Pugs 6.2.8 released! | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 13 July 2005.
brentdax Is there a Perl 6 tail call syntax, and if so is it implemented in Pugs? 00:36
autrijus &sub.goto(...); 00:37
and yes.
brentdax: do you see a problem with Block params default to Item|Pair|Junction -- i.e. any ? 00:42
s/any/Any/
brentdax Just that the for 1|2|3 { ... } behavior is no longer automatic.
autrijus but it is of dubious value anyway 00:43
brentdax Oh...are tail method calls possible?
True.
autrijus tail method calls. hrmph. 00:44
what would you suggest as a syntax?
brentdax Basically, it means that a lot of things that were expected to autothread before won't. But that's not necessarily a bad thing, because those were mostly surprising places to autothread.
autrijus (tail methcalls is only not there because of unspeccedness)
brentdax I don't know...Perl 6 doesn't have an equivalent of a C# delegate (besides general-purpose currying, that is). In the worst case, you could do something along the lines of $obj.can('method').goto($obj, ...); 00:47
But clearly that's rather ugly. Maybe $obj.\method can give you a reference to the method, pre-curried to $obj. 00:50
autrijus maybe .can('method') should precurry to $obj anyway. maybe. 00:51
goto $obj.can('method'): @args;
anyway, try p6l it and we'll see how to implement that :) 00:52
brentdax Will do, soonish.
autrijus eggcellent. 00:53
Khisanth first ./ now .\? :/ 00:58
coral autrijus: with Quantum::Entanglement it's possible to tell it to pick one of the entangled numbers in the junction
how do i do that with (0|1)?
autrijus Khisanth: I don't think it's a good idea :) 00:59
?eval (0|1).pick
evalbot6 1
coral aha
autrijus ?eval (0|1).pick
coral thanks
evalbot6 1
autrijus ?eval (0|1).pick
evalbot6 1
autrijus mmmmm.
coral ?eval ((0|1).pick) xx 10
evalbot6 (0, 0, 0, 0, 0, 0, 0, 0, 0, 0)
coral hmm
Khisanth autrijus: I am envisioning a perl6 variant called toothpick
autrijus Khisanth: I \\ /|\ | \//|\ \
coral ?eval sub decision () returns Int { (0|1).pick }; (decision) xx 3 01:00
evalbot6 (1, 1, 1)
coral ?eval sub decision () returns Int { (0|1).pick }; (decision) xx 50
evalbot6 (0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)
Khisanth that looks like a bridge :)
autrijus coral: xx is merely repeat
coral hm
so it evals left side and THEN repeats
is there a way to force left side to eval *after* the repeat occurs? 01:01
autrijus ?eval map { (0|1).pick } 1..10
evalbot6 (0, 0, 1, 1, 1, 0, 1, 0, 0, 1)
coral aha!
unf!
thank you :)
autrijus np :)
coral ?eval map { (0|1).pick } 1..any(5..10)
evalbot6 ((0 | 1))
coral blinks
?eval map { (0|1).pick } 1..5
evalbot6 (0, 1, 1, 1, 1)
coral ?eval map { (0|1).pick } 1..(any(5,10)) 01:02
evalbot6 ((0 | 1))
coral ?eval any(5,10)
evalbot6 5
coral i don't get it
autrijus ?eval 1..(5|10)
evalbot6 (1, 2, 3, 4, 5)
autrijus ?eval 1..(5|10)
evalbot6 (1, 2, 3, 4, 5)
autrijus ?eval 1..(5|10)
evalbot6 (1, 2, 3, 4, 5)
autrijus ?eval 1..(5|10)
evalbot6 (1, 2, 3, 4, 5)
autrijus hrm.
coral well, i think there's one or two bugs in that somewhere
autrijus what's wrong with the randomizer.
coral ?eval 1..any(1)
evalbot6 (1)
coral ?eval 1..any(10)
evalbot6 (1, 2, 3, 4, 5, 6, 7, 8, 9, 10)
autrijus ?eval map { 1 } 1..2
evalbot6 (1, 1)
coral ?eval 1..any(2,4)
evalbot6 (1, 2)
autrijus ?eval map { 1 } 1..any(2,2) 01:03
evalbot6 ((1))
autrijus ?eval map { 1 } 1..any(3)
evalbot6 ((1))
autrijus ?eval map { 1 } 2..any(3)
evalbot6 ((1))
coral neato
i think 2..any() is busted
autrijus only as argument to map
coral ?eval map { (0|1).pick } 1..(4|5|6).pick
evalbot6 (0, 1, 1, 1)
autrijus I think what happens is that it autothreads over map
coral ?eval map { (0|1).pick } 1..(4|5|6|7|7|7|7|7|7|7|7|7|7|7).pick
evalbot6 (0, 1, 0, 0, 1, 1, 0)
coral hey, that worked. neato. 01:04
autrijus as a list of junctions with one element
and recombine back to any(1,1,1,1,1)
which is ((1))
coral nods
?eval 0^1 01:05
evalbot6 0
coral ?eval 2^4
evalbot6 2
coral ?eval my @code = (0|5).pick..(6|10).pick; @code 01:07
evalbot6 [5, 6]
coral ?eval my @code = (0|5|5|5|5|5|5|5|5).pick..(6|10|10|10|10|10|10).pick; @code
evalbot6 [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
coral bows.
stevan rehi autrijus 01:10
if you have a spare moment, can you please look over perl5/Perl6-MetaModel/t/80_Code.t
I need an (in)sanity check on the class layout :) 01:11
autrijus looks excellent 01:13
stevan autrijus: :)
thank you
autrijus :) 01:14
stevan ok, nothingmuch and I are going to prototype some MMD stuff this weekend
autrijus worries a bit about name clash with existing CPAN Perl6::* modules
stevan yes, I thought about that too
I am punting on it for now though 01:15
autrijus ponders Perl6::Runtime::Signature, P6::Signature, etc
sure
stevan hmm, P6:: might be nice
mugwump stevan: out of interest, why wouldn't you use check_params at new() time ? 01:16
I just hate seeing unmarshalled @_ going into objects ;)
(reminds me of the Tangram code base too much)
stevan mugwump: where specifically?
