»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.
timotimo i'm not going to take a chance on that 00:00
teatime timotimo: wait, when I said Red and Blue, I only meant to choose two random colors, i.e. items with qualitative rather than quantitative differences.
I hope I didn't accidentally reference something bad.
00:00 Kyo91 left
teatime (like Pokémon) 00:00
TEttinger teatime: it's more the two-party system in US politics
timotimo i'd prefer not to spoil your innocence 00:01
or faith in humanity or whatever
teatime TEttinger: Dear god man, you had to keep talking about it until I looked, didn't you?
TEttinger I told you!
I was trying to suggest why you shouldn't look!
00:02 xaltotun joined
TEttinger well since your innocence gauge has dropped to 0, red is the party color for the US Republican party, blue for the US Democratic party, so red and blue would have political connotations 00:03
chartreuse and taupe, less so
(especially because you can't really find a taupe flag in use, because no one can settle on what taupe looks like) 00:04
00:05 Guest18 joined
timotimo TEttinger: that's not the connotation i was going for at all 00:09
anyway, let's forget about it 00:10
TEttinger oh, the halo thing 00:11
Perl Behemoth and Perl Leviathan 00:12
(referring to the evenly-matched great beasts from the biblical Book of Revelations, also the scope of both languages is large) 00:13
which fits with the Apocalypses term used in early Perl 6 lit, right?
geekosaur except "apocalypse" was used more in the technical sense 00:14
00:23 Guest18 left 00:27 Guest18 joined
titsuki zef doesn't show the alternatives when installing Term::termios by "zef install Term::termios": gist.github.com/titsuki/a5ece19802...1a3165ebd9 00:35
Should I delete my module?
00:42 Rawriful left 00:56 pecastro left 01:01 markong left 01:02 Cabanossi left
titsuki ok. I reported the above behavior: github.com/ugexe/zef/issues/227 01:03
01:03 jeek left, jeek joined 01:05 Cabanossi joined 01:06 aborazmeh joined, aborazmeh left, aborazmeh joined
psch fwiw, behemoth and leviathan are coded as first and second in the (really quite amazing) webserial Worm 01:15
so that's still at least some cultural remnant that makes these two ordered 01:16
...but then they're both city busting monsters vOv
teatime psch: I really like The Zombie Knight. it isn't finished though and I ran out of pages again this morning. 01:17
psch teatime: i think i read the first ~5 chapters or so, didn't really grip me 01:18
teatime fair enough
psch teatime: i did enjoy super powereds a fair bit, until there was the reveal that something big had been going on and then the backlog was gone and i didn't keep up with the weekly chapter releases... :S 01:20
midnight moonlight was really great too, and i'm really upset that the author had to drop it for the time being 01:21
01:21 Thrush joined
teatime I'm scared that will happen to zombie knight. He disappears occassionally for months at a time, presumably eventually he'll not come back. 01:21
psch well, upset that the event that made her drop it happened, to be clear
teatime ah 01:22
Thrush Hello! What's the best way to compare two boolean variables in Perl6? Should I use "$var1 == $var2" or "$var1 eq $var2" or something else?
psch Thrush: to be clear, you want True when both are True or both are False? 01:23
Thrush Psch: Last week I needed to check two booleans, when they agreed. I can't remember the exact context, though, but I remember that an xor would not have worked. 01:25
01:26 cgfbee left
psch Thrush: that sounds like an interesting special case, which probably the context would reveal the details of :) 01:26
gfldex Thrush: see docs.perl6.org/type/Junction
psch Thrush: in general, &infix:<==> should be fine 01:27
m: say True == True, True == False, False == False
camelia TrueFalseTrue
psch Thrush: i am definitely curious why xor/&infix:<^^> didn't work though
teatime psch: I assume the worry is that they may not be actual booleans, just something that can coerce to bool
psch oh 01:28
teatime you could so($a) == so($b) or whatever
but presumably there's a cleaner way?
01:28 Zoffix joined
Zoffix m: gist.github.com/zoffixznet/8391d05...840c2bc6ca 01:28
camelia Glass half full
Glass half full
Glass half empty
Zoffix m: gist.github.com/zoffixznet/8391d05...840c2bc6ca
camelia Glass half full
Glass half full
Glass half full
Zoffix Weird. It's just three calls to nqp::rand_n, but it's *always* producing `full` for the first two prints. 01:29
Thrush pcsh: True == True worked for me, but I'm concerned that == is for numbers, which eq is for strings. (True eq True also worked for me.)
Zoffix What am I doing wrong?
01:29 wamba left
Thrush which => while 01:29
psch m: say Bool.^mro
camelia ((Bool) (Int) (Cool) (Any) (Mu))
gfldex m: say so True & True;
camelia True
psch Thrush: Bool isa Int
gfldex if you care about true boolean values use a Junction 01:30
psch m: my $a = 0 but True; say $a & $a
camelia 0
0
psch m: my $a = 0 but True; say so $a & $a
camelia True
psch still have to coerce though
teatime Zoffix: did you know that you only have the string 'Glass half empty' one time in that script? 01:31
01:31 cgfbee joined
teatime gfldex: there's not an all-or-none() junction tho 01:31
psch .oO( one(all(), none()) ) 01:32
definitely needs a builtin
>_>
Zoffix teatime++ thanks
teatime Zoffix: sometimes you need a simple mind to find a simple problem :) 01:33
Zoffix :) 01:34
psch Thrush: in any case, &infix:<==> should work unless your operands have their own .Int
Thrush: well, or they inherit one somewhere i suppose
Thrush: alternatively you can use the wonderfully concise and comprehensible infix operator [&({$^a.so eqv $^b.so})] 01:35
...you probably shouldn't though, instead maybe implement .Bool on your classes and then use &infix:<^^>
i'm making bad jokes, i probably shouldn't have logged in anymore, given the time and all, g'night~ 01:36
Zoffix m: for True, True, False, False, True, False -> $a, $b { say ($a + $b) %% 2 } 01:37
camelia True
True
False
Zoffix m: for True, True, False, False, True, False -> $a, $b { say ?$a === ?$b }
camelia True
True
False
01:38 Zoffix left
gfldex .oO( One needs to dig deep into the typesystem to tell what is true. ) 01:40
the docs are quite thin in that area 01:42
they don't really tell where a True is comming from. They just state: There is an enum of True and False. Good luck! 01:43
I think I better practice with a blogpost first before I take that topic on in the docs. 01:44
Xliff If I have a class that does List, can I override methods? 02:00
m: class A { also does List; method first(+@a) { say "Hi!" } } 02:01
camelia 5===SORRY!5=== Error while compiling <tmp>
List is not composable, so A cannot compose it
at <tmp>:1
Xliff m: class A { also is List; method first(+@a) { say "Hi!" } } 02:02
camelia ( no output )
Xliff m: class A { also is List; method first(+@a) { say "Hi!" } }; A.new.first;
camelia Hi!
Xliff Ah!
