»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:05 znpy joined
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 603b213 | (Zoffix Znet)++ | / (8 files):
Ditch `logo` from the Dists model

We can now auto-generate logosprites from dist names and checking the logotype dir for the actual file, so we don"t need to store that piece of data in DB		 00:06
00:10 psch left 00:15 cdg left
ab5tract m: say +"@a = []".comb < +"@a.clear" 00:22
camelia rakudo-moar : OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '7⏏5\@a.clear' (indicated by 7⏏)␤ in block <unit> at /tmp/uDz2KAbbcX:1␤␤Actually thrown at:␤ in block <unit> at /tmp/uDz2KAbbcX:1␤␤»
ab5tract m: say +"@a = []".comb < +"@a.clear".comb
camelia rakudo-moar : OUTPUT«True␤»
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 76486e8 | (Zoffix Znet)++ | lib/ (4 files):
Toss dists no longer in ecosystem from db on build complete
 00:35
00:38 secwang left 00:42 secwang joined
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 727a71e | (Zoffix Znet)++ | lib/ (2 files):
Refactor: get deleted dists num prior to dumping them instead of fetching all rows twice in db builder
 00:44
00:50 BenGoldberg joined
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 5f9443f | (Zoffix Znet)++ | / (13 files):
Move DbBuilder classes inside the ModulesPerl6:: namespace
 00:53
href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: b617f22 | (Zoffix Znet)++ | t/01-models/02-dists.t:
Fix tests: ::Dists model ->remove_old now returns num of dists dumped instead of invocant
 00:55
rl6-most-wanted: cee0cf6 | (David Warring)++ | most-wanted/modules.md:
Update modules.md

Mention PDF::Grammar,PDF-Tools and Crypt::RC4
01:05 yeahnoob joined 01:06 adhoc_ joined, adhoc_ is now known as adhoc 01:12 znpy left, vendethiel left
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: e58d9ca | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist (3 files):
Add PostProcessor plugin pipeline
 01:15
perlawhirl hellow all... is there a way i can do a 'getch()' type function ( without importing the NCurses module ) 01:20
ie... i want to read one string without waiting for a newline
i mean... one char
gfldex perlawhirl: see github.com/rakudo/rakudo/blob/bceb...le.pm#L171 01:26
uruwi Hold on perlawhirl 01:27
github.com/bluebear94/messing-arou...er/kbd.pl6 01:28
You might want to look at this
It requires a Unix-like environment, though
gfldex, I tried using it and it _does_ wait for a newline; one can solve this by using `stty raw noecho` 01:30
01:32 leont left
uruwi Okay, does anyone know of a portable solution... 01:33
gfldex please note that with stty raw you may have to handle ^C by hand 01:34
uruwi ^ 01:35
though in my example I bind it to `q` :P
01:36 yeahnoob left
uruwi But stty raw is useful if you don't want users (in)advertently exiting the program. 01:37
konobi tput? 01:39
geekosaur ? tput is still unix specific... 01:40
perlawhirl i am on linux so can call stty... will have a go at that... eventually someone will write a Term::ReadKey module :D 01:41
geekosaur also: I wouldn't use stty raw unless you really need to block terminal signals as well. stty -icanon min 1 time 1
konobi geekosaur: not really... the same mechanism works on win32, you just need a win32 termcap thing
geekosaur and tput is output, doesn't help with input 01:43
zengargoyle input and output are the same-ish talking to a tty.
geekosaur tput does not call the terminfo ioctls
er, termios
zengargoyle holdover from the days when display and keyboard were connected to computer over serial.
uruwi * `stty raw -echo time 1`
geekosaur tput is terminfo database lookups 01:44
konobi geekosaur: the node.js "blessed" library seems to cope pretty well with tty's without needing to link to any libraries
zengargoyle all communication between the two is escape codes on that 2-way pipe.
konobi steal stuffz!
geekosaur zengargoyle, changing to character oriented input is not an escape sequence
whatever, this looks like everyone reverted to 4.2BSD when I wasn't looking 01:45
zengargoyle i think it's just moves to the pseudo tty layer.
geekosaur what does that have to do with
never mind
konobi geekosaur: github.com/chjj/blessed
geekosaur konobi, you are remarkably good at missing the point 01:46
when did I say anything abotu libraries
but I think I'm just going to back out because there's so much incorrect and circa-1982 info here that it's not worth it
konobi geekosaur: tput is termcap and terminfo... so you get input and output 01:47
geekosaur konobi, tput is a program, not a library
it queries the terminfo database
it does not do input
konobi mmm... might be the terminology on the package... but i swear it does input and output
geekosaur in particular it does not make stty calls for you
man 1 tput 01:48
konobi well, termcap does the stty stuff, iirc
geekosaur what I suspect is meant there is that it uses the tput program to do output escape sequences and possibly to look up function key definitions
konobi anyways... it's all in plain JS there... so no C library linking needed 01:50
uruwi It uses the node standard library.
The Perl 6 base libraries might not have the same capabilities.
konobi a small amount
01:51 adhoc left
uruwi Which brings me to wondering why the Perl6 NCurses bindings won't work on Cygwin 01:52
01:52 adhoc joined 01:53 xpen joined, z8 joined
konobi uruwi: actually... it only uses fs/path stuff from node core... the rest is all natural 01:55
01:55 kid51 left
uruwi Seems to use event emitters, too 01:56
I don't know if Perl6 has an equivalent...
konobi when it all comes down to it... it's all just escape sequences on a tty anyway
uruwi I know. 01:57
The question is how to disable line buffering.
konobi lemme see 01:58
geekosaur that is not done with an escape sequence
it is done with an ioctl to the tty driver
there *are* escape sequences output afterward, usually, because old terminals used to require stupid stuff, and usually one switches to/from the alternate screen, but outputting those escapes does not switch you out of line mode 02:01
02:01 Some-body_ joined 02:02 DarthGandalf left, Some-body_ is now known as DarthGandalf
konobi geekosaur: ugh... i'd have to delve back into bootloader code to work it out again 02:03
geekosaur it's using stty, not tput 02:04
stty doesn't output escape sequences, it uses driver ioctls
zengargoyle www.linusakesson.net/programming/tty/ - may be an interesting read 02:06
02:09 Actualeyes joined 02:10 Ben_Goldberg joined 02:11 Ben_Goldberg left, Ben_Goldberg joined, Ben_Goldberg left 02:12 Ben_Goldberg joined
geekosaur was the author of the tty section of the first edition of _Unix Power Tools_, fwiw. it's not in later editions because POSIX took over and you no longer needed to know whether you had to use stty(), ioctl(TIOCSTTY), or ioctl(TCSETA)... 02:12
02:12 BenGoldberg left
geekosaur (had to know that even if using the stty command because the different systems had different stty options) 02:14
02:14 adhoc left, dayangkun joined
konobi yeah... solaris tty stuff was a huge PITA initially 02:14
02:23 diakopter___ left
zengargoyle ah, SunOS. accidentally have caps-lock set on your termnal and it thinks you just don't do case at all... W\E\L\C\O\M\E \T\O S\U\U\NOS 02:28
02:29 secwang left
geekosaur STTY -OLCUC -IUCLC 02:30
zengargoyle more like \E\X\I\T (unset caps-lock) login again... ;) 02:31
geekosaur not with the backslashes...
02:33 z8 left 02:35 BenGoldberg_ joined
zengargoyle i don't remember it being that easy. but it was a *long* time ago. 02:36
geekosaur iirc the gotcha there was it also turned off icrnl so you had to use EXIT<control-J> 02:37
or something silly like that
02:37 kaare_ joined 02:38 BenGoldberg_ left, Ben_Goldberg left, BenGoldberg_ joined 02:39 adhoc joined, molaf_ joined
geekosaur but it treated uppercase as lowercase and used backslash to indicate uppercase 02:40
konobi bah... i'll have to follow up.
02:43 molaf left 02:45 adhoc left 02:46 Ben_Goldberg joined 02:47 zengargoyle left, adhoc joined 02:49 BenGoldberg_ left, zengargoyle joined 02:53 Mhoram joined, abaugher left 02:54 BenGoldberg joined 02:56 Ben_Goldberg left, BenGoldberg left, BenGoldberg joined 03:01 telex left 03:02 telex joined
perlawhirl uruwi: That kbd script... is that yours? 03:03
03:09 vendethiel joined 03:18 noganex_ joined, yqt left 03:20 noganex left
skids perlawhirl: gist.github.com/skids/0b1cfabedddd95a824c8 is from some unpublished code I've started to pay with. It is not portable off UNIX but is does handle the cases like cat foo | perl6 interactive-script.pm where stdin != tty. Right now the IO system is still too abstract to make this easy. 03:23
*play
03:29 kaare_ left 03:32 yeahnoob joined, vendethiel left, AlexDaniel left
perlawhirl thanks skids will check it out 03:34
03:37 yeahnoob left 03:38 cdg joined 03:48 hankache joined, hankache left 03:49 hankache joined, jevin joined 03:50 jevin_ left 03:54 yeahnoob joined 04:10 zengargoyle left 04:13 zengargoyle joined 04:34 kmel joined 04:38 hankache left, hobbs left 04:40 hobbs joined 04:45 cdg left 04:54 kmel left 05:03 geraud left, skids left 05:04 geraud joined 05:29 molaf_ left
JimmyZ m: use Test; use Test; 05:33
camelia ( no output )
05:40 perlawhirl left 05:54 khw left 06:03 zengargoyle left 06:30 BenGoldberg left 06:38 dalek left, dalek joined, ChanServ sets mode: +v dalek 06:40 TEttinger joined 06:51 geraud left 06:56 domidumont joined 07:01 domidumont left, domidumont joined 07:11 doublec_ joined 07:12 doublec left 07:13 CIAvash joined 07:19 [Sno] left 07:21 edehont joined 07:24 doublec_ is now known as doublec 07:38 cbk_ left 07:40 _28_ria left
[Tux] again up, two days in a row: 07:44
test 50000 23.517 23.403
test-t 50000 21.276 21.162
csv-parser 50000 29.581 29.468
07:45 cdg joined 07:46 _28_ria joined 07:48 [Sno] joined 07:50 cdg left 07:53 RabidGravy joined
dalek kudo/nom: 8ef5a59 | TimToady++ | src/Perl6/ (2 files):
allow unival terms, because why not?
 08:05
ast: 373ecf3 | TimToady++ | S15-literals/numbers.t:
test for unival literals
08:06 rurban joined 08:12 FROGGS joined
TimToady m: say ⅐ 08:15
camelia rakudo-moar : OUTPUT«0.142857␤»
TimToady m: say ⅐.perl
camelia rakudo-moar : OUTPUT«<1/7>␤»
08:15 darutoko joined
TimToady of course this can be used for obfuscation... 08:22
m: say ⅰi
camelia rakudo-moar : OUTPUT«0+1i␤»
FROGGS uniWAT FTW
TimToady but I think it fits in well with our story of being completely Unicode friendly 08:23
and it's a trivial patch 08:24
cuneiform has some great numbers 08:26
m: say 𒐳
camelia rakudo-moar : OUTPUT«432000␤»
TEttinger "but can language X do this?"
if the question is anything unicode related, for perl it's yes, for many others, no 08:27
well, perl6
TimToady m: say Ⅹ
camelia rakudo-moar : OUTPUT«10␤»
08:31 krakan joined
grondilu lol, is that roman? 08:31
m: say IX
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/rUsVVzjyao␤Undeclared name:␤ IX used at line 1␤␤»
TimToady m: say Ⅸ
camelia rakudo-moar : OUTPUT«9␤»
grondilu oh ok
uruwi m: say １ 08:34
camelia rakudo-moar : OUTPUT«1␤»
TimToady m: say ^Ⅹ Xx Ⅹ
camelia rakudo-moar : OUTPUT«(0000000000 1111111111 2222222222 3333333333 4444444444 5555555555 6666666666 7777777777 8888888888 9999999999)␤»
uruwi m: say 二
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/q3bf0T8gDr␤Undeclared routine:␤ 二 used at line 1␤␤»
TimToady m: say ㆓
camelia rakudo-moar : OUTPUT«2␤»
TimToady m: say ㈡ 08:35
camelia rakudo-moar : OUTPUT«2␤»
uruwi What was the difference?
m: say ⑨
camelia rakudo-moar : OUTPUT«9␤»
TimToady m: say '二'.uniprop
camelia rakudo-moar : OUTPUT«Lo␤»
TimToady m: say '二'.unival
camelia rakudo-moar : OUTPUT«2␤»
TimToady it's a letter, so parses as identifier
uruwi say "二".char 08:36
m: say "二".char #gdi
camelia rakudo-moar : OUTPUT«Method 'char' not found for invocant of class 'Str'␤ in block <unit> at /tmp/SiwxA6vgAQ:1␤␤»
08:36 g4 left
uruwi m: say char("二") 08:36
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/x3QfxRmVUs␤Undeclared routine:␤ char used at line 1. Did you mean 'chars', 'chr'?␤␤»
uruwi m: say chr("二")
camelia rakudo-moar : OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '7⏏5二' (indicated by 7⏏)␤ in block <unit> at /tmp/LLLqxHiKVU:1␤␤Actually thrown at:␤ in block <unit> at /tmp/LLLqxHiKVU:1␤␤»
uruwi Oh screw it 08:37
08:37 g4 joined 08:38 Ven joined
masak morning, #perl6 08:38
uruwi It's 3:38 AM where I am
mrf o/ masak
uruwi Going to bed 08:39
08:39 uruwi is now known as uruwi[A]
TimToady only 12:39 here, but oughta go to bed... 08:39
masak m: say ord("二") - ord("亊") 08:40
camelia rakudo-moar : OUTPUT«2␤»
masak grins
grondilu m: say 三 08:43
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/6rvzLDfUz8␤Undeclared routine:␤ 三 used at line 1␤␤»
grondilu m: say uniname "三" 08:44
camelia rakudo-moar : OUTPUT«<CJK Ideograph>␤»
grondilu m: say uniname "㆓" 08:45
camelia rakudo-moar : OUTPUT«IDEOGRAPHIC ANNOTATION TWO MARK␤»
08:46 zakharyas joined 08:50 mr-fooba_ left 08:52 bjz joined 08:53 edehont left 08:55 azawawi joined
azawawi hi 08:56
and good morning #perl6 08:57
masak m: say ord("三") - ord("亊") 09:00
camelia rakudo-moar : OUTPUT«-129␤»
masak heh.