mugwump oh, wait, duh, this is a test script. I just saw the packages 01:17
and my brain flipped into "you are reading a module" mode
stevan mugwump: yes, it was just a first draft
s/was/is/ 01:18
autrijus hm, I really need to find some breakfast, lest I'll be late, as in late autrijus tang
bbiab & 01:19
mugwump Maybe I should write a version of Class::Tangram::Generator that makes Perl6::Class objects 01:26
hmm, no, it would have to work from T2 objects. 01:31
that seems like an interesting distraction 01:32
autrijus rehi. 02:12
autrijus praises food
cm is hungry 02:16
Khisanth hmm 02:39
mugwump prods svnbot6 03:16
svnbot6 r5722, mugwump++ | A little doodle
r5723, Stevan++ | Perl6::Code - changes execute() to do() because $Larry hates COBOL
stevan mugwump: very interesting
mugwump You like? :D 03:17
stevan so can that make Perl6::MetaModel classes from a DB schema?
mugwump so long as (DB schema) does Tangram::Schema
stevan nice
mugwump I have an example schema that is modelled against an RT schema in the Tangram distribution. There is also a simpler "musicstore" schema 03:20
stevan very nic
e
so using this I can build normal Class::Tangram classes, and they will make Perl6::MetaModel classes? 03:21
svnbot6 r5724, Stevan++ | Perl6.MetaModel : JS port of Perl6::MetaModel
r5724, Stevan++ | * only (partial) Perl6.MetaClass so far, but the C3 MRO is ported
r5724, Stevan++ | - some tests as well
stevan iblech: that one is for you :P
autrijus say what? 03:22
stevan :P
autrijus wow, I see some loving from JSAN
maybe this can indeed hit JSAN ;)
cwest: look! Perl6/MetaClass.js! 03:23
stevan we will see how the actually class builder part turns out, the MetaClass is the easy part
autrijus JSAN-compliant even
stevan: perl5/ ? :)
cwest woo!
autrijus ponders jsan/ or js/
stevan autrijus: I wasnt sure where to put it actually 03:24
mugwump stevan: Class::Tangram generates accessors, it doesn't try to represent them. T2::Class is the "meta" object, if you like
stevan mugwump: ok, I think I get it. I will wait for some more tests to really grok it 03:26
jdv79 how does one generate js? 03:56
autrijus jdv79: see PIL2JS/README
perl5/PIL2JS/README
jdv79 oh, thanks! 03:57
autrijus :) 03:59
brentdax Hmm...are there any parts of Pugs that are especially slow? Each page request is taking several seconds of max CPU right now... 04:03
autrijus yes, the evaluator is especially slow. 04:04
it will be faster in the new evaluator, and is already faster if compiled to PIR.
brentdax Heh. 04:05
autrijus i.e. known issue, working on it :)
(we apologise for the inconvenience)
brentdax (Code base is 700-odd lines, and uses both Perl 5 interop and some simple regexen via Parrot embedding.)
autrijus nod... likely not your problem. 04:06
brentdax Alright. I was thinking that maybe the pattern matches might be a problem, and that I should use :p5 instead. 04:07
autrijus you can bench, but I don't think that is the bottleneck.
brentdax I don't think I'll bother, then--this doesn't have to be fast yet, it just has to work. 04:13
luqui none(): poke 04:16
jdv79 mandel to js in FF ran a little faster this time:) cool. 04:19
luqui there's a js backend now? 04:21
stevan luqui: yes 04:22
brentdax I'm not sure if the JS backend is the coolest or most twisted thing I've ever heard of. Or maybe both. 04:26
cwest brentdax: My vote is both. 04:47
05:57 calvin__ is now known as calvin_
svnbot6 r5725, Stevan++ | Perl6.MetaModel - 06:24
r5725, Stevan++ | * added Perl6.Method class
r5725, Stevan++ | - tests for this
r5725, Stevan++ | * added Perl6.MetaClass.Dispatcher class
r5725, Stevan++ | - added tests for this
r5725, Stevan++ | * added dispatcher support and method support for MetaClass
r5725, Stevan++ | - added tests for that
Darren_Duncan I've just been working on writing my Lightning Talk now ... 08:07
and boy is 5 minutes ever short
near as I can tell, it takes 5 seconds to read a line of text, so I have to max out at 60 full lines, or better yet 60 partial lines to allow pauses and breathing
that's not very much at all
my original intro was 2.5 minutes itself, and I'll have to chop it way down 08:08
anyway, that's all fyi
kks How to call a shell script from a perl cgi? 08:19
QtPlatypus I think that you need #perl rather then #perl6 08:20
kks QtPlatypus, thanks 08:24
QtPlatypus No problem. 08:26
?eval class Foo {...};ref(Foo); 10:49
evalbot6 *** ... - not yet implemented at <eval> line 1, column 12-15
QtPlatypus ?eval class Foo {};ref(Foo);
evalbot6 ::Class
10:52 Aragone is now known as Arathorn
svnbot6 r5726, fglock++ | * Added tests for negative infinity stringification 13:39
r5726, fglock++ | (-Inf ne '-Inf')
r5727, fglock++ | (fixed test count)
r5728, fglock++ | * Tests int() of special values Inf,-Inf,NaN 13:51
dada ?eval ref('Foo'); 14:39
