Pugs 6.2.9 released | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 4 August 2005.
svnbot6 r6262 | fglock++ | * added perl5/Hash Container; tests 00:42
r6263 | putter++ | Prelude.pm: Added support for regex :overlap modifier. But it's disabled until Rul infix:<~~> works in Prelude. 01:10
putter Supaplex: parrot is immature. it does not yet _run_ any interesting language. let alone run so well that the issue of releases would arise. tcl (in the parrot distribution), p6 (pugs), p5 (ponie, somewhere), and perhaps python (in parrot) are perhaps the most serious efforts. With p6/p5 having a community intent to run on parrot. My guess is it will end up being easier for languages to target parrot via p6, rather than directly. 01:11
Supaplex ok 01:14
putter autrijus, everyone: a worrisome observation - I'm seeing localized test instability (sometimes succeeds, sometimes fails) in some rules tests. There is a pugsbug illustrating pod contents determining success/failure, but this is unchanging test files. Perhaps the new use of macros is tickling the earlier problem. Or perhaps this is something new. 01:16
There is always the possibility that significant rules work should be deferred a bit longer, until the new runtime is in. 01:17
Good night all. & 01:30
Supaplex: thanks for your interest.
autrijus rehi 05:21
obra reautrijus 05:23
QtPlatypus Hi autrijus
Khisanth A! 05:25
autrijus :)
autrijus @ work 05:26
vkon++ # pugscode.org/images/lambdacamels.png revised to have the greeks resemble "Perl" 11:48
dinner & 11:51
Aankh|Clone autrijus: Good job on the XML backend. :-P 12:12
12:13 Aankh|Clone is now known as Aankhen``
nothingmuch autrijus: is the lambdacamels picture OK with oreilly? 12:22
nothingmuch just whored massive amounts of office supplies 12:30
fglock ping 12:45
nothingmuch: ping 12:51
nothingmuch pong 12:58
fglock: pong 12:59
fglock nothingmuch: I'm working on the runtime, but I need objects to define Pair and Hash - is it possible to use the metamodel for this? 13:00
nothingmuch fglock: you want to implement Pair and Hash within the metamodel?
that's very possible
probably easy, too
fglock I want to "use" the metamodel to create Int, Str, Pair, ...
nothingmuch s/create/implement/ or s/create/instantiate/ ? 13:01
fglock I have to "box" things before inserting in the Container
nothingmuch fglock: nopaste some pseudocode for how you want the interface to look, i'm not sure I understand what advice I can give yet 13:02
fglock I've got native "int" and "str", and I want to create a Pair ('a' => 3) - I need "Str" and "Int" objects to do this 13:04
I need Int->new( 3 ) and Str->new( 'a' ) 13:05
nothingmuch right
so there is a class Int which has some convenient methods
and 'has int $:value;' 13:06
right?
and you want to know how this might be implemented in the metamodel?
or you want to know if it's already been done?
or...?
fglock Yes, I want to implement this class, probably inside the Perl6::Value modules 13:07
nothingmuch ah
so how would you like me to help?