09:04 cognominal joined 09:05 jonas2 joined
dalek kudo/nom: 2e38ba6 | lizmat++ | src/core/ (6 files):
Introduce Any.gistseen

This method allows a developer to easily create a .gist method that is aware of self-referential structures (such can occur with arrays, lists, hashes and pairs). This replaces the ad-hoc code found so far.		 09:05
09:08 kjs_ joined
lizmat hmmm.... looks like TimToady's unival commit broke the JVM build, because nqp::unipropcode doesn't exist on JVM yet 09:21
gen/jvm/Perl6-Actions.nqp
Error while compiling op unipropcode (source text: "nqp::unipropcode(\"NumericValueNumerator\")"), no registered operation handler
09:21 dakkar joined
lizmat m: use nqp; say nqp::unipropcode("NumericValueNumerator") 09:22
camelia rakudo-moar : OUTPUT«10␤»
lizmat j: use nqp; say nqp::unipropcode("NumericValueNumerator")
camelia rakudo-jvm 4d8a90: OUTPUT«===SORRY!===␤Error while compiling op unipropcode (source text: "nqp::unipropcode(\"NumericValueNumerator\")"), no registered operation handler␤»
lizmat fudging is not that easy, atm, as I don't think Actions/Grammar.nqp *are* currently fudged :-( 09:23
JimmyZ or by nqp::config() or something? 09:28
09:29 ely-se joined 09:33 llfourn joined 09:38 espadrine joined 09:40 Actualeyes left
El_Che a silly question: Can I embed a perl6 compiled regex (rx) within a regex match with the :Perl5 modifier? 09:48
09:49 TEttinger left 09:50 azawawi left 09:51 rurban left, pochi_ joined, pochi left
El_Che tests from the replace sounds like it does 09:52
09:54 RabidGravy left, lucs left, RabidGravy joined 09:55 eliasr joined, lucs joined
lizmat afk until late& 09:57
09:58 Begi joined
grondilu m: my $rx = /foo/; say "foo" ~~ m:p5/$rx/ 09:58
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/kshctkdqAb␤Adverb p5 not allowed on m␤at /tmp/kshctkdqAb:1␤------> 3my $rx = /foo/; say "foo" ~~ m:p5/$rx/7⏏5<EOL>␤»
grondilu m: my $rx = /foo/; say "foo" ~~ m:P5/$rx/ 09:59
camelia rakudo-moar : OUTPUT«｢foo｣␤»
10:00 Actualeyes joined
El_Che m: my $rx = rx:i/<[a..z]>||<[1..9]>/; '.' ~~ m:Perl5/$rx/ ; '.a' ~~ m:Perl5/\.$rx*/ 10:00
camelia ( no output )
El_Che m: my $rx = rx:i/<[a..z]>||<[1..9]>/; '.' ~~ m:Perl5/$rx/
camelia ( no output )
El_Che it does on my repl :)
m: my $rx = rx:i/<[a..z]>||<[1..9]>/; '.a' ~~ m:Perl5/\.$rx/
camelia ( no output )
grondilu you need to say something you know 10:01
10:02 raiph left, yeahnoob left
El_Che grondilu: actually I am trying to find the combination \G + the /c modifier in the perl6 adverbs 10:04
:c is clear. Not so clear about \G
El_Che keeps reading
10:05 rindolf joined
dalek star: 2a70ed7 | (Jens Rehsack)++ | / (3 files): 10:08
star: introduce support for sdkroot and sysroot to support cross-builds
star:
star: To allow building for another location rather the final install prefix,
star: regardless whether chroot based builds, cross-compile or canadian cross sdk
10:08 dalek left 10:09 brrt joined, dalek joined, ChanServ sets mode: +v dalek
FROGGS is a canadian cross sdk something like a cansas city shuffle? 10:11
FROGGS .oO( There was a time. )
arnsholt FROGGS: I'm increasingly thinking that CArrays really shouldn't do the auto-resize thing, and be restricted to a single size. Thoughts on that? 10:12
jnthn morning, #perl6 10:13
arnsholt o/ jnthn
jnthn: Did my Gist hang on your machine as well, BTW? (If you've had time to try it) 10:14
jnthn arnsholt: Didn't have time yet, sorry
arnsholt No problem
El_Che if I rephrase my regex question, is there an equivalent in Perl6 for the perl5 \G anchor 10:15
arnsholt I know you've got a massive pile of stuff on your plate
jnthn Yeah...cleaning up the supplies API at the moment...
FROGGS arnsholt: hmmm
10:16 Actualeyes left
FROGGS arnsholt: what if you want to read a blob of mem until you find something null? 10:16
arnsholt FROGGS: Then it'll be length *
But that's generally for arrays coming from C, which aren't resized anyways
FROGGS still the CArrays grows over the not-owned memory 10:17
arnsholt El_Che: doc.perl6.org/language/regexes#Continue
10:17 Actualeyes joined, Begi2 joined
FROGGS arnsholt: but yeah, when the user creates an array (s)he should perhaps make it a fixed sized array instead 10:18
arnsholt: to avoid realloc hell
RabidGravy I love the way that actually using a library in real code changes the way you think about its interface
FROGGS RabidGravy: in what way? 10:19
10:19 Begi left
RabidGravy in that a nice concise simple interface can often be a pain in the arse when faced with real world problems ;-) 10:20
anyway off to hand over all my worldly goods to the car repairers ;-) 10:21
arnsholt FROGGS: Yeah, I would like it to be so that you always have to declare the size of a CArray when creating one on the Perl 6 side. For arrays coming from C we'll have to allow undefined size, of course
RabidGravy have fun
FROGGS RabidGravy: :o)
arnsholt: can we declare fixed sized CArrays already? 10:22
arnsholt Nope 10:23
El_Che arnsholt: Thx. That corresponds with perl5's /c. I am looking at the perl-ldap code and I see they use a /\G(?: ... )/gcx syntax
10:24 pmurias joined
El_Che maybe it is declaring twice the same thing 10:24
pmurias how should unipropcode be implemented on the JVM, by porting over ucd2c into ucd2java? 10:25
10:25 RabidGravy left
FROGGS arnsholt: then let's play with that thought when it is implemented :o) 10:25
10:26 brrt left 10:28 RabidGravy joined
arnsholt FROGGS: Sounds like a plan. I don't think it should be too hard to implement either 10:30
FROGGS arnsholt: would be nice to steal that task from jnthn++ 10:32
jnthn Task-stealing from jnthn encouraged. :) 10:49
10:52 ely-se left 10:53 Ven left
RabidGravy :) 10:53
okay the car menders didn't want to take my money
jnthn isn't sure whether that's good news or bad... :) 10:54
arnsholt I'm not sure shaped CArrays were ever jnthn's job, TBH =)
10:59 n0xff joined
jnthn Cute 11:01
method flat(Supply:D:) {
supply {
whenever self -> \inner {
whenever inner -> \value {
emit value;
}
}
}
}
:)
moritz nice
11:01 brrt joined
moritz that only flattens one level, though 11:01
is that desired?
jnthn I think so...we don't really have an item/list distinction with supplies 11:02
So you'd not really have a way to stop it 11:03
RabidGravy is "supply" the new "react" or does it do something else I can't work out from the above? 11:05
jnthn supply = returns a supply that, when tapped, runs the block 11:06
react = runs the block immediately and waits until all of the whenevers in it have run their course.
react is more for the "top level"
RabidGravy gotcha
arnsholt jnthn: I think you said something about the "of" syntax for typing aggregate things recently. Is that implemented, or still TODO? 11:10
jnthn arnsholt: I refactored handling of "of" recently, and implemented "is" too
m: my @arr of int; say @arr.WHAT # now works
camelia rakudo-moar : OUTPUT«(array[int])␤»
jnthn m: my @buf is buf8; say @buf.WHAT 11:11
camelia rakudo-moar : OUTPUT«(Buf[uint8])␤»
jnthn That also :)
m: use NativeCall; my @carr is CArray of int8; say @carr.WHAT # maybe this
camelia rakudo-moar : OUTPUT«(IntTypedCArray[int8])␤»
jnthn :)
arnsholt Oh, neat! 11:12
JimmyZ m: my @buf is uint8; say @buf.WHAT
camelia rakudo-moar : OUTPUT«===SORRY!===␤Method 'BUILDALLPLAN' not found for invocant of class 'Perl6::Metamodel::NativeHOW'␤»
jnthn uint8 isn't a container type :)
JimmyZ m: my uint8 @buf; say @buf.WHAT
camelia rakudo-moar : OUTPUT«(array[uint8])␤»
jnthn m: use NativeCall; my int8 @carr is CArray; say @carr.WHAT # cuter way, maybe
camelia rakudo-moar : OUTPUT«(IntTypedCArray[int8])␤»
JimmyZ anyway, LTA :)
jnthn Yeah, not sure what to do about that 11:13
Not sure I feel like demanding everyone "does Container" or some similar role on everything they wish to use as a container type
11:14 khisanth_ joined
jnthn Though it'd be the Java way out of the problem. :) 11:14
El_Che m: my regex foo { . } 11:15
camelia ( no output )
El_Che m: my regex Foo :i { . }
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/LzwzlOpqVy␤Malformed regex␤at /tmp/LzwzlOpqVy:1␤------> 3my regex Foo 7⏏5:i { . }␤»
El_Che adverbs not implemented for regexes?
arnsholt m: my Regex Foo { :i . } # <--
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/2JMFdOlPvu␤Malformed my␤at /tmp/2JMFdOlPvu:1␤------> 3my Regex7⏏5 Foo { :i . } # <--␤»
arnsholt Well, if you spell it right, anyways >.< 11:16
11:16 Khisanth left
El_Che arnsholt: thx. I got it from here: en.wikipedia.org/wiki/Perl_6_rules: regex addition :ratchet :sigspace { <term> \+ <expr> } 11:17
jnthn Then it's wrong, they go inside the curlies 11:18
El_Che jnthn: I'll change it there 11:19
jnthn regex + ratchet + sigspace = rule, also ;)
El_Che++
El_Che what better way to learn perl6 regex than trying to rewrite a perl5 regex mastodont from Net::LDAP :) 11:20
11:21 RabidGravy left, Begi2 left 11:38 rurban joined 11:39 Ven joined 11:58 dayangkun left
arnsholt El_Che: A propos regexes, you can scope adverbs inside groups too: m/foo [:i bar]/ 12:01
12:09 Skarsnik joined 12:14 zengargoyle joined 12:19 znpy joined 12:20 ely-se joined 12:26 BenGoldberg joined 12:30 perlawhirl joined 12:31 CIAvash left
El_Che arnsholt: interesting 12:32
12:36 xpen left
perlawhirl hello everyone 12:36
need a lil help
DrForr Mornin'.
12:36 kjs_ left
perlawhirl in perl5 you could put a map in a hash to generate some KV pairs... i'm trying to do the same with perl 6 and failing ==> pastebin.com/QxtwuK79 12:37
jnthn Try |( map { $_ => $_ } 1 .. 9 )
Or my %h = flat ... 12:38
perlawhirl thanks jnthn... you're a star
was watching your talk on objects ∩ concurrency today... very enjoyable and informative 12:39
jnthn :) 12:40
12:40 rurban left, jojotus left, rurban joined, rurban left 12:42 jojotus joined 12:43 BenGoldberg left
stmuk what's the JVM status? does it sort of work yet? or not so much.. :) 12:44
moritz it works on and off 12:45
zengargoyle my %h = |map { $_ => $_ }, 1..9; say %h; 12:50
m: my %h = |map { $_ => $_ }, 1..9; say %h; 12:51
camelia rakudo-moar : OUTPUT«1 => 1, 2 => 2, 3 => 3, 4 => 4, 5 => 5, 6 => 6, 7 => 7, 8 => 8, 9 => 9␤»
zengargoyle did TimToady go back on loosening precedence of | ?
m: |1..10
camelia rakudo-moar : OUTPUT«WARNINGS:␤Useless use of ".." in expression "|1..10" in sink context (line 1)␤»
zengargoyle m: say |1..10 12:52
camelia rakudo-moar : OUTPUT«1..10␤»
zengargoyle m: say |(1..10)
camelia rakudo-moar : OUTPUT«12345678910␤»
zengargoyle hopes so, likes tight |
12:52 pmurias left
dalek p: 6d3179e | (Stefan Seifert)++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
Fix error messages referring to the wrong op.

Thanks to ilmari++ for catching my copy&pasto!		 12:54
ilmari nine++ # quick fix 12:55
13:03 z8 joined 13:07 brrt left 13:12 z8 left
jnthn m: undef 13:14
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/LjGkdyylqa␤Unsupported use of undef as a value; in Perl 6 please use something more specific:␤ an undefined type object such as Any or Int,␤ :!defined as a matcher,␤ Any:U as a type constraint,␤ Nil as the …»
13:14 pmurias joined 13:19 johan joined 13:22 kjs_ joined 13:26 xenu is now known as pepe, pepe is now known as xenu
dalek kudo/resupply: 461ceab | jnthn++ | / (5 files):
Start cleaning up Supply design/implementation.

This commit substantially refactors how supplies work on the inside, along with a couple of API changes to deal with design issues. cf00a6d | jnthn++ | src/core/Supply.pm: Refactor batch to use whenever.		 13:34
13:34 dalek left
jnthn wrote a long commit message instead of a blog post :P 13:34
github.com/rakudo/rakudo/commit/46...d9d06c83d0 for anyone wanting to read it 13:35
13:35 dalek joined, ChanServ sets mode: +v dalek 13:40 cdg joined
jdv79 that was unusually long 13:42
13:43 AlexDaniel joined
jdv79 i think i followed about half of it 13:43
13:43 rindolf left
timotimo hm, so Publisher is a bit like Vow? 13:44
AlexDaniel m: say ⅐ + 5
camelia rakudo-moar : OUTPUT«5.142857␤»
AlexDaniel finally!
13:44 raiph joined
AlexDaniel m: ∞.say 13:45
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/Qi7jV5dQMM␤Bogus statement␤at /tmp/Qi7jV5dQMM:1␤------> 3<BOL>7⏏5∞.say␤ expecting any of:␤ prefix␤ term␤»
jnthn timotimo: Kinda, sorta :)
timotimo the last two lines look like something unfortunate has happened there
jnthn It's providing the same kind of "security" 13:46
timotimo mhm
jnthn last two lines?
timotimo of the commit message
jnthn oops
timotimo if the publisher doesn't know anything about the tappers, it won't know when all tappers have been closed?
jnthn :)
A publisher *does* have a list of tappers 13:47
timotimo perhaps i ought to re-read?
oh. publisher does, supply doesn't?
jnthn Right
jdv79 as always i look forward to the excellent docs that will just appear someday so i can fully grasp the power:)
13:48 RabidGravy joined
moritz jdv79: learn the virtue of impatience, learn it by trying (and reading the source and design docs), and write the excellent docs yourself! 13:48
AlexDaniel .u 𒐳
yoleaux U+12433 CUNEIFORM NUMERIC SIGN SHAR2 TIMES GAL PLUS MIN [Nl] (𒐳)
AlexDaniel .u １ 13:49
yoleaux U+FF11 FULLWIDTH DIGIT ONE [Nd] (１)
AlexDaniel m: say １１
camelia rakudo-moar : OUTPUT«11␤»
AlexDaniel m: say １１Ⅸ𒐳 13:50
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/OnFft4FMRm␤Bogus postfix␤at /tmp/OnFft4FMRm:1␤------> 3say １１7⏏5Ⅸ𒐳␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement mo…»
AlexDaniel m: say １１ * Ⅸ𒐳
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/wZv5QSC5E7␤Bogus postfix␤at /tmp/wZv5QSC5E7:1␤------> 3say １１ * Ⅸ7⏏5𒐳␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement…»
AlexDaniel m: say １１ * Ⅸ +𒐳
camelia rakudo-moar : OUTPUT«432099␤»
13:50 bjz left
13:57 bjz joined
cognominal why in the REPL, the default is not MONKEY-TYPING and augment ? I fear that the anwer is, go support a ~/.perl6-repl and do it. Not quite there yet. :( 14:09
14:10 Ven left 14:14 kaare_ joined 14:16 raiph left
jnthn The last case I remember of making -e/scripts/REPL work differently caused sufficient confusion we ended up undoing it. 14:18
moritz because the REPL is supposed to behave as similar as possible to the standard behavior
arnsholt Definitely 14:19
moritz jnthn: yes, "no strict" as default caused confusion, iirc
jnthn moritz: Yes, that's the one I was thinking of
arnsholt But supporting a ~/.rakudorepl.rc should be pretty simple though
dalek kudo/resupply: fa096b0 | jnthn++ | src/core/Supply.pm:
Refactor a bunch more methods to use whenever.