evalbot6 ::Str
dada ?eval ref(Foo);
evalbot6 Error: No compatible subroutine found: "&Foo"
dada ?eval my $sub = \&print; ref($sub);
evalbot6 Error: Undeclared variable: "&print"
dada ?eval my $sub := \&print; ref($sub); 14:40
evalbot6 Error: Undeclared variable: "&print"
dada ?eval my $sub = sub { 42 }; ref($sub);
evalbot6 ::Sub
dada ?eval my $array = (1,2,3); ref($array); 14:41
evalbot6 ::Array
nothingmuch home.cwru.edu/%7Ejnt5/Planarity/ 14:47
nothingmuch is already at level 12
productivity is down
but fun is up
dada nothingmuch: Planarity++ (even +++) 15:07
nothingmuch oh man
Khisanth that thing again 15:18
dada ?eval ref 42 15:36
evalbot6 ::Int
dada ?eval ref \42
evalbot6 ::Int
dada ?eval ref 1,2,3
evalbot6 (::Int, 2, 3)
dada ?eval ref(1,2,3)
evalbot6 Error: No compatible subroutine found: "&ref"
dada ?eval ref [1,2,3]
evalbot6 ::Array
QtPlatypus ?eval ref (1,2,3) 15:37
evalbot6 ::Array
Aankhen`` ?eval my $foo = 10; $foo is random_trait; 15:42
evalbot6 undef 15:43
Aankhen`` ?eval my &foo := sub () { "foo" }; &foo is random_trait; foo
evalbot6 'foo'
Aankhen`` ?eval module Foo { my &Foo::("foo") := sub () { "foo" }; &Foo::foo is random_trait; foo } 15:44
evalbot6 Error: unexpected "{" expecting trait, ";" or end of input
Aankhen`` ?eval module Foo { &Foo::("foo") := sub () { "foo" }; &Foo::foo is random_trait; foo }
evalbot6 Error: unexpected "{" expecting trait, ";" or end of input
Aankhen`` ?eval module Foo { &Foo::foo := sub () { "foo" }; &Foo::foo is random_trait; foo }
evalbot6 Error: Undeclared variable: "&Foo::foo"
Aankhen`` ?eval module Foo { &foo := sub () { "foo" }; &foo is random_trait; foo }
evalbot6 Error: Undeclared variable: "&foo"
Aankhen`` ?eval module Foo { }; &Foo::foo := sub () { "foo" }; foo 15:45
evalbot6 Error: Undeclared variable: "&Foo::foo"
Aankhen`` Bah.
dada QtPlatypus: does the space matters?
QtPlatypus dada: Yes
dada awful
QtPlatypus ref(1,2,3); Means call the three argument form of &ref. There is no three arugment form of ref so it fails. 15:46
dada yes, I understand this
osfameron eeek!
QtPlatypus ref (1,2,3); Means call ref on the array (1,2,3). So it works. 15:47
dada but then, "mysub (1,2,3) means "call mysub with a single arrayref parameter?"
QtPlatypus ref .(1,2,3)l Means call the three arguement form of &ref
dada oops
but then, "mysub (1,2,3) means "call mysub with a single arrayref parameter?"
dada is unable to type
QtPlatypus Thats my (and pugs) understanding. 15:48
dada ?eval sub foo(*@args) { join ", ", @args; }; foo(1,2,3);
evalbot6 '1, 2, 3'
dada ?eval sub foo(*@args) { join ", ", @args; }; foo (1,2,3);
evalbot6 '1, 2, 3'
dada mmm
QtPlatypus ?eval sub foo(@args) { join ", ", @args; }; foo(1,2,3); 15:49
evalbot6 Error: No compatible subroutine found: "&foo"
QtPlatypus ?eval sub foo(@args) { join ", ", @args; }; foo (1,2,3);
evalbot6 '1, 2, 3'
dada I'm not sure *@args should take [1,2,3] as (1,2,3)
?eval sub foo(*@args) { join ", ", @args; }; foo (1,[2,3],4); 15:50
evalbot6 '1, 2 3, 4'
QtPlatypus What do you mean?
dada I expected @args to contain only one element, an arrayref 15:51
oh, but my expectation is wrong, I see
(1,2,3) (which is [1,2,3] in p5 terms) is evaluated in list context here, so it correctly derefs 15:52
QtPlatypus It still is the list context.
dada ?eval sub foo(*@args) { join ", ", @args; }; foo \(1,2,3);
evalbot6 '1 2 3'
dada heh
QtPlatypus ?eval sub foo(@args) { join ", ", @args; }; foo [1,2,3]; 15:53
evalbot6 '1 2 3'
dada ?eval sub foo(@args) { join ", ", @args; }; foo \(1,2,3);
evalbot6 '1, 2, 3'
dada this is very, very subtle
QtPlatypus nods "It is" 15:54
dada but wait, why [1,2,3] comes in as @args[0]?
?eval sub foo(@args) { join ", ", @args; }; my @x = (1,2,3); foo @x; 15:55
evalbot6 '1, 2, 3'
dada ?eval sub foo(@args) { join ", ", @args; }; my @x = (1,2,3); foo \@x;
evalbot6 '1, 2, 3'
dada ?eval sub foo(@args) { join ", ", @args; }; my $x = (1,2,3); foo $x;
evalbot6 '1 2 3'
dada ouch 15:56
QtPlatypus The lession here, use the spaceless form unless you realy realy knwo what your doing.
dada I think the spaced form should elicit a warning, unless you "use mastery;" 15:57
gaal hola 16:03
lumi Hola 16:05
stevan hey gaal 16:08
gaal greetings from the tel-aviv hackathon! 16:11
stevan gaal: nice
geoffb waves in the general direction of tel-aviv 16:12
PerlJam tel-aviv hackathon? 16:14
these hackathons are *everywhere* ;)
lumi "Join us..." 16:15
geoffb Need to have one at my house. :-)
stevan PerlJam: if you are going, can you pick me up on the way? My jet is in the shop 16:16
geoffb Of course, not any time soon, what with the coming baby and all . . .
stevan geoffb: hackathon hours are not all that different from new baby hours actually :) 16:18
geoffb Yeah, but both together would have me up the entire time the hackathon was going.