fglock but it should use the object model from the metamodel, not Perl5's
nothingmuch okay
in the minimetamdeol this is very very easy, and I'm now well versed in it
in the real perl 6 metamodel it's probably more sugared 13:08
but i need to reread it, as I'm out of date
fglock I just need a small example on how to start - It can be modified and refined later
nothingmuch basically the process is: you instantiate the class "Class", and get a new class
you fill it up with methods and attributes and stuff
this new class is a subclass of the class "Object" 13:09
then you can instantiate this new class, and get an object
fglock I'll do some tests
nothingmuch in the minimetamodel this would look a bit like my $PairClass = $Class->new('$:name' => 'Pair', '%:methods' => { first => sub { ... }, second => sub { ... } });
uh, '$:superclass' => $object; 13:10
then you can say 'my $this_is_a_pair_instance = $PairClass->new(...)';
you edit the attributes using ::opaque_instance_attrs, (maybe it has a diff name, I can't recall)
want me to walk you through what I wrote yesterday? a good enough intro is the autoloading metaclass 13:11
svn.openfoundry.org/pugs/perl5/Perl...loading.pl
$Class->new creates a class.. This new class is a subclass of 'class'. From this class you create new classes, so it is, in fact, a metaclass too 13:12
it has a name, and some methods
it alters the way method dispatching is performed
then we create $MyAuto, a class that has some_method, and AUTOLOADER
it is a subclass of $Object
and it's metaclass is $Autoloading
instead of being a simple class, $MyAuto is an autoloading one 13:13
thne we create $iAuto, an instance of $MyAuto
and call methods on it
is this becoming clear? 13:14
fglock cool - thanks! I think I can start coding an example
nothingmuch fglock: if something doesn't work nopaste it 13:15
it's slightly confusing at first
but the moment you "get" it things start being easy 13:16
C3 was by var the hardest example
fglock it would be nice to have a common API for the metamodels (it may be too early for that)
nothingmuch fglock: look at what stevan comitted yesterday 13:17
it should be similar to the minimetamodel because it's based on it
it probably has nice sugar to let you create classes in a more declarative way
13:45 Aankh|Clone is now known as Aankhen``
fglock MiniMetamodel doesn't run - "undefined value as a HASH reference", line 33 13:47
autrijus rehi. 13:53
www.cafepress.com/pugscode
fglock hi autrijus
I'm working on the Str, Int .. boxed types 13:55
autrijus cool! 13:57
stevan fglock: you should likely use the Metamodel 1.0
morning all :)
fglock: 2.0 is in very early stages,..
fglock stevan: yes, it passes all tests here
autrijus hey stevan :)
stevan nice t-shirts autrijus :) 13:58
autrijus :D
Limbic_Region autrijus - did you see the question from nothingmuch about O'Reilly being ok with one of the images (presumably it had a camel on it) 13:59
<nothingmuch> autrijus: is the lambdacamels picture OK with oreilly?
mjl69 t-shirt? where? I need more. I wear my black and green perl shirts way too much.
stevan Limbic_Region: that version is not on cafepress
Limbic_Region stevan - I didn't say it was
I was repeating a question to autrijus now that he is here
stevan Pugs t-shirts www.cafepress.com/pugscode | Pugs 6.2.9 released | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net 14:00
Limbic_Region on nothingmuch's behalf 14:00
autrijus Limbic_Region: I asked permission to use that camel when I worked on par.perl.org a while ago
stevan Limbic_Region: ahh, sorry :)
autrijus Limbic_Region: and I don't plan to sell that :)
so I presume it's cool
mjl69 t-shirt link? Need more shirts...
stevan mjl69: see the title 14:01
autrijus mjl69: www.cafepress.com/pugscode
mjl69 oh, dumb me again.
autrijus mugwump: you okay if I put the "powered by phd" there too?
stevan nice 14:02
mjl69 I want to find one of those t-shirts that shows the geometric proof of the pythagorean theorem. I heard they are out there. 14:04
then a t-shirt to match all of the language books on my shelf I plan to study some day. 14:05
I do always work the first chapter at least on the day they come in. Then on the shelf they go.
autrijus mjl69: you mean this? www.cut-the-knot.org/pythagoras/abc.gif 14:06
stevan: I'm uploading the metamodel.png :) 14:08
Limbic_Region mjl69 - geometric proof of the pythagorean theorem - which one, I know of several 14:12
in fact, there are an infinite number of them www.jimloy.com/geometry/pythag.htm 14:14
autrijus stevan: metamodel mouse pad + sticker is on :) 14:19
castaway_ cool! 14:21
autrijus reboots and goes back to coding :) 14:29
cafepress is fun though :)
fglock I've got "Int" working! 14:31
autrijus fglock: woot!
svnbot6 r6264 | iblech++ | * Usual svn props. 14:32
r6264 | iblech++ | * PIL2JS: s/-CPIL/-CPerl5/, i.e. PIL2JS can "parse" PIL again. Next step is to
r6264 | iblech++ | (e.g.) s/->[0]/->{pExpr}/.
autrijus iblech: if you'd like to fix it to use ->{expr}, that could work too
iblech: what do you think? or pExpr is fine?
