»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.
00:01 adu_ left 00:11 Infamous joined 00:12 PacoAir left 00:14 jac50 left 00:21 PacoAir joined, PacoAir left 00:29 adu_ joined 00:33 LlamaRider left
[Coke] rakudo now up to 16 failures, niecza 8, pugs 2. 00:52
(those pugs failures are sporadic) 00:53
rakudo's new failures are in S26-documentation/09-configuration.t 00:56
00:58 adu_ left 01:15 Infamous left 01:22 Chillance left
timotimo i wish rosettacode had some function to give me pages that are tasks, but not in the perl6 group :| 01:26
TimToady you an unimplemented in Perl 6 page? 01:28
you mean*
there is one
rosettacode.org/wiki/Reports:Tasks_..._in_Perl_6
or did you mean something else?
timotimo oooh! 01:30
that is so neat :)
oh, those are many more than i expected 01:31
"the developers thought of everything" i must say
01:32 census left
TimToady there aren't so few of them that we have to worry much about developer collisions yet 01:34
timotimo :) 01:36
01:39 bowtie left
timotimo hm, opengl bindings would not be terrifically hard to make with zavolaj, i imagine ... 01:40
01:40 isomorphisms joined
timotimo i think i'll try to implement this prettily: rosettacode.org/wiki/Brownian_tree 01:44
diakopter bonus if you can beat TI-83 BASIC in codesize 01:45
timotimo i'll try to get a good balance of "short" and "readable". should be totally doable in perl6 ;) 01:47
hm, there seems to be quite a bit of freedom in those exercises 01:50
diakopter well, people take liberties, anyway
timotimo sure 01:52
01:54 marloshouse joined 01:57 PacoAir joined
timotimo r: my %h; %h[1][2] = 5; say %h; 02:02
p6eval rakudo 8ea992: OUTPUT«Index out of range. Is: 2, should be in 0..0␤ in method <anon> at src/gen/CORE.setting:10523␤ in at src/gen/Metamodel.pm:2493␤ in any find_method_fallback at src/gen/Metamodel.pm:2481␤ in any find_method at src/gen/Metamodel.pm:929␤ in block at /tmp/ScWlN0hU…
timotimo hm :|
r: my %h; %h{1}{2} = 5; say %h; 02:03
p6eval rakudo 8ea992: OUTPUT«("1" => {"2" => 5}).hash␤»
timotimo i keep forgetting this
02:04 adu_ joined
TimToady r: my @h; @h[1][2] = 5; say @h; # just checking 02:08
p6eval rakudo 8ea992: OUTPUT«use of uninitialized value of type Any in string context in block at /tmp/nuwzWSmTz0:1␤␤use of uninitialized value of type Any in string context in block at /tmp/nuwzWSmTz0:1␤␤use of uninitialized value of type Any in string context in block at /tmp/nuwzWSmTz…
TimToady heh 02:09
02:09 adu_ left
timotimo use of uninitialized variable $s of type Any in numeric context 02:10
i do not understand why
i don't have a $s anywhere in my code ... ?!
TimToady it'll be somewhere in the setting, probably 02:11
timotimo ah, probably
02:11 adu_ joined
timotimo it seems like it comes from something like this: 02:11
r: my %h; %h{1}{2} == 5;
p6eval rakudo 8ea992: OUTPUT«use of uninitialized variable $v of type Any in numeric context in block at /tmp/YKysIoZaog:1␤␤»
lue is the C<==> a typo in your code? 02:12
timotimo it's in an if in my code
02:13 PacoAir left
lue r: my %h; %h<1><2> == 5; say "I live!"; 02:14
p6eval rakudo 8ea992: OUTPUT«WARNINGS:␤Useless use of "==" in expression "<2> == 5" in sink context (line 1)␤use of uninitialized variable $v of type Any in numeric context in block at /tmp/lMU9S30pUv:1␤␤I live!␤»
lue r: my %h; %h{1}{2} == 5; say "I live!";
p6eval rakudo 8ea992: OUTPUT«WARNINGS:␤Useless use of "==" in expression "{2} == 5" in sink context (line 1)␤use of uninitialized variable $v of type Any in numeric context in block at /tmp/hfIJFWoUs0:1␤␤I live!␤»
lue r: my %h; %h{1}{2} ~~ 5; say "I live!"; 02:15
p6eval rakudo 8ea992: OUTPUT«use of uninitialized variable $a of type Any in numeric context in block at /tmp/sQa0ut8TKt:1␤␤I live!␤» 02:16
timotimo r: my ($x, $y) = 10, 20; my %W; %W{$x-1|$x|$x+1}{$y-1|$y|$y+1} = 1; say %W; # i wonder if this autothreads correctly?
p6eval rakudo 8ea992: OUTPUT«No such method 'STORE' for invocant of type 'Any'␤ in sub AUTOTHREAD_METHOD at src/gen/CORE.setting:2196␤ in at src/gen/Metamodel.pm:2486␤ in sub AUTOTHREAD_METHOD at src/gen/CORE.setting:2196␤ in at src/gen/Metamodel.pm:2486␤ in block at /tmp/srvf7H_rQp:1␤…
timotimo hmm
apparently not?
r: my %h; say %h{10} :exists; 02:18
p6eval rakudo 8ea992: OUTPUT«False␤»
timotimo r: my %h; say %h<<"foo">> :exists; 02:20
p6eval rakudo 8ea992: OUTPUT«False␤»
timotimo r: my %h; my $x = 10; say %h<<$x|$x>> :exists;
p6eval rakudo 8ea992: OUTPUT«Cannot use exists adverb with a slice␤ in method postcircumfix:<{ }> at src/gen/CORE.setting:1591␤ in block at /tmp/NGRqJZFRX0:1␤␤»
timotimo huh?
oh yeah, the first speck of thingie will take a loooong time to settle 02:22
this "redo" isn't working as i expect it... oh, oops :) 02:24
02:25 ggoebel_ left
timotimo oh, yeah, this will be slooooooow 02:28
i have a version now that works and outputs ascii "art" 02:43
02:45 isBEKaml joined 02:48 arlinius joined 02:49 dayangkun joined
timotimo slowperl is slow 02:50
timotimo tries on niecza 02:57
oh, timing the run while compiling niecza doesn't make terribly much sense 02:58
i hope i won't have to change too many things to make it run :| 02:59
changing the somewhat clever position checking code that used junctions to not using junctions makes niecza work and it's significantly faster 03:09
03:09 zacts joined, orafu left, orafu joined
timotimo 3:05.30elapsed <- rakudo | niecza -> 0:12.30elapsed 03:09
for a 60x60 field with 500 particles 03:10
isBEKaml 3 hrs?
timotimo minutes 03:11
isBEKaml *whew* :D
timotimo ;)
i wonder if niecza will get functioning junctions?
colomon n: my $a = 5; say "Now?" if $a = 4 | 5 | 6; 03:12
timotimo niecza is in reduced development mode, right?
p6eval niecza v24-33-gfc6d5fd: OUTPUT«Now?␤»
timotimo n: my ($x, $y) = 10, 10; sub foo($a,$b) { say ($a, $b) }; foo($x-1|$x|$x+1, $y-1|$y|$y+1); 03:13
p6eval niecza v24-33-gfc6d5fd: OUTPUT«9 9␤9 10␤9 11␤10 9␤10 10␤10 11␤11 9␤11 10␤11 11␤»
timotimo oh?
colomon Right now TimToady and I make occasional minor Niecza changes to track new bits of std, and that's about it.
timotimo so it *should* in theory have worked?
colomon timotimo: I don't know what you tried to do.
(I mean, more specifically than "used junctions") 03:14
timotimo while not %W{($x-1|$x|$x+1) ~ "|" ~ ($y-1|$y|$y+1)} :exists { 03:15
this was supposed to check all the neighbours of the particle for existing particles
that check never triggered
oh, indeed. this is interesting 03:17
my &ex = -> $x,$y { say "$x|$y"; say %W; %W{"$x|$y"} :exists; }; while not ex($x-1|$x|$x+1, $y-1|$y|$y+1) { ; this prints out any(29, 30, 31)|any(28, 29, 30) for the "$x|$y" part!
zacts is perl6 usable now for everyday projects?
timotimo so junctions work, but nested junctions don't?
colomon n: my $x = 10; my $y =11; say ($x-1|$x|$x+1) ~ "|" ~ ($y-1|$y|$y+1) 03:18
p6eval niecza v24-33-gfc6d5fd: OUTPUT«any(9, 10, 11)|any(10, 11, 12)␤»
timotimo yup, that's broken indeed
colomon r: my $x = 10; my $y =11; say ($x-1|$x|$x+1) ~ "|" ~ ($y-1|$y|$y+1)
p6eval rakudo 8ea992: OUTPUT«any(any(9|10, 9|11, 9|12), any(10|10, 10|11, 10|12), any(11|10, 11|11, 11|12))␤»
timotimo do you feel like that would be something niecza should get?
lue Where in the spec does it say that rakudo is right? At least initially I feel niecza has the right behavior 03:21
colomon timotimo: well, it seems like. but off the top of my head I don't know that I have any clue where to look to modify such behavior
timotimo oh? huh? 03:22
03:22 adu_ left
timotimo "If two or more arguments are junctive, then the argument that is chosen to be "autothreaded" is:" does that help? 03:23
03:24 Rix left
colomon lue: I would think that infix:<~> is supposed to autothread. 03:24
timotimo gotta run
bbl, though
lue I also expect it to concatenate strings, not magically cause behavior similar to infix:<Z> or infix:<X> 03:25
r: say any(9,10,11) ~ "|" ~ any(10,11,12) 03:28
p6eval rakudo 8ea992: OUTPUT«any(any(9|10, 9|11, 9|12), any(10|10, 10|11, 10|12), any(11|10, 11|11, 11|12))␤»
03:29 Rix joined 03:30 isBEKaml` joined
timotimo but junctions are magic 03:31
lue r: say any(9,10,11) | any(10,11,12)
p6eval rakudo 8ea992: OUTPUT«any(any(9, 10, 11), any(10, 11, 12))␤»
colomon I've got to say I'm not at all sure I understand nested anys.
timotimo so why should tilde magically override the magic of junctions?
