»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:02 dudley5000 joined 00:03 dudley5000 left 00:11 risou is now known as risou_awy 00:18 donri left 00:19 donri joined
soh_cah_toa perl6: #={ I'm a declarator block for $foo } my $foo; say $foo.WHY 00:35
p6eval niecza v9-9-gd637b5c: OUTPUT«===SORRY!===␤␤Action method comment:sym<#=(...)> not yet implemented at /tmp/ImrmJki0I8 line 1:␤------> #={ I'm a declarator block for $foo }⏏ my $foo; say $foo.WHY␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/ni…
..pugs, rakudo 3654b0: ( no output )
soh_cah_toa nom: #={ I'm a declarator block for $foo } my $foo; say $foo.WHY
aw man, i like declarator blocks :\ 00:36
00:38 whiteknight joined
sorear soh_cah_toa: spelled "rakudo:" now 00:38
soh_cah_toa rakudo: #={ I'm a declarator block for $foo } my $foo; say $foo.WHY
p6eval rakudo 3654b0: ( no output )
soh_cah_toa :( 00:39
00:49 plobsing joined
sorear soh_cah_toa: #= is a line comment now. 00:54
also I think it's spelled #|. 00:55
00:55 abercrombie joined
soh_cah_toa well, that's just what's implemented atm right? the spec mentions multi-line ones as well 00:56
sorear soh_cah_toa: the spec is lagging the implementation here. 00:58
talk to tadzik, ey knows the state of the arg
TimToady a block form might be allowed, but you'd need to use #|{} to put it in front 00:59
soh_cah_toa ok
01:12 REPLeffect left, woosley joined 01:17 drbean left 01:26 REPLeffect joined, plobsing left 01:28 jevin left 01:29 jevin joined 01:41 daniel-s joined 02:02 sdadawdaw joined 02:04 sdadawdaw is now known as BosSkurt, BosSkurt left, BosSkurt joined, BosSkurt left, BosSkurt joined 02:05 BosSkurt left 02:13 uvtc joined
uvtc rakudo: my @a = <y z x b c a y z x>; @a.=sort.uniq; say @a 02:14
p6eval rakudo 3654b0: OUTPUT«Method 'uniq' not found for invocant of class 'Array'␤ in <anon> at /tmp/lRGf8vsKdz:1␤ in <anon> at /tmp/lRGf8vsKdz:1␤␤»
uvtc rakudo: my @a = <y z x b c a y z x>; @a .= sort.uniq; say @a 02:15
p6eval rakudo 3654b0: OUTPUT«===SORRY!===␤Confused at line 1, near "@a .= sort"␤»
uvtc rakudo: my @a = <y z x b c a y z x>; @a = @a.sort.uniq; say @a
p6eval rakudo 3654b0: OUTPUT«Method 'uniq' not found for invocant of class 'Parcel'␤ in <anon> at /tmp/lb3R4E565O:1␤ in <anon> at /tmp/lb3R4E565O:1␤␤»
uvtc Hm. On my own R* 2011.07, `@a.=sort.uniq` works, but `@a .= sort.uniq` (with spaces around ".=") fails. 02:16
nom: my @a = <y z x b c a y z x>; @a = @a.sort.uniq; say "@a"; 02:17
TimToady neither of those is the same as the = one, since you don't have .= on the uniq 02:18
sorear uvtc: rakudo: is not R*
uvtc sorear: right, but I've only got R* 2011.07 installed atm.
TimToady but yes, it should be a little more dwimy 02:19
uvtc TimToady: is `@a.=@a.foo` the same as `@a = @a.foo`?
TimToady: oops. I meant `@a.=foo`.
rakudo: my @a = <y z x b c a y z x>; @a = @a.sort.uniq; say "@a"; 02:20
p6eval rakudo 3654b0: OUTPUT«Method 'uniq' not found for invocant of class 'Parcel'␤ in <anon> at /tmp/mCbkmaKpjj:1␤ in <anon> at /tmp/mCbkmaKpjj:1␤␤»
soh_cah_toa uvtc: yes, the .= invocation is mutating while . is non-mutating 02:21
uvtc Then my next question then is (and what I was driving at in the first place): should `@a.=foo` be the same as `@a .= foo`? (spaces around ".=") 02:22
^ s/then//
TimToady the latter is sugar for the former 02:23
it does not really make sense for the precedence of .= to vary depending on whether there are spaces 02:25
uvtc `@a .= foo` is sugar for `@a.=foo`?
TimToady as currently defined
uvtc Ok. Thank you. 02:26
TimToady the only way to write what you wanted earlier was .=sort.=uniq
note the = on the uniq too
uvtc rakudo: my @a = <y z x b c a>; @a.=sort; say "@a"; 02:27
p6eval rakudo 3654b0: OUTPUT«@a␤»
uvtc rakudo: my @a = <y z x b c a>; @a.=sort; say "@a{}";
p6eval rakudo 3654b0: OUTPUT«a b c x y z␤»
uvtc rakudo: my @a = <y z x b c a>; @a .= sort; say "@a{}";
p6eval rakudo 3654b0: OUTPUT«a b c x y z␤»
uvtc Good.
02:29 whiteknight left
uvtc TimToady: I don't understand why you say that " .= " is sugar for ".="... Isn't "$foo op= $bar" always "$foo = $foo op $bar" regardless of spaces? 02:30
TimToady .= exists by analogy to that, but you can't actually say $foo . $bar 02:31
std: $foo . $bar
p6eval std e3c970e: OUTPUT«===SORRY!===␤Variable $foo is not predeclared at /tmp/_XuDco3T0G line 1:␤------> $foo⏏ . $bar␤Unsupported use of . to concatenate strings; in Perl 6 please use ~ at /tmp/_XuDco3T0G line 1:␤------> $foo . ⏏$bar␤P…
uvtc Oh, of course. Thanks. :) 02:32
Just a little extra pixie dust to save you from having to write the array name twice.