iblech autrijus: pExpr is absolutely fine
autrijus 'k 14:33
fglock where can I find object hierarchy docs? 14:36
iblech TSa posted his type lattice to p6l
autrijus fglock: there... is none.
fglock :)
what's the difference between Bit and Bool? 14:37
svnbot6 r6265 | fglock++ | * perl5/ "Int" boxed value, using the MetaModel
autrijus fglock: bool is an enum subtype on Bit
I think
there's a synopsis that talks about it 14:38
mjl69 autrijus: That's the one! Just need to find the t-shirt. 14:41
thanks also Limbic_Regi (delay due to pesky day job, sorry!) 14:42
svnbot6 r6266 | fglock++ | * perl5/ Value - boxed type conversion "$i = $n->int()" works 14:50
14:51 autark is now known as jp-autark
autrijus mjl69: ping 15:03
mjl69 autrijus: hi!
mjl69 goes back to work again 15:05
autrijus mjl69: is this what you want?
cafepress.com/autrijus/
mjl69: I _think_ that's the layout of the original Pythagoras proof 15:06
not sure though
fglock autrijus, stevan: take a look at Perl6-Value-List/t/item-boxed.t
mjl69 wow, you can just put a shirt together just like that?
autrijus mjl69: yeah :)
took me some time to draw that Visio thing though 15:07
mjl69 autrijus: I can really order that shirt?
autrijus mjl69: sure!
svnbot6 r6267 | iblech++ | DrIFT.Perl5: Escape \$@%" in double-quoted strings.
r6268 | fglock++ | * per5/ Value - boxed types mostly done
mjl69 autrijus: you are the ultimate t-shirt man!
autrijus :D
fglock can you assign to a Pair's value - "$x.v = 10"? (I think not, because Pair are not containers, right?) 15:23
autrijus fglock: Pair, like Ref, can point to a container
svnbot6 r6269 | fglock++ | * perl5/ Value - added Num->Inf; Num->NaN 15:24
autrijus and I think it was ruled that $x.value is rw.
which is to say, it returns a lvalue.
which is to say, it can be assigned to
fglock: but I think the .key is a singular value
so you can't assign to .key
fglock Pugs doesn't permit assignment to .value 15:26
autrijus the current runcore doesn't do that, yup.
that's a known bug
fglock ok 15:27
you don't modify an Int - you modify the container, right? 15:29
autrijus right.
see the container.png
fglock isn't Pair a container, then? 15:30
autrijus I tend to think of it like
class Pair { has $.key; has $.val is rw } 15:31
so it's not itself a container
but it does refer to a container
fglock must go for lunch now & 15:33
svnbot6 r6270 | iblech++ | * util/drift.pl: Made it work here. 15:41
r6270 | iblech++ | * PIL2JS: Modified ~~50% of the *.pms to work with the new -CPerl5.
coral -CPerl5? 15:49
dudley coral: compiles to perl5 data structures 15:50
coral interesting 15:51
dudley indeed
coral SV or PMC p5?
that is, perl5 or ponie
autrijus coral: eval p5
coral ok
autrijus as in , "bless(...)"
ods15 i never did get bless(), i thought it was some kind of bad perl joke 15:52
integral if you've read SICP, <SICP> =~ s/tag/bless/ 15:54
iblech autrijus: What's PPos' pExp? (It seems to be always undef...) 16:04
autrijus iblech: it's because Exp doesn't derive Perl5 yet 16:11
iblech: it's the original syntax tree
iblech autrijus: ah, ok
autrijus I presume you don't actually use it
iblech yep
autrijus so all's well then
svnbot6 r6271 | iblech++ | PIL2JS: Ported PStmt, PPos and PStmts to -CPerl5, too. Only PVal, PVar, and 16:25
r6271 | iblech++ | Subs need work now (and bugfixing, of course).
