pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
TimToady note that even in P5, __DIE__ and __WARN__ are not really special tokens, just special keys to %SIG 00:00
also note that in P6, every warn is, in essence, already wrapped by all the handlers in the dynamic scope that might examine it
you only need to wrap warn() if you want to capture control before any of those handlers do. 00:01
but if you merely want to override the default warn handler, that's just the one in the outermost dynamic scope 00:02
so any dynamic scope inside that can override the outer handler
00:12 eternaleye joined 00:14 polettix left 00:19 mj41_ left, armagad left 00:36 alc joined 00:40 Guest39177 left 00:48 japhb left 01:05 cornelius_ joined 01:06 cornelius_ left, cornelius_ joined 01:07 cornelius_ is now known as c9s 01:09 japhb joined
Auzon pugs: my @a = 1 .. 10; my %h = @a; say %h.perl 01:12
exp_evalbot OUTPUT[{("1" => 2), ("3" => 4), ("5" => 6), ("7" => 8), ("9" => 10)}␤]
01:26 Ched- left, Ched- joined
Maghnus pugs: my @a = 1 .. 9; my %h = @a; say %h.perl 01:40
exp_evalbot OUTPUT[*** Odd number of elements found where hash expected: VList [VInt 1,VInt 2,VInt 3,VInt 4,VInt 5,VInt 6,VInt 7,VInt 8,VInt 9]␤ at /tmp/OQq3SCWurd line 1, column 17-27␤]
01:47 alanhaggai joined 02:18 sri_work left, sri_work joined 02:24 bacek_ joined 02:25 sri_work left
Auzon perl6: my $x = 1; $x ++; say $x 02:53
exp_evalbot kp6 r20630: OUTPUT[syntax error at position 14, line 1 column 14:␤my $x = 1; $x ++; say $␤ ^ HERE␤]
..pugs: OUTPUT[2␤]
..rakudo r28040: OUTPUT[2␤]
..elf r20630: OUTPUT[Parse error in: /tmp/sImoTprkeu␤panic at line 1 column 11 (pos 11): Can't understand next input--giving up␤WHERE: my $x = 1; $x ++; say $x␤WHERE: /\<-- HERE␤ STD_red/prelude.rb:98:in `panic'␤ STD_red/std.rb:76:in `scan_unitstopper'␤ STD_red/std.rb:224:in
..`comp_unit'␤ STD_red/std.rb:210:in `_UNIT'␤ ./../STD_red/STD_red_run:1...
02:54 c1sung left
Auzon pugs: sub infix:<++> ($a, $b) {2*($a + $b)}; my $a = 1; my $b = 2; say $a ++ $b; 02:55
exp_evalbot OUTPUT[*** ␤ Unexpected "$b"␤ expecting operator, ":" or ","␤ at /tmp/o0jwpFzpxu line 1, column 72␤]
Auzon pugs: sub infix:<++> ($a, $b) {2*($a + $b)}; my $a = 1; my $b = 2; say $a ++, $b;
exp_evalbot OUTPUT[12␤]
Auzon tsk.
02:55 c1sung joined
Auzon perl6: »*« 1..6 03:10
03:10 exp_evalbot left
Auzon Oops 03:10
03:10 exp_evalbot joined
bacek_ Auzon: its all your fault :) 03:10
Auzon Indeed. I should investigate why that happens.
03:18 Limbic_Region left 03:37 bacek left 03:49 cjfields joined 03:50 cjfields left 04:05 alc left 04:28 wknight8111 left 04:30 nipotan is now known as nipotaway 04:49 FurnaceBoy left 04:58 alc joined 05:03 Maghnus left 05:05 Maghnus_ joined, Maghnus_ is now known as Maghnus 05:14 stevan_ left
bacek_ perl6: say reduce { $^a + $^b }: 100, (1..6) 05:22
exp_evalbot kp6 r20630: OUTPUT[syntax error at position 24, line 1 column 24:␤say reduce { $^a + $^b }: 100, (1..6␤ ^ HERE␤] 05:23
..pugs: OUTPUT[121␤]
..rakudo r28042: OUTPUT[Statement not terminated properly at line 1, near ": 100, (1."␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
..elf r20630: OUTPUT[syntax error at (eval 14) line 3, near ") :"␤ at ./elf_f_faster line 4492␤]
bacek_ rakudo: say reduce { $^a + $^b }, 100, @array 05:24
exp_evalbot rakudo r28042 OUTPUT[Scope not found for PAST::Var '@array'␤current instr.: 'parrot;PCT::HLLCompiler;panic' pc 156 (src/PCT/HLLCompiler.pir:103)␤
bacek_ rakudo: say reduce { $^a + $^b }, 100, (1..6)
exp_evalbot rakudo r28042 OUTPUT[No applicable methods.␤current instr.: '_block11' pc 73 (EVAL_13:27)␤
05:29 halner left 05:46 pochi left
moritz_ good morning 05:54
Auzon good evening moritz_ :)
05:55 jan______ left
Auzon Do you have any ideas why evalbot occasionally segfaults? 05:55
moritz_ no :(
Auzon Did it happen before I added elf support? 05:56
I don't remember it happening before, and I see it mostly happen when people use 'perl6'
moritz_ yes, it happened before
05:58 redicaps joined
Auzon OK. Well, I think I can easily adjust the 'perl6' usage to process and return the output for one implementation at a time. That way, we can at least see which one dies. 06:00
Plus, it'd be nice for the output to not hang on all of them. 06:01
moritz_ aye 06:03
06:11 alanhaggai left
bacek_ pugs: say reduce { }, (1..5); 06:14
exp_evalbot OUTPUT[*** Cannot cast from VList [VInt 1,VInt 2,VInt 3,VInt 4,VInt 5] to Pugs.AST.Internals.VCode (VCode)␤ at /tmp/0s34iW4H9r line 1, column 5-23␤]