arnsholt Just inject eval "~/.rakudorepl.rc".IO.slurp into the code before we start reading code from stdin 14:20
jnthn Not a lot less lines of code with supply/whenever, but certainly feels clearer to me :)
moritz arnsholt: that might not do what people expect wrt scoping 14:21
14:22 zengargoylew joined
arnsholt Oh, right. It'd have to inject the contents of the file in Perl6::Compiler.interactive 14:23
That'd be the same as if the code were typed at the REPL
ely-se concatenate the file and stdin 14:25
arnsholt cognominal: It should be a pretty straightforward feature to implement. In src/Perl6/Compiler.nqp.interactive (lines 122 to 174 in HEAD), you'd have to read in the contents of the RC file, and run it with self.eval($contents) before the call to the superclass at the end of the method 14:26
cognominal Anssering to the similarity requirement. On the other hand, one use of the REPL is incremantal/interactive writing of a class. That was my initial motivation.
arnsholt It's a bit more work than just "$file".IO.slurp, since it's NQP code not Perl 6 code, but not tremendously so
moritz arnsholt: probably nqp::open, nqp::readall 14:27
arnsholt Yeah, that's the order of magnitude I assumed it'd be 14:28
cognominal If I follow my I idea, the prompt would indicate the current entity.
timotimo there's a new bug report on the RT basically about regex literals looking into the lexical environment for a class named "Regex" and when you're inside the class body of Regex and use a regex literal it goes boom
cognominal > class A {
A> method foo {}
A> }
>
timotimo that's not about MONKEY-TYPING or augmenting, though
arnsholt That's harder
cognominal in that sense, no augment here
timotimo that's just our inability to understand end-of-line inside things that aren't finished 14:29
14:29 Ven joined
arnsholt Yeah, which is a shame, but significantly harder to implement 14:30
cognominal I have not yet grasped all of CPS, but what I want mean some ongoing context
timotimo you want to turn our parser into CPS? 14:31
cognominal on the other hand, I want to use the class as well. So my example is only part of the spec.
timotimo, thinkout out loud, not sure it makes sense 14:32
14:32 skids joined
timotimo hmm 14:32
arnsholt Hang on, let's backtrack a bit. Which feature are you wanting here: working multi-line input on the REPL, or "augment class ..." to work at the REPL from the start?
zengargoylew i'd sorta be happy if you could just multi-line a block: {␤code...␤code...␤code..␤}␤
timotimo m: my @foo := gather { "hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }] / }; say @foo[0]; 14:33
camelia rakudo-moar : OUTPUT«Type check failed in binding; expected Positional but got Seq␤ in block <unit> at /tmp/AJ0uG7g2Xv:1␤␤»
timotimo m: my $foo = gather { "hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }] / }; say $foo[0];
camelia rakudo-moar : OUTPUT«hello␤»
arnsholt zengargoylew: Everyone would love that, I think. Unfortunately it's actually a decidedly non-trivial feature
cognominal yea, there is what I want and how to to it. Mixing the two :(
timotimo m: my $foo = gather { "hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }] / }; say $foo[1];
camelia rakudo-moar : OUTPUT«taken␤Nil␤»
timotimo m: my $foo = gather { "hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }] / }; say $foo[2]; 14:34
camelia rakudo-moar : OUTPUT«taken␤Nil␤»
timotimo m: my $foo = gather { "hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }]+ % \s+ / }; say $foo[2];
camelia rakudo-moar : OUTPUT«taken␤taken␤hello hello hello␤»
timotimo m: my $foo = gather { "hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }]+ % \s+ / }; say $foo[1];
camelia rakudo-moar : OUTPUT«taken␤hello hello␤»
zengargoylew played with rlwrap a bit. it's nice to use $EDITOR once in a while. sorta painful when it takes it line by line. :)
timotimo m: my @foo = gather { "hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }]+ % \s+ / }; say @foo[1];
camelia rakudo-moar : OUTPUT«taken␤taken␤taken␤hello hello␤»
zengargoylew hacked up a filter to remove /^> /
14:35 lucasb joined
cognominal ho, I did not noticed the multiline REPL has been removed 14:35
Skarsnik Hello, does someone already tried to compile the clang branch of smokegen? It seem unrelated but it's the tool used for Qt/Kde bindings
timotimo how do i do this stuff again :P
zengargoylew and add custom colored prompt... sorta *looks* a bit better anyways.
timotimo m: my $foo = gather { "hello hello hello hello" ~~ / [<ident> { take $/.Str; say "taken" }]+ % \s+ / }; say $foo.list[1]; 14:36
camelia rakudo-moar : OUTPUT«taken␤hello hello␤»
timotimo ah, of course
$/.Str matches from the start %)
so yeah, grammar parsing can work with gather/take already, but the really difficult part is that we'd have to switch the string being matched against from underneath the parser
14:37 jonas2 left
timotimo otherwise we don't have a benefit to the "CPS transform" 14:37
(in quotes because we're not actually transforming the code. not on moar anyway)
cognominal About this incremental compilation, probably what I ask for is lazy string :)
timotimo yeah, that's a way to do it. 14:38
but that's post-6.c
cognominal I know :)
zengargoylew what's CPS?
timotimo contiunation-passing-style
we had the beginning of a "moreinput" thing that'd let the grammar request extra text 14:39
cognominal from what I understand (very little), this is passing the general context to any routine
s/this is/CPS is/ 14:40
timotimo "continuation passing style" happens when you treat instruction pointers + stack as a first-class entity
cognominal I told the what, timotimo spells the how :) 14:41
timotimo call/cc is a tool to invoke a function and also pass in a "continuation" that, when called, will teleport your instruction pointer back to where the call/cc was in the code
cognominal It means one can snaphote the state of the interpreter and go back to it. 14:42
That makes all sort of backtracking thing for *free*, well at the cost of support CPS
14:43 cbk joined
cognominal I have seen mention of CPS in moarvm 14:43
*snapshot
timotimo moar has single-shot continuations
cognominal what means thing shot?
timotimo after a continuation has been invoked, it can't be invoked again 14:44
cognominal ok
timotimo there's a fun language called "oort" that has first-class labels (as in goto)
ely-se terrible idea with incredibly complexity and no benefits at all 14:47
timotimo oh hey ely-se
dalek kudo/resupply: bae2138 | jnthn++ | src/core/Supply.pm:
Most of the rest of the on -> whenever switch.
timotimo i'm conflicted with regards to my multidim spec tests. they throw an exception currently, so in order to get through them i'd either have to test with lives-ok or fudge with "skip" rather than "todo" :( 14:48
cognominal /msg ely-se a 14:49
I am not sure the allusion to Chopin or the Cure is transparent. 14:50
timotimo ? 14:51
14:51 rurban joined
cognominal I sent a "letter to ely-se" fr.wikipedia.org/wiki/La_Lettre_%C...%C3%89lise www.youtube.com/watch?v=-AUCSkHCrwY 14:53
Some playword need explanation. Like : Ga 14:54
G-grand a-petit for J'ai grand appétit :)
Ven uhm. 14:55
14:56 Hotkeys left, Hotkeys joined
cognominal Ven, valerie.ufoot.org/citations/jeuxdem...us?lang=fr 14:57
14:57 eliasr left
Woodi hallo #perl6 :) 14:58
do REPL have ability to save 'session' to file ? would be nice way to start implementation of new thing :) 14:59
or save session to temp file, use EDITOR on it and restart REPL with edited file... 15:01
Ven it doesn't yet, at least
15:02 telex left
masak maybe serialization could be nmade to do that, I dunno 15:02
it sounds hard/risky to add after the fact
and you get all kinds of fun serialization format versioning difficulties that we already have with precompiled modules 15:03
15:04 telex joined
timotimo masak: i think he really means just a dump of the lines you wrote 15:04
zengargoylew RLWRAP_EDITOR=vim exec rlwrap -m -M .pl6 -r -R perl6 # gets close... saves history, allows editing. ...
can do a bit of multi-line stuff, but it's quirky 15:05
TAB completes on seen things 15:06
15:07 AlexDaniel left
zengargoylew i think you have to remove Linenoise.pm or they conflict a bit. 15:08
lucasb can't I use S/// with the binding operator '~~' ? 15:10
masak timotimo: that would likely be enough in some cases, but any non-deterministic instruction would not restore the state well. 15:11
timotimo that's true
pmurias we could serialize the lexical scope 15:12
dalek c: 788b09c | (Zoffix Znet)++ | doc/Language/grammars.pod:
Fix code example to not emit warnings when run
 15:13
timotimo i think we should concentrate on making the ipython/jupyter integration good because things like the notebook will give us exactly what we need in that case
pmurias but it doesn't seem that usefull and they will bitrot on updates
timotimo because it keeps output around and stuff like that
pmurias ipython integration seem more more usefull than trying to reimplement smalltalk images
timotimo and lets you re-run individual parts on demand
verily. 15:14
15:16 xpen joined
lucasb m: my $s = 'abc'; say $s ~~ S/a/x/ 15:17
camelia rakudo-moar : OUTPUT«Potential difficulties:␤ Smartmatch with S/// can never succeed because the subsequent string match will fail␤ at /tmp/umnqxfm5LR:1␤ ------> 3my $s = 'abc'; say $s ~~ S/a/x/7⏏5<EOL>␤False␤»
masak ipython integration sounds good.
dalek c: e3655a5 | (Zoffix Znet)++ | doc/Language/grammars.pod:
Remove ~, since it is not needed in the example and is distracting from the subject matter, if the reader is not aware of it
 15:20
ely-se c: 15:21
Woodi timotimo: yes, I mean just lines typed into REPL 15:24
what are that ipert/ipython thingies ?
*iperl there
15:25 muraiki joined
JimmyZ jnthn++ # the resupply code is much clear . 15:25
moritz Woodi: ipython is a generic REPL/notebook implmementation
15:25 AlexDaniel joined
Woodi ok, googled it. thanx 15:26
timotimo i should stop calling it "ipython" and using their new name "jupyter" instead, eh? 15:27
15:27 rurban left
Woodi looks like pyWolfram in a tab :) 15:28
15:28 hankache joined 15:29 ZoffixW joined
muraiki jupyter integration would be awesome 15:29
ZoffixW Is token ws { ... } magical? I'm looking at the second example at docs.perl6.org/language/grammars#Action_Objects and I don't see it used anywhere other than being declared, but if I comment it out, the code fails 15:30
pmurias ZoffixW: yes
ZoffixW Also, I don't get why rule pair { <key=.identifier> '=' <value=.identifier> } works at all. I thought rules has :sigspace turned on, so how come it's matching even though there's space between the '=' and tokens, but in the stuff it's parsing there isn't any space 15:31
pmurias, what does it do?
pmurias ZoffixW: search S09 for :sigspace 15:33
ZoffixW Thanks
pmurias ZoffixW: :sigspace inserts <.ws> 15:35
moritz ... wherever there's a space in the regex
pmurias ZoffixW: ws is declared as \h* - so zero space matches that too
ZoffixW Ah
zengargoylew thought <ws> was a bit more than just \h* depending on the characters around it. 15:37
ZoffixW Kinda weird that if I comment out token ws, rule pair { <key=.identifier>'='<value=.identifier>} still works but rule pair { <key=.identifier>'='<value=.identifier> } does not (not the space before the closing brace 15:39
Even though third example here works regardless of the extra space before closing brace in the `rule`: docs.perl6.org/language/grammars#Named_Regexes
pmurias zengargoylew: it's declared as \h* in the example 15:40
zengargoylew ah, gotcha 15:41
but if you comment out the ws in the example doesn't it then fallback to the predefined ws? 15:42
ZoffixW Ah, I see now. The predefined ws takes that \n+ in token TOP. Thanks, zengargoylew++ 15:43
15:44 ab6tract joined
ab6tract o/ #perl6 15:44
regarding object hashes again...
ZoffixW \o
ab6tract m: my $now = DateTime.now; my %h; say %h{$now} if $now === %h.keys[0]; 15:45
camelia ( no output )
ab6tract so it's not actually storing the object in the key after all
or is that because it is now in a new container?
moritz ab6tract: it only autovivifies on write operations 15:46
ab6tract arg, that test case is borked
m: my $now = DateTime.now; my %h; %h{$now} = "now is now!"; say %h{$now} if $now === %h.keys[0];
camelia ( no output )
ab6tract m: my $now = DateTime.now; my %h; %h{$now} = "now is now!"; say %h{$now} if $now === %h.keys[0]; %h.keys[0].WHAT.say 15:47
camelia rakudo-moar : OUTPUT«(Str)␤»
15:47 CIAvash joined
ab6tract so it looks like this is still doing the old $obj.WHICH .. 15:47
moritz: am i still holding it wrong? :( 15:48
JimmyZ It still doesn't Objectk key, iirc
*doesn't support
ab6tract JimmyZ: ah.. TimToady did write the "object hash" a while ago. i asked after the semantics last night, because they are not documented afaict 15:50
m: say "green" === my $green = "green"
camelia rakudo-moar : OUTPUT«True␤»
moritz ab6tract: you should declare 'my %h{Any}' I believe
ab6tract i also thought this was a bit weird... bare string is not the same as containerized string, no? 15:51
moritz m: my $now = DateTime.now; my %h{Any}; %h{$now} = "now is now!"; say %h{$now} if $now === %h.keys[0]; %h.keys[0].WHAT.say
camelia rakudo-moar : OUTPUT«now is now!␤(DateTime)␤»
dalek c: 9e9b82a | (Zoffix Znet)++ | doc/Language/grammars.pod:
Explain why token ws {} is in the example
ab6tract moritz: Ah! thank you 15:52
TimToady m: my %h := :{now => "NOW!"}; say %h.perl
camelia rakudo-moar : OUTPUT«:{:now("NOW!")}␤»
TimToady m: my %h := :{(now) => "NOW!"}; say %h.perl 15:53
camelia rakudo-moar : OUTPUT«:{Instant.new(<425849244517/294>) => "NOW!"}␤»
moritz ab6tract: doc patches are also very welcome :-)
TimToady m: my %h := :{42.000 => "NOW!"}; say %h.perl 15:54
camelia rakudo-moar : OUTPUT«:{42.0 => "NOW!"}␤»
ab6tract moritz: that's my end game :)
15:54 loren joined
TimToady there's your literal form 15:54
ab6tract TimToady: does it require a bind?
loren evening, perl6
TimToady yes, in that case
loren m: say "".elems;
camelia rakudo-moar : OUTPUT«1␤»
loren Why a empty Str's elems is '1' 15:55
15:55 hankache left
ab6tract m: "".comb.elems.say 15:55
camelia rakudo-moar : OUTPUT«0␤»
15:55 khw joined
ab6tract loren: because that is the equivalent of calling elems on a list of one item 15:55
if we looked into the source, we would probably find elems defined on Any such that it calls .List.elems 15:56
loren .., i think it's a little strange 15:57
masak Perl 6 deliberately confuses a 1-element list and a single item.
it's... unusual, yes.
ab6tract (or .list.elems ... still not 100% on the functional differences between the two
15:57 hankache joined
masak but Perl 5 basically does the same, except with less type system. 15:57
ab6tract loren: it's only strange if you are expecting .elems on a string to refer to characters, at which point you would probably also expect to be able to index into the string
loren I haven't noticed that
ab6tract neither of which perl has ever cottoned to, afaik 15:58
loren ..Ok, get it . 15:59
m: say "xxxx".elems; #This will print also '1', aha? 16:00
camelia rakudo-moar : OUTPUT«1␤»
loren ..-_-
ab6tract m: say 16:01
camelia rakudo-moar : OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at /tmp/BLsu1h0rut:1␤------> 3say7⏏5<EOL>␤Other potential difficulties:␤ Unsupported use of bare "say"; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argume…»
ab6tract m: say "xxxx".comb
camelia rakudo-moar : OUTPUT«(x x x x)␤»
TimToady languages that confuse strings with arrays can't really do NFG properly, I suspect
jnthn TimToady: I've been having a bit of a cleanup of Supply, to resolve a number of issues and untangle things.