xinming ?eval { 3+5 } 17:40
evalbot6 8
svnbot6 r6272 | fglock++ | * perl5/ Value - "Pair" type
xinming ?eval { my $x = 3; "a" => 1, $b, $c ==> print };
evalbot6 Error: unexpected ">" expecting term
xinming ?eval { my $x = 3; "a" => 1 };
evalbot6 ('a' => 1) 17:41
xinming ?eval { my $x = 3; "a" => 1; return $x ;};
evalbot6 \3
xinming ?eval { my $x = 3; "a" => 1; return bless $x,"haha"; };
evalbot6 Error: No compatible subroutine found: "&bless"
xinming ?eval { my $x = 3; "a" => 1; }; 17:42
evalbot6 ('a' => 1)
xinming ?eval { my $x = 3; "a" => 1; return $x ;}
evalbot6 \3
xinming anyone here can tell me what that expression exactly returned? a closure or a ref to a scalar ($x) or a hash ref which contains {'a' => 1}? 17:44
autrijus xinming: it returns a scalar $x. the ref thing merely indicates it's lvalue
fglock 'undef' is just an empty Scalar? or is it an object itself? 17:45
xinming autrijus: may I refer the value of the key 'a'? 17:46
Or,If there is a way to get the value of it. 17:47
ods15 geez
that crap looks like perl
autrijus fglock: undef is an empty anything. :) 17:48
fglock: think of it as C's null. 17:49
fglock ok
autrijus ?eval ("a" => 1).key
evalbot6 'a'
autrijus xinming: like that?
xinming autrijus: hmm, that's what I want to know 17:50
if my $t = { my $x = 3; "a" => 1; return $x ;}; 17:51
fglock does .ref returns the class object, or is it always just the class name (string)?
xinming then $t<a> will get the value of a?
iblech fglock: class object, never string 17:52
xinming: No, $t will be a Code ref
But if you wrote my $t = ("a" => 1) then $t<a> would give you 1 17:53
xinming iblech: so the kv pair will be parsed but ignored. right?
iblech xinming: Right.
autrijus fglock: the current implementation is broken, but it's supposed to return a class object.
iblech xinming: That's similar as "sub foo { 1; 2; 3 }" -- the 1 and 2 can safely be optimized away 17:54
fglock should be writing this in Perl6 instead
autrijus they can be optimized away only if they are simple values :) 17:55
fortunately, "overloading the void context" is not yet heard of.
iblech Yep :)
xinming autrijus: Could you please tell me when 6.283 might be released? :-) 17:56
autrijus xinming: good question
xinming: it'll be released when we have proper support for Rules and Grammers, in addition to the object model. 17:57
I'm afraid that's the best answer i have :)
PerlJam autrijus: On p6l, are you asking if roles become classes for instantiation purposes?
autrijus PerlJam: yes.
PerlJam: not the previous "anonymous class gets vivified" foo 17:58
but just "roles are liskov-equivalent to classes"
"roles are just classes that additionally accepts 'does'"
PerlJam interesting and strange. 17:59
autrijus larry's been thinking interesting and strange thoughts. 18:00
xinming I can' think as high as Larry. So, I really don't know what the "roles" do really. just know, It is, really a "class", 18:05
but can't create object. 18:07
mjl69 roles?
autrijus xinming: it can create objects if all its methods are defined 18:09
in that case it's still the same as a class
really, roles are classes with two reuse forms: "is RoleName" and "does RoleName"
while classes only supports "is ClassName"
it seems that's all the difference there is... oh, and roles can be generic (takes type parameters) too. 18:10
putter re pythagoras, www.headmap.org/unlearn/euclid/book1/1.47.htm www.sunsite.ubc.ca/DigitalMathArchi...byrne.html 18:13
xinming hmm... If so, I think, Why "roles" exist might be for joining two (or more) "roles" into a class. 18:14
autrijus putter: have you seen my pythagoras t-shirt at www.cafepress.com/autrijus/ ?
xinming: yes.
mjl69 can create objects if all methods defined...like a C++ class with pure virtual methods? 18:15
putter autrijus: yes, just did. neat!
www.sunsite.ubc.ca/DigitalMathArchi...kI-47.html too
xinming autrijus: May I see your photo? :-) I've went your site. most photos aren't clear I think. 18:16
putter is tempted to create a "How Big Are Things?" T-Shirt www.vendian.org/howbig/ www.vendian.org/howbig/cube/
mjl69 autrijus: I like your t-shirt compared to the other pythagoras shirts because it doesn't say what it is. That makes it more cool. 18:17
fglock re:metamodel - I can't find the syntax to access an instance attribute - "'undefine' => { _('$.value' => undef) }" doesn't work 18:18
stevan fglock: sub _ is probably ignoring the undef 18:22
I think something like if defined $value
to set that
fglock I get "cannot call $?SELF from outside of a MetaModel defined Instance method" 18:23
stevan hmm
can you nopaste the entire code?
