a 2017.03 or earlier rakudos laying around, by any chance. Do you?
As in, does the bug happen there.
alphah I can install latest version, I think I reinstalled rakudo two weeks ago, here is the output of perl6 --version: 16:07
This is Rakudo version 2017.04.3-47-gf0414c468 built on MoarVM version 2017.04-44-gf0db8822
araraloren So make a clean install will be solved ?
ugexe s/<(<sym>)>/<::(<sym>)>/ ?
alphah I'll do that right now and try,,, I will remove ~/.rakudobrew and install again,, is that sufficant? 16:08
Zoffix alphah: no, I meant earlier rakudo. There's some work that went in a few days before the 2017.04 release that we later found had a bug with capture markers in grammars and.... you have capture markers in grammars
timotimo alphah: do you get a better error when you give perl6 --ll-exception?
i.e. a traceback from the "no such method .perl" error?
perlpilot notices that someone is a fan of <( and )> 16:09
I don't think I've seen them used so much :)
jnthn I use them now and then
Zoffix ugexe: isn't that different? <(<sym>)> is <( <sym> )>, no?
jnthn They can save having to use a named capture in various cases
Or capture at all
Zoffix alphah: with rakudobrew, you could do `rakudobrew build moar 2017.03` `rakudobrew switch moar-2017.03` and then you could still switch back to latest 16:10
perlpilot yeah
Zoffix fails to repro any perl on NQPMu on 2017.04.2-6-g894ba82 16:12
alphah: what was the other commit I saw you make, something about changing .ast method to .Str or something? Was that also to workaround some bug? 16:13
alphah timotimo: I got this error message when used --ll-exception: gist.github.com/alphah77/0266d18bb...94bd6e255a
alphah Zoffix: That was the commit: replaced .ast by .Str: gist.github.com/alphah77/0266d18bb...94bd6e255a 16:16
Zoffix alphah: well yeah, why did you replace it?
araraloren Would you mind share your code ? 16:17
Zoffix alphah: anyway, rakudobrew 2017.03 rakudo. Use `rakudobrew switch` to switch between it and latest. Try your bugs with both. If they aren't present in 2017.03, report them, as that's very likely due to Uncurse Curse 16:17
araraloren I want try it on my notebook . 16:18
alphah I thought that the NQPMu error shows because is trying to access the token before it's made. so I just tried to remove .ast and use .Str,,,, I know not logical but it worked
araraloren I install Perl 6 from source .
alphah araraloren: here is the code github.com/alphah77/galaxy 16:19
araraloren alphah, thank u
alphah you will need to run the program as bin/galaxy laws bin/laws rakudostar-0.7-x86_64 to produce the error 16:20
moritz I seem to recall a tool that fat-packages a Perl 6 application into a Windows exe file. Possibly created by jnthn++
anyone got a link for me?
jnthn eco: installer 16:21
aww
moritz: App::InstallerMaker::WiX on the ecosystem :)
moritz jnthn: thanks 16:22
jnthn It doesn't quite do that, but it does make an MSI :)
MasterDuke_ buggable: eco installer
buggable MasterDuke_, Found 3 results: App::redpanda, Sparrowdo::Cpm, App::InstallerMaker::WiX. See modules.perl6.org/#q=installer
alphah trying rakudobrew build moar 2017.03 now
moritz and then samcv++ did some AppImage (iirc?) wizardry... is that usable?
jnthn I'm tempted to make a perl6-winimal "distro" using it that makes a mega simple installer that includes Perl 6, zef, and the debugger frontend or so :) 16:23
moritz in case you can't tell, I'm trying to write a section for my book on how to deploy Perl 6 applications
jnthn Ah :-) 16:24
El_Che moritz: you covered docker, I suppose
jnthn The installers that thing makes are super-minimal, but have been successfully used to deploy a Perl 6 app :-)
timotimo don't spend too many pages on docker, what if docker completely dies next year? :) 16:25
El_Che timotimo: then you update the book and get $ again :)
timotimo haha
jnthn Given how widespread docker has become, not to mention containers in general, that's quite a bold prediction :-)
El_Che at the rate perl6 advances, docker seems to be the only sane way to deploy perl6 16:26
araraloren alphah, no error, in `This is Rakudo version 2017.04.3 built on MoarVM version 2017.04-44-gf0db882
`, installed from rakudo source
jnthn today discovered that docker special-cases connecting to a registry on localhost to bypass SSL certificate verification... :S
araraloren The command what I tried `perl6 -Ilib ./bin/galaxy.p6 laws bin/laws rakudostar-0.7-x86_64` 16:27
alphah araraloren: does the output has value "core => Str|x86_64" , or core => NQPMu ? 16:29
moritz El_Che: re docker, yes
araraloren Yeah, it has
alphah Nice, Im installing this version then
moritz wonders if he should mention moby 16:30
araraloren gist.github.com/araraloren/5598955...d7c71cd0f, alphah
jnthn moritz: yes, when I went source digging today I was like "wait, this isn't the official docker repo" and tried to go to it and was redirected straight back where I came from :P
alphah looks right.
El_Che moritz: as a anecdocte: I have a perl6 app in prod that run from a "global" perl6 install that I don't date to touch. The app will not run on recent perl6 versions :). It would be easy to migrate with docker (what I use elsewhere) 16:31
moritz: becasue I use perl6 packages, the image is created by jenkins in no time
araraloren I agree! :) 16:34
alphah araraloren: would you mind trying again same command above but after checkout 802112596560c64876ea3bb44bd998ce65c52e1e, the reason is latest commit is working fine 17:00
araraloren gist.github.com/araraloren/5598955...0d7c71cd0f 17:05
alphah, I updated output
alphah checked it, so it did not work:token did not match: core => NQPMu. and only change in this commit is definig the token as proto token {8} 17:06
s/8/*/
araraloren I checked too
It's also worked 17:07
fine
alphah building moar 2017.03 and testing on this version...
alphah araraloren: it did not, compare the output of the two, you will se in first: gist.github.com/araraloren/5598955...tput-p6-L5 17:09
second: gist.github.com/araraloren/5598955...tput-p6-L5
line 6 on both files. 17:10
araraloren Yeah, I see it, alphah :) 17:16
alphah nice :) 17:17
also tried on moar 2017.03, still same issue.
araraloren Em, What about master branch ? 17:18
alphah any docs to install it? I usually install rakudobrew then rakudobrew does the rest 17:19
araraloren What
you mean?
Install from source ?
alphah you mean I should install perl6 from master branch?, or I understood wrong? 17:20
araraloren No, I mean your code
alphah aah, master branch work because i just changed the definition of the token (Which i bileve should be exactly the same); in here github.com/alphah77/galaxy/commit/...801486489b 17:21
araraloren Oh, I got it. 17:22
alphah :) 17:23
araraloren alphah, I saw you code, and think you forget write action method 17:36
It's no bug
in rakudo
alphah araraloren: can you point me where I missed action method? I have class Law as Action class, I also have method for that token 17:39
araraloren It's `star-core`
You should write `proto version`
not regular version
alphah I wanna bang my head in the keyboard now : D 17:40
araraloren haha.. :P 17:41
eater is it possible to call a perl6 script within a certain scope in perl6? a bit like PHP's include / require 17:54
araraloren Have you look at require : docs.perl6.org/syntax/require ? 17:56
eater how can I include an file with that :o 17:57
araraloren It just import module at runtime. 17:58
I think it's no mechanism in Perl 6.
eater yeah, and I want to run a script at runtime 17:59
araraloren Yeah, you can use require run a script, maybe .
Just try it.
Geth modules.perl6.org: c9bce27d29 | (Zoffix Znet)++ | 2 files
[REBUILD] Reject too-large logotypes

