»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
diakopter japhb: ^ that's a long time from now 00:00
japhb timotimo: Yeah, that's an irssi fail. xchat-gnome was smart enough to highlight any fallback variant of your nick.
00:00 japhb is now known as japhb_
japhb_ botsnack 00:00
yoleaux 2 Sep 2014 00:34Z <timotimo> japhb_: I'm still wondering why the benchmark plots don't disappear when there is 0 data at all, like the non-nqp-implemented benchmarks in a pure-nqp-comparison
2 Sep 2014 00:37Z <timotimo> japhb_: i wonder when it'll be time to re-implement the perl5 portions of p6bench in pure perl6 :)
2 Sep 2014 14:45Z <brrt> japhb_: i'm back (for 45 minutes or so)
timotimo .in 5d ermergerd blergperst! 00:01
yoleaux timotimo: I'll remind you on 8 Sep 2014 00:01Z
japhb_ Gah, 3 missed messages.
timotimo maybe yoleaux can be taught some nicks belong to the same person?
japhb_ wonders if yoleaux can be taught how to detect nicks owned by the same human, since registering a nick allows that.
heh 00:02
timotimo %)
diakopter .wa japhb 00:03
timotimo japhb_: have you thought about where a functionality to run a single scale of all benchmarks in a row, perhaps without swallowing the output?
yoleaux diakopter: Sorry, that command (.wa) crashed.
timotimo so that i can a) verify nothing crashes, b) extract spesh and/or jit logs and/or profiles
diakopter .wa timotimo 00:04
yoleaux Timothy (male given name): Information for US births: rank: 125th; fraction: 1 in 663 people (0.15%); number: 3016 people per year (US data based on 2013 births and other SSA registrations in the US); Estimates for current US population: expected total number alive today: 980815 people; expected population fraction: 1 in 245 people (0.41%); expected rank: 32nd; most common age: 51 years; (using standard US mortality data)
diakopter .wa TimToady
japhb_ OK, second question first: It'll be time to do the reimplementation of timeall when I trust timing results from Rakudo as much as I do from perl5. Which will be a while, I think. And reimplementing analyze will basically wait until parsing JSON is as fast in Rakudo as it is in perl5.
yoleaux Timothy (male given name): Information for US births: rank: 125th; fraction: 1 in 663 people (0.15%); number: 3016 people per year (US data based on 2013 births and other SSA registrations in the US); Estimates for current US population: expected total number alive today: 980815 people; expected population fraction: 1 in 245 people (0.41%); expected rank: 32nd; most common age: 51 years; (using standard US mortality data)
00:05 chenryn joined
timotimo .tell brrt i wonder why i didn't get any inlines at all in the previous instance i tried it, but now i do see inlines show up in the jitlog. maybe it only shows inlines if it has deopt points nearby? 00:05
yoleaux timotimo: I'll pass your message to brrt.
japhb_ timotimo: I'm not sure what you mean about the single scale of all benchmarks in a row. (The first interpretation of that question that comes to mind makes no sense, so clearly I'm reading it wrong.)
00:06 japhb_ is now known as japhb
timotimo aye, phrasing that wasn't easy %) 00:06
i'd like for the bench utility to run every benchmark a single time, but not start at the lowest scaling
japhb As for empty plots not disappearing, --skip-incomplete may get you what you want 00:08
timotimo didn't last time i tried it :(
japhb timotimo: Given how widely varying the scaling starting points are, how do you propose to choose the scaling they run at? A fixed number of increments above the starting scale? (E.g. running at 3 increments above would make a test with scale => 1 << 7 start at 1 << 10) 00:09
timotimo: Hmmm, that's odd.
timotimo i thought so, too :)
let me retry that
00:09 bjz_ left 00:10 bjz joined
japhb TBH, I may not have tested it with 'bench --format=html_plot compare' but rather with 'bench --format=html_plot history' 00:10
So it may never have worked for standard comparison plots. :-( 00:11
Or to be more precise, bench --format=html_plot --skip-incomplete compare'
GAH, you know what I mean. 00:12
diakopter .gcs "japhb" "TimToady" 00:13
yoleaux [site] "japhb" (121,000), "TimToady" (50,200)
timotimo :)
00:13 mberends joined
timotimo i just found out what i've been doing wrong with bindattrs_* and getattrs_* 00:13
diakopter .gc mberends
timotimo one of these "how could that ever have worked?!" pieces
yoleaux 13,200 (site), 1,420 (api)
japhb wonders how the heck he ended up with a higher gcs than TimToady, and suspects complete the results are complete BS 00:16
s:1st/complete // 00:17
timotimo gcs?
japhb timotimo: diakopter's query from a few minutes ago
timotimo yeah, but what is "gcs"? 00:19
japhb dpk.io/yoleaux, upper left section (Google) 00:20
00:20 chenryn left
timotimo aah, it's like "google fight" 00:21
00:26 ajr_ left
timotimo .chars box 00:29
yoleaux ⌧ ⍽ ⎸ ⎹ ␣ ─ ━ │ ┃ ┄ ┅ ┆ ┇ ┈ ┉ ┊ ┋ ┌ ┍ ┎ ┏ ┐ ┑ ┒ ┓ └ ┕ ┖ ┗ ┘ ┙ ┚ ┛ ├ ┝ ┞ ┟ ┠ ┡ ┢ ┣ ┤ ┥ ┦ ┧ ┨ ┩ ┪ ┫ ┬ ┭ ┮ ┯ ┰ ┱ ┲ ┳ ┴ ┵ ┶ ┷ ┸ ┹ ┺ ┻ ┼ ┽ ┾ ┿ ╀ ╁ ╂ ╃ ╄ ╅ ╆ ╇ ╈ ╉ ╊ ╋ ╌ ╍ ╎ ╏ ═ ║ ╒ ╓ ╔ ╕ ╖ ╗ ╘ ╙ ╚ ╛ ╜ ╝ ╞ ╟ ╠ ╡ ╢ ╣ ╤ ╥ …
╦ ╧ ╨ ╩ ╪ ╫
timotimo i didn't know this command and it's definitely useful!
00:32 PZt left 00:35 chenryn joined 00:42 bjz left 00:50 chenryn left 00:59 _thou left
nick_____ so otter i believe that it would be an nice mascot speakerdeck.com/stevan_little/perl...a-dead-end 01:00
01:01 dolmen left 01:02 FROGGS_ joined, nick_____ left 01:05 FROGGS left 01:10 avuserow joined 01:13 dayangkun joined 01:14 avuserow left 01:20 bcode is now known as we 01:27 avuserow joined
dj_goku_ ggoebel1111112: kind of reminds me of: www.nand2tetris.org 01:38
01:50 grondilu joined 01:51 nbrown__ joined
grondilu so normally we can define a class in its own file by starting with 'class Foo;' but if Foo is supposed to derive from an other role/class, how can we declare it? where to put the 'use Bar'? 01:55
oh I can just write 'use Bar; class Foo is Bar;' apparently
(it's a bit redundant if you ask me) 01:57
01:59 chenryn joined 02:02 nbrown__ left 02:03 noganex_ joined 02:06 noganex left 02:18 Alula_ left, Alula_ joined 02:22 nbrown__ joined
grondilu m: class A { method WHICH { rand } }; my %h{A} of Real; # this works fine 02:25
camelia ( no output )
grondilu m: class A { method WHICH { rand } }; class B { has %h{A} of Real } # this doesn't
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/IhKV4fk0Yx␤Cannot call 'trait_mod:<of>'; none of these signatures match:␤:(Mu:U $target, Mu:U $type)␤:(Routine:D $target, Mu:U $type)␤:(Variable:D $v, Any |c)␤:(Variable:D $v, Mu:U $of)␤…»
BenGoldberg m: class A { method WHICH { rand } }; class B { has (%h{A} of Real) }; 02:26
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/FD3amJLb68␤Unable to parse expression in declarator; couldn't find final ')' ␤at /tmp/FD3amJLb68:1␤------> thod WHICH { rand } }; class B { has (%h⏏{A} of Real) };␤ e…»
02:26 nbrown__ left 02:27 raiph left
TimToady m: class A { method WHICH { rand } }; class B { has Real %h{A} } 02:33
camelia ( no output )
TimToady just doesn't like the 'of' form for some reason
02:34 BenGoldberg left
TimToady mind you, you'll have a hard time looking up any number in that hash 02:34
m: m: class A { method WHICH { rand } }; my Real %h{A}; %h{42} = "hi there"; say %h{42}; # maybe this works 02:37
camelia rakudo-moar 41d7f7: OUTPUT«cannot stringify this␤ in method assign_key at src/gen/m-CORE.setting:10054␤ in sub postcircumfix:<{ }> at src/gen/m-CORE.setting:2799␤ in block <unit> at /tmp/38DGe8Yq30:1␤␤»
TimToady m: m: class A { method WHICH { ~rand } }; my Real %h{A}; %h{42} = "hi there"; say %h{42}; # maybe this works
camelia rakudo-moar 41d7f7: OUTPUT«cannot stringify this␤ in method assign_key at src/gen/m-CORE.setting:10054␤ in sub postcircumfix:<{ }> at src/gen/m-CORE.setting:2799␤ in block <unit> at /tmp/5XMdiswfEv:1␤␤»
02:41 dayangkun left
grondilu "hard time looking up any number" because of the rand? I put it just for an example. Since A has to be a custom class to produce the bug. 02:43
02:43 _thou joined
grondilu m: class B { has %h{Int} of Real } 02:43
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/MyaxutE6_U␤Cannot call 'trait_mod:<of>'; none of these signatures match:␤:(Mu:U $target, Mu:U $type)␤:(Routine:D $target, Mu:U $type)␤:(Variable:D $v, Any |c)␤:(Variable:D $v, Mu:U $of)␤…»
grondilu ah no. My bad
grondilu has to admit he does not fully understand the difference between 'my Real %h{Foo}' and 'my %h{Foo} of Real' 02:44
02:47 _thou left 03:02 telex left 03:04 telex joined
xenoterracide how do I print to stderr 03:41
TimToady note is the simplest 03:44
$*ERR.print and such 03:46
xenoterracide TimToady: ty 03:50
04:04 rindolf joined, [Sno] left 04:06 dayangkun joined 04:08 kaare_ joined
grondilu std: class A does Numeric { }; multi infix:<==>(A $, A $) { False }; say A != A;' 04:08
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unable to parse single quotes at /tmp/Gi_UUCAOyL line 1:␤------> ix:<==>(A $, A $) { False }; say A != A;⏏'␤Couldn't find final "'"; gave up at /tmp/Gi_UUCAOyL line 1 (EOF):␤------> x:<==>(A $, A $)…»
grondilu std: class A does Numeric { }; multi infix:<==>(A $, A $) { False }; say A != A;
camelia std 53b3ca6: OUTPUT«ok 00:01 130m␤»
grondilu m: class A does Numeric { }; multi infix:<==>(A $, A $) { False }; say A != A;
camelia rakudo-moar 41d7f7: OUTPUT«use of uninitialized value of type A in numeric context in block <unit> at /tmp/KuWR6fbevG:1␤␤use of uninitialized value of type A in numeric context in block <unit> at /tmp/KuWR6fbevG:1␤␤False␤»
grondilu n: class A does Numeric { }; multi infix:<==>(A $, A $) { False }; say A != A;
camelia niecza v24-109-g48a8de3: OUTPUT«Potential difficulties:␤ &infix:<==> is declared but not used at /tmp/dIKGt1pz8D line 1:␤------> ass A does Numeric { }; multi infix:<==>⏏(A $, A $) { False }; say A != A;␤␤Use of uninitialized value in numeric context␤…»
grondilu n: class A does Numeric { }; multi infix:<==>(A $, A $) { False }; say A.new != A.new; 04:09
camelia niecza v24-109-g48a8de3: OUTPUT«Potential difficulties:␤ &infix:<==> is declared but not used at /tmp/TOiCjjkPDl line 1:␤------> ass A does Numeric { }; multi infix:<==>⏏(A $, A $) { False }; say A.new != A.new␤␤Unhandled exception: Cannot use value lik…»
grondilu m: class A does Numeric { }; multi infix:<==>(A $, A $) { False }; say A.new != A.new; 04:10
camelia rakudo-moar 41d7f7: OUTPUT«(signal )»
04:10 sidewinder27 joined
grondilu for &infix:<!=>, isn't there a fallback to &infix:<==> in Numeric? 04:11
(line 301 in Numeric.pm) 04:12
04:17 raiph joined
xenoterracide class Logger does LoggerDebug does LoggerInfo (what's the shorter version of this?) 04:28
04:32 _thou joined, chenryn left 04:36 _thou left
xenoterracide gist.github.com/xenoterracide/7cf9...3a7226b929 is my use of requires correct here? also private works like this right? 04:36
grondilu thinks it's 'require' not 'requires' 04:40
std: require Foo; 04:41
camelia std 53b3ca6: OUTPUT«ok 00:01 121m␤»
grondilu std: requires Foo;
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Undeclared name:␤ 'Foo' used at line 1␤Undeclared routine:␤ 'requires' used at line 1␤Check failed␤FAILED 00:01 121m␤»
xenoterracide so no need for the quotes either? 04:44
std: require log(); 04:45
camelia std 53b3ca6: OUTPUT«ok 00:01 123m␤»
xenoterracide std: require log( Str:D $message ); 04:46
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Two terms in a row at /tmp/sXItFxOqCw line 1:␤------> require log( Str:D ⏏$message );␤ expecting any of:␤ infix or meta-infix␤ infixed function␤ statement modifier loop␤Parse failed␤FAILED 00:…»
