»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 diakopter on 14 April 2013.
adu maybe it's not fixed 00:27
00:36 thou joined 00:40 rindolf joined 00:45 adu left 00:46 hlin_laptop joined
lue rn: my @a = [1,2],[3,4]; say @a[0;0]; 01:13
camelia niecza v24-37-gf9c8fc2: OUTPUT«1␤»
..rakudo 71ea14: OUTPUT«WARNINGS:␤Useless use of constant integer 0 in sink context (line 1)␤1 2␤»
01:29 thou left 01:31 dayangkun joined 01:38 FROGGS__ joined 01:40 FROGGS_ left 01:56 berekuk left 02:22 lustlife joined 02:31 frdmn joined 02:39 orafu left, orafu joined 02:50 rindolf left 04:01 toddr joined 04:02 Ben_Goldberg left 04:04 toddr_ left
gtodd m: my @a = [1,2],[3,4]; say @a[0;0]; 04:15
geekosaur it's not "m", its "rn" (r, n; also "nr" will work, I think) 04:18
04:19 kaare_ joined, saxx joined
geekosaur (you can also toss in 'p' to try it in pugs as well as rakudo and niecza, although at this point pugs is pretty far behind the others) 04:19
lue
.oO(Keming strikes again!)
 04:22
geekosaur they look remarkably similar in the font I use, actually; I had to select it to verify
gtodd ha 04:23
are: my @a = [1,2],[3,4]; say @a[0;0];
doh!
autospellcheck
r: my @a = [1,2],[3,4]; say @a[0;0]; 04:24
camelia rakudo 71ea14: OUTPUT«WARNINGS:␤Useless use of constant integer 0 in sink context (line 1)␤1 2␤»
gtodd n: my @a = [1,2],[3,4]; say @a[0;0];
camelia niecza v24-37-gf9c8fc2: OUTPUT«1␤»
04:24 Psyche^ joined
lue looks to me like [0;0] is interpreted as a list of statements as opposed to a semicolon list like it should 04:25
TimToady yes, rakudo doesn't implement slice syntax yet
n: my @a = [1,2],[3,4]; say @a[1;*-1]
camelia niecza v24-37-gf9c8fc2: OUTPUT«4␤»
TimToady well, multi-dim syntax really 04:26
lue S09 has always been a weak point in rakudo :/ 04:27
04:28 Patterner left, Psyche^ is now known as Patterner
TimToady it probably wouldn't be too hard to at least do that, given rakudo already has ** binding 04:29
06:17 SamuraiJack joined 06:28 saxx left 06:39 adu joined 06:50 fgomez left
adu jnthn: are you awake yet? 07:11
it's not Str that's broken, it's try-CATCH that's broken 07:12
nqp: try { nqp::die("hello"); CATCH { say("# ERROR: $_"); } }
camelia nqp: OUTPUT«# ERROR: hello␤»
adu nqp-jvm: try { nqp::die("hello"); CATCH { say("# ERROR: $_"); } }
camelia nqp-jvm: OUTPUT«Exception in thread "main" ␤Exception: java.lang.NoClassDefFoundError thrown from the UncaughtExceptionHandler in thread "main"␤»
labster adu: jnthn isn't usually awake until about 2am 07:18
adu labster: well it's 3:20am for me 07:19
labster: what timezone are you?
FROGGS__ it is 9:19 a.m. at jnthn place
labster California. 12:20 here
adu so 11 jnthn-time
labster \o FROGGS__
FROGGS__ morning 07:20
07:20 FROGGS__ is now known as FROGGS
adu good morning 07:20
jnthn sent me on a wild goose chase :)
FROGGS adu: well, tracking it down to try+catch is good, you just need to fix that instead :P 07:21
adu I have no idea where CATCH is implemented
FROGGS well, grep for CATCH?
labster that's what I tend to do. 07:22
FROGGS it should be in the Actions and should produce some AST, maybe has its own op
and then you have to see how this op/AST is handled
i.e. to what it compiles to 07:23
adu but nqp-jvm t/nqp/44-try-catch.t passes
hmm
FROGGS weird
maybe this try/catch isnt available at that time? 07:24
adu I'm going to add some more test cases 07:26
I really don't understand this test file 07:29
try and catch are never used together
I see try blocks in some test cases
and I see CATCH handlers in other test cases 07:30
labster sounds like an excellent idea, adu.
Well, all of the fun stuff like IO::Path::Win32 is finally working in the core.
adu but I mean every language I know of requires try is paired with catch 07:31
FROGGS nqp-jvm: try { nqp::die("hello") }
camelia nqp-jvm: ( no output )
adu and I think that anyone migrating from those languages is going to use both try & catch without thinking about it
FROGGS nqp-jvm: CATCH { say($_); }
camelia nqp-jvm: ( no output ) 07:32
FROGGS I am a bit curious about that java.lang.NoClassDefFoundError
labster > IO::Path::Win32.new("C:\\foo\\bar\\").parent # yields IO::Path::Win32<C:\foo>
FROGGS labster++ 07:33
that's is a great plus for rakudo
adu t/qregex/01-qregex.t uses both try and CATCH, which is (I believe) why it's failing 07:34
why is CATCH uppercase? 07:41
labster it's a builtin block, like LAST or BEGIN 07:44
I'm not entirely sure how to go about the testing for IO::Path, though. The backend is all in IO::Spec now, which is ironically not specced. 07:46
adu sounds awk-inspired
labster I think I might need to add it into the spec, but I'm not sure. 07:47
adu: who got awk in my perl? 07:48
adu probably the designer 07:49
are module, class, grammar blocks? 07:50
FROGGS adu: CATCH is like a phaser, and all phasers are uc 07:52
labster ah yes, phaser was the word
r: my $g = anon grammar foo { }; say $g ~~ Block; 07:54
camelia rakudo 71ea14: OUTPUT«False␤»
labster I guess not.