02:15 Zoffix joined 02:32 cdg joined
Zoffix New core hacking blog post: "Perl 6 Core Hacking: QASTalicious": rakudo.party/post/Perl-6-Core-Hack...STalicious 02:33
02:37 Zoffix left, cdg left 02:46 ilbot3 left 02:48 itaipu left, Thrush left 02:57 ilbot3 joined, ChanServ sets mode: +v ilbot3 03:07 aborazmeh left 04:25 Cabanossi left 04:27 Cabanossi joined 04:47 eliasr left 05:12 ramortegui joined 05:20 aborazmeh joined, aborazmeh left, aborazmeh joined 05:31 Guest18 left
buggable New CPAN upload: Terminal-Getpass-0.0.5.tar.gz by TITSUKI cpan.metacpan.org/authors/id/T/TI/...0.5.tar.gz 05:35
05:36 opollen joined 05:49 khw left 05:55 opollen is now known as mpollen 05:57 cdg joined, aborazmeh left 05:59 mpollen is now known as lpollen 06:02 cdg left 06:07 lpollen left 06:29 darutoko joined 07:01 Eddward joined
Eddward Where's the best place to open a github issue for the latest star? I'd think rakudo, but last time I saw a problem like this the right answer was moarvm. 07:04
# generate sample input
cat > /tmp/sample.input <<EOF
Musopen%20DVD/Brahms%20-%20Symphony%20No%201%20in%20C%20Major/Symphony%20No.%201%20in%20C%20Minor,%20Op.%2068%20-%20IV.%20Adagio%20-%20PiuC%80%20andante%20-%20Allegro%20non%20troppo,%20ma%20con%20brio.mp3
EOF
# generate sample output
perl -E 'print chr for 77,117,115,111,112,101,110,32,68,86,68,47,66,114,97,104,109,115,32,45,32,83,121,109,112,104,111,110,121,32,78,111,32,49,32,105,110,32,67,32,77,97,106,111,114,47,83,121,109,112,104,111,110,121,32,78,111,46,32,49,32,105,110,32,67,32,77,105,110,111,114,44,32,79,112,46,32,54,56,32,45,32,73,86,46,32,65,100,97,103,105,111,32,45,32,80,105,117,204,128,32,97,110,100,97,110,116,101,32
,45,32,65,108,108,101,103,114,111,32,110,111,110,32,116,114,111,112,112,111,44,32,109,97,32,99,111,110,32,98,114,105,111,46,109,112,51,10' > /tmp/sample.output
# attempt translate 07:05
perl6 -e 'S:g/"%" ( <xdigit>**2)/{ :16(~$0).chr }/.say given "/tmp/sample.input".IO.slurp()' | diff -sq - /tmp/sample.output
# legacy translare
perl -MFile::Slurp -e 'print(read_file("/tmp/sample.input") =~ s/%([0-9A-Fa-f]{2})/chr(hex($1))/egr)' | diff -sq - /tmp/sample.output
That didn't work. The generated /tmp/sample.output was wrong when I copied it out of irc. 07:06
teatime umm, Eddward 07:08
there are better ways to paste (use a paste site)
Eddward yeah. Sorry. I'm up too late. Let me do that right, 07:09
This works. pastebin.com/3xEHcLD5 07:11
The perl6 version results in "Files - and /tmp/sample.output differ". The perl5 gets (the expected) "Files - and /tmp/sample.output are identical" 07:13
I'm probably doing it wrong but I've never found the right way. 07:14
07:22 Zoffix joined
Zoffix Eddward: the difference in in your Perl 5 version, you're just working on bytes, while in Perl 6 version you're working on utf8-decode and normalized strings. 07:23
teatime hrm, I'm not so sure 07:24
Eddward right.
teatime in your sample input you have, at one point, "Piu%CC%80"
What should that become?
Eddward The target has an encoding that perl6 considers "wrong" but that's the name of a file on disk.
Zoffix Eddward: the first problem you can solve by just using a module, e.g. `use URI::Encode; "/tmp/sample.input".IO.slurp.&uri_decode.say`, but you'll still have the issue with normalization, unless you work with bytes in Perl 6 as well and IIRC there's no simple way to do so right now (like, no same convenince as just running a regex on a bunch of bytes) 07:25
Eddward: what's the encoding?
Eddward Musopen DVD/Brahms - Symphony No 1 in C Major/Symphony No. 1 in C Minor, Op. 68 - IV. Adagio - Piu􏿽xCC􏿽x80 andante - Allegro non troppo, ma con brio.mp3
teatime Eddward: No, I don't mean post the output, I mean reason about what it should become.
I think you're working in UTF-8 in Perl6, but not in Perl5.
(but yeah, the sane thing to do is use a decoding module.) 07:26
Eddward I don't know the encoding specifically. The are multiple ways to represent the accented u.
teatime oh, hrm, I take that back. 07:27
Eddward I'm trying to read it from a database and open it in the file system,
teatime 0xCC 0x80 would indeed be a combining grave 07:28
Eddward I think perl5 is just pushing bytes, I don't know how to do that in p6.
teatime that doesn't seem right to me...
yeah, I guess Zoffix was right. My apologies, Zoffix :) 07:29
Eddward URI::Encode doesn't handle other URLs right. (or rather it handles them right but mono doesn't do it right when creating the database. 07:31
07:32 ramortegui left
Eddward The only problem I see when I use the regex to decode the url is the normalizing. 07:33
I think think the source db has urls with + or something in them. URI::Encode turns them into spaces. 07:34
teatime well, when you do chr() separately on CC and 80 you get 2 characters, an I with an accent U+00CC, and an illegal character U+0080
Zoffix m: say "u\x[300]".chars 07:35
camelia 1
Eddward It's a legal filename.
teatime any series of bytes is a legal filename
as long as it doesn't have / or NUL. at least on Unix. 07:36
BUT, that only happens in your Perl6 version.
Eddward I didn't create the file or the db.
teatime chr() in Perl5 seems to do something... different.
Eddward So no easy way to do this for now?
teatime ah, I'm sure you can do it.
Zoffix Eddward: "this" being what? Are you just trying to read the file or are you doing the decoding too? 07:37
Eddward: are the names percent-encoded in the DB?
Eddward Read input data like the sample and produce the expected sample.output. 07:38
Yes. They are file:// url. It trimmed the input to simplify it. 07:39
It's a sqlite3 db produced by banshee. I'm just trying to port my script for exporting playlists. 07:40
07:41 darutoko left 07:42 darutoko joined
Eddward I had this old issue I was never able to figure out : github.com/MoarVM/MoarVM/issues/525 07:42
I saw p6 was getting faster and thought I'd give it another try. 07:43
teatime "call uri_unescape() (we get 2 byte 'ù' here)" that's normalization again 07:45
Zoffix Eddward: lemme try something
Eddward ok. Thanks
07:46 darutoko left
teatime "I wonder if I should be using some datatype other than Str to store path data since Str seems to enforce rules that the filesystem does not." Yes, exactly. 07:46
Eddward I'm using 2017.10 & 6.c
teatime How come you dropped the utf8-c8 solution? afaik that's still valid. 07:49
specifically uri-decode($string, :enc('utf8-c8')) 07:50
Eddward I never got it to work.
I'll try again. 07:51
I don't really understand what utf8-c8. Will that work for any byte string or will it fail with some other encoding? 07:52
teatime it exists almost entirely because normalization breaks filenames :) 07:53
07:53 wamba joined
Zoffix It won't work. utf8-c8 is just utf8 except instead of exploding on invalid utf8, it will shove the bad bits as as is 07:53
teatime "UTF-8 Clean-8 is an encoder/decoder that primarily works as the UTF-8 one. However, upon encountering a byte sequence that will either not decode as valid UTF-8, or that would not round-trip due to normalization, it will use NFG synthetics to keep track of the original bytes involved. This means that encoding back to UTF-8 Clean-8 will be able to recreate the bytes as they originally existed." 07:54
Eddward ok
Zoffix Oh
teatime docs.perl6.org/language/unicode
Zoffix m: say buf8.new(0xCC, 0x80).decode: 'utf8-c8' 07:55
camelia rakudo-moar 9bc26ec3c: OUTPUT: «̀␤»
Zoffix m: say buf8.new(0x75, 0xCC, 0x80).decode: 'utf8-c8'
camelia u􏿽xCC􏿽x80
Zoffix m: say buf8.new(0x75, 0xCC, 0x80).decode('utf8-c8').chars
camelia 3
Eddward perl6 -MURI::Escape ?