03:32 isBEKaml left
lue AFAIK, infix:<~> means string concatenation. I fail to see why this magic happens (but if there's a good reason for it, I won't complain) 03:32
03:32 isBEKaml` is now known as isBEKaml
timotimo if a junction is used as an argument to any routine (operator, closure, method, etc.), and the scalar parameter you are attempting to bind the argument to is inconsistent with the Junction type, that routine is "autothreaded", 03:33
in particular a strong is inconsistent with the junction type 03:34
lue r: say (1|2|3).Str
p6eval rakudo 8ea992: OUTPUT«any(1, 2, 3)␤»
lue wait, so infix:<~> doesn't coerce to Str?
colomon nr: say (1|2|3|) ~ "hello"
p6eval niecza v24-33-gfc6d5fd: OUTPUT«===SORRY!===␤␤Bogus term at /tmp/DGAnIT3ZNi line 1:␤------> say (1|2|3|⏏) ~ "hello"␤␤Parse failed␤␤» 03:35
..rakudo 8ea992: OUTPUT«===SORRY!===␤Bogus statement␤at /tmp/y4ytKNausK:1␤------> say (1|2|3|⏏) ~ "hello"␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-infix␤ prefix or term␤ prefix or meta-prefix␤»…
colomon nr: say (1|2|3) ~ "hello"
p6eval rakudo 8ea992: OUTPUT«any(1hello, 2hello, 3hello)␤»
..niecza v24-33-gfc6d5fd: OUTPUT«Use of uninitialized value in string context␤ at /home/p6eval/niecza/lib/CORE.setting line 1342 (warn @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 266 (Mu.Str @ 15) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/lib/CORE.settin…
timotimo this is the string coercion called on each own item
colomon nr: say (1|2|3) + 3
p6eval rakudo 8ea992, niecza v24-33-gfc6d5fd: OUTPUT«any(4, 5, 6)␤»
colomon hmmm... works on infix:<+> but not infix:<~>? (in niecza) 03:36
timotimo junction does not drive from any
it is a sibling to any
03:37 isBEKaml left
timotimo that should be why it ought to auto thread 03:37
colomon, do you think sorear would agree to fix this? 03:38
colomon timotimo: no idae.
n: say any(9,10,11) + any(9,10,11) 03:39
p6eval niecza v24-33-gfc6d5fd: OUTPUT«any(any(18, 19, 20), any(19, 20, 21), any(20, 21, 22))␤»
lue r: say any(9,10,11) ~ "»ö«" ~ any(10,11,12)
p6eval rakudo 8ea992: OUTPUT«any(any(9»ö«10, 9»ö«11, 9»ö«12), any(10»ö«10, 10»ö«11, 10»ö«12), any(11»ö«10, 11»ö«11, 11»ö«12))␤»
colomon timotimo: may be that I can fix it, if it's specific to infix:<~>
03:40 adu_ joined
lue timotimo: I think I see now. (Still trying to grok the nested any's though). I think your use of "|" (which is also a junction op) is what confused me and made think a bug was afoot :) 03:41
timotimo oh, maybe a , would have been more idiomatic 03:44
i think nested anys are not too bad 03:47
i mean this is mostly about evaluating as booleans
so evaluate a nested any by evaluating all of the inner anys and see if any of then are true 03:48
03:48 telex left
timotimo same way you run functions with then 03:48
lue I think thinking of Junctions as a weird type of array that causes ~ to act like X~ will help for now :)
03:50 telex joined 03:51 xinming left 03:53 xinming joined 03:56 arlinius left
timotimo :) 03:59
rn: say substr("camel", 0|1, 2&3); 04:02
p6eval niecza v24-33-gfc6d5fd: OUTPUT«all(any("ca", "am"), any("cam", "ame"))␤»
..rakudo 8ea992: OUTPUT«any(all(ca, cam), all(am, ame))␤»
timotimo now *that* is interesting 04:03
in this case, the specs say that rakudo is in the wrong and niecza is in the right
TimToady right
timotimo also, i managed to make niecza work correctly like this: 04:04
TimToady no wait
timotimo rn: my %W; %W<10|10>; my ($x, $y) = 11, 11; my &ex = -> Cool $x, Cool $y { say "$x|$y"; say %W; %W{"$x|$y"} :exists; }; while not ex($x-1|$x|$x+1, $y-1|$y|$y+1) { say "bup" } 04:05
TimToady I think rakudo has it right
you resolve the left first
p6eval rakudo 8ea992: OUTPUT«(timeout)10|10␤().hash␤10|11␤().hash␤10|12␤().hash␤11|10␤().hash␤11|11␤().hash␤11|12␤().hash␤12|10␤().hash␤12|11␤().hash␤12|12␤().hash␤bup␤10|10␤().hash␤10|11␤().hash␤10|12␤().hash␤11|10␤().hash␤11|11␤().hash␤11|12␤().hash␤12|10␤().hash␤12|11␤().hash␤12|12␤().hash␤…
..niecza v24-33-gfc6d5fd: OUTPUT«(timeout)10|10␤{}␤10|11␤{}␤10|12␤{}␤11|10␤{}␤11|11␤{}␤11|12␤{}␤12|10␤{}␤12|11␤{}␤12|12␤{}␤bup␤10|10␤{}␤10|11␤{}␤10|12␤{}␤11|10␤{}␤11|11␤{}␤11|12␤{}␤12|10␤{}␤12|11␤{}␤12|12␤{}␤bup␤10|10␤{}␤10|11␤{}␤10|12␤{}␤11|10␤{}␤11|11␤{}␤11|12␤{}␤12|10␤{}␤12|11␤{}␤12|12…
timotimo oops, there's a say too much
TimToady so you should get substr("camel", 0, 2&3) as the first resolution
timotimo but you see how it says 10|11 and such?
diakopter please watch; decide to laugh for yourself: www.liveleak.com/view?i=df8_1362850413 04:06
timotimo perlcabal.org/syn/S09.html#Junctions - i took the example from here
TimToady oh, yes, by that niecza is righter 04:07
I'd forgotten that TheDamian defined them that way 04:08
timotimo rn: my %W; %W<10|10>; my ($x, $y) = 11, 11; my &ex = -> Cool $x, Cool $y { say "$x,$y" %W{"$x|$y"} :exists; }; while not ex($x-1|$x|$x+1, $y-1|$y|$y+1) { say "bup" }
p6eval rakudo 8ea992: OUTPUT«===SORRY!===␤Undeclared name:␤ W used at line 1␤␤»
..niecza v24-33-gfc6d5fd: OUTPUT«===SORRY!===␤␤Undeclared name:␤ 'W' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1443 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) ␤ at /home/p6ev…
timotimo rn: my %W; %W<10|10>; my ($x, $y) = 11, 11; my &ex = -> $x, $y { say "$x,$y"; %W{"$x|$y"} :exists; }; while not ex($x-1|$x|$x+1, $y-1|$y|$y+1) { say "bup" }
that's what i meant to write.
colomon TimToady: any idea why niecza's infix:<~> takes an arbitrary number of arguments?
timotimo see how niecza now passes the inner junctions into $x and $y and they are then not autothreaded
p6eval niecza v24-33-gfc6d5fd: OUTPUT«(timeout)any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)…
..rakudo 8ea992: OUTPUT«(timeout)any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(10, 11, 12),any(10, 11, 12)␤bup␤any(…
TimToady timotimo++ for actually reading the spec :) 04:09
timotimo er, wait, what?
rakudo show sthe same behavior? now i am thoroughly surprised.
maybe parameters to pointy blocks are Mu by default 04:10
rn: say (-> $x, $y { say $x, $y }).signature
p6eval rakudo 8ea992: OUTPUT«:(Mu $x, Mu $y)␤»
..niecza v24-33-gfc6d5fd: OUTPUT«Unhandled exception: Nominal type check failed in binding '' in 'ClassHOW.name'; got Mu, needed Any␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (ClassHOW.name @ 1) ␤ at /home/p6eval/niecza/lib/CORE.setting line 1193 (Parameter.perl @ 9) ␤ at /home/p…
timotimo ah, indeed!