By day 3 my code could be mistaken for fluent Drunkeneese 16:19
stevan geoffb: we could all take turns feeding, burping and changing diapers,... hackers tend to be good at such things :P
geoffb LOL
stevan though in all seriousness, I think gaal has a great idea 16:20
geoffb oh, totally
stevan (mini-local-hcakathons)++
if I didnt think my wife would kill me, I would have one :) 16:22
gaal whoa, gaal also has some grocieries he forgot to put in the fridge. brb 16:41
Jon1 has anybody used placeholders with use perl5:DBI ? 16:48
my @a = $db.selectrow_array('select * from emp where pid = 1'); works for me
my @a = $db.selectrow_array('select * from emp where pid = ?', undef, 1); does not work for me
the parameters are the sql, then a hashref of statement attributes or undef, then the parameters to bind 16:50
in perl 5 I almost always just used undef 16:51
pugs gives me: DBI::st=HASH(0x8968c10)->_prepare(...): attribute parameter 'pugs=SCALAR(0x881e55c)' is not a hash ref at /usr/lib/perl5/site_perl/5.8.4/i686-linux/DBD/Oracle.pm line 321.
geoffb Given what you just said, is there a problem with s/undef/{}/? 16:52
Jon1 what is the p6 syntax for an literal empty hashref? (I've tried various things with no luck) in p5 it was just {} 16:53
geoffb ?eval \{}
evalbot6 \sub {...}
geoffb ah, interesting
?eval {,} 16:54
Khisanth ?eval +{}
evalbot6 Error: unexpected "," expecting ";", statements or "}"
Error: cannot cast from VCode (MkCode {isMulti = False, subName = "<anon>", subType = SubBlock, subEnv = Just (MkEnv {envContext = CxtItem (mkType "Any"), envLValue = False, envLexical = MkPad (padToList [("$?1",[(<ref>,<ref>)]),("$_",[(<ref>,<ref>)]),("$code",[(<ref>,<ref>)]),("$lang",[(<ref>,<ref>)]),("&?BLOCK_EXIT",[(<ref>,<ref>),(<ref>,<ref>),(<ref>,<ref>),(<ref>,<ref>),(<ref>,<ref>),(<ref>,<ref>)]),("&break",[(<ref>,<ref>),(<ref>,<ref>)]),("&
Khisanth !
geoffb ?eval {;}
evalbot6 undef
geoffb !
Jon1 I think in p6 {} is always a closure
geoffb ok
hmmm
Khisanth ?eval hash{}
evalbot6 {('<SubBlock(<anon>)>' => undef)}
geoffb ?eval hash()
evalbot6 {}
geoffb A-ha!
Khisanth ah
so it IS looking more and more like PHP :) 16:55
geoffb evalbot6++ # just for being there when I needed it most
geoffb sees 'PHP' and starts saying 'la-la-la-la-la-la'
Khisanth hmm
?eval {=>}
evalbot6 Error: unexpected ">" expecting term
Jon1 my @a = $db.selectrow_array('select * from emp where pid = ?', hash(), 1); 16:56
DBI::st=HASH(0x8968c10)->_prepare(...): attribute parameter 'ARRAY(0x8968d30)' is not a hash ref at /usr/lib/perl5/site_perl/5.8.4/i686-linux/DBD/Oracle.pm line 321.
Khisanth ?eval {undef=>undef}
evalbot6 {('undef' => undef)}
geoffb Jon1, eww
Jon1, try 'my $empty = hash(); my @a = $db.selectrow_array(..., $empty, 1); 16:57
just to help narrow this one down
Jon1 my @a = $db.selectrow_array('select * from emp where pid = ?', {('undef' => undef)}, 1);
DBI::st=HASH(0x8968c10)->_prepare(...): attribute parameter 'ARRAY(0x8968d30)' is not a hash ref at /usr/lib/perl5/site_perl/5.8.4/i686-linux/DBD/Oracle.pm line 321.
ok
16:57 Maddingue__ is now known as Maddingue
Jon1 my $empty=hash(); my @a = $db.selectrow_array('select * from emp where pid = ?', $empty, 1); 16:58
DBI::st=HASH(0x8968c10)->_prepare(...): attribute parameter 'ARRAY(0x8968d30)' is not a hash ref at /usr/lib/perl5/site_perl/5.8.4/i686-linux/DBD/Oracle.pm line 321.
geoffb pugs seems to have invariance problems with literal things versus same stuffed in variables. Or maybe it is a Perl 6 issue . . .
bah
embed-layer bug, mayhaps 16:59
Khisanth geoffb: perl5 has a bit of problem with that too
geoffb Khisanth, fair enough, but I've still been amazed at how often it comes up in pugs . . . I think my mental model of the internals must be wrong. 17:00
Khisanth well only one I know of in p5 is with undef
geoffb Khisanth, example? 17:01
Aankhen`` ?eval 1 => 2 => 3 => 4
evalbot6 (1 => (2 => (3 => 4)))
geoffb ... and there was much rejoicing 17:02
Aankhen`` Hi.
gaal hey Aanken``!
Aankhen`` pouts.
s/nk/nkh/ 17:03
And: HI gaal. =)
gaal can't type. sorry :)
Aankhen`` What's shakin'?
gaal tlv hackathon's up, yo
Aankhen`` Oooh, awesome.
Man. I'm thoroughly exhausted today. It's only 10:30 PM and I'm already thinking of turning in. 17:04
geoffb would follow up with a similar slangfest, but I just can't bring myself to do so.
Aankhen``, are you in Bangalore?