TimToady it's mostly there for .= new() 02:33
uvtc Ok -- to save you from writing the classname twice (ex. `my Foobar $f = Foobar.new()` vs. `my Foobar $f .= new()`). 02:35
02:37 uasi joined 02:48 uvtc left 02:52 agentzh joined 02:59 drbean joined 03:07 donri left 03:21 beppu left 03:26 Su-Shee left 03:28 Su-Shee joined 03:37 jevin left, jevin joined 03:38 daniel-s_ joined, daniel-s left 04:04 birdwindupbird joined 04:12 kaare_ joined 04:16 colomon_ joined, colomon left, colomon_ is now known as colomon 04:18 puppyuser joined, birdwindupbird left, birdwindupbird joined 04:24 jamtech joined 04:25 dark_x joined 04:32 odoacre joined 04:36 abercrombie left 04:39 soh_cah_toa left 04:40 plobsing joined 04:41 y3llow left, pothos left, y3llow_ joined, pothos_ joined 04:42 pothos_ is now known as pothos 04:43 puppyuser left, y3llow_ is now known as y3llow 05:10 sftp left 05:18 jamtech left 05:20 dark_x left 05:22 molaf joined 05:41 koban` joined, koban` left
dalek ecza: 678102e | sorear++ | lib/ (3 files):
Draft of P6->CLR coercion routines
 05:49
05:53 natureboy left 05:58 jevin left 05:59 jevin joined 06:01 satyavvd joined, MayDaniel joined 06:02 molaf left 06:03 plobsing left 06:04 REPLeffect left 06:06 plobsing joined, MayDaniel left
sorear out 06:14
06:17 c1sung joined 06:19 REPLeffect joined 06:38 wtw joined 06:45 c1sung left 06:56 wamba joined 07:01 frhodes joined
TiMBuS self.bless(*, |%args)!initialize() 07:16
ooOoo
thats nifty, whoever did that
on the other hand, it makes me feel like we left something out of the perl6 OO model, if we need to resort to that to init a class 07:18
PerlJam TiMBuS: but you don't. 07:20
TiMBuS well yeah, submethod BUILD can do the initialization 07:21
but then why did someone change my code to specifically not use it? does nom not have BUILD
PerlJam pretty sure you're talking about a bug right now. 07:25
Where did you see the line of code you mentioned above? 07:26
07:39 wamba left
jnthn morning, #perl6 07:48
moritz morning everybody
07:48 Siddy left
tadzik good morning 07:51
07:52 Siddy joined 07:56 wamba joined 08:00 daxim joined
mberends good morning all 08:00
08:13 thou left 08:19 wamba left 08:20 Siddy is now known as Trashlord 08:21 wamba joined 08:22 colomon left
TiMBuS sorry PerlJam, had a phone call 08:22
I'm talking about IO::Socket::INET in nom 08:23
tadzik I'm pretty sure nom has BUILD, maybe it didn't when someone was moving the code to nom 08:30
08:32 wamba left 08:39 sentinel_ joined 08:43 wamba joined 08:47 lumi_ left 08:49 agentzh left 09:01 drbean left 09:11 orafu left, orafu joined 09:13 domidumont left 09:23 domidumont joined 09:24 SHODAN joined 09:30 uniejo left 09:43 woosley left 09:45 sentinel_ left 09:56 wamba left 10:35 whiteknight joined 10:46 envi_laptop joined 10:54 hatseflats left 10:55 hatseflats joined 10:57 hatseflats left
felher Hm, so '$x .= foo.bar' is the same as '($x = $x.foo).bar' while '$x -= 3 - 3' is '$x = $x - (3-3)' and not '($x = $x - 3) -3'. Why's that? At a first glance '$x .= foo.bar' isn't very DWIM to me. I really thought it be the same as '$x = $x.foo.bar'. Am i forgetting something and its actually consistent? 10:59
flussence if it were foo.bar, then $x.=foo + 3 probably wouldn't DWYM either 11:09
jnthn iirc, .= forces the parser to immediately parse a methodop after it 11:13
Which only swallows the foo
11:15 daniel-s_ is now known as daniel-s 11:16 colomon joined
flussence ooh, I just had a (probably terrible) idea, a HTML tag object where .Str is well-formed XML and .gist is well-formed HTML 11:17
tadzik tee hee
11:23 lumi_ joined 11:25 PZt joined 11:38 JimmyZ joined
moritz some of you might remember perlgeek.de/blog-en/perl-6/how-fast....writeback 11:48
11:48 ruoso left
moritz I just timed it again, this time it was roughly 2 minutes 11:49
11:49 ruoso joined
flussence whoa 11:49
tadzik nicey 11:50
mls hello perl6!
JimmyZ hopes that will be 2 seconds :) 11:51
mls whoa, jnthn++ has been really busy this weekend! jnthn++
jnthn mls: nice :)
er, moritz I meant :)
Well, mls++ did provide a nice patch too ;)
mls I must run valgrind again to see if the 20% rakudo spent in find_method is gone... 11:52
moritz (this timing has been on top of parrot master with whiteknight++'s execellent branch merged into it)
this combination also passes all spectests, fwiw
jnthn mls: Highly likely. 11:53
mls: New boolification protocol + massive method cache miss reduction should have helped with that. 11:54
11:54 cognominal left
mls sounds awesome! 11:54
11:54 cognominal joined
moritz rakudo: run('echo $$') 11:59
p6eval rakudo 3654b0: OUTPUT«32145␤»
JimmyZ rakudo: run('!!'); 12:00
p6eval rakudo 3654b0: OUTPUT«sh: !!: not found␤»
moritz it seems that rakudo's run() is what S29 now specifies to be called shell()
JimmyZ suggests add an invert method to Array, returns a Hash 12:03
TimToady: ^^ 12:04
tadzik hmm 12:05
b: say <a b c>.kv.hash.invert.perl
p6eval b 1b7dd1: OUTPUT«("a" => "0", "b" => "1", "c" => "2")␤»
moritz JimmyZ: there's a reason that Hash.invert returns a List, not a Hash
b: say <a b c>.kv>>.invert.perl 12:06
p6eval b 1b7dd1: OUTPUT«Method 'invert' not found for invocant of class 'Int'␤ in main program body at line 22:/tmp/XZ3XaCwElw␤»
moritz b: say <a b c>.kv>>.inveret.perl
p6eval b 1b7dd1: OUTPUT«Method 'inveret' not found for invocant of class 'Int'␤ in main program body at line 22:/tmp/PvuNVIN3V1␤»
moritz argl
JimmyZ moritz: I want invert key && value 12:08
12:08 satyavvd left
moritz JimmyZ: yes, but returning a Hash is wrong 12:08
JimmyZ what is better? 12:09
tadzik returning a list :)
moritz JimmyZ: because that way you lose information when there are duplicate values
JimmyZ: but @list Z=> @list.keys isn't too complicated either
mls hmm, I must reduce the profiler overhead, it influences the 'ticks' result too much 12:12
pmichaud++ complained that the profiler said that lineof is 20% of the execution time (for ng) 12:13
but reducing lineof to a simple ".return 0" only saves 6 of 120 seconds, that's 5% instead of 20% 12:14
I think it's because lineof execution consists of many many parrot ops, so the profiling overhead per op matters 12:16
I.e. the *profiling* run really is 20% faster with the dummy lineof(), just not the normal run
moritz a classical profiler problem 12:17
JimmyZ moritz: losing information is expected 12:18
moritz JimmyZ: by whom?