TimToady colomon: ~ is list associative
colomon TimToady: ah 04:11
r: say (1|2|3) ~ (10|11|12) ~ (21|22|23) 04:12
p6eval rakudo 8ea992: OUTPUT«any(any(any(11021, 11022, 11023), any(11121, 11122, 11123), any(11221, 11222, 11223)), any(any(21021, 21022, 21023), any(21121, 21122, 21123), any(21221, 21222, 21223)), any(any(31021, 31022, 31023), any(31121, 31122, 31123), any(31221, 31222, 31223)))␤»
timotimo diakopter: the translation must be made up 04:13
i cannot explain it any other way
diakopter I strongly believe the translation is true
timotimo in that case, that's ridiculous 04:14
anyway, i'll go get my almost-daily cup of hot snow
diakopter that a country would give such ludicrous propaganda to its prisoners/victims? 04:15
mmm tasty snow 04:16
timotimo do they expect the people will believe this? or is this a kind of "throw feces enough and something will stick" situation? 04:17
diakopter likely most of them don't have any evidence to the contrary, so why not believe it 04:18
timotimo hmm, probably true
i may just be tired, but i couldn't find any tests that ensure that multiple junctions in the same function call actually work like the spec says 04:41
04:46 adu_ left 04:52 preflex_ joined 04:54 preflex left, preflex_ is now known as preflex 05:05 census joined 05:20 SamuraiJack joined, adu joined 05:35 adu left, adu joined 05:36 adu left, adu joined 06:12 zacts left 06:23 census left 06:24 adu left 06:34 adu joined 06:41 adu left 06:46 am0c joined 07:00 raiph left 07:02 raiph joined 07:13 kaare__ joined 07:16 wk joined 07:18 jaldhar_ left 07:20 jaldhar_ joined 07:24 xinming left 07:26 xinming joined 07:34 FROGGS joined 07:35 jaldhar_ left 07:38 jaldhar_ joined 07:39 SamuraiJack left, SamuraiJack joined 08:19 Myk267 joined, isomorphisms left 08:23 Myk267 left 09:05 wk left 09:07 SamuraiJack_ joined, SamuraiJack left 09:15 _ilbot left 09:17 Khisanth left 09:19 donaldh left 09:21 donaldh joined 09:23 _ilbot joined 09:24 kst joined 09:32 Khisanth joined 09:44 spider-mario joined 09:51 DarthGandalf joined
masak good antenoon, #perl6 09:57
FROGGS hi masak 09:58
masak <zacts> is perl6 usable now for everyday projects? 10:01
this went unanswered in the backlog.
diakopter masak: HI 10:07
masak: zacts is still logged in, just not on this channel :)
masak also, I had no idea America was in such a bad place. 10:08
good thing that the trees are still full of snow to heat.
10:10 jac50 joined
sorear are you referring to the storm system on the other side? 10:10
masak I'm referring to irclog.perlgeek.de/perl6/2013-03-10#i_6570639 10:11
seriously, the North Korean who put that together seems to have an unhealthy fixation with heating snow.
s/with/on/ 10:12
diakopter masak: yes, it's terrible.
there are some classic lines in there
masak "They will be eaten on Tuesday." being one of them? :)
diakopter truly poetic
sorear heating snow? is that what they use in the DPRK because they can't afford coal? 10:13
diakopter sorear: watch the video to join the in-joke :) 10:14
it's short.
masak there's a parallel here to language communities, I guess. somewhere out there I bet there's a community going "yo, we've got features coming out the wazoo an' all the other languages have feature envy, or should" while all the rest of the communities go "dudes, you haven't even stabilized your I/O spec yet" 10:15
diakopter it's so absurd it's difficult to distinguish from parody
jnthn "The other parts of America, often disguised as foreign countries in Europe..."
masak :P
oh, *that's* what he said. 10:16
imagine the work that must have gone into suspending disbelief *prior* to this video.
that's almost more scary. 10:17
FROGGS they are yummi >.<
*g*
hehe
jnthn "These telephones no longer work. There is noone to call"
masak FROGGS: yeah, but there's only one cup of hot snow per person per day.
FROGGS ya, hehe
diakopter jnthn: ROTFL.
masak at least the poor Americans get cake in the end by the kindly North Koreans. 10:18
masak realizes that a single episode of "Friends" would put the whole video into question 10:20
huf 6 idle rich landowners having fun in a cafe all day?
FROGGS masak: but "friends" is just made up
FROGGS makes more hot snow 10:21
diakopter makes yellow snow
FROGGS O.o
eeeeks
masak huf: oh, true. death to the bourgeoisie. 10:22
diakopter ownership is overrated
masak there's probably even enough bums in the series to still make the hot-snow video credible.
diakopter also Seinfeld 10:23
huf masak: now you're talking! :)
masak and Phoebe used to live in a cardboard box.
huf masak: that's just her being excentric
the rich have strange tastes 10:24
sorear
.oO( the moon of Saturn? )
most commonly spelled 'eccentric' around here
masak yes, that's the spelling, everywhere :)
huf huh. my dict lied to me then. 10:25
masak oh, br "excentric".
my bad :)
diakopter reminds me of execrable
masak that word seems to be trending on Twitter recently. 10:26
10:30 jac50 left
diakopter anyone here have access to springerlink? 10:30
FROGGS hmmm, it answeres the question as well if I dont know what it is, I guees :/ 10:32
10:33 fhelmberger joined
masak clearly it has something to do with knights :) 10:34
or links to knights.
or left of knights.
10:35 fhelmberger left
FROGGS or about somebody who is bad to knights 10:35
arnsholt diakopter: Via my university, yeah
10:36 huf left
masak rn: say 0 ** 0 10:42
p6eval rakudo 8ea992, niecza v24-33-gfc6d5fd: OUTPUT«1␤»
10:46 huf joined 11:01 am0c left
grondilu about the new List permutation method: why is it not a Positional method? 11:02
(so that one can get it when implementing Positional)
jnthn Generally, .foo => .list.foo delegators live in Any
masak it's not really "feature envy", it's more like "feature skew". 11:04
jnthn urks 11:05
NQP has support for BUILD. However, since there are no submethods in NQP, one BUILD is seen for all subclasses. 11:06
Which is why the dupe SC registrations...
masak :(
maybe make it a sub? 11:07
jnthn Just gonna make BUILDPLAN only look for methods in exactly that class.
masak *nod* 11:08
jnthn Bug exists in NQP on Parrot too, it just doesn't detect dupe SC registration for some reason.
11:08 crab2313 joined
jnthn Guess I musta just forgot to add the check... 11:08
arnsholt One of the benefits of multiple backends 11:12
masak .oO( port to another platform! you'll find bugs, we promise! ) 11:17
11:19 isBEKaml joined
dalek p: e408df4 | jnthn++ | src/how/NQPClassHOW.pm:
Lookup of BUILD shouldn't see inherited ones.

Doing it this way since NQP has no submethods.		 11:19
p-jvm-prep: ecd78f9 | jnthn++ | nqp-src/nqp-mo.pm:
Get BUILD fix.
 11:20
Ulti diakopter: I have access to springer link afaik 11:23
isBEKaml OHHAI, #perl6! 11:36
11:36 dayangkun left
masak o/ 11:39
isBEKaml masak: You should vacation more often. Your bug-finding-productivity shot up in the first week! :P 11:41
11:45 not_gerd joined
not_gerd o/ 11:45
11:46 wk joined
masak isBEKaml: I know! and yes, I *should* vacation more often! :P 11:48
masak gets on a train 11:49
12:00 woosley1 joined
masak teaching Perl for two days :) 12:22
spider-mario :) 12:23
masak "Let's not forget the things that are good about Perl. It's good at interacting with other programs, and it's good for rapid prototyping. Let's not hassle people when they use Perl the way it was designed to be used." -- perl.plover.com/yak/12views/samples/notes.html 12:25
isBEKaml that's MJD, right? 12:26
masak yes. 12:27
MJD++
isBEKaml ah, yes - I liked his HOP book. :)
12:31 SunilJoshi joined
masak the whole of Talk 4 can be obsoleted with one word nowadays: Moose. 12:31
tadzik hmm, there's nothing like AttributeHOW in the metamodel, is there? 12:36
masak "I went back to the original Christopher Alexander books and found out why: pattern languages are a brilliant idea and Alexander's books about them are works of genius." -- wow. I really should do this.