Aankhen`` Somewhere in India. :-) 17:05
geoffb Dang. Was trying that "Think of a card. Is it the Ace of Spades?" trick
Aankhen`` o_O
With "Bangalore"? 17:06
How does that go?
geoffb You're in the right timezone, and Bangalore is a tech center, so the percentages were good enough to try it. If I'd been right, you'd have gone "How the heck did you know that?" :-)
I think the Ace of Spades and the Queen of Hearts are the most popular cards, so just guessing one of those will get a certain percentage of your audience to be impressed. 17:07
"Think of a number. Is it 3?" Is similar
Aankhen`` Ah, thataway.
?eval (1, 2, 3) >>+ 4 17:08
evalbot6 Error: unexpected ">" expecting operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
Khisanth ?eval (1, 2, 3) >>+<< 4
evalbot6 (5, 6, 7)
Aankhen`` lost his bearings with regard to `>>+` vs. `>>+<<` today. 17:09
Khisanth ?eval sub prefix:<sub prefix:<+> ($val) { $val + 4 } +<< (1,2,3) 17:12
evalbot6 (1, 2, 3)
Khisanth hmm 17:13
oops
?eval sub prefix:<+> ($val) { $val + 4 } +<< (1,2,3)
evalbot6 (5, 6, 7)
Khisanth Aankhen``: need the >> and << unless it's a unary op :)
Aankhen`` Er... wha? 17:14
Khisanth ?eval (1,2,3) >>++ 17:18
evalbot6 Error: Can't modify constant item: VInt 1
Khisanth hmm
?eval my @foo = (1,2,3); @foo >>++ 17:19
evalbot6 Error: Can't modify constant item: VInt 1
Khisanth ?eval my @foo = (1,2,3); ++<< @foo 17:21
evalbot6 Error: Can't modify constant item: VInt 1
svnbot6 r5729, Aankhen++ | * HTTP::Cookies: added `NEXT` block within &add_cookie_header.
r5730, Aankhen++ | * HTTP::Status: (hopefully) better generation of mnemonic subroutines.
r5731, Aankhen++ | * HTTP::Status: fixed mistake in previous commit.
Khisanth why is that a constant? 17:22
Aankhen`` ?eval sub foo (&block) { &block() }; foo(-> { "foo" }) 17:24
evalbot6 'foo'
Aankhen`` ?eval sub foo (*&block) { &block() }; foo(-> { "foo" })
evalbot6 pugs: Internal error: doBindArray: unexpected char: '&' Please file a bug report.
Aankhen`` ?eval sub foo (Code *$block) { &block() }; foo(-> { "foo" })
evalbot6 Error: No compatible subroutine found: "&block"
Aankhen`` ?eval sub foo (Code *$block) { $block.() }; foo(-> { "foo" })
evalbot6 'foo'
Aankhen`` ?eval sub foo (Code *$block) { $block.() }; foo():{ "foo" } 17:25
evalbot6 'foo'
Aankhen`` How come `*&block` doesn't work?
geoffb slurpy block? 17:30
dammit, can't even make unoptimized anymore. 17:32
geoffb goes about trying to clean VM
(sigh)
Aankhen`` Yes, slurpy block.
I would have thought it'd function in the same way as `Code *$block`. 17:33
Khisanth what exactly does a slurpy block mean?
Aankhen`` A block that can be passed as an adverb (see last ?eval above).
G'night. 17:39
svnbot6 r5732, iblech++ | * Usual svn properties.
r5732, iblech++ | * New test added to t/data_types/hash.t:
r5732, iblech++ | my %hash = (a => 1, b => 2, a => 3); say %hash<a> # should be 1, not 3
r5732, iblech++ | (see www.nntp.perl.org/group/perl.perl6....age/22401)
geoffb Hmmm. 17:45
iblech++ # including a link to the ruling 17:46
.oO( But what's wrong with foo(*%defaults, a => 1)? )
gaal hey folks. i'm getting ready to check in the file-based build config stuff; it might break stuff so i need your help testing + fixing :) 17:47
geoffb "file-based build config stuff"?
oh, the YAML thing?
gaal yeah, at last.
geoffb I was in try-to-build mode anyway, so as soon as I can build (aka kill some windows chewing VM), I can help test 17:48
gaal cool, thanks 17:49
geoffb np
gaal sec, i want to make clean once after the svn up to make sure i'm not committing total bollcks :) 17:50
anyone know why $ccdlflags aren't being passed to make ghci today? 17:51
dinner :) & 17:52
iblech stevan: FYI, I'm reading the docs of Perl6[.-]MetaModel now so we can integrate it into PIL2JS today/tomorrow :) 17:55
svnbot6 r5733, iblech++ | PIL2JS: Splitted the main Prelude::JS into smaller parts, e.g. 17:57
r5733, iblech++ | Prelude::JS::Array, Prelude::JS::Code, etc.
gaal hola szabgab 17:59
szabgab hey gaal are you also here ?
gaal i see the hummous got to you. :)
szabgab that and all the other the things I stuffed 18:00
iblech stevan: "Submethods are methods which can only be called by an invocant whose class is the same as which the submethod has been defined in. Meaning they are not inherited."
stevan: But the following does work (I asked Larry): class Foo { submethod blarb {...} }; class Bar is Foo { submethod blarb {...} }; Bar.new.blarb() (calls &Bar::blarb); Bar.new.Foo::blarb() (calls &Foo::blarb) 18:01
Limbic_Region $larry just muddied the decision on hash assignment btw
geoffb It's times like this I remember why I haven't subscribed to p6* 18:02
iblech stevan: The relevant posting is at groups-beta.google.com/group/perl.p...538ded4a3, at the very end 18:03
Hm... WRT to hash creation with duplicate keys: I don't care if we choose the Perl 5 behaviour or Luke's, but I think I definitely want the behaviour to be the same in assignment and binding (consistency...) 18:05
geoffb agreed 18:06
PerlJam iblech: I've come to rely on p5's behavior in a couple of places. I'd prefer that it remain that way :-)
though if it's right-most wins on assignment and left-most on binding I could probably deal.
