6.2.10 released! xrl.us/hxnb | geoffb's column (/. ed): xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 10 October 2005.
eric256 ?eval 3e5.perl 00:03
evalbot_7621 "300000/1"
eric256 ?eval (3e5.5).perl
evalbot_7621 Error: unexpected "." expecting digit, "_", term postfix, operator or ")"
eric256 ?eval (3/5).perl
evalbot_7621 "3/5"
mrborisguy does pugs use the Perl6 Prelude right now? (src/perl6/Prelude/PIR.pm) 00:09
dolmen ?eval (3eR).perl 00:11
evalbot_7621 Error: unexpected "e" expecting digit, "_", "-", "+", exponent, term postfix, operator or ")"
dolmen ?eval (3e5).perl
evalbot_7621 "300000/1"
dolmen ?eval (3e5/5).perl
evalbot_7621 "60000/1"
dolmen ?eval (3e5/1).perl
evalbot_7621 "300000/1"
mrborisguy ?eval eval_parrot( "I0 = 8\nprint IO" );
evalbot_7621 Error: No compatible subroutine found: "&eval_parrot"
dolmen ?eval (3e5/Inf).perl 00:12
evalbot_7621 "0.0"
dolmen ?eval (1/Inf).perl
evalbot_7621 "0.0"
mrborisguy or does the Prelude just not really do anything right now?
dolmen sorry, don't know 00:13
?eval (Inf ^ 2).perl
evalbot_7621 "(2 ^ Inf)"
dolmen ?eval (Inf ^^ 2).perl
evalbot_7621 "bool::false"
dolmen ?eval true
evalbot_7621 Error: unexpected end of input expecting term
dolmen ?eval bool::true 00:14
evalbot_7621 bool::true
dolmen ?eval bool::true.perl
evalbot_7621 Error: unexpected "p" expecting block construct, ":", term, term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
dolmen ?eval (bool::true).perl
evalbot_7621 "bool::true"
eric256 yes it uses the prelude
its compiled during make
dolmen ?eval (4 ^ 2).perl
evalbot_7621 "(2 ^ 4)"
dolmen ?eval (4 ^^ 2).perl
evalbot_7621 "bool::false"
eric256 ?eval (4**2).perl
evalbot_7621 "16/1"
dolmen ?eval (4 ^^ 4).perl
evalbot_7621 "bool::false"
dolmen what is ^ ? 00:15
?eval (4 ^^ 4/bool::true).perl
evalbot_7621 "bool::false"
dolmen ?eval (4 ^^ 4/7).perl
evalbot_7621 "bool::false"
dolmen ?eval bool::true + 2 00:16
evalbot_7621 Error: No compatible subroutine found: "&bool::true"
dolmen ?eval (bool::true) + 2
evalbot_7621 3.0
mrborisguy dolmen: don't have a perl6 of your own to test on? :D
dolmen no
sorry
scook0 dolmen: ^ is a one() junction
?eval ?( (1 ^ 2 ^ 3) == 1 ) 00:17
evalbot_7621 bool::true
scook0 ?eval ?( (1 ^ 1 ^ 3) == 1 )
evalbot_7621 bool::false
mrborisguy dolmen: I don't care if you do... but you can ask Juerd for an account on feather.perl6.nl too 00:24
he gives 'em out so you can play with p6
that's what I do
Juerd mrborisguy: dolmen left 2 minutes before you said that. 00:33
eric256 quite today...except for me and dolmen blabbing to ourselves 01:24
lol
luqui blabbing to evalbot you mean 01:25
eric256 yea, not sure thats much better though
luqui well, we get to see the weird behavior you recover, like: 01:26
?eval bool::false / 2
evalbot_7621 0.0
luqui er, oh!
/ binds tighter than ^^, of course, so that stuff that dolmen did makes sense 01:27
putter hey folks 01:28
luqui hi putter
scook0 hey putter
eric256 hola
putter as dolmen just observed, bool::true/false have parsing issues currently. 01:29
?eval (0 ?? bool::false !! 0)
evalbot_7621 Error: unexpected ")" expecting "x", "d", octal digit, "o", "b", digit, fraction, exponent, term postfix, operator, ":", "," or "!!"
putter ?eval (0 ?? (bool::false) !! 0)
evalbot_7621 0
luqui ?eval 0 ?? bool::false() !! 0 01:30
evalbot_7621 0
luqui the arity of bool::true/::false is not known at compile time
eric256 loves how putter shows up knowing stuff that happened before he came in. lol
luqui that's my guess at least
putter I was so happy I got error line numbers working well when I hit that... :) (new p5 backend) 01:31
scook0 yeah, I don't think 0-ary subs are treated specially by the parser atm
luqui ?eval sub foo() { "hello" } say foo 1
evalbot_7621 Error: unexpected "1" expecting "(", term postfix, operator, ":", ",", postfix conditional, postfix loop, postfix iteration, ";" or end of input
putter re happening, colabti.de/irclogger/irclogger_logs/perl6 is your friend...
luqui no, I think they are
eric256 i know putter, its still funny 01:32
putter :)
?eval sub foo(){say "hi"} (1 ?? foo !! 0) 01:33
evalbot_7621 hi bool::true
scook0 I guess they are then 01:34
putter iblech: from a few? days ago: eval() should be working in PIL-Run. :lang<perl6> and :lang<perl5> at least. Not yaml or haskell. 01:37
scook0 (oh, now I see where they're handled) 01:39
putter autrijus: is there a calendar time estimate on when pil2 might be available? 01:44
scook0 stevan__: ping # quick note re: 'metaclass' 01:47
eric256 ?eval ((1 =>2)).perl 01:48
evalbot_7621 "(1 => 2)"
eric256 ?eval eval(((1 =>2)).perl).perl
evalbot_7621 "(1 => 2)"
putter after burning more hours than I'd like to think about doing a p5 backend overhaul, mostly writing things which will be trashed the minute pil2 exists, I'm reminded of my conviction that pil2 is a prerequisite for doing a real p5 backend.
not that one cant do without. witness pil2js. but since the target is oo, you either approach things the NotRightWay(tm), like PIL-Run, or spend hours like water trying to fake around pil1. 01:49
eric256 "spend hours like water" ... ?
putter lots of it goes by and the scene looks quite unchanged 01:50
;)
eric256 lol. okay
putter ach. /me is just a wee bit frustrated at the moment. 01:51
eric256 any specific? 01:53
eric256 watches as a feathers build process passes my local one started 15min earlier. lol
putter one of those exercises where things are clearly converging, so you keep on making the "ok, continue just a little bit more, rather than switching to other things" (eg, rules engine). but time estimates keep getting blown, so variance grows to the point of uselessness. "done sometime between this evening and next year". and you look back and realize the hours spent could have bought much more useful things. :/ 01:56
stevan__ scook0: pong # what about metaclasses? 01:57
eric256 as long as you learn something from the hours spent then it was worth it. ;) i try to view everything as a learning experience...even things that i know i could have done better in hind site
stevan__ can only ask, then backlog ,.. sorry :) 01:58
scook0 stevan__: please don't equivocate 'metaclass' and 'class metaobject'
it's just that it's really confusing
ideally, I'd like 'class metaobject' to be an object, and a 'metaclass' to be the class of that object 01:59
just a suggestion
eric256 's mind is going to turn to jello if he hears meta, class and object used ever agian. ;)
putter eric256: true. knew someone who treated major life disasters with a "I will put whatever effort needed looking, because there damn well better be something good in here proportional to the pain!". ;) even worked sometimes. :) 02:00
hey stevan__ 02:01
eric256 arg...stepped on tack strip in the carpet and it totaly chewed up my foot...stupid tack strips
putter ouch 02:02
eric256 big wholes and blood.. is this pugs EVER going to finish compiling!!
putter stevan__: any thoughts on why Chaos's sub ::opaque_instance_attr : lvalue (line 40) might sometimes blow out with a Not an ARRAY reference? no doubt I'm abusing it, but I wondered if you'd seen the symptom before. tnx 02:04
eric256 ?eval :b(1).perl 02:08
evalbot_7621 "(\"b\" => 1)"
eric256 ?eval ("b" => 1).perl 02:09
evalbot_7621 "(\"b\" => 1)"
eric256 ?eval :b(1).ref
evalbot_7621 ::Pair
putter ?eval sub f(*%h){} f:a()
evalbot_7621 undef
putter ?eval sub f(*%h){say %h} f:a() 02:10
evalbot_7621 a1 bool::true
putter ?eval sub f($x,*%h){say "$x--",%h} f:a(6)
evalbot_7621 Error: Undeclared variable: "$x"
putter ?eval sub f($x,*%h){say "$x--",%h} f:a (6)
evalbot_7621 6--a1 bool::true
putter ?eval sub f($x,*%h){say "$x--",%h} f:a(3)(6) 02:11
evalbot_7621 6--a3 bool::true
putter ?eval sub f($x,*%h){say "$x--",%h} f(6):a 02:12
evalbot_7621 6--a1 bool::true
putter ?eval sub f($x,*%h){say "$x--",%h} f(6) :a
evalbot_7621 Error: unexpected ":" expecting operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
eric256 ?eval sub x (@a) { for @a { print } }; x(1=>2);
evalbot_7621 Error: Insufficient number of required parameters: 0 actual, 1 expected
eric256 any idea how to fix that putter?
putter hmm...
luqui ?eval sub x (*@a) { for @a { print } } x(1 => 2)
evalbot_7621 undef
luqui ?eval sub x (*@a) { say +@a } x(1 => 2) 02:13
evalbot_7621 0 bool::true
eric256 ?eval sub x (@a) { @a.perl.say; for @a { print } }; x(1=>2);
evalbot_7621 Error: Insufficient number of required parameters: 0 actual, 1 expected
eric256 ?eval sub x (Pair @a) { @a.perl.say; for @a { print } }; x(1=>2);
evalbot_7621 [(1 => 2)] undef
putter "fix" as in workaround? my fuzzy impression is Pair handling is currently a bit kludgy.