isBEKaml masak: yeah, Cristopher Alexander is referred to in books on Design Patterns as well. It's been on my list for a /long/ time. :( 12:37
(of books to read) 12:38
tadzik ph<TAB><TAB> 12:39
phenny, oh, where are you now :(
anyway, I added rosettacode.org/wiki/Documentation#Perl_6 just now :)
not_gerd just fixed sockets
now, if only someone would take on irclog.perlgeek.de/perl6/2013-03-09#i_6569852
masak tadzik: s/as it happened in Perl 5/as in Perl 5/ 12:41
Perl 5 is not gone, and Perl 6 is not replacing it ;)
tadzik ohh, I really wrote that? :D
fixed, thanks 12:42
12:46 bowtie joined 12:47 cognominal joined, SunilJoshi left
jnthn tadzik: Tehre's not an attributeHOW, thething is just called Attribute :) 12:58
13:00 isBEKaml left 13:01 cognominal left 13:04 SunilJoshi joined 13:08 SunilJoshi left 13:21 not_gerd left
masak rn: say pick 1..5 13:26
p6eval rakudo 8ea992: OUTPUT«Nil␤» 13:27
..niecza v24-33-gfc6d5fd: OUTPUT«␤»
colomon rn: say (1..5).pick 13:28
p6eval rakudo 8ea992, niecza v24-33-gfc6d5fd: OUTPUT«2␤»
tadzik jnthn: yeah, I noticed I still haven't fixed WHY there 13:29
it does not segfault now, but the test fails
also: anyone willing to share a cheap room for YAPC::EU?
grondilu tadzik: shouldn't you have used #| instead of #= at the beginning of blocks? 13:34
grondilu talks about rosettacode.org/wiki/Documentation#Perl_6 13:35
13:35 cognominal joined, PacoAir joined
tadzik grondilu: not if I'm talking about working examples :> 13:35
grondilu oh, ok
tadzik I still haven't stopped feeling uneasy about #| :) 13:36
is it spec already?
grondilu yeah, it has been for quite a long time, hasn't it?
tadzik heh, maybe 13:41
13:47 census joined 13:51 census left 14:00 zb left
masak "The only thing more amazing than the degree to which the author missed the point of the exercise here is the degree to which he seems unaware that he missed the point of the exercise." :D 14:02
14:08 jlaire joined
masak "After reading a lot of articles, I discovered that different people had at least eight different notions of what 'strongly typed language' meant. " -- I *knew* it! 14:09
14:13 cognominal left 14:17 census joined 14:19 REPLeffect joined
FROGGS tadzik: rosettacode.org/wiki/Documentation#Perl_6 # say say? 14:21
14:22 Psyche^ joined
masak rn: sub x { return for 1 }; x; say "alive" 14:24
p6eval rakudo 8ea992: OUTPUT«Attempt to return outside of any Routine␤ in block at src/gen/CORE.setting:489␤ in block at /tmp/__4yt4NbsM:1␤␤»
..niecza v24-33-gfc6d5fd: OUTPUT«alive␤»
masak rakudo, what in all the...
masak refuses to swear very loudly, to avoid a long discussion with jnthn about the excellent stability of Rakudo in 2013 :( :( :( 14:25
14:25 domidumont joined
masak submits a rakudobug 14:25
14:25 Patterner left, Psyche^ is now known as Patterner
FROGGS rn: sub x { return 2 for 1 }; my $three = x; say "alive" 14:26
p6eval rakudo 8ea992: OUTPUT«alive␤»
..niecza v24-33-gfc6d5fd: OUTPUT«Potential difficulties:␤ $three is declared but not used at /tmp/VrWBbBK_BI line 1:␤------> sub x { return 2 for 1 }; my ⏏$three = x; say "alive"␤␤alive␤»
FROGGS masak: so it's sink context? 14:27
masak good catch. yes, must be.
FROGGS new features always come with new bugs bundled :o)
masak the RT ticket will not contain any bile against Rakudo, by the way.
I vent here, not on RT. 14:28
rn: sub x { return }; x; say "alive" 14:29
p6eval rakudo 8ea992, niecza v24-33-gfc6d5fd: OUTPUT«alive␤»
14:30 cognominal joined
FROGGS might be good to assign it to moritz 14:30
masak feel free.
r: sub x { return 5 for 1 }; say x 14:32
p6eval rakudo 8ea992: OUTPUT«Attempt to return outside of any Routine␤ in block at src/gen/CORE.setting:489␤ in block at /tmp/QZnjtHvdjL:1␤␤»
masak no, it's not sink context.
masak adds to the ticket 14:33
FROGGS hmmmm
tadzik FROGGS: whoops whoops
FROGGS tadzik: :o)
tadzik fixed, thanks :) 14:34
14:35 mtk left
masak should I unassign moritz? :) 14:35
better not. something tells me that even if this isn't directly related to sink context, maybe it's indirectly related. 14:37
14:38 mtk joined
masak decides to do a bisect 14:38
FROGGS nom: sub x { return 5 for 1 }; say x
p6eval rakudo 8ea992: OUTPUT«Attempt to return outside of any Routine␤ in block at src/gen/CORE.setting:489␤ in block at /tmp/sZuHsqRSJK:1␤␤»
FROGGS masak++ 14:39
masak using a while loop is a decent workaround, by the way.
14:43 not_gerd joined
not_gerd r: sub x { sink return 5 for 1 }; say x 14:43
p6eval rakudo 8ea992: OUTPUT«5␤»
not_gerd looks like sink context to me 14:44
14:44 pmurias joined
pmurias jnthn: hi 14:44
not_gerd I believe it's caused by the fact that rakudo doesn't do lexotic binding of the return to the sub
that's bitten me with gather/take as well 14:45
masak adds this to the ticket
pmurias jnthn: what would be the best way to compile a single file to both javascript and pir (under the new backend seperation scheme)
jnthn sub x { return for 1 }; x; say "alive" 14:47
So, the for is really a map and it's not in sink context.
So the return executes afterwards. 14:48
not_gerd lexotic return should probably force sink context
jnthn ...what? 14:49
After Rakudo got sink context and people discovered these things, something got spec'd to the degree it should tell you about this.
not_gerd if I understand this lexotic stuff correctly, if things like take or return are lexically surrounded by stuff that catches them (gather/sub), they should bind to that
jnthn As for whining about stability, well, it's easy to be stable if we ignore the spec.
not_gerd if the return is inside a lazy map, that probably requires making it eager 14:50
masak I am whining. I think this sucks.
jnthn Then go whine about the spec.
masak yes. 14:51
maybe I simply really dislike this type of unexpected laziness.
clearly, *clearly* it was not my intent as a programmer, to return outside of the routine! arrrgh!
jnthn masak: github.com/perl6/specs/commit/e5b8...08ca9d220a
That's the bit added in response to pepole running into the thing you mentioned. 14:52
Especially the last paragraph.
not_gerd jnthn: see S04 - return is lexotic and should target the sub it's lexically surrounded by 14:53
jnthn not_gerd: BUT WE ALREADY LEFT THAT SUB.
masak "...which may surprise some people." -- YA THINK!?
not_gerd jnthn: that's why I think it needs to force eager evaluation of the map 14:54
masak I swear, I was just idly building my program, and suddenly the for loop was last in the routine.
it was literally a "hey, I can move this return inside of the for loop" moment.
masak glares on the spec
not_gerd lexoticity and least surprise > laziness
pmurias combining laziness and side effects can be expected to lead to craziness ;) 14:56
masak rn: sub x { return 5 for 1; return "ahahaha *sob" }; say x; say "alive"
p6eval rakudo 8ea992, niecza v24-33-gfc6d5fd: OUTPUT«5␤alive␤»
masak adds "always put an unnecessary 'return' after the last 'for' in a routine" to his list of best Perl 6 practices 14:57
not_gerd personally, I'd just make for eager by default 14:58
but what do I know ;)
masak yes, I expect for loops to be eager. 14:59
and it doesn't seem to be just TimToady who likes to indulge in the notion that 'for' and 'map' are exactly synonymous. 15:00
but when a DWIM feature turns a simple refactor into a WAT, clearly something is wrong. 15:01
pmurias how is for being lazy a DWIM feature?
15:02 woosley1 left
jnthn Note that a for loop as the last statemnet of a block, even if you were to evaluate it eagerly rather than sinkily, will build up a list of results, which is potentially just going to be thrown away. 15:03
masak the DWIM in this case is "you get laziness for free, no extra typing needed!" 15:04
jnthn You could get rid of the whole issue by saying that for never produces anything, just Nil
masak +1
jnthn Except then you can't use for the write list comprehensions.