Zoffix teatime: so uri-decode :enc works? 07:56
teatime Zoffix: samcv says it does in the ticket from a year ago
Zoffix Ah, cool
teatime I'm at a bit of a loss for how to plug the pieces together properly, though.
Eddward I never got it to work.
The behavior changed, but I just got a different encoding of the bytes. 07:57
Looks like I need to install URI::Escape. It's not panda any mor, right? 07:58
Zoffix zef install URI::Escape
Eddward thanks 07:59
Says it is installed already.
Zoffix Eddward: yeah, teatime++'s way works. Just need to write to STDOUT in bin mode: $ perl6 -e 'use URI::Escape; $*OUT.write: "/tmp/sample.input".IO.slurp.&uri-unescape(:enc<utf8-c8>).encode("utf8-c8")' | diff -sq - /tmp/sample.output
Files - and /tmp/sample.output are identical
Eddward cool! Thanks! 08:00
I'll wrap that into my p6 version of the export script. I want to see how it performs. It was slow a year ago.
teatime Also, I feel better seeing that samcv was confused for a moment in the same way I was by 0xCC 0x80 :) 08:01
Eddward It sounds like it's a screwy file. I just hate editing originals though. 08:02
samcv hello :)
teatime oh, hey. Sorry to ping you :)
samcv does anyone have a question though :)
teatime Eddward: which?
samcv: not anymore I don't think :)
samcv aww.
if you have one go ahead and let me know :)
teatime samcv: your comments on a ticket lead us to solve it.
Eddward I dug up an old problem.
samcv ah 08:03
teatime samcv++
samcv oh URI::Escape?
teatime yes
samcv yeah i added utf8-c8 capabilities to that module :) you're welcome!
someone else had that issue a while ago and so i took the steps to add that functionality in. since i deemed it pretty essential
Eddward Once I got an updated star it didn't work. Looks like I missed the binary output. 08:04
teatime Eddward: Your filenames on disk aren't wrong. Your database is kindof annoying, but not wrong per-se.
and as you say, it's the data you have.
Eddward samcv I was me. I never got it to work before. I'm getting back to it now. Sorry to be a pest. 08:05
samcv ah
Eddward Looks like we have it worked out.
samcv no worries
teatime Eddward: there's just more than one correct utf8 encoding for most strings, and filenames are in bytes, not any specific normalization etc.
Eddward Thanks again everyone
samcv yep filenames are decoded with utf8-c8 always since we can't assume anything
teatime Zoffix: Did you ever figure out the answer to the question you had that day... it wasn't actually about receiving multiple ^D from stdin, but that was part of the question/example. 08:08
I ended up researching and getting re-educated about ^D, which is always welcome, and now I'm curious about the original question. I can find it in the log probably if you don't recall. 08:10
Zoffix teatime: irclog.perlgeek.de/perl6/2018-01-20#i_15714540 08:13
08:14 Zoffix left
teatime well, that's the one :/ 08:15
08:25 Zoffix joined
Zoffix Eddward: FWIW this version is benching to be 2x faster and doesn't need a module: gist.github.com/zoffixznet/7d0aaef...5477b71ee9 08:25
Eddward cool 08:26
I'd need to adapt it. 08:27
At the moment I'm trying to figure out "No such method 'IO' for invocant of type 'Blob[uint8]'" 08:30
$fh.write: $sth.allrows.map( *.subst: /^ 'file://' /, '' ) 08:31
.map( { S :g / '+' /{uri-escape('+')}/ } )
.map( *.&uri-unescape(:enc('utf8-c8')).encode("utf8-c8") )
.map( *.IO.relative($library_root.IO.absolute) )
.map( * ~ "\n" )
.join;
Zoffix Eddward: remove `.encode("utf8-c8")` 08:32
Eddward I can probably get away with moving the rel->abs before the url unescape since I know the prefix shouldn't be escaped.
Type check failed in binding to parameter '$buf'; expected Blob but got Str ("mod/demo/skaven/seve...) 08:33
encode at the end?
08:33 ZzZombo joined
Eddward That ran... 08:34
Zoffix That's a lot of .maps... why not write a single one? 08:35
Eddward I tried that. Broke it a apart to debug and because I don't really know what I'm doing. 08:36
I'll try to collapse them.
btw, the p6 result watches the p5 result for my entire music library now. 08:37
Zoffix Eddward: also `.subst: :g, "+", "%2b"` will be orders faster than using a regex and sub call
Eddward err watches=> matches
Zoffix * instead of { S :g / '+' /{uri-escape('+')}/ } I meant
Also, I imagine `.starts-with 'file://' ?? .substr: chars 'file://' !! $_` would be loads faster than `*.subst: /^ 'file://' /, ''`. And if they all start with `file://` just use .substr without the conditional 08:38
Eddward subst: :g, "+", "%2b" was actually slower. 0m29.627s v 0m32.654s 08:39
I'll work on collapsing first. 08:40
Zoffix 0_0
ZofBot: IMPOSSIBRU!
ZofBot Zoffix, 'Sixty-fower, sixty-fower, sixty-fow--' The kitchen door opens, and Juliet emerges
ZzZombo "Cannot put a type constraint on an 'our'-scoped variable"
why is that O_o
Zoffix m: my $s = "foo+bar+ber" ~ rand; for ^1000_00 { s:g/ '+' /{'%2B'}/ }; say now - INIT now 08:42
camelia 7.9058585
Zoffix m: my $s = "foo+bar+ber" ~ rand; for ^1000_00 { .subst: :g, '+', '%2B' }; say now - INIT now
camelia 1.38105620
Eddward I'm doing something stupid...
$fh.write: $sth.allrows.map( *.subst: /^ 'file://' /, '' 08:43
.subst: :g, "+", "%2b"
.&uri-unescape(:enc('utf8-c8'))
.IO.relative($library_root.IO.absolute)
~ "\n" )
.join.encode("utf8-c8");
Malformed postfix call (only alphabetic methods may be detached)
I need parens somewhere.
Zoffix Yes, everywhere you've used `:` instead of `()` for a method call
Eddward cool 08:44
.⏏&uri-unescape(:enc('utf8-c8')) 08:45
Still doesn't like this.
Zoffix Eddward: oh, right
Eddward Can't use method syntax? 08:46
Zoffix Eddward: can't break up method chains that contain non-alphabetic methods
like &uri-unescape
Eddward So one line, no spaces? 08:47
$fh.write: $sth.allrows.map( *.subst(/^ 'file://' /, '') 08:48
.subst(:g, "+", "%2b")
.&uri-unescape(:enc('utf8-c8'))
.IO.relative($library_root.IO.absolute)
~ "\n" )
.join.encode("utf8-c8");
$fh.write: $sth.allrows.map( *.subst(/^ 'file://' /, '').subst(:g, "+", "%2b").&uri-unescape(:enc('utf8-c8')).IO.relative($library_root.IO.absolute) 08:49
~ "\n" )
.join.encode("utf8-c8");
that works
ugly, but works.