eric256 nah i have a work around. just though maybe you would have a real fix
putter my fuzzy understanding is real fix is thought to consist of language spec change and non-trivial changes to pugs 02:15
eric256 gotcha...just seemed like since adding Pair fixed it, there might be an easier fix
?eval { a => [1,2,3] }.perl 02:16
evalbot_7621 "\{(\"a\" => [1, 2, 3])}"
eric256 ?eval ~{ a => [1,2,3] } 02:17
evalbot_7621 "a\t1 2 3\n"
putter my fuzzy recollection is there were a bunch of autrijus commit messages "drat, that didnt work, backing out" re Pairs. so easy seems unlikely. ;) 02:18
eric256 lol. alright
?eval ~{ a => [1,2,3] }.perl.eval
evalbot_7621 "a\t1 2 3\n"
putter it's also worth trying pugs -Bjs with stuff like this. pil2js has quite a few 'unexepected successes' on tests. 02:20
eric256 ?eval ~{ :a(1), :b(2), :c(3) } 02:21
evalbot_7621 "a\t1\nb\t2\nc\t3\n"
eric256 ?eval { :a(1), :b(2), :c(3) }.perl.eval.perl
evalbot_7621 "((\"a\" => 1), (\"b\" => 2), (\"c\" => 3))"
putter wont help with parsing or compilation problems, but sometimes the js runtime works better
eric256 ?eval ~{ :a(1), :b(2), :c(3) }.perl.eval
evalbot_7621 "a\t1 b\t2 c\t3"
eric256 ?eval { :a(1), :b(2), :c(3) }.perl
evalbot_7621 "\{(\"a\" => 1), (\"b\" => 2), (\"c\" => 3)}"
eric256 how come evaling that returns a list? 02:22
putter not a list, a hash, no?
eric256 ?eval eval("\{(\"a\" => 1), (\"b\" => 2), (\"c\" => 3)}" ).perl 02:23
evalbot_7621 "((\"a\" => 1), (\"b\" => 2), (\"c\" => 3))"
eric256 ?eval { a => 1 }.perl
evalbot_7621 "\{(\"a\" => 1)}"
eric256 ?eval { a => 1 }.perl.eval.perl
evalbot_7621 "\{(\"a\" => 1)}"
eric256 scratches head 02:24
?eval { :a(1) }.perl.eval.perl
evalbot_7621 "\{(\"a\" => 1)}"
eric256 ?eval { :a(1) }.perl
evalbot_7621 "\{(\"a\" => 1)}"
eric256 ?eval { :a(1), :b(2) }.perl
evalbot_7621 "\{(\"a\" => 1), (\"b\" => 2)}"
eric256 ?eval { :a(1), :b(2) }.perl.eval.perl
evalbot_7621 "((\"a\" => 1), (\"b\" => 2))"
eric256 see? or am i crazy? 02:25
?eval { :a(1), :b(2) }.ref
evalbot_7621 ::Hash
eric256 ?eval { :a(1), :b(2) }.perl.eval.ref
evalbot_7621 ::Array
eric256 ?eval "{(\"a\" => 1), (\"b\" => 2)}".eval.ref
evalbot_7621 ::Array
putter ?eval { :a(1), :b(2) }.perl.eval
evalbot_7621 (("a" => 1), ("b" => 2))
eric256 ?eval "hash {(\"a\" => 1), (\"b\" => 2)}".eval.ref
evalbot_7621 ::Scalar
eric256 ?eval "hash ((\"a\" => 1), (\"b\" => 2))".eval.ref 02:26
evalbot_7621 ::Scalar
putter ?eval my $x = { :a(1), :b(2) }.perl.eval; $x
evalbot_7621 [("a" => 1), ("b" => 2)]
eric256 parse error? or pretty print error? 02:27
putter that does seem excessively listy
eric256 ?eval "{(a => 1), (b => 2)}".eval.ref
evalbot_7621 ::Scalar 02:28
eric256 ?eval "\{(a => 1), (b => 2)}".eval.ref
evalbot_7621 ::Array
luqui ?eval { :a(1) :b(2) }.perl
evalbot_7621 Error: unexpected ":" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or "}"
putter remember { is now magic in strings
eric256 ?eval {(a => 1), (b => 2)}.ref
luqui ?eval { :a(1), :b(2) }.perl
evalbot_7621 ::Block
"\{(\"a\" => 1), (\"b\" => 2)}"
putter ?eval '{(a => 1), (b => 2)}'.eval.ref
evalbot_7621 ::Array
luqui ?eval { :a(1), :b(2) }.perl.say
eric256 i think luqui has it
evalbot_7621 {("a" => 1), ("b" => 2)} bool::true
putter ?eval '\{(a => 1), (b => 2)}'.eval.ref
evalbot_7621 ::Block
eric256 ?eval "{\{(a => 1), (b => 2)}}".eval.ref 02:29
evalbot_7621 ::Str
eric256 hmmm
putter ?eval "{say 'hi'}"
evalbot_7621 hi "1"
luqui ?eval { :a(1), :b(2) }.perl.eval
eric256 ?eval '\{(a => 1), (b => 2)}'.eval.ref
evalbot_7621 (("a" => 1), ("b" => 2))
::Block
luqui ?eval { :a(1), :b(2) }.perl.say 02:30
evalbot_7621 {("a" => 1), ("b" => 2)} bool::true
eric256 or not.... i dunno. something wierd is happening
luqui ?eval eval '{("a" => 1), ("b" => 2)}'
evalbot_7621 (("a" => 1), ("b" => 2))
luqui ?eval \eval '{("a" => 1), ("b" => 2)}'
evalbot_7621 [("a" => 1), ("b" => 2)]
luqui weird
the braces are just disappearing
eric256 can't seem to eval *anything* into a hash.... 02:31
luqui ?eval {("a" => 1), ("b"=>2)}
evalbot_7621 (("a" => 1), ("b" => 2))
luqui there we go
?eval my $x = {("a" => 1), ("b"=>2)}; $x
evalbot_7621 \sub {...}
luqui OH!
Got it
eric256 ?eval { :a(1) }.perl
evalbot_7621 "\{(\"a\" => 1)}"
luqui its the parentheses
eric256 ?eval { :a(1) }.perl.eval.perl
evalbot_7621 "\{(\"a\" => 1)}"
luqui ?eval { a => 1 }.ref
evalbot_7621 ::Hash
luqui ?eval { (a => 1) }.ref 02:32
evalbot_7621 ::Hash
luqui what?
I thought I got it
eric256 lol
luqui ?eval { (a => 1), (b =>1) }.ref
evalbot_7621 ::Block
luqui there we go
eric256 ?eval '\{a => 1, b => 2}'.eval.ref
evalbot_7621 ::Hash
eric256 you were right
;;)
luqui only if the top level of a {} has a pair is it considered a hash
putter ?eval { a => 1, (b =>1) }.ref
evalbot_7621 ::Hash
putter ?eval { (a => 1), b =>1 }.ref 02:33
eric256 goes off to fix it...hopefully
evalbot_7621 ::Block
eric256 thanks luqui
luqui hang on eric256
it may not be broken
eric256 arg....my stupid putty screen to feather
it is. its failing the tests for pretty printing and then evaling
luqui oh right
eric256 when printing a hash it can't make them pairs
luqui but how do you intend to fix it?
oh, you could kill the parens on output 02:34
eric256 dunno yet, haven't looked at the printing for hashes yet
luqui ?eval { a => 1, b => 2 }.perl.say
evalbot_7621 {("a" => 1), ("b" => 2)} bool::true
luqui make that say { "a" => 1, "b" => 2 }
eric256 yep
putter :)
eric256 do you know how to turn off line wraps in vi? 02:35
luqui yeah, let me look at my .vimrc (I've done it)
set nowrap
eric256 hmmm could hashes be being printed by the list routine? 02:38
luqui where are the braces coming from? 02:39
eric256 damn...found them...hmmm
is a hash truly a list of pairs? 02:41
scook0 by the way, I'm currently looking into demagicalizing pairs
eric256 and is there anyother way to force { into a hash?
?eval '\{, (a => 1), (b => 2)}'.eval.ref 02:42
evalbot_7621 ::Scalar
eric256 ?eval '\{ (a => 1), (b => 2)}'.eval.ref
scook0 I have a mostly-complete patch, but I'm running into trouble with primitives
evalbot_7621 ::Block
eric256 ?eval '\{+ (a => 1), (b => 2)}'.eval.ref
evalbot_7621 ::Block
eric256 ?eval '\{, (a => 1), (b => 2)}'.eval.ref
evalbot_7621 ::Scalar
eric256 oh yea..just tried that. lol 02:43
stevan__ scook0: I would prefer not metaclass at all
scook0: I think classes should be instances of Class (note the uppercase)
eric256 hash is treated as a list of pairs, so there is no easy way to drop the paran's....but if there is another way to force it into a hash i could handle that.
?eval 'hash { (a => 1), (b => 2)}'.eval.ref
evalbot_7621 ::Scalar
eric256 ?eval '\hash { (a => 1), (b => 2)}'.eval.ref
evalbot_7621 ::Scalar
stevan__ the meta stuff just makes things more confusing
scook0 stevan__: as in, no distinction between a 'class', and the metamodel-object that represents it?
stevan__ yes
why should there be 02:44
xinming Hi, all, Now I am in ZheJiang... back... :-P
stevan__ we need to throw out the C++/Java notion that a class is anything special
eric256 ?eval 'hash ( (a => 1), (b => 2))'.eval.ref
evalbot_7621 ::Scalar
scook0 yes, but if $Larry *insists* on separating classes and objects, then 'class meta-object' is a better term
eric256 ?eval 'hash ( (a => 1), (b => 2))'.eval.perl
evalbot_7621 "undef"
eric256 ?eval 'hash { (a => 1), (b => 2)}'.eval.perl
evalbot_7621 "undef"
scook0 that was my main point--overall, I probably prefer dropping the distinction
eric256 ?eval { (a => 1), (b => 2)}.perl
evalbot_7621 "sub \{...}"
eric256 ?eval hash ( (a => 1), (b => 2)).perl 02:45
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256 ?eval hash( (a => 1), (b => 2)).perl
evalbot_7621 "\{(\"a\" => 1), (\"b\" => 2)}"
eric256 ahhh... ;)
stevan__ I think $Larry needs to see how powerfull non-seperated class/object is
eric256 ?eval 'hash( (a => 1), (b => 2))'.eval.perl
evalbot_7621 "\{(\"a\" => 1), (\"b\" => 2)}"
eric256 sweet
stevan__ but for now, I need to go watch a movie with my wife :)
stevan__ &
eric256 luqui .. that look about right to you?
luqui eric256, are you going to output as a hash()?
eric256 i was thinking about it 02:46
down side?
luqui not as pretty
eric256 i'm open to suggestions
luqui can you keep the parens off the pairs?
eric256 working++ > pretty-- ;)
luqui eric256, of course
eric256 no
luqui ?eval (a => 3).perl
evalbot_7621 "(\"a\" => 3)"
eric256 i can't, someone else might be able to
luqui have any leads? 02:47
eric256 or provide some other signal to the parser that its a hash and not a block
luqui hash() is probably good for now 02:48
eric256 well for now i'll do that. ;)
objections?
