»ö« | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, alpha:, pugs:, std:, or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by moritz_ on 25 June 2010.
PerlJam what is DO ? 00:01
00:06 mtk left
jnthn sorear: I'm quite sure it's just a block in the normal flow of code execution. 00:07
say 1; { say 2 }; say 3;
is like (aside from the class decl
say 1; class Foo { say 2 }; say 3;
I think Rakudo has it that way too.
And afaik it's spec. Or at least meant to be in there somewhere. :-) 00:08
TimToady package blocks are at DO time
tylercurtis rakudo: say 1; class Foo { say 2; } say 3;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "class Foo "␤»
tylercurtis rakudo: say 1; class Foo { say 2; }; say 3;
p6eval rakudo 7579f7: OUTPUT«1␤2␤3␤»
TimToady sorear: this is specced at S10:88 00:09
anything that wants run at some other time than DO time should be in a phaser 00:14
but pretty much everything important about a class should be expressed by declarations within the class, not the normal code
cognominal thx jnthn for your 22:27 answer 00:22
PerlJam so DO time is "normal execution time"? 00:23
00:23 skids left 00:25 skids joined
PerlJam (ack -wa DO of the synopses doesn't return any results) 00:27
TimToady well, it's "do" not "DO"...
00:28 aubreyja_ joined
sorear I see. 00:33
perl6: my class Foo { ... }; Foo.new.say; my class Foo { };
p6eval rakudo 7579f7: OUTPUT«too few positional arguments: 1 passed, 2 (or more) expected␤ in main program body at line 1␤»
..pugs: ( no output )
sorear Is that legal?
sorear is quite hoping "no" 00:34
TimToady std: my class Foo { ... }; Foo.new.say; my class Foo { };
p6eval std 31614: OUTPUT«ok 00:01 111m␤»
00:34 ashleydev left
tylercurtis std: my $foo; $foo.say; my $foo; 00:35
p6eval std 31614: OUTPUT«Potential difficulties:␤ Useless redeclaration of variable $foo (see line 1) at /tmp/FWZ7DOwJ5s line 1:␤------> my $foo; $foo.say; my $foo⏏;␤ok 00:01 108m␤»
arnsholt Hmm. This given/when stuff was easier than I'd thought
TimToady tylercurtis: the ... makes it a stub decl
tylercurtis TimToady: Ah, right. Didn't notice the ...
00:36 ashleydev joined
arnsholt rakudo: given 5 -> $x, $y { say "alive!" } # Hmm... 00:36
p6eval rakudo 7579f7: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in main program body at line 2:/tmp/pMCr7Cye66␤»
sorear TimToady: std accepts it because the stub makes it in scope. I'm just imagining that, until the real decl, ::Foo is a Failure
pugssvn r31615 | lwall++ | [STDeco] start transitioning to {*} proto stubs 00:39
r31615 | (when bootstrapped and stable, we'll switch from the current {{*}} to {*})
r31615 | implement 'also' declarator, and attach 'also of' trait to proper $*DECLARAND
r31615 | $*OFTYPE is now only used to preserve the 'of' type of a scope declarator until there is a declarand
00:40 aubreyja_ left
TimToady sure, but it's expected that the real Foo must be defined later at compile time, and most code runs at run time 00:40
it's always going to fail to access something at compile time that isn't defined yet...
pugssvn r31616 | lwall++ | [STD] now that reboot is stable, change {{*}} proto stubs to {*} 00:47
sorear TimToady: Am I correct that $STD::ALL is mostly a memory usage hack? 00:52
TimToady no, it's primarily to break the OUTER:: cycles 00:53
sorear that's what I meant 00:54
TimToady :)
but you have to minimally remember all the lexical scopes around YOU_ARE_HERE in the .syml 00:55
and having them all is useful for debugging 00:56
lue afk 00:59
pugssvn r31617 | lwall++ | [STDeco] transit CORE and lib/Test to {*} stubs 01:00
sorear what I want to do is to separate STD.pm6 from the current module system 01:01
because the syml stuff appears to be pretty dependant on p5 semantics
literal: Where in grok is the code which does line wrapping? 01:10
01:13 masonkramer left 01:14 masonkramer joined, masonkramer left, ashleydev left 01:15 skids left
literal sorear: either Pod::Text (Perl 5 Pod) or Perl6::Perldoc (Perl 6 Pod) 01:15
01:18 TiMBuS joined
sorear literal: Where in Perl6::Perldoc? 01:22
literal sorear: actually, I think Perl6::Perldoc doesn't do line-wrapping yet 01:23
sorear so the fact that 'grok' output fits on my screen accurately even after format codes are excised is entirely accidental?
literal most likely 01:24
sorear Hmm. 01:25
sorear wants to add optional dark grey line number annotations to grok output
ingy greetings 01:27
diakopter .
sorear ..
diakopter welcome to our proudly humble abode 01:28
01:29 skids joined, justatheory joined
diakopter .. in theory 01:29
01:31 justatheory left 01:35 whiteknight left 01:38 ashleydev joined 01:40 PenThrive joined 01:47 aubreyja joined 01:50 aubreyja is now known as aubreyja_ 01:51 aubreyja_ is now known as aubreyja
sorear TimToady: would it be reasonable to rename Stash, NAME, etc to STD:: and CursorBase, RE_ast, etc to Cursor::? 01:56
sorear contemplates what STD would look like if it were purged of backtracking (every subrule either returns exactly one cursor or .panics) 02:05
02:19 khb joined 02:21 isBEKaml joined 02:22 aubreyja left 02:23 PZt joined
TimToady Perl 6 requires backtracking 02:24
yes, the internal type could be renamed 02:25
*typs
*e
ingy does rakudo have a perl -I... equiv? 02:39
looks like PERL6LIB works at least 02:42
02:45 masonkramer joined
TimToady I don't think there's a -I 02:57
03:04 aubreyja joined
lue rakudo: class B { has $.a is 0..0xFF }; B.new.a = 300; 03:06
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 20␤»
03:07 aubreyja left
lue can you place limits on attributes? 03:07
.oO(I suppose one could modify the attribute's accessor...)
 03:15
03:21 sftp left
sorear lue: you can spell that 'has uint8 $.a' 03:22
generally speaking you (currently) do that using subset types
power-of-power-of-two-minus-one subsets are predefined 03:23
tylercurtis rakudo: class B { has $.a where 0 <= * <= 255 }; B.new.a = 300;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 20␤»
sorear (at some point I'm going to drag stevan in here and get a real attribute system forward-ported, but that's not likely to be soon) 03:24
tylercurtis rakudo: class B { has $.a where 0 <= * <= 255; }; B.new.a = 300;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 20␤»
tylercurtis rakudo: class B { has $.a where {0 <= $_ <= 255}; }; B.new.a = 300;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 20␤»
lue rakudo: class B { has uint8 $.a }; B.new.a = 300; 03:25
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Malformed has at line 20, near "uint8 $.a "␤»
lue too bad all the sized ints are NYI and their existence is dependant on Parrot (IIRC)
rakudo: class B { has $.a where 0..0xFF }; B.new.a = 300; 03:26
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 20␤»
TimToady rakudo: class B { has $.a of Any where 0..0xFF is rw }; B.new.a = 300
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 20␤» 03:27
TimToady std: class B { has $.a of Any where 0..0xFF is rw }; B.new.a = 300
p6eval std 31617: OUTPUT«===SORRY!===␤Two terms in a row at /tmp/IDUK2e7eBS line 1:␤------> class B { has $.a of Any where 0..0xFF ⏏is rw }; B.new.a = 300␤ expecting any of:␤ bracketed infix␤ constraint␤ infix or meta-infix␤ statement modifier
..loop␤Parse failed␤FAILE…
TimToady std: class B { has $.a where 0..0xFF is rw }; B.new.a = 300
p6eval std 31617: OUTPUT«===SORRY!===␤Two terms in a row at /tmp/QxVDAXfpfl line 1:␤------> class B { has $.a where 0..0xFF ⏏is rw }; B.new.a = 300␤ expecting any of:␤ bracketed infix␤ constraint␤ infix or meta-infix␤ statement modifier loop␤Parse
..failed␤FAILED 00:01…
lue it's not specced, but methinks it should be legal (it's already doable in parameters)
TimToady std: class B { has $.a where 0..0xFF }; B.new.a = 300 03:28
p6eval std 31617: OUTPUT«ok 00:01 111m␤»
tylercurtis std: class B { has $.a is rw where 0..0xFF }; B.new.a = 300
p6eval std 31617: OUTPUT«ok 00:01 111m␤»
sorear lue: approximately seven years ago, TimToady decided what was wrong with the Perl 5 object system and created a new one. 03:29
Two things have happened with this.
Perl 6 has the old version, which has seen nearly no improvement over that time
TimToady rakudo: subset u8 where 0..255; class B { has u8 $.a is rw }; B.new.a = 300; 03:30
sorear The Moose people have backported it to Perl 5 and have been improving it for all this time
p6eval rakudo 7579f7: ( no output )
TimToady some of that is merely to tailor it to the idiosyncracies of p5, and not applicable
I'd like to see a list of what you think isn't in that category
sorear If the current Perl 6 object system looks like a dusty prototype which has been unprioritized for seven years, that's because it has been 03:31
anonymous types
triggers
method modifiers
runtime class creation by API
representation traits 03:32
isBEKaml for instance, one thing I find confusing in the current object system in rakudo is we can instantiate from objects (I might be wrong in phrasing this though)
TimToady I think you're reading some of rakudo's limitations as Perl 6's limitations
anon types are no problem
triggers and method modifiers can be done
isBEKaml rakudo: class A { has $.x is Int}; my $foo = A.new; my $bar = $foo.new; # is this right? 03:33
sorear how do you define an attribute with an anonymous type?
p6eval rakudo 7579f7: ( no output )
TimToady bind it to ::T
tylercurtis rakudo: class Foo { }; my $f = Foo.new; say $f.new.WHAT =:= $f.WHAT
p6eval rakudo 7579f7: OUTPUT«1␤»
lue
.oO(I saw masak do that!)
TimToady roles are full of anonymous types
sorear Our object system has no batteries 03:34
It's all extensibility and no features
TimToady you do like to overstate things, dontchya? :) 03:35
isBEKaml I think saying $foo.new is pretty confusing imho. Objects are created from prototypes(classes) unless you're cloning another object. So .new is misleading here.
03:35 kev joined
lue TimToady: should I be able to use where on attributes in roles/classes ? 03:35
TimToady yes
03:35 kev left
TimToady that's a rakudo nyi 03:35
sorear Actually, my root problem is that I want an 'after' keyword built in 03:37
TimToady also, some of the design philosophy is to let the implementations discover what they need behind .HOW and negotiate something cross-platform.
sorear not necessarily with the same spelling
TimToady you can .wrap any routine
including a method
sorear TimToady: of course. I'm not saying drop everything right now
TimToady and your carping is part of the negotiation process :)
sorear but I don't want lue to think that the object system we have now is final 03:38
TimToady but repr hooks have been envisioned from the start
sorear the perl 6 object system is going to be much more awesome than it is now
despite the fact that some people seem to think it
's perfect as is. imagination, use it
TimToady gee, now you tell me, after ten years... :P 03:39
sorear anyways. 03:40
TimToady the main point of the .HOW barrier is to let the language stabilize on top while we're still experimenting underneath
we encapsulate repr even from the casual user
that's very important, i think
isBEKaml tylercurtis: I wasn't aware of =:=. But I was thinking when doing $foo.new, we were cloning the object. :) 03:41
tylercurtis: my point still stands, it's misleading!
TimToady feel free to play with anything on the other side of .HOW; we want that
sorear I think that the single biggest difference between the Moose and P6 metamodels is that the Moose repr and how are the same object
I'm not sure if this is a good thing
TimToady P6 is completely agnostic on that subject 03:42
so we can try it both ways
ruoso++ was playing with separate RIs for smop
tylercurtis isBEKaml: I agree. On the other hand, sometimes I really like that all the methods of the proto-object are available on any object of the class. 03:43
sorear S12:0791 specs the ability of a single class to interact with multiple, ergo independant, reprs
isBEKaml rakudo: class B { }; my $foo = B.new; say $foo.new =:= $foo;
p6eval rakudo 7579f7: OUTPUT«0␤»
sorear I've also come to the conclusion that the reasonable way to think about protoobjects is as an instance of a nonstandard repr, so every class gets to use at least two of them 03:44
TimToady jnthn++ and I were discussing that earlier today, and I was telling him we could restrict that for efficiency
that's kinda how words work in natural language too 03:45
03:46 xinming joined
TimToady in tagmemics any given syntax structure can have multiple interpretations simultaneously depending on context 03:46
03:46 xinming left
lue rakudo: my $a where 0..0xFF; $a = 999; 03:46
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "my $a wher"␤»
03:47 xinming joined
isBEKaml tylercurtis: if we want all methods of proto-object available, I think we should be doing clone() , not .new - if I'm not wrong.. 03:47
lue rakudo: sub xyzzy($a where 0..0xFF) { say $a }; xyzzy(390)
p6eval rakudo 7579f7: ( no output )
isBEKaml tylercurtis: of course, that would limit the new object to be just using the attribute values that were available in the cloned object. 03:48
tylercurtis isBEKaml: Sometimes, being able to get at "class" methods conveniently from the instance is quite nice. 03:49
isBEKaml tylercurtis: $object.^methods ? 03:50
lue (I'm assuming 'where' is a trait.)