adu r: my $g = anon grammar foo { }; say $g ~~ CompUnit;
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Undeclared name:␤ CompUnit used at line 1␤␤»
FROGGS labster: does it matter that the backend isnt specced? if the "frontend" is specced then this must work as described
adu: these are packages 07:55
adu r: my $g = anon grammar foo { }; say $g ~~ Package;
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Undeclared name:␤ Package used at line 1␤␤»
FROGGS grammar is more like a class then a module though
adu r: my $g = anon grammar foo { }; say $g ~~ Module;
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Undeclared name:␤ Module used at line 1␤␤»
labster I suppose so, but all of the tests are written for File::Spec already :P 07:56
adu labster: how can you use Block?
labster r: use Block;
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Could not find Block in any of: /home/p6eval/nom-inst/lib/parrot/5.2.0-devel/languages/perl6/site/lib, /home/p6eval/nom-inst/lib/parrot/5.2.0-devel/languages/perl6/vendor/lib, /home/p6eval/nom-inst/lib/parrot/5.2.0-devel/languages/perl6/lib, /home/p6ev…
FROGGS adu: what do you mean by "use a block" ?
adu r: my $g = anon grammar foo { }; say $g ~~ Block;
camelia rakudo 71ea14: OUTPUT«False␤»
adu like that 07:57
FROGGS r: my $g = anon grammar foo { }; g.parse( "abc" )
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Undeclared routine:␤ g used at line 1␤␤»
FROGGS r: my $g = anon grammar foo { }; $g.parse( "abc" )
labster r: my $g := { say 'hi there'; } ; say $g ~~ Block;
camelia rakudo 71ea14: OUTPUT«No such method 'TOP' for invocant of type 'foo'␤ in method parse at src/gen/CORE.setting:11162␤ in block at /tmp/sNBmcz3njr:1␤␤»
rakudo 71ea14: OUTPUT«True␤»
adu my question stemed from the use of Block as some builtin type, but I can't find any builtin type for packages 07:58
labster r: my $g := { say 'hi there'; } ; say $g ~~ Block; $g();
camelia rakudo 71ea14: OUTPUT«True␤hi there␤»
FROGGS r: my $g := { say 'hi there'; }; $g()
camelia rakudo 71ea14: OUTPUT«hi there␤»
FROGGS right
07:58 drbean left
FROGGS adu: my $g = anon grammar foo { }; creates an anon object of grammar (like class) foo 07:59
so it isnt a block, it is an object
so you can call methods on it, like parse
adu r: my $g = anon grammar foo { }; say $g ~~ Mu; 08:00
camelia rakudo 71ea14: OUTPUT«True␤»
adu yey it's a Mu
08:00 drbean joined, rindolf joined
FROGGS if you have a real block, you have to call it, like $g(), you will see that in the AST often... ( :op('call'), $block ) 08:00
adu: doc.perl6.org/images/type-graph-Cool.png 08:02
it doesnt tell you much that it is Mu 08:03
adu I know it stands for "most undefined"
what's Cool? 08:04
FROGGS I guess doc.perl6.org can tell you more about that then I can :/ 08:05
adu ah, Cool is the supertype for magic conversions 08:07
08:08 Chillance joined
labster Mu stands for 無. 08:12
08:13 sciurius joined 08:22 crab2313 joined
cognominal in nqp, what is the equivalent of : %h<+> := &int-ctor; It say First child of a 'bind' op must be a QAST::Var 08:29
what is the nqp:: pseudo function I should use?
diakopter %h{'+'} 08:30
sorear nqp: my %h; %h<+> := 2;
camelia nqp: ( no output )
sorear nqp: my %h; %h{'+'} := 2;
camelia nqp: ( no output )
sorear nqp: my %h; say(3) := 2;
camelia nqp: OUTPUT«Error while compiling block : Error while compiling op bind (source text: ":="): First child of a 'bind' op must be a QAST::Var␤current instr.: '' pc 48819 (src/stage2/QAST.pir:17874) (src/stage2/QAST.nqp:2988)␤»
sorear cognominal: that ought to work... 08:31
cognominal nqp: class A { my %h; %h{'+'} := 1 } 08:32
camelia nqp: ( no output )
08:33 crab2313 left
cognominal nqp: class A { my %h; %h{'+'} := -> $_ { $_ } } 08:34
camelia nqp: ( no output )
dalek : 1b8439e | (Tobias Leich)++ | / (2 files):
fixed q, qq and friends
 08:40
masak antenoon, #perl6 08:44
08:44 tgt joined 08:46 adu left
FROGGS hi masak 08:50
08:53 betterworld joined, berekuk joined 08:55 berekuk left, berekuk joined
lizmat good post morpheus, #perl6 08:56
masak and what a morpheus it was! 08:57
FROGGS ?
hi lizmat
masak FROGGS: "...is the god of dreams." en.wikipedia.org/wiki/Morpheus_(mythology) 08:58
FROGGS ahhh, yeah, of course
09:08 spider-mario joined 09:11 dayangkun left 09:14 kaare__ joined 09:15 kaare_ left
lizmat wonders where .WHAT is really specced 09:17
S06:2275 09:18
diakopter at one point it described it as macro-y
lizmat actually: S12:2275 09:19
2289: "These are all actually macros, not true operators or methods." 09:20
diakopter yes 09:22
09:22 sad joined, sad left
lizmat rn: { say $^x + $^x }(5) 09:23
diakopter I bet you search the clogs for WHAT and macro using google you can find more context
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«10␤»
FROGGS std: use v5; require q(./test.pl) # :o(
camelia std 86b102f: OUTPUT«===SORRY!===␤Unable to parse parenthesized expression at /tmp/elsoYIAf1n line 1:␤------> use v5; require q⏏(./test.pl) # :o(␤Couldn't find final ')'; gave up at /tmp/elsoYIAf1n line 1:␤------> use v5; require q(⏏./t…
lizmat would it make sense to add a pattern like "{ say $^x + $^x }(5)" as one of the examples at S04:693 ? 09:26
masak lizmat: yes, think so. 09:27
lizmat: re "These are all actually macros" -- I've yet to see exactly *how* they are macros. in some places in and outside the spec, "they are macros" means "they get special treatment during compilation, and have thunk-like behavior". 09:28
lizmat I thought so, as macros aren't really implemented yet, right 09:29
?