*to
Or do for
masak for some reason I'm fine with statement_mod for yielding results. 15:05
but for the block form to do it feels wasteful and weird.
jnthn Anyway, I'm tired of this whole topic. Rakudo is currently closest to the current spec. If the spec on this changes, I'll be happy to work on changing things to be in line with the updates. In the meantime, the ticket just wants to say that Rakudo shoulda warned (as per current spec) and needs to implement the --> Nil thing. 15:07
masak thank you for your time. sorry about being grouchy when things break my code. 15:08
15:08 woolfy joined, am0c joined
jnthn not_gerd: Does the fix for #942 risk utf-8 chunks getting split, or is that handled further up? 15:11
timotimo jnthn: the qregex test asks for the QRegex::P6Regex compiler; there's a NQPP6QRegexJVM.class, but i can't quite figure out what exactly i have to getcomp to make that work :| i've put use NQPP6QRegex at the top and have tried lots of different combinations of the parts for the getcomp, but none have worked so far
jnthn timotimo: I believe it's registered under the same name 15:12
timotimo oh, ok
jnthn not_gerd: I'm guessing it's just at bytes at this level, so probably it's OK..
timotimo doesn't seem so. at least i'm still getting null pointer exceptions which seem to come from the getcomp returning null or nil or something 15:13
not_gerd jnthn: I don't *think* so - if I recall correctly, we do a partial decode and determine the bounds from that
timotimo scratch that, that wasn't actually correct.
not_gerd (bounds of valid data in buffer, that is)
jnthn not_gerd: OK
not_gerd: Once that gets into Parrot then it's probably worth doing a version bump and seeing if it helps. 15:14
timotimo in fact, my $match := $rxsub($cursor).MATCH; seems to be the line that causes the NPE
jnthn timotimo: Hmm.
timotimo so maybe it's the compilation step that failed, because that woul dhave set $rxsub? or maybe the cursor isn't right ...
not_gerd jnthn: it does fix HTTP::Easy for me
but see the XXX in the rakudo patch 15:15
the problem is that Parrot's readline does not guarantee to actually read a line for sockets
timotimo uh, i added a debug say statement and now the tests run (and fail) ... ?!
not_gerd imo that's a problem with parrot and should be fixed there
however, as whiteknight actually implemented logic to break out of the loop that reads the data, I'm not confident I'm aware of all the issues involved 15:17
jnthn timotimo: I'm wondering if the bug is the same NPE as trying to run 49-regex-interpolation.t gives
timotimo: Does the error you see mention islist anywhere? 15:18
15:19 snuffles joined
timotimo nope, just lots and lots of invoke and invokeInternal 15:19
jnthn hmm 15:20
timotimo when you said it's registered under the same name, you did mean to tell me to run "getcomp('QRegex::P6Regex'), right? 15:21
FROGGS jnthn: when I get a Regex passed to INTERPOLATE, its .NFA gives me an RPMCA of RPMCAs of integers, is that right? 15:22
jnthn timotimo: Yeah 15:23
timotimo how do i safely figure out if the things like $rxsub and $cursor have sane contents?
jnthn timotimo: You can see it being registered under that name at the bottom of NQPP6Qregex.pm
Just checked that bit
Well, can do $obj.HOW.name($obj) to check they're the expected kind of thing 15:24
FROGGS: Yes.
timotimo mhm
jnthn FROGGS: That's a thing you can pass into merge_substates
FROGGS jnthn: really? in the other case I pass QAST::Regexes 15:25
jnthn FROGGS: No, those are passed to add_node, iirc 15:26
timotimo: Something odd...islist does a null check...
FROGGS jnthn: ahh right, I passed the ints to add_node too, my fault 15:27
jnthn timotimo: hm, huh...a CodeRef with a null STable...
FROGGS: Ah, that explains it ;)
timotimo say($rxsub.HOW.name($rxsub)); - gives me a null pointer exception 15:28
FROGGS hmmmm, I think I should write a cronjob that saves `git diffs` of nqp and rakudo every ten minutes 15:29
jnthn timotimo: Yeah, same underlying issue. 15:30
timotimo: Think I'm getting clsoe on it...
pmurias spec discussion aside, it's the error message for return called from outside the routine less than awesome?
timotimo neat :)
15:31 mberends left 15:32 cognominal left
masak pmurias: no, I think it's fine. what surprises me is that an ordinary return in an ordinary for loops can fail in the first place. 15:32
I understand the reason that it does, too. it has to do with iterators and laziness. 15:33
jnthn timotimo: Oh, hm.
timotimo: Seems that the actual issue is that an internals-y compilation unit object is leaking out where an invokable thingy is expected. 15:34
timotimo whoops, okay 15:35
15:40 sqirrel joined
FROGGS jnthn: cool, so nfa is working now for literals and Regexes. What should I do with assertions? do I have to make a regex to get its NFA or should I do something else? 15:40
jnthn FROGGS: Not sure what you mean...example? 15:41
15:43 ruoso joined
FROGGS rn: say "aaaaaaaaaaab" ~~ /<"a+b">+/ 15:43
p6eval rakudo 8ea992: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter < (must be quoted to match literally)␤at /tmp/LWE1tVBTMm:1␤------> say "aaaaaaaaaaab" ~~ /<⏏"a+b">+/␤Unrecognized regex metacharacter " (must be quoted to match literally)␤at /tmp/LW…
..niecza v24-33-gfc6d5fd: OUTPUT«===SORRY!===␤␤Unrecognized regex assertion at /tmp/JrfjZtEQ4J line 1:␤------> say "aaaaaaaaaaab" ~~ /<⏏"a+b">+/␤␤Parse failed␤␤»
FROGGS rn: say "aaaaaaaaaaab" ~~ /<$("a+b")>+/
p6eval rakudo 8ea992: OUTPUT«｢aaaaaaaaaaab｣␤␤»
..niecza v24-33-gfc6d5fd: OUTPUT«Use of Nil as a string␤ at /home/p6eval/niecza/lib/CORE.setting line 1342 (warn @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 449 (Nil.Str @ 4) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/lib/CORE.setting line 2870 (Match.gis…
jnthn There's nothing to decide in that case, though? 15:44
You jsut eval and run?
Are you thinking more like /<@(...)>/ style things?
FROGGS r: say "aaaaaaaaaaab" ~~ /<@("a+b", "\d", "b")>+/
p6eval rakudo 8ea992: OUTPUT«===SORRY!===␤Unrecognized backslash sequence: '\d'␤at /tmp/UXXU2kDe0A:1␤------> say "aaaaaaaaaaab" ~~ /<@("a+b", "\⏏d", "b")>+/␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-infix␤ pref…
FROGGS right
r: say "aaaaaaaaaaab" ~~ /<@("a+b", "\\d", "b")>+/
p6eval rakudo 8ea992: OUTPUT«｢aaaaaaaaaaab｣␤␤»
jnthn Is that not the same as /<@a>/ generally? 15:45
FROGGS jnthn: it is
jnthn timotimo: Think I've got a fix
FROGGS I just asks if there is a cleverer way instead of evaling an anon regex
timotimo masak: do you know when the recording of your talk at fosdem recently can be uploaded or if it already is? 15:47
jnthn FROGGS: No, I don't think so...I think it is a case of doing that and then looking at the NFA of what was produced. 15:48
FROGGS: I mean, technically I guess you could just compile it as far as the AST, try to extract the right bit of that to feed to the NFA builder, and so forth... 15:49
FROGGS: But I'm not sure it's worth it.
FROGGS: You'd have to then go and finish compiling the winner(s)
I think it's probably a premature opt. 15:50
15:50 kst` joined
FROGGS hmmmm, okay 15:51
15:52 kst left
masak timotimo: I have no idea. 15:52
timotimo: I try not to sit around waiting for conf talks to be uploaded. it tends to be a while :) 15:53
timotimo right :(
i think i remembered whoever recorded that said it was at that time unencoded, but otherwise ready to get uploaded
was it lizmat?
dalek p-jvm-prep: 79e176f | jnthn++ | lib/QAST/JASTCompiler.nqp:
A little re-org.
 15:58
p-jvm-prep: ef1be4e | jnthn++ | nqp-src/NQP (2 files):
Don't leak compunits out, in general.

Just give them to things that know what to do with them.
p-jvm-prep: ddf70f4 | jnthn++ | nqp-src/NQPP6QRegex.pm:
Fix handle uniqueness issue.
p-jvm-prep: 43862dd | jnthn++ | t/nqp/49-regex-interpolation.t:
8ac4de5 | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
jnthn heh, it lost two lines
timotimo: Hopefully those help
timotimo i'll test
16:00 am0c left, PacoAir left
timotimo yesterday was the first time i missed loop labels in perl6 16:00
i wanted to redo an outer for loop from inside a while loop and couldn't (at least not easily) 16:01
16:02 PacoAir joined, PacoAir left 16:03 PacoAir joined, census left
ruoso timotimo, you have both control and error exceptions in Perl 6, you can do it that way 16:05
timotimo that's "not easily" to me; i am trying to solve one of the problems on rosettacode and thus want to have pretty and short code :) 16:09
jnthn: that makes matches succeed, cool! 16:10
but now: stack overflow errors happen at some point
masak balanced brackets: gist.github.com/masak/5129165 -- I just seemed to need to get this out of my system for the past two weeks or so. :) 16:12
if anyone knows of a better way to generate uniformly random balanced-bracket strings, I'd be very interested to know.
timotimo interesting, there's a null pointer exception, too ... 16:13
16:15 rindolf joined 16:21 LlamaRider joined
timotimo jnthn: paste.ee/p/yfryu - does this make any sense to you? the first 20 regex tests succeed and this one fails 16:21
it seems like the NPE results from the call to dump_str on the match object 16:22
hm, maybe that's what it does if the regex doesn't match the string or something? 16:23
16:29 pmurias left 16:34 Chillance joined
jnthn timotimo: My best guess is that the CATCH block is hit and then something it does causes an exception. 16:35
timotimo: And given I didn't finish up the exception stuff yet, that has bad consequences. 16:36
timotimo oh, that may very well be. i'll investigate!
jnthn The top mention of it dying in what are telling though
timotimo yes, it spits out the null pointer exception and then enters the CATCH block a bunch of times, then gets the stack overflow errors 16:37
should i comment out all regex tests that cause this behavior as kind of a "todo" thing and commit the rest to the repo?