  - We first get byte size from github, reject right there and then
   if we have more than 32*32*32*2 bytes in logotype
  - If that passes, fetch the pic and take a peek at the size of the
   image and reject it if it's not 32px x 32px		 18:03
araraloren eater, I test it on a example. It's work. 18:04
Just think you have a r.p6: use lib "."; require "m.p6"; and m.p6: say "CALL ME"; 18:05
It will print CALL ME .
TimToady well, there are four or five different things you can mean by "call a script within a certain scope", and sometimes you want 'run', or 'shell', or 'EVALFILE', or even 'EVAL', depending on how much and what kind of dataflow you need in and out of that scope 18:06
it might also mean you need to write a role, if you're thinking of "include" as a way to write generic code 18:08
so I wouldn't get hung up on thinking you have to use a verb called "include" or "require" 18:09
that would be a bit of an XY problem, if so
araraloren TimToady, Which way can import symbol I declare in that perl6 file ? 18:10
require and ?
TimToady well, require can do that, but I don't see eater asking for that, offhand
maybe eater is just looking for qx// or qqx// 18:11
araraloren Em, maybe.
TimToady "run in a scope" is terribly ambiguous
araraloren Yeah. 18:12
eater Im sorry :')
in PHP you have include / require
which will just run the given file in that scope
araraloren Is it like include in c /? 18:13
Import symbol to current scope ?
Geth modules.perl6.org: 40e5cd16fa | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/Source.pm
Return failure from _save_logo when size is no good