bacek_ pugs: say reduce { -1 }, (1..6);
exp_evalbot OUTPUT[*** Cannot reduce() using a unary or nullary function.␤ at /tmp/Y4UdIUNjjl line 1, column 5-26␤]
Auzon moritz_: I took a look at doing that for evalbot, but I'll need to review the docs for Bot::Basicbot. I'll handle that in the morning. night & 06:15
moritz_ Auzon: good night 06:16
06:21 fullermd left 06:30 BinGOs joined 06:34 luqui joined 06:37 justatheory left 06:44 meppl joined 07:08 masak joined 07:21 mj41 joined 07:22 mj41 left 07:29 fullermd joined 07:32 redicaps left 07:37 mj41_ joined 07:43 mj41_ is now known as mj41 07:44 mj41 left 07:45 mj41 joined 08:01 Torment joined 08:10 lg left
masak does prefix ++ in perl 6 return an lvalue? 08:11
moritz_ dunno. Why would anyone want to assign to it? 08:12
masak just wondering if `++$var++` is allowed syntax :)
moritz_ if you want to assign to it, there's no point in incrementing it
masak: don't think so 08:13
masak there is in the above case
(not advocating it, just curious)
moritz_ I can find no evidence in the specs, but I think that it's forbiddne, much like in C and perl 5 08:18
08:18 Jedai left
masak if the specs are silent, it's likely the same as in perl 5, yes 08:22
08:22 elmex joined
bacek_ pugs: my $x=1; say ++$x++; 08:23
exp_evalbot OUTPUT[*** Can't modify constant item: VInt 2␤ at /tmp/dhBKWga9Ub line 1, column 14-20␤]
08:23 pmurias joined
bacek_ perl6: say my $x=1; say (++$x)++; 08:24
exp_evalbot kp6 r20630: OUTPUT[syntax error at position 23, line 1 column 23:␤say my $x=1; say (++$x)++␤ ^ HERE␤]
..pugs: OUTPUT[1␤*** Can't modify constant item: VInt 2␤ at /tmp/nRyP7H6vTZ line 1, column 18-26␤]
..rakudo r28044: OUTPUT[Method 'lvalue' not found for invocant of class 'PAST::Stmts'␤current instr.: 'parrot;PAST::Compiler;as_post' pc 2554 (src/PAST/Compiler.pir:684)␤
..elf r20630: OUTPUT[Can't modify preincrement (++) in postincrement (++) at (eval 14) line 4, near ")++"␤ at ./elf_f_faster line 4492␤]
bacek_ elf++
:)
moritz_ bacek_: that seems like an error message from perl 5 08:25
08:44 bacek_ left 09:20 zamolxes joined
moritz_ pugs: say $?PUGS_VERSION 09:42
exp_evalbot OUTPUT[Perl6 User's Golfing System, version 6.2.13, October 17, 2006 (r18093)␤]
09:48 chris2 joined 09:49 b_jonas joined, ruoso joined 09:50 c9s left 09:59 bacek joined
ruoso :) SMOP grant was approved :) for those who didn't see the news :) 10:09
moritz_ ruoso: congratulations 10:10
(I did see it, but I didn't see you on IRC ;)
ruoso I was on vacations in amsterdam
4 days offline ;)
10:12 masak left
moritz_ a nice city with quite some perl mongers ;) 10:12
ruoso yeah... few friends of mine work at booking... 10:13
10:40 alc left
pmurias ruoso: when will the grant start? 10:43
ruoso I will need a couple weeks before starting...
bacek rakudo: say (1+undef).WHAT; 10:44
exp_evalbot rakudo r28046 OUTPUT[Failure␤]
bacek rakudo: my $a = 1 + undef; say $a; say $a.WHAT;
exp_evalbot rakudo r28046 OUTPUT[1␤Failure␤]
ruoso pmurias, but the grant manager wasn't assigned yet
pmurias will you use kp6 or elf?
bacek Ho! Gotcha!
rakudo: my $a = 1 * undef; say $a; say $a.WHAT; 10:45
exp_evalbot rakudo r28046 OUTPUT[0␤Int␤]
pmurias bacek: that behaviour is correct 10:46
moritz_ maybe infix:<+> is just missing a multi?
pmurias bacek: the second one that is ;)
ruoso pmurias, I should get to creating the Ast object types... 10:48
which means that something might generate C code that will be initialized as the ast tree itself
maybe that will already be something runnable
which would mean that it wouldn't really be AST, but more higher-level SLIME nodes 10:49
my priority is getting p6opaque done 10:50
because that is really the most important milestone
after that, we can have the s1p compiler to write other classes...
and I think we'll then advance in a faster pace...
and maybe even port kp6 or elf to bootstrap in SMOP 10:52
pmurias is the option to have multiple -e's and input files in elf usefull? 11:00
moritz_ pmurias: usually everything after the input file or the first -e is used for @*ARGV
pmurias: how would you distinguish these two cases? 11:01
11:01 pjcj left
pmurias the current elf design uses "--" 11:01
kp6 only allowes multiple -e's 11:02
11:14 hanekomu joined
ruoso pmurias, moritz_, I've just uploaded a new version of the "Perl 6 is just a SMOP" talk slides... would you like to take a look and comment ;) ? 11:14
people.debian.org/~ruoso/SMOP.pdf 11:15
moritz_ ruoso: looking...