hearing non, eric heads out to break...i mean fix output. ;)
(return $ ("hash(":(init (tail str))) ++ ")") 02:51
putter :)
good night &
eric256 i'm reading something wrong on that line
later
putter good luck
eric256 i don't know what the : is doing...anyone know haskell? 02:56
scook0 eric256: : is cons 02:57
PerlJam eric256: The } is cons
er, :
eric256 ahhh. dunno how i forgot that ;)
do you see an error there then? 02:58
perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
scook0 I think you might need to try getting rid of some parens, so it's easier to read
pasteling "eric256" at 24.8.74.230 pasted "parse error." (6 lines, 238B) at sial.org/pbot/13708 02:59
PerlJam I agree with that. Too many parens remind me of lisp
lisp is fine for a computer to read, but the excessing regularity is annoying to this human.
scook0 eric256: is it supposed to be returning a string, or a list of strings?
eric256 i thought a string, but now i think its a list of strings 03:00
scook0 return $ "hash(" ++ (init $ tail str) ++ ")"
eric256 thats the way it was already, i didn't do any of that except the hash( opart
scook0 eric256: what file/line? 03:01
eric256 Prim.hs 1363
trying yours there... looks like its going to make it...dunno what it will look like ;) 03:02
scook0 eric256: you replaced a single-char '{' with a string "hash("
eric256 i don't understand why they are abusing str like that
yea
scook0 which means (:) is no longer what you need
try (return $ "hash(" ++ (init $ tail str) ++ ")")
eric256 : was just a cheap way of joining strings?
lol
PerlJam eric256: it's a "cheap" way of constructing lists of atomic things. 03:03
scook0 (:) only works for adding a single character (element) to the start of a string (list)
PerlJam eric256: in haskell, a string is already a list
scook0 once you start joining two strings (lists), you need (++)
eric256 yea i just forgot what i knew..lol 03:04
?eval "hash((\"1\" => 2))".eval.perl
evalbot_7621 "undef"
PerlJam eric256: "hello" == 'h':'e':'l':'l':'o':[] ( I think, it's been a while since I've hacked haskell)
eric256 ?eval "hash((1 => 2), 2)".eval.perl
evalbot_7621 "\{(\"1\" => 2), (\"2\" => undef)}"
eric256 ?eval "hash((1 => 2))".eval.perl
evalbot_7621 "undef"
eric256 curses agian. lol
scook0 PerlJam: that's correct
eric256 fixed the output...but it doesn't work anyway. lol. i thought i tested that 03:05
?eval "hash((1 => 2), (3=>4))".eval.perl
evalbot_7621 "\{(\"1\" => 2), (\"3\" => 4)}"
eric256 ?eval "hash( (1 => 2) )".eval.perl 03:06
evalbot_7621 "undef"
scook0 does anyone know how named-args get passed to primitives (specifically Object::new)?
eric256 ?eval hash( (1 => 2) )
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256 ?eval hash( 1 => 2 )
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256 ?eval hash( 1 => 2, 3 => 4 )
evalbot_7621 {("1" => 2), ("3" => 4)}
eric256 hash is broken too??
magical pairs are cursing me agian arn't they? 03:07
scook0 &hash is a bit of a hack at the moment
PerlJam Hmm. Last time I looked hash(1=>2) complained about an odd number of elements. I'm guessing someone has been working on hash()? 03:08
scook0 PerlJam: `hash {}` still complains 03:09
PerlJam yeah, I was just looking at Eval.hs 03:10
All those "bloody hacks" need an application of haskell-fu 03:11
eric256 was trying. ;) well actualy mine are probably more on the hackish side. but i'm trying to work on my haskell-fu
PerlJam eric256++ keep it up. 03:12
eric256 ahhh...single item...hmmm 03:14
looking at Eval.hs
Khisanth well hopefully it won't have a hairier core than p5 :)
eric256 no way for it to know that one item is a pair is there? if there is that would/should be a simple fix to add 03:15
of course 3 pairs should still work.
could we just let hash assume the best?
Khisanth that doesn't sound good 03:16
eric256 well you should still get an error if it doesn't realy produce a hash. just not as early
hmmm...silently assumes last odd key has an undef value. not sure thats a feature 03:18
lol
?eval { 1,2,3}.perl
evalbot_7621 "sub \{...}"
eric256 ?eval { 1 => 2,3}.perl
evalbot_7621 "\{(\"1\" => 2), (\"3\" => undef)}"
eric256 same behavior as that if you drop the warning
no comments? 03:20
seems like drop the error and add the error in the actual hash compisition
?eval "\\{ 1, 2}" 03:22
evalbot_7621 "\\1 2"
eric256 ?eval "\\{ 1, 2}".perl
evalbot_7621 "\"\\\\1 2\""
eric256 ?eval '\\{ 1, 2}'.perl
evalbot_7621 "\"\\\\\\\{ 1, 2}\""
eric256 ?eval \\{ 1, 2}.perl
evalbot_7621 Error: unexpected "\\" expecting program
eric256 ?eval '\\{ 1, 2}'.eval.perl 03:23
evalbot_7621 "\\sub \{...}"
scook0 autrijus: ping # in a call to &Object::new, when/where do the named/positional args get segregated? 03:29
stevan scook0: are you asking about pugs, or perl6 in general 03:30
scook0 stevan: pugs internals
stevan couldn't find a good movie to watch
scook0: ahh, can't help you there 03:31
03:31 Supaplex is now known as linuxmammal, linuxmammal is now known as Supaplex
stevan scook0: re: meta junk 03:33
scook0 stevan: yeah? 03:34
stevan if the entire meta-model can sit "on top" of the core interpreter
then it can be much more flexible
this requires that Class (uppercase, proper noun) is accessible and usable in the same way Object is
just having .meta is probably not enough
but then again, if you just put Class at the other end of .meta it is probably okay too 03:35
I think Larry has some really interesting ideas, my issue is that I want to understand them within the context of some meta-model
s/meta/object/ 03:36
eric256 hmmm its doing something special droping the first and last character in Prim.hs .... for some reason that must get called twice on the same bit? seems like a bit of a hack to drop the first and last letter and add your own though../me searches more
stevan I based the current model heavily on CLOS because CLOS was built from the start to be able to encompass the behaviors of some of the other LISP Object systems out there at the time 03:37
since p6 should interact with (at least) p5, and (at most) Python & Ruby, I thought going down this road would be a the right direction 03:38
in the end if you call it meta, that is fine, but it is my hope that the terminology will just be "names" and nothing more
scook0 stevan: the only point I was really trying to make 03:39
was that *iff* classes and objects are distinct kinds of entities
stevan hey nothingmuch :) 03:40
nothingmuch morning
scook0 then if you want a term for "the representation of a class in object-space"
stevan morning
scook0 nothingmuch: hi
eric256 hey nothingmuch
scook0 then that term should be "class metaobject" rather than "metaclass"
eric256 nothingmuch: do you know much about the Pretty printing ?
nothingmuch the pretty printing of what?
stevan scook0: In my mind, Class and Object are no different from Foo and Bar 03:41
eric256 .perl
nothingmuch oh
no clue at all
stevan Class is an instance of Class, and Class is a subclass of Object
eric256 figured it was worth a shot. lol
scook0 stevan: I personally subscribe to that view
nothingmuch has some stupid macho friends
they got slightly drunk 03:42
and we drove this girl home on the way
stevan scook0: my goal for the near future is to get Larry to subscribe to that as well :)
nothingmuch all the time she was in the car they were shouting crap at the top of their lungs trying to impress her
the moment she got out, they became quiet
then one of them got a call from girlfriend, and they talked with her friend 03:43
shouting again
scook0 stevan: I think in Larry's mind, none of Class, Object or Foo are 'proper' objects 03:44
stevan scook0: well I think he is seeing them as Roles of some kind
scook0 so what you think of as Class, Object and Foo, he thinks of as Class.meta, Object.meta and Foo.meta 03:45
i.e. he's making the distinction between class and class-metaobject explicit
whereas you don't believe there's value in that distinction
stevan scook0: my guess is that he is thinking of Class, Object, etc as being creatures of the interperter
and that they need some kind of Perl6 space facade, which is where the Roles come in 03:46
but I might be reading too much into it
nothingmuch anyway, off to bed, lambdacamels
scook0 nothingmuch: bye
stevan shouts at nothingmuch
stevan takes the dog out, be right back 03:48
eric256 i'm missing something obvious...must be...or it could be the fact that there are 15 different coding styles in a language i'm just starting to learn. ;) 03:53
luqui eric256, you mean about haskell? 03:56
eric256 pugs source code in haskell
luqui it's not that you're missing something obvious
it's that you're missing fifty subtle things all at once
that's learning haskell...
eric256 hehe yea i've noticed that already
going...duh thats a sub right there it just looks like an assignment lol
whole new way of thinking 03:57
luqui yep
have you just played with ghci?
eric256 yep
and walked through a tutorial
;) nothing like hands on though
prettyVal d v@(VRef r) = do
what does the @ mean there?
luqui oh, that's just (v = (VRef r)) 03:58
so you're extracting the r inside VRef, but still referring to the whole VRef as v
eric256 so why would the sub then refer to both the v and the r?
oh wait i think i get it 03:59
scook0 eric256: for .perl-ifying hashes, have you tried `hash list (1=>2)`?
eric256 scook..hard part is i can't get it to ouput right yet
cuts off parts of hash( )
scook0 it just avoids having to write `let v = (VRef r)`
eric256 or doubles it up...ended up with hash(ash(1 => 2)) a couple times
scook0 eep 04:00
eric256 lol
yea
scook0 eric256: I wonder if that's related to the (init $ tail str)? 04:01
eric256 definitly is...but removing the init $ tail just makes it stop working at all and returns undef....dunno why 04:02
so there is something subtle going on. i just have to figure out what ;)
prettVal d v@(VRef r) is being called on the value twice as it is being printed..i think. :) 04:03
but i don't know where or why
luqui (init $ tail str) takes off the first and last character of str
eric256 yea
scook0 my guess is that it's trying to remove redundant parens or braces 04:04
eric256 yea...but its doing it while adding them....why remove them and then add? 04:05
remove parens while adding brackets...could be
scook0 seems most likely
eric256 but its also removing redundant brackets....
scook0 maybe the pretty-prenting of a hash (not hashref) has parens
so the pretty-print of a hashref takes that, removes parens, and adds braces instead? 04:06
eric256 and hashref uses it, but removes the parens to add brackets
yea that seems to follow the code there
scook0 exactly
eric256 okay...now to use this new knowledge ;)
thanks. you have to think completely different in haskell...thats hard
not like learning VB or C++ or javascript...totaly differnt 04:07
scook0 eric256: it's not as though the program itself is easy, though
if you were doing something equivalently complex in C++, you probably would have given up by now :)
eric256 probably. i hate c++. lol 04:09
hmm broke something agian. ... wonder what. 04:12
how do i have svk undo all local changes? 04:13
nm found it
eric256__ any idea why if i changed something and it broke something else, then i changed it back and remade...the break didn't go away? do i need to always perl Makefile.PL ? 04:29
luqui eric256__, weird 04:40
i think it depends on what you changed...