geoffb That seriously violates least surprise 18:07
PerlJam well ... talk to $Larry about it. 18:08
PerlJam contemplates writing a "science fiction" story where a large community of programmers follow the lead of their chief language designer despite his occasional bizarre decisions only to find out in the end that the language designer has gone crazy mid-way through the design process ;-) 18:10
geoffb takes comfort in the fact that we can make a 0-day Perl 6+i if need be 18:11
svnbot6 r5734, iblech++ | t/oo/submethods.t -- SomeSubclass.Superclass::some_submethod() should work (and
r5734, iblech++ | does), see groups-beta.google.com/group/perl.p...538ded4a3.
geoffb has mentally brutalizing children's songs going through his head 18:15
.oO( as opposed to all of them that *aren't*? )
Must replace with something opposite . . . hmmm, Ozzy in order, I think 18:16
gaal, my current compile still continuing. How close are you to committing?
dudley is reminded of the duet that Ozzy did with Miss Piggy... 18:17
geoffb Man, that would have been fun to watch
Sigh . . . ghc going on 4 min of CPU time for Pugs.Run 18:19
iwlx Huh, only? 18:20
18:20 iwlx is now known as wilx
wilx I remember AST.hs to take much more time. 18:20
geoffb wilx, *just for that one file* 18:21
Oh, perhaps, but this is pass 2.
whee, linked
(using 290 MB of VM . . . I really need to finish up that separate-link-pass patch)
wilx, oh, and unoptimized . . . I can't build optimized at all on this box 18:22
Limbic_Region I am completely confused by Pugs compile times - my work machine (which is beefier) takes forever while my home machine takes no time at all (relatively speaking) 18:23
geoffb Odd . . . same OS and toolchain versions?
integral Limbic_Region: more memory at home? (more free memory?) 18:24
Limbic_Region integral - beefier = more resources all around 18:25
integral hmm, are you using those resources?
Limbic_Region well, not me personally - but it is possible that there are some uber hidden monitoring devices I don't know about - but if there are I can't detect them 18:26
geoffb gaal, OK, previous pugs made/installed/realcleaned -- ready when you are
Limbic_Region hmmm - that's odd
there very well may be hidden processes 18:27
*shrug* - it isn't like I can complain really - what exactly is Pugs and how does it relate to your job would be a tough question to answer 18:28
geoffb Work machine is locked down, yes? I wonder if they went so far as to run the visible env under a VM supervisor . . . .
Limbic_Region yes, locked down. But it works, and that is what counts. 18:30
geoffb sure . . . just theorizing
?eval my $foo = [1, 4, 9]; say $foo[0]; $foo[0]++; $foo[0]; 18:31
evalbot6 1 \2
geoffb hmm 18:32
Limbic_Region makes sense
1 is the return code of say
geoffb sorry, evalbot was failing to replicate a strangeness in my code
?eval my $foo = [1, 4, 9]; say $foo[0]; $foo[0] += 10; $foo[0]; 18:33
evalbot6 1 \11
geoffb ?eval my $foo = {bar => [1, 4, 9]}; say $foo<bar>[0]; $foo<bar>[0] += 10; $foo<bar>[0]; 18:34
evalbot6 1 Error: Can't modify constant item: VInt 1
geoffb Aha!
Caught you, you bastard
?eval my $foo = {bar => [1, 4, 9]}; say $foo<bar>[1]; $foo<bar>[1] += 10; $foo<bar>[1];
evalbot6 Error: Can't modify constant item: VUndef
geoffb urp?
?eval my $foo = {bar => [1, 4, 9]}; say $foo.perl; say $foo<bar>[1]; $foo<bar>[1] += 10; $foo<bar>[1]; 18:35
evalbot6 {('4' => 9), ('bar' => 1)} Error: Can't modify constant item: VUndef
geoffb wah?
?eval my $foo = (bar => [1, 4, 9]); say $foo.perl; say $foo<bar>[1]; $foo<bar>[1] += 10; $foo<bar>[1];
evalbot6 ('bar' => [1, 4, 9]) 4 \14
coral Limbic: "skills development"
geoffb coral: ? 18:36
?eval my $foo = {bar => [1, 4, 9],}; say $foo.perl; say $foo<bar>[1]; $foo<bar>[1] += 10; $foo<bar>[1];
evalbot6 {('bar' => (1, 4, 9))} 4 Error: Can't modify constant item: VInt 4
Limbic_Region coral ?
geoffb fervently hopes that hash parsing wierdness is a pugs issue and not a design misfeature 18:37
?eval my $foo = {bar => [1, 4, 9], baz => 6}; say $foo.perl; say $foo<bar>[1]; $foo<bar>[1] += 10; $foo<bar>[1]; 18:38
evalbot6 {('bar' => (1, 4, 9)), ('baz' => 6)} 4 Error: Can't modify constant item: VInt 4
geoffb ?eval my $foo = {bar => [1, 4, 9], baz => 6}; say $foo.perl; say $foo<baz>; $foo<baz> += 10; $foo<baz>; 18:39
evalbot6 {('bar' => (1, 4, 9)), ('baz' => 6)} 6 \16
geoffb considers his options: 18:41
1. Bang head against wall until bloody
2. Go crazy and get a visit from the nice men carrying a white coat with long sleeves
3. Table current several projects and actually get around to learning Haskell
Dammit, this isn't an easy decision 18:42
PerlJam Learning Haskell has some side benefits. you should choose option #3
iblech geoffb: FYI, your last two ?evals work fine in PIL2JS :) 18:43
geoffb sigh, you're probably right
Limbic_Region geoffb - you could add an option 4 (which is what I have decided to do)
take a break
use.perl.org/~Limbic%20Region/journal/25807
gaal okay, the build system thing - lumi cause me to think about it a bit
which is good 18:44
because i realized that it sucks
s/cause/caused/
it doesn't suck badly
but it does suck somewhat.