11:16 luqui left 11:18 IllvilJa left
moritz_ ruoso: in Perl 6 you'd say "Int $a = something()" 11:19
not Integer
ruoso ops...
moritz_++, thanks 11:20
ruoso replacing "KP6 backend" by "kp6/elf backend" in the "things to come" slide... 11:24
moritz_ ruoso: how long is the talk? 11:25
ruoso 40 minutes
I realise I'll have to skip some parts
moritz_ don't think so
2 slides per minute on average 11:26
that's fast, but most slides contain only a few words
ruoso some of these slides are entirely entertaining...
moritz_ yes
like mOOOOse
ruoso yeah...
one thing I'd like to see in the near future (as soon as I finish p6opaque) would be SMOP::P6Opaque as a Moose metainstance... 11:28
moritz_ does that mean you have to interact with perl 5 on the C level? 11:30
b_jonas wow,
ruoso only to implement SMOP::P6Opaque...
after that, it would be only in Moose level
b_jonas you mean a uniform p6 interface to any p5 moose object? 11:31
ruoso b_jonas, my point is that this is what P6 is all about...
b_jonas or the reverse, a p5 moose interface for any p6 object?
ruoso I think both scenarios are very likely
both perl5 embeded in SMOP 11:32
as SMOP embeded in perl5
b_jonas rouso: yeah, but uniform interfaces are the part of parrot and p6 are what I least believe in
ruoso b_jonas, not really...
b_jonas I don't like automatic uniform interfaces between languages
ruoso parrot intends to uniform them in the syntax level
b_jonas as such, I don't like the Inline:: p5 modules
ruoso all of them would be compiled targeting parrot
b_jonas really, there are two things I don't like 11:33
ruoso this is different from interfacing with another language's runtime
b_jonas one is _transparent_ interfaces between two language's objects
ruoso which is how I think it should be done...
there's no "transparent" interface... this is something that I still couldn't find to be doable...
you have an "abstraction layer" 11:34
just like XS
SamB what is "transparent"?
ruoso but less ugly
b_jonas that is, I dislike an interface that wants to fool you that the foreign object has the interface similar to native objects
ruoso SamB, transparent means "without typemap"
moritz_ or "hidden"
b_jonas rouso: no, without typemap is the second thing I dislike
that I call automatic interfaces
like when they promise you'll be able to call any C function from perl6 without any wrapper writing 11:35
ruoso b_jonas, the thing is... if you get your protocol to a lower level (meta-object-protocol) you're not hiding...
you're simply communication in a lower level
SamB b_jonas: ludicrous
b_jonas SamB: yeah, they said it a bit differently
and for parrot handling it actually
on the same conference where Leo later answered to my question -- if I remember correctly -- that blocking c extension functions for parrot will have to be stackless because they don't want all the setcontext magic ruby does for threads 11:37
ruoso b_jonas, once you have typing (which you don't in p5), you can have type coercion, then you can coerce Perl 6 Int to C int in the runtime...
only using the MOP
SamB how C int can implement MOP? 11:39
ruoso it's not C int that implements it
but the invocation code that calls the C function
just like XS
SamB why does "XS" make me think of the deathstar? 11:40
ruoso but you can use the runtime's type coercion to always receive Perl 6 native int
and avoid dealing with all the perl magic
b_jonas ruoso: yes, the haskell c interface for example does a quite good job with it
but that doesn't mean you can use any c function
ruoso and that's the big difference from XS
b_jonas there are still lots of requirements a c function you want to call from haskell can do
pmurias it should be possible to use any c function from smop withought writing any wrappers 11:41
SamB lots?
b_jonas (not surprising, since just a 'c function' can do _anything_)
SamB the main restriction is that you need to know the types of the arguments
b_jonas SamB: there's more than that
ruoso SamB, not really... there are IO issues and all other sort of hidden traps 11:42
SamB the second one is that it can't deal in structs...
pmurias SamB: lots refered to?
b_jonas I guess you'll probably still easily be able to wrap _lots_ of c functions
SamB well, that is, it can't deal in directly-passed-or-returned structs
b_jonas like most from libgsl
I think libgsl still doesn't have a full wrapper in _any_ language because it has lots of functions 11:43
even the haskell interface wraps just some of them
it's huge like lapack
anyway, ttl
bye
ruoso SamB, one big difference in P6 is that this struct can be the value itself... not a SV* with an int that happens to be the pointer address of that struct... 11:44
11:44 b_jonas left
ruoso and if the library implements some kind of OO-like API... you can actually map it directly... 11:44
instead of fake-ing everything in p5land to dispatch the methods... 11:45
SamB so you have gtk bindings?
ruoso yes...
you can actually have that today in p5
SamB that use this principle of operation? 11:46
ruoso but only thanks to a whole set of dark magic implemented by the perl-gtk guys
SamB, perl-gtk uses a lot of tricks to make it work...
it would be natural in P6 to have different object systems co-existing
like... 11:47
instead of the p5 interpreter dispatching the method call
to a fake method in p5 land that dispatches the real method in the gtk land
bacek pugs: my $a = 1 + undef; say $a; say $a.WHAT;
exp_evalbot OUTPUT[1␤Int␤]
moritz_ ruoso: re slides, the "only exceptions are native types" lacks a bit context 11:48
ruoso moritz_, ok... I'll work on it... maybe I don't even need to mention it, as it's a technical detail... 11:49
pmurias ruoso: it's possible to encapsulation with other types too, it's just not socially accepted ;) 11:54
* break encapsulation
SamB you don't have anything like __getattr__ in p5?
ruoso pmurias, sorry... I missed the context... 11:55
SamB, __getattr__?
11:56 chris2 left
SamB Python thing... 11:56
ruoso SamB, you mean, from python?
no... actually you can't
SamB p5 hasn't got anything like that?
ruoso because in p6 the interpreter can't assume to know the internal format of the object 11:57
except for the native types
pmurias ruoso: it's possible to break the encapsulation (assume internals knowledge) with non-native types, hoping nothing bad happens 11:58
ruoso it's a gray-area
because SMOP actually uses it for its benefit...