TimToady it's officialy a post_constraint
tylercurtis Of course, you can always do $obj.WHAT.foo(I think this is the correct way).
isBEKaml tylercurtis: huh, why wouldn't $obj.foo() work? Unless I'm mistaken, class methods and object methods are the same... :| 03:53
tylercurtis Right.
isBEKaml: but, I'm saying, that sometimes I like being able to do that.
isBEKaml tylercurtis: following class hierarchy, the compiler would be able to resolve methods present in several parent classes.. 03:54
lue
.oO(not exactly easy to find an exisiting implementation of where in Rakudo)
tylercurtis Another benefit of not having a class/instance method distinction is that for classes where the users of that class will often not care about specific instances(e.g., a Compiler implemented as a class internally to store various of information in attributes), you can more easily allow them to be called either as Foo::Compiler.compile($source) or Foo::Compiler.new.compile($source), just by sticking "self.new.compile($source) unless self.defined" at the 03:56
beginning of the compile method.
isBEKaml lue Grammar.pm I think (not sure)
03:56 juster joined 04:00 PZt left
isBEKaml tylercurtis: that took me sometime to understand. :) 04:00
tylercurtis: you're saying that we just abstract away the instantiations within method implementations? Sure, we can do that. But not doing that would lead to GC problems, I think. I'd rather avoid having to take care of such things. A clear demarcation between objects and classes would be preferable when it comes to instantiating new objects. 04:04
tylercurtis isBEKaml: I agree that being able to do SomeClass.new.new.new.new.new.new.WHAT.new.new.WHAT.WHAT.WHAT.new.new is confusing, but on the other hand, the class/instance method non-distinction can clarify code sometimes. 04:05
04:06 justatheory joined
lue my $a = 3; say $a.WHAT.WHAT 04:07
rakudo: my $a = 3; say $a.WHAT.WHAT
p6eval rakudo 7579f7: OUTPUT«Int()␤»
isBEKaml tylercurtis: I think it was always there. instance objects take to class object methods easily as they don't have their own independent method definitions, unless augmented with MONKEY_TYPING.
tylercurtis isBEKaml: For example, in my say-call-without-parentheses-on-arguments-that-are-positive-integers-with-singly-assigned-lexicals(although "lexical" don't really mean much when you don't have any way to create a new scope :) to LLVM compiler; the SIC::Compiler class keeps some internal variables for things like the numeric suffix for temporary variable names in the generated code. So, you need to create a new SIC::Compiler object to compile any code with 04:08
However, most uses of the SIC::Compiler class are going to consist of "SIC::Compiler.new.compile($the-source-code)". So, at the beginning of the compile method, I use self.defined to check if self is undefined(and thus, in my class's case, at least, is the proto-object), and return the result of calling the method with the same arguments on a newly-created SIC::Compiler object just as if you had done "SIC::Compiler.new.compile($the-source-code)", even
though you just did "SIC::Compiler.compile($the-source-code)".
I could still do that if Perl 6 had separate class and instance methods, but it would be less convenient. I'd have to write a separate class method instead of just sticking the one "return self.new.compile($source) unless self.defined;" line at the beginning. 04:10
isBEKaml tylercurtis: I agree with you there. self.method-name() is not too different from doing this.method-name() in other languages - apart from the fact that the method call happens only within the object's own blueprint of the class' method. 04:15
tylercurtis++ # I now see what he was driving at.. 04:22
tylercurtis I'm not really decided on whether I agree with not having the distinction yet, but I'm definitely finding myself doing stuff like that far more often in Perl 6(and even in NQP) where the method works on either the proto-object or the instance than in other languages. 04:25
And that sentence of mine was badly ordered. put in Perl(and even in NQP) just before "than in other languages" in your mind if it sounds incoherent. 04:26
isBEKaml tylercurtis: self.new has no (AFAIK) equivalent in other languages and offers an advantage here in P6. However, we do object.getClass().newInstance() in Java. 04:27
tylercurtis: so we always get the class method first. 04:28
tylercurtis isBEKaml: right. Other languages can do it. Although, it's not always obvious how or if a language can do some of the things that the unity of "class" and "instance" methods in Perl 6 allows.
In Smalltalk, I hadn't even realized that class methods inherited until just now when I checked, but with Perl 6, it's obvious that "class" methods are inheritable. 04:29
isBEKaml tylercurtis: Java has the concept of Reflection that is not quite obvious/intuitive. Some programmers use it (mostly library developers), most would just avoid it. You have to jump through to get at what you want here. 04:30
tylercurtis: the non-distinguishable nature of class/object methods are somewhat problematic when done incorrectly. (I had referred to GC issues before). I think Perl should just the Right Thing when it sees one. 04:32
tylercurtis: duh.. sorry if I'm being somewhat incoherent. Please let me know.. 04:33
tylercurtis In one of my projects in NQP(it's my GSoC project, in fact), I use a helper class to implement the Tree::Pattern.transform method, which descends through a tree and transforms, using a supplied sub, any parts of the tree that match the pattern. However, I want to be able to create subclasses of both the pattern class and the helper class with additional functionality(for example, to traverse tree subtypes).
So, I did the obvious thing and created a method on the parent class that returns the appropriate helper class. 04:34
Conveniently, I can also call the method on the class to find out what its helper class is. 04:36
It logically feels like a property(not in the "attribute"/"class/instance variable" meaning of the word, but in the "something that describes something") of the class, but it's more convenient to call on the instance in most cases. Best of both worlds. :) 04:38
04:40 justatheory left
isBEKaml Right. :) I think we should restrict the use of class methods to `self` alone and using it on objects are somewhat misleading. 04:43
doing self.new.method looks cleaner than $obj.new.method.
But I guess that would lead to problems in implementation. How do we do $obj.WHAT? Related, we are doing this reflectively... I have no answer.. :) 04:44
tylercurtis: where can I find out about your GSoC project? URL? 04:47
tylercurtis: I'm looking at this. trac.parrot.org/parrot/browser/bran...er/library 04:50
04:58 rv2733 joined
tylercurtis isBEKaml: that's it. I moved it to github.com/ekiru/tree-optimization this weekend so that it's easier for people to try. 05:03
It's also available via plumage("./plumage install tree-optimization"). 05:04
05:04 ashleydev_ joined 05:06 ashleydev left, ashleydev_ is now known as ashleydev
tylercurtis isBEKaml: certain parts of it aren't fully documented(in part because I've got some significant refactors planned for some of those parts and didn't want to have to write and then soon after rewrite it :). 05:10
isBEKaml tylercurtis: ty. But that's alright. I just wanted to look at what you were doing, not necessarily understand it. ;) 05:12
tylercurtis More things about it to look at are available at www.parrot.org/blog/839 , where I've been posting blog posts weekly-ish that have varied from tutorials to descriptions of APIs I was planning to implement to simple descriptions of what I've done for the week. 05:17
isBEKaml tylercurtis: great! that's nice. :) 05:21
05:22 PZt joined
tylercurtis I'll probably be posting another sometime tomorrow, too. :) It's always nice when people sound like they're interested in one's project. :) 05:22
isBEKaml I don't know how that feels. I'm yet to get my own project up. :) 05:23
I'm currently trying to understand parrot and rakudo. So my project won't be up anytime soon. ;) 05:25
tylercurtis: can you tell me what other optimisations are you targeting apart from tailcall recursions? 05:31
05:31 masonkramer left
isBEKaml tylercurtis: I believe there might be more since your first post about PAST Optimizations. :) 05:32
sorear apparently, now that ibus + anthy are working, my compose key /isn't/ 05:34
05:35 ashleydev left 05:37 redicaps joined
tylercurtis isBEKaml: Well, I've mostly been working on the infrastructure, so far, but I've got a working optimization for some tail-calls with an explicit return in NQP-rx(extending it further is blocking on better understanding of PAST::Op(:pasttype<call>) and :pirop<tailcall> nodes), and a prototype of very basic constant-folding of Integer addition for NQP-rx. I've also done quite a bit of constant-folding for PIRATE(which is a PIR compiler written in NQP wit 05:37
PCT).
05:39 hercynium left
tylercurtis Most of the rest of my GSoC is scheduled for implementing specific optimizations, and I have a list somewhere of some possible ones to work on, so expect to see more coming soon. Another thing I might attempt to do is some sort of transformation from PAST to some kind of SSA form to make optimizations simpler. 05:39
05:43 rv2733 left 05:44 Mowah joined
tylercurtis sleep 28800 & 05:49
05:49 tylercurtis left 06:45 cono left 06:48 redicaps left 06:55 cono joined 06:59 Mowah left
sorear perl6: say class Foo { ... } 07:05
p6eval pugs: OUTPUT«*** No such subroutine: "&class"␤ at /tmp/KFCnPQxcsT line 1, column 5 - line 2, column 0␤»
..rakudo 7579f7: OUTPUT«␤»
sorear rakudo: say (class Foo { ... }).WHAT 07:06
p6eval rakudo 7579f7: OUTPUT«error:imcc:syntax error, unexpected '\n'␤ in file 'EVAL_1' line 76␤===SORRY!===␤syntax error ... somewhere␤»
sorear what.
07:06 xinming left
isBEKaml rakudo: say { class Foo {} }.WHAT; 07:07
p6eval rakudo 7579f7: OUTPUT«Block()␤»
isBEKaml rakudo: say { (class Foo {}).WHAT } 07:08
p6eval rakudo 7579f7: OUTPUT«_block64␤»
isBEKaml class Foo {} doesn't seem to be treated as an expression in a saycall.
rakudo: say ({ class Foo {} }).WHAT; 07:09
p6eval rakudo 7579f7: OUTPUT«Block()␤» 07:10
isBEKaml rakudo: say (class Foo{}.new.WHAT) 07:18
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Illegal redeclaration of symbol 'Foo'␤»
isBEKaml rakudo: say ((class Quux{}).WHAT) 07:25
p6eval rakudo 7579f7: OUTPUT«Code()␤»
07:27 mulander left 07:39 tedv left
moritz_ good morning 08:03
isBEKaml moritz_: good morning! at Hackathon? :) 08:05
moritz_ nope
sorear good morning moritz_ 08:06
(is the _ part of your name? You seem to use it unusually often here)
moritz_ sorear: the 'moritz' nick is registered on freenode, so I settled with moritz_ here permanently 08:07
and registered it
08:16 meppl joined
moritz_ speaking of freenode... the FAQ says that if somebody hasn't logged in for three month, it's possible to request that their nick is being freed 08:19
can I find out what was the last login date of a given nick?
dalek ecza: b125222 | sorear++ | (3 files):
Implement a system of auxilliary stacks
ecza: 7943914 | sorear++ | (4 files):
More steps towards class decls. Allow lexget et al to directly access
ecza: 13b25f8 | sorear++ | (7 files):
Add class statements
08:20 xinming joined
TiMBuS moritz_, /ns info moritz 08:22
Last seen : Mar 23 20:57:22 2010 (15 weeks, 4 days, 11:24:19 ago)
moritz_ TiMBuS: thanks
TiMBuS its an 8 year old nick tho, so
yeah you might be stuck with the _ 08:23
sorear 03:24 [freenode] -!- moritz [~moritz@p579F9664.dip.t-dialin.net] 08:25
03:24 [freenode] -!- ircname : Moritz Lenz
03:24 [freenode] -!- moritz_ [moritz@faui2k3.org]
03:24 [freenode] -!- ircname : Moritz Lenz
there are 2 active freenode connections here
moritz_ sorear: yes, just tired it...
*tried
sorear: the nick without the _ is registered by somebody else
sorear according to the freenode FAQ, a nick is considered expired after 60 days without use
moritz qualifies almost twice over 08:26
what does also :foo<bar> mean 08:37
moritz_ doesn't quite understand 'also' 08:38
sorear also is new, less confusing syntax for something old 08:39
class Foo { also is Bar; } # new 08:40
class Foo { is Bar; } # old
class Foo is Bar { } # old and new, equivalent
moritz_ I understood that much, but I don't see how it works, in terms of parsing or underlying semantics
sorear it just lets you put traits inside the class body 08:41
if you don't want to put them in the header 08:42
no semantic or parsing changes required
moritz_ sounds like something you could achieve with a module too 08:43
sorear with a module too?
moritz_ I mean, it sounds like it doesn't need to go into core Perl 6, IMHO 08:44
sorear nope 08:45
98% of the stuff in core Perl 6 doesn't need to be there, and this is no exception 08:46
08:48 eternaleye left 08:57 Ross joined 08:58 Ross left 08:59 Ross joined, Ross left, Ross joined, Ross left 09:07 Ross joined 09:08 Ross left 09:10 tadzik joined 09:16 pmurias joined
pmurias moritz_: ping 09:16
moritz_ pmurias: pong
pmurias moritz_: still blocking on CHI? 09:17
moritz_ pmurias: yes
pmurias comment the dependency out ;)
dalek ecza: 53b6cdf | sorear++ | (2 files):
Implement 'also'
ecza: 6e9becd | sorear++ | (4 files):
Implement superclasses for also
ecza: 676bde0 | sorear++ | (2 files):
Implement in-header trait syntax
ecza: 8da9e3d | sorear++ | (2 files):
With no explicit superclasses, automatically add an Any superclass
ecza: 32d2ad5 | sorear++ | setting:
A bunch of stub classes
sorear niecza just jumped from 5 classes to ~50
sorear since I've got class syntax working... took long enough. 09:18
moritz_ sorear: should I investigate setting niecza up on p6eval?
sorear next up: methods. then, control structures 09:19
moritz_: niecza is going to become much easier to set up very soon
pmurias notices perlito has a python backend
sorear once I can use pmurias++'s STD dist instead of hardcoding a path to my pugs checkout 09:20
moritz_ sorear: then I'll wait... just tell me when you think it's time to add it
sorear (yes, I have hardcoded paths in a public git repo. I feel just a bit dirty for it.)
moritz_ pmurias: perlito seems to have quite a few backends :-)
sorear out. 09:21
pmurias sorear: what do you need to use my STD dist?
it use it for mildew currently
s/it/i/ 09:22
sorear pmurias: it being announced and on CPAN, ideally 09:23
09:24 Ross joined
sorear out for real. 09:24
09:28 agentzh joined
pmurias moritz_: what's the best way to have an optional dependency? 09:28
moritz_ pmurias: the META.yaml file has a "recommends" field, iirc 09:29
09:29 azert0x joined, azert0x left 09:30 azert0x joined, azert0x left
moritz_ oh, that's deprecated :/ 09:30
09:30 azert0x joined
pmurias an optional dependency will show up as a prompt to the user? "do you want to install 1 optional dependency?" 09:30
those are annoying
moritz_ search.cpan.org/~dagolden/CPAN-Meta...pm#prereqs
runtime => { recommends => { CHI => '0.35' } } 09:31
no idea how installer handle it
pugssvn r31618 | pmurias++ | [mildew] temporary disable the CHI dependency for moritz++ 09:33
pmurias moritz_: try installing now
moritz_: i think i'll just move Mildew::Frontend::STD::Cached into a seperate package 09:34
moritz_ tries
09:34 tadzik left
moritz_ Successfully installed Mildew-0.01 09:35
what's next? SMOP?
pmurias yes 09:36
moritz_ then.. Mildew-Setting-SMOP?
pmurias yes
moritz_: smop installed? :) 09:37
moritz_ yes
pmurias good
moritz_ Mildew-Setting-SMOP too
so, how do I run it? :-)
pmurias mildew -e 'say 1'
it's installed into a non-default local::lib?
moritz_ ... and now I have to fix up @INC.
