perl6-projects.org/ | nopaste: sial.org/pbot/perl6 | evalbot: 'perl6: say 3;' | irclog: irc.pugscode.org/ | UTF-8 is your friend!
Set by moritz_ on 4 May 2009.
00:16 justatheory left 00:36 alanhaggai joined, alanhaggai left, amoc joined 00:38 alanhaggai joined 01:00 abra left 01:15 DemoFreak left 01:29 Whiteknight left 01:45 ilogger2_ joined 02:14 eternaleye_ joined 02:36 mikehh_ joined 03:32 eternaleye joined 03:40 orafu joined 03:52 cognominal joined 03:56 meppuru joined 03:57 justatheory joined 04:05 LadyLuna1y joined 05:07 justatheory joined 06:01 meteorjay joined 06:05 M_o_C joined 06:07 xinming joined 06:09 dduncan joined 06:17 dduncan left 06:43 kate21de joined 07:01 r0bby joined 07:25 NoirSoldats joined 07:35 tombom joined 07:57 c9s joined 08:16 someguy joined 08:22 someguy left 08:39 ejs1 joined 08:47 Eevee joined 08:51 iblechbot joined
Tene There are lots of incorrect pages on perlfoundation.org/perl6/ 09:14
09:27 DemoFreak joined 09:48 meteorjay joined 10:03 bacek_ joined 10:11 ejs1 joined 10:28 fridim joined 10:30 pmurias joined 10:40 alanhaggai joined
pugs_svn r26749 | moritz++ | [t/spec] get rid of an instace of "is also" 10:43
moritz_ std: sub f { 3 } sub g { 3 } 10:50
p6eval std 26749: OUTPUT«##### PARSE FAILED #####␤Statements must be separated with semicolon at /tmp/8m2GakLY7m line 1:␤------> sub f { 3 } sub g { 3 }␤ expecting any of:␤ infix or meta-infix␤ infix stopper␤ standard stopper␤ statement␤ statement modifier loop␤
..terminator␤FAILED 00:02 3...
pugs_svn r26750 | moritz++ | [t/spec] get rid of "is also" in unspace.t 10:55
r26751 | moritz++ | [t/spec] get rid of some evals in unspace.t
literal std: rx/ <foo(1, 2, 3)> / 10:58
p6eval std 26749: OUTPUT«ok 00:03 37m␤»
literal std: rx/ <foo_bar-baz(1, 2, 3)> /
p6eval std 26749: OUTPUT«ok 00:02 37m␤»
literal std: rx/ <foo_bar-baz'quux(1, 2, 3)> / 10:59
p6eval std 26749: OUTPUT«ok 00:02 37m␤»
moritz_ std: [=>] <A B>, undef
p6eval std 26749: OUTPUT«ok 00:03 38m␤»
pugs_svn r26752 | moritz++ | [t/spec] correct and simplify chaining.t 11:00
dalek kudo: a74a9cf | moritz++ | t/spectest.data:
we pass S12-methods/chaining.t
 11:03
literal @tell TimToady STD.pm line 2234 doesn't make much sense. $> == $() ? 11:25
lambdabot Consider it noted.
11:28 Whiteknight joined
moritz_ it should probably be '$> variable' 11:33
literal yeah
pugs_svn r26753 | moritz++ | [STD.pm] typo found by literal++ 11:34
moritz_ forgiveness vs. permission ;-)
literal hehe 11:43
11:48 rindolf joined 11:49 abra joined 11:57 cognominal joined 12:11 LadyLunacy joined 12:16 kate21de joined 12:20 minazo joined 12:24 wknight8111 joined 12:27 Psyche^ joined 12:28 Psyche^ joined
literal std: 1 <=> 2 12:55
p6eval std 26753: OUTPUT«ok 00:02 35m␤»
literal rakudo: 1 <=> 2
p6eval rakudo a74a9c: ( no output )
literal rakudo: say (1 <=> 2)
p6eval rakudo a74a9c: OUTPUT«-1␤»
diakopter twimc: borrowing from Perl 6 - gavingrover.blogspot.com/2009/05/gr...admap.html 13:13
jnthn Aye, they directly mention Perl 6 too. :-) 13:18
13:37 masak joined 13:42 skids joined
masak Tene: is Loldispatch available online somewhere? 13:47
14:03 aindilis joined 14:08 hacktor joined 14:10 hacktor left 14:12 fuad joined
diakopter jnthn: right, that's what I meant. 14:13
pugs_svn r26754 | moritz++ | [t/spec] s/is also/augment (also in file name) 14:27
masak moritz_++ 14:28
moritz_ note that it was not s/.../.../g
masak no, it was s/.../... 14:29
:)
kinda open-ended.
moritz_ ;-)
14:32 sri_kraih joined
pugs_svn r26755 | moritz++ | [t/spec] s/is also/augment/ in open_closed.t 14:33
r26755 | moritz++ |
r26755 | moritz++ | Also get rid of lives_ok { ... } around class defintions; the class
r26755 | moritz++ | composition happen at compile time anyway, so that was pretty pointless.
moritz_ ok, 'is also' usage down to 5 files in t/spec, and they really need it (because it's what they test) 14:34
masak moritz_: how do you mean? 14:35
I thought 'is also' was deprecated.
moritz_ masak: I mean I removed all needless occurences of that feature, however it's called syntactically
masak still doesn't understand
I'd better greo for it myself to grok it :) 14:36
moritz_ masak: I didn't remove all occurences of the deprecated 'is also' yet because I don't want to leave that untested in Rakudo 14:37
masak ok.
moritz_ masak: I did s/is also/augment/ in some places that rakudo doesn't run anway...
but most of the occurences weren't necessary for running the tests, so I removed them (or replaced by something simpler) 14:38
masak excellent.
so those 'is also' that remain are those that Rakudo runs, and that can't be replaced by somethine else? 14:39
:q
oops, window focus fail.
I notice none of the test files that use 'is also'/'augment' do 'use MONKEY_PATCHING' 14:40
moritz_ yes, I'm going to change that soonish as well
masak moritz_: do you need anything from Rakudo to be able to go through with it?
14:51 nihiliad joined, ZuLuuuuuu joined
pmichaud fwiw, I'm find if we regress on the 'is also' tests. 14:53
*fine
15:05 payload joined
jnthn Aye. I think the use we make of is also in the setting will soon let us know if we broked it. :-0 15:07
s/0/)/
15:08 pmichaud joined
moritz_ masak: not really; I can simply fudge it out 15:11
masak moritz_: fair enough. but that means we go down in passed tests...