tadzik uhh
ww
JimmyZ moritz: whoever use array.invert :) 12:19
moritz JimmyZ: we didn't make that assuption for Hash.invert
JimmyZ moritz: return List wfm too 12:22
moritz: there is a invert method for hash, it's not bad that there is a invert method for array 12:24
moritz JimmyZ: I generally agree, just don't know if it's used often enough to warrant inclusion 12:25
felher flussence: well, before today i expected '$x .= foo + 3' to be '$x = $x.foo + 3'. It's just that i didn't expect that you can write someting on the right-hand-side of the 'x=' that doesn't affect the left-hand-side because of the low precedence of '='. I think i really need to study the spec on 'x=' again. :)
moritz ie I like the idea
JimmyZ moritz: I always invert a array :) 12:26
moritz: I use cn.php.net/manual/en/function.array-flip.php, which invert a array, hehe 12:28
moritz JimmyZ: I'm curious, could you name a few use cases? 12:29
another way to write it is my %h; %h{@list.keys} = @list # JFYI 12:31
tadzik oh, clever one
moritz I think that's how I usually do it in p5, but I do it so seldom I wonder why others use it often 12:32
(the p5 variant is $h{0..$#list} = @list, fwiw)
though for Perl 6 is Like the @list.keys Z=> @list version better 12:33
JimmyZ moritz: when search a value if it's in a small array and I don't want to write a loop, with a invert , I just write %h = @array.invert; if %h<value> { ... }
mls Some profiler changes brought lineof() down to 8% instead of 5%. I can live with that ;) 12:34
moritz ... and fail for the first array item?
in Perl 6, you'd say if 'value' eq any(@array) { ... } instead 12:35
JimmyZ forgot that one :(
moritz++
rakudo: my @list = <a b c d>; my %h; %h{@list} = @list.keys; say %h.perl; 12:38
p6eval rakudo 3654b0: OUTPUT«("a" => 0, "b" => 1, "c" => 2, "d" => 3).hash␤»
12:42 jerome_ left 12:45 JimmyZ_ joined, jerome_ joined 12:48 JimmyZ left, JimmyZ_ is now known as JimmyZ
JimmyZ rakudo: my @a = 1,2,34, hash(@a).say 12:48
p6eval rakudo 3654b0: OUTPUT«().hash␤»
moritz that looks wrongish 12:53
JimmyZ rakudo: my @a = 1,2,34; my %h = 1 => 'b', 5 => 'd'; sub hasssh(*@a, *%h) { my %b = @a, %h; }; hasssh(@a,%h).say 12:54
p6eval rakudo 3654b0: OUTPUT«("1" => 2, "34" => "1" => "b", "5" => "d").hash␤»
JimmyZ rakudo: my @a = 1,2,34; my %h = 1 => 'b', 5 => 'd'; sub hasssh(*@a, *%h) { my % = @a, %h; }; hasssh(@a,%h).say #
p6eval rakudo 3654b0: OUTPUT«("1" => 2, "34" => "1" => "b", "5" => "d").hash␤»
moritz rakudo: say hash(1, 2, 3).perl 12:55
p6eval rakudo 3654b0: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at src/gen/CORE.setting:4427␤ in sub hash at src/gen/CORE.setting:4486␤ in sub hash at src/gen/CORE.setting:4487␤ in <anon> at /tmp/hxsZZprAiO:1␤ in <anon> at /tmp/hxsZZprAiO:1␤␤»
moritz rakudo: say hash(1, 2, 3, 4).perl
p6eval rakudo 3654b0: OUTPUT«("1" => 2, "3" => 4).hash␤»
moritz rakudo: say hash(ny @ = 1, 2, 3, 4).perl
p6eval rakudo 3654b0: OUTPUT«===SORRY!===␤Non-declarative sigil is missing its name at line 1, near "@ = 1, 2, "␤»
moritz rakudo: say hash(ny @a = 1, 2, 3, 4).perl
p6eval rakudo 3654b0: OUTPUT«===SORRY!===␤Symbol '@a' not predeclared in <anonymous> (/tmp/TWl4NJCf9K:1)␤»
moritz huh? I'm declaring it right there!
rakudo: say hash(ny @a = (1, 2, 3, 4)).perl 12:56
p6eval rakudo 3654b0: OUTPUT«===SORRY!===␤Symbol '@a' not predeclared in <anonymous> (/tmp/8IWukZ873b:1)␤»
moritz b: say hash(ny @a = (1, 2, 3, 4)).perl
JimmyZ rakudo: my @a = 1,2,34; my %h = 1 => 'b', 5 => 'd'; hash(@a,%h).say
p6eval b 1b7dd1: OUTPUT«===SORRY!===␤Symbol '@a' not predeclared in <anonymous> (/tmp/nQ_YvZ_fb6:22)␤»
rakudo 3654b0: OUTPUT«("1" => 2, "34" => "1" => "b", "5" => "d").hash␤»
JimmyZ rakudo: my @a = 1,2,34; my %h; hash(@a,%h).say 12:57
p6eval rakudo 3654b0: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at src/gen/CORE.setting:4427␤ in sub hash at src/gen/CORE.setting:4486␤ in sub hash at src/gen/CORE.setting:4487␤ in <anon> at /tmp/G4r35PDeMe:1␤ in <anon> at /tmp/G4r35PDeMe:1␤␤»
jnthn moritz: ny :) 12:58
moritz ny?