diakopter marco
lizmat but if macro's would have been implemented, would they be macro's? Or still something else? 09:30
dalek pan style="color: #395be5">perl6-examples: b3c75b3 | (L. Grondin)++ | rosalind/mmch-grondilu.pl:
rename mmch
09:31 kaare__ left
diakopter iirc, they were speculated as macros at the very beginning of designing 6model 09:31
dalek : 8dcbdfa | (Tobias Leich)++ | / (2 files):
we can haz summary!
pan style="color: #395be5">perl6-examples: d5f05a4 | (L. Grondin)++ | rosalind/grondilu-mmch.pl:
remove grondilu-mmch
diakopter I think the reasoning was so that the compiler could generate the right "WHAT" for things like primitive types 09:32
dalek ecs: 3acfeda | (Brent Laabs)++ | S32-setting-library/IO.pod:
update IO::Path and IO::FileTests, OS-specific IO::Path::*

rm $.path from filetests and define modified/changed/created/accessed add IO::Path::Unix, IO::Path::Win32, IO::Path::Cygwin
diakopter and exotic other things
09:33 tgt left
lizmat masak: +It's not an error to pass parameters to such a block either: 09:37
+
+ { say $^x + $^x }(5);
+
masak looks good. 09:38
FROGGS { say $^x + $^x }(5, 6); # but they must be used
r: { say $^x + $^x }(5, 6); # but they must be used
camelia rakudo 71ea14: OUTPUT«Too many positional parameters passed; got 2 but expected 1␤ in block at /tmp/STO4LxS8Ya:1␤␤»
masak r: macro twice($stmt) { quasi { {{{$stmt}}}; {{{$stmt}}}; } }; twice say "OH HAI" 09:39
camelia rakudo 71ea14: OUTPUT«OH HAI␤OH HAI␤»
masak lizmat: macros are implemented, and work.
lizmat: they're incomplete, and still buggy, but they're there.
diakopter they do some things, but not all the things
masak lizmat: to the extent *I* understand what macros are, I don't understand how .WHAT is a macro.
lizmat hehe… okokok :-) 09:40
I guess "aren't really implemented yet" was a gross overstatement
my apologies
FROGGS: good point
masak as a lazy git, I'm insulted! :P
09:41 tgt joined
masak lizmat: srsly though, no worries. my point was more that we *do* have some understanding of macros already. 09:41
lizmat: but the terminology in general, and the use of the term "macro" in particular, are sometimes very confusing.
lizmat FROGGS: but isn't that a more general issue? 09:42
masak: s/These are all actually macros/These actually all behave like macros/ ?
FROGGS lizmat: yes, but it was a bit surprising to me when I first come to that 09:43
lizmat ack
09:44 berekuk left
masak lizmat: I'd actually prefer if we left the word "macro" alone for that meaning. 09:44
lizmat: something like "These all actucally have BEGIN-time semantics" seems to capture it just as well. 09:45
(no-one goes around calling BEGIN {} a "macro", do they) :)
diakopter agree ^^ "something like.."
masak while we're on the topic, I think it's ridiculous to call infix:<&&> a "macro". 09:47
lizmat well, the chapter "Introspection" at S12:2271 will need some rewriting then, as the word "macro" occurs a lot in there
masak it's just a short-circuiting operator. if we start to call short-circuiting operators "macros", it's not because we understand macros well, it's because we understand short-circuiting operators badly :) 09:48
lizmat I will keep that in mind when starting to grok S12 09:50
meanwhile: masak/Froggs: +But as always, you must use them all:
+
+ # Syntax error: Too many positional parameters passed
+ { say $^x + $^x }(5,6);
+
masak "These are all actually macros, not true operators or methods." -- this sentence reveals another confusion. "macro" and "(true) operator" are orthogonal concepts. 09:51
r: say Macro.^mro
camelia rakudo 71ea14: OUTPUT«(Macro) (Routine) (Block) (Code) (Any) (Mu)␤»
masak r: say Sub.^mro
camelia rakudo 71ea14: OUTPUT«(Sub) (Routine) (Block) (Code) (Any) (Mu)␤»
masak r: say &infix:<+>.^name
camelia rakudo 71ea14: OUTPUT«Sub+{<anon>}+{Precedence}␤»
lizmat likes introspection
masak macros are routines. operators are subs are routines.
macros can be operators. 09:52
(hm, in which case they'd not be subs, but still routines) :)
r: macro infix:<!+!>($a, $b) { quasi { 42 } }; say 4 !+! 2
camelia rakudo 71ea14: OUTPUT«42␤»
masak r: macro infix:<!+!>($a, $b) { quasi { 42 } }; say &infix:<!+!>.^name
camelia rakudo 71ea14: OUTPUT«Macro+{Precedence}␤» 09:53
masak r: macro infix:<!+!>($a, $b) { quasi { 42 } }; say &infix:<!+!>.^mro
camelia rakudo 71ea14: OUTPUT«(Macro+{Precedence}) (Macro) (Routine) (Block) (Code) (Any) (Mu)␤»
masak had personally missed that '+{Precedence}' part
interesting.
cognominal In the doc, except in the macro section, macro seems to mean, "it looks like a method or a function but it is implemented otherwise" 09:54
masak yeah. which is a pity. 09:58
(because macros are fairly well-defined.)