I thought there might be a good way to prepare for it. 15:12
moritz_ masak: sure, implemnt 'use MONKEY_PATCHING' and 'augment' ;-)
masak :)
15:17 justatheory joined 15:22 fuadaw joined
pugs_svn r26756 | moritz++ | [t/spec] rename is_also_instead.t and update syntax 15:32
15:33 ZuLuuuuuu left
pmichaud how many tests do we go down? 15:35
(approx) 15:36
10? 100?
moritz_ I'm removing all, then I'll run the harness and tell you
perhaps 50 to 100
pmichaud I'll still live with the regression.
we can always start to ship a MONKEY_PATCHING.pm module to get "use MONKEY_PATCHING" to not fail :-P 15:37
15:37 M_o_C joined
moritz_ that's something that can easily be fudge away 15:37
don't worry on that front
pmichaud that too.
15:37 fuadaw left
masak pmichaud: an actual module would be dependent on the cwd of the user, and/or PERL6LIB. 15:38
15:39 justlooks joined
pmichaud masak: I'm thinking we might end up with a standard lib-location for rakudo-distributed modules. 15:39
masak \o/
pugs_svn r26757 | moritz++ | [t/spec] remove rest of "is also" on classes
moritz_ pmichaud: first step, move Test.pm to lib/, IMHO
masak pmichaud: did I tell you about the time I tried to remove Test.pm from the November repo?
pmichaud I think you mentioned it, but I forget the punchline 15:40
masak turns out I couldn't bring myself to do it, because November's Test.pm is still objectively better than Rakudo's.
15:40 PhatEddy joined
moritz_ what's the difference? 15:40
diagnosis?
masak I don't recall exactly in which ways, but likely candidates are got/expected and is_deeply.
those are the two things we added. 15:41
pmichaud I think got/expected in in Rakudo's Test.pm now.
masak and they're deeply useful in our test suite.
moritz_ I'm pretty sure we backported is_deeply to Rakudo.
masak :)
moritz_ pmichaud: I don't think so
pmichaud: I tried to incorporate some of Ovid's patches, but had a few problems
pmichaud goes to look at November's Test.pm 15:42
afk for a bit 15:43
dalek kudo: cbc1bf6 | moritz++ | t/spectest.data:
regress on a few tests tests that now use "augment" instead of "is also"
 15:44
Tene ack 15:46
I just missed him
pmichaud back 15:47
15:48 Psyche^ joined
pmichaud From a cursory look, I think I'd be fine with adopting November's Test.pm in place of Rakudo's. 15:48
PhatEddy I've got a short regex program that hangs rakudo here: token chunk { <[a..z]>* }; say 'ok' if 'index/' ~~ /[ <chunk> '/'?]*/ 15:53
anything to do beyond mailing in a rakudobug? 15:54
moritz_ PhatEddy: it's known that PGE hangs on quantified zero width matches
15:56 eternaleye_ joined
pmichaud you might want 15:57
token chunk { <[a..z]>+ }; ... / [ <chunk> ** '/' ]? / 15:58
PhatEddy sorry - just didn't occur to me right off to search rt from PGE ... 15:59
moritz_ PhatEddy: it's known, but I don't know if it's in RT 16:04
16:07 baest joined, smtms joined, c1sung joined 16:08 pdc303 joined 16:09 broquaint joined, spx2 joined 16:10 krunen joined, Eevee joined 16:13 payload1 joined, H1N1[S] joined 16:18 cognominal joined 16:19 integral joined 16:26 cavelife^ joined 16:30 amoc joined 16:37 masak joined 16:56 mberends joined
mberends masak: another nice blog, about yarn :) 16:57
masak mberends: oh hai
thanks. :) 16:58
mberends does there need to be a yarn pumpkin as well as a proto pumpkin?
masak nah, not yet, methinks. 16:59
in fact, I'm not sure there "needs" to be a proto pumpkin... I just like the thought of the pumpkin enough to keep it up. :P
mberends it's quite worrisome sleeping without a pumpkin, not knowing whether the world will be the same as when I drifted off... 17:00
masak *lol* 17:01
jnthn
.oO( mmm....pumpkin soup )
 17:02
masak rolls pumpkin over on mberends' desk
mberends will use git pull --rebase more often in future
masak mberends: I meant to write that recursive-deps stuff today, but I was distracted by yaks.
Infinoid damn yaks 17:03
mberends I'm going to cage-clean HTTP::Daemon now 17:04
masak mberends++
17:09 mikehh_ joined
dalek kudo: e6363e8 | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 383 files, 11085 passing, 1 failing

   S32-io/IO-Socket-INET.t 2 - echo server and client		 17:14
17:15 knewt joined
s1n i think this is a parse error: 17:23
rakudo: class A { my $.foo; sub test() { say "oh hai!"; } }; my A $a .= new; $a.test();
p6eval rakudo e6363e: OUTPUT«Could not locate a method 'test' to invoke on class 'A'.␤current instr.: 'die' pc 16816 (src/builtins/control.pir:225)␤»
s1n notice the misplaced 'my' seems to screw with how the rest of the class is being parsed 17:24
any thoughts?
jnthn s1n: The error is completely correct. 17:25
A has no method test.
It has a sub test though. :-)
rakudo: class A { my $.foo; method test() { say "oh hai!"; } }; my A $a .= new; $a.test();
p6eval rakudo e6363e: OUTPUT«oh hai!␤»
s1n jnthn: i was getting that with a multi-method... 17:26
rakudo: class A { my $.foo; multi method test() { say "oh hai!"; } }; my A $a .= new; $a.test();
p6eval rakudo e6363e: OUTPUT«oh hai!␤»
s1n wtf
what exactly is the difference between sub and method? 17:27
moritz_ ok, 11074 tests still passing 17:28
s1n: a method has an invocant
in $a.test, $a is the invocant
jnthn Also, a method gets entered into the methods table and can be called as one. A sub does not. 17:29
s1n why doesn't p6 just dwim and add the invocant if i called it with one? :)
i really don't want to have to know that much, i just want to define a method on a class
jnthn Then define one!
Using the keyword designated for that purpose. :-)
moritz_ s1n: because perl 5 didn't do the distinction, and it was a mess 17:30
s1n: if you are forced to declare it, optimizations and introspections can work
s1n i would just like it to automatically know "hey, that's a method on the class, let's add the invocant" or "hey, that's a stand-alone method, don't add the invocant" 17:31
i already defined the class scope, why do i need to use a different keyword? 17:32
jnthn Because you can also, quite legitimately, put a sub in a class too.
s1n which do you think is more common?
jnthn Which is something that is called without an invocant.