SamB eh?
ruoso but it only does that after checking for a known responder interface 11:59
for bootstrapping reasons
SamB __getattr__ is implemented by the object...
ruoso SamB, oh sorry... I misunderstood it
SamB you must be thinking of getattr()
;-P
if you think those look kinda similar, you'd be right ... they are quite closely related 12:00
ruoso but the "get attr by name" thing might not be doable in p6
SamB hmm. what did you mean by "fake method"?
ruoso because there are no "public attributes" in p6
SamB, a method that simply proxies to the real method 12:01
SamB where do they come from?
ruoso SamB, the fake methods? they are declared somewhere in gtk initialization...
but I'm not really sure... 12:02
SamB ah, so they aren't transient?
ruoso transient? 12:03
SamB eh, I should read up on perl objects... 12:04
bacek perl6: say (1..4).map({$^a+$^b+$^c})
exp_evalbot kp6 r20630: OUTPUT[syntax error at position 4, line 1 column 4:␤say (1..4).map({$^a+$^b+$^c}␤ ^ HERE␤]
..pugs: OUTPUT[64␤]
..rakudo r28046: OUTPUT[64␤]
..elf r20630: OUTPUT[Global symbol "$c" requires explicit package name at (eval 14) line 3.␤ at ./elf_f_faster line 4492␤]
SamB what was the manpage for that again?
moritz_ S12 (for Perl 6)
12:07 pbuetow joined
ruoso moritz_, you think I could remove the "native types" part of the presentation/ 12:08
?
moritz_ ruoso: yes
ruoso nice
12:14 Simor joined
ruoso updated the slides... 12:14
12:15 wknight8111 joined
bacek rakudo: say (1+undef)+undef 12:16
exp_evalbot rakudo r28046 OUTPUT[MMD function __add not found for types (102, 102)␤current instr.: 'infix:+' pc 9730 (src/gen_builtins.pir:6757)␤
SamB that's really ugly
why does it say types 102,102 12:17
ruoso SamB, that's probably the id of the type...
bacek pugs: say (1+undef)+undef
exp_evalbot OUTPUT[1␤]
SamB well it's a stupid-looking sort of id
moritz_ SamB: I agree, but what would you do about anonym types? 12:18
SamB hmm. 12:19
not sure...
avoid having them?
ruoso SamB, that's absolutely impossible 12:20
moritz_ SamB: no way
SamB: $variable but $role; creates an anynyom type
ruoso there's no way to bootstrap a type system without anonymous types
SamB hehehe
I have not seen a single anonymous type in Haskell or Coq or Agda 2... 12:21
moritz_ SamB: usually that's the point of them, you don't see them at all ;-)
ruoso ok... which type does the namespace have?
SamB namespace? 12:22
ruoso I mean... the namespace implementation
you need an anonymous type to implement namespace
SamB points out that none of those systems is object oriented
ruoso that's not the point...
they are not bootstrapped type systems 12:23
that's the point
SamB oh
not familiar with that terminology
bacek rakudo: say ~(1,2,3,4).map({ $^a+$^b+$^c+$^d+$^e })
exp_evalbot rakudo r28047 OUTPUT[10␤]
ruoso in SMOP, for instance, everything is SMOP__Object*
even the interpreter
bacek rakudo: say ~(1,2).map({ $^a+$^b+$^c+$^d+$^e })
exp_evalbot rakudo r28047 OUTPUT[MMD function __add not found for types (102, 102)␤current instr.: 'infix:+' pc 9730 (src/gen_builtins.pir:6757)␤
ruoso SamB, that means that even the SMOP Object is implemented using SMOP Object... 12:25
that's a bootstrapped type system
pmurias isn't it a bootstrapped object system? 12:26
ruoso and all of those types are anonymous so far, as SMOP doesn't have namespaces yet
pmurias, I can't really tell the difference between type system and object system...
SamB oh, a bit like Agda with type-in-type? apparantly that admits a form of russel's paradox...
i.e. it gives an inconsistant logic 12:27
ruoso that's the precise reason of why it's named "bootstrap"
SamB so, I hope you weren't planning on using Perl 6's typesystem as a formal logic 12:28
;-P
bacek perl6: say undef + 1 + undef 12:30
exp_evalbot kp6 r20630: OUTPUT[1␤]
..pugs: OUTPUT[1␤]
..rakudo r28047: OUTPUT[MMD function __add not found for types (102, 91)␤current instr.: 'infix:+' pc 9730 (src/gen_builtins.pir:6757)␤
..elf r20630: OUTPUT[Use of uninitialized value in addition (+) at (eval 14) line 3.␤Can't call method "Str" on an undefined value at ./elf_f_faster line 617.␤ at ./elf_f_faster line 4492␤]
12:31 cjfields joined 12:32 cjfields left 12:50 vixey joined
ruoso one interesting thing to do (that anyone could try) would be to implement the default HOW of Perl 6 in Perl 6 12:58
www.perlfoundation.org/perl6/index....mop_oo_api describes the API it should implement... 12:59
lambdabot Title: SMOP OO API / Perl 6
ruoso svn.pugscode.org/pugs/v6/smop/src-s1p/Object.pm is an example of how that code could look like 13:00
maybe the .^! calls in Object.pm should be turned into .^ calls
or maybe even the code that call those .^! methods should reside into the HOW implementation 13:01
bacek perl6: say undef - 1
exp_evalbot kp6 r20630: OUTPUT[-1␤]
..pugs: OUTPUT[-1␤]
..rakudo r28047: OUTPUT[MMD function __subtract not found for types (102, 91)␤current instr.: 'infix:-' pc 9743 (src/gen_builtins.pir:6765)␤
..elf r20630: OUTPUT[Can't call method "Str" on an undefined value at ./elf_f_faster line 617.␤ at ./elf_f_faster line 4492␤]
bacek pugs: say undef - "123" 13:03
exp_evalbot OUTPUT[-123␤]
13:04 alanhaggai joined 13:14 FurnaceBoy joined 13:16 meppl left
pugs_svnbot r20631 | ron++ | just a subversion test 13:20
r20632 | ron++ | just a subversion test
13:33 eternaleye left 13:43 rdice joined
rakudo_svn r28049 | pmichaud++ | [perl6]: 13:46
r28049 | pmichaud++ | * Fix a number of operators to do proper casting.