08:50 rindolf joined
Eddward 0m31.993s 08:50
08:53 dogbert11 joined 08:55 setty1 joined, dogbert17 left
Zoffix Eddward: you don't need to write it all in one line. Just don't put whitespace between method call dots (or use unspace (`\`)). gist.github.com/zoffixznet/5f8e4f6...dc9463cb54 08:57
08:57 Zoffix left
Eddward got it 08:58
09:00 syntaxman left 09:02 cgfbee left, syntaxman joined 09:03 syntaxman is now known as Guest79180
Eddward A little faster ... 0m25.846s 09:07
09:07 statisfiable6 left, statisfiable6 joined, ChanServ sets mode: +v statisfiable6 09:08 cgfbee joined 09:20 parv joined 09:23 greppable6 left, greppable6 joined 09:24 ZzZombo left
AlexDaniel reportable6: uptime 09:28
reportable6 AlexDaniel, 1 day, 17 hours, 20 minutes, and 16 seconds, 1566.617188MiB maxrss. This is Rakudo version 2018.01-29-ga2499c90f built on MoarVM version 2018.01 implementing Perl 6.c.
AlexDaniel uhhh…
09:28 reportable6 left, bisectable6 left, unicodable6 left, reportable6 joined, ChanServ sets mode: +v reportable6, bisectable6 joined, unicodable6 joined 09:29 ZzZombo joined 09:36 lasse_ joined 09:37 Eddward left
lasse_ Hi, I have a PHP job scheduling system. I want to call perl6 from it. PHP arrays carring needed info, Those can be serialized or var_exported and written to disk before perl6 invokation and read back after return. Can anyone advice how to do this? 09:38
09:53 Zoffix joined
Zoffix lasse_: do you have PHP end sorted? What about using json_encode() on the data and sending it to perl6 to STDIN, and then reading it from STDOUT. Then on perl6 end, you'd do `use JSON::Fast; my @data := from-json $*IN.slurp; .... process it....; put to-json @result' 09:57
m: use QAST:from<NQP>; BEGIN $*W.compile_time_evaluate: $, QAST::Op.new: :op<call>, :name<&say>, QAST::Op.new: :op<call>, QAST::Block.new: QAST::SVal.new: :value<meows> 09:59
camelia meows
Zoffix Is it mandatory that QAST::Blocks be declared somewhere and then cloned and p6capturelex'ed into a QAST::WVAL? Is the above call fine? (also, does it need p6capturelex?) 10:00
m: use QAST:from<NQP>; BEGIN $*W.compile_time_evaluate: $, QAST::Stmts.new: QAST::Op.new(:op<bind>, QAST::Var.new(:name<$foo>, :scope<lexical>, :decl<var>), QAST::SVal.new: :value<meows>), QAST::Op.new: :op<call>, :name<&say>, QAST::Op.new: :op<call>, QAST::Block.new: QAST::Var.new: :name<$foo>, :scope<lexical> 10:02
camelia meows
lasse_ What do you mean by PHP end sorted? Yes I can Json_encode. Bus how do i invoke Perl6 and make it read my JSON?
Zoffix Seems to work, but I wonder if it's doing too much work with QAST::Block 10:03
lasse_: just like any other program. ##php would be able to give more competent advice than me, but quickly glancing at the docs, looks like proc_open() can get the job done: php.net/manual/en/function.proc-ope...ample-4258 10:05
lasse_: you'd just `fwrite($pipes[0], json_encode($the_stuff));` and then `$result = stream_get_contents($pipes[1]);` to get the results 10:06
+ `fclose($pipes[0]);` in between :)
PHP++ # easy to program-by-googling
10:07 lookatme left, lookatme joined
lasse_ Tx Zoffix, I try it. Now I just have to learn Perl 6:) 10:09
Zoffix :) 10:10
lasse_: this seems to work (reverses the array and uppercases the stuff in it): gist.github.com/zoffixznet/ef6dd97...f0e9e1d5d6 10:20
Hope that helps \o
10:20 Zoffix left 10:36 ZzZombo left 10:37 ZzZombo joined 10:43 leah2 left 10:48 shinobi-cl left 11:05 lookatme left 11:06 lookatme joined 11:08 markong joined 11:11 quotable6 left 11:13 lookatme left 11:14 lookatme joined 11:17 Grauwolf left 11:18 Grauwolf joined
Geth doc: 9292ce5b45 | (JJ Merelo)++ | t/pod6.t
Refactor to turn into a single test

And maybe it should be merged into another file to make tests a bit faster.		 11:27
11:34 wamba left 11:35 quotable6 joined 11:36 itaipu joined 11:37 setty1 left 11:39 eliasr joined 11:40 quotable6 left 11:42 quotable6 joined 11:46 quotable6 left 11:51 wamba joined, lookatme left 11:52 lookatme joined 12:10 parv left 12:12 lookatme left 12:14 quotable6 joined, ChanServ sets mode: +v quotable6 12:21 lookatme joined 12:36 |oLa|1 left 12:43 rindolf left
ZzZombo <ZzZombo> "Cannot put a type constraint on an 'our'-scoped variable" 12:57
<ZzZombo> why is that O_o
moritz 'our'-scoped variables are globals, so no one piece of code has authority over it 12:59
ZzZombo what do you mean global?? 13:00
moritz m: { our $x = 42 }; { say our $x }
camelia 42
moritz our $x creates a lexical $x that is an alias for $x in the current namespace 13:01
ZzZombo Anyway, I don't get it why does that change anything regarding "no one piece of code has authority over it". 13:02
moritz change from what? 13:03
ZzZombo Err? That shouldn't change anything, and constrained `our` variables should work like any other type of them. 13:04
El_Che that's the point of being global 13:05
moritz ZzZombo: so, if you have a lexical variable "my Int $x = 42", there can be only one declaration of it, and its type is clear 13:06
but, globals aren't like that, tied to a lexical location of your program
13:07 shinobi-cl joined
moritz they just come into existence whenever somebody references it 13:07
which access determines its type, then?
multiple declarations or run-time reads and writes could have conflicting or more narrow or more specific types 13:09
13:09 markong is now known as supernaut, supernaut is now known as suprnaut
moritz and in the end, globals are simply not such a good idea for most purposes, so it doesn't seem to be worth it to come up with an elaborate scheme for dealing with it all 13:09
13:10 suprnaut is now known as markong 13:14 ChoHag joined
ZzZombo Sorry, I don't follow you. I have a package X with a variable `our y`, to use it I need to import the said package first, while to use a true global variable I just reference it. 13:15
And there can only be one package named such. 13:16
moritz I don't understand your distinction between a package variable a "true global variable"
m: say $X::y 13:17
camelia (Any)
13:17 jeromelanteri joined
moritz see? I could reference that before doing any use or import 13:17
before any type constraint for it could even be known
m: $Just::discovered = 42; 13:18
camelia ( no output )
ZzZombo Why does that not fail the same way as accessing undeclared `my` variables?
moritz if i run that, and *then* load module Just, which declares "our Array $discovered"; what then?
13:18 p6lert left
moritz because it's a global 13:18
and globals don't need declarations
ZzZombo ugh
moritz only their lexical aliases 13:19
if you don't like 'em, don't use 'em :)
ZzZombo except I have to?
13:19 p6lert joined
ZzZombo what kind of response is that? 13:19
moritz what do you need them for?