In classes? methods. 17:33
s1n so... why not have dwim keyword which automatically figures it out, unless i want something unusual
unless method was intended for that 17:34
because i keep screwing it up and placing subs/methods incorrectly and spend hours debugging it
jnthn method is the keyword for "I'm defining something that can be invoked as a method on this class"
In a class, you probably usually will want to be writing methods.
s1n jnthn: the point is i already defined the class, everything inside it is _logically_ defined on the class by default, right? 17:35
otherwise, class is just a hackish way of doing packages from p5
i want a class, when i define 'class A { .. }' everything inside that closure should default to be defined against that class 17:36
jnthn A class is a type of package, just as a role, a module, a grammar, etc is.
s1n everytime i define a class, i constantly have to define the class contents within the contents of the class
jnthn Myself, I'd rather not having subs suddenly changing their meaning in different types of package. If something is different it's probably best off if it looks different. 17:37
s1n personally, i hate having to keep saying what's inside the class when i've already said what the class' scope is
jnthn Anyways, if it bugs you and others enough, it's only a proto routine_declarator:dwimcode away. :-)
s1n all i'm getting at is the class is already defined in the closure, everything in it _should_ belong to it by default 17:38
jnthn It does, but there's a difference between "belongs in the package of the class" and "belongs in the method table of the class". 17:39
Tene s1n: It's also useful to define class-scoped subs. You don't want those showing up as methods on the class.
s1n Tene: i know what class scoped methods are
Tene s1n: I said class-scoped *subs*. That aren't methods on the class.
class A { sub helper_function_one { ... } } 17:40
s1n i can't help but think java got this right, i have to add 'static' to define something against the class
Tene "Inside of a class declaration, 'sub' gets a new meaning, and we have a different keyword we have to use for what 'sub' used to mean." 17:41
Is a bad rule.
jnthn Java doesn't have a sub keyword.
s1n i think i've mistakenly used sub/method incorrectly about 10 times, and each time it takes me hours to figure out what i've done wrong
moritz_ jnthn: java doesn't have subs at all 17:42
s1n to me, the sub/method thing is unintuitive and i will always screw it up 17:43
jnthn moritz_: I know, I was more drawing out that there's no keyword at all in Java for declaring a chunk of code.
moritz_ that it's (IMHO bad) C++ heritage 17:44
s1n you can, but it has to either be static on a class, or an anon closure
(that is, you can do { code } in both languages) 17:45
you just can't call it, it's just called in whatever closure it's currently in
code { more code } even more code => just plows through each of those in order 17:46
moritz_ ie it's not a declaration, just a new scope 17:47
s1n moritz_: yeah, basically
Tene s1n: You really think that it would be a good idea for 'sub' to magically mean something different in classes, and require 'sub-yes-really-a-sub' to mean what 'sub' used to mean? You think that would be more intuitive and lead to fewer mistakes? 17:50
s1n Tene: i dunno, but i'm sure i can't be the only one screwing that up 17:51
Tene That's the first time I've heard anyone mention it. That doesn't necessarily mean much, though.
s1n it seems intuitive to me to have one keyword for all methods and one that says "this one is just a named closure and isn't defined on the class" 17:52
Tene My feeling is that it's not so much unintuitive as it is just a (needed, IMO) change from Perl 5.
s1n and everything else? 17:53
Tene s1n: 'sub' in a class means the same thing as 'sub' in a namespace and 'sub' anywhere else.
s1n i think i'm failing to express my point
Tene how is "'sub' means exactly the same thing everywhere" less intuitive than "'sub' means the same thing everywhere, except in class blocks, where you have to say 'really-sub' to mean what 'sub' means everywhere else"? 17:54
Hm. I think I already said that.
Or I'm failing to understand.
s1n class A { method-on-A; method-on-A; method-not-on-A; } versus class A { method; method; method-not-on-A } 17:55
Tene I have no idea what that means.
s1n i've already declared the class, everything in it is in the class unless i say otherwise
jnthn s1n: I think what you're missing is that there's different ways to be "in the class". 17:56
Tene a sub isn't a method, though!
s1n then why is it called "submethod"?
why not "code { }"
Tene erm, submethod is different from sub
jnthn s1n: As soon as you want sub to mean something else in there, you then have to go through all the other package types (role, grammar, etc) and say "so what does sub mean in there?"
s1n Tene: how does that make sense??
jnthn And then we get back to "memorizing lists of things".
Which is something Perl 6 tries to avoid. 17:57
s1n jnthn: i'm not sure i understood what you mean
jnthn In a world where people can declare their own package types and routine types in derived forms of Perl 6, I think keeping the meanings of routines and the meanings of packages orthogonal is probably a win, even if it's at the expense of some dwim-i-ness. 17:58
s1n we have method, sub, and submethod and they all mean something pretty different. no one else sees that as confusing?
jnthn No, just as I don't see "we have module, grammar, class, role all meaning something pretty different" as confusing either.
Tene Different names for different things. it would be a lot more confusing if we only had one name for all of those things, or all of those names for one thing.
exactly
jnthn needs to eat - bbiab 17:59
s1n okay, then i hope that works out for everyone else because they all sound the same to me and no one else seems to think it's even the slightest bit confusing
of course, you both have worked _on_ rakudo, so your views may be slanted 18:00
moritz_ s1n: casual users don't have to care about submethods
Tene s1n: do you actually have a proposal for how things should be instead? I might have missed it.
I came in a bit late.
s1n moritz_: how are casual users going to know the difference?
Tene i'm still not quite sure what your complaint is. 18:01
s1n Tene: i thought i did but i guess not
moritz_ s1n: the just read a tutorial which covers the basics, and it will deal with subs and/or methods. They'll use what they'll be taught.
it worked the same for Perl 5: the documentation uses some idioms, and people copied them
s1n moritz_: i would agree, which means people are less likely to use or understand sub/submethod 18:02
because when they define a class, just just want to define a function on the class
Tene s1n: can you explain "on the class"? 18:03
moritz_ they want to define a method on the class ;-)
and nearly all other languages call that thing "method" as well 18:04
s1n okay, lemme try with another example in another language
class A { void foo() { ... } }; #c++, foo is defined as a method on instances of A 18:05
requires no additional knowledge other than i put it inside A's definition
moritz_ s1n: at the cost of less flexibility 18:06
s1n moritz_: how so?
moritz_ s1n: you can't declare a function in a class in C++
s1n it defaults inside
18:07 smtms joined
s1n moritz_: class A { static void foo() { ... } }; 18:07
moritz_ s1n: that's a static method, not a function.
s1n wtf is the difference?
Tene that's a... what mroitz said
moritz_ s1n: how do you call a static method?
s1n A::foo()
moritz_ well, in Perl you call it as A.foo() 18:08
and a sub as A::foo()
or, if imported, simply as foo()
s1n so what does A.foo() mean? shouldn't it be $a.foo()?
moritz_ A.foo() means "call a method on class A" 18:09
s1n then what is A::foo()?