04:49 chenryn joined
grondilu if you want to assert that your role "requires a method", you just stub it. 04:58
role Logger { method log {...} }
or something like that
'require' is a variant of 'use', it has little to do with role, mixins and stuff. It's about module importation. 05:00
S11:323
synopsebot Link: perlcabal.org/syn/S11.html#line_323
grondilu m: role A { method foo {...} }; class B does A { }; 05:02
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!===␤Method 'foo' must be implemented by B because it is required by a role␤»
grondilu ^see? rakudo tells you that B *must* implement foo
xenoterracide grondilu: oh right 05:03
05:06 raiph left 05:13 mr-foobar left 05:16 SamuraiJack joined 05:27 [Sno] joined 05:38 xenoterracide left, dayangkun left 05:49 xenoterracide joined
grondilu n: class A does Numeric {}; multi infix:<*>(A $, A $) { rand }; say A.new**2;' 05:50
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Bogus statement at /tmp/inUYvtSlYk line 1:␤------> ix:<*>(A $, A $) { rand }; say A.new**2;⏏'␤␤Parse failed␤␤»
grondilu std: class A does Numeric {}; multi infix:<*>(A $, A $) { rand }; say A.new**2;'
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unable to parse single quotes at /tmp/UALGYOzhoh line 1:␤------> ix:<*>(A $, A $) { rand }; say A.new**2;⏏'␤Couldn't find final "'"; gave up at /tmp/UALGYOzhoh line 1 (EOF):␤------> x:<*>(A $, A $) …»
05:50 gfldex joined
grondilu std: class A does Numeric {}; multi infix:<*>(A $, A $) { rand }; say A.new**2; 05:50
camelia std 53b3ca6: OUTPUT«ok 00:01 130m␤»
grondilu n: class A does Numeric {}; multi infix:<*>(A $, A $) { rand }; say A.new**2;
camelia niecza v24-109-g48a8de3: OUTPUT«Potential difficulties:␤ &infix:<*> is declared but not used at /tmp/mpEeGO8whv line 1:␤------> class A does Numeric {}; multi infix:<*>⏏(A $, A $) { rand }; say A.new**2;␤␤Unhandled exception: Cannot use value like A as …»
05:51 dolmen joined
grondilu p: class A does Numeric {}; multi infix:<*>(A $, A $) { rand }; say A.new**2; 05:52
camelia rakudo-parrot 41d7f7: OUTPUT«maximum recursion depth exceeded␤current instr.: 'print_exception' pc 142790 (src/gen/p-CORE.setting.pir:61593) (gen/parrot/CORE.setting:11857)␤called from Sub 'Numeric' pc 232640 (src/gen/p-CORE.setting.pir:96683) (gen/parrot/CORE.setting:1058)␤ca…»
grondilu j: class A does Numeric {}; multi infix:<*>(A $, A $) { rand }; say A.new**2; 05:53
camelia rakudo-jvm 41d7f7: OUTPUT«(timeout)»
05:57 chenryn left, anaeem1_ joined 06:00 anaeem1_ left 06:01 gfldex left 06:06 xenoterracide left 06:09 chenryn joined 06:11 mr-foobar joined 06:16 PZt joined 06:20 dolmen left 06:21 _thou joined 06:25 _thou left 06:27 domidumont joined 06:32 dayangkun joined 06:35 darutoko joined 06:37 sidewinder27 left, dayangkun left 06:46 aindilis` left, aindilis joined 06:48 sorear left 06:49 dayangkun joined 06:51 sorear joined 06:55 dayangkun left 06:57 kaleem joined, virtualsue joined 07:03 dayangkun joined 07:04 eternaleye left 07:05 jaffa4 joined
jaffa4 hi 07:05
I have question concerning native call. Are callbacks supported? 07:06
moritz I think so, yes
jaffa4: there's even a test file for callbacks in the NativeCall repo 07:07
jaffa4 The doc says they are going to be supporteed in the future
The doc says they are going to be supported in the future
moritz then probably the docs are out of date 07:09
jaffa4 zavolaj? 07:10
07:10 eternaleye joined
moritz zavolaj! 07:12
grondilu m: role A {}; class B does A {}; multi f(A $a, $x) { f($x, $a) }; multi f($x, B $b) { $x }; say f B, pi;' 07:13
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/JZFqesm2LN␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/JZFqesm2LN:1␤------> multi f($x, B $b) { $x }; say f B, pi;'⏏<EOL>␤ expectin…»
grondilu m: role A {}; class B does A {}; multi f(A $a, $x) { f($x, $a) }; multi f($x, B $b) { $x }; say f B, pi;
camelia rakudo-moar 41d7f7: OUTPUT«3.14159265358979␤»
grondilu role A { multi f(A $a, $x) is export { f($x, $a) } }; import A; class B does A {}; multi f($x, B $b) { $x }; say f B, pi;' 07:14
m: role A { multi f(A $a, $x) is export { f($x, $a) } }; import A; class B does A {}; multi f($x, B $b) { $x }; say f B, pi;'
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/j3jE0PGySP␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/j3jE0PGySP:1␤------> multi f($x, B $b) { $x }; say f B, pi;'⏏<EOL>␤ expectin…»
grondilu m: role A { multi f(A $a, $x) is export { f($x, $a) } }; import A; class B does A {}; multi f($x, B $b) { $x }; say f B, pi;
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/lZLRwlWz5r␤Calling 'f' will never work with argument types (B, Num)␤ Expected any of: ␤ :(Any $x, B $b)␤at /tmp/lZLRwlWz5r:1␤------> es A {}; multi f($x, B $b) { $x }; say [33…»
07:15 fdd__ joined
grondilu n: role A { multi f(A $a, $x) is export { f($x, $a) } }; import A; class B does A {}; multi f($x, B $b) { $x }; say f B, pi; 07:15
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Action method statement_control:import not yet implemented at /tmp/vMeb9Tv6vp line 1:␤------> $x) is export { f($x, $a) } }; import A⏏; class B does A {}; multi f($x, B $b) {␤␤Unhandled e…»
fdd__ hi could some one help me please I have no idea what I doing but I need to read a file x amount of pages long but I need to view page at a time. This code just makes it scroll off the screen in one go using putty justpaste.it/gxtw 07:18
thnaks
07:19 we is now known as u
grondilu what do you mean by "page"? 07:19
I guess it's a constant number of lines, right? 07:20
fdd__ well i dont want the whole file to scroll by in one go
i want to hit enter and read it by page at a time 07:21
so yes a number of lines would be good
as long as I can move on to the next number when i press enter
grondilu then make a counter and do a prompt every n lines
07:23 virtualsue left
fdd__ ok thank you that gives me some thing google 07:23
grondilu m: role A { multi f(A $a, $x) is export { $x } }; import A; class B does A {}; f B, pi;' 07:24
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/KWa__n0GBx␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/KWa__n0GBx:1␤------> ; import A; class B does A {}; f B, pi;'⏏<EOL>␤ expectin…»
grondilu damn it I always forget to remove the quote at the end
m: role A { multi f(A $a, $x) is export { $x } }; import A; class B does A {}; f B, pi;
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/X__lDE61n6␤Undeclared routine:␤ f used at line 1␤␤»
grondilu so, can't I import a multi??
m: role A { sub f(A $a, $x) is export { $x } }; import A; f A, pi;' 07:25
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/Qh9INtECEZ␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/Qh9INtECEZ:1␤------> is export { $x } }; import A; f A, pi;'⏏<EOL>␤ expectin…»
grondilu m: role A { sub f(A $a, $x) is export { $x } }; import A; f A, pi;
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/ufK4QLtBtz␤Undeclared routine:␤ f used at line 1␤␤»
grondilu m: class A { sub f(A $a, $x) is export { $x } }; import A; f A, pi; 07:26
camelia ( no output )
grondilu m: class A { sub f(A $a, $x) is export { $x } }; import A; say f A, pi;
camelia rakudo-moar 41d7f7: OUTPUT«3.14159265358979␤»
grondilu so importation works from a class but not from a role. Isn't that a bug?
m: import (role { sub duh is export { "duh" } }) <duh>; say duh; 07:31
camelia rakudo-moar 41d7f7: OUTPUT«===SORRY!=== Error while compiling /tmp/Texobmqx5t␤A symbol '&duh' has already been exported␤at /tmp/Texobmqx5t:1␤------> ␤»
07:40 jack_rabbit left 07:41 zakharyas joined
moritz grondilu: roles are rather special/complicated 07:49
does to roles actually being some sort of multi
and a role not doing much until it's punned or applied
grondilu ok 07:52
07:54 virtualsue joined 07:55 yeahnoob joined
yeahnoob Someone can help with it? Please see stackoverflow.com/q/25638528/3132217 07:57
Processing a large text file with Perl6 , but it's running very slow. And I don't know why... 07:58
grondilu well for one, the idiomatic way to read a file line by line is 'for lines open "file" {...}' 08:03
08:03 dayangkun left
grondilu $line.split(' ') is better written $line.words 08:03
08:04 u is now known as bcode
grondilu it does not make sense to write "{%dict{$p1}} {$p2}".words; You're concatenating and the split again?! 08:04
just write %dict{$p1} = (%dict{$p1}, $p2) 08:05
yeahnoob the code just make a dictionary from a text file 08:06
08:07 kjs_ joined
grondilu well: my %dict; for lines open "file" { my ($a, $b) = .words; %dict{$a} //= $b } 08:07
yeahnoob for example abc.txt = " a b \n c d\n a e\n a cc\n", it get a dictionary like "a-> b e cc" 08:08
grondilu oh 08:09
08:09 dayangkun joined
grondilu well: my %dict; for lines open "file" { my ($a, $b) = .words; push %dict{$a}, $b } 08:09
08:09 dakkar joined, _thou joined
yeahnoob OK. i would try your code modification:) 08:10
08:10 FROGGS_ is now known as FROGGS 08:14 _thou left, fdd__ left 08:17 kjs_ left
yeahnoob Still running two slow...even after the modification. 08:20
grondilu open("file").lines.classify(*.words[0])
also: perl6 --version
yeahnoob perl6 --version 08:21
This is perl6 version 2014.08 built on MoarVM version 2014.08
i have tried the 2014.04 / 2013.12 versions
the 2013.12 version coming from Ubuntu's package, and 2014.04 from the rakudo-star 08:22
08:24 domidumont left
yeahnoob confused... 08:25
grondilu 2014.08 on MoarVM is not far from the best there is. 08:26
08:27 anaeem1_ joined
sergot hi o/ 08:30
08:31 jaffa4 left
yeahnoob Perl 5 code deal with it about 200ms...I have tried in the past day. Really don't know why V6 running 20+sec, and still no finish. 08:37
tadzik Because it's slower in many cases 08:38
dalek rl6-most-wanted: 024137c | (Kamil Kułaga)++ | most-wanted/modules.md:
HTML documentation like htmlify.p6
 08:41
08:41 kjs_ joined 08:42 yeahnoob left 08:47 anaeem1_ left, anaeem1_ joined 08:50 ribasushi left 08:54 ribasushi joined 08:58 fhelmberger joined 08:59 mr-foobar left, yeahnoob joined 09:09 virtualsue left 09:11 pecastro joined
timotimo o/ 09:15
FROGGS o/ 09:17
yeahnoob o/
09:17 aindilis` joined 09:18 aindilis left 09:19 dolmen joined 09:21 domidumont joined 09:29 mr-foobar joined 09:33 denis_boyun joined 09:37 dayangkun left 09:42 yeahnoob left 09:50 SamuraiJack left 09:51 SamuraiJack joined, xinming_ joined 09:54 rindolf left, xinming left 09:58 _thou joined 10:03 _thou left, dayangkun joined 10:08 kjs_ left 10:15 dayangkun left 10:16 virtualsue joined 10:19 kjs_ joined, dolmen left
colomon wo\ah, 20+ new module fails overnight 10:20
just tested, NativeCall is failing 10:26
t/09-nativecast.t ...... 1/10
# Failed test 'casting int * to CArray[uint32] works'
# at t/09-nativecast.t line 12
# expected: '1 2 3'
# got: '1'
10:27 dayangkun joined
mathw_ at least nativecall failing is a good reason for lots of other things to fail... 10:27
10:27 mathw_ is now known as mathw
colomon mathw: NativeCall is pretty important these days! 10:28
mathw yes, still can't quite believe how easy it is too
I need some ideas, my local perl mongers group are finally having a technical talks meeting and they want me to talk about Perl 6. I'm struggling to find the right sort of way in which shows them things which are beneficial to them instead of just different 10:29
sergot afaik it's because of this: github.com/rakudo/rakudo/commit/bf...6560f7ecba 10:30
or...
is it? 10:31
colomon Migfht be because of the reason that patch was needed?