09:38 tadzik joined
tadzik wklej.org/id/363056/ can someone show me the way? (line #37) 09:39
moritz_ tadzik: try: method clear { $!z = 0; nextsame; } 09:40
pmurias moritz_: eval $(perl -I$HOME/perl5/lib/perl5 -Mlocal::lib)? 09:41
moritz_ pmurias: p6eval@debian:~$ ./mildew-install/bin/mildew
Can't locate MildewCORE.lex in @INC (@INC contains: lib mildew-install/lib/perl5//x86_64-linux-gnu-thread-multi mildew-install/lib/perl5/ /etc/perl /usr/local/lib/perl/5.10.0 /usr/local/share/perl/5.10.0 /usr/lib/perl5 /usr/share/perl5 /usr/lib/perl/5.10 /usr/share/perl/5.10 /usr/local/lib/site_perl .) at mildew-install/lib/perl5//CursorBase.pm line 320.
find mildew-install/ -name MildewCORE.lex # no result
pmurias you installed Mildew-Setting-SMOP?
moritz_ yes
pmurias that's bad
pmurias tried removing his local lib and installing everything on his box 09:42
moritz_ I can try again
pmurias * tries
moritz_: wait with that
moritz_ Successfully installed Mildew-Setting-SMOP-0.01
pmurias nopaste the ~/.cpanm/build.log? 09:43
(hint App::Nopaste is really usefull) 09:44
moritz_ pmurias: nopaste.snit.ch/21931 09:45
tadzik: nopaste.snit.ch/21932 09:47
09:48 Trashlord left
pmurias hmm 09:48
moritz_: i found a different bug in Mildew-Setting-SMOP 09:49
09:49 Trashlord joined
pmurias moritz_: can you find Mildew and SMOP filed in mildew-install? 09:49
moritz_ mildew-install/lib/perl5/Mildew/Setting/SMOP.pm 09:50
mildew-install/lib/perl5/SMOP.pm
when looking for SMOP.pm 09:51
pmurias moritz_: STD.pmc?
moritz_ mildew-install/lib/perl5/Mildew.pm
mildew-install/lib/perl5/STD.pmc
pmurias moritz_: what i suspect is the Build process die'ing because of badly set @INC 09:55
dzil build;cd Mildew-Setting-SMOP;perl Build.pl;perl Build 09:56
should return more info
cd Mildew-Setting-SMOP-0.1
* 0.01
pmurias fixes an other bug in the build process... 09:57
moritz_ rakudo: die 2 09:59
p6eval rakudo 7579f7: OUTPUT«2␤ in main program body at line 20:/tmp/iWSoE1QX01␤»
tadzik moritz_: so before would be { nextsame; $.z = 0 } ?
moritz_ tadzik: actually it's the other way round 10:01
tadzik or no, that'd be after
yep
then around is just a method with nextsame() in the middle. Makes sense, even better than in Moose
moritz_ nextsame just calls the next available method with the same name, and the same parameters as the current method 10:02
tadzik so it looks up the parent class, the roles etc?
moritz_ and multi methods from the same class, if applicable 10:03
tadzik mhm
moritz_ roles are flattened into classes at composition time, so they don't need to be considered specifically
tadzik I see. Thanks, moritz_++ 10:04
moritz_ you're welcome
you can also specifically call a method from another class
so self.Point::clear()
arnsholt Right, my given/when for NQP is pretty much done 10:14
I got commit access to the repo at some point, so does pushing the branch to Github sound like a resonable way to submit it for review?
10:24 isBEKaml left
moritz_ yes 10:26
arnsholt Excellent 10:27
And done 10:30
moritz_ arnsholt++
10:44 timbunce joined
TiMBuS if i add two roles and they have a conflicting method, is there a way to disambiguate? or does the flattening make this not possible 10:48
Method 'dothing' collides and a resolution must be provided by the class 10:52
oh ok
that warning isnt thrown when using 'but'
or, i guess, any kind of runtime composition 10:53
10:58 whiteknight joined
pugssvn r31619 | colomon++ | [t/spec] Update fudging. 11:07
11:07 colomon joined 11:09 tadzik left, agentzh left
pugssvn r31620 | pmurias++ | [mildew] fix a bug in the build process 11:25
pmurias moritz_: maybe that will help with the installation of mildew (install mildew and mildew-setting-smop again)
pugssvn r31621 | pmurias++ | [mildew] add basic tests that check if Mildew-Setting::SMOP builds correctly 11:37
11:40 colomon left
moritz_ pmurias: # Failed test at t/01-tmp-is-created.t line 2 11:52
all 6 subtests fail
11:53 Guest38594 joined, Guest38594 is now known as pragma_
pmurias moritz_: dzil build 11:53
moritz_: in Mildew-Setting-SMOP
then perl Build.pl
then perl Build
and perl Build test 11:54
that will allow us to see when do the errors happen
moritz_ doesn't reveal more details 11:55
building works
the test doesn't
no error messages except for "Failed test..." 11:56
szabgab I understand proto is now out and one should use pls ?
where is pls ?
pmurias moritz_: is tmp created
?
moritz_ szabgab: pls is only a proof-of-concept so far 11:57
pmurias moritz_: run the test manually
moritz_ szabgab: so continue running proto for now
no tmp/ created
that's why the test fails :-)
szabgab: and FYI, 'pls' is a branch in the proto repo 11:58
pmurias moritz_: perl -MSTD works?
szabgab ty
pmurias moritz_: make sure you don't have the STD from src/perl6 in your @INC 11:59
moritz_ ugh, I had a relative path in PERL5LIB
which didn't work out for installing Mildew
pmurias: now perl -MSTD -e 1 works, but no tmp created 12:00
pmurias you ran perl Build again
moritz_ yes
pugssvn r31622 | pmurias++ | [mildew] remove a file which crept up into the repo 12:03
pmurias Mildew-Setting-SMOP-0.01.tar.gz was creted by dzil build?
* created
moritz_ pmurias: yes, that's where I run perl Build.PL
pmurias an old one crept into the repo so that's why i ask 12:04
moritz_ btw is the 'diff' file there on purpose?
moritz_ tries again
still no tmp 12:05
pugssvn r31623 | pmurias++ | [mildew] remove trash found by mortiz++ 12:06
pmurias moritz_: check if ACTION_code is called by inc/MyBuilder.pm
inserting a warn should do
pmurias updates Module::Build 12:07
moritz_ pmurias: it is not 12:08
moritz_ has Module::Build version 0.3603 12:09
pmurias has 0.3607 12:11
moritz_: try perl Build code? 12:12
moritz_ seems like a noop 12:13
szabgab moritz_: looking at the book I see some of the chapeters in src/ that are not included in the Makefile , is that on purpose?
pmurias moritz_: that's worring 12:14
moritz_: could you nopaste your Build.PL
?
moritz_ szabgab: afaict only builtins.pod shouuld not be included
pmurias: nopaste.snit.ch/21934 12:15
szabgab: are others missing too?
szabgab and why not include it with a big todo?
pmurias moritz_: that's bad
szabgab checking
pmurias moritz_: that's seems to be the problem
moritz_ szabgab: we had some ongoing discussions whether it should be a proper chapter or an appendix, and then I forgot 12:16
szabgab: feel free to change though
12:16 clintongormley joined
pmurias moritz_: could you dzil build in smop and nopaste the Build.PL from there? 12:16
szabgab roles and subtypes are also missing from Makefile
moritz_ szabgab: that's bad... fixing...
szabgab I can fix it 12:17
I was just asking
moritz_ szabgab: even better :-)
szabgab as trying to get involved
but I though I should first read what is there already :)
what about creating a perl script insted of the make file ?
moritz_ what would be the benefit? 12:18
szabgab checking for files not included in makefile
pmurias moritz_: maybe your dist::zilla version is different
szabgab or the files could be renamed to 01_prefix.pod etc 12:19
and build the list of files on the fly
moritz_ nopaste.snit.ch/21935
szabgab: if you want to do that, please check with lucs on #perl6book first, he might have some planned changes
szabgab: and I don't like the number prefix, because we might still add chapters 12:20
12:20 masak joined
masak oh hai, #perl6! 12:20
moritz_ pmurias: Dist::Zilla version 4.101900 12:21
masak: oh hai. I've added some crude line wrapping code into XML::Writer
masak ok. 12:22
pmurias moritz_: nopaste your dist.ini
masak last night after logging off, instead of sleeping, I wrote that new Yapsi runtime. :)
it was exactly the fun I had pictured it to be.
pmurias moritz_: from Mildew-Setting-SMOP
moritz_ pmurias: isn't that under version control?
pmurias yes 12:23
moritz_ well, I didn't change it :-)
Last Changed Rev: 31607
pmurias moritz_: you have an incorrect Build.PL
moritz_: you have App::nopaste installed? 12:25
* App::Nopaste
moritz_ no
but that's easyenough to change :-)
pmurias then you could nopaste stuff with "nopaste file" 12:27
moritz_: check the version NAME ^ 12:29
of Dist::Zilla::Plugin::ModuleBuild
moritz_: it seems our Dist:Zilla's behave differently
12:30 timbunce left
moritz_ 2.101241 12:31
pmurias what? so your Dist::Zilla is made of different versions?
perl -MDist::Zilla::Plugin::ModuleBuild -e 'print $Dist::Zilla::Plugin::ModuleBuild::VERSION' 12:33
moritz_ huh
that's 4.101900
I got my version from looking into the module
package Dist::Zilla::Plugin::ModuleBuild; 12:34
BEGIN { $Dist::Zilla::Plugin::ModuleBuild::VERSION = '2.101241';
}
seems to get overridden later on
moritz_ shudders
masak jnthn: ping
12:35 f00li5h joined 12:36 f00li5h left, f00li5h joined, f00li5h left, f00li5h joined 12:37 f00li5h left, f00li5h joined
masak f00li5h: stop that. :) 12:37
pmurias moritz_: that's a leftover file from an old Dist::Zilla 12:38
moritz_: www.yourcargo.org/Mildew-Setting-SM....01.tar.gz 12:40
12:40 lichtkind joined
lichtkind cheers 12:41
masak \o
pmurias moritz_: you could try using that tarball if you Dist::Zilla is a strange hybrid
12:42 s1n joined
moritz_ which depends on yet another new module. *sigh* 12:45
Can't locate Config/AutoConf.pm in @INC
12:46 PenThrive left
f00li5h masak: turns out that i forgot about /connect -! ... 12:46
masak f00li5h: no worries. :) 12:47
gist.github.com/471521 # yay! new runtime!
12:47 takadonet joined
takadonet morning all 12:48
masak takadonet: \o
takadonet masak: how are u?
masak takadonet: it's way too warm here in my part of the world, but I'm inside, and coding Perl 6, so I guess I'm fine. :)
takadonet: and u? 12:49
takadonet masak: at ISMB right now waiting for the keynote speaker to start
masak nice.
takadonet masak: you do bioinformatics correct?
masak I try to. :) 12:50
takadonet :)
12:51 Ross left
takadonet masak: www.slideshare.net/bosc2010/fields-...10-bioperl 12:52
moritz_ pmurias: with your tarball I get "Can't stat tmp: No such file or directory at /opt/perl-5.12.1/lib/5.12.1/Module/Build/Base.pm line 4858
masak looks
moritz_ "
which is non-fatal
takadonet people were happy to see some bioperl6 code :)
moritz_ but later it dies with Can't locate object method "compile" via package "Mildew::Backend::C::So" at /home/p6eval/mildew-install/lib/perl5//Mildew/Compiler.pm line 21
(while running ./Build)
masak takadonet: yoy, Perl 6 is mentioned, at least :) 12:53
takadonet masak: with some of my code! hehe
12:54 sftp joined
takadonet masak: with help from you and moritz_ 12:54
moritz_ bows
takadonet bbl presentation is starting
12:55 takadonet left
pmurias moritz_: is Mildew::Backend::C::So installed? 12:56
12:56 agentzh joined
masak takkadonet: very nice. takadonet++ 12:56
pmurias moritz_: the non fatal warning is expected 12:57
moritz_: who own's the evalbot box?