macros are routines that execute as soon as their argument list has been parsed. they accept ASTs or program strings as paramters, and they emit ASTs or program strings. 10:00
10:00 berekuk joined
dalek ecs: 432411a | (Elizabeth Mattijsen)++ | S04-control.pod:
Additional examples in "Statement-level bare blocks"
 10:01
cognominal masak: but I interpret that, maybe wrongly, as "they should have been written as macro, if macro were available at the time it was written." 10:03
masak yes, and my point is that that's not true. 10:05
or if that's true, the mechanism by which it is true eludes me.
in some sense it feels like the burden of proof rests on the one claiming that .WHAT is a macro, to show an example implementation of it. because I don't see it. 10:06
diakopter masak: my point above was that if the compiler knows the variable or expression results in a value of native type, then it can't have a 6model .WHAT method called on it, per se... if the compiler is treating it as a more ethereal "variable"/"memory location" and not a scalar container (if it's able to assume as such for optimization purposes) 10:07
... so the macro would generate the thing that returned the right type for that primitive type 10:08
I suspect that was the motivation between the handwavey macroization
er, *behind 10:09
lizmat rn: my $a="WHAT"; say { 1 }.$a 10:10
camelia rakudo 71ea14: OUTPUT«No such method 'postcircumfix:<( )>' for invocant of type 'Str'␤ in block at /tmp/i_vlrVN1ld:1␤␤»
..niecza v24-37-gf9c8fc2: OUTPUT«Unhandled exception: Unable to resolve method postcircumfix:<( )> in type Str␤ at /tmp/a2r78Stc2A line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4300 (module-CORE @ 583…
lizmat according to spec, this should work. Which also means to me that .WHAT is *not* a macro
as this would be purely runtime, no? 10:11
diakopter well there was always going to be a WHAT method behind the scenes anyway on things that could have methods 10:12
lizmat rn: my $a="say"; { 1 }.$a 10:14
camelia niecza v24-37-gf9c8fc2: OUTPUT«Unhandled exception: Unable to resolve method postcircumfix:<( )> in type Str␤ at /tmp/zm2lCnveie line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4300 (module-CORE @ 583…
..rakudo 71ea14: OUTPUT«No such method 'postcircumfix:<( )>' for invocant of type 'Str'␤ in block at /tmp/mrZvm7Q8cW:1␤␤»
lizmat rn: { 1 }.say 10:15
camelia niecza v24-37-gf9c8fc2: OUTPUT«{ ... }␤»
..rakudo 71ea14: OUTPUT«Block.new()␤»
10:16 SamuraiJack left
lizmat brb 10:16
FROGGS hmmm, should be easy to fix that 10:21
rn: { 1 }."say"
camelia niecza v24-37-gf9c8fc2: OUTPUT«===SORRY!===␤␤Unsupported use of . to concatenate strings or to call a quoted method; in Perl 6 please use ~ to concatenate, or if you meant to call a quoted method, please supply the required parentheses at /tmp/EUU2bnMD48 line 1 (EOF):␤…
..rakudo 71ea14: OUTPUT«===SORRY!===␤Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.␤at /tmp/2w4O_MZmDP:1␤------> { 1 }."say"⏏<EOL>␤»
10:25 hlin_laptop left
lizmat rn: say { 1 }.WHAT; my $a="WHAT"; eval "say { 1 }.$a" 10:25
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«(Block)␤(Int)␤»
diakopter heh.
lizmat even shorter: 10:26
say { 1 }.WHAT; eval "say { 1 }.WHAT"
rn: say { 1 }.WHAT; eval "say { 1 }.WHAT"
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«(Block)␤(Int)␤»
lizmat rn: say { 1 }.WHAT; eval 'say { 1 }.WHAT' 10:27
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«(Block)␤(Block)␤»
lizmat aha, interpolation, of course… :-)
jnthn Afternoon, #perl6
lizmat jnthn++ ! 10:28
dalek : 866a4aa | (Tobias Leich)++ | / (2 files):
this should fix: require q("./test.pl")
 10:29
jnthn .WHAT is compiled go a low-level operation of the implementation's choosing. Same with .HOW. They aren't methods.
FROGGS hi jnthn 10:30
jnthn They're handled inside of the compiler. Some may say that makes them macros. I don't care if you call them magical unicorn sugar puffs.
o/ FROGGS, lizmat
lizmat ah, the infamous musp's ! 10:31
10:31 berekuk left
jnthn I suspect if the spec says "macro" about them it doesn't literally mean the CORE.setting has a macro WHAT or so. 10:31
lizmat jnthn: but they also exist as runtime method, no?
jnthn No.
Not at the moment, anyway. We could add them as methods but you'd barely reach them. 10:32
lizmat so anything like: $a="WHAT", { 1 }.$a would never work?
jnthn Well, that means "WHAT"({ 1 }) which means you're trying to invoke a string :)
lizmat (the spec seems to imply that they would)
10:33 woosley left
jnthn $a = "WHAT"; { 1 }."$a"() could work if we did also put method WHAT etc. somewhere. 10:33
Where does it say they should work, ooc?
If they do then they'll be method WHAT() { self.WHAT } :)
lizmat then I guess S12:2298 is in error 10:34
jnthn looking
10:35 tgt left
jnthn lizmat: That bit of S12 is talking about how to bypass the fact that .WHAT and friends are treated specially. 10:35
But I think that:
And if you don't know the method name in advance, you'd be using the
variable form anyway:
$obj.$somemeth
...could be clarified as:
10:36 tgt joined
jnthn $obj."$somemethname"() # where $somemethname is a method name 10:36
$obj.$somemeth # where $somemeth is a method object
Or so
diakopter rn: my $a="WHAT"; say { 1 }."$a"() 10:37
lizmat hmmm… ok
camelia niecza v24-37-gf9c8fc2: OUTPUT«Unhandled exception: Unable to resolve method WHAT in type Block␤ at /tmp/0lhf4RSQLo line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4300 (module-CORE @ 583) ␤ at /home…
..rakudo 71ea14: OUTPUT«No such method 'WHAT' for invocant of type 'Block'␤ in block at /tmp/slbEKeSkqY:1␤␤»
lizmat I had it shorter:
rn: say { 1 }."WHAT"()
camelia niecza v24-37-gf9c8fc2: OUTPUT«Unhandled exception: Unable to resolve method WHAT in type Block␤ at /tmp/blKnrV91Km line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4300 (module-CORE @ 583) ␤ at /home…
..rakudo 71ea14: OUTPUT«No such method 'WHAT' for invocant of type 'Block'␤ in block at /tmp/kMyWKBb4_u:1␤␤»
grondilu rn: say join ",", (my @)[0] = [^2], [<foo bar>] 10:38
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«0 1 foo bar␤»
lizmat wonders how to create a method object
grondilu was expecting "0 1,foo bar" 10:39
rn: say join ",", [^2], [<foo bar>]
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«0 1,foo bar␤»
grondilu see?