Tene s1n: the issue is that @larry feel strongly that keywords suddenly having different behaviors in different places is a very bad thing.
moritz_ that's calling a function in A's namespace
which is *not* a method call
Tene because in Perl 5 it has lead to a large list of arbitrary things to memorize. 18:10
moritz_ and thus determined by lexical scope and namespace
Tene instead of "This construct behaves this way", you had to memorize "This construct behaves in these three ways in these three different situations"
moritz_ a method call is always independent of namespace and scope
s1n forget it, it's not worth it 18:11
18:13 masak joined
Tene hi masak 18:13
masak: I added LolDispatch to web. it maybe needs a less silly name?
18:21 ejs1 joined
moritz_ is a bit annoyed that his blog still doesn't show up in the ironman feed, and mst doesn't respond to emails anymore 18:24
masak Tene: I think the name is very fitting. :) 18:25
Tene: but you should definitely blog about it, and about the pastebin.
s1n let's say i wanted to introduce or change some syntax (like a grammar) within the language, do i do that with protos? 18:34
moritz_ usually not 18:35
either with macros 18:36
or something like
augment slang Regex { token backslash:std<\Y> { YY }; }
that would introduce a new \Y token, which matches the literal YY 18:37
(very silly, but what I used in a test recently)
rakudo: rx:P5//
p6eval rakudo e6363e: OUTPUT«Colons cannot be used as delimiters in quoting constructs at line 1, near ":P5//"␤␤current instr.: 'parrot;PGE;Util;die' pc 129 (runtime/parrot/library/PGE/Util.pir:86)␤»
moritz_ rakudo: m:P5// 18:38
p6eval rakudo e6363e: OUTPUT«Colons cannot be used as delimiters in quoting constructs at line 1, near ":P5//"␤␤current instr.: 'parrot;PGE;Util;die' pc 129 (runtime/parrot/library/PGE/Util.pir:86)␤»
pugs_svn r26758 | moritz++ | [t/spec] improve fudging of capture.t; vim modeline 18:41
r26759 | moritz++ | [t/spec] avoid newlines in test descriptions
s1n moritz_: can i augment any keyword or STD expression? 18:46
moritz_ s1n: yes 18:47
s1n so i can do something like 'augment slang If { ... }'? 18:48
moritz_ no
s1n #i don't know what the syntax in STD is for 'if'
well, i was trying to say, can i augment a keyword like 'if', is that just not possible? 18:49
moritz_ augment slang MAIN { token statement_control:sym<if> { IFF } } 18:50
then you can write 'IFF 3 < 4 { say "yes" } '
s1n can i change IFF to be if with some other change? 18:51
that looks like i'm only allowed to alias things, i want to actually augment the existing tokens
moritz_ actually my example was bogus 18:52
s1n heh
moritz_ you'd neet do put in the regex to match a whole `if' statement, including condition, else, elsif blocks etc.
s1n even when you're just augmenting it?
moritz_ see STD.pm, lines 1256ff
s1n goes off to look 18:53
moritz_ what do you mean by "just augmenting"?
s1n well, augment, to me, means i'm just augmenting existing behavior in some way, not redefining it
moritz_ so you want to change the semantics, not the syntax? 18:54
s1n moritz_: both :)
moritz_ s1n: if you want to change both, you'd be better off with a macro
s1n are there spectests for macros?
moritz_ yes, a few
s1n Missing test file: t/spec/S12-class/is_also_instead.t 18:56
moritz_ uhm, did I forgot to push? 18:57
18:58 abra joined, fuad joined
moritz_ s1n: cbc1bf6140f9c31872b51ddf9c6a7161845c789c should have fixed that 18:58
s1n moritz_: ahh, okay, i checked that out yesterday, so i updated 18:59
rakudo: macro f { '1 + 1' }; f.say; 19:00
p6eval rakudo e6363e: OUTPUT«Could not find non-existent sub f␤current instr.: '_block14' pc 64 (EVAL_16:49)␤»
moritz_ rakudo doesn't do macros
s1n oh, bummer :(
moritz_ which is why there are so few tests
s1n okay, well, can i do this then: macro if { True }; if False { say "oh hai"; } 19:02
moritz_ no 19:03
if you tell me what you want to achieve, maybe I can help you better
19:04 jan___ joined
s1n moritz_: i want to alter the syntax and semantics of a keyword, let's start with the if statement 19:11
let's see, we can maybe add something like the comma operator from c?
(just an example to get me started)
moritz_ so what does the comma operator from C do? 19:12
s1n kinda like shortcuircuting, so you can have multiple clauses
bad way of doing && :)
moritz_ well, the comma is taken in Perl 6 already 19:13
so that's kinda hard
s1n so let's invent one
btw, it's not entirely like &&, just trying to give an example: en.wikipedia.org/wiki/Comma_operator 19:14
moritz_ let's say you want ,, to do the same thing that && does
then you could do something along these lines:
macro infix:<,,>($a, $b) { quasi {{{ $a && $b }}} } 19:15
s1n i could then do 'if True ,, False { say "never happens" }' 19:16
19:16 payload joined
moritz_ right 19:16
s1n but that's not modifying 'if', it's adding a new infix operator 19:17
i'm trying to think of something else we could do to modify sym:if... 19:18
19:18 M_o_C joined
s1n maybe always return True 19:18
maybe if is a bad example 19:19
maybe make 'loop' behave like 'while'?
(one Bool clause)
19:20 cspencer joined
moritz_ augment slang MAIN { token statment_control:loop is parse &statment_control:while { quasi {{{ while $/<xblock><EXPR>.compile { $/<xblock><pblock>.compile } }}} } } # perhaps something along these liens 19:22
dalek kudo: 496e1f0 | moritz++ | t/spectest.data:
a few more files for t/spectest.data
 19:25
19:26 jferrero joined 19:29 masak joined
masak does 'augment slang' require MONKEY_PATCHING, or something similar? 19:30
fuad masak:hello friend
moritz_ masak: dunno
masak fuad: greetings, fuad. also known as usr. :)
fuad yeah:) 19:31
im here again
moritz_ only knows urs
masak fuad: me too. :)
fuad: as you can tell, the network I use in the evenings is flaky.