ab6tract yeah, some languages like to treat strings as lists of characters
loren Yeah, i come from c++, so .. 16:02
zengargoylew m: put
camelia rakudo-moar : OUTPUT«␤»
TimToady jnthn: so I gather
jnthn No, gather is for lists :P
ab6tract TimToady: so what's your take?
zengargoylew wonders if say error should mention put also.
jnthn TimToady: Very little of it affects existing code, BUT the place it will is a big one :(
TimToady I take it I should read your very long commit message 16:03
jnthn Yes, please :)
TimToady should get to it pretty soon in the backlog, if y'all don't keep saying interesting things :)
jnthn Ignore the last line of the message; I did a mis-edit.
arnsholt jnthn: Is there a synonym of the shape syntax that lets me use a $ sigil instead of the @?
jnthn arnsholt: No, you'd need to pass :shape to the consturctor 16:04
arnsholt Right. Thanks!
timotimo m: class Foo { has @.test[10] }
camelia ( no output )
timotimo ^- we can support that in NativeCall now! ohmygosh!
hankache evening 16:05
:)
Skarsnik timotimo, maybe add a test in NC for it? (whatever it is!) xD 16:08
16:10 xpen left, perlawhirl left
hankache how's everyone doing? 16:10
moritz can't speak for everyone 16:11
16:11 zakharyas left
CIAvash didn't know there was a `put` routine! 16:12
arnsholt jnthn: Oh, right. 'Cause the shape isn't part of the type, is it? Only the type parameter is
jnthn arnsholt: It is under the hood; we hide it somewhat 16:13
arnsholt: To be clear: the number of dimensions are part of the type
The sizes of them aren't 16:14
TimToady so we can have dynamically sized dimensions eventually
arnsholt Right, right. That meshes with what I've seen in Array.pm, come to think of it. So then something like "my CArray[int] $foo" would be unspecified as regards to number of dimensions, right? 16:15
jnthn m: my ($a, $b) = 3, 4; my @x[$a;$b]; say @x.shape; # Like this, TimToady? :)
camelia rakudo-moar : OUTPUT«(3 4)␤»
TimToady oh, jnthn distributes the future! 16:16
loren good night, everyone
TimToady *ooh
dalek c: 910fa06 | (Zoffix Znet)++ | doc/Language/exceptions.pod:
Fix grammar
rl6-roast-data: 45a472a | coke++ | / (9 files):
today (automated commit)
16:17 loren left 16:18 ZoffixW left
TimToady .tell zengargoyle yes, I decided to keep | where it was, and if we want a slipper at so/not precedence, call it something else, like "sl" or some such, but then its not really saving keystrokes over |(1..10) 16:18
yoleaux TimToady: I'll pass your message to zengargoyle.
jnthn There's already slip at listop precedence :) 16:19
TimToady yes, but that doesn't help inside the comma
jnthn true :)
16:20 dwarring left
TimToady it's a little late to claim "slip" for that, and make "slip-all-the-things" for the listop... 16:20
so we'll probably leave things a the |(1..10) status quo 16:21
16:21 ely-se left
TimToady *at 16:21
**unless you're French
dalek c: 9bb938f | (Zoffix Znet)++ | doc/Language/exceptions.pod:
Grammar / clarification
 16:22
16:24 g4 left
arnsholt jnthn: For a declaration like "my @foo[2; 3] is Foo of Bar" the Foo.new constructor gets called twice, first without and then with the shape parameter. Whence the first one? 16:29
16:29 domidumont left
jnthn Um... :) 16:30
timotimo jnthn: hum. so we can't actually do CArray with @.test[10]?
jnthn arnsholt: I'm not actually sure.
arnsholt timotimo: Not *yet* =) 16:31
timotimo well, if the size of the dimensions isn't part of the type ... then what?
16:33 Ven left
arnsholt jnthn: Hmm, nm. On closer inspection looks like I'm wrong =) 16:34
I think
TimToady jnthn: supply vs publish is a bit of a mixed metaphor; did you consider "Supplier"? 16:37
16:37 dwarring joined
[Coke] is '⅓'.unival.WHAT.gist, '(Rat)', "'⅓' is a Rat"; --- that shouldn't need a gist to check if it's a Rat 16:37
TimToady probably predates 'is' handling type objects 16:39
16:40 FROGGS left
lucasb m: my @a[-1] 16:40
camelia rakudo-moar : OUTPUT«Memory allocation failed; could not allocate 18446744073709551608 bytes␤»
TimToady m: use Test; is ⅓.WHAT, Rat, "I smell a rat!"
camelia rakudo-moar : OUTPUT«ok 1 - I smell a rat!␤»
arnsholt jnthn: Actually, I was wrong about being wrong. There are two calls:
m: class Foo does Positional { method new(:$shape) { say "shaped {?$shape}/{$shape.^name}"; nqp::create(self) }; }; my @foo[2;3] is Foo 16:41
camelia rakudo-moar : OUTPUT«shaped False/Any␤shaped True/List␤»
arnsholt So if I make the shape required it fails:
m: class Foo does Positional { method new(:shape($)!) { say "shaped"; nqp::create(self) }; }; my @foo[2;3] is Foo
camelia rakudo-moar : OUTPUT«===SORRY!===␤Required named parameter 'shape' not passed␤»
jnthn TimToady: I pondered Supplier, and worried if it'd be a tad close to Supply 16:43
timotimo "SupplySource"
jnthn I wanted one word. :)
TimToady: It does fit nicely though. 16:44
TimToady how often will the "Publisher" name occur in normal code?
jnthn I'm not entirely sure (more)
mst Quartermaster
timotimo Sourcepply 16:45
jnthn If code that's doing async I/O, time-based stuff, and working with existing supplies, nearly never
*In
In code that's exposing live supplies (like, a UI framework that wants to provide a supply of UI events), maybe it'd show up somewhat. Depends how you factor it.
In the "I want to expose a supply people can tap for status updates", it'd be common
timotimo .o( we can encode the size of an inlined array in CArray as the number of dimensions ) 16:46
jnthn I went with publisher largely because in the case you use it, you actually are wanting to do a kind of publish/subscribe model.
TimToady we could put them on the Dole :)
jnthn The other thing on this
16:47 raiph joined
zengargoylew Producer 16:47
jnthn We'd like a method where you can take one supply (an on-demand one typically, but either way one where you're doing some processing) and be able to share its results out (as a live supply).
And publish is a nice name for such a method. 16:48
TimToady "share" is shorter, and you just used that word :) 16:49
jnthn Very true. :)
I could happily call it share. But I don't want to call the noun a SharingPlatter :)
Sharer feels really odd :) 16:50
I'd rather go with Supplier than either of those :)
mst I liked Quartermaster :( 16:51
"the thing you get your supplies from"
jnthn Dealer
timotimo Treasurer
16:51 domidumont joined
mst (I am not seriously proposing it because it's probably non-obvious to many people but ...) 16:51
arnsholt timotimo: Not if we want to support multi-dimensional C arrays (which are a thing too =) 16:53
TimToady I suspect people could easily learn to distinguish Supply from Supplier, and the relationship would, I suspect, be more obvious 16:54
jnthn arnsholt: The memory layout of MultiDimArray matches how C lays things out, I believe
timotimo arnsholt: damn it.
jnthn (I actually looked that up when working on the REPR)
timotimo jnthn: yeah, we just can't easily encode multi-dimensional array sizes in the number of dimensions
so that CStruct can work with inlined arrays 16:55
TimToady though to be sure, "share" doesn't imply doling things out in a sane fashion, like an allowance
ab6tract mst: Quartermaster reminds me of a Jamie Oliver recipe where he offhandedly suggests that you ask your fishmonger to debeard your mussels 16:56
TimToady maybe there's a better verb for how suppliers feed a supply chain
16:56 domidumont left
TimToady unfortunately, "supply" is taken... 16:56
jnthn TimToady: Maybe so. I think we can go with Supplier as the name, anyway.
geekosaur provision?
jnthn (For what I called Publisher)
distribute maybe works... 16:57
lucs dispenser?
16:57 domidumont joined
arnsholt jnthn: Looks like the first, shape-less invocation of the constructor comes from Perl6::World; install_lexical_container calling build_container. Ring any bells? 16:57
TimToady distribute isn't too bad
16:57 domidumont left 16:58 domidumont joined
jnthn arnsholt: Hmmm 16:58
Yeah, I know what bit of code that is :)
TimToady though in the commercial metaphor, distributor is downstream rather than upstream like your suppliers, but maybe that's just fine here, since a supplier's downstream entities are distributors to them 16:59
automakers are distributors of batterys :) 17:00
17:01 lucasb left
arnsholt jnthn: Since it's being constructed by ::World, there's a good reason for it I can't avoid? =) 17:02
jnthn arnsholt: There may not be a good reason 17:03
arnsholt: Could you try this patch: gist.github.com/jnthn/34266189529166caa286 17:05
(Bit busy on something else at the moment... :))
arnsholt Yeah, I'm semi-busy right now too, but I'll try it later tonight and get back to you. Cheers! 17:06
jnthn ok, cool :) 17:08
17:12 FROGGS joined 17:13 vytas left, vytas joined 17:14 vytas left 17:16 ZoffixW joined
ZoffixW Should unit class IRC::Client::Plugin::HNY:ver<1.001001> does IRC::Client::Plugin; be sufficient or do I actually have to `use IRC::Client::Plugin;` first? 17:17
I'm getting "Invalid typename 'IRC::Client::Plugin'"
jnthn Need a use first.
ZoffixW aw :(
TimToady we could make 'unit' special that way, I suppose...but then people might think it's 'class' that is doing the magic 17:19
ZoffixW I'd kinda expected `does` to do the magic 17:20
TimToady hard to know where the balance point is on that, since you could pull in the wrong thing by accident with a typo
though usually a typo won't exist 17:21
jnthn It'd be a bit surprising to have a unit => curlies refactor break the code.
17:22 donaldh joined
jnthn ZoffixW: In Perl 6, class/role <=> compilation unit are fairly decoupled. 17:22
donaldh pwd
jnthn /home/donaldh/illegal_things
ZoffixW :)
donaldh ahem 17:23
jnthn
.oO( So that's where he keeps his secret Perl 7 plans... )
donaldh Does perl6 have a way of augmenting USAGE with POD?
TimToady autouse would also encourage people to spread dependency info all over the program, which could be construed as an antipattern, unless it's a keep-things-nearby pattern :)
ZoffixW Is there something wrong with this? I'm getting "/home/zoffix/devbox/tmp/perl6-IRC-Client/lib/IRC/Client.pm6" and looks like it on the code inside the tap. I thought it'd close over or something 17:24
m: class Foo { multi method timer () { 1 }; multi method timer(Str) {say "42"} }; my @plugins = Foo.new; Supply.interval( .timer ).tap({ .timer("43") }) for @plugins.grep(*.^can('timer'));
camelia ( no output )
jnthn I think if you put declarator docs on the MAIN subs, they go into usage.
ZoffixW And I can't repro in camelia :/
timotimo ZoffixW: .timer is $_.timer, so maybe you want something else there? 17:25
17:25 ab6tract left
timotimo i mean in the Supply.interval( ... ) thing 17:25
m: say .timer
camelia rakudo-moar : OUTPUT«Method 'timer' not found for invocant of class 'Any'␤ in block <unit> at /tmp/I3IpnmpcYz:1␤␤»
ZoffixW timotimo, hehe... well, inside the .tap({}) I do want the $_.timer, but where $_ is an item from @plugins
17:26 z8 joined
timotimo ooh! 17:26
yeah, now i see the postfix for :)
i hallucinated a ; there
17:28 hankache left, z8 left
ZoffixW This works: fpaste.scsys.co.uk/501918 17:29
I guess there's no way to do a postfix for() AND have something to close over.
timotimo please be more specific?
oh!
well, duh, { $_.foo } will always give you a closure that takes a single argument and puts it into $_ 17:30
try OUTER::$_ ?
m: $_ = "hi"; do { say OUTER::$_ }
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/JpJ_Ji_R7C␤Malformed lookup of ::$_; please use ::('$_'), ::{'$_'}, or ::<$_>␤at /tmp/JpJ_Ji_R7C:1␤------> 3$_ = "hi"; do { say OUTER7⏏5::$_ }␤»
timotimo m: $_ = "hi"; do { say $OUTER::_ }
camelia rakudo-moar : OUTPUT«hi␤»
timotimo m: $_ = "hi"; say $OUTER::_ for <1 2 3>
camelia rakudo-moar : OUTPUT«hi␤hi␤hi␤»
timotimo m: $_ = "hi"; say $_ for <1 2 3>
camelia rakudo-moar : OUTPUT«1␤2␤3␤»
timotimo does that help?
mst ZoffixW: timer seems a really weird name for a method providing an interval? 17:31
also now there's no way for a plugin to decide maybe it doesn't want to have one
ZoffixW Yeah, $OUTER::_ works :) Thanks, timotimo++ (Perl 6)++
timotimo mst: it's even stranger to me to have a candidate that takes a string and returns a constant string :D
mst timotimo: hm?
timotimo not important
mst ZoffixW: seems like you'd be better allow the interval value to return 0 and then providing a default implementation that does that in the plugin role 17:32
timotimo should there be a way for things to declare they "don't want to have one"? why should a child be allowed to specify it doesn't want to have broccoli?
sometimes you really do know better!
timotimo AFK for a bit
mst timotimo: uhhh, he's testing can()
timotimo oh!
mst so he's already decided it's optional
ZoffixW Yeah, it is optional.
mst I'm just proposing a nicer way to make it optional
timotimo well, shows how much of the code example i read!
mst lol
ZoffixW And yeah, currently I have an issue that I want my role to have something like: multi method interval() { ... } and then have ... stuff 17:33
k, I'll do the return 0 stuff
and call it interval mst++
TimToady on the subject of Unicode-isms, I think I'm also in favor of building in the simple literal powers with superscripts as postfixes, for similar reasons: 1) It's trivial to implement, 2) it's a well-known notation, 3) the tight binding of postfixes actually works out more intuitively than **, 4) most powers in real code are literal (indeed, most powers are ² or ³ :), and 5) mathies will love it to pieces.
ZoffixW But.. but... Math::Operators::Superscripts ... 17:34
17:34 raiph left
ZoffixW modules.perl6.org/repo/Math::Operat...perscripts 17:34
:D
TimToady well, it's a nice prototype :) 17:35
ZoffixW TimToady, but I agree. Since there's the fractions stuff it'd kinda expected powers would work too.
TimToady but one of our big principles is, "Math just is."
the question to me is actually whether we draw the line at making 1½ work :) 17:36
ZoffixW lol
Hotkeys lol
TimToady and I'm of two minds about that... 17:37
17:37 pmurias left
Hotkeys I can't be bothered to figure out how to easily type superscripts so I'll stick to the x ** y notation :p 17:37
17:37 pmurias joined, dakkar left
TimToady someone will eventually write a "pretty this to Unicode" program 17:38
ZoffixW .tell hahainternet I recall you really wanted those superscript powers: irclog.perlgeek.de/perl6/2015-11-25#i_11598216
yoleaux ZoffixW: I'll pass your message to hahainternet.