'undefine' => { _('$.value' => undef) } should be 'undefine' => sub { _('$.value' => undef) } 18:24
maybe that is it
I made that mistake several times myself
Perl 6 tends to stick with your
s/r$//
fglock fixed! thanks (i forgot the "sub" :( 18:25
autrijus wouldn't it be nice to
package HASH; use overloas '&{}'
sadly that doesn't quite work.
stevan LOL
autrijus may be possible with autobox.pm though.
autrijus stops thinking bad ideas
stevan scolds autrijus for even going there :P 18:26
mjl69 is this the channel with the bot that evaluates statements or was it #haskell?
Limbic_Region ?eval say 'here'
evalbot6 here bool::true
mjl69 oooh, I thought it was here.
integral #haskell has @eval
mjl69 oh, it's both. 18:27
autrijus #haskell also has @plugs
putter re T-Shirts and objectsm an idea -- date them, eg, 2005. Or "Pugs 2005". "Pugs - The First Year". "Fall 2005". motivates sale closing because you wont be able to get the high-something "Summer 2005" (wow!) versions a month from now.
autrijus good idea.
mjl69 I suggested in #ruby-lang they should have one too since Ruby lends itself to that sort of learning/testing. I guess the fear is it opens the bot's system to attack. 18:28
buy the t-shirt without the date right away. Limited edition.
putter ;) 18:29
stevan autrijus: you could date *and* version them
get your 6.2.9 t-shirts now before 6.2.10 comes out
mjl69 convert the obfuscated camel code from the back of the perl 5 t-shirt to perl 6... 18:30
PerlJam If you date and version them, put a feature list on the back so that either it continues to grow or (if it always takes up the entire back) the font keeps getting smaller :-) 18:31
putter re obfuscated, good point. the largest target market is p5 people, so want to hook them...
stevan ChangeLog t-shirts!!!
putter lol :)
mjl69 like concert tour dates 18:32
PerlJam exactly
how does cafepress work exactly? Does it cost to update or put up a new entry or just when people buy?
If it only costs when they buy you can automatically update the shift from the repository ;-) 18:33
er, s/shift/shirt/
autrijus mm changelog t-shirts 18:45
very crazy
svnbot6 r6273 | iblech++ | PIL2JS: Remaining .pms ported (PVar, Subs, PVal). Now bugfixing. 18:46
autrijus ok. I'll date everything to expire at 1st november 18:47
18:47 spinclad___ is now known as spinclad
autrijus done. 18:55
putter neat. err, just to ensure no misunderstanding, I was thinking of dates on the objects, rather than the webpage... but webpage is nice too. 19:01
putter ponders what to work on... 19:03
nothingmuch: ping? 19:05
coral cafepress lets you put up one t-shirt and sell it for free 19:12
dunno about zazzle prices
s/one t-shirt/one entry for each type of item/
putter sigh. nobody is smoking p6 rules. :( 19:13
autrijus coral: s/for each type of item/$& for each store/ 19:19
coral: and because a person can create +Inf of stores, that's not a severe restirction
gah. past 3am already.
autrijus uploads a brief journal and goes to sleep 19:20
fglock how do I get the class, inside a metamodel object?
autrijus waves &
fglock good night autrijus
xinming ?eval { my $t = 3; 'a' => 1 }; 19:27
evalbot6 ('a' => 1)
xinming which value will return in this expression? 3 or 'a' or 1 ?