no, nevermind 04:41
you should never have to Makefile.PL unless you rearrange files
eric256__ i commented out the # of elements check in Eval.hs and then pugs started returning undef for all hashrefs...so i uncommented it and recompiled but it still did it. 04:42
luqui that was the only change you did between recompiles? 04:43
eric256__ yep 04:44
commented, make, tested, uncommented, make, tested
before the comment it worked, after both tests failed 04:45
luqui very strange
don't know what to tell you
eric256__ hehe. i thought i'd done something wrong, thats why i revereted all files. didn't help
well it di, it started working right. but then when i tried agian same thing happened
like it wasn't noticing my change toe Eval.hs 04:46
lol or i'm stupid
lol. interactive pugs doesn't print the value
you ahve to {1=>5}.perl.say;
blah
thats an hour of life wasted 04:47
luqui :-/
luqui had a similar experience
I was debugging one of my modules for two hours only to find that the error was due to an error in the .t
PerlJam luqui: always check both sides of the equation :) 04:48
eric256__ hmmm.. works fine. except for loop seems to do something realy strange 04:59
?eval my @t = (1, {1=>2} ); for @t {$_.perl.say;}
evalbot_7621 \1 {("1" => 2)} undef
eric256__ ?eval my @t = (1, 1=>2 ); for @t {$_.perl.say;}
evalbot_7621 \1 Error: Insufficient number of required parameters: 0 actual, 1 expected
eric256__ ?eval my @t = (1, hash(1=>2) ); for @t {$_.perl.say;}
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256__ ?eval my @t = (1, hash(1=>2) ); for Any @t {$_.perl.say;}
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256__ ?eval my @t = (1, hash(1,2) ); for Any @t {$_.perl.say;} 05:00
evalbot_7621 Error: Pattern match failure in do expression at src/Pugs/Eval.hs:419:4-12
eric256__ ?eval my @t = (1, hash(1,2) ); for @t {$_.perl.say;}
evalbot_7621 \1 {("1" => 2)} undef
eric256__ why is that \1?? its a ref to it?
luqui a ref to 1
luqui dunnos why
eric256__ hmmm. so if you have a hashref in an array, then $_ ends up being a ref to a ref?... explains the odd behaviour at least. 05:01
?eval hash( hash( 1, 2) ).perl
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256__ hehe. things start to make a sick kind of sense...why they coded it the way they did 05:02
luqui ?eval hash( a => hash( b => c ) ) 05:08
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
luqui ?eval hash( a => hash( b => 'c' ) )
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256__ whole thing is screwy . lol
luqui yeah
I've been wanting to port some of my modules to perl 6 05:09
but then I start and I run into these bugs. it's very frustrating
pugs is not very good with nested data strucutures... at all
?eval hash( a => 'b' ) 05:10
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
luqui umm, I swear that used to work
eric256__ ?eval hash( a , 'b' )
luqui ?eval hash( a => 'b' ).perl
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256__ arg. wish escape wouldn't close the window
?eval hash(1,2) 05:11
evalbot_7621 {("1" => 2)}
luqui ?eval hash('a', 'b')
evalbot_7621 {("a" => "b")}
luqui ?eval hash('a' =>'b')
evalbot_7621 Error: cannot cast from VUndef to Handle (VHandle)
eric256__ it sees it as a pair and a single element. dunno why evalbot isn't catching the warning though
luqui ?eval hash('a' =>'b', 'c') 05:12
evalbot_7621 {("a" => "b"), ("c" => undef)}
luqui ?eval hash('a' =>'b', 'c', 'd', 'e' => 'f')
evalbot_7621 {("a" => "b"), ("c" => "d"), ("e" => "f")}
eric256__ i'm confused...how can a hashref be pointing to a hashref? shouldn't a hashref have to point to the hash itself? 05:14
or all refs that point to a hash in the end considered a hash?.....arg. 05:15
luqui what do you mean a hashref pointing to a hashref
is there somewhere you are looking
eric256__ when i do that for loop thing...right now i wrap hashrefs in hash( ) instead of { } 05:16
inside a for loop it becomes
hash ( hash ( ) )
only way i can get there in the code is if there is a ref (of type hash) pointing to a ref (of type hash)... 05:17
luqui ... for Any @t ?
maybe the reftype code dereferences until it finds something that is not a ref 05:18
(just guessing here)
?eval (\\\1).ref
evalbot_7621 Error: unexpected "\\" expecting expression
luqui ?eval (\ \ \1).ref 05:19
pasteling "eric256" at 24.8.74.230 pasted "pugs> my @t =(1, {2=>2}); for" (5 lines, 91B) at sial.org/pbot/13709
evalbot_7621 ::Int
luqui yeah, like that
?eval \ \ \ 1
evalbot_7621 \1
eric256__ hmmm 05:20
well that presents a problem. lol
but it explains what they did...gonna sleep on this. i think there might be an easy way around it.
luqui perhaps
maybe autrijus will be around when you get back 05:21
so he can guide you
eric256__ i'm assumeing that if there is a way to jump to the final type that you could also jump to a final value. at least i know alot more than 4 hours ago
luqui is meaning to get back into pugs hacking at some point 05:22
thanks for inspiring me
eric256__ hehe thanks for the help. feel free to fix this if you can figure it out! i just picked tests that were failing and started. ;)
later
luqui later 05:23
mrborisguy anybody know the story on 'src/perl6/Prelude/PIR.pm' ? 05:28
does it somehow get included, or is it just there for the future? 05:29
luqui you could grep the source 05:34
mrborisguy to be honest, I don't even know where to start looking. src/Pugs ?
runinshine Is the source of Pugs Hashell? 05:36
mrborisguy runinshine: yeah, Haskell
runinshine I don't understand it , can you give me some advise? 05:38
mrborisguy you don't understand what? Haskell? 05:39
you don't understand Perl6?
scook0 runinshine: did you have a question before? 05:57
runinshine yeah 05:58
who can give me some advice of learning the Hashell ? 06:00
luqui go to #haskell and ask where to start 06:01
give them your background as a programmer
runinshine thanks 06:02
Khisanth haskell looks very nice and appealing, just need to find something to do with it! :)
scook0 Khisanth: like writing a Perl 6 implementation! :P 06:03
Khisanth well I mean personally 06:04
runinshine why is it suitable for writing the implementation of Perl6?
mrborisguy scook0: the problem with starting to write a Perl 6 implementation is that it's not really a good place for the beginner 06:05
scook0 mrborisguy: *grin*
mrborisguy I tried to learn a little Haskell this summer, and I can do some stuff in it, but I poked around in the internals, and I'm clueless! 06:06
(although, it might help if I knew what each file was trying to accomplish :D)
scook0 mrborisguy: well, don't hesitate to ask
luqui runinshine, because that's what autrijus wrote it in
scook0 not everybody has to figure it out the hard way
luqui :-)
mrborisguy true, true. 06:07
sometimes I think... just give me something to hack away on for a while until I figure it out, maybe I could help 06:08
but I never know what... any ideas?
luqui the high numbers of :todo<bug>s in .t 06:09
er, t/
scook0 attention: anyone who uses alternative backends 07:03
I have a patch that demagicalizes pairs/named-args in the Parser and the normal Haskell runcore 07:04
but it will almost certainly break named-args in the various backends
luqui scook0, cool 07:08
svnbot6 r7622 | scook0++ | * Abracadabra, shalakazam, demagicalize pairs!
r7622 | scook0++ | (This will probably break all the other backends.)
r7622 | scook0++ | * A few minor style/readability tweaks
luqui scook0++ 07:09
(3 isn't enough)
scook0 luqui: it *almost* passes the entire test suite
(except for a few cases that are really caused by other parser bugs)
but, as I said, it will break the other backends :(
luqui how come
did you change PIL?
scook0 luqui: no--that's why they'll break 07:10
luqui oh...
scook0 I added (Sym "named"), which the compiler won't recognise
luqui I see
well, it had to be done sometime
scook0 but, it's possible to temporarily undo the parser changes with a one-line fix 07:11
(basically, I have a 'named' combinator that wraps around the parsers for pairs)
if you replace that combinator's definition with "id", named-pairs get parsed as normal pairs again
by the way, did somebody check in some tests for demagicalized behaviour? 07:12
I seem to remember someone adding some, but I might have imagined it 07:13
luqui ??
luqui will grep and then add if not
did you use the semantics from the recent p6l thread?
that is foo(a => b) # named foo((a => b)); # normal pair 07:14
scook0 luqui: yeah
but not everything works: e.g. `foo.(:a)` is broken
because of a bug in `foo.()` parsing 07:15
luqui Retrieving log information from 7119 to 7622 # it's been a while
a day or two ;-) 07:16
gaal kolibrie: pong
luqui we'll catch up to parrot pretty soon here
scook0_ urk, net troubles :( 07:17
luqui ?eval sub foo($x) { say $x } foo(x => 4) 07:18
evalbot_7622 4 bool::true
luqui ?eval sub foo($x) { say $x } foo((x => 4))
evalbot_7622 x4 bool::true
luqui woot
?eval sub foo($x) { say $x.ref } foo((x => 4))
evalbot_7622 Pair bool::true
luqui ?eval sub foo($x) { say $x.ref } my $p = (x => 4); foo($p) 07:19
evalbot_7622 Pair bool::true
luqui ?eval sub foo($x) { say $x.ref } my $p = (x => 4); foo(*$p)
evalbot_7622 Array bool::true
luqui ?eval sub foo($x) { say $x.ref } my %p = (x => 4); foo(%p)
evalbot_7622 Hash bool::true
luqui ?eval sub foo($x) { say $x.ref } my %p = (x => 4); foo(*%p)
evalbot_7622 Array bool::true
luqui scook0_, didn't do the splat?
i.e. the remagicalizing operator
scook0_ luqui: I don't think hash-splat ever worked before 07:20
luqui fair enough
scook0_ luqui: are you writing tests, or shall I start some? 07:25
luqui I was updating pugs 07:30
but I shall start writing tests now
scook0_, looks like there are already tests in t/syntax/pairs.t 07:32
scook0_ iblech++ # demagicalized pair tests 07:34
luqui how do I praise and not always see my own name with svk? 07:35
07:36 scook0_ is now known as scook0
scook0 luqui: you don't happen to know the exact syntax of sub calls wrt where you can put adverbs, do you? 07:38
(I'm writing parse-tests for various sub-call forms) 07:39
Aankhen`` foo($bar, :baz :quux)
coral . 07:40
ww
scook0 I'm just wondering about these sorts of things: 07:41
foo:adv()($arg) # probably not
Aankhen`` I agree with the comment. :-)
scook0 foo():adv
foo() :adv # with space
Aankhen`` Those last two are probably possible. They'd need to be for operators which take adverbial modifiers, e.g.: 1..10 :by(2) 07:42
scook0 foo(:alfa :bravo $arg :charlie :delta):echo :foxtrot
Aankhen`` Of course, I'm out of the loop, so I could be dead wrong.