I'd be really surprised if that patch was the cause
10:31 kjs_ left
timotimo what colomon said 10:32
sergot irclog.perlgeek.de/perl6/2014-09-02#i_9283186 10:33
colomon: yes, I was wrong 10:34
colomon sergot: good catch
sergot github.com/rakudo/rakudo/commit/ef380d29f9 maybe? 10:35
10:38 bjz joined 10:44 chenryn left 10:45 kjs_ joined
carlin panda's projects.json still not updating? :( 10:47
tadzik isn't it?
oh, starman was being stupid for some reason 10:48
carlin hasn't updated since August 27, according to the HTTP headers
tadzik :o
I poked the indexer 10:49
10:50 kaleem left
tadzik carlin: should be better now 10:51
carlin tadzik++ thanks
tadzik no idea what caused it though, I'll monitor if it breaks again
10:51 FROGGS[mobile] joined 10:52 kaleem joined 11:03 oetiker left 11:06 chenryn joined 11:14 dayangkun left
jnthn
.oO( wo\ah is a pretty alarm-ing exclamation... )
 11:15
11:18 oetiker joined
jnthn mathw: What kind of audience is it aimed at? 11:19
mathw: Dunno if any of my recentish talks can be of some inspiration...
Does anybody know what happened to DESTROY in S12? 11:20
I went to familiarize myself with the API it wanted so I can help out nine++'s work, and found it seems to have vanished... :S
And the only recent history I could see was removing fossils related to it whihc suggests it somehow got removed long ago. 11:21
I guess somebody conflated "not timely" with "useless" :P 11:22
moritz jnthn: it looks to me that they were only specified as method DESTROYALL / submethod DESTROY
jnthn: and then (more or less accidentally) removed 11:23
jnthn Oh...
So accidental. OK.
That makes me worry a bit less )
11:24 pmurias joined
pmurias won't the moment DESTROY is called be very implementation dependent? 11:24
jnthn pmurias: Yes. 11:25
pmurias: And it may not be called at all.
But that's likely 'cus the VM is shutting down, in which case anyting you're holding on to is about to go away also :)
Or at least, anything you're planning to do last-gasp cleanup of with DESTROY 11:26
lizmat wonders it should be called DESTROY 11:31
as the semantics are different from P5 11:32
maybe CLEANUP would be better ?
nine What's the difference in semantics? 11:33
jnthn I think cleanup has the wrong nuance 11:37
It's more like an event, really 11:38
COLLECT or so
I dunno :)
I do know it's dinner time, though. :)
&
11:40 pmurias left
lizmat REAP ? 11:40
moritz REAPer, grim 11:43
carlin class Reaper is Grim does Reap 11:44
11:46 labster left 11:47 _thou joined, labster joined 11:52 _thou left
flussence I see this DESTROY thing as more of an advisory hook into the GC - if something's important enough that it *needs* to always run at a specific time, you'd just write it as normal code. 12:00
12:04 yeahnoob joined
flussence otoh, a separate CLEANUP might be useful to have, so the GC can politely ask things like caching classes to trim themselves a bit... 12:04
yeahnoob Need Help Please..stackoverflow.com/q/25638528/3132217 12:09
Processing a large text file using Perl6. And it's running Very Very Slow.
flussence what does "perl6 -e 'say $*VM'" print on your system? 12:11
yeahnoob But it only cost less than 200ms with Perl 5, almost the same code.
i have tried JVM/MoarVM/Parrot 12:12
now its ouput is jvm(1.7)
output 12:13
ing a large text file using Perl6. And it's running Very Very Slow.
perl6 -e 'say $*VM'
jvm (1.7)
flussence what about perl6-m -e 'say $*VM' ?
yeahnoob i don't build MoarVM on my PC. 12:14
So no perl6-m now. But i have tried MoarVM with perl6-m 12:15
FROGGS[mobile] yeahnoob: .get is very very slow, a grammar might be better there
colomon yeahnoob: %dict{$p1} = "{%dict{$p1}} {$p2}".words; strikes me as a VERY inefficient way to do that 12:16
yeahnoob i know its not a efficient code
i have modified it with some suggestion.
But no better result 12:17
Such as:
use v6;
my %dict;
for open('wordpairs.txt',:r).lines {
my ($key, $words) = .words;
push %dict{$key}, $words;
}
## Test
Or: 12:18
...
use v6;
my %dict;
my $a;
my $b;
for lines open "wordpairs.txt" {
($a, $b) = .words;
push %dict{$a}, $b;
}
## Test
FROGGS[mobile] dont paste here
please use a paste service
colomon yeahnoob: can you upload the wordpairs file somewhere so we can try wit hthat? 12:19
yeahnoob Sorry
mathw jnthn: competent, active Perl 5 developers. Most of them seem to do nothing but websites in it, so I'm not sure how enthused they'd be by your brilliant ones about asynchronous programming.
yeahnoob I host the whole thing at my github
FROGGS[mobile] I can reply to your stackoverflow question later when I'm at my pc... then I can propose a grammar
yeahnoob github.com/yeahnoob/perl6-perf
colomon yeahnoob: brill 12:20
yeahnoob Pre-Thanks for any answer:)
mathw jnthn: I might throw in something about NativeCall though, because last I checked you couldn't do it that easily in 5... 12:21
12:21 chenryn left
colomon yeahnoob: it's entirely possible the answer (at the moment) is just rakudo is slow. but we can at least learn more about it…. 12:23
12:23 dwarring left
yeahnoob Yes, it's maybe the rakudo's. The code work no problem with a small test.txt file 12:25
colomon yeahnoob: it takes almost two minutes to run on my Mac 12:28
profiling now 12:30
yeahnoob en..it's not a reasonable time cost, which is 160ms with the same function Perl 5 test code on my PC. 12:32
colomon Hmm, just got it down to 51 seconds. 12:36
FROGGS[mobile] colomon: I bet I get it down to <5s 12:37
:p 12:38
colomon FROGGS[mobile]: with a grammar? 12:39
12:40 FROGGS[mobile] left 12:43 dayangkun joined 12:46 FROGGS[mobile] joined
FROGGS[mobile] colomon: aye 12:47
colomon FROGGS[mobile]: Trying now...
I dunno, my first stab at a grammar seems painfully slow 12:53
oooo, maybe not 12:54
12:55 PZt left
FROGGS[mobile] it should be a pretty easy grammar (with actions) 12:56
I mean, lines and words... right? 12:57
colomon FROGGS[mobile]: pretty sure it wasn't the grammar that was painfully slow, it was the call to Match.gist
My grammar can succesfully scan the file in 21 seconds. 12:58
FROGGS[mobile] ohh, true :o)
hmmm, how big is the file?
colomon 1.5 meg 12:59
no, waitr, that was the profile output!
1,474,282 bytes
FROGGS[mobile] how many lines?
colomon 138,523 13:00
FROGGS[mobile] because we can 20k lines Perl 6 with a much more complex grammar in 35s on my box
colomon okay, that was 46 seconds to parse and count
13:00 _thou joined
colomon should maybe start $working at some point.... 13:02
FROGGS[mobile] please put it in a gist, I try to improve it in abit 13:03
13:03 telex left
colomon heh, 97% of the running time is the my <word-pair> rule. 13:04
13:04 telex joined, kaleem left
colomon www.harmonyware.com/files/profile-1....0399.html 13:05
half the running time is in prefix:<~>
FROGGS[mobile] where is the code? :o) 13:06
13:06 kaleem joined
colomon WHAT 13:07
13:07 guru joined, guru is now known as Guest39660 13:08 Guest39660 is now known as ajr_
colomon one-third of the running time is going to join?! 13:09
FROGGS[mobile] can I see the grammar?
colomon got it down to 37 seconds
gist.github.com/colomon/ce21bb3f199637814f00 13:10
profiling that one now
13:11 salv0 left
colomon profile: www.harmonyware.com/files/profile-1...50828.html 13:11
hmmm 13:13
13:13 Akagi201 left
colomon 33 seconds 13:15
dalek kudo/pod-fixes: 6ac67de | (Rob Hoelz)++ | src/Perl6/Grammar.nqp:
Convert two more END-related things in POD parsing

It's finish now
kudo/pod-fixes: 74fd61d | (Rob Hoelz)++ | src/Perl6/Grammar.nqp:
Fix paragraph and abbreviated comments

Except for table comments, that is
13:15 Akagi201 joined
colomon new one gist.github.com/colomon/c2d2c1940b92d8d9df01, profile www.harmonyware.com/files/profile-1...18343.html 13:16
colomon is trying to build moar jit to see what that does to the timings 13:22
13:23 raiph joined 13:24 kaleem left, prevost joined
colomon hmmm, JIT doesn't help at all 13:25
13:25 rindolf joined
timotimo ah, caphash is the top time-user there. some time soon MATCH will be rewritten in rakudo to include the stuff CAPHASH used to do and that'll make stuff faster, because it won't build an intermediate data structure just to throw it away again 13:27
colomon profile of the last with JIT is www.harmonyware.com/files/profile-1...51175.html 13:28
timotimo ah, 73.8% jitted frames isn't terribly bad
5% deoptimizations seems to be quite a lot
huh, using the jit makes it slower, but the time spent doing dynamic optimization is basically zero ... 13:29
13:32 smash_ is now known as smash
timotimo doesn't FROGGS[mobile] want to implement the proper MATCH for rakudo? :P 13:32
13:35 chenryn joined
FROGGS[mobile] timotimo: not that I know :o) 13:36
timotimo do i remember correctly that you did that for NQP already?
FROGGS[mobile] timotimo: also, I am implementing C globals just fo you already :p 13:37
timotimo ooooh!
for me? :D
FROGGS[mobile] jnthn did
surprise /o/
timotimo i'm not sure if i even need c globals for any project i have going on at the moment 13:38
FROGGS[mobile] you said that yoz'd need that for OpenGL IIRC
timotimo oh
i hope you won't expect me to build an opengl binding immediately now :) 13:39
FROGGS[mobile] much disappointment *g*
timotimo what i needed for opengl was casting the return value of functions to invocables
FROGGS[mobile] ahh, true
timotimo and having mutable bufs would be very interesting for opengl 13:40
FROGGS[mobile] that should be doable too, since we have nativecast and callbacks
Buf is not mutable?
timotimo buf or blob or whichever 13:41
not yet, needs the native array thingie jnthn has somewhere near the top on his todo list
FROGGS[mobile] we have that now me thinks
ahh
I see
[Coke] mathw: I did a presentation on six for a bunch of fivers a few months ago, you can reuse that if you like.