moritz confused
jnthn you typed ny instead of my :)
moritz oh.
b: say hash(my @a = (1, 2, 3, 4)).perl
p6eval b 1b7dd1: OUTPUT«{"1" => 2, "3" => 4}␤»
moritz b: say hash(my @a = 1, 2, 3, 4).perl
p6eval b 1b7dd1: OUTPUT«{"1" => 2, "3" => 4}␤» 12:59
moritz rakudo: say hash(my @a = 1, 2, 3, 4).perl
p6eval rakudo 3654b0: OUTPUT«("1" => 2, "3" => 4).hash␤»
JimmyZ rakudo: my @a = 1,2,34, 5 ; hash(@a).say 13:02
p6eval rakudo 3654b0: OUTPUT«("1" => 2, "34" => 5).hash␤»
JimmyZ rakudo: my @a = 1,2,34 ; hash(@a).say
p6eval rakudo 3654b0: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at src/gen/CORE.setting:4427␤ in sub hash at src/gen/CORE.setting:4486␤ in sub hash at src/gen/CORE.setting:4487␤ in <anon> at /tmp/MN3VN9kDaB:1␤ in <anon> at /tmp/MN3VN9kDaB:1␤␤»
13:04 woosley joined 13:12 colomon left 13:19 abercrombie joined, woosley left 13:30 benabik left 13:32 bluescreen10 joined, colomon joined 13:33 colomon left 13:39 PacoLinux_ joined 13:44 pmurias joined 13:46 JimmyZ_ joined 13:47 SHODAN left 13:49 pmurias left 13:50 JimmyZ left, JimmyZ_ is now known as JimmyZ 13:56 Reaganomicon joined
dalek kudo/nom: f34aa80 | moritz++ | src/core/control.pm:
rename &run to &shell

It should have been that way for quite some time, according to the spec. Add a stub &run to ease the transition		 14:12
ast: ab15f81 | moritz++ | S32-trig/pi.t:
simplify pi.t
ast: 143c569 | moritz++ | / (2 files):
track spec change: &run -> &shell
14:17 wtw left 14:18 donri joined 14:21 amil joined
mls find_method is down from 20% to 2.01%. Good job, jnthn++ 14:25
14:26 woosley joined
tadzik nice 14:26
what's up?
I mean, you know... ;)
14:27 drbean joined
mls the "top" function is stll get_new_pmc_header. 27% 14:27
still
moritz so we're creating too many objects, right? :-) 14:28
14:28 risou_awy is now known as risou
mls right. Note that a function call needs at least two PMCs 14:28
one for the continuation, and one for the args
moritz so, we call too many functions too :(
mls maybe 14:29
Parrot_Sub_invoke: 20% 14:30
JimmyZ oh, that's 47%
moritz though not all new PMCs come from Parrot_Sub_invoke, right? 14:31
mls don't add them. the 20% and 27% include the cost of the called function
invoke calls get_new_pmc_header when it allocates the continuation 14:32
moritz: about 50%
invoke->pmc_new 13.78%, all pmc_new calls: 27.26% 14:33
14:35 cognominal_ joined
mls SixModelObject_get_bool: 5.44%. Is that good or bad? 14:35
JimmyZ is not sure 14:36
mls (It's actually 7.4% instead of 5.44%) 14:37
14:38 cognominal left
mls seen jnthn 14:38
aloha jnthn was last seen in #perl6 1 hours 39 mins ago saying "you typed ny instead of my :)".
mls jnthn: seems like we still have many BOOL_MODE_CALL_METHOD usages. Expected? 14:42
sorear good * #perl6 14:44
mls o/
sorear wonders if BOOL_MODE is based on Niecza's vtable system
jnthn Note that get_new_pmc_header almost certainly includes the time we spend running GC as well as the time we spend allocating. 14:48
sorear: no
sorear: Though if we came up with something similar that's maybe reassuring ;)
mls: Not entirely unexpected. I've a further optimization to do there, and after that if there's still a lot it'd be interesting to know where they're coming from. 14:49
mls jnthn: from the 18.1% of gc_gms_allocate_pmc_header, 13.93% are spent in GC
jnthn mls: ok, that's still a fair amount in allocation. 14:50
moritz: yes, we certainly call too many things...that's why we need an inliner. :) 14:51
14:53 uasi left, uasi joined 14:57 uasi left 14:59 [Coke] left 15:00 [Coke] joined 15:02 daniel-s left, daniel-s joined 15:05 thou joined 15:06 birdwindupbird left, im2ee joined
im2ee Hi! :) 15:07
tadzik hi inliner :) 15:09
abercrombie Hi #perl6, how to set the file pointer to the beginning of the file? 15:12
mux use seek(FILEHANDLE, 0, SEEK_SET) 15:13
oh, perl6 might have changed this though.
tadzik I suppose it did. The only thing on the top of my head is "close and open" 15:14
I'm sure there's a better way though
abercrombie Yes, right now I am using close and open. But I think it's better if I can reset the file pointer
15:14 molaf joined 15:16 thou left 15:36 woosley left 15:38 benabik joined 15:44 phenny left, phenny joined
abercrombie niecza: my $a='abc'; if $a ~~ s/(.*)b(.*)/$0$1/ {say $a} 15:48
p6eval niecza v9-10-g678102e: OUTPUT«ac␤»
abercrombie rakudo: my $a='abc'; if $a ~~ s/(.*)b(.*)/$0$1/ {say $a}
p6eval rakudo f34aa8: OUTPUT«abc␤»
abercrombie How to do it in rakudo?