diakopter r: say (our method foo() { }).WHAT
camelia rakudo 71ea14: OUTPUT«(Method)␤»
lizmat rn: my $a= method { }; say $a.WHAT
ack
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«(Method)␤»
diakopter r: say (our method foo() { }).name 10:40
camelia rakudo 71ea14: OUTPUT«foo␤»
jnthn r: say Int.^find_method('abs').WHAT.say
camelia rakudo 71ea14: OUTPUT«(Method)␤True␤»
jnthn hm, I said too much... 10:41
FROGGS uhh, good that we arn't mafiosos :P 10:42
lizmat for now I'll suggest this change to S12:
- $obj.$somemeth
+ $obj."$methodname"()
+
+or:
+
+ $obj.$methodobject
jnthn Speak for yourself
grondilu perl6 flattens too much. Everytime I try to use lists of arrays, I get troubles :(
jnthn uh, wait, I mean... :)
FROGGS hehe
jnthn lizmat: +1
lizmat actually, also: 10:43
- $obj."WHERE"
+ $obj."WHERE"()
right?
jnthn Yes, pretty sure the () are mandatory
std: my $obj; $obj."WHERE"
camelia std 86b102f: OUTPUT«===SORRY!===␤Unsupported use of . to concatenate strings or to call a quoted method; in Perl 6 please use ~ to concatenate, or if you meant to call a quoted method, please supply the required parentheses at /tmp/aMvLxZvcte line 1:␤------> my $o…
jnthn r: my $obj; $obj."WHERE"
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.␤at /tmp/SlDP2yHIhH:1␤------> my $obj; $obj."WHERE"⏏<EOL>␤» 10:44
dalek ecs: dd82127 | (Elizabeth Mattijsen)++ | S12-objects.pod:
Clarafication/Expansion in examples of Introspection section
 10:45
11:05 kaare_ joined
masak yes, used to be $obj."method" worked, but it was deemed to confusable with Perl 5 concatenation, so the () were added, and any $obj."method" without the () in the spec is a fossil. 11:22
11:25 lustlife left 11:30 greg____ joined
lizmat S02:4215 probably needs this diff then: 11:34
- $foo $bar: # ILLEGAL -- use $bar.$foo for indirection
+ $foo $bar: # ILLEGAL -- use $bar."$foo"() for indirection
am not sure what to do with the next line though:
(foo bar) baz: 1 # ILLEGAL -- use $baz.$(foo bar) for indirection
grondilu rn: say join(":", (my @)[0] = [^2], [<a b>]), " differs from ", ((my @)[0] = [^2], [<a b>]).join: ":"
camelia rakudo 71ea14: OUTPUT«0 1 a b differs from 0 1:a b␤»
..niecza v24-37-gf9c8fc2: OUTPUT«Potential difficulties:␤ Useless redeclaration of variable @ (see line 1) at /tmp/f7f1xnreEu line 1:␤------> [^2], [<a b>]), " differs from ", ((my @⏏)[0] = [^2], [<a b>]).join: ":"␤␤0 1 a b differs from 0 1:a b␤»
11:35 tgt left 11:36 tgt joined
grondilu rn: (my @a)[0] = [^2], [<a b>]; say @a[0].join(":"); say join ":", @a[0]; 11:36
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«0 1:a b␤0 1 a b␤»
grondilu rn: my $a = [^2], [<a b>]; say $a.join(":"); say join ":", $a; 11:37
camelia rakudo 71ea14: OUTPUT«WARNINGS:␤Useless use of "," in expression "my $a = [^2], [<a b>]" in sink context (line 1)␤0:1␤0 1␤»
..niecza v24-37-gf9c8fc2: OUTPUT«0:1␤0 1␤»
grondilu rn: my $a = ([^2], [<a b>]); say $a.join(":"); say join ":", $a;
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«0 1:a b␤0 1 a b␤»
grondilu wtf 11:38
isn't $x.join: $sep supposed to be the same as join $sep, $x ??
masak good question. 11:42
11:42 PacoAir joined
masak there's different amounts of flattening going on there, but I'm still not convinced there should be a difference. 11:42
specifically, I don't understand the result of the second join. 11:43
grondilu rn: my $x = <foo bar>; say join ", ", $x; say $x.join: ", "; 11:46
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«foo bar␤foo, bar␤»
grondilu hum, I think I get it.
rn: my $x = <foo bar>; say join ", ", @$x; say $x.join: ", ";
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«foo, bar␤foo, bar␤»
grondilu makes sense 11:47
the method call version forces the evaluation as a list, while the function call version considers the argument as an array. Or something like that. 11:48
11:49 kaare__ joined, kaare_ left 11:50 greg____ left
grondilu or in ther words, when writing join $sep, $x, $x is only the first item of a list that could have had several. Like join $sep, $x, $y. In which we probably wanted $x ~ $sep ~ $y. So it really has a different semantic than $x.join($sep) 11:51
s/In which/& case/ 11:52
masak yeah. 11:53
that does make sense.
11:54 kaare__ is now known as kaare_
masak rn: my $x = <foo bar>; say join(", ", |$x); say $x.join: ", "; 11:54
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«foo, bar␤foo, bar␤»
masak there you go.