jnthn timotimo: You could do that, yes
timotimo i'll do that later today 16:38
jnthn timotimo: Or maybe see if it does make it into the CATCH block and see where it explodes in there
timotimo it probably jumps into the true branch and tries to call nqp::index(~$_, $expect_substr) 16:39
jnthn oh, I wonder if the ~$_ is to blame...
timotimo not sure what the $_ is though
jnthn The exception I guess...
timotimo could be, yes
jnthn nqp-jvm: try { nqp::die('omgz'); CATCH { say("here") } } 16:40
p6eval nqp-jvm: OUTPUT«here␤»
jnthn nqp-jvm: try { nqp::die('omgz'); CATCH { say(~$_) } }
p6eval nqp-jvm: OUTPUT«Exception in thread "main" ␤Exception: java.lang.NoClassDefFoundError thrown from the UncaughtExceptionHandler in thread "main"␤»
timotimo fwiw, when i comment out that code the tests all run (and of course that's not a fix at all)
jnthn nqp-jvm: try { nqp::die('omgz'); CATCH { say($_.WHAT) } }
p6eval nqp-jvm: OUTPUT«␤»
jnthn nqp-jvm: try { nqp::die('omgz'); CATCH { say($_.HOW.name($_)) } } 16:41
p6eval nqp-jvm: OUTPUT«BOOTException␤»
jnthn nqp-jvm: try { nqp::die('omgz'); CATCH { say(nqp::getmessage($_)) } }
p6eval nqp-jvm: OUTPUT«omgz␤»
timotimo should i use getmessage instead of ~$_?
jnthn You can try that, for sure. 16:42
timotimo when i just replace ~$_ with nqp::getmessage($_) i get null pointer exceptions every time i enter the catch block, followed by the catch block being entered 16:43
i'll be afk for a bit
16:43 cognominal joined
jnthn timotimo: When you get back, ^^ may help (still need the getmessage) 16:46
16:51 crab2313 left 16:54 SamuraiJack_ left, SamuraiJack__ joined 17:08 simcop2387 left
lizmat timotimo: thanks for the reminder about FOSDAM talks. 17:08
not all of them got recorded, or completely recorded 17:09
most of them are not quite in focus, and in HD
I'm wondering whether it makes sense to downgrade the video quality even further for downloadable / streamable experience 17:10
and the audio is not too good either. This has taught me that using a Nikon D800 is not really a good way of recording presentations 17:12
lizmat afk for most of the evening 17:14
17:19 simcop2387 joined, simcop2387 left, simcop2387 joined
timotimo oh, right 17:19
i'll check that change after a commute, jnthn :) 17:20
17:20 skids joined 17:21 PacoAir left
grondilu rakudo warns me about useless use of [] with metaoperator? 17:21
rn: sub infix:<m+> { ($^a + $^b) mod 13 }; say ^13 Z[m+] ^13;
p6eval niecza v24-33-gfc6d5fd: OUTPUT«0 2 4 6 8 10 12 1 3 5 7 9 11␤»
..rakudo 8ea992: OUTPUT«Potential difficulties:␤ Useless use of [] around infix op␤ at /tmp/AKRpun407P:1␤ ------> :<m+> { ($^a + $^b) mod 13 }; say ^13 Z[⏏m+] ^13;␤0 2 4 6 8 10 12 1 3 5 7 9 11␤»
17:22 PacoAir joined
grondilu using [] is not useless even when not necessary, is it? It can always be used for clarity. 17:22
jnthn std: say ^13 Z[m+] ^13;
p6eval std 86b102f: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/icKAMFYlcu line 1:␤------> say ^13 Z[m+⏏] ^13;␤ expecting any of:␤ infix or meta-infix␤ regex atom␤ semicolon list␤ statement end␤Parse failed␤FAILED 00:00 45m␤»…
jnthn oops
std: sub infix:<m+> { ($^a + $^b) mod 13 }; say ^13 Z[m+] ^13; 17:23
p6eval std 86b102f: OUTPUT«ok 00:00 46m␤»
colomon n: sub infix:<m+> { ($^a + $^b) mod 13 }; say ^13 Z[m+] ^13;
p6eval niecza v24-33-gfc6d5fd: OUTPUT«0 2 4 6 8 10 12 1 3 5 7 9 11␤»
jnthn grondilu: I incorporated that message from STD yesterday. It's not meant to trigger inside of a meta-op.
grondilu well, it does, apparently.
17:24 isomorphisms joined 17:26 not_gerd left 17:28 jac50 joined
grondilu r: say ^3 Z[+] ^3 17:30
p6eval rakudo 8ea992: OUTPUT«0 2 4␤»
jnthn huh, it only triggers it on a user-defined op? 17:39
17:39 James_ joined, James_ is now known as Guest71434 17:43 jac50 left 17:45 bbkr joined 17:49 Guest71434 is now known as jac50
masak rn: gist.github.com/masak/5129165 17:49
p6eval niecza v24-33-gfc6d5fd: OUTPUT«Unhandled exception: System.NullReferenceException: Object reference not set to an instance of an object␤ at Builtins.InvokeSub (Niecza.P6any obj, Niecza.Variable[] pos) [0x00000] in <filename unknown>:0 ␤ at Niecza.Kernel.ToComposable (Niecza.STable arg… 17:50
..rakudo 8ea992: OUTPUT«[[]][[]][[]]␤»
masak hm, I wonder what Niecza stumbles on there.
that's a weird error. reproducible locally. 17:51
rn: constant C = 41, 42, 43; say C[1] 17:54
p6eval niecza v24-33-gfc6d5fd: OUTPUT«Unhandled exception: System.NullReferenceException: Object reference not set to an instance of an object␤ at Builtins.InvokeSub (Niecza.P6any obj, Niecza.Variable[] pos) [0x00000] in <filename unknown>:0 ␤ at Niecza.Kernel.ToComposable (Niecza.STable arg…
..rakudo 8ea992: OUTPUT«42␤»
masak there it is.
17:55 cognominal left
jnthn rn: constant C = (41, 42, 43); say C[1] 17:56
p6eval niecza v24-33-gfc6d5fd: OUTPUT«Unhandled exception: System.NullReferenceException: Object reference not set to an instance of an object␤ at Builtins.InvokeSub (Niecza.P6any obj, Niecza.Variable[] pos) [0x00000] in <filename unknown>:0 ␤ at Niecza.Kernel.ToComposable (Niecza.STable arg…
..rakudo 8ea992: OUTPUT«42␤»
jnthn (just checking it's not a precedence discrepancy) 17:57
17:57 stevan_ left, stevan_ joined 18:02 cognominal joined, stevan_ left
dalek kudo/nom: 5963753 | jnthn++ | src/Perl6/Grammar.pm:
Turn off check giving false positives for now.
 18:06
kudo/nom: 8d510e3 | jnthn++ | src/Perl6/ModuleLoader.pm:
One pir:: => nqp::.
 18:09
18:11 census joined, marloshouse left 18:12 cognominal left 18:13 PacoAir_ joined 18:14 PacoAir left, PacoAir_ is now known as PacoAir
timotimo jnthn, were you going to make exceptions work soon? 18:15
jnthn timotimo: Finishing that lot up is fairly high on my todo list, yes. 18:16
(Probably next few days.)
timotimo neato
jnthn afk for a bit
timotimo maybe i will be able to get the regex tests in acceptable shape today 18:17
18:18 cognominal joined 18:21 lichtkind_ joined 18:23 pmurias_ joined 18:24 pmurias joined
pmurias jnthn: what is needed for automatic precompilation for rakudo & nqp? 18:24
18:24 pmurias_ left, lichtkind left 18:30 LlamaRider left
arnsholt Apparently the Rakudo dyncall library is "hardly maintained". That was news to me ^_^ 18:33
18:33 wk left 18:35 Targen left
jnthn pmurias: The tricky part is knowing that nothing you depend on has been invalidated before loading the pre-compiled version. It's not a Horribly Impossible problem, just needs care to get right. 18:36
pmurias: Then there's the "where do we store the pre-compiled things" which needs to be configurable. 18:37
pmurias: An environment variable specifying that is a good start.
arnsholt: Um, who said it was hardly maintained? o.O
arnsholt jnthn: lists.parrot.org/pipermail/parrot-d...07420.html 18:38
18:39 James_ joined
jnthn oh, parrot-dev... 18:39
18:40 James_ is now known as Guest40611 18:43 jac50 left
arnsholt tries to ask about it in #parrot 18:45
moritz o/ 19:00
19:00 PacoAir_ joined 19:01 PacoAir left, PacoAir_ is now known as PacoAir
arnsholt \o moritz 19:02
masak moritz! \o/ 19:03
"oh, parrot-dev" pretty much sums it up.