sorear huh 15:49
tadzik b: my $a='abc'; if $a ~~ s/(.*)b(.*)/$0$1/ {say $a}
p6eval b 1b7dd1: OUTPUT«Any()Any()␤»
tadzik pff
TimToady rakudo: my $a='abc'; if $a.subst(/(.*)b(.*)/, -> $/ { "$0$1" }) { say $a } 15:50
p6eval rakudo f34aa8: OUTPUT«abc␤»
TimToady hmm
b: my $a='abc'; if $a.subst(/(.*)b(.*)/, -> $/ { "$0$1" }) { say $a }
p6eval b 1b7dd1: OUTPUT«abc␤»
TimToady double hmm
jnthn o.O 15:51
TimToady b: my $a='abc'; if $a.subst(/(.*)b(.*)/, -> $/ { say $0; say $1; "$0$1" }) { say $a }
p6eval b 1b7dd1: OUTPUT«a␤c␤abc␤»
jnthn oh, .subst won't mutate $a
TimToady duh
jnthn hands TimToady another coffee :) 15:52
TimToady b: my $a='abc'; if $a.=subst(/(.*)b(.*)/, -> $/ { "$0$1" }) { say $a }
p6eval b 1b7dd1: OUTPUT«ac␤»
sorear niecza has $a.subst(PAT, SUBST, :inplace) for this
TimToady rakudo: my $a='abc'; if $a.=subst(/(.*)b(.*)/, -> $/ { "$0$1" }) { say $a }
p6eval rakudo f34aa8: OUTPUT«abc␤»
TimToady bug
jnthn indeed. 15:53
TimToady hands pmichaud a DP
sorear DP?
jnthn rakudo: my $a='abc'; $a.match(/(.*)b(.*)/).perl.say
p6eval rakudo f34aa8: OUTPUT«␤»
jnthn rakudo: my $a='abc'; $a.match(/(.*)b(.*)/).say
p6eval rakudo f34aa8: OUTPUT«␤»
TimToady Dr Pepper
jnthn rakudo: my $a='abc'; say ?$a.match(/(.*)b(.*)/)
p6eval rakudo f34aa8: OUTPUT«Bool::False␤» 15:54
jnthn rakudo: my $a='abc'; say ?$a.match(/(.*?)b(.*?)/)
p6eval rakudo f34aa8: OUTPUT«Bool::True␤»
abercrombie jnthn: I know subst won't mutate the original string. How about s/// form in perl6?
jnthn Backtracking FAIL
abercrombie: s/// mutates.
15:54 benabik left
abercrombie so currently b: and rakudo: all fail on s/// ? 15:54
15:55 phenny left, phenny joined
jnthn Seems in b the s/// runs, but $0 and $1 aren't available. 15:55
TimToady rakudo: $_ = 'entirely'; s/^/not /; .say 15:56
p6eval rakudo f34aa8: OUTPUT«not entirely␤»
16:00 JimmyZ left 16:13 MayDaniel joined 16:18 daniel-s left
abercrombie I thought the default loop var $_ in a for loop is writable, but it seems not true in Rakudo July 16:19
TimToady b: my @a = 1,2,3; for @a { $_++ }; say @a 16:20
p6eval b 1b7dd1: OUTPUT«234␤»
TimToady looks writable to me 16:21
rakudo: my @a = 1,2,3; for @a { $_++ }; say @a
p6eval rakudo f34aa8: OUTPUT«2 3 4␤»
TimToady can you give a counterexample?
abercrombie my $fh = open 'a.txt'; .say if .=subst(...) for $fh.lines; 16:22
I got the following error: Cannot modify readonly value
TimToady sorear: re irclog.perlgeek.de/perl6/2011-09-04#i_4372584, I think you'll find that Alaska is further from Quebec than SD is, and fits your criteria
abercrombie if I use $some=.substr(..) instead, I am fine
s/substr/subst/
TimToady the values returned from .lines are readonly because they have no container 16:23
16:23 icwiener joined
abercrombie Isn't the returned value stored in $_ ? 16:23
jnthn Essentially, You're trying to modify values coming back from an iterator rather than looping over an array.
TimToady Perl 5 did everything with containers, so you can modify any temp values. In Perl 6 we're trying to keep better track of immutability so that the optimizer will (someday) have more to work with. 16:24
jnthn I *think* we're binding to $_ in that case, so there's no container. 16:25
16:25 MayDaniel left
jnthn Well, we'd have to bind 16:25
Otherwise we wouldn't be able to modify original things when there are containers. :)
abercrombie Oh, I got it 16:26
TimToady rakudo: .say if s/am/pm/ for [$*IN.lines].list 16:28
p6eval rakudo f34aa8: OUTPUT«Cannot assign to a readonly variable or a value␤ in sub infix:<=> at src/gen/CORE.setting:6870␤ in block <anon> at /tmp/YGzjR0Hb5U:1␤ in method reify at src/gen/CORE.setting:3601␤ in method reify at src/gen/CORE.setting:3506␤ in method reify at src/gen/CORE.se…
TimToady hmm 16:29
rakudo: .say if s/am/pm/ for [$*IN.lines].values
p6eval rakudo f34aa8: OUTPUT«Cannot assign to a readonly variable or a value␤ in sub infix:<=> at src/gen/CORE.setting:6870␤ in block <anon> at /tmp/R9oo1Ij1JU:1␤ in method reify at src/gen/CORE.setting:3601␤ in method reify at src/gen/CORE.setting:3506␤ in method reify at src/gen/CORE.se…
TimToady rakudo: .say if s/am/pm/ for @([$*IN.lines])
p6eval rakudo f34aa8: OUTPUT«Cannot assign to a readonly variable or a value␤ in sub infix:<=> at src/gen/CORE.setting:6870␤ in block <anon> at /tmp/5JxvpLO3qa:1␤ in method reify at src/gen/CORE.setting:3601␤ in method reify at src/gen/CORE.setting:3506␤ in method reify at src/gen/CORE.se…
TimToady jnthn: what's up with that ^^
rakudo: .say for @([$*IN.lines]) 16:30
p6eval rakudo f34aa8: OUTPUT«Land der Berge, Land am Strome,␤Land der Äcker, Land der Dome,␤Land der Hämmer, zukunftsreich!␤Heimat bist du großer Söhne,␤Volk, begnadet für das Schöne,␤vielgerühmtes Österreich,␤vielgerühmtes Österreich!␤␤Heiß umfehdet, wild umstritten␤liegst dem Erdteil du inmi…
jnthn rakudo: [<a b c>][1] = 'd'
p6eval rakudo f34aa8: ( no output )
jnthn hm
TimToady rakudo: .say if .=subst(/am/, 'pm') for @([$*IN.lines])
p6eval rakudo f34aa8: OUTPUT«Cannot assign to a readonly variable or a value␤ in method dispatch:<.=> at src/gen/CORE.setting:623␤ in block <anon> at /tmp/RBKI2SuFjC:1␤ in method reify at src/gen/CORE.setting:3601␤ in method reify at src/gen/CORE.setting:3506␤ in method reify at src/gen/C…
jnthn rakudo: [<a b c>][1] .= subst('b', 'x') 16:31
p6eval rakudo f34aa8: ( no output )
TimToady say?
