pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, elf: etc.) (or perl6: for all) | irclog: irc.pugscode.org/
Set by Tene on 29 July 2008.
meppl good night 00:00
ruoso work & 10:08
Juerd pmurias: Same things you can do anywhere else 11:17
pmurias is coding in the hotel lobby ;) 11:20
Juerd I'm going to explore the parking lot behind the venue again 11:21
To see if I can legally park there
Someone said yes, but yesterday some other cph inhabitant said no.
pmurias Juerd: what are you doing after that? 11:22
Juerd After that I will try to find an ATM
I'm out of kroner
pmurias can one pay in euro here? 11:24
Juerd No 11:24
I have sufficient euro cash with me
But the silly danes never joined the euro zone :) 11:25
My tomtom has ATMs for Sweden, not for Denmark 11:29
So if I ask it to navigate to one, it'll send me to Malmö... :)
pmurias do you know where other people are? 11:36
pugs_svn r21894 | pmurias++ | [elfX] start of a m0ld backend
pmurias looking around the city& 11:42
ron anyone around for questions ? 13:38
[particle] yes 13:40
Tene Yes! 13:43
ron perl6: class Foo { method fooey {say 'fooey'} }; class goo { method gooey {say 'gooey'} }; sub goo() { return Foo }; my $goo = goo.new; $goo.fooey; $goo.gooey; 13:44
p6eval rakudo 29834: OUTPUT[fooey␤Method 'gooey' not found for invocant of class 'Foo'␤current instr.: '_block11' pc 72 (EVAL_13:25)␤]
..elf 21894: OUTPUT[Can't locate object method "gooey" via package "Foo" at (eval 127) line 16.␤ at ./elf_f line 3861␤fooey␤]
..pugs: OUTPUT[fooey␤*** No such method in class Foo: "&gooey"␤ at /tmp/Ecdv4HcSwa line 1, column 142-152␤]
ron perl6: class Foo { method fooey {say 'fooey'} }; class goo { method gooey {say 'gooey'} }; sub goo() { return Foo }; my $goo = goo.new; $goo.fooey;
p6eval elf 21894, pugs, rakudo 29834: OUTPUT[fooey␤]
ron have done similar tests under perl5 and believe perl5 resolves at compile time. 13:45
have read that perl6 does not use bar words but class tests seem to use them.
should I start to document some of this stuff somewhere?
bar -> bare 13:46
perl6: class Foo { method fooey {say 'fooey'} }; class goo { method gooey {say 'gooey'} }; my $goo = goo.new; $goo.gooey; 13:47
p6eval elf 21894: OUTPUT[Undefined subroutine &GLOBAL::goo called at (eval 125) line 13.␤ at ./elf_f line 3861␤]
..rakudo 29834: OUTPUT[invoke() not implemented in class 'goo'␤current instr.: '_block11' pc 20 (EVAL_12:13)␤]
..pugs: OUTPUT[gooey␤]
ruoso I think pugs is right in both cases... 13:48
bare "goo" geos through standard symbol lookup 13:50
and it finds the closest definition
which is sub goo
ron From what I could tell with --target=past|post|pir in last case rakudo still looking for sub and not sure whether lookup is done. 13:51
ruoso in that case, rakudo is presuming class must have a uppercase first letter... which is wrong iirc 13:52
ron token typename { <?before <.upper> > <name> } 13:52
Lorn hi, a answer about perl6 in perlmonks: www.perlmonks.org/?node_id=703877 13:54
ron Thanks, think maybe I should still document someplace if I can't fix. 13:55
ruoso pugs: { sub foo { say 'foo1' }; { sub foo { say 'foo2'}; foo(); } foo(); }; 13:58
p6eval pugs: OUTPUT[*** ␤ Unexpected "foo"␤ expecting operator or "}"␤ at /tmp/CjGkoE9n0I line 1, column 61␤]
ruoso pugs: { my sub foo { say 'foo1'; }; { my sub foo { say 'foo2'}; foo(); }; foo(); }; 13:59
p6eval pugs: OUTPUT[foo2␤foo1␤]
ruoso pugs: { sub foo { say 'foo1'; }; { my sub foo { say 'foo2'}; foo(); }; foo(); };
p6eval pugs: OUTPUT[foo2␤foo1␤]
ruoso pugs: { sub foo { say 'foo1'; }; { sub foo { say 'foo2'}; foo(); }; foo(); };
p6eval pugs: OUTPUT[foo2␤foo1␤]
ruoso rakudo: { sub foo { say 'foo1'; }; { sub foo { say 'foo2'}; foo(); }; foo(); }; 14:04
p6eval rakudo 29834: OUTPUT[./parrot: error while loading shared libraries: /home/evalenv/parrot/blib/lib/libparrot.so.0.6.4: file too short␤]
ruoso elf: { sub foo { say 'foo1'; }; { sub foo { say 'foo2'}; foo(); }; foo(); };
p6eval elf 21894: OUTPUT[Subroutine foo redefined at (eval 127) line 4.␤foo2␤foo2␤]
ruoso one could argue that unless you have 'my sub' it should store it in the package...
(which doesn't solve the goo.gooey problem) 14:05
(but is just a note about pugs implementation)
elf: { my sub foo { say 'foo1'; }; { my sub foo { say 'foo2'}; foo(); }; foo(); }; 14:06
p6eval elf 21894: OUTPUT[Subroutine foo redefined at (eval 127) line 4.␤foo2␤foo2␤]
ron pugs: { my sub foo { say 'foo1'; }; } ; foo(); 14:07
p6eval pugs: OUTPUT[*** No such subroutine: "&foo"␤ at /tmp/zu7ZdK0StI line 1, column 36-41␤]
ron pugs: { sub foo { say 'foo1'; }; } ; foo(); 14:08
p6eval pugs: OUTPUT[foo1␤]
ruoso pugs: { sub foo { say 'foo1'; }; { sub foo { say 'foo2'}; foo(); }; foo(); }; foo() 14:10
p6eval pugs: OUTPUT[foo2␤foo1␤foo2␤]
ruoso pugs: { sub foo { say 'foo1'; }; { my sub foo { say 'foo2'}; foo(); }; foo(); }; foo()
p6eval pugs: OUTPUT[foo2␤foo1␤foo1␤]
ruoso I think a warning would be ok when not using 'my sub'