In case something actually uses that value
eater araraloren: not really, php's include is at runtime while C's is at compile time 18:14
araraloren Oh, except that ? 18:15
TimToady what are you expecting it to do to the current scope after it has run?
and do you mean dynamic scope or lexical scope? 18:16
or some other scope?
what kind of data flows back and forth? 18:17
or are you expecting it to define something?
eater dynamic scope 18:18
TimToady another degree of freedom is whether you want it to reparse the file every time you call it 18:20
eater as long as it's parsed everytime I run perl itself I'm happy :> 18:21
TimToady do you really need it reparsed every time, or just if it changes? 18:22
eater ^
ye
Geth doc/molecules-patch-1: 3794e8dcea | (Christopher Bottoms)++ (committed using GitHub Web editor) | doc/Type/Callable.pod6
Added multi-positional use of assuming

Borrowed content from brian d foy's Stack Overflow post stackoverflow.com/a/43669837/215487		 18:23
TimToady you do you expect it to relate to your current program? how does the data flow? what symbols get defined? 18:24
eater what do you mean with "how does the data flow"?
TimToady what are you trying to do with the file that is run? 18:25
Geth doc: molecules++ created pull request #1296:
Added multi-positional use of assuming		 18:26
eater i'm trying to create a program that reads an configuration script that's also in perl6, it defines tasks and those are then executed by the main program
TimToady okay, so you want it to build a data structure that you can access after you call it, which means you probably want a symbol to be imported so you can access that data structure 18:28
eater yes
masak nine: thanks for the tip about the p5ish way. that did not occur to me. I might still use that to make the code less obnoxious.
TimToady in that case, either 'use' or 'require' can do that
Geth modules.perl6.org: 8ce873a75d | (Zoffix Znet)++ | 7 files
[REBUILD] De-bitrot Mojolicious stuff

  - Bump Mojolicious prereq to 7.31 and fix all Mojo::File bitrot
  - Bump Mojolicious::Plugin::AssetPack to 1.42
  - Fixes #69		 18:29
TimToady you call it a configuration "script", but do you ever run it by itself? 18:29
maybe it's just a module
in which case plain old "use" should work 18:30
TimToady (maybe with a "use lib '.';" or some such to tell it where to look 18:31
)
eater I could make it a module, but I rather make it just an script with everything preset for creating tasks in that case
but if that isn't possible
I can make it a module
TimToady a module is a kind of script 18:31
TimToady I don't understand what you mean when you say "just a script" 18:32
what are you expecting a script to do that a module can't do? 18:33
eater true 18:33
I'm not expecting it to be able to do more
but I would like more control over how it's run 18:34
Geth doc: 1f7cc4efa0 | (Christopher Bottoms)++ (committed by Zoffix Znet) | doc/Type/Callable.pod6
Added multi-positional use of assuming (#1296)