jnthn TimToady: Didn't want output, just wanted to see if it complained about lack of rw-ness
(trying to figure out if it's just down to the iterator) 16:32
rakudo: @([$*IN.lines]).perl.say
p6eval rakudo f34aa8: OUTPUT«Array.new("Land der Berge, Land am Strome,", "Land der Äcker, Land der Dome,", "Land der Hämmer, zukunftsreich!", "Heimat bist du großer Söhne,", "Volk, begnadet für das Schöne,", "vielgerühmtes Österreich,", "vielgerühmtes Österreich!", "", "Heiß umfehdet, wild um…
jnthn rakudo: [$*IN.lines][0] = 'foo'
p6eval rakudo f34aa8: ( no output )
jnthn rakudo: @([$*IN.lines])[0] = 'foo'
p6eval rakudo f34aa8: ( no output )
jnthn hmmm. 16:33
rakudo: for @([$*IN.lines]) { $_ = 'foo' }
p6eval rakudo f34aa8: ( no output )
jnthn rakudo: for @([$*IN.lines]) { .= say } 16:34
p6eval rakudo f34aa8: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 1␤»
jnthn rakudo: for @([$*IN.lines]) { .=say }
p6eval rakudo f34aa8: OUTPUT«Land der Berge, Land am Strome,␤Land der Äcker, Land der Dome,␤Land der Hämmer, zukunftsreich!␤Heimat bist du großer Söhne,␤Volk, begnadet für das Schöne,␤vielgerühmtes Österreich,␤vielgerühmtes Österreich!␤␤Heiß umfehdet, wild umstritten␤liegst dem Erdteil du inmi…
jnthn rakudo: for @([$*IN.lines]) { .=subst('am', 'pm') }
p6eval rakudo f34aa8: ( no output )
jnthn .=subst('am', 'pm') for @([$*IN.lines])
TimToady seems to be only the modifier form
jnthn rakudo: .=subst('am', 'pm') for @([$*IN.lines])
p6eval rakudo f34aa8: OUTPUT«Cannot assign to a readonly variable or a value␤ in method dispatch:<.=> at src/gen/CORE.setting:623␤ in block <anon> at /tmp/iCFojRSfI1:1␤ in method reify at src/gen/CORE.setting:3601␤ in method reify at src/gen/CORE.setting:3506␤ in method reify at src/gen/C…
jnthn TimToady: Yeah, that's what I was just concluding too.
16:35 MayDaniel joined
jnthn ah, I think I see what might be going on... 16:35
TimToady is something decontainerizing? 16:36
jnthn TimToady: Yeah, the thunk we pass to map.
Decontainerizes the $_ 16:37
I should be $dayjob-ing right now; I'll patch it this evening.
16:37 phenny left
TimToady 'k, thanks 16:37
16:38 phenny joined 16:41 frhodes left 16:44 phenny left 16:45 phenny joined 16:50 mberends left 16:57 zby_home_ joined 17:01 phenny left 17:02 frhodes joined, phenny joined 17:08 masak joined
masak evenin', channel. 17:08
jnthn good evening, human. 17:09
masak glances nervously around him 17:10
tadzik good evening masak
masak tonight's plan: do a few of the things I planned to do in the weekend, but didn't :) 17:11
notably, I'll try blogging a bit. 17:12
17:16 sftp joined, amil left 17:36 wamba joined 17:37 thou joined 17:41 ashleyde1 left, ashleydev joined
sorear o/ masak 17:47
TimToady: what is the functional difference between not having a container and a SvREADONLY container?
flussence er, this? metacpan.org/module/Readonly::XS 17:51
sorear flussence: vaguely 17:52
flussence: SvREADONLY is a core Perl 5 feature
17:52 icwiener left 17:55 frhodes left 18:06 benabik joined 18:10 colomon joined 18:14 molaf left 18:16 benabik left 18:17 thou left 18:22 benabik joined 18:24 benabik left, benabik joined 18:29 risou is now known as risou_awy 18:32 amil joined 18:36 mj41 joined 18:50 daxim left 18:54 envi_laptop left
masak nice email on p6u. 18:54
masak replies to it
rakudo: my $c = 0..20; say +$c 18:58
p6eval rakudo f34aa8: OUTPUT«21␤»
masak rakudo: my $c = 0..20; say ~$c
p6eval rakudo f34aa8: OUTPUT«0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20␤»
18:58 frhodes joined 18:59 frhodes left
masak rakudo: my $c = 0.5 .. 5.5; say +$c 19:04
p6eval rakudo f34aa8: OUTPUT«6␤»
masak rakudo: my $c = 0.5 .. 5.4; say +$c
p6eval rakudo f34aa8: OUTPUT«5␤»
tadzik masak++ # responses 19:06
sorear We need to coordinate this better 19:07
I responded at the same time
to the other thread
also, my responce directly contradicted masak's on one point.
19:08 wamba left
masak sorear: I posted here on the channel that I was responding. I figgered that'd reduce duplicated effort. :P 19:08
apparently not.
sorear masak: I checked this channel before I started writing, and checked it again after I was done. 19:09
clearly, *I* should have posted that I was responding
masak yes. :) 19:11
though I liked your responses, too. and I don't see the point of our disagreement.
er. I mean I didn't find any. 19:12
lue hello planet o/
masak good time, lue.