19:31 [particle] joined
fuad hi moritz_ 19:32
moritz_ hi fuad ;-)
s1n moritz_: that's an interesting macro, i'd like to see how that works out :) 19:33
moritz_ s1n: me too ;-)
s1n i'm kinda interested in being able to do things like that, to see what i can get the lang to do :) 19:34
19:37 abra joined
masak s1n: re subs/methods, I've been using classes and methods for almost a year now, and I don't experience the issue you're describing -- accidentally combining class/sub, then trying to call the sub as a method and then spending hours searching for the problem. 19:38
s1n masak: then i must be an idiot
masak s1n: I suggest you examine whether this might not be a correct design in Perl 6 after all.
s1n: I didn't say that. 19:39
s1n masak: what do you mean examine if it's a correct design?
masak s1n: it might simply be that you have more Perl 5 experience than I do.
s1n: I mean that for me, it works very well.
s1n: I use methods and subs interchangably in classes, and I don't often confuse them.
s1n masak: well, i have more c/c++ experience than anything else, so that's where i'm coming from
masak ok. 19:40
jnthn masak: If you're using methods and subs interchangably, you might have issues. ;-)
masak jnthn: how so?
s1n masak: it's a matter of how every other language treats methods
masak jnthn: I can call subs from methods. no harm in that. 19:41
s1n: Perl 6 doesn't necessarily aspire to conform to how other languages do things. :)
s1n masak: right, and everyone else here seems bent on keeping the current behavior 19:42
jnthn masak: Ah, that was a sentence with more than one reading. :-)
I get the one you meant now.
masak s1n: including me. but I wouldn't describe it as 'bent on'. I just don't find a problem with it.
s1n my main point was that sub/method/submethod all sound the same and mean the same thing to me mentally. i just wanted to define a function in the class (instance function), i really didn't want to have to think "gee, what kind of function is this" because i already put it inside the class 19:44
masak jnthn: would it be possible to issue a warning message in s1n's confusion case, something like "did you mean to declare 'test' as a method?" ?
s1n: I see your point. 19:45
s1n: I use the distinction sub/method to my advantage within classes, so I'm not sure I would welcome your solution.
moritz_ once macros and syntax changes work, the obvious solution is to define your own Perl 6 slang
masak "there will be modules."
s1n well, if i put the class as a single line declarator, that makes sense to have to say, but i put the function _inside_ the class closure, it just made sense that it would figure it out for me 19:46
masak s1n: there's still some dissonance between your model and mine, it seems. 19:47
I cannot make sense of your last utterance.
s1n heh
masak "figure it out for me" -- please elaborate.
jnthn masak: The error message is already very clear. "There's no method called foo on class A." Well, that suggests I think quite clearly enough "go look through class a for method foo". :-)
*A 19:48
masak jnthn: agreed.
jnthn There's lots of Rakudo errors that aren't great, but that's one of the less bad ones, IMO.
19:48 explorer__ joined
fuad jnthn:Hi mate 19:48
masak jnthn: nod. 19:49
s1n when i think of code, i think in psuedocode, then translate. i thought "gee, i want to add a function to instances of this class", i accidentally translated that in perl6 to a 'sub' and i spent forever trying to find where i messed up on the function
jnthn fuad: hi :-)
s1n i put the sub inside the class but it didn't know that i meant method because it's inside class A{}
sub and method are too mentally similar to me and i messed it up
fuad jnthn:im yesterday`s guy which have used nickname usr,so from now im with this nickname;)
masak s1n: ok, so this is from the perspective of someone learning Perl 6, and that someone doesn't know subs from methods? 19:50
jnthn fuad: Ah, OK. :-)
s1n masak: this is from the perspective of someone wanting to put a function inside a class and perl6 not knowing that i will likely never mean sub and usually mean method
subs will probably go outside classes, methods inside 19:51
unless i've yet again misunderstood things
the only reason why i even discussed this was in the hopes that _someone_ might see my frustrations and say "hmm, is this the best way to do this?" 19:53
masak s1n: does your usage of the word 'function' include both subs and methods?
s1n: it's quite brave of you to enter a community in the hopes of changing its mind. :)
s1n masak: my usage of function is a closure, where i put it or explicit definition is what separates method/sub
masak s1n: ok, 'closure' and 'function' is another pair of terms which could create confusion in discussions between us. :) 19:54
s1n masak: not trying to change it's mind, i just want the best product that will be fun to use and will attract lots of users, anything i can do to achieve that, the better
masak: how about "a block of code that can be called from somewhere by name" heh
masak s1n: again, I'm one of those users. I'm happy with this aspect of Perl 6. you seem not to be. 19:55
s1n: that definition works.
s1n: it differs from my view of a closure as something ordinarily anonymous.
s1n masak: right, like i said, i seem to be the only one 19:56
masak: a closure is just that block without a name :)
masak s1n: yes. but a sub or a method in Perl 6 is not just a block with a name.
s1n i always think conceptually before i write anything, it allows me to work with _many_ languages on a regular basis :)
masak s1n: my only recommendation is to try using the 'method' keyword with an open mind. 19:57
s1n masak: all i can really do is provide feedback 19:58
masak s1n: sure. that's what #perl6 is for. 19:59
s1n i don't necessarily have a solution (Tene) or maybe even a really good explanation (moritz_), i'm doing the best i can with the mental converters for perl6 that i have :)
i hunted down the same mistake after a number of hours enough to want to say something
no one agrees, so i guess the world keeps turning 20:00
masak s1n: yes, I've often come into #perl6 with complaints having had similar experiences.
s1n: but not with this one.
s1n heh
masak s1n: it might be your unique C/C++ background that makes you turn down this particular garden path.
s1n quits climbing uphill
skids skips technical nitpicking on what "closure" actually means because he'd probably just get owned by a CS major :-) 20:01
masak skids: :)
skids: a closure is just a thunk. :P
skids thunk is an overloaded term even by CS standards. 20:02
s1n i have to think conceptually like this because i work with no less than 5 languages at work, perl6 for a hobby, and the graduate degree that has no language focus (only conceptual ideas with implementations in whatever)
at the very least, i'd like to be able to alter this behavior 20:05
20:06 LadyLuna1y joined
skids So you have no intention of ever using class-global subs (no invocant, but part of a class?) 20:07
s1n no i might, but i'd like something like 'static method foo' instead of 'sub foo' 20:08
but that's just a c/c++ weenie talking :)
that way all methods in a class are really instance methods for the class, and there is no sub :) 20:09
jnthn s1n: The other thing to remember is that if you declare a method foo you can't just call it as a subroutine foo() in Perl 6. Sub dispatch and method dispatch are different things. 20:10
s1n that way i can merrily do "class A { method foo{}; method bar{}; static method baz{} }" and mentally seperate baz out as something different 20:11
jnthn: i have no idea what you meant by that lol
masak mberends: so, are you using that pumkin right now?
20:11 FurnaceBoy joined
skids s1n: you could do method foo is export... just you'd always have to have the invocant as first arg. 20:12
s1n skids: sadly, i once again didn't understand that 20:14
skids method foo is export will also generate an exported sub in the namespace. multi method foo is export will let you do it from more than one class with the same method name. 20:15
jnthn s1n: class A { method foo() { }; method bar() { foo() } } # This call to foo() is an error, you need to call it as a method
s1n: class A { sub foo() { }; method bar() { foo() } } # This call to foo() works fine since you're using sub-calling syntax to call a sub
s1n jnthn: to me, that's a failure in design, because self should be implied
skids You don't always want self on helper functions. 20:16
jnthn s1n: It's just about avoiding confusion between method dispatch and sub dispatch.
s1n jnthn: still doesn't sound like a good idea 20:17
jnthn: you honestly think people aren't going to make that mistake enough to be put off by it?
jnthn s1n: The maning of "push @foo, 42" changing because your class happened to define a push method to me seems like a worse one.