13:43 mr-foobar left 13:46 yoleaux joined, ChanServ sets mode: +v yoleaux
[Coke] (though you'd need to familiarize yourself with the slides before using it; I was trying to hit all the highlights for "neat stuff fivers mights want". (and didn't have enough time to touch grammars, unfortunately) 13:48
timotimo oh, that's the one where i manually colored syntax in the slides, right? :D
[Coke] m: sub postfix:<!> { [*] 2..$^a }; say 0!, 1!, 2! 13:49
camelia rakudo-moar 41d7f7: OUTPUT«112␤»
[Coke] timotimo++ yes, I think so!
timotimo that was fun
except it was stupid work ;)
jnthn oh wow, timotimo provides Slide Highlighting as a Service? :D 13:51
timotimo i really want to find a better way to do that for the next time.
jnthn :) 13:52
timotimo it was also a very last-minute decision
14:02 slavik left, dalek left 14:03 dalek joined, ChanServ sets mode: +v dalek
jnthn Yay. Startup time is lower than last time I was looking at it... 14:06
14:07 salv0 joined 14:13 Ven joined
Ven o/, #perl6 14:13
timotimo: slides.com highlighted my p6 correctly
timotimo oh, neato
Ven timotimo: i.e. slides.com/namidoc/road2p6#/9 14:14
[coke]: I'd be very interested in your material too :-) 14:15
timotimo that presentation was on google docs
14:15 slavik joined
Ven timotimo: which one? [coke]'s one? 14:16
[Coke] ven, mathw : docs.google.com/presentation/d/1ls...sp=sharing
timotimo yes
Ven thanks! [coke]++
[Coke] Happy to to allow folks edit access if they're up for it. (timotimo, looks like we ended up with a white copy and a black copy) 14:17
timotimo ah
yes, i think so
[Coke] I should go through and delete one of those. 14:19
(and make sure everything is up to date in teh other)
Ven well, cognome++ asked me to talk about 6 a bit too, but I really don't want to explain the internals 14:20
"who cares" :-)
[Coke]: no hard feeling if I skip the opening slides :p?
[Coke] Ven: of course not, share and enjoy. 14:22
if you have improvements, let me know. :)
colomon is very glad he is listening to some nice fiddling and so not tempted to sing the share and enjoy song
[Coke] definitely need updated moar benchmark shots! 14:24
I see anonymous ibex and anonymous liger are reading the doc. :P
Ven I'll read through it later. 14:29
I also don't really know what 5ers want, since I'm not a 5er, but I'm gonna use that :p
moritz perl 5 programmer, presumably 14:30
Ven moritz: wut? 14:31
timotimo i think he misread your "want" as "are"
Ven Oh, I think you read my "want" as "are"
timotimo %)
Ven timotimo++ # 2spooky
moritz Ven: sorry for the confusion 14:32
dalek kudo/nom: 358c749 | jonathan++ | src/Perl6/Metamodel/BOOTSTRAP.nqp:
Avoid around 6,000 allocations at startup.

These chewed through around a 16th of a nursery.		 14:35
jnthn Also is a small win for the initial run through every multi dispatch.
14:35 dj_goku_ left
timotimo huh. 14:41
interesting
14:42 mr-foobar joined 14:43 anaeem1_ left
timotimo Ven: "rakudo-js, unfortunately abandoned" - not actually true, but you may know that by now already 14:43
Ven timotimo: it was when I wrote this 14:44
this is the talk I gave at the fpw2014
timotimo yeah, i saw the "3 month ago"
14:45 itz joined
timotimo (i'm using google translate to read your slides 14:45
)
sub postfix ($ N) {<!> [*] 2 .. $ N ; } <- that's very well translated, google! 14:46
m)
14:46 itz_ left
Ven timotimo: not sure what that gives %) 14:48
timotimo: think about woolfy++ and lizmat++, they sat through it without translators %)
14:49 pmurias joined
hoelzro morning #perl6 14:49
pmurias masak: I have thought a bit about your take an example and refactor it into a grammar thing 14:50
masak: what I think would be most usefull instead of having it as standalone web things would be to integrate it into an editor 14:51
masak pmurias: yes, quite probably
yoleaux 1 Sep 2014 19:45Z <Ven_> masak: you might think my "Map[K, V]" naming is unfortunate, but we can't type hashes.
2 Sep 2014 00:28Z <jepeway> masak: hey, I'm poking at RT 122319; would you have time to peek at clog around irclog.perlgeek.de/perl6/2014-09-01#i_9282046?
Ven actually, maybe my Foo %Bar{Baz} works but no clue what it does 14:52
pmurias masak: we could even do something like automatically find a rule that matches a given block of selected text 14:53
[Coke] someone was looking for things to do recently: here's a boring list of work that is a good way for folks to contribute: gist.github.com/coke/ac078396e8f216b83e9a 14:54
pmurias masak: so you could select {"foo":123} and have it "refactored" into <json>
[Coke] happy to have that go somewhere more easily editable. 14:55
timotimo Ven: i live so close to the french border, maybe i'll swallow my pride and learn some french at some point ... 14:56
Ven timotimo: well, if you want to drop by at some point, n'hésite pas %)
timotimo is nancy very far from you? that's the closest big town i think
Ven timotimo: almost 400km between those, yes – I'm in paris. 14:57
14:58 itz_ joined
Ven *(that's where the french mongueurs are) 14:58
timotimo mhm
that's pretty far :\
Ven indeed 14:59
15:00 itz left
pmurias masak: the biggest problem with integrating it into an editor is choosing the editor we want to integrate into ;) 15:01
masak kinds of doubts that's the biggest problem :)
kind* 15:02
Ven: `my Foo %Bar{Baz}` works, AFAIK. generally when I have no clue what something does, I read the spec.
15:03 itz joined
Ven masak: I don't know, I just read that syntax 30min ago :P 15:03
15:04 itz_ left
masak anyway, your flippant ignorance of the actual way to write a typed hash does not inspire in me confidence about your choice of the name `Map`. 15:04
nwc10 jnthn: Rakudo startup seems to be 3% less than "last time I measured it" 15:08
15:08 ajr_ left, guru joined, guru is now known as Guest31904 15:09 Guest31904 is now known as ajr_ 15:10 chenryn left 15:11 denis_boyun left
Ven masak: what was that about? 15:11
15:12 kaleem joined
Ven masak: as I said - this is barely pseudo code anyway. 15:12
no apparently-flippant compiler is capable of handle `has`'s `where`s 15:13
masak Rakudo parses them but ignores them at present.
as deficiencies go, there are worse :)
Ven handling* 15:14
15:14 chenryn joined
jnthn m: my Int %h; %h<a> = 'omg' 15:14
camelia rakudo-moar 41d7f7: OUTPUT«Type check failed in assignment to '%h'; expected 'Int' but got 'Str'␤ in method assign_key at src/gen/m-CORE.setting:9716␤ in sub postcircumfix:<{ }> at src/gen/m-CORE.setting:2799␤ in block <unit> at /tmp/V93rnBru1J:1␤␤»
jnthn Doesn't look very ignored to me ;-)
masak that part works.
Ven m: my %h{Str}; %h{1} = 5; 15:15
camelia rakudo-moar 41d7f7: OUTPUT«cannot stringify this␤ in method assign_key at src/gen/m-CORE.setting:10054␤ in sub postcircumfix:<{ }> at src/gen/m-CORE.setting:2799␤ in block <unit> at /tmp/Ngws364YOd:1␤␤»
Ven well, at least it's not as terrible a syntax as go's.
masak .oO( hammer time! )
jnthn nwc10: I'll take 3% :) 15:16
nwc10 me too 15:17
15:17 Ven left 15:22 firefish5000 left
[Coke] .words is slooow. 15:33
15:37 PhilH joined
colomon [Coke]: yeeeeessssssss 15:40
[Coke] looking at the SE code that came through here a day or 2 ago 15:41
stackoverflow.com/questions/2563852...ow-2014-09
[Coke] updates to latest moar JIC. 15:42
colomon [Coke]: irclog.perlgeek.de/perl6/2014-09-03#i_9290176 and beyond
irclog.perlgeek.de/perl6/2014-09-03#i_9290559 is good. ;) 15:43
but not great
jnthn 'night o/ 15:45
15:46 virtualsue left
timotimo aaw, jnthn is already gone again :S 15:47
15:48 itz_ joined
tadzik which timezone is jnthn in? :o 15:48
timotimo +7 or +8 i think
15:50 zakharyas left, itz left 15:52 itz_ left, fhelmberger left 15:53 itz joined
yeahnoob +8 zone, me too. Sleeping time now. 15:56
15:57 yeahnoob left
[Coke] colomon: looks like 23% of the time is spent in reify (... using my local unpublished version of his code which uses .IO.lines and then calls ".words" 138K times. 15:59
16:02 kjs_ left
colomon [Coke]: pretty sure using .words is a bad idea for performance 16:03
16:04 mattp__ joined 16:05 prevost_ joined
colomon doesn't mean to say we shouldn't work on optimizing .words performance, mind you 16:05
16:05 hoelzro_ joined 16:06 mads-_ joined, [Coke]_ joined, Juerd_ joined 16:07 arnsholt_ joined
masak 'night, #perl6 16:07
moritz \o
timotimo gnite masak!
16:07 yukko_ joined 16:08 clkaoud left, ClarusCogitatio_ left, prevost left, hoelzro left, mads- left, Celelibi left, ggherdov left, arnsholt left, Juerd left, mtj_ left, [Coke] left, skarn left, mattp_ left, Juerd_ is now known as Juerd, yogan joined, camelia joined 16:09 mtj_- joined, hoelzro_ is now known as hoelzro, ClarusCogitatio joined, clkaoud_ joined, Celelibi joined 16:10 ChanServ sets mode: +v camelia, vukcrni joined 16:11 skarn joined, skarn is now known as Guest10538 16:12 ggherdov joined 16:17 pecastro left 16:27 MilkmanDan left 16:29 raiph left, raiph joined 16:32 MilkmanDan joined 16:33 spider-mario joined, spider-mario left 16:35 spider-mario joined 16:38 dj_goku joined 16:41 SamuraiJack left, SamuraiJack joined 16:42 dakkar left 16:43 ajr_ left 16:45 dayangkun left 16:47 guru joined, guru is now known as ajr_, pmurias left 16:49 raiph left 16:50 SamuraiJack left 16:57 mads-_ is now known as mads-
FROGGS colomon: I was just able to make it about 20% faster: github.com/yeahnoob/perl6-perf/pull/1 16:57
PerlJam Is RT #118755 still valid? (i.e. should first return the mutable containers?) 16:59
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=118755
17:04 anaeem1_ joined 17:05 chenryn left
TimToady on a 3300 word file, .words creates 16000 hashes 17:07
well, BOOTHashes
9900 BOOTIters
FROGGS we are quite fast at allocating all these things /o\ 17:08
TimToady 10000 BOOTCodes, 39000 BOOTInts, and 86000 BOOTArrays
FROGGS ohh dear 17:09
TimToady well, .words is defined as .comb(/\S+/) but it's stupid to implement it that way 17:10
FROGGS I know of somebody who has a commit bit *g* 17:11
TimToady just needs to complain a little first
FROGGS sure
I'd like to fix rakudo's MATCH which should help here too, the problem is that I don't know what to fix :o( 17:12
I think jnthn++ said that we could get rid of three allocations
TimToady obviously we're paying a very high price for the abstraction here 17:13
17:14 bjz left
FROGGS well, the machines have to pay that price... not the devs :o) 17:14
17:14 bjz joined
TimToady we pay a price in time 17:14
FROGGS intel should give us some money, because we indirectly sell CPUs :o) 17:15
TimToady then they should be paying us to *not* work :P 17:17
since we're trying to sell fewer CPUs
dalek kudo/nom: 07ef21e | duff++ | src/core/Any.pm:
Fixes RT #118755
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=118755
17:28 jack_rabbit joined
vendethiel "computational expressions" is what happened when f# realized it wanted haskell's do-notation but didn't have hindley-milner. 17:30
at least it's cute...