* owns 12:59
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say $h_ref<a>; 13:00
p6eval rakudo 7579f7: OUTPUT«Null PMC access in type()␤ in main program body at line 20:/tmp/y9fWXYyJS1␤»
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say $h_ref.perl;
p6eval rakudo 7579f7: OUTPUT«\("b" => 2, "a" => 1)␤»
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say ${h_ref}<a>; 13:04
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "say ${h_re"␤»
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say $$h_ref<a>;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "say $$h_re"␤»
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say %h_ref<a>;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Symbol '%h_ref' not predeclared in <anonymous> (/tmp/HBwUba7bcT:20)␤»
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say %{$h_ref}<a>;
p6eval rakudo 7579f7: OUTPUT«␤»
moritz_ pmurias: diakopter owns it 13:05
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say %$h_ref<a>; 13:06
p6eval rakudo 7579f7: OUTPUT«␤Null PMC access in find_method('HOW')␤ in main program body at line 1:/tmp/MEHuxodZza␤»
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; my %g = Hash($h_ref); say %g<a>; 13:09
p6eval rakudo 7579f7: OUTPUT«Could not find sub &Hash␤ in main program body at line 20:/tmp/uq2GdO8qo_␤»
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; my %g = %$h_ref; say %g<a>;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "my %g = %$"␤»
cosimo mm
pmurias diakopter: ping 13:12
moritz_: try perl -MMildew::Backend::C::So -e 1 13:14
cosimo rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; my %g = $h_ref; say %g<a>;
p6eval rakudo 7579f7: OUTPUT«1␤»
cosimo shiny
but isn't there a direct way? 13:15
i'd have expected $h_ref.<a> to work
rakudo: my %h=('a'=>1,'b'=>2); my $h_ref=\%h; say $h_ref.{'a'}
p6eval rakudo 7579f7: OUTPUT«Null PMC access in type()␤ in main program body at line 20:/tmp/NRIhf8BtHO␤»
cosimo mmh
13:18 tadzik joined
masak submits rakudobug 13:23
by the way, why isn't .invert in Cool? I wanted to use it on a List last night, and it turns out it's only defined for Hash. 13:24
moritz_ pmurias: fails, with 80 lines of "Couldn't load class" error messages. And App::Nopaste refuses to install 13:25
afk
masak rakudo: say (\my %h)<a> 13:28
p6eval rakudo 7579f7: OUTPUT«Null PMC access in type()␤ in main program body at line 20:/tmp/7teqOLAaKD␤»
cosimo rakudo: my $x = "someThing"; say $x ~~ m:i/something/;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "say $x ~~ "␤»
cosimo rakudo: my $x = "someThing"; say $x ~~ m/something/;
p6eval rakudo 7579f7: OUTPUT«␤»
cosimo rakudo: my $x = "someThing"; say $x ~~ m/someThing/;
p6eval rakudo 7579f7: OUTPUT«someThing␤»
cosimo are regex modifiers broken? 13:29
masak rakudo: my $x = "someThing"; say $x ~~ m/:i something/;
p6eval rakudo 7579f7: OUTPUT«someThing␤»
masak cosimo: NYI on the outside, that's all.
cosimo masak: NIY you mean?
masak I hear both abbreviations.
NYI is the more common, though. 13:30
cosimo ok :)
masak they mean the same.
cosimo but the "Correct(tm)" form is m:i// then?
masak yes, I'd say so. 13:31
IIRC, the tricky part with implementing it was to change the regex after it had been compiled.
since ':i' on the inside is how it "actually works", at least in PGE and nqp-rx. 13:32
13:32 whiteknight left
moritz_ there is no "change the regex after it had been compiled" 13:38
the :i must be passed on the regex compiler 13:39
currently the tricky thing with implementing adverbs in rakudo is not to make anything unrelated parsefail
masak ah. 13:42
moritz_ I tried adding a simple ':g'? after the 's' parsing for s///
and it would then parsefail identifies starting with s. 13:43
"huh".
jnthn masak: pong
cosimo rakudo: my @x = ('5c', 'Blah', '10', 'Blah'); say @x.shift.Int; 13:44
p6eval rakudo 7579f7: OUTPUT«5␤» 13:45
cosimo cool
how do I convert, say '5c' to decimal?
rakudo: say '5c'.hex;
p6eval rakudo 7579f7: OUTPUT«Method 'hex' not found for invocant of class 'Str'␤ in main program body at line 20:/tmp/hruYr0rHDd␤» 13:46
moritz_ rakudo: say :16('5C')
p6eval rakudo 7579f7: OUTPUT«92␤»
masak jnthn: remember we talked about SIC and explicitly putting in an 'outer' pointer from an inner block to its outer block?
jnthn Yes 13:47
masak jnthn: I've forgotten your reson for why that was necessary; and now I have things working without it... :) 13:48
jnthn Well, pointer sounds runtime-y
masak so there's probably something I'm missing.
yes, the pointer became entirely runtime-y.
jnthn OK, and how do you know from the IL which block is nested inside which?
cosimo rakudo: my $hex = '5c'; say :16($hex);
masak but PIR has a static :outer() thingy.
p6eval rakudo 7579f7: OUTPUT«92␤»
cosimo rakudo: my $hex = '5c2f'; say :16($hex); 13:49
masak jnthn: "IL"?
p6eval rakudo 7579f7: OUTPUT«23599␤»
jnthn SIC?
cosimo moritz_: masak: thanks
masak SIC currently doesn't know, doesn't care.
jnthn Um
masak the compiler cares, and the runtime cares.
but not SIC.
jnthn So how do you know how to set up the links between blocks?
That information must be *somewhere*?
masak well, the outer block takes a closure to the inner block at runtime. 13:50
at that point, an 'outer' pointer is made.
this will work for subs as well as simpler things like immediate blocks. 13:51
jnthn: here's an example of it working: gist.github.com/471521 13:52
jnthn Right, so that's where you have said info in the bytecode. It builds the links like that.
pmurias shouldn't we have a nopaste mentioned in the topic?
masak jnthn: yes. so now I'm left wondering what PIR uses its :outer() attribute for :) 13:53
jnthn: since the blocks in SIC are about as static as static lexpads, and don't have much relation to the lexpads at runtime. 13:54
jnthn masak: e
You aren't assuming that caller = outer here are you? 13:55
masak gosh, I hope not. :)
13:55 rv2733 joined
masak for a while yesterday, I think I was. 13:55
but now I think I'm not, since it's all about when the closure is taken. it's taken at block entrance.
after that, it might be called at any time, and will still have the right 'outer'. 13:56
jnthn rakudo: my $x = 42; sub foo { our sub bar { say $x } }; { my $x = 69; bar() }
p6eval rakudo 7579f7: OUTPUT«42␤»
jnthn masak: Will you get that right under your model?
Here the foo block was never entered.
masak hm.
that'll take the closure at CHECK time or something like that, I assume. 13:57
rakudo: my $x = 42; sub foo { $x = 5; our sub bar { say $x } }; bar; foo; bar; 13:58
p6eval rakudo 7579f7: OUTPUT«42␤5␤»
masak and it'll re-rake the closure if foo is ever entered.
jnthn I guess it just feels odd to me to set up the static chain dynamically, that's all.
masak yes, I think I have the model right in my head. :)
willing to be proven wrong, of course.
jnthn: well, it just feels wrong to put in something that I never end up using... 13:59
rakudo: sub foo { my $y; our sub bar { say $y } }; bar
p6eval rakudo 7579f7: ( no output )
masak this is problematic in my model. :/
and seemingly in Rakudo's, too :) 14:00
locally, Null PMC access.
masak submits rakudobug
jnthn Ugh
Yeah, that's why we need the proto-lexpads.
masak yep. 14:01
let's call them 'static lexpads', like pmichaud has started doing.
I'm going to have to think about that case some more.
but probably when the closure is taken at CHECK time, it'll use the static lexpad as its .outer 14:02
and then it works.
14:02 slavik joined
jnthn Anyway, my answer in summary is, with the model you've adopted you don't need to say in the code 'my outer is X'. However, I know which model I prefer. :-) 14:02
masak noted. :) 14:03
well, it just seemed to me there was a category in there somehow, at least in the way PIR does it. the :outer() thingies refer to other subs, i.e. static things in the code. whereas I needed to refer to outer lexpads, live things in the runtime. 14:04
moritz_ as soon as you start to optimize, you will be glad for every piece of information
masak the same block of code may have zero, one or several lexpads "associated" with it. 14:05
slavik so, how much openmp type stuff will there be in Perl6? ;)
masak so pointing back to the parent *block* doesn't really tell me much anyway, it seems.
s/parent/outer/
moritz_ slavik: as much as people implement
slavik :( 14:06
masak slavik: sorry, expecting a free lunch? :P
slavik: is there funding money involved?
slavik masak: I wish ;)
moritz_ slavik: well, it's true for any other feature too
slavik honestly though, I wish there was a finalized spec
hint hint wink wink 14:07
*cough* TimToady *cough*
moritz_ considering how many quirks we found in the exisiting spec so far, finalizing it before it was implemented seems like a direct way to failure
slavik the moment there is, I think I will start to use Perl6 more seriously.
masak slavik: I used to wish for that, too. now I wish the spec freezes gracefully.
slavik moritz_: I am not accounting or implementation in that statement. 14:08
masak slavik: the fact that you aren't... :)
slavik masak: I think you put it in better terms :)
masak that's why the spec *is* changing nowadays is mainly because of the implementations!
moritz_ slavik: you can't separate spec from implementation to the extend you like
slavik moritz_: not talking about separating them. 14:09
masak we know what we want from Perl 6, but some ideas need tweaking when the rubber meets the road.
moritz_ so what do we want from Perl 6? :-)
slavik moritz_: it's in the sense of learning something and not having it change the next day ;)
moritz_: cake
and flying pigs
moritz_ had cake today 14:10
slavik >.>
moritz_ strawberry and cream
slavik yea? well, one of my friends is a baker! take that!
moritz_ there's still some in the kitchen :-)
moritz_ read "bake that" :-)
slavik haha
pmurias moritz_: diakopter is the one who could give me a acoount on the evalbot box? 14:11
moritz_ pmurias: yes. Normally I could too, but I forgot my password :/
masak moritz_: I want a simple, reliable model that I can't get into my head, and that my mind doesn't reject because it's inelegant. lately, with things coming together in Rakudo with nesting/OO/different kinds of variables, I've been getting a tast of that. it's wonderful. 14:12
s/can't/can/
14:13 kid51 joined
pmurias moritz_: why does App::Nopaste refuse to install? 14:13
moritz_ wonders if Perl 6 fits into a model that fits into your head :-)
or mine
pmurias: a dependency fails, HTTP::Server::Simple or so
even though I don't want to run a server, just a client
cosimo rakudo: my @x = 1, 2, 3; @x[2]:delete; say @x; 14:14
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "@x[2]:dele"␤»
cosimo rakudo: my @x = 1, 2, 3; @x[2].delete; say @x;
p6eval rakudo 7579f7: OUTPUT«Method 'delete' not found for invocant of class 'Int'␤ in main program body at line 20:/tmp/GSSwm0mafY␤»
moritz_ operator adverbs NYI
rakudo: my @x = 1, 2, 3; @x.delete(2); say @x.perl
p6eval rakudo 7579f7: OUTPUT«[1, 2]␤»
moritz_ rakudo: my @x = 1, 2, 3; @x.delete(1); say @x.perl
masak simple example: in github.com/masak/proto/blob/pls/t/s...ith-test.t , the @actions variable is threaded into the scopes of methods in mocking subclasses of the classes I want to test. as a consequence, I can run the code I want to test, and then inspect @actions. such a simple thing, but having all the components for it to work... is great.
p6eval rakudo 7579f7: OUTPUT«[1, Any, 3]␤»
cosimo moritz_: thanks 14:15
masak moritz_: by the way, do you agree that .invert could well be on List, instead of on EnumMap?
cosimo rakudo: my @x = 1, 2, 3; splice(@x, 1); say @x;
p6eval rakudo 7579f7: OUTPUT«1␤»
14:15 nathan06 joined
cosimo rakudo: my @x = 1, 2, 3; splice(@x, 1, 1); say @x; 14:15
p6eval rakudo 7579f7: OUTPUT«13␤»
moritz_ masak: on both, actually :-)
cosimo rakudo: my @x = 1, 2, 3; splice(@x, 1, 1); say @x.perl;
slavik not exactly perl6 related, but is there a gui/browser type app for cpan? (like synaptic but for cpan)
p6eval rakudo 7579f7: OUTPUT«[1, 3]␤»
pmurias moritz_: could you nopaste the 80 line error? 14:16
masak puts it on List
pmurias if it's so long surely it must be very usefull ;)
14:16 takadonet joined
masak moritz_: what'd be the failure mode for trying to .invert an odd-elemented List? 14:17
moritz_ masak: none. .invert just returns a list of Pairs
masak doesn't follow
moritz_: but what if the input List isn't a list of Pairs?
moritz_ oh wait
rakudo: say {a => 1, b => 2}.perl 14:18
p6eval rakudo 7579f7: OUTPUT«{"a" => 1, "b" => 2}␤»
moritz_ rakudo: say {a => 1, b => 2}.invert.perl
p6eval rakudo 7579f7: OUTPUT«(1 => "a", 2 => "b")␤»
moritz_ see, list of pairs
masak oh, I guess it'd just do $list.hash.invert, and the error'll come from .hash
moritz_: we're talking past each other.
moritz_ I'd think List.invert would turn <a b c> into a => 0, b => 1, c => 2
masak yes, exactly. 14:19
so $list.kv.hash.invert, then.
and it needs to be spec'd :)
14:19 christine joined
moritz_ except that .hash would stringify the keys/indexes 14:19
masak hm. 14:20
moritz_ it's more like $list.pairs.map({ .value => .key})
masak ah, better.
masak likes
jnthn Comments welcome from @all on news.perlfoundation.org/2010/07/hag...eta-m.html 14:21
pmurias moritz_: i'll try to install Mildew on a fresh 5.12.1 so that all the undeclared dependencies will be exposed
masak read "hugme grant" :)
moritz_ jnthn: I have that open in my browser tab all day, in order not to forget commenting
jnthn masak: :) 14:22
moritz_: I only just noticed it was posted :-)
moritz_ wrote hugme completly without a grant
pmurias: nopaste.snit.ch/21938 14:23
I kinda think that different installation procedures took different perl versions
but then I don't understand why noen of the unit tests failed
*none
masak oooh, S09!