*meaning
20:18 frew|work joined
s1n jnthn: i dunno, it seems unnatural that self is not implied 20:19
jnthn Only if you want to overvalue object orientation at the expense of other paradigms.
s1n jnthn: does perl6 have a full fledge object model or a second class object model? seems to me that method dispatch takes precedence 20:21
20:21 pmurias joined
jnthn I'm not even sure how you'd distinguish the two. 20:22
I think it's safe to say that Perl 6 is a multi-paradigm language, and that object orientation is one paradigm that it provides a lot of support for. But it's not the only one that matters either.
frew|work jnthn: couldn't we have both sub and method? 20:24
s1n it's okay that no one else agrees with me now, but the first thing i'll do with macros is obsolete sub and change method to behave as i mentally mean it to
jnthn frew|work: We do.
frew|work but method doesn't imply self?
jnthn frew|work: Yes, and also that the thing is callable as a method.
s1n frew|work: only if the invocant is specified
jnthn Which if you're calling it as self.foo() it is :-) 20:25
s1n even within the class itself, self has to be specified
frew|work but if I have this I'm good:
jnthn Right. If you're calling another method.
s1n again, bring on the macros :)
frew|work method(:$self, $foo) {}
jnthn frew|work: Did you mean for the colon to be after $self ?
frew|work can't remember invocant syntax 20:26
sure?
yeah
jnthn ($self: $foo)
frew|work that's not so bad. And it should be easy to do what s1n wants with a grammar extension right?
jnthn std: method(:$self, $foo) {}
p6eval std 26759: OUTPUT«##### PARSE FAILED #####␤Unexpected block in infix position (two terms in a row, or previous statement missing semicolon?) at /tmp/xqquTAUi16 line 1:␤------> method(:$self, $foo) {}␤ expecting any of:␤ infix or meta-infix␤ infix stopper␤ standard
..stopper␤ termina...
frew|work when that works 20:27
jnthn Ah, yes. Space needed if you're declaring an anonymous method.
std: method (:$self, $foo) {}
p6eval std 26759: OUTPUT«##### PARSE FAILED #####␤Can't put required parameter after variadic parameters at /tmp/4p6wHXhHnY line 1:␤------> method (:$self, $foo) {}␤ expecting any of:␤ parameter␤ post_constraint␤ trait␤ whitespace␤FAILED 00:02 36m␤»
jnthn std: method ($self:, $foo) {}
gah!
p6eval std 26759: OUTPUT«##### PARSE FAILED #####␤Unable to parse multisig; couldn't find final ')' at /tmp/Pqv2c0yv3f line 1:␤------> method ($self:, $foo) {}␤ expecting any of:␤ parameter␤ signature␤ whitespace␤FAILED 00:02 36m␤»
frew|work nice try.
jnthn lol, OK, I give up
20:27 szbalint joined
jnthn std: method ($self: $foo) {} 20:27
p6eval std 26759: OUTPUT«ok 00:02 36m␤»
jnthn phew
frew|work std: method ($self: $fo) {}
p6eval std 26759: OUTPUT«ok 00:02 36m␤»
20:28 antiphase joined
skids yeah, the colon, comma equivalence is going to take some habituating here too. 20:28
frew|work it makes sense
s1n rakudo: class A { method foo($self: $fo){ say "oh hai" }; method bar { foo; } }; my A $a .= new; $a.bar;
p6eval rakudo 496e1f: OUTPUT«too few arguments passed (0) - 2 params expected␤current instr.: 'parrot;A;foo' pc 173 (EVAL_23:82)␤» 20:29
s1n is that wrong?
jnthn s1n: Yes.
erm
well, the code is wrong, Rakudo is right.
pmurias s1n: btw what languages do you use at work?
skids It does make sense, it's just getting my autonomous spurious-punctuation-and-overparenthesization system trained for it. :-)
jnthn rakudo: class A { method foo($self: $fo){ say "oh hai" }; method bar { self.foo(42) } }; my A $a .= new; $a.bar; 20:30
s1n pmurias: c, c++, (c-tran even), perl, shell scripting, java, python
p6eval rakudo 496e1f: OUTPUT«oh hai␤»
s1n jnthn: i was wanting to not have to say self
jnthn rakudo: class A { sub foo($fo){ say "oh hai" }; method bar { foo(42) } }; my A $a .= new; $a.bar;
p6eval rakudo 496e1f: OUTPUT«oh hai␤»
jnthn Then write that second one.
But remember that declares a sub, not a method. 20:31
frew I think it would be cool if method defaulted the invokant into $self if one wasn't specified
pmurias what's the point?
jnthn frew: It's always available as "self". :-)
s1n frew: jnthn just recommended otherwise
pmurias you have self
frew didn't know that
what's the point of defining $self if self is available?
jnthn frew: Not much. :-) 20:32
frew: method foo(@list: $x) { ... } can be useful though if your invocant is an array for example.
But the scalar from is less useful.
pmurias method foo(&f) {...} 20:33
frew well yeah, but if you are building a regualr class...
skids It's for when you don't want to code in the first person :-)
jnthn ;-)
frew regularly doesn't code in first person, but that's javascript
prototypical OO, though weird, can be very sexy
pmurias s1n: if one is confusing methods and subs wouldn't having them under a joint keyword confuse him even more 20:34
?
frew agrees with pmurias 20:35
now that I know that self is always available I don't see why you want to do what you want to do s1n
skids method talks($he: $babble) # coded by extrovert ... method talk($you: $babble) # coded by dominatrix 20:36
frew does anyone know if we could have some kind of function case statement in p6? so that you could do something like this:
my $foo = given $foo { when 'bar' { 'baz' } }; 20:37
s1n pmurias: it wouldn't be under one keyword, the class-based method would need a modifier like multi
Tene got a basic blog app running in 20 mins on rakudo. :)
s1n frew: fyi, self isn't _always_ available, not in BUILD
pmurias class-based method = ? 20:38
Tene I heard masak did a speed-run blog app earlier, and I wanted to try.
frew but self doesn't really exist in BUILD yet, does it?
pmurias it should
frew: submethods are just methods you don't inherit 20:39
frew it should exist or self should work?
pmurias it exists
Tene will post when I get to my hotel.