I believe you can't put a positional parameter in the middle of a bunch of named parameters... 07:43
scook0 maybe I'll write up a big list of all the corner cases, and send it to p6l
foo(:adv $arg)
Aankhen`` Positional parameters before named parameters! 07:44
foo($arg, :adv);
foo $arg, :adv; # dunno about this
scook0 Aankhen``: I think that should work too 07:45
luqui Aankhen``, I'm trying to kill that
not sure what Larry thinks
Aankhen`` OK.
scook0 it's not really necessary any more
luqui he's wanted "zones" for a while, but that may have been his implementation-head, which the demagicalized pairs proposal destroyed
Aankhen`` But if you kill it, what happens to operators that can take adverbial modifiers?
scook0 now that named-arguments can be determined statically 07:46
luqui Aankhen``, what do you mean?
they still exist
Aankhen`` They're not affected?
luqui no
scook0 Aankhen``: he was talking about "Positional parameters before named parameters!"
Aankhen`` Ah, I see.
scook0 not your example
luqui ohh
scook0 foo $arg, :adv; # should work 07:47
Aankhen`` spanks his lag.
scook0 just like: foo $arg, adv=>1;
and as a matter of style, mixing named and positional args is still bad news 07:48
luqui maybe
scook0 but it no longer needs to be illegal, from an implementation POV 07:49
luqui: i.e. foo($pos, :named, $pos, :named)
luqui foo 1, 2, 3, :parallel, { big block of code }
it's an endweight thing
gaal Hey Aankhen``, long time no moose.
Aankhen`` Hiya gaal. :-D
scook0 luqui: yes, your particular example was good 07:50
putter a quick log check before sleep yields...
scook0 which is why it's nice that they don't need to be illegal
but things like *my* example should be frowned upon by P6 style guides
luqui yeah probably
putter on adverbs, re "# probably not", remember rx:perl5// rx:x(4)// etc.
scook0 putter: but in the case of rules, you can choose your own delimiters 07:51
with sub calls, you're stuck with parens
luqui putter, those are for quoters though
putter hmm....
scook0 foo:adv<>($arg) # illegal ?
Aankhen`` ponders the perverseness of whoever termed these things delimiters.
luqui I think so
scook0 foo :adv .($arg) # insane? 07:52
putter ;)
luqui s/\./,/ and it should work 07:53
but for a very different reason :-)
putter ok. quoters, good point. good night &
Aankhen`` Quoters? 07:55
Nah.
Delimiters is the right word. It's just a strange word. :-)
luqui "quoter" speaking about the rx part, not the //, which are delimiters 07:56
Aankhen`` Ah, I see. I overcompensated for my lag, then.
wolverian Juerd, ah, thanks. 08:33
iblech scook0++ # demagicalizing pairs 11:40
scook0: I'm fixing Pugs.Compile and then PIL2JS currently
svnbot6 r7623 | iblech++ | Pugs.Compile: Compile the new Syn "named". scook0++ for demagicalizing pairs! 11:43
r7624 | iblech++ | * t/syntax/pairs.t: Per 12:46
r7624 | iblech++ | www.nntp.perl.org/group/perl.perl6....age/23532, the keys of
r7624 | iblech++ | syntactical pairs should get stringified.
r7624 | iblech++ | * PIL2JS: Demagicalized pairs!
r7624 | iblech++ | * libjs/PIL2JS.js: For now, special PIL2JS.NamedPair class -- named pairs
r7624 | iblech++ | are no longer PIL2JS.Pairs, but PIL2JS.NamedPairs.
r7624 | iblech++ | * libjs/PIL2JS.js, PIL::*: s/pairs/namedpairs/ where appropriate.
r7624 | iblech++ | * PIL::PApp, PIL::PVal: Efficient &Pugs::Internals::named_pair.
r7624 | iblech++ | * P6 Prelude::JS:
r7624 | iblech++ | * *$normal_pair evaluates to a magical pair.
r7624 | iblech++ | * Enabled code for *%hash (was always there, but because of a small
r7624 | iblech++ | oversight not used).
r7624 | iblech++ | * This means PIL2JS passes 16/17 of pairs.t! (And the single failing test
r7624 | iblech++ | is because of Pugs.Parser, not PIL2JS). Hurray for demagicalized pairs :))
scook0 iblech: which pairs.t test-fail is caused by Parser? 13:05
iblech scook0: foo @array => 42 -- this doesn't get a Syn "named" 13:06
scook0: (It's the last test of pairs.t) 13:07
scook0 now that I check...I don't think *any* pairs that don't have a bareword on the lhs get parsed properly
iblech yep, the all seem to get converted into &infix:«=>»($lhs,$rhs) 13:08
scook0 seems the rule I'm using only matches the bareword variant
other ones are probably matched using normal operator parsing 13:09
iblech ah, yes, makes sense
svnbot6 r7625 | scook0++ | * More sub-call parse tests 13:16
r7626 | iblech++ | t/syntax/pairs.t: More tests: 13:28
r7626 | iblech++ | * :a is a named, but (:a) is pair.
r7626 | iblech++ | * $arrayref => ... should stringify $arrayref (but ($arrayref => ...) should not).
eric256__ what do demagicalized pairs by us? 13:59
?eval sub x (@y) { @y.perl.say }; @y( x=> 2); 14:01
evalbot_7626 Error: Undeclared variable: "@y"
eric256__ ?eval sub x (@y) { @y.perl.say }; @x( x=> 2);
evalbot_7626 Error: Undeclared variable: "@x"
eric256__ ?eval sub x (@y) { @y.perl.say }; x( x=> 2);
evalbot_7626 Error: Insufficient number of required parameters: 0 actual, 1 expected
eric256__ ?eval sub x (@y) { @y.perl.say }; x( (x=> 2) );
evalbot_7626 [("x" => 2)] bool::true
eric256__ ?eval sub x (*@y) { @y.perl.say }; x( x=> 2 ); 14:02
evalbot_7626 [] bool::true
eric256__ ?eval sub x (*@y) { @y.perl.say }; x("test", x=> 2 );
evalbot_7626 ["test"] bool::true
eric256__ ?eval sub x (Pair *@y) { @y.perl.say }; x("test", x=> 2 );
evalbot_7626 ["test"] bool::true
eric256__ ?eval "hello".trans("h" => "t"); 14:03
evalbot_7626 "hello"
eric256__ you broke trans!!!
;)
?eval "hello".trans(( "h" => "t" ));
evalbot_7626 "hello"
eric256__ ?eval "hello".trans( ("h" => "t") );
evalbot_7626 "hello"
eric256__ hmmm
?eval "hello".trans( (h => "t") ); 14:04
evalbot_7626 "hello"
iblech eric256__: Demagicalized pairs are extremele important. In short: foo(bar => $baz) does now always pass $baz by the name "bar" to &foo. foo((bar => $baz)) does now always pass the pair (bar => $baz) positionally to &foo. foo($pair) does now always pass $pair positionally to &foo, even if $pair happens to contain a Pair object.
eric256__ gotcha...so how would we get trans( Str $self; Pair *@intable) to always take a list of pairs and not named parameters? 14:05
iblech eric256__: Consider you wrote a generic subroutine, sub foo ($thing) { ...; bar($thing); ... }. Your &foo would probably break if somebody passed a pair to &foo, as this pair would get intepreted as a named arg by &bar. 14:06
eric256__ right. i understand, i think. ;) 14:07
iblech Larry meant we should use <== (so "hello".trans(<== "h" => "t", ...), IIUC)
eric256__ but how do i say, my sub doesn't accept named args so take named args as a list of pairs?
iblech Or we have to add parens around each pair ("hello".trans((a => "b"), (c => "d")))
eric256__ okay thats what i was wondering
i wonder why the parns aren't working right now though. 14:08
?eval "hello".trans( (h => "t") );
evalbot_7626 "hello"
eric256__ ?eval sub x (Str $t, Pair *@intable) { @intable.perl.say }; x("t", (h => t)); 14:09
evalbot_7626 Error: No compatible subroutine found: "&t"
iblech Yep, seems that the vanilla core has some kind of bug
eric256__ ?eval sub x (Str $t, Pair *@intable) { @intable.perl.say }; x("t", (h => 't'));
evalbot_7626 [("h" => "t")] bool::true
eric256__ there are two cores? thats the exact same signature and it works outside perlude.pm 14:10
?eval sub x (Str $t, Pair *@intable) { @intable.perl.say }; x("t", (h => 't', q => 'r'));
evalbot_7626 [("h" => "t"), ("q" => "r")] bool::true
iblech Yep, there is the vanilla Haskell core, the JavaScript compiler, and the compiler to Perl 5 14:11
ah! Try rebuilding the Prelude
eric256__ still feels like there should be a way to say i'm expecting pairs so don't pars as args, but i'm guessing thats at the heart of magical vs. unmagical
iblech++
hadn't thought of that../me starts full rebuild
do i have to make clean to force it to rebuild prelude? 14:17
iblech eric256__: make clean works for sure, but modifying src/perl6/Prelude.pm (e.g. insert a space somewhere) will suffice 14:18
eric256__ slaps forhead....yea that would have worked, and triggered feather to rebuild too. lol
where you around last night when i was working on .perl ? 14:19
iblech no, but I did saw your checkins
eric256__ do you know why when printing refs...it prints each layer and breaks the brackets off the previous layer until it reach the value? fi you have $ref1 -> $ref2 -> Hash; it doe sthe first one, then does the second one breakcing the { } off the first and adding its own 14:20
wouldn't it make sense just to skip printing anything till you get to the actual value? 14:21
iblech ?eval my %hash = (a => 1); my $ref2 = \%hash1; my $ref1 = \$ref2; $ref1.perl.say
evalbot_7626 Error: Undeclared variable: "%hash1"
iblech ?eval my %hash = (a => 1); my $ref2 = \%hash; my $ref1 = \$ref2; $ref1.perl.say
evalbot_7626 {("a" => 1)} bool::true
eric256__ you can't see it there because its removing the { } and then adding them agian. ever level 14:22
but i changed the front { to hash... and i get hash(ash(1=>2))
iblech Hm. The vanilla core always derefs references -- this might be important here 14:23
eric256__ which then took me an hour to trace around enough to realize that the multiple levels for ref created by for @z ( where driving me nuts
nothingmuch morning
eric256__ morning
iblech Hi nothingmuch :) 14:24
eric256__ if you look Prim.hs 1349 you'll see the function i'm talking about. it gets called twice if you do sub x (@y) { for @y ( $_.perl.print ) }; x({hello=>1}); 14:25
?eval sub x (@y) { for @y ( $_.perl.print ) }; x({hello=>1});
evalbot_7626 \"perl6"Error: Pattern match failure in do expression at src/Pugs/Eval.hs:419:4-12
eric256__ ?eval sub x (@y) { for @y ( $_.perl.print ) }; x(1);
evalbot_7626 \"perl6"Error: Pattern match failure in do expression at src/Pugs/Eval.hs:419:4-12
iblech ?eval sub x (@y) { for @y { $_.perl.say } }; x(1) 14:26
evalbot_7626 \1 undef
eric256__ thats what i was going for. lol spaces ?