ZzZombo I want to store some state in my package.
moritz there are lots of ways to do that 13:20
singletons come to mind, for example
ZzZombo Singletons are for classes, right? 13:21
moritz and classes are a nice way to store structured, type-safe data
which seem to do want to do
ZzZombo No, I don't need nor want any initialization/clean up that using objects brings. 13:22
moritz I don't think I can help you, nor do I feel this discussion is being productive for anybody 13:23
ZzZombo what
fuck this shit as well, holy crap, what a load of bullshit was that?? 13:24
13:29 shinobi-cl left 13:33 camelia left 13:35 jeromelanteri left 13:41 leah2 joined 13:46 leah2 left 13:57 shinobi-cl joined, Eddward joined
teatime Eddward: Welcome back! 13:58
14:02 aborazmeh joined, aborazmeh left, aborazmeh joined 14:03 moritz joined 14:09 aborazmeh left
Eddward teatime: Good morning! 14:10
14:22 Guest79180 left 14:23 syntaxman joined, cgfbee left, syntaxman is now known as Guest77349 14:24 Eddward left 14:29 cgfbee joined 14:42 itaipu left 14:54 itaipu joined 14:55 Ven`` joined 14:57 pecastro joined 15:01 cdg joined 15:02 itaipu left 15:03 cdg left, cdg joined 15:06 rindolf joined 15:07 cdg left
shinobi-cl hi all... what is the X:: exception for Out Of Memory? I could not find it in docs.perl6.org 15:23
El_Che shinobi-cl: does it exist? 15:24
timotimo shinobi-cl: when you run out of memory, moarvm just exits
El_Che Out of memery is mostly too late
o
shinobi-cl oh, ok. 15:25
timotimo an out-of-memory situation is exceedingly difficult to properly recover from
15:25 BuildTheRobots joined
El_Che most likely the kernel will kill your process 15:25
timotimo especially when you're dealing with multiple threads; we always have the spesh thread
oh, hold on
m: my $foo = buf8.allocate(999999999999); CATCH { .perl.say }
evalable6 (exit code 1) MoarVM panic: Memory allocation failed; could not allocate 999999999999 bytes
timotimo hm, ok 15:26
i thought perhaps we throw that kind of error
i mean, when you're requesting 1 gig of memory and the system can't give you what you want, that doesn't mean that the next 8 byte allocation will also fail
m: say 999999999999 / 1024 / 1024 / 1024 15:27
evalable6 931.32257461455
timotimo in this case, though, i was requesting almost a terabyte of ram
shinobi-cl I have a big table, and the user has the chance to request creation of indexes for all the columns if he/she wants to. I was thinking on returning a False if the index ended being too big for RAM. 15:29
timotimo you can check if asking for a given chunk of memory would succeed by NativeCalling into malloc (and of course free right after that) and deciding based on the result 15:30
shinobi-cl obviously it is better to return a False before trying to allocate too much memory, but in the meantime i thought that capturing the exception would have been good enough.
timotimo actually, if your index data is a buf8 or blob8 or similar already, you can just use what malloc gives you 15:31
El_Che will the application be portable?
timotimo malloc is available everywhere we have moarvm, i believe 15:32
El_Che aren't you binding libc? 15:33
(aka will it work on Windows?)
timotimo if you "is native(Str)" it will use what's loaded in the process already
shinobi-cl the data is not buf8 of blob8. It is actually anything that perl6 can represent. I will skip this for the moment i guess. BTW i'm developing on Linux 64bit.
timotimo oh, in that case calculating what the necessary size would be for the buffer is rather difficult 15:34
if your index is just like an array of native ints or nums it'd be simple, but with proper objects you'll go through moarvm's own gc-managed heap
shinobi-cl Well... the index is just a .Str of the data in the cells, and the contents are just Ints indicating the row number where the value can be found 15:35
like "A" -> (1, 3, 4, 1000) 15:36
like "A" => (1, 3, 4, 1000)
we can found the value "A" in the rows 1, 3, 4 and 1000 for that column.
timotimo so a list of pairs from Str to a list of Int?
shinobi-cl yes 15:37
timotimo if you want to be memory-efficient and don't mind a bit of extra work, you can make that a list of strings and three lists of ints
the list of strings is what has your "A" and friends in it, the first list of ints tells you where in the third list your actual data begins and the second list of ints tells you how many entries belong to that 15:38
i.e. for just the A example you have it would be ["A"], [0], [4], [1, 3, 4, 1000]
15:38 Rawriful joined
shinobi-cl %index<ColumnA> = :A(1,3,4,100), :B(2,500), :C(9,10,11,540) .... etc 15:38
thanks timotimo, i will see how i can make it work. will try to use the least amoung of hashes and only use arrays when possible. Just one last question. Is there a method or way to ask the acual size of an object? 15:42
actual size*
timotimo apart from taking heap snapshots and measuring memory usage before and after creating it, not really, no 15:43
Ulti ///^H^H^H^H/win 5
timotimo especially since if you recently freed objects of the "right size", you'll fill slots back up that were used before, so memory usage will appear to be 0
shinobi-cl that sounds fine for a test i am doing. What can i found docs about this (my guess is that this has to be done in NQP) 15:44
Where can i find*
(sorry my english sucks today :P )
timotimo which do you mean? 15:46
measuring memory usage before and after?
shinobi-cl yes
timotimo the telemetry module in rakudo exposes that, for example
shinobi-cl I will use that one, seems ok for what i need :) thanks timotimo++ 15:49
timotimo good luck! 15:50
i'll be afk for a bit
15:52 cdg joined
wictory[m] where is &next defined? 15:59
moritz src/core/control.pm 16:00
wictory[m] thanks!
moritz found through "git grep 'sub next'" :-)
you're welcome
16:01 setty1 joined
wictory[m] hmm, I was grepping but I didnt find it through that .. I'm probably getting tired. Also good to hear that other people use the same method as I do :) Thanks again! 16:02
16:09 esh left
timotimo s: "next", \() 16:10
SourceBaby timotimo, Something's wrong: ␤ERR: Cannot resolve caller sourcery(Str, Capture); 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: &next, \()
SourceBaby timotimo, Sauce is at github.com/rakudo/rakudo/blob/c6cc...rol.pm#L80
timotimo wictory[m]: ^
16:10 esh joined
moritz never figured out how to use that bot 16:13
wictory[m] timotimo: 👍
16:17 igor__ joined
igor__ Hi everyone! Can my someone explain why sort(@a) ~~ sort(@a) is False? 16:18
What comparison operator should use in this case?
psch igor__: depending on the contents, probably &infix:<eqv> 16:19
igor__ I just want to compare these two lists
I don't understand why ~~ doesn't work here
psch igor__: the content dependency is "are there custom classes in there that maybe eqv weirdly"
psch wonders where camelia fluttered off to 16:20
evalable6: help
evalable6 psch, Like this: evalable6: say ‘hello’; say ‘world’ # See wiki for more examples: github.com/perl6/whateverable/wiki/Evalable
psch evalable6: my @a = 3,1,5,2,4; say sort(@a) ~~ sort(@a)
evalable6 False
geekosaur also ~~ is not a short way to say eqv. it has special behaviors that may be getting in the way 16:21
do NOT assume it's the way to write a generic "=="!