jnthn welcome back, moritz :) 19:06
pmurias jnthn: is it to be expected that defining HLL::Backend::JavaScript in a seperate file causes problems? 19:07
timotimo doesn't understand 19:08
pmurias jnthn: it's not seen, but HLLBackendJavaScript is
jnthn pmurias: Yeah, NQP's nested package handling is naive... 19:10
pmurias: Though not entirely sure why that case ain't handled... 19:12
19:15 woolfy left 19:20 xinming left 19:22 xinming joined 19:32 SamuraiJack__ left 19:37 wk joined 19:38 sqirrel left
pmurias jnthn: isn't it a similiar reason to why you use QAST::CompilerJAST instead of QAST::Compiler::JAST? 19:39
jnthn pmurias: Yes :)
bbkr star: use JSON::RPC::Client; JSON::RPC::Client.new( url => "bbkr.org:8080" ).locate('perlgeek.de').say # paid/free GeoIP database support :)
p6eval star 2013.02: OUTPUT«("area_code" => 0, "city" => "Fürth", "continent_code" => "EU", "country" => "Germany", "country_code" => "DE", "dma_code" => 0, "latitude" => "49.466702", "longitude" => "10.966700", "postal_code" => "90763", "region" => "Bayern", "region_code" => "02", "time_zone"…
19:39 jac50 joined
masak whoa. 19:41
we don't accept system calls and file I/O on p6eval, but arbitrary *network* calls are OK? :P
jnthn masak: Sure, they're going to somebody else's system :P 19:42
masak oh, indeed. 19:43
bbkr github.com/bbkr/GeoIPerl6 - please test on your machines if you have some spare time
19:44 Guest40611 left
bbkr so far OS X 10.7 and Ubuntu 12.04 works with paid and free (from repositories) databases 19:44
pmurias jnthn: what we would need is a better load_module in src/ModuleLoader.pm?
arnsholt bbkr: Sorry for being so slow to respond to your Zavolaj bug reports, BTW
My time on Macs has been greatly reduced recently =/
jnthn pmurias: If the problem is in global merging, yes 19:45
bbkr arnsholt: that's fine, I also didn't have much time for Perl6 lately :) 19:46
19:48 PacoAir_ joined, PacoAir left, PacoAir_ is now known as PacoAir
dalek p: a4fa7e4 | (Arne Skjærholt)++ | src/6model/reprs/VMArray.c:
Serialize and deserialize VMArray REPR data.
 20:00
p: b553620 | (Arne Skjærholt)++ | src/6model/reprs/VMArray.c:
Refactor get_pos_{int,float} logic.
p: e45782b | (Arne Skjærholt)++ | src/6model/reprs/VMArray.c:
Refactor null_pos and set_pos_{int,float} logic.
p: 7399985 | (Arne Skjærholt)++ | src/6model/reprs/VMArray.c:
Serialize and deserialize VMArrays.
jnthn arnsholt: Does your serialization format match what I did for VMArray on the JVM the other day, ooc? :) 20:04
jnthn figures it'll be pretty close :)
arnsholt That's a very good question! 20:05
repr_data is serialized as only the type object. The rest of it (bits and kind) is reconstructed from that 20:06
The object itself is ->elems and then the elements
Oh, and the serialization/deserialization loops branch on type for each element ATM 20:09
Not the most efficient way to do it, but saved me some implementation time. If performance in the serialization logic matters I can fix that once I get things working
20:20 PacoAir left 20:22 PacoAir joined 20:26 domidumont left
Ulti FROGGS: vimeo.com/61476623 sorry the quality is kind of low... the source video is better 20:30
20:33 kaare__ left
timotimo i get really weird behavior from the CATCH block, things like a pyramid of "not " before the "ok" :D 20:33
i think i shall wait for exceptions to work properly
FROGGS Ulti++ 20:35
20:39 James_ joined 20:40 James_ is now known as Guest14786 20:42 szabgab joined
szabgab good daytime 20:42
I have not been here for a while.... 20:43
20:43 jac50 left
szabgab rakudo: my $x = "text"; say "{ $x.WHAT }" 20:43
p6eval rakudo 8d510e: OUTPUT«use of uninitialized value of type Str in string context in block at /tmp/nV69nBrwtL:1␤␤␤»
Ulti FROGGS: the original video is here mattoates.co.uk/files/rakudo_git.mp4 works in VLC for me but not some other players :S 20:44
szabgab method calls don't work in interpolation any more?
timotimo hello mister gabo :)
szabgab: the problem is different, the value of WHAT is the type class and those are undefined 20:45
r: my $x = "text"; say "{ $x.WHAT.^name }" 20:46
p6eval rakudo 8d510e: OUTPUT«Str␤»
szabgab rakudo: my $x = "text"; say $x.WHAT
p6eval rakudo 8d510e: OUTPUT«(Str)␤»
arnsholt jnthn: What's the correct way to make an op create a sixmodel object?
szabgab timotimo: the thing is that I noticed this with some of the method calls in an application
after compiling rakudo for the first time in monthes 20:47
hmm, but I just wrote a simple example an it works there, strange, I need to look into those classes 20:48
timotimo well, feel free to paste it and i'll have a look (not that i'd be an expert or anything) 20:49
have you tried the debugger for rakudo yet?
it's quite spiffy
FROGGS Ulti: I only use VLC, so that's pretty fine :o) 20:52
thanks!
szabgab oh, I am using qq{} in the code and not ""
rakudo: my $x = "text"; say $x.WHAT
p6eval rakudo 8d510e: OUTPUT«(Str)␤»
jnthn arnsholt: allocate, then initialize 20:53
szabgab err. sorry
jnthn arnsholt: See any of the box opes, for example
*ops
szabgab class X { has $.name };my $z = X.new(name => 'Foo'); print " {$z.name} "; print qq{ {$z.name} } 20:54
rakudo: class X { has $.name };my $z = X.new(name => 'Foo'); print " {$z.name} "; print qq{ {$z.name} }
p6eval rakudo 8d510e: OUTPUT« Foo {X<-518084861>.name} »
szabgab so qq{ } is not the same as "" 20:55
?
jnthn szabgab: qq and "" imply the same thing
szabgab: Trouble is that you picked {...} as your delimeters
szabgab it used to work 20:56
jnthn Yes, Rakudo used to not follow the standard grammar properly there.
Or maybe STD had it the other way around once too...
szabgab ah, so it was incorrect back then?
jnthn Yeah
szabgab oh my
fixing the perl6maven.com generating code
...
thanks! 20:57
jnthn I'm pretty sure Rakudo has it as STD does, since I recall discussing the rationale for it being this way with TimToady++ as I worked on it.
timotimo szabgab: yeah, your perl6maven posts appeared in my reader and the links were all broken :( 20:58
jnthn You can just pick a different quoter though
r: say qq[ abc { "omgz".flip } def ] 20:59
p6eval rakudo 8d510e: OUTPUT« abc zgmo def ␤»
szabgab timotimo: sorry or that :) now you know why ...
I am fixing it right now
jnthn: I go with []
timotimo jnthn: so is {} inside qq{} not working a bug in rakudo? i don't fully understand 21:00
arnsholt jnthn: Oh, right. I forgot about nqp.ops. No way to do it directly in NQP via QAST/Operations.nqp?
masak r: say qq{omg}
p6eval rakudo 8d510e: OUTPUT«omg␤»
geekosaur timotimo, as I read the above discussion, it's per spec
masak r: say qq{omg { 2 + 2 } wtf}
p6eval rakudo 8d510e: OUTPUT«omg { 2 + 2 } wtf␤»
masak timotimo: no, I think it's to spec.
jnthn timotimo: No, it's not a bug.
21:01 rindolf left
masak std: qq{omg { $variable } wtf} 21:01
p6eval std 86b102f: OUTPUT«===SORRY!===␤Variable $variable is not predeclared at /tmp/n03I4ldgLt line 1:␤------> qq{omg { ⏏$variable } wtf}␤Check failed␤FAILED 00:00 42m␤»
masak wtf :)
FROGGS hehe
masak oh right.
timotimo ok
masak yes, it still interpolates, of course.
jnthn yeah, if you take a look at STD's nibble
It does
|| <starter> <nibbler> <stopper>
|| <escape>
Meaning that a nesting of the delim beats and escape sequence. 21:02
Closures are an escape.
*an
The top line is why you can do stuff like 21:03
r: say q{ if $x = 42 { say "it's the answer" } }
p6eval rakudo 8d510e: OUTPUT« if $x = 42 { say "it's the answer" } ␤»
jnthn And the nested { and } are counted
21:04 bruges_ left
masak *nod* 21:05
21:06 bruges joined 21:10 Guest14786 left 21:18 not_gerd joined
not_gerd golfed my module compilation failure as far as I could 21:20
see github.com/gerdr/rakudobug if anyone wants to reproduce
21:23 PacoAir left 21:24 PacoAir joined
masak not_gerd: nice golf. 21:25
not_gerd: ooc, how much of the module hierarchy is needed?
does the bug trigger even without the PASM and/or PASM::Backend prefixes on the modules? 21:26
not_gerd masak: the hierarchy is necessary
masak both levels? 21:28
not_gerd I believe so, but feel free to experiment
masak it just seemed unlikely to my spider sense that *two* levels would be necessary to trigger the bug. 21:32
21:33 domidumont joined
not_gerd masak: I just tried again to remove either level, and both appear necessary 21:36
masak ok. 21:39
if I were you, I'd map out the order in which packages/classes are defined. might turn up something interesting. 21:40
jnthn not_gerd++ # putting the thing in a git repo, which'll make it easy for me to recreate
21:40 domidumont left
jnthn Does it have an acompanying RT? 21:40
masak like, sometimes when you include A::B, the A package will get a stub that can be replaced later.
jnthn I did fix a bunch of this stuff a while back. Thought I'd nailed these issues. :( 21:41
not_gerd module load order is consistent: gist.github.com/gerdr/5130590 21:45
I can create an RT ticket if so desired
jnthn not_gerd: Plesae do create the RT, and link to the github repo in it. 21:47
not_gerd: It's the best way to ensure it doesn't get lost.