Borrowed content from brian d foy's Stack Overflow post stackoverflow.com/a/43669837/215487
eater define import modules and methods before it's run
s/define//
kyclark Given a hash my %h = :one(1), :two(2), :three(3);, how could I sort first by length of key then by key? E.g., so I get "one two three"?
gfldex eater: you can do that with require
kyclark m: my %h = :one(1), :two(2), :three(3); say %h.keys.sort(*.chars || *.Str)
camelia (two one three)
TimToady if you want to define methods, then it's probably a class 18:35
[Coke] .tell zoffix please check my comment on doc pr # 1296 18:35
TimToady (or a role, if it's supposed to be generic)
yoleaux [Coke]: I'll pass your message to zoffix.
Geth doc: 7e84db63a2 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Type/Callable.pod6
Move ref to SO to license
TimToady but classes and roles are also just modules/scripts that you pull in with 'use' or 'require'
gfldex m: my %h = :one(1), :two(2), :three(3); say %h.keys.sort(*.Str).sort(*.chars)
camelia (one two three)
gfldex no wait, that's silly 18:36
TimToady afk for a bit &
Geth doc: e8ae3b2521 | (Zoffix Znet)++ (committed using GitHub Web editor) | README.md
Move ref to SO to LICENSE section
 18:37
MasterDuke_ m: my %h = :one(1), :two(2), :three(3); say %h.keys.sort({$^a.chars <=> $^b.chars || $^a cmp $^b})
camelia (one two three)
gfldex m: my %h = :one(1), :two(2), :three(3); say %h.keys.sort({$^a.chars cmp $^b.chars || $^a.Str cmp $^b.Str}) 18:38
camelia (one two three)
[Coke] Zoffix++ # thanks. 18:38
timotimo kyclark: you can just return a little list of two results 18:43
be sure to only take one argument, otherwise it'll make a comparator function out of it
i.e. try { $^e.chars, $^e.Str }
but not perl6-try, but more like english-try, or "attempt" or something
araraloren night, o/ 18:44
gfldex m: my %h = :one(1), :two(2), :three(3); say %h.keys.sort({$^a.chars, $^a.Str})
camelia (one two three)
kyclark timotimo, that works, but I don't understand why! What is the $^e? 18:45
timotimo { $^e.foo, $^e.bar } is equivalent to -> $e { $e.foo, $e.bar }
kyclark Default arg like $^a/$^b?
timotimo not exactly, but similar
kyclark Is "e" different from a/b?
timotimo yes 18:46
e is e, but a is a and b is b
you can use any name you want
gfldex timotimo: is a list return value for the callable speced (in `sort`)
?
timotimo $^foo, $^meh, $^a, $^OMG-WTF-BBQ
kyclark Just don't use a/b?
timotimo for every different $^foo you put you'll get one entry in the signature for the block that contains the $^foo
gfldex kyclark: alphabetical order is taken into account for autoarguments 18:47
timotimo the entry will be for $foo instead of $^foo, so if you have an innerer block you can just refer to $foo so it won't make a signature for the innerer block
aye, the argument will be sorted in the resulting signature
gfldex sub f(){ say $^b, $^a }; f(1,2) 18:48
m: sub f(){ say $^b, $^a }; f(1,2)
camelia 5===SORRY!5=== Error while compiling <tmp>
Placeholder variable '$^a' cannot override existing signature
at <tmp>:1
------> 3sub7⏏5 f(){ say $^b, $^a }; f(1,2)
timotimo yeah, don't have () there
gfldex m: sub f { say $^b, $^a }; f(1,2)
camelia 21
gfldex a list as return value for the callable of sort is ENODOC
gfldex goes to fill the form 18:49
timotimo gfldex: it's just sorting the resulting lists
it's just a schwartzian transform of one element to a little list for each of the elements
gfldex true but not exactly obvious
timotimo feel free to point it out in the docs
Geth doc: 5dfc1effe3 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/operators.pod6
Add ...^ and …^ ops to index