s1n i'm going to just take a backseat on this argument, it's clear my opinion is at odds here, i'd rather just deal with it until i get macros
frew hopes macros don't end up being sketch like filters are in p5 20:41
I know they are supposed to be real macros, but still
s1n if they're anything like c++ macros that let me fundamentally change anything as i see fit, i'll be a happy camper :) 20:42
frew but shouldn't we theoretically be able to change syntax in p6 w/o a macro? 20:43
just modify the grammar right?
pmurias yes
frew which can be done in code
jnthn rakudo: my $tester = { given $^condition { when 42 { say "the answer" }; when 0 { say "nothing" } } }; $tester(42); $tester(0);
s1n yeah, but macros let you do it in a module without redistributing the perl6 STD grammar
p6eval rakudo 496e1f: OUTPUT«the answer␤nothing␤»
frew oooh
jnthn frew: See above in answer to your question.
rakudo: my &tester = { given $^condition { when 42 { say "the answer" }; when 0 { say "nothing" } } }; tester(42); tester(0); # may work too 20:44
p6eval rakudo 496e1f: OUTPUT«the answer␤nothing␤»
frew woo!
jnthn yay
frew awsome
jnthn That's a bit neater. :-)
frew this means I will be writing dispatch tables 90% less often!
jnthn frew: Don't forget there's multi-dispatch too. :-) 20:45
frew oh I know
and I miss it at work
jnthn In this case it's more code, though.
s1n actually, that was something i was looking for exactly for something else! thanks!
frew :-)
20:45 donaldh joined
frew I am writing a web based test thing for at work, and multi dispatch and that version of given/when make the code even neater 20:46
s/for at/for use at/
20:59 zamolxes joined 21:00 fridim joined 21:01 nemstep joined
moritz_ I've added enough new tests that the overall number of passing tests increased ;-) 21:13
11179
frew haha
fuad moritz_ im off to bed dude,have a nice time.Take care:)
frew how good do you think the coverage is for the language as a whole?
fuad jnthn:take care dude,im out:) 21:14
moritz_ frew: dunno, perhaps 60%
jnthn fuad: See you. :-) 21:15
moritz_: Nice! :-)
moritz_ but most of the coverage isn't really good 21:17
for example macros and object pipes are nearlly not tested, only a tiny bit
but that's really hard to do as long as no compiler implements them 21:18
patmat i wish everyone a good night!! 21:21
moritz_ good night
fuad g`night patmat
jnthn night
moritz_: Yeah, I agree on the "hard to test without a compiler getting close to attempting them" thing. 21:22
moritz_ jnthn: speaking of which, are there any new features you want to do soonish, and which need more tests?
jnthn moritz_: In theory, everything in news.perlfoundation.org/2009/04/hag...t_tra.html 21:23
;-)
moritz_ jnthn: yeah, I looked at it some time ago, and picked the item from the bottom ;-) 21:24
21:24 meppl joined
jnthn moritz_: There's this section called the project plan that gives an odering... ;-) 21:25
erm, ordering... :-) 21:26
moritz_: Traits tests are good though of course. :-)
D3 should be not too hard to write tests for.
But we may already have some.
moritz_ takes a look 21:27
jnthn D2 has some coverage.
pugs_svn r26760 | moritz++ | [t/spec] don't use outer lexicals in submethods.t 21:28
r26761 | moritz++ | [t/spec] fudge capture.t for rakudo
r26762 | moritz++ | [t/spec] bring new-symbols.t more in line with S02 and S05
r26763 | moritz++ | [t] update TASK, tests for unary [!=] etc
mberends general question, is it wrong that 'constant FOO =42;' in a module does not put FOO in the script using the module? Also tried adding 'is export', no dice. 21:30
jnthn mberends: It's wrong, and is export will be the way to do it. 21:31
moritz_ rakudo: module A { constant FOO is export(:DEFAULT) = 3 }; say FOO
p6eval rakudo 496e1f: OUTPUT«Null PMC access in isa()␤current instr.: 'parrot;List;!flatten' pc 6781 (src/classes/List.pir:215)␤»
jnthn But is export on a constant currently doesn't work.
moritz_ ticket!
21:31 eternaleye joined
jnthn oh noes where's masak?! 21:31
moritz_ where's masak when you need him? ;-)
mberends shall I report it? ... ok moritz_
moritz_ mberends: go ahead please
jnthn ...I love how "ticket" and "masak" are about synonyms now.
mberends masak's neighbour probably switched off their communal router. 21:32
moritz_ "can you masak it?" ;-)
mberends just a poor imitation, i'm afraid ;-)
skids notes small edit distance masak --> mistake --> ticket
jnthn Rakudo has so many masaks. 21:34
.oO( But I'm working to get rid of them. Mwaha. )
moritz_ does $obj.?method returns a list? 21:35
I know .+ and .* does
jnthn moritz_: Not AFAICR. 21:37
moritz_: Just undef or the result.
moritz_ jnthn: thanks 21:38
jnthn moritz_: Feel free to review my .WALK tests too.
moritz_ rakudo: say 3.3.int 21:44
p6eval rakudo 496e1f: OUTPUT«3␤»
pugs_svn r26764 | moritz++ | [t/spec] improve and extend parallel dispatch tests 21:51
moritz_ rakudo: use Test; plan 1; is_deeply((1, 2), [1, 2]) 21:54
p6eval rakudo 496e1f: OUTPUT«1..1␤ok 1 - ␤» 21:55
frew rakudo: use Test; plan 1; is_deeply([1, 2), [1, 2])
rakudo: use Test; plan 1; is_deeply([1, 2], [1, 2])
p6eval rakudo 496e1f: OUTPUT«Statement not terminated properly at line 1, near "([1, 2), ["␤␤current instr.: 'parrot;PGE;Util;die' pc 129 (runtime/parrot/library/PGE/Util.pir:86)␤» 21:56
rakudo 496e1f: OUTPUT«1..1␤ok 1 - ␤»
moritz_ I guess is_deeply has no way to distinguish these
frew rakudo: use Test; plan 1; is_deeply((1, 2), (1, 2))
p6eval rakudo 496e1f: OUTPUT«1..1␤ok 1 - ␤»
frew hm.
well isn't it supposed to be the same?
moritz_ not quite
if you assign it to a scalar, (...) is promoted to a Capture (which is immutable) 21:57
but [...] is a (mutable) Array
frew rakudo: my $foo = [1,2,3]; 21:58
p6eval rakudo 496e1f: ( no output )
frew why is that legal?
rakudo: my $foo = (1,2,3);
moritz_ why shouldn't it be?