I'll try and look at it during this week though.
timotimo if i want to build my own "is foobar", i'll add a multi trait-mod:<is>(Routine $a, :foobar), right? 21:48
jnthn _, not - 21:49
and :$foobar!
timotimo oh, ok
jnthn Make it required
timotimo why the $ though?
jnthn std: sub foo(:foobar) { } 21:50
p6eval std 86b102f: OUTPUT«===SORRY!===␤Malformed parameter at /tmp/PmFKsqmcyb line 1:␤------> sub foo(:⏏foobar) { }␤ expecting any of:␤ formal parameter␤ named parameter␤Parse failed␤FAILED 00:00 42m␤»
timotimo ah, duh :)
jnthn Because it's not valid syntax without ;)
21:51 szabgab left
masak maybe it should be... :) 21:53
std: sub foo(:$foobar) {}
p6eval std 86b102f: OUTPUT«ok 00:00 43m␤»
masak we admit sub foo(42), why not sub foo(:foobar) ?
not_gerd jnthn: just submitted as #117117 21:54
++jnthn for looking into it
colomon masak: it has to have a named parameter with that name, we don't care what the value is? 21:56
masak right. 21:57
jnthn std: sub foo(:foobar()) { }
p6eval std 86b102f: OUTPUT«===SORRY!===␤Malformed parameter at /tmp/GBmxKAeYUI line 1:␤------> sub foo(:foobar(⏏)) { }␤ expecting any of:␤ named_param_term␤ new name to be defined␤ parameter␤ routine_def␤ signature␤Parse failed␤FAILED 00:00 42m␤»… 21:58
jnthn std: sub foo(:\foobar) { }
p6eval std 86b102f: OUTPUT«ok 00:00 43m␤»
jnthn not_gerd: thanks
'night, #perl6 22:03
diakopter o/ 22:04
colomon \o
22:04 PacoAir left
lue hello o/ 22:05
diakopter hi 22:06
masak lue! \o/ 22:07
timotimo i can use signatures in given/when, right? what's the signature for that? :| 22:12
i've tried ten things so far :|
diakopter (the_eleventh_thing)
not_gerd r: given \(42, :foo<x>) { when :($pos, :$foo!) { say 'works' } } 22:13
p6eval rakudo 8d510e: OUTPUT«works␤»
timotimo oh, a ":" 22:14
i've tried \ and |, but neither worked. i did not expect :. thanks!
masak no-one expects the colon inquisition! 22:15
not_gerd but everyone wants the colon
I don't think that's true for the inquisition 22:16
masak though maybe for Spanish... :)
timotimo hehe
r: given \(42, :foo<x>) { when :($pos, :$foo!) { say 'works'; say $pos } }
p6eval rakudo 8d510e: OUTPUT«===SORRY!===␤Variable '$pos' is not declared␤at /tmp/2S9Ne0gBs7:1␤------> :($pos, :$foo!) { say 'works'; say $pos⏏ } }␤ expecting any of:␤ postfix␤»
timotimo do i need a pointy block?
lue r: given \(42, :foo<x>) { when :($pos, :$foo!) { say 'works'; say $_[0] } } 22:18
p6eval rakudo 8d510e: OUTPUT«works␤42␤»
timotimo that's unfortunate :| 22:19
well, not really surprising though
masak I dunno. I could argue that one both ways.
it would be a great thing for pattern matching if that one actually *did* bind $pos in the block. 22:20
timotimo you mean you wouldn't want names to be clobbered by signatures?
masak right, that could be an argument against.
but it seems a weak argument to me.
timotimo that's what i intended to do
not_gerd r: given \(42, :foo<x>) { when :($pos, :$foo!) { (-> $pos, :$foo { say $pos })(|$_) } }
p6eval rakudo 8d510e: OUTPUT«42␤»
masak because the clobbering isn't really clobbering. it's lexical overriding.
timotimo ah, so a pointy block does make it work! (but very ugly :D)
lue may Signature<pos_var> ? 22:21
masak $^pos ?
make it a little shorter...
anyway, 'night, #perl6
lue r: sub foo($bar, :$bar) { say "I'm alive!" }; foo(1,:bar(2));
p6eval rakudo 8d510e: OUTPUT«===SORRY!===␤Redeclaration of symbol $bar␤at /tmp/xSY0IIybRR:1␤------> sub foo($bar, :$bar⏏) { say "I'm alive!" }; foo(1,:bar(2));␤»
timotimo 'night masak!
lue So I'd be fine with $_<pos> in timotimo's example code 22:22
not_gerd perhaps matching against a signature should do something useful with $/
or is $/ supposed to always be tied to regex matching? 22:24
lue not_gerd: I'd guess so. Also it wouldn't work if part of your when { } block does some pattern matching.
timotimo does something speak against the syntax when -> $foo, :$bar { ... }?
is that evil and terrible?
22:26 pmurias left
not_gerd you could also get rid of given/when and just do a multi dispatch on local subs 22:28
timotimo but i can't have anonymous multi subs ;) 22:29
lue r: anon multi sub marine($a) { "expect failure" };
p6eval rakudo 8d510e: OUTPUT«===SORRY!===␤Cannot use 'anon' with individual multi candidates. Please declare an anon-scoped proto instead␤at /tmp/b1hYn0kgO0:1␤------> ulti sub marine($a) { "expect failure" }⏏;␤ expecting any of:␤ postfix␤ s…
lue ?? /me is intrigued with anon proto ... 22:30
timotimo oh that's interesting
"Cannot put proto on anonymous routine" 22:31
22:31 cognominal left
lue r: anon proto sub foo(|) {*}; my &a = multi sub foo($a) { say $a }; my &b = multi sub foo($a, $b) { say $a + $b; }; a(5); b(4,5); 22:32
p6eval rakudo 8d510e: OUTPUT«5␤9␤»
timotimo but that's not anonymous, that's called foo!
lue the anon makes in anonymous
timotimo so you have to identify it with a name, but the name gets discarded? 22:33
lue perlcabal.org/syn/S02.html#Scope_declarators
It gives pretty names for errors and allows for recursive anonymous subs. 22:34
timotimo hmm 22:36
doesn't seem very helpful
22:38 cognominal joined 22:52 felipe left 22:53 spider-mario left
not_gerd you could also get rid of given/when and just do a multi dispatch on local subs 22:53
looks like there's an echo in here ;) 22:54
timotimo: the closest I got to 'anonymous' dispatch using multis is gist.github.com/gerdr/5130877 22:55
22:56 xinming left
not_gerd bye, #perl6 22:56
22:57 nyuszika7h is now known as 13WAAQHWO, not_gerd left
diakopter o/ 22:57
22:58 xinming joined
timotimo that is ugly :( 23:01
23:04 13WAAQHWO is now known as nyuszika7h 23:28 census_ joined
timotimo hm, how do i check that a junction turns into the right thing in a test case? :| 23:29
23:29 census left
timotimo i suppose i have to do some magic with Mu and such .. ? 23:31
but i thouht when i try to get the values out of a junction i'm doing it wrong
oh, i can .Str it! 23:32
rn: sub tp($a, $b, $c) { "$a $b $c" }; my Mu $res = tp("dog", 1|2, 10&20); say $res.Str; 23:35
p6eval rakudo 8d510e: OUTPUT«any(all("dog 1 10", "dog 1 20"), all("dog 2 10", "dog 2 20"))␤»
..niecza v24-33-gfc6d5fd: OUTPUT«all(any("dog 1 10", "dog 2 10"), any("dog 1 20", "dog 2 20"))␤»
timotimo that's good. 23:36
23:36 cognominal left
geekosaur something demorgan-ed wrong? 23:37
23:42 dayangkun joined 23:43 cognominal joined
timotimo no, not quite 23:43
it's just the rule that the leftmost & junction will be run before the leftmost | junction
so, i've added this test, should i let it fail with rakudo or add a todo fudge already?
23:44 cognominal left, cognominal joined
geekosaur I was commenting in the difference between niecza and rakudo 23:45
colomon timotimo: if you know if fails in rakudo, add a todo 23:49
timotimo good. 23:50
i may even add a fix soon-ish? who knows! :)
23:51 cognominal left
dalek ast: b88a9d9 | (Timo Paulssen)++ | S03-junctions/autothreading.t:
autothreading order tests (& before |)
 23:51
colomon timotimo++ 23:52

Powered by Cro and the Raku® Programming Language.

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

Thank you!