pugs_svnbot r20633 | diakopter++ | [yap6] reverting to earlier MJD system; using CPS instead...
diakopter ruoso: congrats
ruoso diakopter, thanks
pmichaud yes, ruoso, congrats
I'm glad to see it funded
ruoso :) 13:47
did you two see the slides I'm preparing... 13:48
for the portuguese perl workshop
pmichaud, diakopter, people.debian.org/~ruoso/SMOP.pdf 13:49
moritz_ pugs: say {$^a**2}.reduce: 1, 2, 3, 4 13:55
exp_evalbot OUTPUT[<SubBlock(<anon>)>␤*** Cannot cast from VList [VInt 1,VInt 2,VInt 3,VInt 4] to Pugs.AST.Internals.VCode (VCode)␤ at /tmp/LLg4znQI5A line 1, column 1 - line 2, column 1␤]
ruoso moritz_, the ":" indicates the invocant... 13:58
are you sure that was your intention? 13:59
bacek pugs: say (1..2).reduce({$^a})
exp_evalbot OUTPUT[*** Cannot reduce() using a unary or nullary function.␤ at /tmp/8sPcrFY8Qz line 1, column 5 - line 2, column 1␤]
moritz_ ruoso: I know, I just didn't look up in which class the reduce() was defined
but List makes much more sense than Code
bacek pugs: say 'for pmichaud :)'; say (1..2).reduce({$^a}) 14:00
exp_evalbot OUTPUT[for pmichaud :)␤*** Cannot reduce() using a unary or nullary function.␤ at /tmp/L24ZyG9TBV line 1, column 28 - line 2, column 1␤]
14:00 Simor left 14:01 Lorn joined 14:03 cmarcelo joined
ruoso pugs: sub foo (*@args) { @args.join(',') }; (1..5).reduce(&foo) 14:04
exp_evalbot OUTPUT[*** Cannot reduce() using a unary or nullary function.␤ at /tmp/OORfmYsbKK line 1, column 39 - line 2, column 1␤]
ruoso pugs: sub foo (*@args --> String) { @args.join(',') }; say (1..5).reduce(&foo) 14:05
exp_evalbot OUTPUT[*** Cannot reduce() using a unary or nullary function.␤ at /tmp/VSwh18zPrI line 1, column 54 - line 2, column 1␤]
ruoso what does "unary or nullary function" means here?
moritz_ pugs: say $*PROGRAM_NAME 14:06
exp_evalbot OUTPUT[␤]
ruoso perl6: sub foo (*@args --> String) { @args.join(',') }; say (1..5).reduce(&foo) 14:07
diakopter ruoso: I hope you don't mind I stole smop's former name (yapc)
pugs_svnbot r20634 | diakopter++ | [yapc] removing calculator POC
14:07 exp_evalbot left
diakopter yap6 14:07
ruoso diakopter, of course I don't...
14:07 exp_evalbot joined
diakopter ok thanks 14:07
ruoso what are you using it for?
diakopter a continuation-passing-style recursive-descent parser for Perl 6 in Perl 5
ruoso cool
diakopter based on MJD's code in Higher Order Perl
14:07 meppl joined 14:08 iblechbot joined
diakopter exp_evalbot: come back 14:08
oh, it's back
pmichaud pugs: multi sub foo() { 0; } multi sub foo($a, $b) { $a + $b }; say (1..10).reduce(&foo); 14:09
exp_evalbot OUTPUT[55␤]
moritz_ is glad he started evalbot in a 'while true; do $start_evalbot; done' loop
pmichaud pugs: multi sub foo() { 0; }; say (1..10).reduce(&foo);
exp_evalbot OUTPUT[*** Cannot reduce() using a unary or nullary function.␤ at /tmp/ZvRPKwxZSt line 1, column 31-51␤]
pmichaud pugs: multi sub foo($a, $b) { $a + $b; }; multi sub foo() { 0; }; say (1..10).reduce(&foo); 14:10
exp_evalbot OUTPUT[55␤]
ruoso pugs: multi sub foo (*@args --> String) { @args.join(',') }; say (1..5).reduce(&foo) 14:13
exp_evalbot OUTPUT[*** Cannot reduce() using a unary or nullary function.␤ at /tmp/GSMNAjBLcf line 1, column 60 - line 2, column 1␤]
ruoso pugs: multi sub foo ($a, $b --> String) { @args.join(',') }; say (1..5).reduce(&foo)
exp_evalbot OUTPUT[*** ␤ Unexpected "."␤ expecting "::"␤ Variable "@args" requires predeclaration or explicit package name␤ at /tmp/48Pu6zAtl2 line 1, column 42␤]
ruoso pugs: multi sub foo ($a, $b --> String) { ($a,$b).join(',') }; say (1..5).reduce(&foo)
exp_evalbot OUTPUT[1,2,3,4,5␤]
ruoso weird...
pugs: multi sub foo ($a, $b, $c --> String) { ($a,$b,$c).join(',') }; say (1..5).reduce(&foo) 14:14
exp_evalbot OUTPUT[1,2,3,4,5␤]
ruoso pugs: multi sub foo ($a, $b, $c --> String) { ($a,$b).join(',') }; say (1..5).reduce(&foo)
exp_evalbot OUTPUT[1,2,4␤]
ruoso weird...
moritz_ ruoso: why is that weird?
ruoso the *@args version was supposed to wok
work
moritz_ at least the first seems very intuitive to me
ruoso or is that a known bug?