14:24 tylercurtis joined
pmurias moritz_: that's a missing Devel::Declare 14:24
masak jnthn: I'm reading the "Benefits..." section, and thinking "wow, yes! send the guy money!"
"Accessing object attributes by index offset" \o/ -- today's theme, clearly. 14:26
pmurias moritz_: and MooseX::Declare is mentioned as Moose prereq
p6eval: did you ran Build in the tarball with the same @INC settings as when installing Mildew?
14:26 xlat joined
pmurias s/p6eval/moritz_/ 14:27
moritz_ yes
though I'm not entirely sure
pmurias you installed the things before with cpanm -L mildew-install?
moritz_ yes 14:28
masak jnthn: all in all, looks very good.
jnthn++
moritz_ masak: then also comment on the blog post
masak does so
14:31 Ross joined
pmurias moritz_: try running cpanm -L mildew-install . in the unpacked tar ball 14:31
moritz_ pmurias: I'll try tomorrow, now I'm frustrated and don't want to touch any CPAN-like things for a while :/ 14:33
pmurias just run the command if it failes i'll ask diakopter for a commit bit and finish the Mildew installation myself ;) 14:34
moritz_ yes, fails 14:35
masak time to wash some clothes. will be back later tonight.
14:35 masak left
tylercurtis jnthn: if the new P6object doesn't support multiple inheritance, Parrot will need to either keep the old P6object around or have a standard work-around that does. Or at least, if it doesn't, I'll complain loudly. :) 14:35
moritz_ hopes that NQP-rx will be that workaround 14:37
jnthn tylercurtis: The comment is a bit speculative. But more importantly, I doubt that most things will use the very core directly. They'll use a meta-object for classes that does support multiple inheritance. 14:40
Oops, about to fail at being on time for church 14:41
jnthn back later
moritz_ I just wonder about the usefulness of natively typed attributes when we don't have variables to store them 14:42
but I guess we have to start with one of them :-) 14:43
szabgab: any success with adding those chapters to the Makefile? 14:55
14:57 masonkramer joined
cosimo rakudo: my $x = "加"; say $x.chars; 15:04
p6eval rakudo 7579f7: OUTPUT«1␤»
cosimo rakudo: my $x = "加"; say $x.bytes;
p6eval rakudo 7579f7: OUTPUT«3␤»
cosimo cool
15:06 takadonet left
moritz_ rakudo: 3 does True; 15:08
p6eval rakudo 7579f7: ( no output ) 15:09
moritz_ rakudo: say 3 does True;
shouldn't that fail, because 3 is immutable?
p6eval rakudo 7579f7: ( no output )
moritz_ says 3 here locally
dalek ok: 92dcca7 | moritz++ | src/roles.pod:
apply some editorial advise; short explanation for infix:<but>
 15:23
15:25 nathan06 left 15:34 Mowah joined
cosimo I just finished chunked transfers support for LWP::Simple -> bit.ly/bZW1xa 15:35
\o/
moritz_ \o/ 15:37
cosimo++
15:38 skids left
cosimo I find myself *wanting* to write Perl 6 code more and more 15:38
moritz_ the patch doesn't look too painful 15:39
cosimo actually, it looks clean doesn't it? i'm sure it can be improved, but... 15:42
moritz_ these days when I write Perl 6 code I spend far more time on the application logic, testing and debugging than on working around rakudobug or NYI features
tadzik cosimo: still no binary file support, no?
getstore will work only for plaintexts
moritz_ cosimo: yes, looks clean... I meant it wasn't full of workarounds or so 15:43
moritz_ is looking for ideas for his weekly Perl 6 challenges 15:45
15:45 skids joined 15:46 Lawn joined 15:47 Mowah left 15:49 skids left 15:51 kid51 left 15:55 takadonet joined, skids joined
takadonet can a grammar with action method be lazy? In sense that once we have one 'full' record we stop parsing the file and allowed to resume later? 15:58
cosimo tadzik: why not?
tadzik: or, why it doesn't work for non-plaintext? 15:59
16:00 xlat left
moritz_ takadonet: in ideal Perl 6, yes 16:01
takadonet moritz_: but not currently?
moritz_ at least I think so
not currently
tadzik cosimo: I remember writing similar getstore for your LWP some time ago. Downloaded binary files were just recognised as text files by the OS 16:02
moritz_ currently you need to emulate the laziness by matching one item, lazily record that (via gather/take), and then anchor the next match
with $str.match($regex, :p($previous_position) 16:03
takadonet currently i'm using gather and take but perhaps I'm not anchor the next match
cosimo tadzik: a couple of weeks ago I didn't have HTTP headers parsing either 16:05
so it should be different now, but I haven't tried binary files yet
16:06 avar joined
avar Hi, who has access to the perl6 github account? Please remove my commit access to perl6/book. I keep getting github issue spam about it. 16:07
Someone added me to that account without me knowing about it
moritz_ sorry, forgot to remove you :( 16:08
will do
avar WHEE
oh yeah, forgot about having asked about it earlier. /me goldfish memory
Just noticed it because someone filed 5 issues or something :) 16:09
tadzik Hmm. Is there a way to get the subroutine parameters like in Perl 5? How to get unknown number of params? 16:10
16:10 takadonet left
slavik data dumper? 16:10
moritz_ sub f(*@a) { ...}
tylercurtis rakudo: sub foo { say @_; } foo 1, 2, 3, 4;
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "sub foo { "␤»
tylercurtis rakudo: sub foo { say @_; }; foo 1, 2, 3, 4; 16:11
p6eval rakudo 7579f7: OUTPUT«1234␤»
tylercurtis rakudo: sub foo (*@positional, *%named) { say @positional; say %named }; foo 1, 2, 3, 4, :hey<there>; 16:12
moritz_ avar: you should be gone now (I only have API access, so I'm not quite sure if it worked)
p6eval rakudo 7579f7: OUTPUT«1234␤hey there␤»
avar moritz_: whee 16:13
thanks! 16:15
16:15 avar left
tylercurtis rakudo: sub foo ($one, $two, *@many) { say $one; say $two; .say for @many; }; foo("One...", "Two...", "I don't know how to count higher than two..., but at least these examples answer your question, eh, tadzik?"); 16:16
p6eval rakudo 7579f7: OUTPUT«One...␤Two...␤I don't know how to count higher than two..., but at least these examples answer your question, eh, tadzik?␤»
tadzik yep, got it :)
16:17 agentzh left
ingy greetings 16:17
16:18 isBEKaml joined 16:19 tylercurtis left 16:23 justatheory joined 16:30 justatheory left
ingy is writing a simple YAML dumper for p6, because he's sick of .perl not dumping objects :\ 16:31
cosimo rakudo: my $status = 'HTTP/1.0 302 Found'; say $status ~~ m/30[12]/; 16:33
p6eval rakudo 7579f7: OUTPUT«␤»
cosimo rakudo: my $status = 'HTTP/1.0 302 Found'; say $status ~~ m/302/;
p6eval rakudo 7579f7: OUTPUT«302␤»
cosimo rakudo: my $status = 'HTTP/1.0 302 Found'; say $status ~~ m/30[2]/;
p6eval rakudo 7579f7: OUTPUT«302␤»
cosimo rakudo: my $status = 'HTTP/1.0 302 Found'; say $status ~~ m/30[12]/;
p6eval rakudo 7579f7: OUTPUT«␤»
cosimo mmh
rakudo: my $status = 'HTTP/1.0 302 Found'; say ($status ~~ '301' || $status ~~ '302'); 16:34
lue ohai o/
p6eval rakudo 7579f7: OUTPUT«0␤»
cosimo rakudo: my $status = 'HTTP/1.0 302 Found'; say ($status ~~ m/301/ || $status ~~ m/302/);
p6eval rakudo 7579f7: OUTPUT«302␤»
cosimo rakudo: my $status = 'HTTP/1.0 301 Found'; say ($status ~~ m/301/ || $status ~~ m/302/);
p6eval rakudo 7579f7: OUTPUT«301␤»
lue rakudo: my $status = 'HTTP/1.0 301 Found'; say $status ~~ m/30[1|2]/ 16:35
p6eval rakudo 7579f7: OUTPUT«301␤»
cosimo rakudo: my $status = 'HTTP/1.0 301 Found'; say $status ~~ m/30 <[12]>/;
p6eval rakudo 7579f7: OUTPUT«301␤» 16:36
cosimo rakudo: my $status = 'HTTP/1.0 30A Found'; say $status ~~ m/30 <[12A]>/;
p6eval rakudo 7579f7: OUTPUT«30A␤»
ingy rakudo: { die 'foo'; CATCH { say $! } } 16:44
p6eval rakudo 7579f7: OUTPUT«foo␤»
ingy rakudo: { die 'foo'; CATCH { die $! } }
p6eval rakudo 7579f7: ( no output )
lue rakudo: warn ('bad things'); say $!
p6eval rakudo 7579f7: OUTPUT«bad things at line 20:/tmp/K1aVbYwf4c␤Any()␤»
ingy TimToady: is this on purpose? ^^
cosimo TimToady: maybe this can be useful for newbies like me? nopaste.snit.ch/21945 16:45
diakopter is what on purpose? 16:46
TimToady rakudo: die "message" 16:47
p6eval rakudo 7579f7: OUTPUT«message␤ in main program body at line 20:/tmp/oP5txIC0nC␤»
TimToady looks like a bug
rakudo: die 'foo'; CATCH { when /bar/ {} } 16:48
p6eval rakudo 7579f7: ( no output )
TimToady looks like it's not rethrowing correctly
we don't tend to use exceptions heavily in p6 because it's counterproductive under parallelism 16:49
so it's not exercised much
we fail much more often than we die :)
diakopter ours is but to do and fail 16:50
ingy TimToady: thanks 16:52
lue is fail() less bad than die() ? [it is in English, just wondering here] 16:53
pugssvn r31624 | lwall++ | [S05] redirect to new char class syntax with patch from cosimo++
TimToady depends on whether you've done a 'use fatal', in which case fail does a die
but usually fail just returns an unthrown exception as a form of undef 16:54
16:54 TiMBuS left
TimToady if you try to use it as a value, then it dies 16:54
ingy TimToady: where did you lift that concept from? 16:55
it's quite interesting... 16:56
lue rakudo: my $a where 0..0xFF; $a = 300; # is this use of 'where' legal?
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Confused at line 20, near "my $a wher"␤»
TimToady every so now and again, I invent something from scratch
std: my $a where 0..0xFF; $a = 300; # is this use of 'where' legal?
p6eval std 31623: OUTPUT«ok 00:01 111m␤»
ingy if true {TimToady++}
TimToady yes :P
ingy: though it was a synthesis in the sense that there have always been languages that have the semi-predicate problem on either side of succeed/fail; I just decided to have it on neither side 16:58
16:58 dual joined
TimToady so you can return interesting values of both success and failure 16:58
lue [well, that changes where I should look to put it (excuse the pun) . Now to find where post_constraints for variables lie.]
another example of a quantum language :) 16:59
ingy TimToady: heads or tails? 17:00
17:00 aubreyja joined 17:03 jaffa4 joined
pugssvn r31625 | lwall++ | [S12] Nail down class representations sooner to make jnthn++ happier 17:03
jaffa4 hi
ingy greetings jaffa4 17:04
jnthn TimToady: \o/ 17:05
.oO( maybe I should read the commit pre-\o/ :-) )
jaffa4 hi ingy
How is Rakudo star?
jnthn jaffa4: Coming along. The REPL now remembers variables and stuff between lines, closures are fixed, various other bits too. Still some stuff to go. 17:09
lue would this work too: my $a; $a where 0..0xFF; ? [applying 'where' after declaration]
jnthn But the must-have things on the ROADMAP are more default with than not
lue: no
pugssvn r31626 | lwall++ | [S12] discuss pessimizing optimizations with augment
lue I figured /me off to find where variables are first created in the grammar 17:10
TimToady jnthn: did you see the CATCH problem earlier?
jnthn TimToady: no 17:12
Didn't backlog quite yet :-)
TimToady rakudo: die 'foo'; CATCH { when /bar/ {} } 17:13
p6eval rakudo 7579f7: ( no output )
TimToady that should rethrow
rakudo: die 'foo'; CATCH { die "bar" }
so should that
p6eval rakudo 7579f7: ( no output )
lue std: my $a where 0..0xFF is copy
p6eval std 31625: OUTPUT«===SORRY!===␤Two terms in a row at /tmp/ZeGXZHrlza line 1:␤------> my $a where 0..0xFF ⏏is copy␤ expecting any of:␤ bracketed infix␤ constraint␤ infix or meta-infix␤ statement modifier loop␤Parse failed␤FAILED 00:01 114m␤»
lue std: my $a is copy where 0..0xFF 17:14
p6eval std 31625: OUTPUT«ok 00:01 111m␤»
jnthn TimToady: Seems it swallows stuff too eagerly, yes.
moritz_ phenny: tell takadonet to look at <nopaste.snit.ch/21946>, which is a working example of lazy matching with Rakudo
phenny moritz_: I'll pass that on when takadonet is around.