I don't think | works without the parens, though.
11:55 domidumont joined
masak rn: my $x = <foo bar>; say join ", ", |$x; say $x.join: ", "; 11:55
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«foo, bar␤foo, bar␤»
masak seems it does :)
grondilu I don't see why it wouldn't have.
masak sometimes I suffer from the ghosts of old NYI that have since gone away. 11:56
lizmat rn: my $x = <foo bar>; say (|$x) 11:59
camelia niecza v24-37-gf9c8fc2: OUTPUT«\("foo", "bar")␤»
..rakudo 71ea14: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&prefix:<|>' called (lines 1, 1)␤»
lizmat feels like a rakudobug to me
rn: my $x = <foo bar>; say |$x
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«foobar␤»
12:00 rking left
lizmat without parens it's ok 12:00
masak yes, that's what I remember happening.
12:00 woolfy joined
lizmat seems you got it backward, though, cause it's working *without* the parens 12:01
12:01 ozmq joined
lizmat and it isn't *with* the parens 12:01
jnthn Well, what do you expect it to do there?
jnthn isn't sure what the expected behavior of using | out of the context of an argument list is. 12:02
lizmat well, in either case its a prefix |, is it not?
I mean "say |$x" is a prefix | 12:03
or am I missing something here?
jnthn Yes, but it's not actually turned into a call to prefix:<|>. Rather, it's used by the thing constructing the AST for the call to know to mark the arg as flattening.
So that it actually does get flattened. 12:04
12:04 domidumont left
lizmat so it's more like a macro in that context? 12:04
jnthn If we're calling "soemthing the compiler handles" a macro, then I guess so :)
12:05 domidumont joined
masak lizmat: now you're just doing it to tease me. 12:05
lizmat pleads guilty 12:06
jnthn :D
masak .oO( the whole bloody COMPILER is "more like a macro"! )
jnthn masak: To be fair, a big thing that makes macros interesting is that the compiler invokes them ;)
masak I'm not disagreeing about that. 12:07
GlitchMr thedailywtf.com/Articles/Global_Spaces.aspx 12:09
my @space-char := map " " x *, ^Inf; 1;
lizmat rn: sub prefix:<|> { @_.join }; my $x = <foo bar>; say (|$x)
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«foo bar␤»
GlitchMr Reimplementing stuff from this website in Perl is so fun
Perl 6*
lizmat oddly enough, $x is already stringified before prefix:<|> is called 12:10
afk for a few hours& 12:12
masak lizmat: no, it's not. that's the wrong conclusion :) 12:15
rn: sub prefix:<|> { say @_[0].^name; @_.join }; my $x = <foo bar>; say (|$x)
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«Parcel␤foo bar␤»
masak it gets stringified the moment you call .join on it. and because it's a Parcel, it gets a space in-between its elements. 12:16
cognominal lizmat, jnthn, I am confused too about the not really a prefix:<|>. I remember trying using it outside arguments
12:19 felher left 12:35 ozmq left 12:58 xinming left 13:00 xinming joined, pecastro left, spider-mario left 13:02 pecastro joined, fgomez joined 13:04 Chillance left 13:10 Chillance joined 13:40 saxx joined 14:00 rking joined 14:01 saxx left, AMENDEES_ joined 14:15 felher joined
dalek : 1e60268 | (Tobias Leich)++ | lib/Perl5.pm:
we dont need that
 14:20
FROGGS jnthn: I'm in serious trouble atm. The way v5 inherits from Perl6::Grammar doesn't allow me to override the pod directives, this breaks everything that uses Perl 5 pod blocks... :o( 14:21
jnthn: Is there a better way to inherit from Perl6::Grammar? 14:22
jnthn FROGGS: Inheriting from Perl6::Grammar sounds wrong. 14:30
FROGGS: I'd expect inheriting from HLL::Grammar
FROGGS that's what I do too, but if it doesnt inherit from Perl6::Grammar I end up: P6opaque only supports type changes where the MRO of the original type is a suffix of the MRO of the new type 14:31
jnthn That sounds like you're trying to rebless the current cursor to do a language switch rather than just doing something like LANG does. 14:32
FROGGS well, token statement has these reblessing
I can remove it for a test though 14:33
jnthn I don't know you should be ending up in there.
I'd expect after you see use v5 you do a langauge switch
The statement rebless is when you tweaked. 14:34
(e.g. added a new operator)
The case you have is more like when we switch to the Q langauge to aprse a quote, or the Regex language to parse regexes. 14:35
14:35 PacoAir left 14:36 PacoAir joined
FROGGS I'm commenting out the reblessing in rakudo and v5 and see what happens 14:37
14:38 tgt left
FROGGS when hitting a quote there is no real language switching too, it just nibbles/quibbles use a different language 14:39
but I cant do that
hmmm 14:42
after doing that it uses the P5Actions, but still Perl6::Grammar 14:43
jnthn Regexes use a different langauge, though?
FROGGS so setting %LANG<MAIN> alone has no effect
jnthn: but for regexes you still know the delims 14:44
so you can use sibble/tribble/...
jnthn How do you know today when to stop parsing P5 and return to P6?
Looking at STD, it declares a new %*LANG and sets it up. 14:45
FROGGS well, use Module only affects the block it is in, so { perl6... { use v5; ... } } is just working 14:46
and if v5 is the top thing in a file we are clear too...
FROGGS is looking at STD now
jnthn FROGGS: HLL::Grammar defines a <LANG(...)> which is used for language switching. 14:47
Or is a way to do it
That is much closer to what you want, I think.
FROGGS hmmm, sounds good
:my %*LANG; <-- STD
and then it does a reblessing too
at least that is how I read line 1445 14:48
jnthn yeah, we can't do it that way.
It suspect it relies on a P5ish view of bless under the hood, given it's translating to P5.