geoffb iblech, sure, but I have some very cool things I'd like to do in the pugs world that A) require Perl 5 embedding right now, and B) were making me happy to think about doing *soon*
gaal, fair enough
gaal here's the deal 18:45
it has two purposes:
1. factor out duplicate lines in Makefile.PL
2. centralize config options
(making it possible to put your personal prefs in a file)
(and override them temporarily from env) 18:46
so, #2 achived.
problems is, i removed about 20 lines from Makefile.PL
by adding about 200 elsewhere :)
geoffb I've shelved some of my makefile tweaks in the hope that your stuff would do #1, to make my changes easier
stevan iblech: ping
iblech stevan: pong 18:47
stevan iblech: re: submethods
geoffb gall, Well, just not having 20 different variations on ghc's command line params would be nice . . .
gaal so maybe i'll ci this and let y'all try and see if you like it?
geoffb gaal, That's the Pugs Way.
stevan yes, what larry said was that if you call the submethod bypassing the dispatch table, then all bets are off
see Perl6::SubMethod::force_call 18:48
gaal just have to add a little dep on the config file so that editing it then running make triggers a rebuild.
iblech stevan: ah, ok then
geoffb gaal, cool 18:49
stevan iblech: it is not the idea solution, but it works for BUILDALL DESTROYALL, and I punted on the other details
mostly because I dont have proper packages in the prototype 18:50
iblech: also, i will be updating the Perl6.MetaModel (JS) more this weekend, but it is not yet complete enough for integration
I still need to figure out how to do Perl6.Object 18:51
and Perl6.Class
basically, JS does not have AUTOLOAD
which is the crux of how I deal with method dispatching
iblech Hm... we could emit .dispatch("subname")() instead of .subname(), then .dispatch can deal with AUTOLOAD appropriately 18:53
stevan iblech: that would work
or I can try to pre-populate the prototype of the generated class
iblech Ah, basically AUTOLOAD at compile-time :) 18:54
stevan yes 18:55
bloated, but it would work
however, It would make classes closed
unless I wrapped .meta.add_method to add into the class as well
but then I would have to handle all classes
iblech Hm... couldn't you extend the prototype at runtime, if necessary?
stevan yuk 18:56
iblech: yes, but say I add to Perl6.Object, I need to mess with every prototype in the system
because I am not using the prototpyes for inheritance
I am doing it on my own
iblech Ah..., ok
stevan oh, wait
Object.watch might do what we need 18:57
I am investigating right now
iblech Object.watch?
Ah! 18:58
coral Limbic: pugs at work. skills development. 19:01
Limbic_Region coral - that's an extreme stretch given that I am not a programmer by trade 19:02
geoffb Limbic_Region, what is your trade then?
Limbic_Region geoffb - well, since working as a government employee I don't really do much of anything 19:03
geoffb *chuckle*
Besides that, silly man
Limbic_Region but my previous job titles included "Systems Engineer"
geoffb I can definitively say (having hired several "system engineers") that programming is Most Definitely skills development 19:04
Limbic_Region basically - I design solutions to problems which may involve hw, sw, networks, etc
right - but as I said, I am not a systems engineer anymore
geoffb It stretches the brain in a way that makes for much better solutions. 19:05
so what does the government think you do, how about that phrasing?
stevan iblech: I don't think Object.watch will do what we need
Limbic_Region geoffb - I review documentation, sit on an engineering review board to render decisions - um - hard to say 19:06
stevan Limbic_Region: that sounds like *gasp* management??
Limbic_Region our IT staff consists of 2 people - myself and a network goon - everything else is contracted out 19:07
geoffb yikes
Limbic_Region stevan - nope, not people mgmt nor project mgmt - just decisions on if a particular approach meets the contractual requirements
stevan Limbic_Region: suuuuurrreeee :) 19:08
Limbic_Region does get to code some perl sometimes when data needs to get converted from 1 format to another in a hurry
go figure
stevan iblech: I have to do $work now, but I will be doing some MetaModel stuff more tonight (approx 5-6 hours from now) 19:09
geoffb apt-get install hugs # The adventure begins . . . 19:10
<newbie_mode>Does Haskell have a way to combine multiple lines into one? (as with ; in Perl) 19:21
gaal ; ? you mean \ ? 19:22
geoffb, one function application?
geoffb No, I meant I want to be able to do a pile of definitions on one line, like the custom when feeding evalbot6. These tutorials all show prettyprinted files, and I'm a one-liner guy 19:23
gaal ahhh, you want non-layout
{ a ; b ; c }
you need the braces too
geoffb ok, /me tries that . . .
Hmm, hugs doesn't seem to like that 19:25
bah, ok, food has arrived
&
kolibrie hmm, larry is rewriting method and attribute syntax 19:50
geoffb o_O 19:51
Link, kolibrie ?
iblech stevan: re. Great :) 19:52
geoffb: www.nntp.perl.org/group/perl.perl6....uage/22404
kolibrie iblech: thanks for providing the link 19:57
svnbot6 r5735, fglock++ | * Added a "Set::Symbols" role that implements unicode
r5735, fglock++ | operators in Span.pm and Recurrence.pm (it could be
r5735, fglock++ | used in Set.pm too)
kolibrie geoffb: sorry, I was on the phone
geoffb kolibrie, oh, no prob. 19:58
Interesting -- so now convention becomes authoritative
er, wrong word there 19:59
"becomes binding"?