ruoso or even when declaring a 'my sub' that overrides a package sub... 14:12
or a pre-declared class... 14:13
and also when declaring a sub that overrides a class declaration... 14:15
since subs and classes can both be used without sigil (since that classes doesn't have a sigil) 14:16
ron pugs: sub foo { print "foo1\n"; }; { sub foo { print "foo2\n"; }; foo(); }; foo();
p6eval pugs: OUTPUT[foo2␤foo1␤]
ron in perl5 with warnings there is a sub redefined warning 14:17
does the evalbot print warnings?
[particle] most of the perl 6 implementations don't do warnings yet
ron rakudo: sub foo { print "foo1\n"; }; { sub foo { print "foo2\n"; }; foo(); }; foo(); 14:17
p6eval rakudo 29834: OUTPUT[foo2␤foo2␤]
ruoso that might be right, as bare 'sub' is not the same as 'my sub' 14:18
[particle] perl6: say $*ERR: "howdy"
p6eval rakudo 29834: OUTPUT[Statement not terminated properly at line 1, near ": \"howdy\""␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤]
..elf 21894: OUTPUT[Undefined subroutine &GLOBAL::infix_58_58 called at (eval 119) line 3.␤ at ./elf_f line 3861␤]
..pugs: OUTPUT[howdy␤]
pmurias ruoso: why does every SMOP__Object need to have a data?
ruoso pmurias, it was to allow responder interfaces to be treated as regular objects.... 14:19
but I'm not sure anymore that this is actually required
considering that ->data is only accessed by the object itself 14:20
we could move the data pointer to a SMOP_LOWLEVEL_OBJECT_BASE struct
since smop_lowlevel uses that pointer to hold lock and refcnt informatoin
but that is a quite extensive refactoring... 14:21
pmurias loads of search and replace ;)
ron rakudo: say $*ERR: "HOWDY";
p6eval rakudo 29834: OUTPUT[Statement not terminated properly at line 1, near ": \"HOWDY\";"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤]
ron perl6: say $*ERR: "HOWDY"; 14:22
p6eval rakudo 29834: OUTPUT[Statement not terminated properly at line 1, near ": \"HOWDY\";"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤]
..pugs: OUTPUT[HOWDY␤]
..elf 21894: OUTPUT[Undefined subroutine &GLOBAL::infix_58_58 called at (eval 119) line 3.␤ at ./elf_f line 3861␤]
ruoso pmurias, I'm neglecting that question because we're talking about just one pointer, and mostly because most objects so far are using it... 14:23
(all objects that use smop_lowlevel do)
ruoso but we might get into this issue before "1.0" ;) 14:24
pmurias, btw... about code object... I forgot one attribute when we were talking about it... 14:25
class Code { has $.outer; has $.mold; has $.signature; method postcircumfix:<()> {... } } 14:26
and I'm thinking about having signature like:
class Signature { method ACCEPTS (Capture $c) {...}; method BIND (Capture $c, LexicalScope $l) {...} }; 14:27
ruoso this way we can have a lowlevel implementation of the default block signature... 14:27
we can even compile lowlevel implementations for the Signatures we use in s1p... 14:28
pmurias ruoso: i think we should try to have as little stuff in C as possible 14:29
ruoso we can have a signature compiler 14:30
and not write it in C
instead of instantiating a Signature object, we create a new type specifically for that block :)
or re-use for identical signatures... 14:31
(specially the default block signature
pmurias the JIT can wait a bit
ruoso no... it's not just in time... it's compilation to static C code... 14:32
s1p doesn't have a real interpreter yet...
and won't have...
it will be compiled to static C code...
pmurias ruoso: what do you mean by a real interpreter one which compiles and runs lines one by one 14:34
?
ruoso I mean... ( $_ is rw = OUTER::<$_> ) would be compiled to a smop responder interface that implements ACCEPTS and BIND for that specific signature...
pmurias re real interpreter i reread and got it 14:35
i'm actually thinking of taking the .pbc spec and adapting it to smop
but it's just an optimalisation so it can wait 14:36
ruoso I think so
ron so I'm guessing that there may need to be some kind of compile time symbol table lookup for some of these bareword cases ... If anyone actually knows I'm curious ... 14:37
ruoso ron, after bootstrapping, yes... but s1p is a 1st stage compiler that won't have this kind of lookup...
pmurias ruoso: do you know how to compile a C file using smop withought putting it into the Makefile? 14:43
ruoso pmurias, what's the problem of putting it into Makefile 14:44
pmurias i'm testing switching p6 code from slime to mold 14:46
pmurias ruoso: and once we get something running on smop it would be awkward to require putting the users script into our build system 14:53
pugs_svn r21895 | pmurias++ | [smop] .p6 tests use mold 15:20
pmurias ruoso: how should we migrate use v6 blocks to mold, a use v6-mold syntax or take a big leap? 15:22
ruoso take a big leap... 15:33
maybe creating an additional placeholder for the molds initialization... 15:34

Powered by Cro and the Raku® Programming Language.

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

Thank you!