17:14 Ross left
jnthn I'll have to glance S04 to check when we auto-rethrow. 17:14
17:15 Ross joined
TimToady any time you fall of the end, basically; it's only succeed that bypasses 17:15
jnthn TimToady: r31625 says just what I wanted 17:16
- layout P6opaque, P6hash, P5hash, P5array, PyDict, Cstruct, etc.
+ repr P6opaque, P6hash, P5hash, P5array, PyDict, Cstruct, etc.
17:16 tedv joined
jnthn I guess this is the repr that was set with the "is repr" trait? 17:16
So if you :repr(*) then .repr on the HOW returns *?
TimToady I presume 17:17
at least, for the class it would
an instance might return something more specific, I suppose
jnthn Yeah, that was going to be my next question.
Do we have a per-instance version too. :-)
.REPR macro or some such.
Anyway, +1 on the changes. Thanks. 17:18
slavik where can I read about diff between .MACRO and .method?
jnthn TimToady: btw, I liked the "also" change yesterday. 17:19
slavik was Java's "everything implicitly extends Object" considered? Looks like a decent way of getting base functionality into an object
sorear good * #perl6
moritz_ good now, sorear :-) 17:20
std: now
slavik o/
p6eval std 31625: OUTPUT«ok 00:01 107m␤»
slavik std: now + 1
p6eval std 31625: OUTPUT«ok 00:01 109m␤»
slavik std: now + 1
p6eval std 31625: OUTPUT«ok 00:01 109m␤»
sorear I interpret the 'repr' trait on classes as being a specialization pragma
slavik :(
sorear The class body is allowed to assume that a specific repr will be used
moritz_ slavik: std does syntax check.
lue rakudo: say now
slavik oh
p6eval rakudo 7579f7: OUTPUT«Could not find sub &now␤ in main program body at line 20:/tmp/i5MV7S6h5i␤»
slavik rakudo: now
p6eval rakudo 7579f7: OUTPUT«Could not find sub &now␤ in main program body at line 20:/tmp/ChBxXBjQxq␤»
jnthn slavik: The default type of everything is Any 17:21
slavik aww :(
jnthn slavik: Which in turn inherits from Mu
moritz_ facts: 1) gather/take is lazy
jnthn sorear: Yeah, I figure you can't take advantage of that knowledge until you parsed the whole class body.
moritz_ 2) my @a = 1..*; # also lazy
jnthn sorear: In case also repr(*); shows up along the way
moritz_ 3) assigning the result from gather/take to an array evaulates it eagerly 17:22
jnthn But that's fine I guess.
slavik moritz_: unless it's ==> ?
moritz_ slavik: I'm talking about the 'is' situation in Rakudo, not the 'should be' situation
slavik oh
moritz_ sorry if that wasn't clear
sorear pmurias: Is your dist a fork of std or are you intending to merge it back at some point?
slavik moritz_: I have no idea what the is situation is ;)
isBEKaml rakudo: class Foo{}; Foo.^parents.perl.say;
moritz_ slavik: facts 1) to 3) describe it :-) 17:23
p6eval rakudo 7579f7: OUTPUT«(Any, Mu)␤»
isBEKaml slavik: ^^
lue [why is the basis for everything named Mu?]
slavik moritz_: I mean wouldn't op(<==) make it lazy?
17:23 tadzik left
slavik it's a quark, and smaller than an atom, which is Any 17:24
or so I think
jnthn lue: S02
moritz_ slavik: it would. But I'm kinda surprised that assignment doesn't work lazily with gather/take, even though my @a = 1..*; *is* lazy 17:25
so, the real question is: BUG?
sorear TimToady, jnthn: die in CATCH is a longstanding Parrot bug 17:26
trac.parrot.org/parrot/ticket/1631 17:27
jnthn sorear: I was kinda hoping it was mostly a code-gen bug but maybe it's not so simple
e.g. we needed to emit a pop_eh where we're not
jnthn checks the ticket
sorear it's mostly likely a PAST bug.
jnthn sorear: ah, yes 17:28
OK
Taht's probably easier to fix
In PAST::Compiler somewhere
(easier than if it was deep in Parrot) 17:29
ops.'push_pirop'('pop_eh') 17:30
hmm
Looks like it is
ingy "Error: Too many positional parameters passed; got 2 but expected 1" ... what subroutine???!!! 17:31
moritz_ ingy: less than awesome, and reported. Happens for method dispatch
isBEKaml moritz_: about my @a= 1..*; maybe you missed the parens around rvalue? I recently hit on this and pmichaud++ pointed out that '=' binds tighter than '..'. (What's the situation here? I sorely lack context here.. :()
jnthn moritz_: huh, when? 17:32
ingy error reporting is teh big suck right now :\
pmurias sorear: i could merge back the few changes i made, and make Dist::Zilla make a dist from the pile of files in src/perl6
jnthn Doesn't the backtrace have the current method in?
ingy at least it keeps me around :\
moritz_ isBEKaml: that's only the case in $scalar = 1..*;
17:32 Lawn left
moritz_ rakudo: class Foo { method x() { } }; Foo.x(5) 17:32
jnthn rakudo: class C { method m { } }; C.m(42)
moritz_ seems that p6eval is borked again 17:33
Too many positional parameters passed; got 2 but expected 1
that's what I get locally
p6eval rakudo 7579f7: ( no output )
jnthn I get
Too many positional parameters passed; got 2 but expected 1 in 'C::m' at line 1 in main program body at line 1
Oh, do you mean in the REPL?
isBEKaml moritz_: Ah, I see..
moritz_ jnthn: yes
jnthn Oh 17:34
moritz_ jnthn: but I've also seen that creeping up in normal code
jnthn moritz_: Got a ticket for a non-REPL case?
ingy rakudo: say Boolean::true
p6eval rakudo 7579f7: OUTPUT«Can not find sub Boolean::true␤ in main program body at line 1␤»
ingy rakudo: say Boolean::True 17:35
p6eval rakudo 7579f7: OUTPUT«Can not find sub Boolean::True␤ in main program body at line 1␤»
moritz_ jnthn: Regex::Match.from(1, 2, 3);
lue ingy: that's Bool::True
isBEKaml rakudo: say (1 < 0).perl
ingy thx
p6eval rakudo 7579f7: OUTPUT«Bool::False␤»
moritz_ jnthn: gives a backtrace with line number, but without the method name
jnthn sorear: oh, may have found it 17:36
moritz_: oh
ingy rakudo: use Bool; say True
jnthn moritz_: That's an NQP class?
p6eval rakudo 7579f7: OUTPUT«===SORRY!===␤Unable to find module 'Bool' in the @*INC directories.␤(@*INC contains:␤ lib␤ /home/p6eval/.perl6/lib␤ /home/p6eval//p1/lib/parrot/2.5.0-devel/languages/perl6/lib␤ .)␤»
moritz_ jnthn: or even PIR
jnthn moritz_: ah, I can totally believe we may be crappy on those
isBEKaml jnthn: irclog.perlgeek.de/perl6/2010-07-11#i_2538819
this is another.
ingy is Bool::True the shortest form of Bool::True? 17:37
jnthn ingy: Normally just True works
moritz_ rakudo: say True
p6eval rakudo 7579f7: OUTPUT«1␤»
moritz_ rakudo: say ?1 # even shorter
p6eval rakudo 7579f7: OUTPUT«1␤»
ingy why does it Str to '1'? 17:38
seems arcane
moritz_ masak recently said about the same
ingy </bitchin>
17:39 Kodi joined
ingy phenny: tell masak /me hugs masak 17:40
phenny ingy: I'll pass that on when masak is around.
17:43 whiteknight joined 17:44 masonkramer left
jnthn > { die 'foo'; CATCH { die $! } } 17:45
foo
ingy rakudo: use Test; is "foo\nfoo\n", "bar\nbar\n", 'Baz!';
p6eval rakudo 7579f7: OUTPUT«not ok 1 - Baz!␤# got: "foo\nfoo\n"␤# expected: "bar\nbar\n"␤»
jnthn spectests and goes to make some noms
ingy is there a way to tell test not to serialize newlines as \n?
jnthn ingy: Suspect it's just doing .perl on them 17:46
ingy nod :\
I need to get my waiver in to pmichaud so I can get meh bit! 17:47
moritz_ ingy: what would you have it serialize to instead?
ingy moritz_: I would do it same as p5 17:48
sorear moritz_: I have root on feather1, can I help you with your password problems?
ingy got: "foo
moritz_ sorear: no, wrong matchine
ingy # foo
"
like that
moritz_ ingy: that's a bit bad if there are spaces after the newline
but now that .subst properly works for zero-width assertions, it should be SMOP to patch Test.pm if you really want it 17:49
ingy moritz_: testing yaml dumps is almost impossible with \n
and TestML supports diff output of got/expected 17:50
moritz_ ingy: just get it right :-)
pragma_ I disagree.
moritz_ pragma_: with what?
ingy anyway, I just want it to be the same as p5 by default, and have options to do whatever pragma_ wants ;) 17:51
pragma_ concedes.
ingy hugs pragma_
\@_o/ 17:52
moritz_ ingy: should it chomp trailing newlines? 17:54
ingy whatever p5 does. ( I think that is no)
jnthn moritz_: S02-literals/quoting.t has two fails on Windows - iirc this is the one you asked me to check?
moritz_ jnthn: yes... qx and qqx at the very end 17:55
jnthn not ok 145 - qx
# got: "world\r\n"
# expected: "world\n"
not ok 146 - qqx
# got: "world\r\n"
# expected: "world\n"
ingy rakudo: say [].WHAT.WHAT
p6eval rakudo 7579f7: OUTPUT«Array()␤»
moritz_ jnthn: does it work if you .chomp the result from qx, and remove the \n from the RHS?
lue
.oO(darn those crazy windows)
ingy rakudo say Array(1,2) 17:56
rakudo: say Array(1,2)
p6eval rakudo 7579f7: OUTPUT«Could not find sub &Array␤ in main program body at line 20:/tmp/Xa7gAanl93␤»
pugssvn r31627 | Kodi++ | [S32/Temporal] Changed to use a different way of specifying time zones, which is hopefully saner than my last proposal.
ingy rakudo: say Array()
p6eval rakudo 7579f7: OUTPUT«Could not find sub &Array␤ in main program body at line 20:/tmp/GixAlIescH␤»
17:56 masak joined
lue ohai masak o/ 17:57
ingy masak: \ (o) (x) /
jnthn moritz_: Probably will since I fixed .chomp for windows a while back
masak Oddly, Perl 6 doesn't seem to be mentioned at en.wikipedia.org/wiki/Semipredicate_problem
someone should add it there. :)
phenny masak: 17:40Z <ingy> tell masak /me hugs masak
jnthn moritz_: nom time here
masak submerges again
17:57 masak left
ingy jnthn: nom nom 17:57
what kind of thing does .WHAT return? 17:58
17:58 aubreyja left
moritz_ a type object 17:59
lue masak: Perl 6 would get its own section on that page :)
ingy rakudo: say Array() 18:00
p6eval rakudo 7579f7: OUTPUT«Could not find sub &Array␤ in main program body at line 20:/tmp/zJ9HUJWQrQ␤»
moritz_ rakudo: say Array
p6eval rakudo 7579f7: OUTPUT«Array()␤»
ingy rakudo: say Array.WHAT
p6eval rakudo 7579f7: OUTPUT«Array()␤»
ingy whay not Type() ? 18:01
moritz_ coercion syntax NYI
rakudo: say Array.perl # if you want runnable code...
p6eval rakudo 7579f7: OUTPUT«Array␤»
18:02 aubreyja joined
ingy rakudo: say Array.perl 18:02
p6eval rakudo 7579f7: OUTPUT«Array␤»
18:02 aubreyja_ joined
ingy rakudo: say Array.perl.WHAT 18:02
p6eval rakudo 7579f7: OUTPUT«Str()␤»
moritz_ rakkudo: say "ab\ncd".subst(rx/^^/, '# ', :g)
ingy rakudo: say [].WHAT eq 'Array()' 18:03
p6eval rakudo 7579f7: OUTPUT«1␤»
18:03 aubreyja_ left
moritz_ rakudo: say "ab\ncd".subst(rx/^^/, '# ', :g) 18:03
p6eval rakudo 7579f7: OUTPUT«# ab␤# cd␤»
sorear mm rakkudo
18:03 isBEKaml left
moritz_ rakudo: use Test; isa_ok Mu, Mu, 'foo' 18:04
p6eval rakudo 7579f7: OUTPUT«No applicable candidates found to dispatch to for 'isa_ok'. Available candidates are:␤:(Mu $var, Any $type)␤:(Mu $var, Any $type, Any $msg)␤␤ in main program body at line 20:/tmp/pGmGDYICap␤»
dalek kudo: b48fb92 | moritz++ | Test.pm:
allow Mu type checks in isa_ok
 18:10
kudo: 582dc43 | moritz++ | Test.pm:
[Test] make diagnostics in is() like in Perl 5, ingy++
ingy moritz_++ 18:17
\o/
moritz_: thanks 18:18
ingy updates rakudo. prays that other shit don't break.