19:12 envi_laptop joined
lue should get in the habit of checking his email on a semi-regular basis 19:13
sorear masak: I said that if {...} { say "yes" } was correct (it said yes). You said bug. 19:14
masak hm. 19:15
I think I agree that it should say "yes".
masak re-reads what he wrote
lue rakudo: say ?{...}
p6eval rakudo f34aa8: OUTPUT«Bool::True␤»
masak I posted a correction. sorear++ masak-- 19:17
19:21 birdwindupbird joined
lue I wonder if I should sign up for p6l/p6u again... 19:28
moritz p6u is low traffic
masak moritz! \o/
19:31 mj41 left 19:33 frhodes joined 19:34 donri left
lue how about die("of a awful fate" :noloc) or something like that? 19:35
masak sans TTIAR, perhaps :)
moritz lue: read my mail to p6u
sorear niecza: die "of an awful fate" :noloc 19:36
jnthn die-quietly "of a awful fate"
p6eval niecza v9-10-g678102e: OUTPUT«===SORRY!===␤␤You can't adverb that at /tmp/91CmRUmJD9 line 1 (EOF):␤------> die "of an awful fate" :noloc⏏<EOL>␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 685 (CORE die @ 2) …
moritz and p6l
masak lue: it's generally considered a design smell to have a negatively-phrased adverb.
sorear niecza: (die "of an awful fate") :noloc
p6eval niecza v9-10-g678102e: OUTPUT«===SORRY!===␤␤You can't adverb that at /tmp/zcrgTY4JkJ line 1 (EOF):␤------> (die "of an awful fate") :noloc⏏<EOL>␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 685 (CORE die @ 2…
sorear hrhmf
moritz std: 'foo' :a
p6eval std e3c970e: OUTPUT«ok 00:01 119m␤»
lue I read the p6u email, but didn't get to your condemnation of the named argument before saying that :)
moritz std: sort(1, :foo :bar)
p6eval std e3c970e: OUTPUT«ok 00:01 121m␤»
moritz named arguments aren't bad per se -- they are just a hint to investigate if there's something we could do better on another level 19:37
lue masak: TTIAR? [Ten Times In A Row ?]
masak Two Terms.
lue: I bet you know this by now :) 19:38
lue yes, just not the acronym :)
.oO(For quiet &die, perhaps &faint [implying it died (or at least went unconscious] before it could say everything]?)
 19:40
masak lue: there's already &exit and &sleep 19:41
lue Hm. (mortiz is right, figuring out the way to do &die without backtrace is interesting) 19:42
tadzik passout() 19:43
jnthn whine 19:45
lue
.oO(if it's a user error (and you don't want backtrace), it's like they gave a nasty blow to the poor program, perhaps &ko for quiet form of &die?)
I like the sound of &whine :) 19:46
masak likes the taste of beer :) 19:51
19:52 MayDaniel left, mux left
tadzik :P 19:52
lue
.oO(whine "you never pick the right card! Never never never never!";)
tadzik whine() should randomly say "you don't love me anymore" or "I'm not talking to you" 19:53
masak it should play discordant tunes on the PC speaker.
man, Perl 6 gonna be *great*. 19:54
blog post! strangelyconsistent.org/blog/dash-n...p-part-two 19:55
TimToady: that one's for you :)
I think that might be the most confused Perl 6 blog post I've ever written. but that's how I feel about the {YOU_ARE_HERE} approach to -n and -p. 19:57
in a way, the current AST approach is an immunoresponse to the abject handwavery of the spec.
19:59 thou joined, mux joined
sorear meh, the spec is not that ad 20:00
bad
masak how does Niecza implement it?
sorear CORN.setting: for lines {YOU_ARE_HERE} 20:01
I'm not crazy about TimToady's choice of names
a module does not have one setting, it has a stack of them
SETTING means UNIT::OUTER 20:02
in a typical program, SETTING === CORE
typically you have NULL -> CORE -> (program)
in an eval, you need SETTING::SETTING, because it's NULL -> CORE -> (program) -> (eval)
with -n you have NULL -> CORE -> CORN -> (program) 20:03
Niecza interprets -n, -p, and -L as applying to the main program only; there is currently no way to override the setting for modules, but it would be LHF 20:04
20:04 birdwindupbird left
masak ooh! 20:04
I'm glad I posted my confused rant, if the result was this nice answer. 20:05
sorear the interesting bit comes in where this interacts with the compiled module cache, if you have some people wanting to use CORE->CORN and other people wanting to use CORE->CompanySetting->CORN on the same niecza install
masak that's what TimToady meant by saying that the settings are nested...
sorear I haven't actually read your rant yet. Let me do that now
masak well, it's basically obsolete at this point :)
I need to write a part three now. 20:06
lue considers using the name 'frotzlexity' for a P6-coded IF
masak oddly appropriate. 20:07
TimToady note that CORN is really the name of the new language, which includes its entire pedigree CORE+N 20:08
masak nodnod
TimToady CORE->CompanySetting->CORN is non-sensical
CORE+CompanySetting+N is its true identity
sorear oh!
20:09 envi_laptop left
TimToady I keep saying this, and I'm glad to keep saying it: we must always know what language we are in while we're parsing 20:09
so relative language names don't cut it, when you need a universal naming scheme 20:10
masak maybe we don't understand you because we don't know what language you're speaking :P
20:12 bluescreen10 left
TimToady At this point I believe I am speaking Axiomatic rather than Handwaving... 20:13
cotto pmichaud, ping
TimToady problem is, both of those languages always leave something unsaid...
sorear TimToady: would it be reasonable for me to a) rename COR{N,P}.setting to {N,P}.setting; b) change -L to accept a +-path line CORE+P, CORE+N; and c) change -n,-p to append +P or +N to the language name 20:15
?
masak TimToady: thanks for explaining this.
TimToady sorear: sure, that'd be reasonable
and more composable, as masak++ wishes
though I'd be a little unhappy if people start abusing settings when they should just have a 'use', where a 'use' can adequately tweak the grammar 20:16
-n and -p require a setting because 'use' can't retroactively put a loop around UNIT without cheating 20:17
sorear in CORE.setting: my $_runner = *(); $_runner.({YOU_ARE_HERE}); 20:19
in N: IMPORT { $CALLER::_runner = sub ($code) { for lines { $code() } } }
easy. 20:20
also, I kind of liked the eskimo operator. 20:22
masak I admire it as a neat hack. 20:25
I don't find it particularly useful or worth preserving.