pugs: multi sub foo (*@args --> String) { @args.join(',') }; say (1..5).reduce(&foo) 14:15
exp_evalbot OUTPUT[*** Cannot reduce() using a unary or nullary function.␤ at /tmp/9qeKf8xSJy line 1, column 60 - line 2, column 1␤]
ruoso this is not a unary function
moritz_ ruoso: probably .arity doesn't return the right thing for slurpy signatures 14:16
ruoso pugs: multi sub foo (*@args --> String) { @args.join(',') }; say &foo.arity
exp_evalbot OUTPUT[1␤]
ruoso pugs: multi sub foo ($a, $b --> String) { }; say &foo.arity 14:17
exp_evalbot OUTPUT[2␤]
ruoso hmm
moritz_ it should probably return Whatever for a slurpy sig
ruoso actually Infinity
moritz_ or better, a range
ruoso: no, it doesn't accept infinitely many arguments
ruoso moritz_, actually it does
my $capture = some_funky_sub(); foo(|$capture) 14:18
moritz_ a range can describe things like sub foo($a, *@rest_args) more accurately than a simple Infinite
ruoso but you can't figure out what the end of the range is..
moritz_ the end of the range is Infinity, sure 14:19
ruoso oh...
ok... I see.
but :(*@args) is still just Inifinity
but :($something, *@args) is then 1 .. Infinity 14:20
moritz_ why? it won't fail if you pass 1 arg
and 1 != Infinity
ruoso but you can actually pass 0 arguments as well... 14:21
but it's just semantics, actually
14:21 FurnaceBoy is now known as FurnaceBoyAway
ruoso returning a range kinda implies the ($some, $foo, *@args) scenario... 14:21
moritz_ ruoso: the specs are a bit more explicit in the case of the 'want' builtint... 14:22
Their difference is that C<.arity> considers only mandatory parts,
while C<.count> considers also optional ones, including C<*$>:
ruoso as well as for ($some, $foo?, $bar?)
ah...
ok
that would make the arity for *@args to 0
moritz_ and .count to Infinity 14:23
ruoso pugs: multi sub foo (*@args --> String) { }; say &foo.count
exp_evalbot OUTPUT[*** No such method in class Sub: "&count"␤ at /tmp/52l5Q1Iqin line 1, column 45 - line 2, column 1␤]
ruoso so reduce should use .count instead of .arity 14:24
moritz_ if (
urks
ruoso: yes
14:27 zamolxes left, cmarcelo left 14:31 elmex left 14:48 polettix joined 14:49 froh-doh joined 14:50 polettix left, froh-doh is now known as polettix 15:12 TJCRI joined 15:13 hanekomu left 15:17 Psyche^ joined 15:25 Ched- left 15:33 Patterner left, Psyche^ is now known as Patterner 15:46 jhorwitz joined 15:48 IllvilJa joined
pugs_svnbot r20635 | moritz++ | [spec] cleared old pugs specific environment variable 15:52
15:54 justatheory joined
pugs_svnbot r20636 | moritz++ | [spec] removed superfluous TODO marking from radix.t 15:55
pmurias reduce with a 3 argument sub only works on odd number of elements? 15:57
ruoso pmurias, multipliers of 3, maybe.... 15:59
pmurias ?
how would you reduce 6 elements? 16:00
moritz_ actually S29 says the function has to work with two items at a time 16:01
pmurias just seen it 16:02
;)
16:02 alanhaggai left
pmurias but it's possible to write a reduce version for a sub which takes $n elements 16:02
the problem is would it be usefull? 16:03
moritz_ $res = &$expression($res, $cur);
is the '&' mandatory here?
pmichaud no.
moritz_ good
pmichaud my $foo = sub { ... }; ..... $foo();
ruoso moritz_, S03 'Reduction operators' seems to imply that it might receive as many arguments as it wants 16:05
it might actually receive the entire list 16:06
moritz_ ruoso: actually reduction operators don't rely on reduce() at all
ruoso sure sure...
moritz_ ruoso: for example they take associativity into account
ruoso but it says that, 1 foo 2 foo 3, can be written as [foo] 1, 2, 3 -- wich seems to imply that 1,2,3 are arguments to foo 16:07
moritz_ I don't understand that conclusion 16:08
16:08 TJCRI left
pmurias infix operators take 2 arguments 16:08
pmichaud infix:, has list associativity, and thus may take more than 2 arguments 16:09
pugs_svnbot r20637 | pmurias++ | [elf] -I's are .push'ed instead of being .unshift'ed and reversed 16:11
pmurias pmichaud: infix list operators take all the argments
* arguments 16:12
pmichaud i was simply pointing out that some infix: operators don't have 2 arguments.
or
that there exist some infix: operators that work with more or less than 2 arguments.
pmurias pmichaud: how can they work with less arguments?
s/less arguments/1 argument/ 16:13
pmichaud if they're invoked directly, as in &infix:<,>($x)
or consider the case of (1,)
pmurias are ones which take exactly 3 allowed? 16:14
pmichaud I don't know about that. Doesn't seem like it.
ruoso but...
it wouldn't be that nice to have that kind of instrospection for every infix operation... 16:15
maybe do something at bind-time 16:16
to realise how many arguments were bound
pmichaud afk (errands, lunch)
ruoso and "consume" the items of the list to the next call 16:17
16:17 FurnaceBoyAway is now known as FurnaceBoy
ruoso that actually might work very nicely 16:17
clkao win 53 16:18
moritz_ win 53, loose 52 - that's life.
ruoso the capture sent to the operator would keep the state of which arguments were already subject to bind 16:19
and the capture would then be reused in the next call
moritz_ sounds kinda tricky ;)
ruoso sounds the kind of tricks that will make Perl 6 cool ;) ;) ;) 16:20
moritz_ hehe 16:21
ruoso it's even performance-friendly
as it would keep using the same capture as argument
instead of creating a new capture everytime
moritz_ but not very parrot-friendly
because it doesn't use explict captures for all sub/method/op calls
ruoso capture doesn't support foreign captures? 16:22
like... a capture that is actually bound to a C int array?
moritz_ actually I don't know
perhaps there is a way to reuse captures in parrot, and I don't know about it 16:23
ruoso the question is... is the invocation in parrot static? 16:24
if so, rakudo will have to implement an abstraction layer on top of it
to support different invocations...
moritz_ only hopes that parrot supports powerful incantations 16:25
Auzon ruoso: congrats on your grant :) 16:36
ruoso Auzon, thanks
:)
spinclad ruoso: 'What is in the way?': do you mean 'What is blocking the way', or 'What is on the way, coming' ? 16:37
ruoso what is on the way, coming...
would "What is on the way?" be more correct? 16:38
spinclad a case of english's 'in/on' distinction... yes, 'on the way'
ruoso thanks... 16:39
ruoso fixing...
spinclad 'in the way': sitting in the road; 'on the way': moving down the road
ruoso spinclad++, thanks... I've just uploaded a new version with that fix... 16:41
16:41 Psyche^ joined
Auzon pugs: say log 0 16:41
exp_evalbot OUTPUT[-Inf␤]
spinclad 'i was on my way till this cow got in my way' 16:44
pmurias ruoso: new version of what? 16:46
ruoso pmurias, the slides...