18:21 eternaleye joined
ingy rakudo: say 'ingy'.times(3) 18:23
p6eval rakudo 7579f7: OUTPUT«Method 'times' not found for invocant of class 'Str'␤ in main program body at line 20:/tmp/rpkvq3boRA␤»
ingy rakudo: say 'ingy' x 3
p6eval rakudo 7579f7: OUTPUT«ingyingyingy␤» 18:24
ingy rakudo: say 'ingy' x 'foo'
p6eval rakudo 7579f7: OUTPUT«␤»
moritz_ rakudo: say ('ingy' xx 3).perl
ingy guessed TimToady would be more creative ;)
p6eval rakudo 7579f7: OUTPUT«("ingy", "ingy", "ingy")␤»
ingy nice 18:25
rakudo: say (1,2).WHAT
p6eval rakudo 7579f7: OUTPUT«Parcel()␤»
ingy :D
Parcel!
moritz_ rakudo: say ('ingy', ('ingy', 'foo') xx *).[^5].join(', ') 18:26
p6eval rakudo 7579f7: OUTPUT«ingy, ingy, foo, ingy, foo␤»
ingy is that a term from the gay community?
ingy dicks
I mean *ducks*
moritz_ ingy: Parcel = Parenthesis Cell
ingy ha!
TimToady++
only TimToady could come up with that 18:27
I <3 the Perl community
screw Python... (even if they have a more usable language)
ingy is BACK!
at least for now... 18:28
only Perl is -Ofun
lue afk 18:30
18:30 [Coke] joined
diakopter interesting: cacm.acm.org/magazines/2010/7/95061...g/fulltext 18:33
(but over my head) 18:34
ingy is there an example of a perl module dist with `make install` capability? 18:38
moritz_ ingy: all of my Perl 6 modules are installable, provided you generate a makefile with ufo (see <github.com/masak/ufo/>) 18:39
ingy moritz_: example url? 18:40
arnsholt Just see ufo's docs
ingy wants to grab moritz_' modules
18:42 Chillance joined
ingy cues up Side A of Heart's Dreamboat Annie, FTW! 18:43
moritz_ ingy: github.com/moritz/Math-RungeKutta 18:45
ufo && make install 18:46
ingy :-*
18:46 takadonet joined 18:47 nadim joined, nathan06 joined
ingy moritz_: I already have my own Makefile which handles 'make test' and 'make t/specific.t' 18:50
does ufo, do make test?
moritz_ make test, yes
ingy guess I can find out and patch if necessary
moritz_ make t/specific.t not
hugme: add ingy to ufo 18:51
hugme moritz_: sorry, I don't know anything about project 'ufo'
moritz_ uhm.
just submit a patch :-)
ingy to the mothership?
moritz_ that masakship, yes :-)
ingy XD 18:52
moritz_: this would be a nice patch: ufo - | make -f - install 18:58
pugssvn r31628 | Kodi++ | [t/spec] Overhauled DateTime.t. Rakudo can pass all of these with my latest version of DateTime.pm.
moritz_ ingy: and a SMOP... 19:01
diakopter wow, 1000 commits to pugscode svn in the last 2 months
19:02 envi^home left
ingy moritz_: SMOP forthcoming :) 19:02
moritz_ ingy: I'm faster :-)
SMOP already written and tested
... and pushed
ingy :P 19:03
moritz_++ # I guess :\
;)
19:04 cono left
moritz_ that MAIN sub makes it really easy to have command line arguments with default values 19:04
ingy moritz_: you realized that '-' means stdin, yeah?
or stdout
depending
make -f - (reads stdin)
moritz_ is open('-') specced? 19:05
ingy I just meant that `ufo -` prints to stdout
19:05 cono joined
moritz_ yes, that's what my patch did 19:06
ingy ok, the log msg didn't say that
sorry, should have looked at the code
moritz_ it's alien technology. Of course it's magic :-) 19:07
ingy neat, MAIN binds to @*ARGS?
moritz_ yes, with some additional magic 19:08
ingy <3
moritz_ like possible multi subs
ingy nod
moritz_ and named arguments bein --foo options
ingy ossum
ingy puts on the fIREHOSE 19:09
moritz++ # now I can install TestML to write YAML tests 19:10
xoxo 19:11
moritz_ \o/ 19:12
ingy > ufo - | sudo make -f - install clean 19:15
works!
moritz_ \o/
but why sudo? are you installing into a non-standard location? 19:16
default is ~/.perl6/lib/ 19:17
19:17 takadonet left 19:19 takadonet joined
ingy moritz_: O 19:21
moritz_: I thought stuff went where Test.pm goes...
moritz_: take a look at: > perl -e 'use Test::More tests => 1; is "foo\nfoo\n", "foo\nfoo", "A test";' 19:22
your new output++ is a little off
moritz_ you mean no leading/trailing ''? 19:23
19:23 payload joined 19:25 Kodi left
ingy moritz_: right 19:27
moritz_: also the period after Failed test 1 of 1. :) 19:28
dalek kudo: d60d064 | moritz++ | Test.pm:
[Test] quote strings in diagnostic output
 19:31
19:33 masak joined
masak ahoy! 19:33
moritz_ aloha
masak so now there's two of us believing that enum values maybe should stringify to their names. ingy++ \o/ 19:34
ingy masak: say Bool::True; 19:35
masak ingy: True 19:36
ingy :D
masak , the living evalbot
ingy oui
masak lives
masak: I think I meant more that Bool should be first class in Perl 6 19:37
moritz_ it's an enum as currently specced
ingy but I like where you are going.
masak ingy: I like that it's an enum. pmichaud seems grudgy over it being an enum. 19:38
ingy masak: I like it too.
masak: I'm writing a YAML dumper right now
after I get it working ok, we should hack in .yaml support 19:39
then .perl will be sad
:)
masak :)
19:43 takadonet left
ingy masak: did you like our ufo tweak? 19:44
masak haven't studied it in detail. 19:45
ingy ufo - | make -f - install clean
no Makefile!
masak huh :)
why?
ingy case I already have a Makefile
and you say to throw it away 19:46
so why ever make one?
except as a speed op :P
masak feel free to prepare a patch, and I'll review it :)
ingy masak: too late 19:47
moritz++ !!!
moritz_ already pushed one
masak o.O :)
I'll review it pushed, then. :)
ingy already pulled it
it works great over here ;) 19:48
masak moritz_++ # cute change
is assigning $*OUT equivalent in all ways to binding $*OUT? 19:49
sorear assigning $*OUT is a fatal error
moritz_ sorear: where is that specced?
masak sorear: not assigning *to* $*OUT.
moritz_ masak: except when one modifies the variable it is assigned to 19:50
masak ah, right.
sorear it's not specced in so many words, but S03:1569 strongly implies that all bindings to $*OUT are readonly
ingy moritz_: your latest Test patch is great. thanks 19:51
moritz_ sorear: no.
sorear: it just describes ::=
sorear: and only because it was RO-bound it's not modifiable in this example 19:52
sorear it establishes the standard way to bind $*OUT as ::=, not :=
moritz_ no. It's just an exmaple for ::=. 19:53
sorear along with S02:2705
Wait, how are you doing that
19:54 IllvilJa joined
sorear I tried to take the synopses at face value once, there wasn't nearly enough to go on 19:54
moritz_ $*OUT is not a user-defined dynamic variable
ingy rakudo: if 0 {} elsif 1 {say "OK"} 19:55
p6eval rakudo 582dc4: OUTPUT«OK␤»
ingy elsif lives!
worst name evar
'elsif' is teh worst name evar -- gwall 19:56
gloria actually told me that. (paraphrased)
moritz_ things that X_CREAT is worse (don't remember what X is)
*thinks
masak kinda likes 'elsif' 19:59
it's pronounced the way it's spoken! :P 20:00
ingy el `Sif
masak std: if 0 {} else if 1 { say "OK" } 20:01
p6eval std 31628: OUTPUT«===SORRY!===␤Please use 'elsif' at /tmp/7j3yGrZiSI line 1:␤------> if 0 {} else if ⏏1 { say "OK" }␤Confused at /tmp/7j3yGrZiSI line 1:␤------> if 0 {} else if ⏏1 { say "OK" }␤ expecting statement_control␤Parse
..failed␤FAILED…
20:01 rv2733 left
masak though in the above case, I'm sorely tempted to advocate a "just understand what I want already" kind of approach :) 20:02
if people want to write it out in long form, why not just let them? 20:03
moritz_ if we allow too much stuff, we end up with crappy error messages, long documentation and lettle consistency 20:05
ingy wow: I almost asked how to do a simple thing (because I couldn't find doc on it) but then I just tried: for %node.keys.sort -> $key {
\o/
perl6 is sometimes intuitive
moritz_ ingy: you can also say %node.sort, and it sorts by key too, and returns a list of paris
*pairs 20:06
moritz_ can't type anymore
ingy moritz_: I tried pairs but messed it up
masak is there a way to do nested signatures on pairs?
moritz_ rakudo: .value.say for { a => 1, b => 4, d => 8, c => 0 }
p6eval rakudo 582dc4: OUTPUT«Method 'value' not found for invocant of class ''␤ in main program body at line 20:/tmp/R7ByJh2yA_␤»
masak because there totally should be.
moritz_ masak: sure, with .key and .value
rakudo: .value.say for { a => 1, b => 4, d => 8, c => 0 }.sort
p6eval rakudo 582dc4: OUTPUT«1␤4␤0␤8␤» 20:07
masak moritz_: no, I mean binding the key and value in each pair to a parameter.
moritz_ masak: sure, $pair ($key, $value) 20:08
masak \o/
TimToady++ jnthn++
moritz_ which works through the general mechanism of methods calls
masak the section "When is a closure not a closure" is strange and exhilarating. today I understand a little bit more of it than earlier. 20:09
jnthn rakudo: { die 'foo'; CATCH { die $! } } 20:22
p6eval rakudo 582dc4: ( no output )
jnthn rakudo: { die 'foo'; CATCH { die $! } }; say "ok"
p6eval rakudo 582dc4: ( no output )
moritz_ the first segfaults 20:23
jnthn ah, ok
Does anyone know the ticket #?
I don't spot it.
(Got a fix locally, gets through the Parrot tests and Rakudo spectest is looking good - almost there.)
masak jnthn: I think there is one. 20:25
looking.
jnthn: rt.perl.org/rt3/Ticket/Display.html?id=64262 20:26
20:27 timbunce joined 20:33 wilx joined 20:34 jaffa4 left
ingy how do I change a grammar token at runtime? 20:44
(for my own grammar)
lue ohaio o/ 20:46
masak ingy: I imagine the answer is the same as "how do I change a method at runtime?" 20:47
moritz_ rakudo: class A { method foo { say "old" } }; my $x = A.new() but role { method foo { say "new" } }; $x.foo 20:50
p6eval rakudo d60d06: ( no output )
masak std: until 42 -> $a { ... }
p6eval std 31628: OUTPUT«ok 00:01 114m␤»
moritz_ says "new" here
masak std: unless 42 -> $a { ... } 20:51
p6eval std 31628: OUTPUT«ok 00:01 111m␤»
moritz_ so "but" seems to work for replacing methods
ingy rakudo: grammar G1 { our $t = 'xxx'; token t { $t } rule TOP { <t> } } G1.parse("aaa xxx bbb")
p6eval rakudo d60d06: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 20␤»
jnthn does to do it in place
moritz_ right
masak ingy: semicolon.
moritz_ two of them :-)
masak right.
ingy: the rule is simple.
ingy rakudo: grammar G1 { our $t = 'xxx'; token t { $t }; rule TOP { <t> } }; G1.parse("aaa xxx bbb");
p6eval rakudo d60d06: ( no output )
ingy rakudo: grammar G1 { our $t = 'xxx'; token t { $t }; rule TOP { <t> } }; say G1.parse("aaa xxx bbb"); 20:52
p6eval rakudo d60d06: OUTPUT«␤»
ingy rakudo: grammar G1 { our $t = 'xxx'; token t { $t }; rule TOP { <t> } }; say G1.parse("aaa xx bbb");
p6eval rakudo d60d06: OUTPUT«␤»
ingy rakudo: grammar G1 { our $t = 'xxx'; token t { $t }; rule TOP { <t> } }; say G1.parse("xxx bbb");
p6eval rakudo d60d06: OUTPUT«xxx ␤»
ingy rakudo: grammar G1 { our $t = 'xxx'; token t { $t }; rule TOP { <t> } }; say G1.parse("xx bbb"); 20:53
p6eval rakudo d60d06: OUTPUT«␤»
ingy rakudo: grammar G1 { our $t = 'xx'; token t { $t }; rule TOP { <t> } }; say G1.parse("xx bbb");
p6eval rakudo d60d06: OUTPUT«xx ␤»
ingy that seems like an ok way to do it for my needs
moritz_ you know that rules also can take parameters? 20:55
20:55 aubreyja left
moritz_ rakudo: grammar A { token TOP { <b("foo")> }; token b($x) { $x } }; say A.parse('foo') 20:55
p6eval rakudo d60d06: OUTPUT«foo␤»
moritz_ rakudo: grammar A { token TOP { <b("foo")> }; token b($x) { $x } }; say A.parse('bar') 20:56
p6eval rakudo d60d06: OUTPUT«␤»
ingy rakudo: grammar G1 { our $t = '+++'; token t { $t }; rule TOP { <t> } }; say G1.parse("+++ bbb");
p6eval rakudo d60d06: OUTPUT«+++ ␤»
ingy rakudo: grammar G1 { our $t = '+++'; token t { $t }; rule TOP {<t>} }; say G1.parse("+++ bbb"); 20:57
p6eval rakudo d60d06: OUTPUT«+++␤»
20:58 nathan06 left
lue rakudo: my $a where 0..10; $a = 20; 20:59
p6eval rakudo d60d06: OUTPUT«===SORRY!===␤Confused at line 20, near "my $a wher"␤»
lue > my $a where 0..10; $a = 20;
Contextual $*PARAMETER not found
moritz_ std: my $a where 0..10; $a = 20;
p6eval std 31628: OUTPUT«ok 00:01 111m␤» 21:00
21:00 timbunce left
lue ever closer! I can paste what I changed so far, 21:00
moritz_ rakudo: my Int $a where 0..10; $a = 20; say $a
p6eval rakudo d60d06: OUTPUT«===SORRY!===␤Confused at line 20, near "my Int $a "␤»
lue [and yes, I remembered a separate branch this time :)]
masak rakudo: repeat while 42 -> $x! { say "OH HAI" } 21:04
p6eval rakudo d60d06: OUTPUT«OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH HAI␤OH
..HAI␤OH…
masak rakudo: repeat while 42 -> $x! { say $x }
p6eval rakudo d60d06:
..OUTPUT«␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤
masak hm. I guess that is according to spec. 21:05
jnthn masak: It's only saying exactly the same as -> $x would.
masak aye.