20:26 bluescreen10 joined
moritz one can replace it by a LAST phaser quite elegantly and more readable, no? 20:30
masak aye.
even in Perl 5, '}{ ...' is just a funky way to say 'END { ... }'
20:31 kaare_ left
lue b: say "'hai\"" ~~ /$bq=(\"||\') <?before $bq> \N* / 20:31
p6eval b 1b7dd1: OUTPUT«===SORRY!===␤Confused at line 22, near "say \"'hai\\"␤»
moritz don't use p5 regex syntax in p6 20:32
std: /$bq=(\"||\') <?before $bq> \N* /
p6eval std e3c970e: OUTPUT«===SORRY!===␤Variable $bq is not predeclared at /tmp/Ggk10Fntqo line 1:␤------> /$bq⏏=(\"||\') <?before $bq> \N* /␤Variable $bq is not predeclared at /tmp/Ggk10Fntqo line 1:␤------> /$bq=(\"||\') <?before $bq⏏> \N* …
lue
.oO(I accidentaly used P5? Wow.)
 20:34
20:35 frhodes left
moritz not quoting = is what I actually meant :-) 20:38
20:38 im2ee left, im2ee joined 20:41 supernovus joined
moritz masak++ # blag 20:42
masak henceforth I'll not hesitate to post "I'm so confused!" :P 20:43
20:50 whiteknight left
lue
.oO(Now is a good time I think to start coding Frotzlexity's world model)
 20:57
TimToady sorear: your 'lines' needs a ()
masak std: for lines { .say } 20:59
p6eval std e3c970e: OUTPUT«===SORRY!===␤Function 'lines' needs parens to avoid gobbling block at /tmp/6MtcddYvQ1 line 1:␤------> for lines⏏ { .say }␤Missing block (apparently gobbled by 'lines') at /tmp/6MtcddYvQ1 line 1:␤------> for lines { .say }[33…
21:03 Trashlord left 21:07 amil left, donri joined
dalek kudo/nom: 493e23e | jonathan++ | src/Perl6/Actions.pm:
Fix rw-stripping bug in statement modifier form of for.
 21:08
21:09 frhodes joined
masak 'night, #perl6 21:11
21:11 masak left
im2ee good night! :) 21:21
tadzik o/ 21:22
21:25 im2ee left 21:36 supernovus left 21:37 zby_home_ left
dalek kudo/optimizer: 5ab4e11 | jonathan++ | / (4 files):
Add an optimizer stage to the compiler, which for now we only enable when --optimize is passed.
 21:50
kudo/optimizer: de1b900 | jonathan++ | src/Perl6/ (3 files):
Start sketching out inlining of statements of simple blocks. Many issues, but handles a basic example case.
kudo/optimizer: 3f25af9 | jonathan++ | src/Perl6/Optimizer.pm:
Start to harden and improve the immediate block inline handling a bit.
kudo/optimizer: 471fc0d | jonathan++ | src/Perl6/Optimizer.pm:
Topic preservation over immediate block inlines. A little refactoring. Preliminaries for detecting cases where we can 'inline' onlystar protos (though it won't be so much an inlining as just jumping right into the multi-dispatcher).
jnthn Mmm. Oktoberfest bier.
tadzik wooo 21:53
jnthn: does this work?
jnthn tadzik: The beer seems to work pretty well, yes.
tadzik: The optimizer...well, it only implements one optimization. I didn't try it on anything serious yet (like the setting).
This is mostly me having some -Ofun rather than working on the stuff I probably should be to get a nom-based release in shape... :) 21:54
jnthn checks if CORE.setting builds on it, just out of curiosity. 21:56
(e.g. with --optimize)
21:59 zby_home joined
jnthn Well, it builds a setting that compiles 22:00
as in, we make it through the build
Survive hello world too.
Claims it inlined 174 immediate blocks. 22:01
benabik jnthn++
tadzik to benchmark? :) 22:02
well, measure
jnthn Guess could always see if it makes any impact on the spectests, though I'd be skeptical of much given it's only doing one thing. :) 22:03
spectest is looking OK so far.
ah, it's caused some fails 22:07
ah, hm 22:18
benabik hmm? 22:20
jnthn think I tracked down what's going on. 22:21
dalek kudo/optimizer: 5b1e31b | jonathan++ | src/Perl6/Optimizer.pm:
Ensure that topic restoration doesn't clobber block result!
 22:28
22:29 plobsing left 22:30 plobsing joined 22:31 djanatyn joined
djanatyn Heyo. 22:33
jnthn o/ djanatyn 22:34
Yay, tests looking better after that fix.
felher jnthn++ 22:38
jnthn: where do i have to pass the --optimize-flag? to Configure.pl? 22:39
tadzik jnthn: is there a performance win?
jnthn felher: Are you talking about the current work I'm doing in the optimizer branch? 22:40
felher jnthn: yeah :)
jnthn felher: It's really to early to sanely use it for anything. But if you want to play, you just pass Rakudo a --optimize flag
If you want to try and optimized setting, just hack the appropriate line in Makefile
*an 22:41
felher: This is *extremely* early and experimental work at the moment, though.
felher jnthn: Yeah, np. I just want to play with it :)
jnthn: thnx
jnthn :) 22:42
22:44 bluescreen10 left 22:52 benabik left
jnthn ok, sleep time & 22:52
tadzik jnthn: g'night
22:57 daniel-s joined 22:59 Trashlord joined 23:00 frhodes left 23:03 zby_home left
tadzik ooh, win 23:03
23:04 contempt_ joined
tadzik gist.github.com/1196124 23:04
b: say (34.89 - 27.12) / 34.89 23:05
p6eval b 1b7dd1: OUTPUT«0.222699914015477␤»
tadzik goo.gl/acAxg
23:05 contempt_ left
felher tadzik: hui, quite nice. :) 23:08
23:08 whiteknight joined
tadzik mandelbrot does not react to --optimize at all 23:15
sorear niecza: my $start = times[0];my $i = 0; for 1..100000 { $i = $i + $_; }; say $i; say times[0] - $start; 23:17
p6eval niecza v9-10-g678102e: OUTPUT«5000050000␤4.136259␤»
sorear disappointing
oh right I completely forgot to optimize Ranges
tadzik g'night #perl6 23:19
felher good night tadzik 23:21
23:35 Psyche^ joined 23:38 Patterner left, Psyche^ is now known as Patterner 23:39 uasi joined 23:43 uasi left 23:44 uasi joined 23:48 uasi left

Powered by Cro and the Raku® Programming Language.

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

Thank you!