?eval 1
evalbot_7626 1
iblech Actually, this call should die -- 1 is not an array. [1] or (1,) should work. Or, alternatively, the signature should read (*@y)
eric256__ see how the 'for' makes a ref to 1 though 14:27
iblech eric256__: It's for @array {...}, not for @array (...)
Ah
eric256__ that ref adds on top of exisiting refs, so if you pass it an array ref, you get a ref to an array ref
duh
lol
iblech This is not a "real" ref, I think. It's because $_ is a full container, not only a value. See:
?eval 1
evalbot_7626 1
iblech ?eval my $foo = 1
evalbot_7626 \1
eric256__ hmmm.. well if my moded version { x => 1 }.perl = "hash( (x => 1) )" while if use that sub above it prints "hash(ash( (x => 1))" 14:29
iblech Ah, I get your problem now, I think. But anyway, { x => 1 } should .perl to either '{ x => 1 }' or '\hash((x => 1))'. And hash((x=>1)) should .perl to 'hash((x=>1))' 14:31
eric256__ yea and since i couldn't get the inner pairs to drop parents, i was going with hash or \hash...didn't matter till i could get it to print right though 14:32
iblech ah, ok 14:34
eric256__ so i was trying to figure out a way in haskell to get to the bottum value instead of printing each one and stripping the { } off at different levels.
eric256__ scratches his head....actualy it looks like it might be fixed already...oddnees 14:36
iblech scook0: I just ran a new JS smoke with the demagicalized pairs patch in -- it seems that the only problem is that foo("x" => 1) is not recognized as a named (same bug as with foo(@array => 1)). 14:39
svnbot6 r7627 | iblech++ | PIL2JS: PIL::PApp: Properly handle complex structs (especially coderefs) as the
r7627 | iblech++ | RHS of a named (e.g. FETCH => {...}) (one-word patch).
iblech scook0: But no other regressions :)
svnbot6 r7628 | iblech++ | src/perl6/Prelude.pm: Extremely minor cosmetical fix, so the Prelude gets 14:44
r7628 | iblech++ | rebuilt (needed by scook0++'s demagicalized pairs patch).
eric256__ ?eval '\hash(1,2)'.eval.ref 14:48
evalbot_7627 *** No compatible subroutine found: "&eval" at -e line 11, column 5-47
eric256__ ?eval \hash(1,2).ref
evalbot_7627 *** No compatible subroutine found: "&eval" at -e line 11, column 5-37
eric256__ ?eval hash(1,2).ref
evalbot_7627 *** No compatible subroutine found: "&eval" at -e line 11, column 5-35
eric256__ ?eval 1; 14:49
evalbot_7627 *** No compatible subroutine found: "&eval" at -e line 11, column 5-21
eric256__ ohh....okay. ?
?restart 14:50
?eval 1 14:51
evalbot_7628 1
eric256__ ?eval hash(1,2).ref
evalbot_7628 ::Hash
eric256__ ?eval \hash(1,2).ref
evalbot_7628 \::Hash
eric256__ ?eval {1,2}.ref
evalbot_7628 ::Block
eric256__ ?eval {1 => 2}.ref
evalbot_7628 ::Hash
eric256__ ?eval \{1 => 2}.ref
evalbot_7628 \::Hash
nothingmuch eric256__: this is discussed in s04 i think 14:52
geoffb: ping
iblech eric256__: \ binds stronger than method application, i.e. you've to use \(hash(...).ref)
nothingmuch iblech: that seems odd 14:56
it just doesn't look right syntactically
iblech It's the same with ~long.expression(here) 14:57
(And, FWIW, it does look right to me, but I don't have a strong opinion on this)
eric256__ in pugs source...readRef gets the next layer of ref...unless i'm still way off base, is there a function that gets the actual value of the thing eventualy referenced? 14:58
iblech eric256__: The problem is: you may only dereference exactly one real level, where "real levels" get created by prefix \, see t/var/autoderef.t. So a function which will dereference as often as needed to get the actual value would probably be against the spec 15:01
eric256__ hmmm. okay. so your saying if i'm in a for that is forcing an additionaly layer of ref, then when i perl print that it should be the extra ref? 15:02
iblech eric256__: (But I do see your problem -- the vanilla core does a) not make a difference between real refs as created with \$foo and containers, and b) the vanilla core dereferences unlimitedly
eric256__ hehe..great 15:03
iblech eric256__: Quite. The "additional layer of ref" you're seeing is not a real reference (which should get .perled), but only an indicator that you're .perling a container, not value. And this is a bug in pugs's internals, which is probably hard to fix
eric256__ but if i ignore the extra layer of ref then i might be ignore a true reference somewhere else? but would a reference to a reference to a value, always be a reference to a value? (as far as .perl is concerned) 15:04
iblech Exactly this is the problem, you can't know which refs are true refs and which aren't... 15:05
Yep. (\\5) should .perlify to \\5. 15:06
eric256__ ?eval \\5.ref
evalbot_7628 Error: unexpected "\\" expecting program
eric256__ ?eval (\\5).perl
evalbot_7628 Error: unexpected "\\" expecting expression
eric256__ ?eval (\ \5).perl
evalbot_7628 "\\5" 15:07
eric256__ ?eval (\ \5).ref
evalbot_7628 ::Int
eric256__ the docs say .perl just needs to output something that will eval back to the same ref type.
eric256__ decides it must be time for cerial.
iblech Here you see the bug -- (\\5) should be a Ref of Ref of Int, not a Int 15:08
eric256__ ohhh
that seems like a pretty serious bug
?eval (\5).ref
evalbot_7628 ::Int
eric256__ ?eval \(\5).ref
evalbot_7628 \::Int
eric256__ ?eval \5.ref
evalbot_7628 \::Int
iblech Yep. The new Haskell runcore will fix this (and it is already fixed in the JavaScript runcore)
eric256__ when does the new one come out? becuase i'm not going to bother on .perl for now if its going to be fixed better soon ;) 15:09
iblech autrijus has already started (see src/PIL*). IIRC porting stevan's metamodel to Haskell has got the highest priority ATM 15:10
eric256__ gotcha...guess i'll look for other tests to fix then. thanks 15:11
drove me nuts all last night. ;)
iblech nod... 15:12
rafl Juerd: ping 15:31
gaal hello everybody! has anyone here built their own ghc? 15:34
i want Hoogle to do GHC libraries, and the easiest way to get that is to run feather:~gaal/src/hihoo/hihoo.pl on a compiled GHC tree :-) 15:35
but ghc doesn't build correctly on feather or on my machine at home.
rafl Juerd: ping 15:36
gaal rafl: you're actually a likely candidate and you were logged out exactly as i was asking. do you happen to have ghc build from source yourself? 15:37
apt-get source ghc6 doesn't build out of the box. 15:38
Juerd rafl: pong
nothingmuch is irc.perl.org dead? 15:42
gaal colabti.de/irclogger/irclogger_logs/perl6 is alive...
eric256__ irc.pugscode.org maybe?
gaal oh, .perl.org, you're looking for #parrot? 15:43
ESCdd then.
nothingmuch esc dd?
oh
eric256__ how do you maodify the env for commands you are going to run with system? 15:44
gaal env CAT=foo command
oh, from within Perl?
eric256__ yea
gaal { local $ENV{CAT} = 'foo'; system ... } # p5 15:45
nothingmuch eric256__: you can '{ local $ENV{xxx} = "foo"; system(@cmd_and_args) }'
gaal { temp %*ENV<CAT> = 'foo'; ... } # p6, but won't work yet
nothingmuch for a channel with 188 users #irssi is pretty dead
wolverian nothingmuch, probably because it's just a help channel nowadays 15:48
nothingmuch i guesso
wolverian nothingmuch, if you want development, #irssi2 (although I don't know how active it is either)
nothingmuch i don't really =)
i wanted help
eric256__ oh i was hoping to do it in p6...could i system('env CAT=foo command') 15:49
nothingmuch eric256__: not portably
eric256__ okay. i think i have a way around it. if a set the evn in a shell script, then run a perl script from the shell script, the env should carry over right? 15:50
iblech eric256__: %*ENV<CAT> = "foo" does work, only temp on a hash element does not
rafl gaal: That should be an RC bug. 15:51
gaal: I'll try to reproduce it.
eric256__ ohh. i don't care if i can temp it...do i?
gaal this RC bug?
iblech I.e. you have to my $old_cat = %*ENV<CAT>; %*ENV<CAT> = "foo"; system "..."; %*ENV<CAT> = $old_cat
gaal s/this/what is an/ 15:52
rafl gaal: Release critical. I'll report it for you if you want.
Juerd: I wanted to ask if I can install tetex-extra and mzscheme on feather, but I just did it and removed it afterwards. 15:53
gaal: What's the error you get when building ghc? 15:54
gaal rafl: okay. I just tried ./configure ; make
and there were a few C errors
Juerd rafl: You can install anything via apt-get
rafl: And anything else in your home directory
rafl gaal: Do dpkg-buildpackage -us -uc -rfakeroot
gaal the first one looked pretty silly, an undeeded static declaration, which i removed
but there were other problems later on. 15:55
Juerd rafl: No debian packages without repositories in /etc/apt/sources.list, please though
That sucks when libraries are upgraded :)
rafl Juerd: Aye.
gaal rafl: what do you need me to do? 15:57
rafl gaal: Do apt-get source ghc6; cd ghc6.4-4.1; dpkg-buildpackage -rfakeroot -us -uc and tell me if it produces some errors. 15:58
gaal it does, but i believe they don't have anything to do with my compilation errors.
because the compilation errors are very cish. 15:59
rafl gaal: Paste them. 16:01
gaal does debian have an automatic build system?
paste in a sec
rafl Yes. It's started by doing dpkg-buildpackage -rfakeroot -us -uc in the src directory.
pasteling "gaal" at 192.115.25.249 pasted "ghc6 source build error" (13 lines, 1.2K) at sial.org/pbot/13713
gaal i mean, does every source upload get built for the maintainer? 16:02
rafl: so, this just looks like a bug in the GHC code, apparently simple to fix: remove the 'static' keyword from the declaration from static_objects' 16:03
rafl There are no source-only uploads. Maintainers upload a source package + one binary build for their architecture. Other architectures gets build by build daemons.
gaal but then there are different errors down the road
rafl buildd.debian.org
gaal rafl: so presumably the maintainer ought to know pretty fast about such a breakage? 16:04
rafl I'm also compiling it ATM. Let's see if I can reproduce it.