ooops.I mean the clous*
maybe this sub
?eval sub { my $t = 3; 'a' => 1 };
evalbot6 sub {...} 19:28
xinming ?eval my $x = sub { my $t = 3; 'a' => 1 }; $x();
evalbot6 ('a' => 1)
xinming anyone here has a clue? 19:29
?eval my $x = sub { my $t = 3; 'a' => 1 }; $x().key; 19:30
evalbot6 'a'
xinming hmm, Maybe It just returns a anonymous hash ref.
iblech xinming: re. It returns a pair, but pairs can .key, .value, .kv, .pairs, too 19:40
fglock re: metamodel - how do I get the class of an instance? 19:43
xinming ?eval my $x = sub { my $t = 3; 'a' => 1 }; my $w = { 'b' => 2, $x() }; $w.keys 19:44
evalbot6 ('a', 'b')
xinming ?eval my $x = sub { my $t = 3; {'a' => 1} }; my $w = { 'b' => 2, $x() }; $w.keys
evalbot6 ('a 1', 'b')
xinming ?eval my $x = sub { my $t = 3; {'a' => 1} }; my $w = { 'b' => 2, 'hashref' => $x() }; $w.values 19:45
evalbot6 (2, (('a' => 1)))
xinming iblech: thanks. 19:46
iblech fglock: Generally, .ref. dunny about Perl5-MetaModel 19:48
xinming: np
fglock I'm implementing .ref :)
iblech oh :) 19:50
putter autrijus: re "Day 190 (r6272): Pugs Store opens!", err, no. My suggestion was to have a date _on the shirts/mugs/etc _themselves__. Sorry I was unclear. 19:54
xinming ?eval class TMP { has $.tmp = 3; } 20:02
evalbot6 Error: unexpected "{" expecting trait, ";" or end of input
xinming ?eval class TMP { has $.tmp; }
evalbot6 undef
xinming ?eval class TMP { has $.tmp; }; my $t .= new TMP; 20:03
evalbot6 Error: unexpected ";" expecting letter or digit, ":", term postfix, operator or "," The .= operator needs a type specification.
xinming ?eval class TMP { has $.tmp; }; my TMP $t .= new; 20:04
evalbot6 \TMP.new();
xinming ?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp = 1; 20:05
evalbot6 \1
xinming ?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp = 1; $t.tmp.say
evalbot6 1 bool::true
xinming ?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp( 1 ); $t.tmp.say
evalbot6 Error: No compatible subroutine found: "&tmp"
xinming ?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp = 1; my $t = $t.tmp; $t.tmp.say 20:06
evalbot6 Error: No compatible subroutine found: "&tmp"
xinming ?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp = 1; my $y = $t.tmp; $y.say 20:07
evalbot6 1 bool::true
fglock stevan: ping 20:08
xinming ?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp = 1; my $y = $t.tmp(); $y.say
evalbot6 1 bool::true
xinming hmm... By the way, Why can't auto-generated accessor method pass a value as a parameter? 20:09
?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp = 1; $t.tmp(2); $t.tmp.say 20:10
evalbot6 Error: No compatible subroutine found: "&tmp"
iblech Because they're declared as method tmp () is rw {...} 20:16
I.e. they can only be used as lvalues
theorbtw1 Not quite -- because they take zero parameters, other then the object, and in $t.tmp(2), you're calling it with one. 20:18
xinming iblech: In fact, That's why I wonder why the default isn't method tmp( $x ) { if( $x ) { $.tmp = $x }; $x }
theorbtw1 The rvale isn't a parameter in an lvalue call, for better or for worse.
xinming I think This might be more often for a writable accessor. 20:19
theorbtw1 xinming, because if you want to assign something, you should do something that looks like an assignment.
putter later &
theorbtw1 I'm off to bed. 20:22
svnbot6 r6274 | fglock++ | * perl5/ Container - added "Scalar" type; tests.
r6274 | fglock++ | .ref() doesn't work yet
xinming theorbtw1: hmm... thanks, In fact, sometimes, I am thinking in perl 5 for my personal habits. :-)
theorbtw1: good night.
iblech I'm too.