ZoffixW Hotkeys, sure, but if you type them all the time and you're typing them off from some paper that uses proper powers and your program will be looked at by something who's not "really" a programmer... :) 17:39
TimToady for sure we can't make ⅔r³ work without introducing a bad prefix vs term ambiguity, but 1½ has no such difficulty 17:40
ZoffixW Oh, it's Operators::Math::Superscripts
Hotkeys ZoffixW: fair enough
ZoffixW m: multi postfix:<½> ($n) { $n - .5 }; say 1½ 17:42
camelia rakudo-moar : OUTPUT«0.5␤»
ZoffixW ehehe. Perl 6++ :)
minus the math error :P 17:43
Hotkeys isn't 1½ 1.5?
ZoffixW yeah
Should be + instead of - :)
Hotkeys just checking for my sanity's sake
:p
gtodd after christmas I can't wait for perl6 (and 5) start saving the planet ... the way it saved Human Genome Project :-)
worrydream.com/ClimateChange/
ZoffixW \o/ 17:44
Hotkeys perl 6 saved the human genome project?
ZoffixW Wait, perl saved the genome... that
gtodd perl5 :-D
ZoffixW I guess this is based on a true story: xkcd.com/208/
:)
Hotkeys I didn't click the link but I knew exactly what it was 17:45
also this comic describes me when I learned regex 17:46
ZoffixW :P
Skarsnik 250Mb? now we need a big data DB to handle this at least!
ZoffixW m: multi prefix:<⅔> ($n) { $n * 2/3 }; multi postfix:<⅔>($n) { $n + 2/3 }; multi postfix:<³>($n){ $n**3 }; my \r = 2; say [1⅔, ⅔r³ ] 17:47
camelia rakudo-moar : OUTPUT«[1.666667 5.333333]␤»
ZoffixW This is magical :P
Hotkeys now make it work for any power/fraction :p 17:48
gtodd hehe ... in the section on model-driven [reading/authoring/debate] on worrydream.com/ClimateChange/ Mr. Victor asks "What if there were an npm for scientific models?" ... but but why can't CPAN be the gold standard?!
17:48 pmurias left
Hotkeys m: ⅔.uninames 17:49
camelia rakudo-moar : OUTPUT«Cannot call uninames(Rat); none of these signatures match:␤ (Str:D $str)␤ in block <unit> at /tmp/ed9LMIRzMf:1␤␤»
Hotkeys huh
ZoffixW m: ⅔.Str.uninames
camelia ( no output )
ZoffixW m: say ⅔.Str.uninames
camelia rakudo-moar : OUTPUT«(DIGIT ZERO FULL STOP DIGIT SIX DIGIT SIX DIGIT SIX DIGIT SIX DIGIT SIX DIGIT SEVEN)␤»
Hotkeys m: ⅔.nude
camelia ( no output )
Hotkeys er
ZoffixW heh 17:50
Hotkeys m: say ⅔.nude
camelia rakudo-moar : OUTPUT«(2 3)␤»
Hotkeys fancy
I didn't realize these already worked as rats
ZoffixW m: "⅔".uninames
camelia ( no output )
ZoffixW curses
m: say "⅔".uninames
camelia rakudo-moar : OUTPUT«(VULGAR FRACTION TWO THIRDS)␤»
ZoffixW Oh, I didn't realize I was being vulgar...
Hotkeys lol
jnthn TWO &*%!ING THIRDS 17:51
Hotkeys m: say "³".uninames
camelia rakudo-moar : OUTPUT«(SUPERSCRIPT THREE)␤»
jnthn Time for dinner; bbl 17:52
gtodd m: my %thing.push: <q w e r t y> ; say "there" if "q" ~~ %thing ; say "there" if %thing.values ~~ /w/ ;
Hotkeys m: say 3³
camelia rakudo-moar : OUTPUT«there␤there␤»
rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/jvNkWWErA1␤Bogus postfix␤at /tmp/jvNkWWErA1:1␤------> 3say 37⏏5³␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement modifier␤ …»
Hotkeys I like the use of bogus 17:53
gtodd why doesn't: say "there" if "w" ~~ %thing.values work?
or why do I feel like it should work :-\ 17:54
Hotkeys m: my %thing.push: <q w e r t y>; say %thing
camelia rakudo-moar : OUTPUT«e => r, q => w, t => y␤»
ZoffixW m: my @a = 1, 2, 0, False, 42, True; @a.grep(*).say 17:55
camelia rakudo-moar : OUTPUT«(1 2 0 False 42 True)␤»
ZoffixW ENOTMAGICALENOUGH
m: my @a = 1, 2, 0, False, 42, True; @a.grep(so *).say
camelia rakudo-moar : OUTPUT«(1 2 42 True)␤»
ZoffixW cool 17:56
Hotkeys so?
ZoffixW m: say 'foo' ~~ /foo/; say so 'foo' ~~ /foo/;
camelia rakudo-moar : OUTPUT«｢foo｣␤True␤»
ZoffixW Converts to boolean
Hotkeys I didn't know this existed
I was looking for something like this a couple weeks ago 17:57
ZoffixW :)
17:57 molaf joined 17:58 vendethiel joined, donaldh left, raiph joined, rurban joined 17:59 lucasb joined
ZoffixW Hm. I wonder how that new CompUnit stuff will work with circular references. I've got IRC::Client that `use`s IRC::Client::Plugin just for the sake of saying has IRC::Client::Plugin @.plugins; and in IRC::Client::Plugin I have `use IRC::Client` for the sake of having a method interval (IRC::Client) { ... } :/ 18:02
Maybe I should avoid using those as types.
ilmari can't you forward-declare the class? 18:03
ZoffixW Not sure what that is... is it class IRC::Client { ... }
literally
nine yes
ZoffixW Thanks. I'll try
Hotkeys I've been doing irc plugin stuff too what a coincidence 18:04
and by doing
I mean not doing
18:04 raiph left
Hotkeys but intend to do at a future time 18:04
because I was trying to figure out the best way to load and unload plugins and then just kind of stopped working on my irc thing 18:05
ZoffixW I'm making a Perl 6 version of my New Years bot (metacpan.org/pod/App::IRCBot::NewYears) and showcase Perl 6 at Freenode's New Years party. People constantly notice the bot and ask what lang it's written in. Would be nice to broadcast Perl 6's existence :P
Hotkeys I'm just making (and by making I mean the same as I did with doing above) a general IRC bot for shenanigans etc. 18:06
ZoffixW Cool :) 18:07
18:07 raiph joined
Hotkeys Do you know a simple way to load and unload plugins? 18:08
specifically the unloading part
ZoffixW Seems like I *can't* forward-declare the class: "The following packages were stubbed but not defined: IRC::Client::IRC::Client::Plugin"
Hotkeys, what do you mean by unloading?
Just have an $!is-active attribute and set/unset it
Hotkeys I guess that would be the smart thing to do 18:09
Skarsnik unloading the code :)
Hotkeys what about like 18:10
hot reloading plugins
is that an unadvisable thing
18:11 rindolf joined, raiph left
ZoffixW Probably :) 18:11
No idea how to do that in P6
Hotkeys Me either
Skarsnik The issue is decision like purging the symbole table for the module or not
18:11 psch joined
Skarsnik require reload the file 18:11
like in p5
pink_mist I think hot reloading plugins is an inadvisable thing in every language I've heard of :P if it wasn't in P6, I'd be (even more) amazed (than I already am) 18:12
Skarsnik in perl5 to do that it was something like that : delete the module entry in @INC, requiere on it
pink_mist Skarsnik: that wouldn't be clean ... you really ought to do more than that 18:13
Hotkeys I know someone who did it with a python bot
not sure if it was advisable
but
he did :p
Skarsnik but it work :)
ab5tract pink_mist: is there an actually "clean" way to do it in P5?
as i understand it, is a long open and unsolved problem 18:14
*it is
unless you had the forethought to never import a sub and always use it fully namespaced
Skarsnik and it really make sens for a IRC BOT
gtodd m: my %thing.push: <q w e r t y> ; say "there" if "q" ~~ %thing ; say "there" if "w" ~~ any(%thing.values) ;
camelia rakudo-moar : OUTPUT«there␤there␤»
nine That ClimateChange site is somewhat depressing
gtodd why is the "any" needed for %thing.values ( alist?) but not for %thing ? is there a "default" for "~~" against arrays ? 18:15
ZoffixW Oh wow. Two modules using each other actually freezes perl6 18:16
nine ZoffixW: yes, we did have protection against that, but it's broken. It's on my TODO list for curli
ZoffixW k :)
gtodd nine: ... we are going to have fight on the same team to survive :-|
ZoffixW nine++ # fixing things
gtodd nine: should become obvious at some point ... but possibly a bit too late 18:17
18:18 domidumont left
nine gtodd: it's the interactive "world carbon emissions (in GtCO2/yr) and carbon budget" diagram that makes it seem almost hopeless 18:18
ZoffixW plans to just hitch a ride with Elon Musk to Mars 18:19
gtodd :-D you said "almost" =-D
Hotkeys gfodd what's wrong with "thing" ~~ any @foo
18:19 spider-mario joined
TimToady too much carbon on Mars as well, relative to oxygen :) 18:20
Hotkeys smart match "thing" with any of @foo
ZoffixW :)
nine TimToady: well you can hardly make it worse there ;)
Hotkeys we just need to put more oxygen there TimToady no problem
TimToady well, really you mostly need more hydrogen
ZoffixW TimToady, not true actually! There're two oxygens for every carbon! :P 18:21
Hotkeys lol
TimToady and install a magnetic field, while you're at it
DrForr So it's already got a head start on carbon sinks.
Hotkeys we just need to reactivate mars geologically
no problem
get that core all hot and bothered
TimToady that would take...some patience
DrForr Bruce Willis will save us! 18:22
ZoffixW Meh, dig a hole. Underground city. Problem solved :P
Hotkeys I prefer dome cities
TimToady till the next fire drill
Hotkeys much more futuristic
gtodd Hotkeys: nothing wrong with it ... just wonder why the junctions aren't automagical ... I thought they were at one point 18:23
ZoffixW And there's likely more of that briny water they recently discovered. Split that up with electrolysis. Boom: you got oxygen to breath and hydrogen to make fuel (CH4) and moar oxygen
gtodd m: (1,2,3,4,5) ~~ (*,2,3,*)
camelia ( no output )
DrForr That *does* increase the risk that Arnold Schwarzenegger will have to save you, not worth the risk IMO.
ZoffixW :)
gtodd m: say "magic" if (1,2,3,4,5) ~~ (*,2,3,*)
camelia ( no output )
ZoffixW m: say (1,2,3,4,5) ~~ (*,2,3,*)
camelia rakudo-moar : OUTPUT«False␤»
TimToady m: say (1,2,3,4,5) ~~ (**,2,3,**) 18:24
camelia rakudo-moar : OUTPUT«True␤»
ZoffixW 0.o whoa!
TimToady * now is just a single smartmatch
gtodd oh right
forgot about that change
TimToady m: say (1,2,3,4,5) ~~ (**,2,*,4,**)
camelia rakudo-moar : OUTPUT«True␤»
18:24 cdg left
TimToady m: say (1,2,3,4,5) ~~ (**,Int,*,Int,**) 18:25
camelia rakudo-moar : OUTPUT«True␤»
TimToady because we made that work
ZoffixW 0.o
psch ZoffixW: your forward declaration for IRC::Client::Plugin didn't work because it was inside IRC::Client, btw
18:25 cdg joined
ZoffixW I thought I tried the other way too 18:25
ZoffixW tries again
gtodd and "**" is read as "any number of smartmatches" ? 18:26
psch the other way should work, although i don't know how that interacts with the statement level package declarator
TimToady kinda ties in with the multi-dimensional notion
ZoffixW psch, "Too late for unit-scoped class definition; Please use the block form." **uses block form** "The following packages were stubbed but not defined: IRC::Client::Plugin"
psch ZoffixW: huh. is that gisted or so somewhere? 18:27
ZoffixW psch, here: github.com/zoffixznet/perl6-IRC-Client 18:28
psch, you can run examples/bot.pl6 to see the error
arnsholt jnthn: Patch removes the shape-less constructor. Running spectest now
ZoffixW psch, and here I'm doing the forward declaration and I only really need it for line 13: github.com/zoffixznet/perl6-IRC-Cl...ent.pm6#L2 18:29
Hotkeys pff what kind of nick limit is 9 chars
psch ZoffixW: i'll look closer in a moment 18:30
ZoffixW Hotkeys, RFC1459? :)
Hotkeys pff
18:31 Hotkeys is now known as Ihavemorethan9ch
Ihavemorethan9ch close enough 18:31
nine Seems like build on JVM is broken again: Error while compiling op unipropcode (source text: "nqp::unipropcode(\"NumericValueNumerator\")"), no registered operation handler
FROGGS nine: I'd continue with writing tests, or is there anything else I should look into?
18:31 Ihavemorethan9ch is now known as Hotkeys
gtodd smartmatch against a hash seems to match "any (%hash.keys)" automagically as in my %hash = <A B C D> ; "A" ~~ %hash ; 18:31
ZoffixW m: say "Ihavemorethan9ch".chars
nine FROGGS: tests sound good :) I have cleaned a couple of spec tests yesterday in the roast/curli branch
camelia rakudo-moar : OUTPUT«16␤»
psch nine: i'm working around that right now 18:32
FROGGS nine: my plan was to write tests for creating CURI's and installing into them
psch with dummy getuniprop_str and unipropcode (and only matching unum when when we're on moar)
FROGGS nine: dunno what tests in that area exist already
nine FROGGS: other missing pieces are $?REPO and detecting loops when loading modules. The latter involves communicating a list of loading modules to the precomp process 18:33
gtodd but then for the values you need to explicitly say "B" ~~ any (%hash.values) ; ... gaak I don't need to worry if it once upon a time used to not be like that :-)
FROGGS nine: okay, I stick with tests :o)
Hotkeys interesting, rfc 2813 also has "
Each user is distinguished from other users by a unique nickname having a maximum length of nine (9) characters. "
nine FROGGS: acking for CompUnit in roast is enlightening :)
Hotkeys oops newline
FROGGS nine: oh, good idea 18:34
dalek kudo/curli: 9345857 | jnthn++ | src/core/ThreadPoolScheduler.pm:
Fix thinko in scheduler that swallowed exceptions.

Deals with a bunch of silent failures that made debugging harder.
18:34 dalek left
ZoffixW heh 18:34
18:35 dalek joined, ChanServ sets mode: +v dalek
FROGGS nine: cool, there are no install tests yet 18:35
nine FROGGS: the tests that are still failing may also give some inspiration 18:36
FROGGS true
18:38 kjs_ left, rigsby joined 18:39 vike left, kjs_ joined
psch oh ffs 18:40
r: say nqp::getcomp('perl6')
pink_mist <ab5tract> pink_mist: is there an actually "clean" way to do it in P5? <-- sorry, was off ... there's modules that can remove things from a namespace, but I don't know if they're 100% at what they do
camelia rakudo-jvm 4d8a90: OUTPUT«===SORRY!===␤java.lang.NullPointerException␤»
..rakudo-moar: OUTPUT«Method 'gist' not found for invocant of class 'Perl6::Compiler'␤ in block <unit> at /tmp/tmpfile:1␤␤»
psch no .gist i'm ok with
grmbl
zengargoylew somebody once pointed out that the thing to do was load the module in a Safe and then dispose of the Safe when done with the module... 18:42
18:42 kjs_ left
Hotkeys a Safe? (not in docs.perl6.org) 18:43
ZoffixW It's in a safe place... 18:44
badum tsss
Hotkeys oh you
zengargoylew nah, in P5
Hotkeys oh
whoops
is there anything similar in the six? 18:45
hahainte1net oh thanks ZoffixW 18:47
zengargoylew i sorta expect/expected p6 to be all restricted scopelike that a use/require of something is only visible to the block that does it (and inner things).