Maybe you also want to try to build it in a clean chroot (pbuilder).
gaal: No. You don't get mails if the build fails.
gaal rafl: the newest one at buildd.debian.org/build.php?arch=i3...p;pkg=ghc6 is from a while ago! 16:06
rafl gaal: Yes. The 6.4-4.1 version was uploaded using binary-only NMUs for each architecture because it needed by-hand-bootstrapping. 16:07
(ghc6 build-depends on ghc6 and ghc6 wasn't installable)
OK, I need to go. See you in some hours. 16:08
gaal bye rafl. 16:09
xerox Searching for GHC .debs? 16:10
I'd say, use haskell-unsafe: deb haskell-unsafe.alioth.debian.org/archive/i386 . unstable testing stable 16:11
gaal xerox: no, i want a built source tree
xerox Maybe it has deb-src ?
gaal to slurp function signatures from the .hi files
to create a .HOO file
xerox Hoogle does that, iirc
gaal so that Hoogle can finally show libraries 16:12
xerox Ah!
You should talk with ndm. He's doing that for Hoogle3.
gaal # no.perlcabal.org/~gaal/hihoo.pl
(very rough first take) 16:13
xerox: he doesn't seem to be online on #haskell now.
xerox Check #haskell-blah, he talked some mins ago. Though he said something like "I'm drunk" :-P
svnbot6 r7629 | iblech++ | * t/syntax/pairs.t: foo("a" => 42) is named, foo(("a" => 42)) is not. 16:15
r7629 | iblech++ | * Pugs.Parser: Pairs with complex LHS may now be parsed as a Syn "named":
r7629 | iblech++ | foo @array => 42; # named
r7629 | iblech++ | foo (@array => 42); # positional pair
r7629 | iblech++ | * PIL2JS:
r7629 | iblech++ | * PIL::PApp, PIL::PVal: Stringify the LHS of a named pair.
r7629 | iblech++ | * This means PIL2JS passes t/syntax/pairs.t 27/27! :) scook0++
iblech ?eval sub foo { say caller().perl }; foo 16:22
evalbot_7629 Error: Must only use named arguments to new() constructor
iblech Ah, the Prelude needs to be recompiled again
svnbot6 r7630 | iblech++ | src/perl6/Prelude.pm: Again a trivial cosmetical change so the Prelude gets 16:23
r7630 | iblech++ | rebuilt (necessary because of r7622 and r7629).
brother ?eval "ABC".trans("a" => "A", "b" => "B", "c" => "C") 16:28
evalbot_7629 *** No compatible subroutine found: "&eval" at -e line 11, column 5-98
iblech brother: Try again in a minute, feather is probably rebuilding pugs currently 16:29
wolverian wow. that looks a wonderful interface. pairs are so natural for that sort of a thing. 16:31
(really.)
does it work if the value of a pair is more than one character? can I give it closures which it runs on every replacement?
"ABC".trans(a => { ("A".."C").pick })
brother Probaly not with a closure 16:32
wolverian why not? 16:33
brother but "<foo>".trans( '<' => "&lt;", '>' => "&gt;" ) is supposed to work 16:34
?eval "ABC".trans("a" => "A", "b" => "B", "c" => "C")
evalbot_7629 *** No compatible subroutine found: "&eval" at -e line 11, column 5-98
wolverian right, that is similar use to that
brother still not there
wolverian and has nothing to do with my question :)
i.e. why wouldn't closure work there (given that it could be implemented)? 16:35
gaal Juerd: mind if I install + register hs-plugins on feather?
brother ok, wrong answer. It doesn't work (now), not that it wouldn't work (eventually). 16:36
?eval "ABC".trans("a" => "A", "b" => "B", "c" => "C") 16:40
evalbot_7630 "ABC"
brother dohhh
wolverian brother, ah. thanks!
hahaha.
brother you could have told me that first time I tried
wolverian ?eval "ABC".trans(A => "a")
evalbot_7630 "ABC"
brother ?eval "abc".trans("a" => "A", "b" => "B", "c" => "C")
evalbot_7630 "abc"
wolverian oh, wow, => autoquoting doesn't work yet, I guess
s,/wow, //
brother ?eval "abc".trans("a" => "A", "b" => "B", "c" => "C")
evalbot_7630 "abc"
wolverian erm. 16:41
oh, it's broken.
sorry to confuse.
brother ?eval "abc".trans("a" .. "c" => "A" .. "C")
evalbot_7630 "abc"
brother ?eval "abc".trans(["a" .. "c"] => ["A" .. "C"])
evalbot_7630 "abc"
brother ?eval "abc".trans(list ["a" .. "c"] => ["A" .. "C"]) 16:42
evalbot_7630 Error: Unknown syntactic construct: Syn "named" [Syn "\\[]" [Pos (MkPos "<eval>" 1 19 1 29) (App (Var "&infix:..") Nothing [Cxt (CxtItem (mkType "Str")) (Val (VStr "a")),Cxt (CxtItem (mkType "Str")) (Val (VStr "c"))])],Syn "\\[]" [Pos (MkPos "<eval>" 1 35 1 45) (App (Var "&infix:..") Nothing [Cxt (CxtItem (mkType "Str")) (Val (VStr "A")),Cxt (CxtItem (mkType "Str")) (Val (VStr "C"))])]]
wolverian weird that it doesn't error either
brother wow
did I do that?
wolverian well, now it did ;)
?eval list [1]
evalbot_7630 ([1])
wolverian ?eval list [1] => [2]
evalbot_7630 Error: Unknown syntactic construct: Syn "named" [Syn "\\[]" [Pos (MkPos "<eval>" 1 7 1 8) (Val (VInt 1))],Syn "\\[]" [Pos (MkPos "<eval>" 1 14 1 15) (Val (VInt 2))]]
wolverian there you go.
probably not your fault.
iblech ?eval "abc".trans((a => "A")) 16:43
evalbot_7630 "abc"
brother Hmmm, "abc".trans(["a" .. "c"] => ["A" .. "C"]) 16:44
work on my pugs
xerox Here does work too.
iblech ?eval $?PUGS_VERSION 16:45
evalbot_7630 \"Perl6 User\'s Golfing System, version 6.2.10, October 10, 2005 (r7630)"
brother r7628 here
iblech ?eval "abc".trans((["a" .. "c"] => ["A" .. "C"])) 16:46
evalbot_7630 "ABC"
xerox r7573 here
iblech Since r7622, pairs are demagicalized (scook0++). Since r7629 complex named arguments are possible too. So .trans(["a".."c"] => ...) does not pass a Pair object to .trans, but the named argument "a b c" 16:47
Should pugs emit a warning in this case? 16:48
brother So, for a short time r7617 till r7622 trans actually worked 16:49
iblech It should still work if you add parens around the pairs
(It's a bug of the vanilla core that .trans(a => "A") doesn't die with "Unexpected named argument 'a'") 16:51
wolverian oh, but that usage is so nice!
?eval "abc".trans( (["a"] => ["A"]) ) 16:52
evalbot_7630 "Abc"
wolverian I guess I can live with that if I can remove the arrayreferences :)
(I think that was the spec?) 16:53
(wrt named arguments)
brother ?eval "ABC".trans( 'A'=>'a', 'B'=>'b', 'C'=>'c' ) 16:59
evalbot_7630 "ABC"
brother ?eval "XYZ".trans( 'XYZ' => 'xyz' )
evalbot_7630 "XYZ"
iblech wolverian: Yep, removing the [] ought to work
brother: You either need (the not yet implemented) <== or you have to add parens around the pairs. .trans("A" => "a") is parsed as passing the *named argument* "A" (with value "a") to .trans, not as passing the pair ("A" => "a") positionally to .trans 17:00
brother ?eval "ABC".trans( ('A'=>'a'), ('B'=>'b'), ('C'=>'c') )
evalbot_7630 "ABC"
iblech yep, that should work.
wolverian do I need to put the <== inside the parens? 17:01
would this work: "ABC".trans <== a => "A";
iblech dunno 17:02
wolverian I think that works for normal functions
foo ==> map { } ==> bar
probably should work here too. I'll check for :todo tests
brother ok, I give up.
iblech www.nntp.perl.org/group/perl.perl6....uage/23541 (message from Larry, "self" is now an alias for $?SELF) 17:07
www.nntp.perl.org/group/perl.perl6....uage/23541 (message from Larry, "self" is now an alias for $?SELF) 17:10
err, wrong key
r0nny_ hoi there :) 17:15
gaal ahoy r0nny_!
iblech food & 17:17
Limbic_Region anyone having problems building pugs from scratch on winders? 17:19
Limbic_Region just got a GHC out of heap error
I am restarting the build now - but I have my doubts 17:20
mrborisguy anyone here on feather? 17:35
what's the best way to build pugs to test a change? 17:36
I'd suppose it needs to be built locally in my /home somehow
Limbic_Region has an acct on feather but doesn't use it for anything other than IRC when I am at work 17:38
ok - succesfully built 17:44
any particular reason decimals are now showing up as fractions?
gaal mrborisguy: read perlcabal.org/~autrijus/
Limbic_Region ?eval my $foo = 11 / 10; say $foo
evalbot_7630 1.1 bool::true
gaal mrborisguy: then cd into your pugs dir, and do perl Makefile.PL ; nice make
iblech Limbic_Region: you mean: 17:45
?eval say (11/10).perl
evalbot_7630 11/10 bool::true
gaal whoa. is that intended? 17:46
Limbic_Region ?eval my $foo = .1 * (int (rand 2) ?? 1 !! -1); say $foo
evalbot_7630 0.1 bool::true
Limbic_Region ?eval my $foo = .1 * (int (rand 2) ?? 1 !! -1); $foo
evalbot_7630 \1/10
Limbic_Region that's what I mean
gaal ?eval my $foo = .1 * (int (rand 2) ?? 1 !! -1); "$foo"
Limbic_Region ?eval my $foo = .1 * (int (rand 2) ?? 1 !! -1); $foo
evalbot_7630 "0.1"
\1/10
Limbic_Region ?eval my $foo = .1 * (int (rand 2) ?? 1 !! -1); $foo
evalbot_7630 \1/10
iblech Yep, eric256 implemented this yesterday. The only requirement for .perl is, that $thing.perl.eval eqv $thing. Thus showing fractions as x/y is ok and doesn't loose precision
gaal indeed. cool hack! 17:47
iblech The stringification of fractions is unchanged, i.e.:
gaal ?eval my $foo = .1 * (int (rand 2) ?? 1 !! -1); "$foo"
iblech ?eval ~11/10
evalbot_7630 "0.1"
1.1
gaal :-p
Limbic_Region this is definatley not what I mean though
iblech eric256++ # indeed
Limbic_Region if I multiply a decimal by an integer
I don't expect a fraction
?eval my $foo = .1 * (int (rand 2) ?? 1 !! -1); $foo
evalbot_7630 \1/10
Limbic_Region oh wait 17:48
nevermind
mrborisguy gaal: thanks.