psch yeah, i was wondering if the LHS topicalization is to blame
geekosaur++
El_Che eqv also is False
igor__ Yes, eqv works indeed
psch evalable6: my @a = 3,1,5,2,4; say sort(@a) eqv sort(@a)
evalable6 True
igor__ I know that it is not a short way, I thought that ~~ is used to compare elements 16:22
and if elements are equal, then ~~ is true
psch igor__: smartmatch is not comparison. it's asking the RHS if the LHS is acceptable
evalable6: my &h = -> $a { $a.uc }; say &h ~~ &h 16:23
evalable6 (exit code 1) No such method 'uc' for invocant of type 'Block'
in block <unit> at /tmp/Ki1IIGAwks line 1
psch evalable6: my &h = -> $a { $a.gist.uc }; say &h ~~ &h
evalable6 -> $A { #`(BLOCK|94066313049264) ... }
igor__ Quote from a Perl6 book that I'm reading now: "To know whether they contain the same elements, use the smart match operator" 16:24
timotimo evalable6: my @a = 3,1,5,2,4; say sort(@a).List ~~ sort(@a).List
evalable6 True
El_Che evalable6: my @a = 3,1,5,2,4; @a ~~ @a
evalable6
timotimo igor__: ^
El_Che evalable6: my @a = 3,1,5,2,4; say @a ~~ @a 16:25
evalable6 True
timotimo evalable6: my @a = 3,1,5,2,4; say @a.sort.WHAT
evalable6 (Seq)
psch timotimo: ah, so it's Seq.ACCEPTS that's surprising?
timotimo i think so
igor__ Ok, that means that I have to add .List here
timotimo i don't know what Seq.Accepts should do, in general
igor__ convert to a list explicitly
psch it does Cool.ACCEPTS currently, apparently
timotimo igor__: or assign to an @ variable earlier
psch as in, Seq.pm doesn't have a definition
+src/core/ 16:26
oh hm, maybe Iterable or Sequence have one
El_Che it true it's not alway straight forward when a .List convertion is required
timotimo psch: sourcery can help us with that :)
s: (1,2,3).sort, "ACCEPTS", \() 16:27
SourceBaby timotimo, Something's wrong: ␤ERR: Could not find candidate that can do \()␤ in sub sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 29␤ in block <unit> at -e line 6␤␤
timotimo er, of course
s: (1,2,3).sort, "ACCEPTS", \(Seq)
SourceBaby timotimo, Sauce is at github.com/rakudo/rakudo/blob/8afd...Any.pm#L19
psch ah, Any
timotimo oops
that's an undefined seq right there
psch timotimo: yeah i stopped looking manually after checking Seq, Sequence, and Iterable
timotimo s: (1,2,3).sort, "ACCEPTS", \(Seq:D)
SourceBaby timotimo, Sauce is at github.com/rakudo/rakudo/blob/8afd...Any.pm#L19
timotimo s: (1,2,3).sort, "ACCEPTS", \(Seq.new())
SourceBaby timotimo, Something's wrong: ␤ERR: Too few positionals passed; expected 2 arguments but got 1␤ in block <unit> at -e line 6␤␤
timotimo s: (1,2,3).sort, "ACCEPTS", \((1,2,3).sort)
SourceBaby timotimo, Sauce is at github.com/rakudo/rakudo/blob/8afd...Any.pm#L18
timotimo multi method ACCEPTS(Any:D: Mu:D \a) { self === a }
that's the one it uses 16:28
psch mhm, eqaddr should probably fail there
...unless we started interning Seqs? :o
igor__ Is there any universal method to say, when I should convert something to a list and when not?
psch igor__: i'm not quite sure the current behavior you observed should stay like this 16:29
igor__ By the way, it is enough to convert the RHS only
16:29 Rawriful left
psch as in, at least fully reified Seqs should probably ACCEPTS like List does 16:29
igor__ so sort(@a) ~~ sort(@a).List is True
psch says my gut
mst psch: I think <thing> ~~ <thing> for any value of <thing> should be expected to be true unless you've got a *really* good reason for it not to be, yeah 16:31
psch mst: like <thing> being a Routine
El_Che mst++ 16:32
igor__ mst: I agree, that Is what I expected
16:32 domidumont joined
psch evalable6: say Regex.^mro 16:33
evalable6 ((Regex) (Method) (Routine) (Block) (Code) (Any) (Mu))
mst psch: because putting a Routine on the RHS of smartmatch is explicitly asking for it to be invoked and its result used, right/
psch well, Code is probably the type graph diversion that makes it special
mst ah, yeah, it's marked as "(Code) method ACCEPTS" but it's in the Routine docs, hence me getting the wrong one 16:34
igor__ Is there any method shat I should use for reliable comparison of tqo seqs? Just convert the RHS to a List?
psch mst: to be fair, i think that really is the one execption, Code being magically as smartmatch RHS 16:35
igor__: eqv as suggested earlier seemed to work?
teatime wonders if the similarity between Seq and Sikh is one of those odd Perl/Larry jokes or not
mst or "you made an object of a class where you defined your own ACCEPTS method in order to intentionally alter the behaviour"
Code behaving magically is really just a shorthand way of doing that 16:36
(arguably, in my head)
16:36 chakli joined
igor__ eqv will not work if you compare sorted([1,2,3,4]) with [1,2,3,4] though elements are equal 16:36
psch oh, yeah, but Code is the one exception we have in CORE afaik
igor__: but that's not two Seqs ;)
igor__: i see your point though.
igor__ yes, I said it wrong 16:37
two iterables or whatever
mst I guess the problem is that you can't be consistent for all Seqs, because not all Seqs can be effectively compared, because they could be lazy and/or infinite?
chakli Is there a builtin class/function to measure time elapsed
timotimo mst: Seq is a stateful thing, so checking against another thing should consume the seq and you're screwed :)
chakli m: my $start = time; sleep 3;say time -$start
evalable6 3
chakli like this? 16:38
timotimo m: my @a = 9, 4, 2, 5, 6; my $foo = @a.sort; say $foo ~~ $foo; .say for $foo
evalable6 True
(2 4 5 6 9)
psch m: my $start = now; sleep 1; say now - $start
16:38 domidumont left
evalable6 1.0017287 16:38
timotimo but apparently we cache that?
oh, we don't compare at all, right
m: my @a = 9, 4, 2, 5, 6; my $foo = @a.sort; say $foo ~~ $foo.List; .say for $foo
evalable6 (exit code 1) False
The iterator of this Seq is already in use/consumed by another Seq
(yo…
timotimo, Full output: gist.github.com/c10d4b00c8389e82e5...eb534a78d2
timotimo mst: ^- check it out
16:38 evalable6 left, evalable6 joined, domidumont joined
moritz chakli: I think lizmat built something for instrumentation/timing, but I can't remember what it's called 16:39
timotimo moritz: it's "Telemetry" and "snapper" 16:40
i.e. if you -Msnapper your script it'll give you a nice table/summary at the end
igor__ So, may I think that the best method is to ~~ + convert RHS into a list?