Auzon pugs: say ('A' .. 'Z').elems 16:53
exp_evalbot OUTPUT[26␤]
Auzon pugs: say elems('A' .. 'Z')
exp_evalbot OUTPUT[26␤]
16:56 nipotaway is now known as nipotan 16:57 Patterner left, Psyche^ is now known as Patterner 17:09 justatheory left 17:11 apeiron left 17:12 jhorwitz left 17:16 Lorn left 17:18 stevan_ joined 17:26 justatheory joined 17:27 sri_work joined 17:46 Jedai joined
moritz_ pugs_svnbot: you're late 17:49
pugs_svnbot r20638 | moritz++ | [irclog] #parrotsketch should link to its logs
moritz_ pugs_svnbot: thank you
17:55 justatheory_ joined 18:01 Torment left 18:04 justatheory left
meppl good night 18:05
18:08 sscaffidi left, sscaffidi joined, sscaffidi left, sscaffidi joined 18:10 meppl left, pmurias left 18:11 b_jonas joined 18:17 sscaffidi left, sscaffidi joined 18:18 jhorwitz joined 18:20 sri_work_ joined
ruoso home & 18:31
18:31 ruoso left 18:32 wknight8111_ joined 18:33 wknight8111_ left 18:36 sri_work left 18:40 justatheory_ left 18:43 justatheory joined 18:47 andy123 joined
pugs_svnbot r20639 | diakopter++ | [yap6] 19:02
r20639 | diakopter++ | Nested error handling (and reporting!) working; the parser author
r20639 | diakopter++ | can define helpful syntax error messages. Also, trace output
r20639 | diakopter++ | considerably improved.
19:02 sscaffidi left 19:03 andy123 left 19:07 sscaffidi joined 19:13 cjfields joined 19:17 jhorwitz_ joined 19:28 alester joined 19:29 sscaffidi left 19:30 kcwu left 19:32 jhorwitz left 19:40 sscaffidi joined 19:46 pmurias joined
avar $ ../../parrot perl6.pbc -h 19:52
perl6.pbc
Usage: perl6 [switches] [--] [programfile] [arguments]
19:52 ruoso joined
avar Is this rakudo still?:) 19:52
19:54 Ched- joined
moritz_ avar: what else could it be? ;-) 19:56
avar nothing really, I just checked out with git-svn and figured I might be getting something ancient:) 19:59
I thought they'd updated -h to say rakudo even if the directory was still called perl6
moritz_ avar: I think it's mostly the standard HLL compiler template thingy 20:01
[particle] no, i modified it for rakudo i think, but when you build rakudo, it makes an executable called 'perl6' 20:03
not called 'rakudo'
avar Oh the humanity! 20:05
pmurias avar: do you think the kp6 CLI could be improved? 20:15
avar No wai man, it's PERFECT
>:)
vixey What's CLI?
pmurias command line interface
avar: ;) 20:16
elf's CLI sucks, so i thought it might be a good idea to steal kp6's
avar yeah it was pretty neat 20:17
I remember cleaning it up, could still use some more cleanup:)
stole most of it from pugs's
20:19 mncharity joined
pmurias mncharity: hi ;) 20:20
s/;)/:)/
mncharity hi pmurias. just stopping by briefly to ping stevan_.
:)
pmurias avar: you mean cleaning up internals or implementation? 20:21
avar the command line switches
e.g. -B ad -C
pmurias mncharity: i'm replacing the CLI is elf's abilitity to load multiple files at once ./elf file1 file2 usefull? 20:22
diakopter use no warnings 'recursion'; # finally found it 20:26
no warnings 'recursion'; # I mean 20:27
20:29 sscaffidi left
Auzon Is printf() dead? 20:32
rakudo: say sprintf('%%');
exp_evalbot rakudo r28052 OUTPUT[Could not find non-existent sub sprintf␤current instr.: '_block11' pc 28 (EVAL_10:17)␤
Auzon hm. 20:33
pmurias Auzon: printf lives :( 20:34
20:34 jferrero joined
Auzon It's not in S29. 20:35
Why the sad face though?
It's easy enough to implement in P6, I think.
20:35 chris2 joined
pmurias it's in S29 20:35
it's in IO.pod
Auzon: it's trival with sprinf 20:36
Auzon Yeah.
doh. I don't have a local S29.
moritz_ svn.pugscode.org/pugs/docs/Perl6/Sp...ctions.pod 20:37
pugs_svnbot r20640 | diakopter++ | [yap6] 20:38
r20640 | diakopter++ | Still further improved error handling/reporting.
r20640 | diakopter++ | Non-trace-mode source filter speeds up execution by lots%
r20640 | diakopter++ | no warnings 'recursion'; # eliminates that parse-time warning.
r20640 | diakopter++ | Next task: produce an AST instead of evaluating arithmetic expressions
r20640 | diakopter++ | synchronously.