svnbot6 r5736, fglock++ | * svn properties 20:18
Khisanth hmm why not just method foo is private ... ? 21:31
geoffb
.oO( He wants two uses for the same thing (_ == privacy) so that people don't complain about it being a singleton concept in the language? )
 21:33
21:35 Maddingue__ is now known as Maddingue
iblech Night all :) 21:38
svnbot6 r5737, iblech++ | PIL2JS:
r5737, iblech++ | * Added some new primitives and fixed some bugs.
r5737, iblech++ | * Refactored &run_pugs, run_pil2js, etc. into separate module lib/PIL2JS.pm.
r5737, iblech++ | * runjs.pl automatically precompiles the Prelude now, if needed.
r5737, iblech++ | * Test.pm is now precompiled, too. Now tests using Test.pm compile and run
r5737, iblech++ | with a satisfactoring speed -- if s/use Test//:
r5737, iblech++ | "use Test" still loads and compiles Test.pm -- bad. BEGIN {
r5737, iblech++ | %*INC{"Test.pm"}++ } would work, but: pugs -E "BEGIN {...}" t/.../foo.t does
r5737, iblech++ | (of course) not work...
ocarina greetings 22:18
autrijus, i'm reading the interview chromatic made you at perl.com... pretty cool :) 22:19
22:19 ocarina is now known as Excalibor
geoffb OK, end of _Haskell Tutorial for C Programmers_, time for a break before _Tour of the Haskell Syntax_ 22:29
Excalibor geoffb, :) 22:34
Excalibor still reading _Functional Programmin with Haskell_ :-P 22:37
geoffb I'm trying to keep my partially-read book stack small, but I'm pretending that online tutorials do not constitute pushing onto that stack 22:40
Excalibor geoffb, yeah, keep lieing to yourself... ;) 22:43
geoffb yep, it's completely OK that I'm several levels deep in Haskell and Forth tutorials because of harrorth. Damn nothingmuch++. :-) 22:44
Khisanth you can't damn someone while increasing their karma! 22:47
geoffb Certainly you can. "Damn you for being right" makes total sense to me. Besides, damnation and karma aren't on the same vector, though they have non-zero dot product. 22:49
Oh dear, I'm losing my mind.
Khisanth since we are damning things I would like to damn wikipedia :) 22:51
geoffb For managing to replicate the whole surfing-off-into-the-sunset problem of the web as a whole, all in just one site?
Excalibor geoffb, haha only serious? 22:53
geoffb nodnod 22:54
Excalibor well... time for bed to me
geoffb g'night 22:55
Excalibor take care... thx
gotta keep trying to understand those monads in pugs, though... some IO module in mind, no way to do w/o them though, lol 22:56
i'll keep studying :)
laters!
szabgab ?eval "gabor" ~ " szabo" 23:00
evalbot6 'gabor szabo'
szabgab (gaal here showing off safe mode)
?eval open "/etc/passwd" 23:01
evalbot6 Error: No compatible subroutine found: "&Pugs::Internals::openFile"
Khisanth geoffb: neither 23:02
geoffb neither?
geoffb can't remember having asked a disjunction
Khisanth oops "no" :)
geoffb So why *are* you damning wikipedia? 23:03
Khisanth because I read something on it before and now I can't recall the name!
geoffb heh
?eval use perl5:Cwd; cwd 23:04
evalbot6 pugs: *** No compatible subroutine found: "&require_perl5" at src/perl6/Prelude.pm line 110, column 30-59
Khisanth ah en.wikipedia.org/wiki/Programming_paradigm should lead me to the right place :)
geoffb Cute: [10,9..1] in Haskell 23:10
wolverian ?eval 3..1 23:12
evalbot6 ()
wolverian :(
Khisanth geoffb: why is that cute? 23:17
[10,8..1] is even nicer :) 23:18
hmm 23:19
?eval 3..1:by(-1)
evalbot6 Error: unexpected ":" expecting digit, fraction, exponent, term postfix, operator, "does", "but", "is", "^..^", "cmp", "<=>", "^..", "..^", "..", postfix conditional, postfix loop, postfix iteration, ";" or end of input
geoffb OK, so I hadn't come up with the most powerful example. I just meant the idea of determining the step by showing two in series, rather than some "by" syntax, is a cool idea
Khisanth bah, do both for the sake of TIMTOWTDI! :) 23:20
geoffb well, yes, there is that
Khisanth I see nothing particularly wrong with :by though 23:21
geoffb OK, this is beginning to bug me, and the tutorials so far have not cleared this up. Is there either A) a base type that unifies with every other type, or B) a data structure in Haskell that is variable-length like a list and contains differently-typed contents, like a tuple?
Khisanth [10..] :)
heh a list of tuples? :P 23:22
geoffb Hmmm, doesn't seem to work. 23:23
"Cannot infer instance"
So I guess that kills A as a likely possibility, or it would have Just Worked
Khisanth heh I think I had/have the same problem 23:24
geoffb I was appalled to see the definition of Eq on tuples, with only tuples of length 1 to 10 or so defined
For a language that prides itself on flexible typing, that's just odd 23:25
Khisanth isn't zip only define for up to 3 lists? :) 23:29
geoffb That's just wrong.
"Language designers should be forced to pass language ideas past the average 5 year old . . . " 23:30
Khisanth geoffb: although I seem to recall autrijus writing a oneliner that does the same as zip but works for an arbitrary number of lists :) 23:31
nothingmuch seen ingy
jabbot nothingmuch: ingy was seen 2 days 6 hours 31 minutes 5 seconds ago
nothingmuch blast
thunder
geoffb Khisanth, yeah, for a list of lists . . . but what about an arbitary length tuple of lists? 23:32
ingy nothingmuch: whuh? 23:36
geoffb _Tour of the Haskell Syntax_ down . . . next! 23:47

Powered by Cro and the Raku® Programming Language.

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

Thank you!