Limbic_Region I am using .perl
Limbic_Region kicks himself
iblech Limbic_Region: But if you say() or ~ $foo, you'll see normal 123.456-style output
Limbic_Region this code is a bit old
17:48 r0nny_ is now known as r0nny
Limbic_Region I forgot how it worked 17:48
ok - so nmake required 2 runs for it to actually get all the way through
I ran out of ghc heap
but other than that - groovy 17:49
TTFN all
svnbot6 r7631 | iblech++ | examples/network/evalbot/{evalbot.p6,evalhelper.p5}: Support the other 17:50
r7631 | iblech++ | backends, e.g.
r7631 | iblech++ | ?eval 5 + 3 # normal backend
r7631 | iblech++ | ?eval:js 5 + 3 # PIL2JS
r7631 | iblech++ | ?eval:p5 5 + 3 # PIL-Run
r0nny is there somewhere a description of the pugs thread interface ? 17:53
iblech r0nny: No... you may grep for Thread in src/Pugs/Prim.hs
r0nny darn 17:54
iblech ?eval "does evalbot still work?" 17:55
evalbot_7631 "does evalbot still work?"
iblech ?eval:js "and does -BJS work?"
evalbot_7631 *** Precompiled Prelude doesn't exist yet; precompiling... (You can safely ignore the 'useless use of constant' warnings.) pugs: *** No compatible subroutine found: "&use" at -e line 1, column 1 Use of uninitialized value in print at /home/fibonaci/devel/pugs-base/perl5/PIL2JS/lib/PIL2JS.pm line 142. *** Reading input from "/tmp/gJJDX52T0F"... *** Compiling PIL to JavaScript... Can't bless non-reference value at /home/fibonaci/devel/pugs-ba
r0nny ?eval:c++ cout << "meh - wont work" << endl; 17:56
evalbot_7631 Error: unexpected "c" expecting operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
iblech ?eval:p5 "hello from pil-run"
evalbot_7631 Can't locate Scriptalicious.pm in @INC (@INC contains: /home/fibonaci/devel/pugs-base/perl5/PIL-Run/lib /home/fibonaci/devel/pugs-base/perl5/PIL-Run/.. /home/fibonaci/devel/pugs-base/perl5/PIL-Run/../Perl6-Value/lib /home/fibonaci/devel/pugs-base/perl5/PIL-Run/../Perl6-Container/lib /home/fibonaci/devel/pugs-base/perl5/PIL-Run/../Perl6-MetaModel /home/fibonaci/devel/pugs-base/perl5/PIL-Run/../Perl6-MetaModel/lib /etc/perl /usr/local/lib/perl/5.8.7
gaal w00t! 17:58
?eval say "iblech++ " xx 3 17:59
evalbot_7631 iblech++ iblech++ iblech++ bool::true
iblech ?eval say "thanks" xx 3 # :) 18:00
evalbot_7631 thanksthanksthanks bool::true
r0nny OMG 18:03
pugs-threads are messy to understand 18:04
can anyone tell me, if there is another way, then async, to create a thread ? 18:07
gaal r0nny: nope. these are kind of uncharted water, too. 18:09
r0nny hmm
is there a way, to pass params in a nice way to a asynv block
gaal dunno 18:11
funny pic of autrijus: www.syntaxpolice.org/pics/estoniaBe...ia/14.html
xerox We want names :D 18:14
gaal names?
xerox Of all those people! :-) 18:15
gaal ah, ask syntaxninja on #haskell :)
integral r0nny: pass params? you just use closure surely? sub foobar ($arg) { async { say $arg } } # ie, the sub generates asyncs that are closed over $arg 18:16
r0nny integral: cool - thx 18:17
so i can do a cmdline shell, and "fork" irc bots for different servers to the bg :) 18:18
gaal r0nny: why would you not want to use separate processes for that? 18:19
r0nny gaal: "fork" not fork ;P 18:21
just let them be in the bg as threads
gaal yes, i understand. why *not* use separate processes though? what do threads buy you? 18:25
iblech brother: I think I have fixed $str.trans in r7632
svnbot6 r7632 | iblech++ | * src/perl6/Prelude.pm: Fixed $str.trans (trivial one-char patch).
r7632 | iblech++ | * t/var/constant.t: The initial value of a constant should get evaluated at compile-time.
iblech gaal: inter-bot-communication :)
r0nny no 18:26
its one big bot handles different servers in different threads
gaal iblech: that is a scary thought :)
iblech :D
r0nny btw - ho does variable-sharing betwen threads work in pugs ? 18:27
iblech ?eval my $var; my sub forker () { async { $var++ } }; forker() for 1..20; $var 18:28
evalbot_7631 \19
iblech The framework for threading Pugs uses, STM, guarantees that the internal state etc. doesn't get corrupted 18:29
?eval my $var; my sub forker () { async { $var++ } }; forker() for 1..20; 1 for 1..1000; $var
evalbot_7631 \20
r0nny so basically everything is not local is shared
gaal '1 for 1..1000'? is sleep unsafe? 18:30
?eval sleep 1; say "perhaps to dream"
evalbot_7631 Can't exec "./pugs": Permission denied at examples/network/evalbot//evalhelper.p5 line 46.
gaal ?eval sleep 1; say "perhaps to dream" 18:31
evalbot_7631 *** No compatible subroutine found: "&eval" at -e line 11, column 7-59
gaal heh.
iblech gaal: Yep, it is. Else you could ?sleep 1_000_000 (and this isn't trapped by BSD::Resource/ulimit, as sleep() doesn't burn any cpu time (IIRC))
gaal it had better not ;)
r0nny ?eval my $var; ny sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var
?eval my $var; ny sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var 18:32
evalbot_7631 *** No compatible subroutine found: "&eval" at -e line 11, column 7-229
gaal "ny sub"?
r0nny ops
?eval my $var; my sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var
evalbot_7631 *** No compatible subroutine found: "&eval" at -e line 11, column 7-229
iblech r0nny: Try again in a minute, the Prelude is currently recompiling, thus &eval isn't there yet
r0nny ok 18:33
?eval my $var; my sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var
evalbot_7631 *** No compatible subroutine found: "&eval" at -e line 11, column 7-229
r0nny hmm
iblech s/try again in a minute/try again in two minutes/ ;) 18:34
r0nny ?eval my $var; my sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var
evalbot_7631 *** No compatible subroutine found: "&eval" at -e line 11, column 7-229
r0nny argh
i got no patience
brother iblech: nice, didnt see that 18:35
r0nny jumps around way to nervous, and breaks the src tree
can i try now *dogy eyes*
?eval my $var; my sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var 18:36
evalbot_7631 *** No compatible subroutine found: "&eval" at -e line 11, column 7-229
r0nny o0
theorbtwo Hm, it was doing the 'No compatible subroutine found: "&eval"' thing the other day, too.
r0nny damn - compile it faster 18:37
r0nny changed his mind-state to pre-insane
?eval my $var; my sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var 18:38
evalbot_7631 *** No compatible subroutine found: "&eval" at -e line 11, column 7-229
theorbtwo r0nny: you could get your own copy...
You can tell when evalbot fixes itself; it'll log out then log back in with a new username.
gaal why, btw, are you declaring your subs 'my' at all?
theorbtwo Yeah, like that, but it's supposed to log back in agin... 18:39
r0nny ?eval my $var; my sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..10000; $var
theorbtwo Thank you, evalbot.
r0nny hope it eats it
evalbot_7632 (no output) 18:40
iblech ?eval "abc".trans((a => "A"))
evalbot_7632 "Abc"
theorbtwo r0nny: Why are you trying to "eat" evalbot?
iblech r0nny: s/10000/1000/ (evalbot has a cpu time limit)
r0nny ehm - mean hope evalbot eats my code, and wont spill it out
?eval my $var; my sub bsleep() { 1 for 1..1000;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..1000; $var
evalbot_7632 \undef 18:41
r0nny ?eval my $var; my sub bsleep() { 1 for 1..10;} ; my sub forker () { async {bsleep for 1..rand(20); $var++ } }; forker() for 1..20; 1 for 1..1000; $var
evalbot_7632 \20
r0nny ?eval use xxo;
evalbot_7632 pugs: *** No compatible subroutine found: "&use" at Prelude.pm line 61, column 30-59
r0nny ?eval my sub test(){use xxo;xxo->new;}; 18:42
evalbot_7632 pugs: *** No compatible subroutine found: "&use" at Prelude.pm line 61, column 30-59
r0nny it is secure
svnbot6 r7633 | iblech++ | t/operators/adverbial_modifiers.t: Fixed some tests WRT pairs now being 18:52
r7633 | iblech++ | demagicalized.
r7634 | iblech++ | New docs/journal/2005-10-15.kwid: scook0 demagicalized pairs; I updated PIL2JS 18:58
r7634 | iblech++ | accordingly; PIL2JS passes 27/27 of t/syntax/pairs.t.
geoffb nothingmuch, I have not backlogged yet . . . but I saw you pinged? 19:00
nothingmuch no longer relevant 19:01
because I have to go
geoffb ah
nothingmuch but I wanted to start writing ;-)
geoffb nothingmuch, I am gone most of today. Back online tomorrow
nothingmuch okies
Juerd gaal: If it won't break automatic upgrades, go ahead 19:22
gaal: If it's available through debian packages, use those.
wolverian hm, now I have my own "feature" in perl6 as well. 19:24
Juerd WHich?
wolverian self is called 'self' by default.
Juerd And .foo is $_.foo unconditionally?
wolverian I assume so. the decision only pertained to naming, but that makes sense. I'll follow up if you want. 19:25
actually, I guess I'll follow up anyway
as it's important to me as well
Juerd, or did you already? 19:26
Juerd Nafaik
wolverian how can you not know? :)
Juerd I have faulty memory 19:27
wolverian oh, okay. well, sending.
Juerd It's not okay
Not remembering things sucks
wolverian I'm sorry. 19:28
sent. 19:29
rafl gaal: ghc6 (sid version) built fine in my chroot. 20:31
gaal: I don't know where your problem is. Please also try to build it in a clean environment.
svnbot6 r7635 | iblech++ | Moved primitive_named_params.t and passing_the_pair_class_to_a_sub.t out of 21:05
r7635 | iblech++ | pugsbugs/, as they pass now (because of the demagicalizing pairs patch).
wolverian does S02 really mean to imply that you can use regexes as q:to (heredoc) terminators? 22:38
rafl Pugs is in Debian testing, yay! (6.2.8, though. 6.2.10 needs to go though the NEW queue first) 22:45

Powered by Cro and the Raku® Programming Language.

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

Thank you!