17:33 ajr_ left 17:34 zakharyas joined
dalek kudo/nom: 03b9a09 | moritz++ | src/Perl6/Metamodel/BOOTSTRAP.nqp:
[Bootstrap] natively-type a variable
 17:39
kudo/nom: 4d18460 | moritz++ | src/Perl6/Metamodel/BOOTSTRAP.nqp:
[Bootstrap] remove an unused variable
kudo/nom: 962a993 | moritz++ | src/Perl6/Actions.nqp:
Cache a constant expressions
kudo/nom: 8ef5c55 | moritz++ | src/Perl6/Grammar.nqp:
[grammar] use some more native types
17:41 cognome left
lizmat moritz: maybe those fixed need a Moar or NQP revision bump?? 17:43
*fixes
moritz lizmat: what makes you think so?
lizmat Stage parse : moar(47427,0x7fff74dba310) malloc: *** error for object 0x7f85024f8290: pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug
make: *** [CORE.setting.moarvm] Abort trap: 6
is what makes me suspect :-) 17:44
moritz lizmat: I have tested it with moarvm master, but there's nothing in there the looks like it needs a new MoarVM
lizmat just before your commits, I had just built without any problems
17:44 cognome joined
lizmat retrying with moar master 17:45
nope, same failure 17:48
retrying with nqp=master
17:49 kjs_ joined 17:50 itz left
lizmat nope, still same problem :-( 17:52
going back in commits now
nine (git bisect)++
lizmat moritz: seems 8ef5c5564c is the reason 17:53
if I remove all the native str typing, it builds 17:56
further refining now
17:56 itz joined
woolfy Hi all. Remark from a friend from outside Perl 6: "I am interested in Perl6 but until there is an easy way to install it on a Windows machine, my interest is expressed from a safe distance only." 18:01
So I went looking for a MSi and found this: rakudo.org/downloads/star/
vendethiel woolfy: a new pal :P !
a windows user
woolfy There's a MSI file for Rakudo Star 2014.03. Not yet for 2014.08.
Any idea when?
vendethiel: actually, this is regarding a Perl-friend of mine that I know for 10+ years... 18:02
lizmat moritz: is it legal to do a eq 'f' on a str in nqp ?
woolfy And next question: how "easy" (I know, many definitions of what easy is) is it to install Rakudo Star with this MSI file? 18:03
vendethiel woolfy: it's a one-click installer
woolfy (I am not on Windows, so I cannot try it myself and I really would hate myself going back to Windows again)
vendethiel it's very well packaged.
it's just unsuitable for prod because it's so slow 18:04
(it's on parrot :P)
lizmat moritz: it appears that this line is causing the problem:
woolfy vendethiel: Really that simple? So, at the end, my friend would have what, Parrot and MoarVM and JVM and NQP and Rakudo and could just go ahead?
lizmat my str $last := nqp::substr($str, nqp::chars($check) - 1, 1);
woolfy Oh. Just parrot.
vendethiel woolfy: well, 2014.03 only had parrot
lizmat it feels like it is causing a problem on the next line (3449)
$last eq ')' || $last eq '}' || $last eq ']' || $last eq '>'
woolfy Hm. I'm not going to advice him to work on Parrot. No way. 18:05
vendethiel because every rakudo star was only-parrot, not just the windows one
so, technically, this new R*for 2014.08 could definitely use moarvm :-)
woolfy vendethiel: OK, so it would be in the 2014.08 release?
vendethiel I hope so! But no reason not to, we know moar works on windows -- jnthn++'s on it
woolfy So, back to my first question... Any idea when? (I mean: "when is rakudo-star-2014.08.msi going to be available?"). Days, weeks, or wait for 2014.09 or 2014.10 or so... ballpark figure would be nice. 18:07
18:08 BizarreCake joined
lizmat I think the main problem is that there are few people with MSI-building fu :-( 18:09
18:10 MilkmanDan left
woolfy fu? 18:10
nine woolfy: knowledge 18:11
18:11 SamuraiJack joined
woolfy Ah. That went way over my head. 18:11
18:12 kaleem left, molaf_ joined
FROGGS woolfy: "when is rakudo-star-2014.08.msi going to be available?" - hours 18:13
lizmat Suffix[edit]
-fu
(slang) Expertise; mastery.
My Google-fu is weak!
Aragorn uses Ranger-fu to figure out that Sam and Frodo have taken a boat.
woolfy: from en.wiktionary.org/wiki/-fu 18:14
timotimo FROGGS: pretty awesome! :)
18:14 MilkmanDan joined
FROGGS timotimo: I also did a 2014.04-moar.msi and -parrot.msi 18:14
woolfy Yeah yeah... as said, it went way over my head. First time that I saw somebody use this lingo outside the "kung fu" area of expertise. 18:15
18:15 zakharyas left
FROGGS but I did not upload it because I did not have a single tester... 18:15
huf "expert in kung"
i wonder what that is..
FROGGS but now I just upload anyway
woolfy FROGGS++
18:16 molaf__ left
dalek kudo/nom: 0be2374 | (Elizabeth Mattijsen)++ | src/Perl6/Grammar.nqp:
Unbust on (at least) OSX, caused by 8ef5c5564c

No idea why this makes a difference, but suspect the eq on the next line to have some influence in this. Error seen was:
Stage start : 0.000 Stage parse : moar(51746,0x7fff74dba310) malloc: *** error for object 0x7faa4c7aad90: pointer being freed was not allocated
  *** set a breakpoint in malloc_error_break to debug
make: *** [CORE.setting.moarvm] Abort trap: 6		 18:16
18:17 SamuraiJack left
woolfy OK, FROGGS, thanks. I just mentioned it to my Perl5-friend (who I leave to be anonymous for now). I guess he did not expect this. I hope he likes it. 18:18
FROGGS I already started to make the msi's, so another funny coincident :o)
PerlJam FROGGS++ the best way to get testers is to make a release and tell everyone where to get it :) 18:19
FROGGS *g*
yeah, that might work too :o)
18:21 molaf_ left
moritz lizmat++ # fixing up my mess 18:23
lizmat moritz: is that supposed to work in nqp? eq on a native str ?
FROGGS nqp-m: my str 'foo'; say($foo eq 'foo')
camelia nqp-moarvm: OUTPUT«Confused at line 2, near "my str 'fo"␤ at gen/moar/stage2/NQPHLL.nqp:481 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPHLL.moarvm:panic:105)␤ from gen/moar/stage2/NQP.nqp:913 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/nqp.moarvm:comp_unit:872)␤ fro…»
FROGGS nqp-m: my 'foo'; say($foo eq 'foo') 18:24
camelia nqp-moarvm: OUTPUT«Confused at line 2, near "my 'foo'; "␤ at gen/moar/stage2/NQPHLL.nqp:481 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPHLL.moarvm:panic:105)␤ from gen/moar/stage2/NQP.nqp:913 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/nqp.moarvm:comp_unit:872)␤ fro…»
FROGGS nqp-m: my str $foo := 'foo'; say($foo eq 'foo')
camelia nqp-moarvm: OUTPUT«1␤»
FROGGS yes
moritz lizmat: I dearly hope so :-)
lizmat hmmm... well... then I have no idea...
moritz Cursor.INTERPOLATE has lots of vars that could become native
lizmat but undoing the "str" typing unbusted the build for me
moritz but when I tried to make it that way, core compilation hangs 18:25
lizmat so, definitely something fishy goin on there
vendethiel reads some "go sucks post" again and chuckles
FROGGS++ thanks for taking care of us windows users
well, I don't really code on windows anymore, but I know I have an incredible luck 18:26
dalek kudo/nom: c524e7d | (Elizabeth Mattijsen)++ | src/core/Parcel.pm:
Simplify Parcel.WHICH, now that we don't flatten
ar: 0644ed9 | (Tobias Leich)++ | / (2 files):
pass --backends option to rakudo's configure
 18:30
18:32 gfldex joined 18:35 telex left 18:36 guru joined, guru is now known as ajr_, telex joined
vendethiel can I encode church's numbers in P6's type system? 18:38
probably not :(
TimToady m: say set(set()) 18:39
camelia rakudo-moar 8ef5c5: OUTPUT«set(set())␤»
TimToady interesting datapoint: I rewrite .comb(/S+/,$limit) in very raw nqp using nqp::findcclass and nqp::findnotcclass, and it runs twice as slow... 18:48
FROGGS hehe 18:49
TimToady despite the fact that .words is entirely jitted
FROGGS the funny thing here was that my window size made the line break after the 'as'... so the 'slow' was kinda unexpected *g* 18:50
TimToady would be lovely to have a statement-by-statement profiler...
PerlJam TimToady: Seems like it wasn't so stupid to implement .words with .comb(/S+/) after all ;)
TimToady thinks findcclass must be doing something stupid now, like reloading the Unicode database each time 18:52
[Coke]_ TimToady: you using moar, I assume? 18:53
PerlJam If you make findcclass and findnotcclass faster, that will have broad, positive impact 18:54
18:54 [Coke]_ is now known as [Coke]
PerlJam crosses his fingers that TimToady's changes have a happy ending. 18:54
18:57 sjohnson left
diakopter istr writing findcclass 19:00
19:00 darutoko left, itz left, itz joined
TimToady oh, gee, the first thing findcclass does is length = MVM_string_graphs(tc, s); 19:03
diakopter I thought graphs was cached 19:04
on the string
object
TimToady maybe it is, though strings have changed somewhat since then
timotimo PerlJam: did you nqp::unbox_s the target string once? 19:05
if the string is utf8, it may have to scan through the string to get to the $start-index (byte vs codepoint) 19:06
diakopter it never stores as utf8
19:07 rindolf left
timotimo oh, right, we have 32bit storage for codepoints anyway 19:07
19:07 rindolf joined
TimToady findcclass does appear to cache the grapheme count, but it does create a new iterator every time; dunno whether that's expensive or not 19:16
diakopter I'd guess so 19:17
TimToady the regex engine at least knows it's gonna look for the same thing again, so it naturally saves the previous cursor to start from; maybe that's the difference... 19:18
19:19 _thou left
TimToady I suppose we could cache the iterator, as long as the string doesn't change, and start from where we left off before, if appropriate 19:19
or add some kind of simplecomb instruction that knows it's going to alternate on some boolean test 19:20
diakopter that wouldn't be threadsafe, unless you laboriously made it so 19:22
(the cached iterator)
PerlJam depends on where the cache lives :)
diakopter true
19:23 zakharyas joined
timotimo right, the threadcontext could hold it and it'd be fine 19:24
TimToady well, this is probably too micro of an optimization to be looking at it right now 19:25
FROGGS woolfy: 19:26
rakudo.org/downloads/star/rakudo-st...8-moar.msi
rakudo.org/downloads/star/rakudo-st...parrot.msi
vendethiel woolfy: FROGGS++ just updated the msi :-)
uploaded*
TimToady especially since it's lunch time
vendethiel sorry FROGGS, I'm too fast for you ;-)
FROGGS O.o
timotimo in the compilation process of rakudo-parrot, there was one file where updating the line position cache took much longer because the file was utf8 and it was re-doing the iterator from the beginning each and every time
FROGGS how on earth...
vendethiel did I know you pushed those?
FROGGS vendethiel: yeah
vendethiel I wonder :-)
timotimo he sat on the f5 button
FROGGS or he just pulled the internet and saw it in the diff 19:27
woolfy :-)
timotimo %)
vendethiel actually I just saw the github notifications
cause I see those instantly
19:28 itz_ joined
timotimo .. github notifications? 19:28
vendethiel timotimo: FROGGS++ posted something on github about that *before* noticing woolfy++ here.
FROGGS I replied to a panda issue 19:29
vendethiel and I basically saw it insta
(I spend far too much time on github :/)
FROGGS but we <3 github! 19:30
I'd work for them if they'd need Perl ppl
19:30 itz left
vendethiel FROGGS: well, I do too. But still, I reached 531 days of streak -- that's far too much time spent on that 19:30
woolfy OK, I told my friend and he is happy. Now he needs some tuits (well, hours) to start playing with it. But I am sure (meaning "I really really really hope) he is not the only one that will use it. 19:31
timotimo FROGGS: after installing it, the windows start menu shortcut wouldn't work 19:32
vendethiel woolfy: trust me, he's not. even if only with the people I introduced to p6 ;-)
and I introduced *a lot* of them and they're mostly using windows
timotimo it'd say "missing file" or "broken shortcut" or something and tells me to look for "perl6-m.exe", but there's only perl6-m.bat
FROGGS timotimo: there is one? O.o 19:33
woolfy I am NOT on Windows so there is no way I am going to try that!
timotimo yes.
FROGGS timotimo: I see, that seems to be a relict of parrot-ish days, when we had .exe files only 19:34
timotimo :)
would be neat if you fixed it anyway :)
FROGGS yeah...