Or something like that.
oh, in fact it's completely replacing the current curosr... 14:49
Anyway, <LANG(...)> is probably the way to go.
FROGGS ya, will test that now 14:50
thanks!
15:05 bwk joined 15:15 Bzek left 15:17 bwk left, bwk joined 15:21 bwk left 15:22 AMENDEES__ joined, AMENDEES_ left 15:29 tgt joined 15:43 pecastro left
timotimo should rakudo introspect comments to see if there are vim or emacs configs for tabstop size, so that $?TABSTOP can be set properly? :P 16:14
dalek pan style="color: #395be5">perl6-examples: 1c124a8 | (L. Grondin)++ | rosalind/eubt-grondilu.pl:
(rosalind) EUBT
 16:15
grondilu ^ this one was tough
FROGGS jnthn: okay, so I put my v5 in %LANG<Perl5>, and added a <statementlist=.LANG('Perl5', 'statementlist')>, looks good so far 16:26
jnthn FROGGS: Sounds righter :) 16:27
FROGGS btw, if we had a $GRAMMAR (like $ACTIONS), it would help too
$*GRAMMAR that is 16:28
so the Perl5.pm just sets this
16:41 SamuraiJack joined 16:50 frdmn left
timotimo grondilu: my @left = grep none(@picked), @species; - clever! surprised me that it would work tbh. grep does ~~ internally? 16:56
16:58 cognominal left
colomon timotimo: yes 16:59
16:59 cognominal joined
colomon timotimo: possibly not explicitly, but grep smartmatches. 16:59
17:03 crab2313 joined
grondilu timotimo: yes, grep does ~~ internally. I've been using this "grep none(@)" construct for a while now. 17:05
Well, a junction is supposed to be exactly like any scalar, so I thought I should be able to do grep $junction, @array as I would do grep $scalar, @array. 17:07
colomon r: (1..10).grep(3|4)
timotimo it could theoretically have autothreaded the grep call and returned a junction of grep results for each of the junctionpars
camelia rakudo 71ea14: ( no output )
colomon r: say (1..10).grep(3|4)
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Unable to parse expression in argument list; couldn't find final ')'␤at /tmp/BLVZAbque0:1␤------> say (1..10).grep(3|4⏏)␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-infix␤»…
timotimo r: say (1..10).list.grep(3|4\) 17:08
camelia rakudo 71ea14: OUTPUT«===SORRY!===␤Unable to parse expression in argument list; couldn't find final ')'␤at /tmp/R_ZCAJqupm:1␤------> say (1..10).list.grep(3|4\⏏)␤ expecting any of:␤ postfix␤»
timotimo r: say (1..10).list.grep(3|4)
camelia rakudo 71ea14: OUTPUT«3 4␤»
timotimo there we go
r: say (1...10).grep(3|4) # should be the same
camelia rakudo 71ea14: OUTPUT«3 4␤»
colomon must have stuck some sort of control characters in there. Colloquy--
timotimo i wonder if the 1..10.grep behavior is to spec
colomon rn: say (1..10).grep(3|4) 17:09
timotimo looks like an italic backslash to me
camelia rakudo 71ea14, niecza v24-37-gf9c8fc2: OUTPUT«3 4␤»
colomon there you go
geekosaur my client shows a control character, FS (0x1c) 17:10
colomon and mine shows nothing. :\ 17:12
geekosaur (I use an xchat hook to make control characters visible) 17:13
17:15 crab2313 left
timotimo i should really get a compose combination for the ␤ character 17:16
17:17 AMENDEES__ left 17:24 tgt left 17:28 gtodd left 17:43 spider-mario joined 18:10 census joined 18:12 sftp left 18:13 kaare_ left, sftp joined 18:32 grondilu left 19:03 domidumont left 19:15 sqirrel joined
cxreg so, niecza does concurrency using the threads provided by the CLR? 19:15
what is rakudo going to do on jvm in that space? anything? 19:16
tadzik I guess it'll use threads provided by JVM 19:17
19:20 tgt joined
nwc10 cxreg: I know that jnthn has thought about how to do Rakudo concurrency on the JVM in some detail, but I don't know much more than that. 19:22
but, first one needs Rakudo on the JVM :-) 19:23
19:35 kaare_ joined 19:57 kaare_ left 20:03 berekuk joined
dmol Why JVM, not Parrot? 20:07
FROGGS dmol: what exactly do you mean? 20:09
20:10 rindolf left
masak I think the short answer is that exploring Perl 6 concurrency on Parrot isn't a priority for anybody right now. 20:10
nwc10 partly, if I understood what I've been told, because Parrot's threading model is very different from the JVM's (and JVM and .NET are fairly similar). And because Parrot's threading implementation is very new, and nothing (else) is using it 20:13
flussence Threading is a solved problem on the JVM. It's easier to port Rakudo to that than to wait for Parrot to catch up. 20:14
lue (nor is it, as I've heard, nearly as easy as it would be on JVM, a platform that has threads for a while)
nwc10 so, it's easier to do JVM first - you know that any bugs you find are going to be in the Rakudo code, because someon else has already hit the JVM bugs
then, once Rakudo's threading support is solid, you come back and port it to Parrot
masak there's a disconnect here between Rakudo and Parrot devs, I find. Parrot devs say "what are you talking about, Parrot has threads, they're awesome!". pmichaud has repeatedly said "if that's so, then please show me how we can do hyper addition, `@a >>+<< @b`, using Parrot's primitives." 20:17
lue
.oO("We're not rakudo programmers! How are we supposed to create an example like that?")
 20:18
lue goes to actually read the PDD on threads, if it exists 20:20
dmol Thanks for explanation. I'm just curious about p6 world, and try to understand it's tendency.