18:47 hahainte1net is now known as hahainternet
hahainternet and thanks TimToady for unicode superscripts, but you should not leave out $x₁ and $x₂ being legit variable names :) 18:48
yoleaux 17:38Z <ZoffixW> hahainternet: I recall you really wanted those superscript powers: irclog.perlgeek.de/perl6/2015-11-25#i_11598216
18:48 Sqirrel joined
zengargoylew sorta like: blah; ...; { use Foo; blah; ...; } # Foo is totally not a thing out of the block. 18:48
TimToady hahainternet: I'm kinda dubious of that one, in particular due to 1-based vs 0-based assumptions 18:49
Hotkeys just let $x_0 also exist :p
hahainternet TimToady: alphabetic subscripts are also permitted
so $x_a and $x_b for example
i just think these characters should be permitted in variable names 18:50
TimToady Hotkeys: $x_0 already exists :P
hahainternet rather than any special handling
Hotkeys I meant with subscript :p
I just cba to figure out how to type subscripts in the windows
without too much effort 18:51
hahainternet depends on how your compose key works, i'm not sure how to do subscript as on Debian
Hotkeys I don't think I have a compose key
hahainternet but regardless, doing y_a and y_b for your different variable names are common
Hotkeys is that like the any key?
uruwi[A] x₀
18:51 uruwi[A] is now known as uruwi
hahainternet Hotkeys: it's shift + altgr for me 18:51
zengargoylew compose + _ + 1
uruwi has WinCompose
18:52 vike joined
gtodd m: say "Ƹ̵̡􏿽xD3􏿽x9C􏿽xCC􏿽xB5􏿽xCC􏿽x84􏿽xCC􏿽xA8􏿽xC6􏿽xB7" ; 18:52
camelia rakudo-moar : OUTPUT«Ƹ̵̡Ӝ̵̨̄Ʒ␤»
zengargoylew wanted to update the How to input Unicode part of the docs, but decided it was too hard. :P 18:53
ZoffixW Hotkeys, not knowing how to type it and cba to learn how to type it are poor arguments for letting $x₀ be a thing :P
FROGGS looks a bit ill in my client
Hotkeys did you mean not letting 18:54
and if you did I am not at all against them
zengargoylew plus, Debian and probably others using iBus has sorta broken XCompose functionality anyways.
hahainternet FROGGS: the subscript and superscript fonts are mostly ignored
and i typed those on debian so
ZoffixW Hotkeys, no I meant "for letting" :) If you have trouble typing them, that means you'll likely won't use them :P 18:55
hahainternet the only way these will get any attention is when something properly supports their use
and i see no difference between $y1, $y2 and $y₁ and $y₂
Hotkeys They should still be allowed though
ZoffixW m: multi term:<x₀> { 42 }; say x₀
camelia rakudo-moar : OUTPUT«42␤»
zengargoylew hahainternet: what do you mean fonts are ignored? i'm on debian and see sub/super scripts just fine. 18:56
ZoffixW hahainternet, the latter is much harder to type :P
hahainternet zengargoylew: certain alphabetic subscripts don't exist in unicode at all, and many are not included in your 'regular' fonts
Hotkeys if we can't get subscripts on scalars can we use them for arrays ;p
hahainternet so they look weird as they're brought in from other fonts
ZoffixW Even if you set them as compose sequences, it's more time 18:57
hahainternet ZoffixW: so?
it also takes longer to use uppercase letters
Hotkeys eg @x₁
hahainternet should they be restricted in variables?
Hotkeys :D
ZoffixW lol fair enough :P
hahainternet i mean, it really doesn't take long to do shift + altgr + _ + 0
zengargoylew ah, codepoint vs formatting.
Hotkeys that combo uses my whole hand :P 18:58
gtodd .u U+1F98B
yoleaux No characters found
hahainternet for ² and ³ on debian you can cheat and just hold altgr + press 2,3
18:58 uruwi left
hahainternet but they are 'third level keys' and are documented so incredibly poorly you'd be lucky to find a list anywhere 18:58
18:58 uruwi joined, hankache joined
uruwi wonders why there is still no superscript -1 18:58
hahainternet $x¯¹ 18:59
it's two characters
ZoffixW I don't have those
The "hold altgr + press 2,3" I mean
hahainternet ZoffixW: what are you on?
ZoffixW LSB Version:core-2.0-amd64:core-2.0-noarch:core-3.0-amd64:core-3.0-noarch:core-3.1-amd64:core-3.1-noarch:core-3.2-amd64:core-3.2-noarch:core-4.0-amd64:core-4.0-noarch
Distributor ID:Ubuntu
Description:Ubuntu 12.04 LTS
Release:12.04
Codename:precise
uruwi Actually, I'd prefer an option that's one character wide. 19:00
ZoffixW Oopps Sorry. I thought that would be one line
hahainternet ah i don't know anything about ubuntu i'm afraid
Hotkeys who is it that mentioned wincompose
just got it and this is wonderful
uruwi I mentioned it
hahainternet uruwi: then you should go talk to the unicode standard types, because i want $x_c too :)
ZoffixW hahainternet, it's a fork of debian, so I'd imagine stuff would be similar
hahainternet ZoffixW: it was a fork of debian a long time ago
19:00 uruwi is now known as uruwi[A]
hahainternet now it's a testing platform for mark shuttleworth's broken ideas ;) 19:00
ZoffixW hehe :P
hahainternet anyway yeah, holding altgr is the 'third level' selection, so your distro may configure it differently 19:01
FROGGS m: use Test; is ::('Foo'), Failure, '...' # :/
camelia rakudo-moar : OUTPUT«not ok 1 - ...␤␤# Failed test '...'␤# at /tmp/FWeST2YcbN line 1␤# expected: (Failure)␤# got: (Failure)␤»
hahainternet for compose sequences for superscripts it's hold shift + altgr, ^ then the number
ZoffixW hahainternet, oh, I was just trying my Compose key
hahainternet ¹³³⁷
zengargoylew depends on tweaks, my compose is rightwindowskey on one machine and capslock on another. 19:02
hahainternet anyhow TimToady my feeling is, if it's in unicode and it has unambiguous use in maths or computing, there's no reason we shouldn't support it, but i know it's easy to say that without me doing the implementation
timotimo :þ
ZoffixW dammit... Worked once with Compose+shift+6, but now it doesn't :S something's borked
hahainternet so i will try and put some more time into learning nqp
Hotkeys uruwi[A]: subscripts won't work for me :(
hahainternet ZoffixW: shift + altgr first, in that order
ZoffixW timotimo, was the Weekly published for this week?
hahainternet it's a bit wacky
ZoffixW hahainternet, same difference 19:03
hahainternet ZoffixW: weird, wish i could help i'm afraid
gtodd U+1F98B (butterfly) is recommended for Unicode 9
ZoffixW ᦋ
hahainternet i assume … is ... in p6 too
ZoffixW That looks like a ballsack on my screen
hahainternet i don't have that font char :( 19:04
ZoffixW hahainternet, irclog.perlgeek.de/perl6/2015-11-25#i_11598956
hahainternet wait, i'm confused
what's with the link?
El_Che had a great few hours at work transposing perl5 regex into the perl6 dialect 19:05
nice
ZoffixW hahainternet, oh. Usually when I can't see chars on IRC, they show up fine in the browser.
Hotkeys m: say (1…3).perl
camelia rakudo-moar : OUTPUT«(1, 2, 3).Seq␤»
Hotkeys ayy
hahainternet ZoffixW: oh really? that shouldn't be the case!
the same font selection should be carried out
bloody ubuntu :p
ZoffixW hahainternet, not if you use different fonts for IRC/Browser :P 19:06
I don't want fixed-width fonts in the browser :P
hahainternet ZoffixW: it doesn't actually use a single font you know 19:07
when a font doesn't have a particular glyph
it pulls said glyphs from other fonts
ZoffixW I highly doubt that
Hotkeys wincompose isn't doing anything that requires shift key for me
:(
ZoffixW hahainternet, anyway. here's that char: i.imgur.com/JSvLkVp.png
geekosaur depends on the client
hahainternet ZoffixW: en.wikipedia.org/wiki/Fallback_font
wiki.archlinux.org/index.php/Fonts...r_with_X11 19:08
it's a real thing, believe me or not ;)
geekosaur I note that hexchat on most Debian-derivatives does not handle codepoints > U+FFFF
19:10 cdg left, cdg joined
psch welp, there goes my workaround for <unum> on jvm 19:10
Hotkeys oh
psch we don't have the compiler object yet during Grammar or Actions compilation, which i probably should've been aware of... :)
Hotkeys I figured it out
ZoffixW hahainternet, ah OK. You meant a fallback font. I thought you meant it goes over a bunch of my installed fonts, trying to find some chars.
Hotkeys you dont have to hold the compose key 19:11
I'm a goof
ZoffixW You don't?
Whoa!
Hotkeys for wincompose
ZoffixW Hotkeys++
Hotkeys idk about linoox
hahainternet ZoffixW: that is ultimately how it works, see the second link
Hotkeys but if it works
then np
:p
hahainternet you can control the order in which it searches :)
ZoffixW Damn. And now the superscript stuff works ⁶¹²
hahainternet and yeah sorry was i not clear about compose, it's shift+hold, press altgr and release both on Debian 19:12
swank :D
Hotkeys $x₃
hahainternet it's a unicode party!
19:12 cdg left
Hotkeys ¹²³⁴⁵ 19:12
this is so useful
psch i suppose the alternative is throwing NYI when we get dummy values back from unipropcode and getuniprop_str...
Hotkeys Imean not incredibly but
yknow
ZoffixW hahainternet, I've been holding compose while typing "—" for a good 5-6 years :P
Hotkeys haha
hahainternet ZoffixW: D: i'm so sorry
psch 'cause i'm not implementing those on r-j right now vOv 19:13
hahainternet psch: i don't know if i mentioned i saw what you did, and i thank you greatly for it
psch hahainternet: uh, what? :)
hahainternet psch: your module the other day 19:14
when we were discussing unicode super/subscripts etc
psch hahainternet: oh, the superscript slang? that was just poking around towards making it work - it doesn't (or didn't for me) work
hahainternet still deserves a thanks
Hotkeys thanks uruwi[A] now I am a unicode god
ok brb going to lecture
psch hahainternet: FROGGS++ can probably give you pointers towards actually making it work, fwiw
FROGGS umm, what? 19:18
ahh, slangs :o) 19:19
psch gist.github.com/peschwa/7202b9f622de517cd881 this one
psch afk for a bit
flussence I don't see it mentioned above so: all the compose keycombos are in /usr/share/X11/locale/en_US.UTF-8/Compose (seems to be the case here even though that's not my system locale) 19:20
19:21 espadrine left
zengargoylew time to pack up from work and start the awesome 4½ day weekend 19:21
TimToady it is quite certain that, if keyboard manufacturers added an explicit Compose key, they'd put it in the wrong place
ZoffixW flussence++ sweet 19:22
arnsholt jnthn: There's a spectest break with your patch. t/spec/S02-types/is-type.t bails with the error "Cannot modify an immutable BagHash" after test number 8 (t/spec/S02-types/is-type.rakudo.moar:39)
19:22 MadcapJake joined
hankache zengargoylew 4 and a half!! I am jealous 19:22
zengargoylew maybe that will be enough free time to figure out adding =data to POD! :) 19:23
ZoffixW Doesn't seem to be accurate tho. I see "<Multi_key> <backslash> <o> <slash> : "🙌"", but pressing that sequence does nothing
Hotkeys TimToady: even if they just renamed the right alt button that'd be okay
I never use right alt anyway
ZoffixW uses right Windows key for Compose
nine Note to self: if you set ENV vars _after_ running the sub process, no debugging in the world will make it see the variable... 19:24
retupmoca zengargoylew: have you seen Data::Section::Simple ?
Hotkeys has no right windows key
flussence 🙌 wfm
Hotkeys only left
retupmoca has no idea if it actually works, fwiw
Hotkeys and I use left windows key to open things
19:24 cdg joined
Hotkeys hit left windows key, type part of name, hit enter 19:24
zengargoylew just remap it to whatever you want. caps-lock seems to suit me best.
Hotkeys hope windows got what I wanted to open right
MadcapJake What am I doing wrong in this: gist.github.com/MadcapJake/0089315a03c13e30c514
19:25 leont joined
Hotkeys that seems reasonable zengargoylew, the only real use for capslock is yelling 19:25
ZoffixW MadcapJake, age is private?
Hotkeys even then I often hold shift
ZoffixW MadcapJake, oh, I'm blind never mind
19:26 cdg left
MadcapJake ZoffixW: Just trying to play with declaring custom accessors 19:26
zengargoylew sadly you can't really mix XCompose and iBus
retupmoca MadcapJake: $!age is private - you either need to make it public '$.age' or write a custom submethod BUILD
MadcapJake oh you can't do it with `new`?
retupmoca self.bless(...) will only initialize public attributes by default 19:27
zengargoylew iBus 'borrows' some of the standard XCompose mappings but not all, and doesn't read ~/.XCompose
ZoffixW How about that. I was right :P
MadcapJake ohh, what exactly is the difference between `new` and `BUILD`?
arnsholt new has to actually construct the object; for example using low-level primitives like nqp::create 19:28
BUILD only has to initialize an already constructed object
ZoffixW MadcapJake, this might be helpful: docs.perl6.org/language/objects#Submethods
19:29 yqt joined, Sqirrel left
zengargoylew retupmoca: i'm not sure Data::Section::Simple is actually supposed to be possible. still need to ask TimToady about that. 19:30
arnsholt docs.perl6.org/language/objects#Obj...nstruction (linked to from the submethod section) is more relevant
zengargoylew the =finish section it's using is *supposed* to be POD according to the S26 19:31
retupmoca zengargoylew: Ah, ok. I don't actually know anything about it, other than "it exists"
19:32 kjs_ joined
zengargoylew i'm not sure if current handling of =finish is an accident in implementation or a change in spec.... 19:32
MadcapJake so it goes `new` -> `bless` -> `BUILDALL` -> { `Mu.BUILD`, `Any.BUILD`, ... , `self.BUILD` } 19:35
19:35 CIAvash left 19:36 kjs_ left
zengargoylew commute & 19:36
19:36 zengargoylew left
MadcapJake so when would you ever want to override `new`? 19:37
19:37 ZoffixW left
jnthn arnsholt: Hm, are those tests about "is Foo" on state vars? 19:39
19:39 cdg joined
jnthn arnsholt: If so, one was passing bogusly and one was failing and todo'd already, so I'd be fine with skipping them both 19:39
TimToady MadcapJake: if you wanted a positional interface, for one thing
hankache MadcapJake i do it to pass positional arguments instead of named 19:40
MadcapJake ok cool! I get it, which route would you take for doing a "named constructor" paradigm, perhaps turn `new` into a multi method? 19:42
19:47 xfix joined
retupmoca I usually do 'multi method new', but I've also done 'method otherName { self.bless(...) }' as well 19:48
dalek ast/curli: a138e08 | FROGGS++ | S11-repository/curli-install.t:
add tests for installing into a CURI
 19:49
FROGGS nine: ^^
not much, but it is a start 19:50
dalek kudo/curli: 7155a27 | FROGGS++ | t/spectest.data:
run test file S11-repository/curli-install.t
nine FROGGS: a good start :) 19:51
19:53 ZoffixW joined
ZoffixW Is the colon instead of () for method calls only for builtins? 19:54
arnsholt jnthn: Yeah, it's the state %h is BagHash one
ZoffixW $.ssay: $msg; # $.ssay:⏏ $msg; expecting any of: colon pair
Kinda weird we have both "Classes and Objects" and "Object Orientation" in the docs 19:57
19:57 darutoko left, eliasr joined
psch m: class A { method ssay($a) { say $a }; method foo { self.ssay: "foo" } }; A.new.foo 19:58
camelia rakudo-moar : OUTPUT«foo␤»
psch ZoffixW: $. in a class doesn't exactly mean self, only somewhat
(i'm not sure on the exact difference...)