Night all :)
xinming night
role X { is Y; }; class Y does X { }; What happens if we do something like this? 20:27
fglock xinming: there should be an error - circular inheritance or such 20:28
xinming hmm: so role X { is SOME_CLASS; } will turn SOME_CLASS into a role, right?
fglock (unless it is a syntax error) 20:29
xinming ?? 20:30
fglock: Could you tell me if I am thinking right.
fglock xinming - I don't know if it is allowed for a role to inherit from a class 20:31
xinming A role is allowed to declare an additional inheritance for its class when that is considered an implementation detail:
This is from Synoposis
role Pet { 20:32
is Friend;
}
fglock ah, ok - but I think the inheritance only applies to the methods in the role. 20:33
svnbot6 r6276 | iblech++ | PIL2JS: Stopped PIL2JS from dieing, but more work needed to make it work again (-> tomorrow).
xinming There are too many perl tricks I think useful in perl 6. :-) 20:36
s/think/like/
at least, we don't need to my $self = shift; for every method in a package anymore. 20:37
?eval class TMP { has $.tmp is rw; }; my TMP $t .= new; $t.tmp = 1; my $y = $t.tmp; $y.say 20:45
evalbot6 1 bool::true
xinming ?eval class TMP { has $:tmp is rw; }; my TMP $t .= new; $t.tmp = 1; my $y = $t.tmp; $y.say
evalbot6 Error: No compatible subroutine found: "&tmp"
xinming ?eval class TMP { has $.tmp; }; my TMP $t .= new; $t.tmp = 1; my $y = $t.tmp; $y.say 20:46
evalbot6 1 bool::true
xinming ?eval class TMP { has $.tmp;method tmp { "perl 6".say}; }; my TMP $t .= new; $t.tmp; 20:47
evalbot6 perl 6 bool::true
xinming it's said that all function call need the parenthesis. So is this wrong for pugs? 20:48
?eval class TMP { has $.tmp;method tmp { "perl 6".say}; }; my TMP $t .= new; tmp $t;
evalbot6 Error: Wrong number of invocant parameters: 0 actual, 1 expected in "&TMP::tmp"
xinming ?eval class TMP { has $.tmp;method tmp { "perl 6".say}; }; my TMP $t .= new; tmp $t();
evalbot6 Error: cannot cast from VObject (MkObject {objType = (mkType "TMP"), objAttrs = <ref>, objOpaque = Nothing, objId = 52}) to Pugs.AST.Internals.VCode (VCode)
xinming ?eval class TMP { has $.tmp;method tmp { "perl 6".say}; }; my TMP $t .= new; tmp( $t );
evalbot6 Error: Wrong number of invocant parameters: 0 actual, 1 expected in "&TMP::tmp"
xinming ?eval class TMP { has $.tmp;method tmp { "perl 6".say}; }; my TMP $t .= new; tmp $t:; 20:50
evalbot6 perl 6 bool::true
xinming ?eval class TMP { has $.tmp;method tmp { "perl 6".say}; }; my TMP $t .= new; tmp $t; 20:51
evalbot6 Error: Wrong number of invocant parameters: 0 actual, 1 expected in "&TMP::tmp"
xinming seen autrijus 20:52
jabbot xinming: autrijus was seen 1 hours 32 minutes 18 seconds ago
putter xinming: class TMP { has $.tmp;method tmp { "perl 6".say}; }; my TMP $t .= new; tmp($t:); 21:05
& 21:08
mugwump autrijus: yes, put the powered by PhD logo there! 21:59
it is at wellington.pm.org/archive/200506/pe...ide02.html ... the one on the next page is interesting, too 22:00
I think the reverse side of a powered by PhD top would have to have a * followed by a little bibliography 22:01
including paper names and references, perhaps we can brainstorm some here, stevan ? 22:02
mugwump reads the journal entry 22:04
svnbot6 r6277 | fglock++ | * perl5/ - implemented .ref() in Num,Int,Str,Bit,Pair,Scalar 22:09
r6278 | fglock++ | * perl5/ - new methods .perl, .increment, .decrement 23:20
r6278 | fglock++ | partially implemented

Powered by Cro and the Raku® Programming Language.

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

Thank you!