19:35 itz_ left, itz joined 19:37 domidumont left
moritz 19:40
19:40 BizarreCake left, beastd joined
moritz t/spec/integration/99problems-51-to-60.t segfaults for me 19:41
Ulti one issue with the MSI is it's not signed so modern Windows platforms complain a lot 19:42
vendethiel nothing's signed, these days :/
Ulti well that's the first time I have ever seen both a browser warning and a full windows 8.1 prompt refusing to run an installer 19:43
FROGGS Ulti: if someone would help me to accomplish that... 19:44
Ulti yeah if I could I would, I have the means but not the know how to do rakudo builds on windows
TimToady okay, now I have a .words that is six times faster than the .comb :)
diakopter I think pmichaud has a keypair for it or somethingg
Ulti I use this machine just for games ;P
moritz I'll happily sign any Star .msi
... with GPG :)
FROGGS :o)
PerlJam TimToady++ 19:45
Ulti FROGGS++ well it runs >:3 19:46
on windows 8.1
FROGGS built on WinXP :P
Ulti using panda wasn't so easy fatal: could not create work tree dir 'C:\Users\Matthew\.panda-work\1409773602_1'.: No such file or directory 19:47
FROGGS yes, that's github.com/tadzik/panda/issues/98 19:48
Ulti oh ok
FROGGS that seems easy to fix... dunno why it isnt already
dalek line-Perl5: e8a8e57 | nine++ | / (2 files):
Fix all sorts of strange errors on exit caused by double free'd string.

SvPV returns a pointer to the string buffer in an SV. NativeCall seems to free strings passed to it, so both Perl 5 and NativeCall were freeing the same string leading to different error messages.
Pass a copy of the string instead.		 19:49
nine YES!
This one took hours to track down...
FROGGS I can imagine :/
nine++
timotimo there's a trait for string management
nine timotimo: is it documented somewhere? 19:51
timotimo not sure 19:53
github.com/jnthn/zavolaj/blob/mast...l.pm6#L323
19:55 lastofthe joined
nine timotimo: thanks. 19:56
hoelzro timotimo: how are you even supposed to use that? 19:59
it doesn't look like a trait to me
dalek ar: cbe310b | (Tobias Leich)++ | tools/build/star-product.wxs:
install windows start menu shortcut compatible with moar
 20:02
20:02 prevost_ left
FROGGS timotimo: I updated the msi so that the shortcut works 20:02
20:02 colomon left
timotimo cool, thanks 20:03
why you no timotimo++ in your commit message? :P
FROGGS hehe, sorry timotimo++ :P
vendethiel FROGGS++ # thanks! 20:13
20:14 opendevnet joined 20:15 gtodd joined 20:18 opendevnet left 20:20 spider-mario left 20:22 denis_boyun joined
tadzik nine: hmm, maybe you can bump the refcount and avoid copying? 20:22
nine++ # tracking down rough stuff 20:23
20:24 [Sno] left
nine tadzik: I probably could if I find out how to keep NativeCall from freeing the string. 20:25
20:25 [Sno] joined, [Sno] left
tadzik nine: or tell it to decrease the refcount instead 20:25
maybe Nativecall should have something like "call this instead of free()"
nine Oh! I just figured that since callbacks reference the objects, that means the GC still knows that they are in use and I don't have to do anything :) 20:26
20:26 spider-mario joined
tadzik it'd probably also help things like gtk 20:26
glib in general
nine But this brings me to the question: when are NativeCall callbacks actually freed?
I probably have to do that myself. Can I just free() them? in my Perl6::Object destructor? 20:27
20:27 spider-mario left
tadzik FROGGS: as for the issue: I don't really know why it doesn't work 20:27
so it's not easy to fix to me :)
FROGGS coward! :P 20:28
dalek kudo/nom: 39ee9fa | TimToady++ | src/core/Str.pm:
don't implement .words with .comb

Runs about six times faster now.		 20:29
tadzik TimToady++
FROGGS TimToady++ 20:30
timotimo is it also still faster for short strings? 20:31
20:33 denis_boyun left, mr-foobar left
TimToady probably 20:40
dalek kudo-star-daily: f10f88a | coke++ | log/ (14 files):
today (automated commit)
 20:43
rl6-roast-data: efe0fe5 | coke++ | / (6 files):
today (automated commit)
[Coke] lots of failures across platforms in S32-temporal/local.t 20:44
parrot still dying after the 1/0 update (and fix) 20:45
kjs_ TimToady: Just wondering (as a lurker): the original code was 2 lines, which I would say is impressive, compared to the new 20-line implementation. Is this just a temporary replacement until .comb gets faster?
20:45 spider-mario joined
TimToady well, if .comb gets considerably faster, we can consider it :) 20:45
timotimo right
20:46 colomon joined
TimToady there are probably a general class of .comb patterns that reduce to what .words is doing, and could be optimized the same way 20:46
kjs_ ok. just hard to imagine that so much code is 6 times faster than the 2-line implementation :-) 20:47
20:47 Ven joined
[Coke] bulding rakudo behind a firewall. using --git-protocol to rakudo's configure. can't build moar because of the submodules using their own protocol. 20:47
sjn algorithmic complexity ftl
[Coke] kjs_: take a look at .comb
TimToady you probably don't realize how hard P5 works to avoid calling the regex engine :)
[Coke] it's shorter now than it was.
lizmat TimToady: I wonder whether pack/unpack would benefit from such a treatment 20:48
TimToady what is pack/unpack doing that you would like to avoid? 20:49
20:49 FROGGS[mobile] left, kaare_ left
kjs_ [Coke]: oh right - you mean when you consider match as well. 20:52
TimToady regex has a lot of setup/teardown at the best of times
20:54 rindolf left 20:56 kurahaupo joined 20:58 treehug88 joined
lizmat TimToady: for $template.comb(/<[a..zA..Z]>[\d+|'*']?/) -> $unit { 20:58
hmm... that's really just set up 20:59
21:03 zakharyas left, _thou joined 21:08 anaeem1_ left, _thou left 21:10 dwarring joined 21:11 colomon left 21:12 Ven left 21:13 Rotwang joined, ab5tract joined, Rotwang left
dalek p: f6ceab6 | (David Warring [email@hidden.address] | examples/rubyish/ (3 files):
[examples/rubyish] disable interpolation in single quoted strings etc
 21:15
ab5tract m: my @greetings = '\o\', '\o/', '/o/' 21:17
camelia rakudo-moar c524e7: OUTPUT«===SORRY!=== Error while compiling /tmp/k1DP__aMpB␤Missing semicolon.␤at /tmp/k1DP__aMpB:1␤------> my @greetings = '\o\', '\⏏o/', '/o/'␤ expecting any of:␤ postfix␤»
ab5tract m: my @greetings = '\o\', '\o/', '/o/';
camelia rakudo-moar c524e7: OUTPUT«===SORRY!=== Error while compiling /tmp/6FQeQZCvfB␤Missing semicolon.␤at /tmp/6FQeQZCvfB:1␤------> my @greetings = '\o\', '\⏏o/', '/o/';␤ expecting any of:␤ postfix␤»
ab5tract m: my @greetings = '\o\', '\\o/', '/o/'; 21:18
camelia rakudo-moar c524e7: OUTPUT«===SORRY!=== Error while compiling /tmp/Mdzspluk1b␤Missing semicolon.␤at /tmp/Mdzspluk1b:1␤------> my @greetings = '\o\', '\⏏\o/', '/o/';␤ expecting any of:␤ postfix␤»
ab5tract seems a bit odd 21:19
anyway, greetings!
dg m: my @greetings = '\o\\', '\o/', '/o/';
camelia ( no output )
ab5tract m:'\o\', '\\o/', '/o/'.roll(6); 21:20
21:20 gfldex left
ab5tract m: '\o\\', '\o/', '/o/'.roll(6); 21:20
camelia ( no output )
lizmat m: < \o/ o/ /o >.roll(6).say 21:21
camelia rakudo-moar c524e7: OUTPUT«o/ o/ /o /o /o \o/␤»
lizmat m: < \o/ o/ /o >.pick(*).say
camelia rakudo-moar c524e7: OUTPUT«\o/ o/ /o␤»
lizmat m: < \o/ o/ /o >.pick(*).say 21:22
camelia rakudo-moar c524e7: OUTPUT«/o o/ \o/␤»
lizmat *phew*
ab5tract m: < \o/ /o/ \o\ >.pick(*).say 21:24
camelia rakudo-moar c524e7: OUTPUT«\o\ \o/ /o/␤»
ab5tract ok
so just stick to < > for lists of single quoted text :)
i think i can do that
i had a different question in mind though 21:25
m: use MONKEY_TYPING; augment class Cool { method como { say "{$.name}.perl is perl" } }; "this".como; 21:27
camelia rakudo-moar c524e7: OUTPUT«No such method 'como' for invocant of type 'Str'␤ in block <unit> at /tmp/EHFbbmDbDD:1␤␤»
ab5tract m: use MONKEY_TYPING; augment class Str { method como { say "{$.name}.perl is perl" } }; "this".como;
camelia rakudo-moar c524e7: OUTPUT«No such method 'name' for invocant of type 'Str'␤ in method como at /tmp/2RgrTdEHnI:1␤ in block <unit> at /tmp/2RgrTdEHnI:1␤␤»
ab5tract m: use MONKEY_TYPING; augment class Str { method como { say "{$..perl is perl" } }; "this".como; 21:28
camelia rakudo-moar c524e7: OUTPUT«===SORRY!=== Error while compiling /tmp/vnSt0brP6l␤Two terms in a row␤at /tmp/vnSt0brP6l:1␤------> tr { method como { say "{$..perl is perl⏏" } }; "this".como;␤ expecting any of:␤ argument list␤ …»
ab5tract m: use MONKEY_TYPING; augment class Str { method como { say "{$.perl is perl" } }; "this".como;
camelia rakudo-moar c524e7: OUTPUT«===SORRY!=== Error while compiling /tmp/RqDAej_z1E␤Two terms in a row␤at /tmp/RqDAej_z1E:1␤------> class Str { method como { say "{$.perl ⏏is perl" } }; "this".como;␤ expecting any of:␤ statement …»
ab5tract ok sorry guys
lizmat at the moment, class are closed by default 21:29
with MONKEY_TYPING you *can* augment a class with a method
ab5tract m: use MONKEY_TYPING; augment class Str { method como { say "{$.perl} is perl" } }; "this".como;
camelia rakudo-moar c524e7: OUTPUT«"this" is perl␤»
lizmat however, and this is a known bug, that method is *only* visible in the class it is made in
ab5tract ok
lizmat it is *not* known in any of its subclasses
ab5tract makes sense, but also kind of puts a damper on a project i had in mind 21:30
lizmat m: use MONKEY_TYPING; augment class Str { method foo { say "foo" } }; "bar".foo 21:31
camelia rakudo-moar c524e7: OUTPUT«foo␤»
ab5tract which was to create an .explore method and mix it in at the Mu level :)
lizmat m: use MONKEY_TYPING; augment class Cool { method foo { say "foo" } }; "bar".foo # not seen on Str
camelia rakudo-moar c524e7: OUTPUT«No such method 'foo' for invocant of type 'Str'␤ in block <unit> at /tmp/C_jCxP68Cn:1␤␤»
ab5tract yeah, that was my question :)
it didn't seem very perl 6-y to have it work in the first case but not in the second 21:32
lizmat not sure if this helps:
use MONKEY_TYPING; augment class Cool { method foo { say "foo" } }; augment class Str { }; "bar".foo
m: use MONKEY_TYPING; augment class Cool { method foo { say "foo" } }; augment class Str { }; "bar".foo
camelia rakudo-moar 39ee9f: OUTPUT«foo␤»
lizmat the main reason for this issue, is that classes currently are very bad parents: 21:33
ab5tract so the it just sort of needs a bit of a jump start, eh?
lizmat children know who their parents are, but parents do not know who their children are
ab5tract ohh
how very Perl ;)
lizmat so, when augmenting Cool, it doesn't know who to notify of the changes made
this *will* be fixed at one point, but is not high on the list atm 21:34
ab5tract that would also put a barrier on the .explore idea
on what layer would it need to be addressed? is it an issue in the rakudo nqp code or is it on a vm level? 21:35
lizmat what would such an .explore method do?