16:40 zakharyas joined
psch the build-in module for timing is Telemetry, isn't it 16:41
timotimo yes
chakli Telemetry seems more like perf :) ill stick to "time - $start" :) 16:46
Thanks 16:47
psch m: say time.WHAT; say now.WHAT 16:48
evalable6 (Int)
(Instant)
psch chakli: &now gives you milliseconds, fyi
moritz m: say now 16:49
evalable6 Instant:1517158200.001190
chakli oh, thanks. ill use that then
moritz looks like microseconds, even
chakli is there something like c++ destructor for perl6 classes? 16:52
moritz there's DESTROY, but it only runs when the GC reclaims the object
psch object destruction isn't deterministic 16:53
there's probably a long explanation by jnthn++ somewhere in the irclog
all i know is "spooky lowlevel magic happens sometimes" 16:54
chakli not when it the object goes out of scope
timotimo correct
geekosaur it becomes *eligible* if it goes out of scope with no remaining references. gc will happen when memory is needed at some point afterward 16:55
I don;t know exactly what gc strategy or mechanism moarvm uses 16:56
timotimo if the object dies young, the gc that will get it will happen rather a bit sooner, because moarvm does minor and major collections
wictory[m] is there any nice way to format a Duration as hh:mm:ss, for example, or does everybody roll their own formating every time? 17:00
17:02 Cabanossi left, zakharyas left 17:03 khw joined
wictory[m] p6: say(DateTime.new(DateTime.now.later(:1day)-DateTime.now).hh-mm-ss) 17:03
17:04 Cabanossi joined
wictory[m] m: say(DateTime.new(DateTime.now.later(:1day)-DateTime.now).hh-mm-ss); 17:06
evalable6 23:59:59
17:09 itaylor57 joined
psch .o( days are only 23:59:59 long??! ) 17:10
moritz seems about right :-)
17:10 khisanth__ left
moritz the two execution of DateTime.now won't return the same result 17:10
*executions
wictory[m] a day is actually something like 23:59:59.999287, depending on your perl6 build.. 17:13
moritz m: given DateTimew.now { say DateTime.new(.later(:1day)-$_).hh-mm-ss
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/bhWVSLxhbq
Missing…
moritz, Full output: gist.github.com/49ade66e5ef99e603c...d68d53f6b0
igor__ Ok, thank you very much for the help! I created a question on Stackoverflow (before joining this channel), so if you have time and want to answer it for all, please: stackoverflow.com/questions/484883...n-operator
moritz m: given DateTimew.now { say DateTime.new(.later(:1day)-$_).hh-mm-ss }
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/WWT6E4x5xo
Undeclared name:
DateTimew used at line 1. Did you mean 'DateTime'?
igor__ And thank you very much for the help once again
moritz m: given DateTime.now { say DateTime.new(.later(:1day)-$_).hh-mm-ss } 17:14
evalable6 00:00:00
17:14 leah2 joined 17:22 khisanth__ joined 17:23 igor__ left
moritz igor__: I've written an answer, though I suspect it might not be very satisfactory to you 17:23
Geth doc: f74384849d | (Moritz Lenz)++ | doc/Type/Any.pod6
&sort actually returns a Seq, not a List

see stackoverflow.com/questions/48488381		 17:27
synopsebot Link: doc.perl6.org/type/Any
17:28 leah2 left
teatime moritz: from the other guy: "It would not make sense to run through two Seqs to see if they equivalent because then they'd be exhausted." seems like there would be plenty of situations where that's all you want/need to know about two seqs, and are happy to let them be exhausted after comparison. *shrug* 17:30
timotimo teatime: then you can still use .List on them and discard the result immediately, or use something like >>==<< 17:31
moritz teatime: the other option would be to use .cache in ACCEPTS
teatime: for the plenty of cases where you don't want the Seq to be exhausted
the other question is: is it really worthwhile to return a Seq from sort, and not just a List? 17:32
you have to exhaust the input list anyway for sorting, so the performance gains from returning a lazy list are limited by nature
teatime timotimo: to your first point, they might be really long. but >>==<< would seem to satisfy me. 17:33
17:33 zakharyas joined
moritz m: say (1, 2) >>==<< (1, 2, 3) 17:34
evalable6 (exit code 1) Lists on either side of non-dwimmy hyperop of infix:<==> are not of the same length
left: 2 elements, right: 3 elements
in block <unit> at /tmp/bhurDAHHbt line 1
moritz might not be what you want
psch m: say (1,2) <<==>> (1,2,3)
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/esu9jrTNFo
Unsuppo…
psch, Full output: gist.github.com/f8e42d06acaa84869d...6336373a25
psch oh can't dwim both sides or something..? 17:35
ah, no
m: say (1,2) <<[==]>> (1,2,3)
evalable6 (True True False)
17:37 trnh joined 18:00 leah2 joined 18:03 zakharyas left 18:08 domidumont left 18:09 domidumont joined 18:16 zakharyas joined, domidumont1 joined 18:18 domidumont left 18:20 Ven`` left
lizmat Files=1233, Tests=76124, 302 wallclock secs (15.06 usr 5.48 sys + 2070.09 cusr 200.43 csys = 2291.06 CPU) 18:22
wow ^^^
18:32 igor__ joined
igor__ It's me again (the question about ~~ of seqs) 18:33
Probably you knew it, but it was a surprise for me,
result of ~~ depdends on implementation: Perl6 from Debian returns True, 18:34
psch igor__: that's the same implementation, just an older version
igor__ and perl6 that I built using rakudobrew returns False
psch bisectable6: help
bisectable6 psch, Like this: bisectable6: old=2015.12 new=HEAD exit 1 if (^∞).grep({ last })[5] // 0 == 4 # See wiki for more examples: github.com/perl6/whateverable/wiki/Bisectable
teatime igor__: it might be such an old version that sort doesn't return a seq. 18:35
psch bisectable6: old=2015.12 new=HEAD my @a = 2,4,1; say sort(@a) ~~ sort(@a)
bisectable6 psch, Bisecting by output (old=2015.12 new=8afd791) because on both starting points the exit code is 0
psch, bisect log: gist.github.com/455b8ae06915031698...7da6b2cbef
psch, (2016-12-28) github.com/rakudo/rakudo/commit/43...296ecab5be
igor__ yes, you seem to be right
ic$ perl6 --version This is Rakudo version 2017.12 built on MoarVM version 2017.12.1 implementing Perl 6.c. ic$ /usr/bin/perl6 --version This is perl6 version 2015.11 built on MoarVM version 2015.11
AlexDaniel 6c: my @a = 2,4,1; say sort(@a) ~~ sort(@a)
committable6 AlexDaniel, ¦2015.12,2016.01.1,2016.02,2016.03,2016.04,2016.05,2016.06,2016.07.1,2016.08.1,2016.09,2016.10,2016.11,2016.12: «True␤» ¦2017.01,2017.02,2017.03,2017.04.3,2017.05,2017.06,2017.07,2017.08,2017.09,2017.10,2017.11,2017.12,2018.01,HEAD(8afd791): «False␤»
psch AlexDaniel++
igor__ wow 18:36
that was cool
AlexDaniel
.oO( what am I being ++-ed for? The bots are doing all the work )
igor__ 6c makes magic
6c: my @a = 2,4,1; say sort(@a) ~~ sort(@a).List
committable6 igor__, ¦6c (27 commits): «True␤»
psch AlexDaniel: knowing and working on the bots :p 18:37
igor__ Wow, that was cool
How can I use this bot offline? Or better to say, how can I use without IRC?