ZoffixW psch++, ok, thanks
20:02 zacts joined
flussence just replaced some ridiculous «$sock.print:("rn" RX~ @lines).join» hack in an old one-off script with a way more readable lexical «use newline :crlf» :D 20:02
mst RX~ ? 20:03
flussence R[X[&infix:<~>]] 20:04
arnsholt mst: R reverses the argument order of an infix
leont That is rather awful…
yoleaux 22 Nov 2015 09:01Z <FROGGS> leont: I try to golf it down to a testcase
flussence looks cute but takes 15 minutes to figure out wtf it does first time...
arnsholt Yeah. Why not "@lines X~ "\r\n"" though?
TimToady why no tjust RX~ ? 20:05
oh, that's what you have
flussence arnsholt: @lines was a bunch of quoted strings in the actual code, it *was* slightly easier to follow by putting the line ending up front instead of having it dangling
TimToady sure, end-weight is important sometimes, and the usual reason for using R 20:06
arnsholt Ah, right. That makes sense, yeah
ZoffixW Someone said P6's S/// is the same as P5's s///r, but I don't see S/// in the docs nor specs and I can't figure out how to actually use it.
m: say 'foo' ~~ S/foo/foos/;
camelia rakudo-moar : OUTPUT«Potential difficulties:␤ Smartmatch with S/// can never succeed because the subsequent string match will fail␤ at /tmp/J7GZIfOmVj:1␤ ------> 3say 'foo' ~~ S/foo/foos/7⏏5;␤False␤»
TimToady R/ often has a big hairy thing to be divided by something simple
ZoffixW s///r is "return the result instead of trying to assign it" 20:07
TimToady ~~ is not "the binding operator" in P6
it's the smartmatch operator
it happens to topicalize as well
flussence that error message could probably use a "Did you mean '='?" then 20:08
wait, that's wrong
dalek ast: b287fb8 | arnsholt++ | S02-types/is-type.t:
Skip bogus is Foo on state variables.
TimToady but you can't use it *only* for topicalization unless your RHS returns something that is a known success/failure value
ZoffixW Then what's the right way to use S///?
TimToady m: say S/foo/foos/ with 'foo'; 20:09
camelia rakudo-moar : OUTPUT«foos␤»
ZoffixW Oh
ZoffixW never heard of with
TimToady m: say S/foo/foos/ given 'foo';
camelia rakudo-moar : OUTPUT«foos␤»
flussence S/// works on $_, for everything else you'd need .subst
ZoffixW m: say S/foo/foos/ for 'foo'
camelia rakudo-moar : OUTPUT«foos␤»
ZoffixW cool
TimToady m: say ("foo" andthen S/foo/foos/) 20:10
camelia rakudo-moar : OUTPUT«foos␤»
flussence m: given 'foobar' { my $baz = S[(.)$0] = 'xx'; dd $_, $baz }
camelia rakudo-moar : OUTPUT«Str $var = "foobar"␤Str $var = "fxxbar"␤»
colomon /me did not realize S/// was a thing. s///, sure. But S?
TimToady m: "foo" ~~ say S/foo/foos/) 20:11
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/TjHOET_HLw␤Unexpected closing bracket␤at /tmp/TjHOET_HLw:1␤------> 3"foo" ~~ say S/foo/foos/7⏏5)␤»
TimToady m: "foo" ~~ say S/foo/foos/
camelia rakudo-moar : OUTPUT«foos␤»
TimToady so there's lots of ways to do it, just not the one you want 20:12
but I figgered it would be better to fail in that case than to try to dwim
ZoffixW I went with: $!debug and say ".privmsg({$msg.subst("\n", "␤", :g)})";
dalek kudo/nom: 6313051 | arnsholt++ | src/Perl6/World.nqp:
Be more specific about when to create container types.

  jnthn++ for the patch.		 20:13
Skarsnik .subst is easier to use x)
ZoffixW It reminds me of JavaScript... and I'm not a huge fan of JavaScript :)
flussence m: use newline "␤".pred; say 'hi!'; 20:14
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/m9RZGS6Hi3␤Error while importing from 'newline':␤no EXPORT sub, but you provided positional argument in the 'use' statement␤at /tmp/m9RZGS6Hi3:2␤------> 3".pred7⏏5; say 'hi!';␤»
flussence m: OUR::<$?NL> := "␤".pred; say 'hi!'; 20:15
camelia rakudo-moar : OUTPUT«hi!␤»
timotimo ZoffixW: it wasn't publihed for this week yet :(
flussence m: BEGIN OUR::<$?NL> := "␤".pred; say 'hi!';
camelia rakudo-moar : OUTPUT«hi!␤»
ZoffixW timotimo, ok :)
flussence m: BEGIN OUR::<$?NL> := "␤".pred; say $?NL.perl; 20:16
camelia rakudo-moar : OUTPUT«"\n"␤»
ZoffixW timotimo, wait until next week and do a double-feature :)
flussence m: BEGIN OUR::<$?NL> := "\x[2423]"; say $?NL.perl;
camelia rakudo-moar : OUTPUT«"\n"␤»
20:16 rurban left
dalek kudo/curli: 5344dfd | (Stefan Seifert)++ | src/core/CompUnit (3 files):
Fix circular module loading detection and error messaging
 20:17
nine ZoffixW: ^^^ 20:18
ZoffixW weee
nine++
nine, but does that mean it's impossible to have two modules `use` each other?
nine ZoffixW: yes and that's a good thing ;)
ZoffixW: the solution to your problem really is predeclaring the class. 20:19
ZoffixW Hm. But that means I have a real usecase for which P6—with all its might—doesn't exist.
nine, but I've tried that and I'm getting errors
s/doesn't exist/can't solve/;
nine, irclog.perlgeek.de/perl6/2015-11-25#i_11598614 20:20
timotimo ZoffixW: i'm afraid i'll end up doing it double-feature style next week, yeah :|
dalek kudo/nom: 6ca14ad | peschwa++ | src/Perl6/ (2 files):
Restrict <unum> matching to moar.

NQP-J doesn't have (proper) getuniprop_str or unipropcode yet, so we work around that.		 20:22
psch uhm 20:23
no one build HEAD on jvm right now... :P
20:23 MadcapJake left 20:24 lichtkind joined
dalek p: efd48d6 | peschwa++ | src/vm/jvm/ (2 files):
Implement stubs for getuniprop_str and unicodeprop.
 20:25
kudo/nom: d4b93ab | peschwa++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for nqp::unipropcode and nqp::getuniprop_str stubs.
psch there, all better \o/ 20:26
20:27 TEttinger joined 20:28 FROGGS left
nine ZoffixW: your issue must be solvable. I'm sure of that. I'm also too tired right now to find the solution :/ 20:29
ZoffixW Alright. I'll mention it again some time in the future :)
jnthn arnsholt: Feel free to fudge those and push the patch. :)
psch ZoffixW: the obvious fix would moving Plugin.pm inline into Client 20:30
ZoffixW: that's a bit eh, though, for obvious reasons
arnsholt jnthn: Already done! =D
ZoffixW right
psch ZoffixW: but stubbing the *role* IRC::Client::Plugin outside of IRC::Client works as expected :P 20:31
jnthn arnsholt: Indeed, now I caught up with the backlog :) 20:32
ZoffixW: If two things are so tightly coupled that they want a typed circular dependency, they probably deserve to be in the same compilation unit (e.g. module)
20:33 kid51 joined
psch ZoffixW: as in s/class/role/ in lib/IRC/Client.pm6:2 20:33
ZoffixW psch, oohhh.. Damnit. right :D
jnthn Oh, or was pre-decl just being done wrong and they were in the same... :) 20:34
ZoffixW jnthn, nah, all I needed is a way to say has My::Role $bar; in one of my classes and it can be done by stubbing. The reason I was getting errors is because I was stubbing it with class My::Role { ... } instead of role My::Role { ... }
psch++
jnthn Aye, I thought you were trying to have them over two files :)
ZoffixW jnthn, yeah, they *are* in two different files :) 20:35
20:35 geraud joined
ZoffixW And stubbing lets you avoid `use`ing the other file 20:35
Or at least it works right now :P (it gets `use`d by other stuff). 20:36
Skarsnik do you have a gist of how your irc client work? I nearly want to rewrite one for me x) 20:38
ZoffixW Skarsnik, examples/bot.pl6 in this repo: github.com/zoffixznet/perl6-IRC-Client/ 20:42
Skarsnik, for some values of "work" though. I merely managed to get it join a channel and spam with an Interval every 5 seconds :) 20:43
Skarsnik "Nicks can only be 9-chars long max." that untrue for most irc server x) 20:44
ZoffixW Skarsnik, yeah, I'll remove that restriction soon.
hahainternet if you're writing an irc client
do not hardcode line length
while most ircds are limited to 512 or whatever
it's now common to change that
bloody irssi 20:45
ZoffixW noted ;)
pink_mist hahainternet: is there any way the server can communicate line length to you upon connection though?
hahainternet pink_mist: it does so as part of its capabilities doesn't it?
lucasb m: say Mu === Mu 20:46
camelia rakudo-moar : OUTPUT«Cannot call infix:<===>(Mu, Mu); none of these signatures match:␤ ($?)␤ ($a, $b)␤ (Int:D \a, Int:D \b)␤ (int $a, int $b)␤ (Num:D \a, Num:D \b)␤ (Num $ where { ... }, Num $ where { ... })␤ (num $a, num $b --> Bool)␤ (Str:D …»
pink_mist I've never seen that bit in the caps
lucasb should '===' be defined to work with Mu too?
ZoffixW pink_mist, Freenode sends NICKLEN=16
pink_mist ZoffixW: no, line length, not nick lenght
ZoffixW err
pink_mist *length
jnthn lucasb: Generally operators are not.
lucasb: If they were, junction auto-threading wouldn't work. 20:47
geekosaur don't see a line length in freenode's stuff, but freenode uses the rfc default
hahainternet pink_mist: just hopping on perl irc, no line length there i can see
lucasb jnthn: ok, thanks
pink_mist hahainternet: yeah, I've never heard of any ircd that didn't have 512 as limit
ZoffixW pink_mist, don't see anything: fpaste.scsys.co.uk/501930 20:48
lucasb the proto is: proto sub infix:<===>(Mu, Mu)
hahainternet pink_mist: there are many, i've had people whining at me to modify so they can do improved ascii spam for example
lucasb but the actual multi doesn't use Mu, so I think it defaults to Any
hahainternet so, worth making it configurable
20:48 zacts left
ZoffixW Hm. There's "CHANNELLEN=50 NICKLEN=9 TOPICLEN=490 AWAYLEN=127 KICKLEN=400" I'm sure there's a *LEN for line lengths 20:48
pink_mist hahainternet: well without knowing what the capability is called, it's hard to know how to configure it :P 20:49
hahainternet pink_mist: consider sending test lines
that's probably the best way to tell
pink_mist that's ... not something I'd want a client doing every time it connects
hahainternet why not? that's a standard part of any connection 20:50
negotiating capabilities
it's literally part of IP
ZoffixW I get ERROR :Request too long
pink_mist of IP, sure, but not IRC ... the server may not even allow you enough of a buffer to do it
hahainternet if the server doesn't give you enough of a buffer to send a 1k long line 20:51
then that ircd is broken
i think 512 is a good default
but all i'm saying is, don't make it hardcoded
as i've had to deal with the result of that and it's bloody annoying :D
ZoffixW Don't see anything: www.irc.org/tech_docs/005.html 20:53
And there's like a million *LEN
Including the maximum *targets* of a PRVMSG -_-
CRAP! 20:54
Missed my bus -_
Lol. Now I'm gonna have to sit here for another 15 minutes :)
20:56 Ribbit joined
masak ZoffixW: time enought to write at least one Perl 6 script ;) 20:57
enough*
ZoffixW :P
20:58 raiph joined
dalek kudo/resupply: d781934 | jnthn++ | src/core/Supply.pm:
Tap.close should be something True.
 21:01
kudo/resupply: fd732ce | jnthn++ | src/core/Supply.pm:
Rename Publisher to Supplier.
Skarsnik ZoffixW, how I will write a irc bot stuff gist.github.com/Skarsnik/734f88458584af90ba54 if you want some idea xD 21:02
ZoffixW Skarsnik, how un-Perl-6-ish to use on_disabled instead of on-disabled :P 21:03
OMG XML! 21:04
Skarsnik, I'd definitely choose JSON :P
Skarsnik I don't know xml/yml syntax
xml is dumb to write x)
*Json 21:05
ZoffixW What?
What does "*Json" mean? You don't know JSON syntax? JSON is dumb to write?
Skarsnik let's do it again. I am not familiar with JSON syntax
and XML is dumb/easy to write 21:06
ZoffixW Skarsnik, I'll teach you: [ "a", "list", "of", "things" ], { key: "value", "or anther key": "value", "even": { "nested": ["stuff", "is", "fine"] } } 21:07
Skarsnik, end of lesson :)
hahainternet not quite, json has some slight weirdnesses
ZoffixW Skarsnik, and you have to learn it anyway. P6 distros META files are in JSON :)
Skarsnik I copy paste everytime!
21:07 skids left 21:08 kjs_ joined
ZoffixW Skarsnik, JSON also has true/false/null that are written literally, without quotes 21:08
hahainternet also has no date format
also has no way to order hash keys
ZoffixW leaves to catch the bus this time
hahainternet also has a licensing restriction that you must do nothing 'bad' or 'evil'
can't remember which
Skarsnik I have nothing against json lol, it was just easier to write xml quicly here xD
21:08 ZoffixW left
hahainternet i still use json regardless as it's the best of a bad bunch imho 21:08
yaml is probably better structured
but json is missing a bunch of things that should have been better thought out 21:09
but javascript lol
timotimo our json parser is already slow, better not make the format more complicated :(
Skarsnik I had a weird perl5 IRC 'sdk' it's like 50 lines, all it does is connecting to irc, load/unload pm file, you can write the whole thing without stopping your bot once x) 21:11
related: I should port my Acme::IRC::Art on perl6 21:12
21:12 [Sno] left
Skarsnik but irc color is a nightmare to debug with all irc client not acting the same 21:13
21:14 diakopter___ joined 21:21 _28_ria left 21:24 _28_ria joined 21:33 hankache left
dalek kudo/resupply: 0a8ab76 | jnthn++ | src/core/Supply.pm:
API cleanup; avoid some is-* and other not.
 21:33
kudo/resupply: 0c4e50f | jnthn++ | src/core/ (5 files):
Update code to use new Supplier model.
gfldex S32 says about .getc "Returns a Str type object if no more lines to be read." 21:35
rakudo does: nqp::chars($c) ?? $c !! Nil
who is right?