I think it would be at the nqp level, with maybe some VM hooks involved 21:36
jnthn would know better...
ab5tract list available methods, documentation on said methods, and explore inter-relationships between the various classes and roles
lizmat m: say Str.^methods 21:37
camelia rakudo-moar 39ee9f: OUTPUT«WHY BUILD Int Num chomp chop substr pred succ match ords lines samecase samespace trim-leading trim-trailing trim words encode wordcase trans indent codes path unival univals WHICH Bool Str Stringy DUMP ACCEPTS Numeric gist perl comb subst split␤»
lizmat m: say Str.^methods>>.WHY
camelia rakudo-moar 39ee9f: OUTPUT«(Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any) (Any)␤»
ab5tract and -- the most important part -- leave bits of wisdom and even code examples by also mixing in some methods for adding text
lizmat no explanation yet :-(
ab5tract then synchronizing those somehow (optional) 21:38
so .WHY would explain where the method come from, i gather? 21:39
21:39 virtualsue joined
ab5tract *came 21:39
lizmat yes, that's the idea 21:40
look at S26:01
synopsebot Link: perlcabal.org/syn/S26.html#line_01
lizmat hoelzro has recently done a lot in that area...
hoelzro looks 21:41
ab5tract so yeah .explore would rely on this stuff
hoelzro lizmat: I'd like to start documenting the core methods, but I don't want to kill parse time =/ 21:42
lizmat FWIW, it feels like .explore would be a development tool, so it will probably never live in core
hoelzro jnthn had the idea of having docs stored elsewhere, perhaps even generated from doc.perl6.org stuff
agreed
there's nothing that says that WHY must come exclusively from #|/#= comments 21:43
ab5tract m: Str.trim.WHEREFORE
camelia rakudo-moar 39ee9f: OUTPUT«Invocant requires an instance, but a type object was passed␤ in method trim at src/gen/m-CORE.setting:6744␤ in block <unit> at /tmp/uvwJYpY7n3:1␤␤»
lizmat hoelzro: indeed...
m: "foo".trim.WHEREFORE 21:44
camelia rakudo-moar 39ee9f: OUTPUT«No such method 'WHEREFORE' for invocant of type 'Str'␤ in block <unit> at /tmp/I11mU1sobl:1␤␤»
hoelzro WHEREFORE is on POD objects
lizmat m: "foo".trim.WHY.WHEREFORE
camelia rakudo-moar 39ee9f: OUTPUT«No such method 'WHEREFORE' for invocant of type 'Any'␤ in block <unit> at /tmp/HGhegGE04_:1␤␤»
21:44 colomon joined
ab5tract gotcha 21:44
hoelzro m: #|(hello) sub foo {} ; say(&foo.WHY.WHEREFORE)
camelia rakudo-moar 39ee9f: OUTPUT«sub foo () { #`(Sub|53068504) ... }␤»
lizmat m: #|(hello) sub foo {} ; say &foo.WHY # pod at runtime! 21:48
camelia rakudo-moar 39ee9f: OUTPUT«hello␤»
ab5tract oh snap!
i like the look of this synopsis
21:50 erkan left
ab5tract and i am definitely into this built in introspection 21:51
hoelzro ab5tract: you might also be interested in the meta object API
21:51 erkan joined
hoelzro ex: 21:51
ab5tract like so much of my experience with perl 6 so far, the feature is just there already
hoelzro m: .say for Str.^methods 21:52
camelia rakudo-moar 39ee9f: OUTPUT«WHY␤BUILD␤Int␤Num␤chomp␤chop␤substr␤pred␤succ␤match␤ords␤lines␤samecase␤samespace␤trim-leading␤trim-trailing␤trim␤words␤encode␤wordcase␤trans␤indent␤codes␤path␤unival␤univals␤WHICH␤Bool␤Str␤Stringy…»
hoelzro eh
m: say Str.^methods
camelia rakudo-moar 39ee9f: OUTPUT«WHY BUILD Int Num chomp chop substr pred succ match ords lines samecase samespace trim-leading trim-trailing trim words encode wordcase trans indent codes path unival univals WHICH Bool Str Stringy DUMP ACCEPTS Numeric gist perl comb subst split␤»
hoelzro that's better.
ab5tract yeah, i want to make an interface for this. maybe even render it to HTML and serve it on a local port 21:53
lizmat m: say Str.^methods.sort
camelia rakudo-moar 39ee9f: OUTPUT«Code object coerced to string (please use .gist or .perl to do that) in any p6sort at src/vm/moar/Perl6/Ops.nqp:346␤␤Code object coerced to string (please use .gist or .perl to do that) in any p6sort at src/vm/moar/Perl6/Ops.nqp:346␤␤Code object …»
ab5tract though something in ncurses could be cool as well ;)
lizmat m: say Str.^methods>>.gist.sort
camelia rakudo-moar 39ee9f: OUTPUT«ACCEPTS BUILD Bool DUMP Int Num Numeric Str Stringy WHICH WHY chomp chop codes comb encode gist indent lines match ords path perl pred samecase samespace split subst substr succ trans trim trim-leading trim-trailing unival univals wordcase words␤»
hoelzro m: Str.^methods.grep(*.name ne '<anon>').sort(*.name)
camelia ( no output )
hoelzro m: say Str.^methods.grep(*.name ne '<anon>').sort(*.name)
camelia rakudo-moar 39ee9f: OUTPUT«ACCEPTS BUILD Bool DUMP Int Num Numeric Str Stringy WHICH WHY chomp chop codes comb encode gist indent lines match ords path perl pred samecase samespace split subst substr succ trans trim trim-leading trim-trailing unival univals wordcase words␤»
ab5tract m: say Str.^methods.grep(*.name eq '<anon>').sort(*.name) 21:55
camelia rakudo-moar 39ee9f: OUTPUT«␤»
hoelzro there happen to be none on Str =)
ab5tract so what's the metamagic to insert some pod onto those methods 21:56
hoelzro well, there's nothing spec'd 21:57
but I happen to know the secret incantations to do so ;) 21:58
actually, since we could stick that "incantation" on the end of CORE.setting... 22:01
we might be able to easily document core classes/methods without imposing too much extra parse time
it could even be a Configure.pl option 22:02
dalek kudo/nom: cf69d8b | (Elizabeth Mattijsen)++ | src/core/List.pm:
Use int in List.ACCEPTS

For 1000 element identical lists, this is 14% faster. Benchmark:
   my @a = ^1000; my @b = ^1000; my $a=0; for ^1000 { $a++ if @a ~~ @b }; say $a
From 20 CPU secs to 17.3.		 22:03
lizmat sleep& so good night, #perl6!
22:04 ajr_ left
ab5tract m: my $ex = "trim"; for Str.^methods -> $m { "$ex.$m = '|#(where) $m.perl'".EVAL } 22:04
camelia rakudo-moar 39ee9f: OUTPUT«Code object coerced to string (please use .gist or .perl to do that) in block <unit> at /tmp/E5wz6xDqRj:1␤␤Code object coerced to string (please use .gist or .perl to do that) in block <unit> at /tmp/E5wz6xDqRj:1␤␤Not enough positional parameters…»
hoelzro night lizmat
ab5tract o/ lizmat !
22:04 _thou joined
ab5tract i'm pretty sure that road leads to evil 22:05
hoelzro ab5tract: which one?
ab5tract m: for Str.^methods -> $m { "Str.$m = '|#(where) $m.perl'".EVAL }; say Str.^methods>>.WHY; 22:06
camelia rakudo-moar 39ee9f: OUTPUT«Code object coerced to string (please use .gist or .perl to do that) in block <unit> at /tmp/Dyy6Qxn6Ze:1␤␤Code object coerced to string (please use .gist or .perl to do that) in block <unit> at /tmp/Dyy6Qxn6Ze:1␤␤Invocant requires an instance, b…»
ab5tract m: for Str.^methods -> $m { "Str.$m.perl = '|#(where) $m.perl'".EVAL }; say Str.^methods>>.WHY;
camelia rakudo-moar 39ee9f: OUTPUT«Code object coerced to string (please use .gist or .perl to do that) in block <unit> at /tmp/fXAwQK1rRy:1␤␤Code object coerced to string (please use .gist or .perl to do that) in block <unit> at /tmp/fXAwQK1rRy:1␤␤Invocant requires an instance, b…»
hoelzro ab5tract: you probably want "$m.perl()" 22:07
ab5tract m: for Str.^methods -> $m { Str.$m = "'|#(where) $m.perl'".EVAL }; say Str.^methods>>.WHY;
camelia rakudo-moar 39ee9f: OUTPUT«Invocant requires an instance, but a type object was passed␤ in method WHY at src/gen/m-CORE.setting:5960␤ in block <unit> at /tmp/EGZbE1lWcV:1␤␤»
hoelzro m: my $m = 'foo' ; say "$m.perl"
camelia rakudo-moar 39ee9f: OUTPUT«foo.perl␤»
hoelzro m: my $m = 'foo' ; say "$m.perl()"
camelia rakudo-moar 39ee9f: OUTPUT«"foo"␤»
ab5tract anyway, as i said, evil :)
hoelzro very >:)
ab5tract m: for Str.^methods -> $m { Str.$m = "'|#(where) $m.perl()'".EVAL }; say Str.^methods>>.WHY;
camelia rakudo-moar 39ee9f: OUTPUT«Invocant requires an instance, but a type object was passed␤ in method WHY at src/gen/m-CORE.setting:5960␤ in block <unit> at /tmp/UAq2FHbrTb:1␤␤»
ab5tract so what's the magic, doc? 22:09
:D
m: for Str.^methods -> $m { "Str.$m.perl() = '|#(where) $m.perl()'".EVAL }; say Str.^methods>>.WHY;
camelia rakudo-moar 39ee9f: OUTPUT«===SORRY!=== Error while compiling EVAL_0␤Two terms in a row␤at EVAL_0:1␤------> Str.method ⏏WHY (Str:D: *%_) { #`(Method|63575496) .␤ expecting any of:␤ method arguments␤ postfix␤ …»
hoelzro m: Str.^find_method('trim').set_why('trims the whitespace off of a string') 22:10
camelia ( no output )
hoelzro m: Str.^find_method('trim').set_why('trims the whitespace off of a string') ; say Str.^find_method('trim').WHY
camelia rakudo-moar 39ee9f: OUTPUT«No such method 'set_docee' for invocant of type 'Str'␤ in method WHY at src/gen/m-CORE.setting:3501␤ in block <unit> at /tmp/_4OUQlLWJC:1␤␤»
ab5tract hoelzro: now that's more like it!
hoelzro interesting
ab5tract: that method may disappear any day now
ab5tract yeah it should be spec'd 22:11
hoelzro so don't get too comfortable with it!
ab5tract (the solution i mean)
hoelzro well, I'd rather something like Str.^find_method('trim').WHY = 'trims a string'
agreed
ab5tract yeah, .WHY should have a .= 22:12
hoelzro I tried doing that, but my NQP fu was weak at the time
22:12 lastofthe left
hoelzro I still might not be able to do it 22:12
someone like jnthn probably could
ab5tract i think some process of making it even easier could be a good idea too 22:13
it wouldn't even need to interact directly with the Pod looked up by .WHY 22:14
(though the .WHY explanation could be derived from the larger Pod)
22:14 firefish5000 joined 22:15 virtualsue left
hoelzro ab5tract: what exactly do you mean? 22:15
ab5tract i'm envisioning a way of making it very easy to add documentation around Perl 6 22:17
something like a small web application backed up by a git repository 22:18
by backed up i mean the it would source it's content from the git repository
and serve it to your browser over a local port
(or a cool CLI interface, though I think the HTML is easier short term) 22:19
kind of like a distributed perlmonks, i guess
hoelzro ab5tract: so the WHYs of libraries could be populated by such a Git repo? 22:20
ab5tract theoretically, yes, but not necessary either
if it were in Rakudo * then it could be the substrate of the shipped docs 22:21
hoelzro ab5tract: that's an interesting idea 22:22
we'll have to see what TimToady thinks 22:23
ab5tract hoelzro: glad you think so!
i'm curious to hear what he will say 22:24
basically i've been spending a lot of time blowing my own mind in the moar repl
and thinking "i wish i had somewhere useful to stash that example" 22:25
other than just a flat text file growing in perl 6 examples 22:26
22:27 jack_rabbit left
hoelzro like an image? 22:27
ab5tract i'd say more like a network
hoelzro ah ha 22:28
sorry, I have to go for now
I have class in a bit!
ab5tract okay, cheers hoelzro !
hoelzro later! 22:29
22:31 Psyche^ joined 22:35 Psyche^_ left 22:38 treehug88 left
firefish5000 How do you extract rules from captures? something like { $_<rule>.say } for $/[*]; ? 22:40
22:44 sivoais left 22:46 ren1us joined
ren1us anyone know if there's p6 syntax highlighting available for sublime? 22:46
timotimo i don't think so :( 22:53
ren1us i found this last night, but it doesn't seem to actually do anything: github.com/cognominal/sublime-perl6 22:55
22:57 pippo joined, spider-mario left 22:58 Guest10538 left, Guest10538 joined 22:59 Guest10538 is now known as skarn 23:00 xragnar_ joined, xragnar left, xragnar_ is now known as xragnar
japhb .ask lizmat I saw your benchmark in the comments for the List.ACCEPTS optimization. Care for a perl6-bench commit bit? :-) 23:01
yoleaux japhb: I'll pass your message to lizmat.