Just to check my dumb question before asking them 18:38
AlexDaniel psch: what I find more impressive is that evalable6 figured out by itself that camelia is not here, and started responding to m:
18:38 kjk joined
AlexDaniel never thought that this feature would be as useful :) 18:38
18:38 zakharyas left
AlexDaniel igor__: unfortunately there is no way currently, I have some ideas on how to make that happen but not much done yet 18:39
psch yeah, that's pretty great
AlexDaniel igor__: you can of course compile rakudo on every release, that by itself will take a bit of time
igor__ Yes, I would do it
And provide a public service doing this
AlexDaniel mm what do you mean? :) 18:40
igor__ I'm working on a documentation service for different languages (cheat.sh; just do curl cheat.sh to see what it looks like)
psch igor__: maybe make it an irc bo- wait a second
18:41 Kyo91 joined
igor__ perl6 is not yet integrated, but I'm working on it now 18:41
18:41 rindolf left
igor__ so it will be possible to get perl6 doc using cheat.sh (and all its interfaces like vim plugins, emacs plugins etc.). Probably it is not really useful for perl6, 18:42
teatime igor__: nice.
igor__ because it has its own documentation mechanisms and they are really cool,
psch igor__: nah, it sounds like it could be useful
igor__ BUT
AlexDaniel interesting project
igor__ I want to integrate perl6 support in cheat.sh anyway,
and now I had an idea:
kjk hello, 10.rand works but shouldn't rand(10:) also work? 18:43
igor__ integrate not only doc, but execution also
AlexDaniel huggable: eval
huggable AlexDaniel, nothing found
igor__ and this feature (something like c6)
psch kjk: i think you mean 'rand 10:'
igor__ would be really cool here
18:43 rindolf joined
teatime igor__: I'm actually rather impressed, because the commands I've tried so far actually have useful cheatsheets already. 18:43
psch m: say rand 10:
kjk psch: doesn't work either
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/qRLhDwjZ8w
Unsuppo…
psch, Full output: gist.github.com/7cef0e5f02b365d844...d99f22e40f
psch kjk: ah, precedence there probably 18:44
m: say rand 10::
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/TMhaVa_tKI
Unsuppo…
psch, Full output: gist.github.com/bf49a5fdc984c3984a...7a79330489
18:44 leah2 left
psch hmm, maybe not vOv 18:44
ooh, it's a per5-ism warning there
kjk: yeah, it should work but we don't reach the... whatever we call the : there during parse because we hit a perl5-ism complaint first 18:45
m: say 10:
evalable6 10
kjk perl5-ism warning messes up the grammar?
igor__ teatme: you cal try cheat.sh/perl/ by the way
psch kjk: FSDO "mess up," i suppose :)
...i think that colon is also an invocant marker? the term seems bound to signatures in my head though 18:46
18:50 igor__ left
kjk psch: rand being a term gives it higher precedence than calling like rand(10:) ? 18:51
psch kjk: the parens are always wrong i think
m: say(10:) 18:52
evalable6 10
psch uh
...or i'm just confused
m: class A { method foo { say "foo" } }; foo(A:)
evalable6 foo
kjk m: pick(^10:)
evalable6
psch kjk++
of course the parens are allowed, because : is just an invocant marker and it's still a normal invocation... >_> 18:53
m: say &rand 18:54
evalable6 sub rand ( --> Num:D) { #`(Sub+{Callable[Num:D]}|93966989900360) ... }
psch m: say &term:<rand>
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/JuJI2NSlYq
Undeclared routine:
term:<rand> used at line 1. Did you mean 'term:<time>', 'term:<now>'?
psch kjk: rand doesn't seem to be a term
well, we have a 'token term:sym<rand>' in Perl6::Grammar 18:55
that's why we compile-time SORRY
kjk saw this in the doc: sub term:<rand> (--> Num:D)
psch huh, i wonder why that's there 18:56
the impl of &rand is a normal sub
github.com/rakudo/rakudo/blob/mast....nqp#L3426 is the relevant bit of the core grammar
that's where we compile-time bail 18:57
kjk I don't really understand the difference between a term and a normal sub. is it precedence?
psch terms have different parsing rules 18:58
and allow more than subs as their identifier too 18:59
...actually i think that is the parsing difference
i might be getting tired :)
m: sub term:<dice> { 1..6 .pick }; say dice + dice
evalable6 12
psch m: sub dice { 1..6 .pick }; say dice + dice
evalable6 (exit code 1) Too many positionals passed; expected 0 arguments but got 1
in sub dice at /tmp/DJhj6EhYGP line 1
in block <unit> at /tmp/DJhj6EhYGP line 1
psch kjk: ^^^ that's another difference actually 19:00
from docs.perl6.org/routine/term:%3C%3E
kjk psch: oh ic, thanks 19:01
TEttinger m: sub term:<dice> { 1..6 .pick }; say dice + dice + dice + dice 19:03
evalable6 12
TEttinger m: sub term:<dice> { 1..6 .pick }; say dice + dice + dice + dice
evalable6 12
TEttinger uh
kjk ha I think (1..6).pick is needed
teatime TEttinger: abhishekkr.github.io/slides/slide-...random.gif
psch oh right
prefix:<^> vs infix:<..>
TEttinger but that wouldd be 3 right? 19:04
psch the former has the space trick
...or something?
i give up, time for bed o/
TEttinger m: sub term:<dice> { (1..6).pick }; say dice + dice + dice + dice
evalable6 14
TEttinger better
m: sub term:<dice> { (1..6).pick }; say dice + dice + dice + dice + dice
evalable6 17
TEttinger still statistically unlikely
err no 19:05
that's average, or close
m: sub term:<dice> { (1..6).pick }; say dice + dice + dice + dice + dice
evalable6 15
TEttinger ok, good
19:10 leah2 joined 19:12 kjk left 19:29 leah2 left 19:34 greppable6 left 19:35 greppable6 joined 19:37 pecastro left 19:41 domidumont1 left 19:51 Rawriful joined 20:12 leah2 joined 20:30 ChoHag left
DrForr m: sub infix:<d>($a,$b){[+]{(1..$b).pick) xx $a}} 20:32
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/sS0f4AFe1Y
Unable …
DrForr, Full output: gist.github.com/b84be515d7b8a115c4...fa8ddee656
20:33 Kaiepi left 20:36 Kaiepi joined
AlexDaniel m: sub infix:<d>($a,$b){[+]{(1..$b).pick xx $a}} 20:36
evalable6
20:55 teatime left 21:00 rindolf left 21:02 rindolf joined 21:05 |oLa| joined 21:07 markong left 21:13 colomon left 21:14 colomon joined, MasterDuke joined 21:22 MasterDuke left 21:33 lasse_ left 21:40 rindolf left 21:45 MasterDuke joined 21:56 MasterDuke left 21:58 MasterDuke joined 22:04 markong joined, MasterDuke left 22:05 Aaronepower joined 22:10 itaipu joined 22:39 setty1 left 22:42 shinobicl joined 22:43 shinobi-cl left 23:04 SourceBaby left 23:05 p6lert left 23:06 synopsebot left, dalek left, Geth left, psch left, Util left 23:07 sergot left, masak left 23:09 Ven`` joined, Ven`` left, dalek joined, ChanServ sets mode: +v dalek, synopsebot joined, Geth joined, ChanServ sets mode: +v synopsebot, ChanServ sets mode: +v Geth, p6lert joined 23:11 colomon left, Ven`` joined 23:17 cdg left 23:24 shinobi-cl joined 23:25 shinobicl left
buggable New CPAN upload: P5-X-0.0.1.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.1.tar.gz 23:25
23:25 shinobicl joined 23:28 shinobi-cl left 23:30 Util joined
lizmat ^^ supporting -r -w -x -e -f -d -s -z -l 23:30
still a WIP, but usable so far 23:31
23:31 Eddward joined 23:32 Ven`` left 23:34 wamba left
buggable New CPAN upload: P5built-ins-0.0.3.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.3.tar.gz 23:35
23:50 comborico1611 joined 23:54 colomon joined 23:59 trnh left

Powered by Cro and the Raku® Programming Language.

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

Thank you!