Juerd jnthn: Is there a chance that the new Supply stuff makes my mqtt module less slow?
jnthn Juerd: Maybe, depends where the bottleneck is. 21:36
timotimo the bottleneck is somewhere different 21:37
psch gfldex: getc return value on EOF isn't tested in roast, from what i can see
timotimo the mqtt thing only uses 5% cpu
somewhere it's sleeping
Juerd jnthn: I wish I knew what its bottleneck was. Given a burst of TCP packets, they trickle slowly in the supply... :|
jnthn Ah, then likely not 21:38
But supply throughput has got better, so far as I can tell
Something that ran for nearly 7s before now runs in 0.7s now.
timotimo whoa!
how the hell?!
jnthn Design. :P
Juerd (For perspective: the packets are tiny, and it's only a few dozen in a second)
21:39 rindolf left
Juerd jnthn: Impressive! What causes this speedup? 21:39
timotimo design!
say it with me, y'all!
21:39 bjz left
jnthn Juerd: Probably avoiding a load of locking 21:40
By adopting a partially-lock-free thing on the common path 21:41
Juerd That'll be great for almost everything. Great job :)
jnthn Throughput of supply pipelines in general should get a bunch better when you start chaining operations on. 21:42
(Compared to how it was before)
Juerd I notice in the mqtt thing that sometimes, but only sometimes, it'll output 10 to 20 lines at once, whereas normally it does about two per second.
So it can be done :) 21:43
Is there a way to see what it's blocking on, in all the threads?
jnthn Not easily; our debugging tooling for threaded stuff is nearly non-existent so far. :( 21:44
timotimo jnthn: where do i look to figure out how to make my Int @foo[2;2] = (1, 2; 3, 4); work?
jnthn Array.pm
timotimo just looked through Rakudo::Internals
jnthn It'll likely be a fail in there somewhere
I can't quite guess what. 21:45
Juerd jnthn: That's okay, I'll just try this project again in a year or so :)
Built what I needed in Perl 5. It's good to have that as a fallback :D
timotimo 'k
jnthn Damn, I'm a factor of two off being able to say that if you turn off the safety valve you can shove a million messages per second through a supply. 21:46
TimToady get a faster computer
or better yet, just do some more design! 21:47
jnthn :)
Juerd jnthn: use less <safety>? :) 21:48
use more <design>;
timotimo jnthn: it hits the STORE(Array:D: Iterable:D \iterable) candidate in Array rather than the STORE(::?CLASS:D: Iterable:D \in) candidate in ShapedArray
Juerd Together, it'll beat quantum computers.
jnthn If we really want to nail supply performance somebody should perhaps design us a trace JIT...
timotimo: oops. :)
timotimo this couldn't possibly have something to do with what someone else found today about the constructor being called twice with different things? 21:49
jnthn timotimo: That's the problem then. Can you stick a proto in to "force" it to disregard the other one?
No
timotimo the proto is right in front of that candidate :)
jnthn ohhh
Does the method new inside of TypedArray have a clue about shapes? 21:50
If not, then I suspect *that* is the fail
timotimo let's see. possibly not :)
yeah, doesn't accept a shape parameter
i'll cargocult from Array.new, probably
21:50 kaare_ left
jnthn yeah 21:51
21:51 muraiki left
timotimo should i also port the "\values or **@values via new-internal" thing from Array to TypedArray as well? 21:52
jnthn timotimo: Yeah, sounds reasonable 21:54
21:55 xfix left
dalek ast/resupply: 46231a7 | jnthn++ | / (22 files):
First round of updates for supply changes.
 21:55
jnthn 10 tests still need attention, but I think the number of root causes is smaller
21:56 bjz joined
dalek kudo/resupply: 8f9a7a9 | jnthn++ | src/core/Supply.pm:
Supplies treat values through them as items.

So use .push, not .append.		 21:56
21:56 Zoffix joined 21:57 pdcawley left, _28_ria left
leont Does that mean Supply gets split in an emitting and a receiving end? 22:04
timotimo a little bit, yeah
jnthn leont: Somewhat, yes. Supply.new now becomes Supplier.new, and the object you get back has a .Supply method. You emit/done/quit on the Supplier, and (as before) tap the Supply. 22:05
leont I think I could like that, yes 22:06
jnthn leont: This will only make a notable difference to code that used supplies in this way. Supply.interval, Supply.from-list, etc. are all just the same.
As is working with async sockets, processes, etc.
leont C++ has a similar split on its promise implementation (emitting end being called promise, receiving is future) 22:07
jnthn There's also stronger enforcement of the supply protocol (emit* [done|quit])
timotimo uh-oh ... bad news
Stage optimize : Cannot call ACCEPTS; no signatures match / at gen/moar/m-BOOTSTRAP.nqp:2353 (blib/Perl6/BOOTSTRAP.moarvm::741) / from gen/moar/m-CORE.setting:883 (<ephemeral file>:ACCEPTS:36) / ... / from gen/moar/m-BOOTSTRAP.nqp:472 (blib/Perl6/BOOTSTRAP.moarvm:bind_one_param:811) 22:08
ever seen something like this?
jnthn o.O
No
timotimo inside INITIALIZE_DYNAMIC, oddly 22:09
diakopter___ timotimo: maybe you're still using my moar patch ;)
timotimo hah, i don't think so
ah, this is all under "blib/Perl6/Optimizer.moarvm:report_inevitable_dispatch_failure:161"
22:09 spider-mario left
timotimo oh my 22:10
seems like that's what you get when you have two multi methods with the exact same signature
inside a role inside a class
OSLT
jnthn yawns and figures it's time to rest 22:11
Juerd Good night!
timotimo good night jnthn!
d^_^b Zoffix: I saw a MR on learnxinyminutes. You know if anyone is planning to add concurrency examples to the markdown?
jnthn See y'all tomorrow 22:12
o/
timotimo timo@schmetterling ~/p/rakudo (nom)> perl6 -e 'my Int @foo[2;2] = (1, 2; 3, 4); say @foo'
[[1 2] [3 4]]
Zoffix d^_^b, none that I know of. Ven++ was planning to do some work on my PR, but I doubt they'd mind if you contribute :P
d^_^b Zoffix: ok, there are a few bits I don't get how to use/or what they are for. 22:13
Zoffix: this one? github.com/adambard/learnxinyminut...ssues/1390 or 1434
22:14 bjz left
Zoffix The open one, 1390 22:14
d^_^b ok cool
Zoffix timotimo, neat 22:15
timotimo running spec tests on the change now
hopefully it doesn't break anything else
22:16 lucasb left
psch ...i'd really like camelia to build a new r-j :/ 22:17
as in, nom/HEAD definitely compiles atm
timotimo ooooh 22:20
t/spec/S02-types/array-shapes.rakudo.moar - TODO passed: 15 22:22
Zoffix \o/
timotimo um 22:24
but the test code then tries to pop from and push to a shaped array
it's just a one-dimensional array, but shaped arrays immediately die when you try to pop or push
so maybe my patch is wrong?
psch m: my @a[2]; @a.push: 0; 22:25
camelia rakudo-moar : OUTPUT«Cannot push a fixed-dimension array␤ in block <unit> at /tmp/gNKfLdNtf6:1␤␤»
timotimo maybe after popping from a shaped array it's supposed to be possible to push again? 22:29
the code so far has the pop TODO'd (it's supposed to throw an exception, but used to not do that), then the pop after it used to work
but that's only because typed arrays with shapes didn't actually get the shape set on them at all
22:35 sno joined 22:36 vendethiel left 22:38 |Tux| left 22:40 diakopter___ left
ab5tract timotimo: that seems like a reasonable point, but at the same time the properties of arrays in perl 5 that allowed the queue/stack semantics 22:41
was the variable length
(plural / singular mismatch.. variable (dynamic! egads!) length of arrays was the property that enabled push/pop) 22:42
22:42 bbkr left 22:43 bbkr joined 22:48 bbkr_ joined
timotimo mhm 22:48
so ... what are you saying? %)
22:49 pdcawley joined
Hotkeys apparently vulgar fractions are called that because it's the old use of vulgar meaning common 22:51
it's just fractions that aren't bigger on the top than the bottom
22:52 bbkr left 22:58 FROGGS joined
Mouq doesn't think it really makes sense to push/pop from a shaped array… 23:03
TimToady Mouq! Happy T-weekend! :) 23:05
ab5tract timotimo: I haven't read enough of the backlog to know about your commit 23:07
Mouq Hey! And to you as well ^^ 23:08
ab5tract but it almost sounds to me like the test is wrong
23:08 FROGGS left
timotimo the commit isn't up yet 23:09
ab5tract locally push/pop are not enabled for shaped arrays 23:10
Mouq I'm on break right now… I'm interested in looking at what I can do to speed up the peg engine, but the code is, heh, a little dense
ab5tract is your commit about adding it?
i mean, i am totally down with variable length shaped arrays :D 23:11
we can make a trait for when you want to keep them 100% static length
23:12 FROGGS joined
timotimo no, my commit is about adding shapes to non-natively typed arrays 23:13
ab5tract ahh. cool :) 23:14
timotimo m: my Int @foo[2;2] = (1, 2; 3, 4); # this doesn't work because nothing actually sets the shape on that 23:18
camelia rakudo-moar : OUTPUT«Type check failed in assignment to ; expected Int but got List␤ in block <unit> at /tmp/SMwkbEYSfI:1␤␤»
timotimo m: my Int @foo[2;2]; say @foo.shape
camelia rakudo-moar : OUTPUT«(*)␤»
timotimo m: my int @foo[2;2]; say @foo.shape
camelia rakudo-moar : OUTPUT«(2 2)␤»
timotimo m: my @foo[2;2]; say @foo.shape
camelia rakudo-moar : OUTPUT«(2 2)␤»
Mouq m: my int @foo[2;2] = (1, 2; 3, 4); 23:19
camelia ( no output )
psch ...what's a T-weekend? 23:21
23:21 FROGGS left
TimToady the weekend that starts with T-day 23:22
psch ohh
RabidGravy they all go out and do vast amounts of crystal meth?
23:25 kjs_ left
dalek kudo/nom: 2e8830b | TimToady++ | src/ (3 files):
implement postfix:<ⁿ>

Works for both positive and negative integer powers.		 23:26
TimToady but only for literals, no variables, despite the name :) 23:27
uruwi[A] What the hell?
So, for example, 3² = 9?
TimToady yeah
uruwi[A] > no variables: I wish
23:28 uruwi[A] is now known as uruwi
TimToady any variable would have to be named i or n, unless we cheated somehow 23:28
uruwi m: my int @foo[2;2] = (1, 2; 3, 4); say @foo.shape;
camelia rakudo-moar : OUTPUT«(2 2)␤»
uruwi Modifier characters, or are they counted as part of the grapheme before?
m: my int @foo[2;3;4]; say @foo.shape; 23:29
camelia rakudo-moar : OUTPUT«(2 3 4)␤»
uruwi m: my int @foo[2;3;4]; say @foo;
camelia rakudo-moar : OUTPUT«[[[0 0 0 0] [0 0 0 0] [0 0 0 0]] [[0 0 0 0] [0 0 0 0] [0 0 0 0]]]␤»
uruwi What's the difference between int and Int? 23:30
TimToady native int vs bigint
m: say 2³ 23:31
camelia rakudo-moar : OUTPUT«8␤»
TimToady m: say 2³²
camelia rakudo-moar : OUTPUT«4294967296␤»
TimToady m: say 2¯¹
camelia rakudo-moar : OUTPUT«0.5␤»
TimToady this one shouldn't break JVM 23:32
23:32 FROGGS joined 23:35 Oatmeal left
AlexDaniel TimToady: wow! Another awesome thing today! 23:36
m: my $var = 5; say $var²; # just checking the error message… 23:37
camelia rakudo-moar : OUTPUT«25␤»
TimToady m: my $baʳ = 42; say $baʳ
camelia rakudo-moar : OUTPUT«42␤»
AlexDaniel ah 23:38
TimToady uruwi: yes, that's the problem with the modifier letters
but I figure ** is good enough for when there are variables
so we'll restrict this to convenient literals
uruwi Do you see anyone using this? 23:39
AlexDaniel I think that what we have now already satisfies most of the use cases
TimToady sure, I'll probably use it :)
or are you asking about variables?
uruwi No 23:40
Not about variables
TimToady m: say [+] (3 1 4 1 5 9 7)»²
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/NJmjErxwp9␤Two terms in a row␤at /tmp/NJmjErxwp9:1␤------> 3say [+] (37⏏5 1 4 1 5 9 7)»²␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ statement modif…»
TimToady oops 23:41
m: say [+] <3 1 4 1 5 9 7>»²
camelia rakudo-moar : OUTPUT«182␤»
TimToady sum of squares
23:43 rigsby left, FROGGS left
uruwi I would personally write it as [+] ((3, 1, 4, 1, 5, 9, 7) »**» 2), but hey, TMTOWTDI 23:44
,: [+] ((3, 1, 4, 1, 5, 9, 7) »**» 2)
m: [+] ((3, 1, 4, 1, 5, 9, 7) »**» 2)
camelia ( no output )
uruwi m: say [+] ((3, 1, 4, 1, 5, 9, 7) »**» 2)
camelia rakudo-moar : OUTPUT«182␤»
TimToady outer parens not necessary, btw
AlexDaniel m: say ½ # :)
camelia rakudo-moar : OUTPUT«0.5␤»
TimToady now if we only had vulgar superscript fractions :) 23:45
23:45 rigsby joined
AlexDaniel m: say ½² 23:45
camelia rakudo-moar : OUTPUT«0.25␤»
uruwi m: say ⅑*⅛*⅐*⅙*⅕*¼*⅓*½
camelia rakudo-moar : OUTPUT«0.0000028␤»
AlexDaniel I love it 23:46
TimToady m: say (⅑*⅛*⅐*⅙*⅕*¼*⅓*½).perl
camelia rakudo-moar : OUTPUT«<1/362880>␤»
23:46 Skarsnik left, skids joined
TimToady now we just need to teach .perl to use vulgar fractions where appropriate ;) 23:47
uruwi Not all terminals might display them well
TimToady hence the winky 23:48
uruwi For instance, if they display them only as wide as Latin letters, they could be harder to read
23:48 Oatmeal joined
TimToady needs a nap; anyone should feel free to write tests for the superscript exponents 23:49
zzz &
uruwi Happy napping
TimToady or rather, 💤 23:50
AlexDaniel uruwi: well, indeed: files.progarm.org/2015-11-26-01502..._scrot.png
Hotkeys I wonder if Task::Star plays well with windows 23:54
before Ifinished typing it got angry
RabidGravy if I can't work out how to type it with a maximum of three key strokes it's unlikely I'll remember the codepoint as a number 23:55
uruwi compose, numerator, denominator 23:56
Hotkeys it only works for a limited number of fractions I think 23:57
like all of the irreducable fractions up to 7/8
uruwi 6/7 doesn't have its own
RabidGravy oh yes, xchat just being shit
uruwi Only ⅐ for a denominator of 7 23:58
Hotkeys do any /7
I can't even see that on hexchat
hexchat is terrible at unicode for me on windows
uruwi Time to get another client.
AlexDaniel ⅞ ⅒ 23:59
Hotkeys might be the font
what's a good unicode-happy mono font
uruwi VL Gothic
AlexDaniel m: say ↉
camelia rakudo-moar : OUTPUT«0␤»
uruwi shows up for me fine
AlexDaniel wow

Powered by Cro and the Raku® Programming Language.

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

Thank you!