23:02 kjs_ left 23:12 lastofthe joined 23:16 avuserow left 23:17 beastd left
pippo m: grammar G { rule TOP {<line>}; rule line {\N+}}; G.parse('Hello World'); say $<line>; 23:18
camelia rakudo-moar cf69d8: OUTPUT«｢Hello World｣␤␤»
pippo m: grammar G { rule TOP {<line>}; rule line {\N+}}; G.parse('Hello World'); say A ~ $<line>; 23:19
camelia rakudo-moar cf69d8: OUTPUT«===SORRY!=== Error while compiling /tmp/BkZ3UBIcMo␤Undeclared name:␤ A used at line 1␤␤»
23:19 BenGoldberg joined
pippo m: grammar G { rule TOP {<line>}; rule line {\N+}}; G.parse('Hello World'); say ~A ~ $<line>; 23:19
camelia rakudo-moar cf69d8: OUTPUT«===SORRY!=== Error while compiling /tmp/kudVtopjse␤Undeclared name:␤ A used at line 1␤␤»
pippo m: grammar G { rule TOP {<line>}; rule line {\N+}}; G.parse('Hello World'); say 'A' ~ $<line>;
camelia rakudo-moar cf69d8: OUTPUT«AHello World␤»
pippo m: grammar G { rule TOP {<line>}; rule line {\N+}}; G.parse('Hello World'); say $<line>;
camelia rakudo-moar cf69d8: OUTPUT«｢Hello World｣␤␤» 23:20
pippo firefish5000: ^^ does this help?
firefish5000 Unfortunatly no. My case is something like... 23:23
m: grammar G { rule TOP { (<word>)+ }; rule word {\S+}}; G.parse('Hello World'); for $/[*] { say $_<word>;}
camelia ( no output )
firefish5000 m: grammar G { rule TOP { (<word>)+ }; rule word {\S+}}; G.parse('Hello World'); for $/[0..*] { say $_<word>;}
camelia ( no output )
firefish5000 m: grammar G { rule TOP { (<word>)+ }; rule word {\S+}}; G.parse('Hello World'); say $/[0]<word>; 23:24
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
pippo m: grammar G { rule TOP { (<word>)+ }; rule word {\S+}}; G.parse('Hello World'); say $<word>;
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
firefish5000 But I keep getting (Any) or 'postcircumfix:<{ }> not defined for type' Any/Array 23:25
23:25 mberends left, BenGoldberg left
pippo m: grammar G { rule TOP { <word>+ }; rule word {\S+}}; G.parse('Hello World'); say $<word>; 23:25
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
firefish5000 m: grammar G { rule TOP { (<word>)+ }; rule word {\S+}}; G.parse('Hello World'); say $/.perl; 23:26
camelia rakudo-moar cf69d8: OUTPUT«Any␤»
pippo m: grammar G { rule TOP { <word>+ }; rule word {\S+}}; my $result = G.parse('Hello World'); say $result<word>;
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
23:26 BenGoldberg joined
firefish5000 m: grammar G { rule TOP { (<word>)+ }; rule word {\S+}}; G.parse('Hello World'); say $/; 23:26
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
firefish5000 mayby I mistyped something this time
pippo m: grammar G { rule TOP { <word>+ }; rule word {\w+}}; my $result = G.parse('Hello World'); say $result<word>; 23:27
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
pippo m: grammar G { rule TOP { <word>+ }; rule word {\w+}}; my $result = G.parse('Hello World'); say $/;
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
firefish5000 m: grammar G { rule TOP { (<word>)+ }; token word {\S+}}; G.parse('Hello World'); say $/;
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
firefish5000 m: grammar G { rule TOP { <word>+ }; token word {\S+}}; G.parse('Hello World'); say $/;
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
firefish5000 m: grammar G { rule TOP { <word>+ }; token word {\S+}}; my $res = G.parse('Hello World'); say $res.perl; 23:28
camelia rakudo-moar cf69d8: OUTPUT«Any␤»
firefish5000 Ahh, stupid mistak here... I never match the space 23:30
m: grammar G { rule TOP { (<word>+\s+) }; token word {\S+}}; my $res = G.parse('Hello World'); say $res.perl;
camelia rakudo-moar cf69d8: OUTPUT«Any␤»
firefish5000 m: grammar G { rule TOP { (<word>\s*)+ }; token word {\S+}}; my $res = G.parse('Hello World'); say $res.perl;
camelia rakudo-moar cf69d8: OUTPUT«Match.new(orig => "Hello World", from => 0, to => 11, ast => Any, list => ([Match.new(orig => "Hello World", from => 0, to => 6, ast => Any, list => ().list, hash => EnumMap.new("word", Match.new(orig => "Hello World", from => 0, to => 5, ast => Any, list …»
firefish5000 m: grammar G { rule TOP { (<word> \s+)+ }; rule word {\S+}}; G.parse('Hello World'); for $/[0..*] { say $_<word>;} 23:31
camelia ( no output )
firefish5000 m: grammar G { rule TOP { (<word> \s+)+ }; rule word {\S+}}; G.parse('Hello World'); say $/.perl;}
camelia rakudo-moar cf69d8: OUTPUT«===SORRY!=== Error while compiling /tmp/H_36zwMLkt␤Unexpected closing bracket␤at /tmp/H_36zwMLkt:1␤------> }}; G.parse('Hello World'); say $/.perl;⏏}␤»
firefish5000 m: grammar G { rule TOP { (<word> \s+)+ }; rule word {\S+}}; G.parse('Hello World'); say $/.perl; 23:32
camelia rakudo-moar cf69d8: OUTPUT«Any␤»
TimToady that can't match
\s+ requires a space after World 23:33
firefish5000 m: grammar G { rule TOP { (<word> \s*)+ }; rule word {\S+}}; G.parse('Hello World'); say $/.perl;
camelia rakudo-moar cf69d8: OUTPUT«Match.new(orig => "Hello World", from => 0, to => 11, ast => Any, list => ([Match.new(orig => "Hello World", from => 0, to => 6, ast => Any, list => ().list, hash => EnumMap.new("word", Match.new(orig => "Hello World", from => 0, to => 5, ast => Any, list …»
firefish5000 Ahh, Thanks TimeToady.
TimToady and you don't really need it there anyway 23:34
firefish5000 *TimToady.
pippo m: grammar G { rule TOP { (<word> \s*)+ }; rule word {\S+}}; G.parse('Hello World'); say $<word>;
camelia rakudo-moar cf69d8: OUTPUT«(Any)␤»
TimToady m: grammar G { rule TOP { (<word> )+ }; rule word {\S+}}; G.parse('Hello World'); say $/.perl;
camelia rakudo-moar cf69d8: OUTPUT«Match.new(orig => "Hello World", from => 0, to => 11, ast => Any, list => ([Match.new(orig => "Hello World", from => 0, to => 6, ast => Any, list => ().list, hash => EnumMap.new("word", Match.new(orig => "Hello World", from => 0, to => 5, ast => Any, list …»
TimToady since 'rule' already implies that whitespace matches \s*
I wouldn't use rule for word though, it's obviously just a token 23:35
m: grammar G { rule TOP { (<word> )+ }; token word { \S+ }}; G.parse('Hello World'); say $/.perl;
camelia rakudo-moar cf69d8: OUTPUT«Match.new(orig => "Hello World", from => 0, to => 11, ast => Any, list => ([Match.new(orig => "Hello World", from => 0, to => 6, ast => Any, list => ().list, hash => EnumMap.new("word", Match.new(orig => "Hello World", from => 0, to => 5, ast => Any, list …»
TimToady m: grammar G { rule TOP { <word> + }; token word { \S+ }}; G.parse('Hello World'); say $/.perl; 23:36
pippo m: grammar G { rule TOP { (<word> )+ }; token word { \S+ }}; my $result = G.parse('Hello World'); say $result<word>;
camelia rakudo-moar cf69d8: OUTPUT«Match.new(orig => "Hello World", from => 0, to => 11, ast => Any, list => ().list, hash => EnumMap.new("word", [Match.new(orig => "Hello World", from => 0, to => 5, ast => Any, list => ().list, hash => EnumMap.new()), Match.new(orig => "Hello World", from …»
rakudo-moar cf69d8: OUTPUT«(Any)␤»
TimToady m: grammar G { rule TOP { <word> + }; token word { \S+ }}; G.parse('Hello World'); say $/<word>;
camelia rakudo-moar cf69d8: OUTPUT«｢Hello｣␤ ｢World｣␤␤»
firefish5000 Well, its just a test case, so rule should work
m: grammar G { rule TOP { (<word> )+ }; rule word {\S+}}; G.parse('Hello World'); for $/[*] { say $_<word> ; } ;
camelia rakudo-moar cf69d8: OUTPUT«postcircumfix:<{ }> not defined for type Array␤ in method gist at src/gen/m-CORE.setting:13326␤ in sub say at src/gen/m-CORE.setting:14281␤ in block <unit> at /tmp/44w8QIaKf6:1␤␤»
TimToady see what I did 23:37
<word> +
m: grammar G { rule TOP { [<word> ]+ }; token word { \S+ }}; G.parse('Hello World'); say $/<word>; 23:38
camelia rakudo-moar cf69d8: OUTPUT«｢Hello｣␤ ｢World｣␤␤»
TimToady there's another way
if you don't like the ws before the +
pippo m: grammar G { rule TOP { [<word> ]+ }; token word { \S+ }}; G.parse('Hello World'); say $/<word>.WHAT;
camelia rakudo-moar cf69d8: OUTPUT«(Array)␤»
pippo m: grammar G { rule TOP { [<word> ]+ }; token word { \S+ }}; G.parse('Hello World'); .say for $/<word>; 23:39
camelia rakudo-moar cf69d8: OUTPUT«｢Hello｣␤ ｢World｣␤␤»
firefish5000 But, in the case of [(<rule1>||<rule2>)\h*\n?]+, is there a way to keep the order?
pippo I learned a new thing. Thank you TimToady. 23:40
23:40 xenoterracide joined
TimToady [$<rule>=(<rule1>||<rule2>)\h*\n?]+ or so 23:40
or just write a rule that does rule1||rule2 23:41
don't try to pack too much into one regex
firefish5000 hmm, I suppose that would work.. but it makes the last 2hr of struggling seem so pointless. 23:42
TimToady you have to enjoy struggling if you're to make it in this line of work :) 23:43
timotimo i decided to do a btrfs relocation across devices and not have a backup before starting it ... this will be ... interesting 23:45
firefish5000 I enjoy it well. It's just that when the simple answer is given by someone else, I cannot help but feel defeated. 23:46
firefish5000 guesses that asking for help would be pointless if he doesn't want the answer
TimToady you are allowed to feel bemused, but you are not allowed to feel defeated :) 23:47
lastofthe firefish5000: the path to simple is long and winding :)
TimToady indeed, it's taken me several decades to get here :) 23:48
you kind of have to expect to try a thousand ways not to invent the light bulb, and then when you do invent the light bulb, you'll be wrong after another century or so :) 23:50
23:50 leont joined
TimToady the only true defeat is one that you learn nothing from 23:51
that's why it sucks to be dead after one :) 23:52
leont What's the equivalent of perldoc for rakudo?
BenGoldberg Nicola Tesla invented the flourescent lamp a century ago, but it took until a decade ago for someone to think of twisting it into a helix for more compactness. 23:53
leont It seems I'm doing something wrong with spurt, that I think used to work a year ago, but I don't know how I should use it
firefish5000 well, i'll do my best. although my light bulb usualy needs a centurys worth of refinement.
TimToady: lol
pippo night #perl6!
leave
23:54 pippo 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!