20:21 sqirrel left
masak dmol: it is a fair question. 20:23
lue Parrot's ideas on concurrency are interesting (multiple concurrency models can be used at the same time), but I wonder how much of it is actually implemented. 20:29
20:31 jnthn joined 20:32 mathw left, _sri joined, mathw joined 20:34 frdmn joined 20:35 avar left 20:37 yx joined, Bucciarati joined
cxreg masak: did parrot ever respond with some examples? 20:41
tadzik rurban++ commited some code to nqp repo
it was never fully implemented though, I think
some things were lacking
20:41 adu joined, SamuraiJack left, _sri left, _sri joined, avar joined, avar left, avar joined 20:42 ilogger2 joined, ChanServ sets mode: +v ilogger2
dalek : 88cfd7d | (Tobias Leich)++ | / (5 files):
instead of inheriting from Perl6:Grammar [...]

we just execute a statementlist in the given language.		 20:49
20:50 p5eval joined
dalek : f1e39fb | (Tobias Leich)++ | t/test_summary:
strip backtrace
 20:54
[Coke] loses some backscroll, and can't hit the log website to find it! *sadface* 21:00
dalek rl6-roast-data: 9af76ea | coke++ | / (3 files):
today (automated commit)
rl6-roast-data: 4f577ad | coke++ | / (3 files):
today (automated commit)
adu hi all 21:01
masak hi adu
adu what's the difference between AST and QAST? 21:04
masak AST is language-independent, QAST is what nqp is using. 21:05
[Coke] AST is a generic ter.
timotimo masak: when reviewing the t2 results, did you think the producing grammar for pieces of speech has been implemented in an optimal way? i thought it was pretty concise and good; did you see any obvious improvements?
[Coke] m
timotimo ter.m? :)
adu doc.perl6.org/type/AST 21:06
masak timotimo: I did, but I didn't put it in the review.
timotimo give me the short version of what you thought could be done better? :)
masak timotimo: it occurs to me that the grammar will not generate all possible trees with equal probability.
timotimo: (because it works top-down, and so the distribution will be biased because of that) 21:07
timotimo oh, that's a good point. reminds me of the time when the minichallenge to make random brace pairs was in the air
masak yeah, that one is clearly related.
random brace pairs and binary trees are equal up to isomorphism. 21:08
adu random brace pairs?
dalek ast: 48308c5 | coke++ | S12-construction/destruction.t:
pugs fudge
 21:09
timotimo rosettacode.org/wiki/Balanced_brackets - i think this one
masak yes, but that's one about verification, not random generation.
cognominal can someone tell me the difference between LALR and operator precedence parsing? I think that LALR is faster because it is compiled into an automata but operator precedence is more general when combined with other engines like in Perl 6
can someone correct me or elaborate? 21:10
masak adu: gist.github.com/masak/5129165
cognominal: "an automaton" ;) 21:11
that was the correction -- hold on and I'll try to elaborate.
cognominal: LALR seems more comparable to recursive descent than to operator precedence. 21:13
cognominal: and Wikipedia's summary seems to be "LALR is hard to understand, recursive descent is easier".
cognominal mask: well with recursive descendant parsing in Perl 6, once you start using procedural alternatives, there is so many of them 21:14
I am writing material and I said that with reduce/reduce and shift/reduce conflicts, LALR was tricky to use. 21:15
masak ah, ok.
yes, maybe that's where the trickiness comes from. I can neither confirm nor deny. 21:16
cognominal my understanding is that with recursive descendant parsing for handling || and && , operator precedence parsing for, well operator stuff and NFA for the rest. Perl 6 has found a sweet spot. 21:17
the synopsis are silent about operator precedence. And I don't even think that rakudo implement that. It is just two stacks juggling (operators and operands) according to precedence. 21:20
I think operator precedence means generating some kind of automata
masak yes, Rakudo and nqp do operator precedence. 21:21
cognominal or not, it seems that the juggling I describe is en.wikipedia.org/wiki/Shunting_yard_algorithm
masak and that's the only bit that *doesn't* involve generating some kind of automata.
cognominal I must confuse with some automata based system to deal with operators.
masak yes. 21:22
cognominal Yes, I spent the afternoon reading the code. And I get the gist of it.
masak I think the shunting yard algorithm is closely related but not identical.
cognominal btw, I need the trick to use C<.method(args)> as postfix to avoid left recursion in my bast grammar. 21:23
different things but the vocabulary is the same, so my confusion : en.wikipedia.org/wiki/Operator-prec...ce_grammar en.wikipedia.org/wiki/Simple_precedence_grammar en.wikipedia.org/wiki/Operator-precedence_parser 21:27
21:28 telex joined
adu cognominal: I'm not a fan of LALR 21:29
I am a big fan of PEG tho
cognominal and packrat?
adu cognominal: I guess
what's the difference? 21:30
cognominal not sure
don't drink PEG, btw :)
adu oh packrat = PEG + memoization
cognominal ok
I thought they were related 21:31
I don't think there is a good book that is up to date with all these kind of grammar.
adu PEG just requires an ordered '|', as opposed to longest-match '|'
cognominal I was told that the last edition of the dragon book was not worth the price. 21:32
100$ is a total rip-off 21:33
ingy is a big fan of PEGs, I think 21:35
21:55 kurahaupo joined
pmichaud expression parsing in NQP and Rakudo is a shift/reduce parser 21:57
with infinite levels of precedence available
21:59 grondilu joined
cognominal thx 22:00
pmichaud or, more accurately, an "operator precedence parser". 22:02
22:06 tgt joined
grondilu wow, there is a v5 module now, as specced. Nice. 22:14
22:22 felher joined 22:29 Vlavv_ joined
masak 'night, #Perl6 22:30
#perl6*
lizmat night masak, and #perl6
adu cognominal: well I'm a big fan of PEGs because they're provably composable 23:10
I have a question 23:27
so Str has index and rindex methods, but instead of ltrim and rtrim there's trim-trailing and trim-leading? 23:29
[Coke] masak: boingboing.net/2013/04/27/automated...rce=feedly 23:32
23:44 crab2313 joined 23:49 tgt left 23:50 berekuk joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!