Fixes #1292		 18:51
eater okay 19:07
apperently I was looking for EVALFILE
Geth doc: 6e1482811b | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Type/Any.pod6
Remove needless prose in .flat

  - Flat returns a Seq with an iterator fed by original stuff, so it solving any
   halting problems is irrelevant.
  - Flat used to have a bug where it was failing to propagate .is-lazy in certain
   cases, causing full reification, which might be what the original author hit
   upon, to surmise there are any issues with flattening infinite things.
  - Fixes #1202		 19:10
timotimo eater: please be aware that when you EVALFILE something the lexical variables you have in those files will not be available outside of the EVALFILE 19:17
eater timotimo: sure, thanks :) 19:19
masak moritz: by the way, the blog post about "infix:<< > >> should be outlawed" (featured in this week's Perl Weekly) felt to me like a good example of "I can't think of a use case for this feature, so let's ban it" 19:20
moritz masak: indeed 19:22
masak: my main thought on the topic is that comparisons often include on relatively fixed, known quantity
masak: and the other one that's being compared 19:23
timotimo so never use > again, always use !< ?
moritz masak: and I like to put the lesser-known quantity on the LHS
if $attempts > $threshold { die "nope, not workin'"; }
moritz timotimo: the author argued that you should use $b < $a instead of $a > $b 19:24
timotimo huh
we can outlaw > and force people to use R< for that :P
moritz timotimo: llewellynfalco.blogspot.de/2016/02/...gn-in.html 19:25
timotimo: re R< sure :-)
moritz though of course in his example, we'd just write 5 < x < 10 :-) 19:26
or 10 > x > 5, neither of which is confusing, IMHO
or even x ~~ 5^..^10
gfldex "The way things work in Llewellyn's world" this title clearly states the intend of the author 19:27
this is brilliant too: llewellynfalco.blogspot.de/2017/01/...esign.html 19:30
timotimo both $low < $x < $high and $x ~~ $low..$high are helpful for this
Geth doc: ef19149a8b | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Type/Any.pod6
Improve C<flat> docs

  - use more detailed examples that show actual flattening
  - include that Maps get flattened too, since they're Iterable
  - Get rid of broken and convoluted `deepmap` example; it destroys Maps and Pairs		 19:30
gfldex you don't even need to read the post, the first two lines give if away
timotimo what constitutes a line on your end? 19:31
gfldex err sentence (I got 3 hours sleep last night) 19:32
timotimo ah 19:32
Zoffix hoelzro: Sent you a PR fixing IO::String for latest Rakudo's: github.com/hoelzro/p6-io-string/pull/7 21:24
hoelzro just saw it - thanks! will look in depth after work
Zoffix ^ that's used by our doc site's `make xtest`, so if anyone were having trouble with it, you can run `zef --force install github.com/zoffixznet/p6-io-string...x-say.zip` in the meantime 21:25
hoelzro ok 21:26
Geth doc: 5b1627785e | (Zoffix Znet)++ | 8 files
Fix xtest
 21:58
Geth doc: 8da79bb26e | (Zoffix Znet)++ | doc/Language/pragmas.pod6
Document `fatal` pragma

Fixes #1275		 23:19
Geth doc: 3ba0c02390 | (Samantha McVey)++ | lib/Pod/Htmlify.pm6
Append .html to pages ending in a dot

You are not supposed to end URL's with a dot		 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!