Auzon Ah. It's in Pugs, not whatever syn repo I'm using
mncharity pmurias: re CLI, my impression of what's about to happen is elf/ is going to get very conservative, moving towards having a release process, and most elf-ish development is going to move into a new elfish/on_perl5/ .
pmurias we can do a release at any point 20:39
mncharity more specifically, re is useful, yes, definitely.
diakopter mncharity: where can I find a specification/description of the AST format that elf expects
mncharity pmurias: but one of the objectives of the split is to allow folks to do arbitrary changes without the worry of "is it the right thing" and "will it break every one who is using elf". 20:40
so we avoid the longstanding pugs issue of "you want to refactor aggressively, and exploratorily, but there are other people (including yourself when you are working on other things) who depending on stability, so refactoring/exploration is badly impaired". 20:41
pmurias it think just creating a stable branch would be sufficent 20:42
mncharity: i'm not sure which calling convention should i implement
mncharity a single unitary unstable branch requires intensive communication and commonality of vision among all the developers. part of the reason git is displacing svn is that is expensive and hard to come by. 20:43
pmurias CAPTURE($capture), NAMED([...],{...}) and FAST($arg1,$arg2,$arg3,[@slurpy])?
mncharity diakopter: re "AST format that elf expects", one problem is divergence of what parrot and elf call an ast. the Match tree resulting from parsing 20:44
is generated by STD_red's std.rb, based on STD.pm (old, with exceptions noted by NONSPEC flag comments), and consumed by IRx1_FromAST_create.pl. No other spec/desc. The IR nodes are 20:46
described in IRx1_nodes_create, with some additional (and soon to grow) methods in IRx1_Analysis. Semantics are implicit in IRx1_Analysis and EmitSimpleP5, and somewhat in flux (ie, undef vs empty array for missing or non-applicable list-of-something fields). No other spec/desc. 20:48
diakopter mncharity: thanks! :)
mncharity IRx1_FromAST_create has a couple of code massage regexps at bottom. 20:49
s/Semantics are implicit in(.*)/IRx1_FromAST, $1/ 20:50
pmurias mncharity: POSITIONALS_ONLY() would be propably usefull to for p5 compat
20:51 Blasi left
mncharity pmurias: re calling convention, there are a couple of possibilities. but it may be better to defer named arguments for some days, as my impression is Emit is about to be gutted, with much of it moved into IRx1_Analysis. 20:53
or, a different answer, might be to set up an elfish/on_perl5 now (a makefile like showcode's can be used) and start experimenting... 20:57
pmurias mncharity: i'm thinking of add a different backend 20:58
mncharity me too. :) 20:59
moritz_ parrot? smop? lisp? javascript? 21:00
mncharity which is the main motivation for moving anything p6-ir-and-semantics-related out of Emit.
pmurias it thought a p5 with named arguments ;)
but any of the ones moritz wants are fine 21:01
smop is a bit too young
Patterner CamelFish is the JavaScript in Parrot, right?
spinclad ecmascript methinks 21:02
moritz_ Patterner: there's 'eclectus' and 'ecmascript' in languaes/ in parrot trunk
pmurias mncharity: re javascript oo, it's crapy so i wouldn't bet on it maping to anything cleanly
mncharity re "p5 with named arguments", neat. elfish/on_perl5_wna/ :) I don't currently expect to do much IR change, so simply copying Emit and modifying it should work. 21:04
elfish/on_perl5_named/
eh 21:05
pmurias mncharity: why keep it in a seperate directory?
Auzon elf: 31 -12 21:07
exp_evalbot RESULT[19␤]
pugs_svnbot r20641 | pmurias++ | [elf] a new incomplete command line iterface 21:08
r20641 | pmurias++ | (checking it in so i don't accidently screw up anything ;)
mncharity elf/ is the production compiler. it just happens to be one folks can easily make derivatives of. such derivative projects go elsewhere, just as pugs derivatives have gone elsewhere than src/pugs/. the elf/ compiler will draw changes from the various derivatives, in a git like manner, but can be depended upon to always work, unlike, obviously, the derivatives. 21:09
Auzon pugs: say sprintf('%u', -42);
exp_evalbot OUTPUT[*** Ix{Integer}.index: Index (-42) out of range ((0,15))␤ at /tmp/xAH0EbFQBA line 1, column 1-23␤]
Auzon kp6: say sprintf('%u', -42);
exp_evalbot kp6 r20640: OUTPUT[syntax error at position 11, line 1 column 11:␤say sprintf('%u', -42)␤ ^ HERE␤]
Auzon elf: say sprintf('%u', -42);
exp_evalbot OUTPUT[4294967254␤]
Auzon elf: say sprintf('%04u', -42); 21:10
exp_evalbot OUTPUT[4294967254␤]
mncharity the other question is why misc/elfish/on_foo/ rather than misc/whatever_foo. and that's to simplify collective regression testing and such.
Auzon Evalbot must run on a 32 bit computer...
mncharity i'm out. back... probably tomorrowish. last questions? 21:11
spinclad mncharity: having fun? 21:12
mncharity :/
spinclad :\
mncharity moritz_: re langauges, not sure yet.
cheers & 21:13
21:13 mncharity left 21:21 rdice left 21:22 jhorwitz_ left
pmurias moritz_: re languages, any you're willing to help with? ;) 21:24
21:27 Torment joined 21:31 b_jonas left
moritz_ pmurias: good question. I have to admit that I know nearly no language besides perl 5 good enough 21:33
there are a few rather unpopular (like Eiffel) that don't seem to fit at all
and currently I'm selling my soul to the test suite and rakudo, so there's just a little tiime slice left - not enough to really influence the choice of a backend language 21:35
21:36 cjfields left 21:43 Jedai left 21:47 stevan_ left 21:50 iblechbot left 21:55 chris2 left 22:50 jferrero left 22:54 Limbic_Region joined 22:59 polettix left 23:07 polettix joined 23:20 drbean_ left 23:25 pmurias left, poletti joined 23:31 cjfields joined 23:34 polettix left

Powered by Cro and the Raku® Programming Language.

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

Thank you!