21:05 timbunce joined
masak I kinda thought it'd fail to bind. 21:05
jnthn Why? 21:06
masak because the first iteration doesn't pass in a value.
or so I thought.
jnthn ?
masak in a 'repeat while' loop, one iteration is run, then the condition is tested. 21:07
jnthn Ah, yes
masak so in the first iteration the condition hasn't been tested.
thus (I thought) no parameter would be passed in.
rakudo: repeat while my $a = 42 { say $a } 21:08
p6eval rakudo d60d06:
..OUTPUT«Any()␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤42␤
21:09 Trashlord left
lue [ IIUC, here's what I did: gist.github.com/471833 ] 21:09
moritz_ masak: re last section of S04, I agree it's a tough read 21:10
masak moritz_: I'm not sure I understand what 'cloning' means exactly here. and to the extent that I (think I) understand, I'm not sure it's a good word for it. :) 21:11
moritz_ "cloning" means "take a snapshot of the lexical environment" 21:12
masak do you have a concrete example?
moritz_ my $x = 3; my $y = { ... }; # $x and $y are the lexical environment of the block 21:14
+ the setting, of course. But that's immutable anyway
masak I don't really see where "snapshot" comes in. one keeps around a pointer to that lexpad. 21:15
"snapshot" sounds like things are photographed and then the photo never changes.
that's not what closures do at all.
you can emulate mutable objects with closures, precisely because it's not just a snapshot, but a living link back to the surrounding lexpad. 21:16
jnthn It's a snapshot of that link. :-)
moritz_ masak: then think of it as a "branch" in the version control sense
masak I'm at the point where I think I got the theory right, but the terminology is confusing me. 21:17
moritz_ masak: have you seen the "illguts", the illustrated guts of perl 5?
masak no.
ah, found it: www.cpan.org/authors/id/GAAS/illguts-0.09.pdf 21:18
moritz_ hm, it's not in there anyway... sorry
I'm looking for an illustration I've once seen
masak I saw that PDF the other day.
moritz_: maybe you're thinking of HOP. it has a nice picture of lexpads, IIRC. 21:19
ingy masak: does ufo make install just install .pir, and is that enough?
masak ingy: per default, yes.
ingy: moritz_ disagrees-ish with me on that.
ingy install lib/TestML/Runner/TAP.pir ~/.perl6/lib/TestML/Runner/TAP.pir 21:20
masak ingy: there's another install target that installs the .pm files, too.
moritz_ install-src
masak right.
ingy Unable to find module 'TestML::Runner::TAP' in the @*INC directories.
masak huh.
21:20 Ross left
moritz_ masak: if you have HOP available, page 76 describes how Perl 5 does it 21:20
ingy (@*INC contains: t lib /Users/ingy/.perl6/lib 21:21
masak moritz_: I don't right now, but I'll look it up when I get the chance.
jnthn masak: Installing just the PIR is not enough
masak jnthn: why not?
jnthn Because we look for the .pm files first 21:22
Because the .pir versions don't have a way for us to pull the version info out
masak huh.
jnthn And because .pir files are just cached compiled versions
masak well, then clearly ufo must install both .pir and .pm files.
jnthn The original source is authoritative.
masak changes ufo
ingy \o/ 21:23
jnthn masak++
ingy masak++
masak ingy++ # field-testing
jnthn There was a HUGE storm here earlier. I hope it's nice and cold out now.
moritz_ it's as simple as adding a install-src dependency to install:
jnthn goes to take a quick walk and see
ingy feels like the first man in space
moritz_ btw I got the same error too, but attributed it to my own stupidity
ingy moritz_: for what? 21:24
masak moritz_: good idea.
moritz_ ingy: pir-installed-only module
masak pushed. 21:25
ingy: try now.
lue rakudo: say $*PARAMETER 21:26
p6eval rakudo d60d06: ( no output ) 21:27
21:29 plobsing joined
ingy can I get 'use Module' to not import the exports? 21:29
lue
.oO(I wonder why it couldn't find contextual $*PARAMETER...)
 21:30
ingy masak: seems to work 21:31
masak yay
moritz_ ingy: 'need Module;' iirc
masak std: my @a; my @b; zip(@a;@b) 21:33
p6eval std 31628: OUTPUT«ok 00:01 113m␤»
21:33 slavik left
masak std: loop (;;) -> $a { ... } 21:35
p6eval std 31628: OUTPUT«===SORRY!===␤Missing block at /tmp/SPmRet1WfV line 1:␤------> loop (;;) ⏏-> $a { ... }␤Parse failed␤FAILED 00:01 107m␤»
21:35 ashleydev joined
lue Grammar.pm is the one place where a document of some sort would be helpful in understanding all the weird little <tokens> and whatnot 21:36
ingy moritz_, masak, everyone: CHECK IT!! -> github.com/ingydotnet/yaml-pm6/blob...t/dump.tml
YAML.pm passes 4 tests writtn in TestML
masak nice.
moritz_ \o~ 21:37
ingy it's STOOPID PHRESH \o/
lue oooh! aaah! 21:38
ingy masak++ # thanks for ufo++ 21:39
moritz_++ # thanks for fixing Test etc
ingy writes more tests and refactors
YAML::dump will be soon! 21:40
21:40 ashleydev left
lue Can somebody point me in the right direction? I'm stuck on that error concerning $*PARAMETER 21:41
21:43 ashleydev joined, ashleydev left
jnthn lue: For now probably just put parens around the decl. 21:46
lue > my ($a where 0..10); $a = 20; 21:47
20
jnthn Ah, guess constraint types on variables are not yet re-implemented. :-( 21:51
lue so it's no longer a grammar issue? [see gist.github.com/471833 for what I did] 21:52
masak loliblogged! use.perl.org/~masak/journal/40442 21:55
& 21:57
masak sleeps 22:02
22:02 masak left
lue jnthn: what's weird is that 'where' works in parameters 22:09
rakudo: sub xyzzy($a where 0..10) { say $a }; xyzzy(5); xyzzy(15); 22:10
p6eval rakudo d60d06: OUTPUT«5␤»
arnsholt Does p6eval no longer print STDERR?
jnthn lue: No, it'll be more than a grammar issue. 22:12
lue: Not weird.
The way signatures are built is a bit different.
lue The tricky bit [for me, considering my inexperience with Rakudo] is finding where the heck a reimplementation of 'where' for !signatures goes 22:15
jnthn lue: I'm not quite sure how to do it off hand.
arnsholt It'd be in several places I guess? 22:16
lue arnsholt: that's exactly what I don't like :D
jnthn lue: It'll finally need to be attached in declare_variable 22:17
arnsholt Off the top of my head, both binding and assignment should heed constraints I think
lue jnthn: I wouldn't expect you to know. After all, it's NYI
jnthn arnsholt: It's not that they aren't heeding them.
arnsholt: It's that the where constraint ain't being attached to the container.
arnsholt Oh, right =)
jnthn So that's "all" that needs doing :-)
arnsholt For some value of all, yeah
jnthn It shouldn't be too bad - I suspect we may want a fake $*PARAMETER to collect the where clauses. 22:18
But that also feels a tiny bit messy
May work though
Thing is, at the time I think STD didn't parse a where clause outside of a signature.
lue rakudo: our Fake $*PARAMETER;
p6eval rakudo d60d06: OUTPUT«===SORRY!===␤In "our" declaration, typename Fake must be predeclared (or marked as declarative with :: prefix) at line 20, near " $*PARAMET"␤» 22:19
jnthn lue: will have to be in the grammar IIRC
:my $*PARAMETER; or some such
See where similar shows up inside Signature
er, signature
22:20 pjcj joined
lue signature as in signature.(pm|pir) ? 22:21
jnthn no
the rule in grammar.pm
lue ah 22:22
dalek kudo: 62ee6d9 | jonathan++ | build/PARROT_REVISION:
Bump PARROT_REVISION to get PAST::Compiler fix for exception handler code
22:24
lue there's only a signature *token*, and :my $*PARAMETER doesn't show up there. [it does show up in the token below it, which is token parameter] 22:25
.oO(let's add it and see what happens!)
jnthn lue: aha yes
Taht's the one I meant
pugssvn r31629 | jnthn++ | [t/spec] Add test by KyleHa++ that didn't seem to have made it from the RT ticket to the tests for the CATCH not catching its own exception bug. 22:26
22:30 azert0x left 22:33 kid51 joined
lue > my $b where 0..10; $b = 39; 22:37
39
at least I don't need parentheses for that result anymore :)
22:48 payload left 22:51 timbunce left
araujo lue, was it needed before? 22:59
lue to get it to fail silently like that, yes. Or else an actual error message popped up.
But now, it just fails silently :) 23:00
23:04 elmex joined 23:09 elmex left 23:13 pmurias left
sorear jnthn: kyleha writes tests too? 23:14
kid51 is preparing an introductory talk on Rakudo* 23:18
first audience may be a local Linux user group
the talk is intended to be more of an announcement than an explanation
kid51 would welcome feedback on slides: thenceforward.net/perl/talks/rakudo...index.html 23:19
(obviously adapted from pmichaud's YAPC talk, with his permission)
Tene kid51: the "Star" part also came from the usage of * in Perl 6 as "Whatever"
kid51 Tene: yes I think pmichaud mentioned that in that talk 23:20
but that's a bit too fine a point for this talk 23:21
Tene 'k
kid51 ... so fine that I don't recall it from pmichaud's talk
23:24 snarkyboojum joined
snarkyboojum kid51: on slide6 you mention Parrot and Pugs under the "Multiple implementations" bullet point. I have find that a bit misleading, perhaps Rakudo (on Parrot) and Pugs listed there? 23:25
kid51 Well, Rakudo is the Perl 6 implementation on the Parrot virtual machine ... correct? 23:26
You can have multiple implementations of Perl 6. Those are the two most prominent. 23:27
snarkyboojum yeah, just pointing out that on my reading you were implying that Parrot is an implementation of Perl 6 (which I realise isn't your intention)
23:27 justatheory joined
kid51 okay, I see what you're driving at; will fix 23:28
snarkyboojum s/you were/you seemed to be/
jnthn sorear: Yes :-)
sorear I knew he ran an automatic test crossposter 23:29
23:29 shade_ joined 23:31 justatheory left
jnthn No, he's written quite a lot of tests too :-) 23:31
kid51 snarkyboojum: fixed in the source code (though not on the web site) 23:33
23:34 colomon joined
kid51 afk & 23:34
jnthn o/ colomon :-) 23:38
ingy rakudo: my $h = {}; $h{'foo'} ||= 0; say $h.perl; 23:39
p6eval rakudo 62ee6d: OUTPUT«{"foo" => 0}␤»
ingy rakudo: my $h = {}; $h{'foo'}++ ||= 0; say $h.perl; 23:41
p6eval rakudo 62ee6d: ( no output ) 23:42
23:42 masonkramer joined
ingy rakudo: my $h = {}; $h{'foo'}++ ||= 1; say $h.perl; 23:42
p6eval rakudo 62ee6d: ( no output )
ingy rakudo: my $h = {}; $h{'foo'}++; say $h.perl;
p6eval rakudo 62ee6d: OUTPUT«{"foo" => 1}␤»
23:43 meppl left
ingy rakudo: my $h = {}; my $n = $h{'foo'}++; say $n; 23:43
p6eval rakudo 62ee6d: OUTPUT«Any()␤»
ingy rakudo: my $h = {}; my $n = ++$h{'foo'}; say $n;
p6eval rakudo 62ee6d: OUTPUT«1␤» 23:44
ingy rakudo: given 3 { when 4 { say "four" } } 23:52
p6eval rakudo 62ee6d: ( no output )
ingy rakudo: given 3 { when 3 { say "four" } }
p6eval rakudo 62ee6d: OUTPUT«four␤»
23:55 Psyche^ joined 23:59 Patterner left, Psyche^ is now known as Patterner

Powered by Cro and the Raku® Programming Language.

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

Thank you!