p6eval rakudo 496e1f: ( no output )
frew well, I thought we didn't have references like we did in p5, so I figure that $foo = [1,2,3] kinda kills the point of sigils
moritz_ actually it's the other way round: we don't need Perl5esq references because everything is an object anyway 21:59
you can still use @a and %a to enforce particular types
you can also write $a = [1, 2, 3]; is Perl 5 22:00
frew rakudo: my $foo = [1,2,3]; $foo[1];
p6eval rakudo 496e1f: ( no output )
frew rakudo: my $foo = [1,2,3]; say $foo[1];
p6eval rakudo 496e1f: OUTPUT«2␤»
frew so all @ really does is force it to be an array 22:01
moritz_ it still does a little more
for example 'my @a' is not only a type constraint
but it instantiates an array object
'my Array $x; $x.push(3)' wouldn't work, because there's no Array object in $x yet 22:02
frew weird...
moritz_ (unless by some magic autovivification; but it sure wouldn't work in the general case)
frew so $foo = [1,2,4] is like... a lightweight version 22:03
moritz_ lightweight compared to what?
frew I dunno, the real object?
moritz_ the object in $foo is just a real array object like the one in @foo
(when you do my @foo = 1, 2, 3;'
frew but not yet instantiated? 22:04
moritz_ s/'/)/
sure it's instantiated
frew oooh
sorry
I didn't read what you said carefully enough
that makes sense 22:05
so my Array $a .= new; is equiv (more or less) to my @a;
moritz_ more or less, yes ;-) 22:06
there are still other effects that I forgot
for example @a = ... imposes list context
$a = ... imposes item context, even if there's an Array stored in $x 22:07
frew this making () and [] unnecessary
mberends oops, rakudobug sent, forgot to prefix subject with [BUG]
frew but ultimately this means that @ is much more sugar than it was before, correct?
moritz_ yes 22:08
rakudo: say (1, 2, 3).perl; say (1, [2, 3]).perl
p6eval rakudo 496e1f: OUTPUT«[1, 2, 3]␤[1, [2, 3]]␤»
moritz_ # not quite unnecessary ;-)
you *can* write Perl 6 like PHP, with only the $ sigil 22:09
but it's not idiomatic, and not recommended
frew what does that even mean? 22:10
frew doesn't know PHP
moritz_ PHP uses $ on all variables
frew oooh
moritz_ they have a bit less syntax, and try to make up for it with more than 3000 functions in the default namespace ;-) 22:11
frew yeah, I knew about that
moritz_ but ridiculing PHP is no fun; it's too easy 22:12
frew I just think that something annoying could happen that could confuse people ( sub foo { return [1,2,3] }; my $bar = foo )
is there a way to avoid that?
moritz_ yes, return (1, 2, 3) instead ;-) 22:13
frew will that give an error?
moritz_ no
frew but [] will give an error? 22:14
rakudo: sub frew { return [1,2,3]; } my $station = frew; say $station; 22:15
p6eval rakudo 496e1f: OUTPUT«Statement not terminated properly at line 1, near "my $statio"␤␤current instr.: 'parrot;PGE;Util;die' pc 129 (runtime/parrot/library/PGE/Util.pir:86)␤»
frew rakudo: sub frew { return [1,2,3]; }; my $station = frew; say $station;
p6eval rakudo 496e1f: OUTPUT«1 2 3␤»
frew rakudo: sub frew { return (1,2,3); }; my $station = frew; say $station; 22:16
p6eval rakudo 496e1f: OUTPUT«1 2 3␤»
frew isn't sure what the difference is
pugs_svn r26765 | moritz++ | [t/spec] set a few svn props 22:17
r26765 | moritz++ |
r26765 | moritz++ | (git-svn)-- for not handling them, and svn-- for not having a sane default
moritz_ rakudo: sub a { (1, 2, 3) }; sub b { [1, 2, 3] }; for a { print "a" }; for b { print "b" };
p6eval rakudo 496e1f: OUTPUT«Could not find non-existent sub for␤current instr.: '_block14' pc 90 (EVAL_17:59)␤»
moritz_ rakudo: sub a { (1, 2, 3) }; sub b { [1, 2, 3] }; for a() { print "a" }; for b() { print "b" };
p6eval rakudo 496e1f: OUTPUT«aaab»
moritz_ frew: that's the difference
frew weird... 22:18
so why doesn't for iterate over the list?
because [] means list right?
moritz_ because [] means "a scalar holding an Array"
and () means "a list" 22:19
frew ok
got it.
moritz_ believe me, that has been confusing to all of us.
frew it's a little strange to my perl5 shaped brain, but one day I'll get used to it and maybe see the benefit
do you grok it and dig it now?
moritz_ the basics 22:20
jnthn [ 1, 2, 3 ] gives you a reference to an array. (1, 2, 3) gives you a list that will flatten. 22:23
moritz_ don't mention references, you evil thing!
frew I'm just gonna stop thinking about it.
jnthn moritz_: Yeah, I know, but it's how it looks underneath. :-P
22:24 sparc joined
moritz_ the problem with thinking in references is this: 22:24
rakudo: sub a(@b) { print "b" for @b }; a([1, 2, 3])
p6eval rakudo 496e1f: OUTPUT«bbb»
moritz_ if we pass a reference in, where does the dereferncing happen? 22:25
jnthn It doesn't, it's just a case of binding. 22:26
moritz_ so binding dereferences? ;-)
frew zomg 22:27
frew is still confused
don't explain to me
I'm tired of thinking about it really
jnthn frew: It Just Works. ;-) 22:28
moritz_: Perhaps best to think of it is
frew I just think that putting an array in a scalar is confusing and maybe shouldn't work.... It just doesn't seem like something I'd ever want to do
jnthn rakudo: my @a = [1,2,3]; for @a { .say } 22:29
p6eval rakudo 496e1f: OUTPUT«1 2 3␤»
jnthn rakudo: my @a := [1,2,3]; for @a { .say }
p6eval rakudo 496e1f: OUTPUT«1 2 3␤»
jnthn So it's not about binding.
But note that the @ on the LHS puts the RHS in list context. 22:30
A [...] doesn't interpolate into a list.
mberends rakudo: my $a = [1,2,3]; for $a { .say }
jnthn However, when you pass it as an argument, there's no list context stuff going on there. So the array reference just gets bound.
p6eval rakudo 496e1f: OUTPUT«1 2 3␤»
jnthn gah, I said reference again! 22:31
jnthn thinks we should write the dictionary of Perl 6 swear words.
"length" :-O
"reference" :-O
frew good idea
jnthn It wasn't mine. :-)
.oO( if in dobut, disclaim all responsibility )
 22:32
*doubt
skids
.oO(if in donut, eat way out)
 22:50
dalek kudo: a27bbb6 | moritz++ | build/PARROT_REVISION:
bump parrot revision to get Unicode identifiers in named arguments, chromatic++
jnthn chromatic++ :-)
rakudo: sub мат(:$добре) { say $добре ?? "ГРОВНО!" !! "ЕБАТ" }; мат(:добре); мат(:!добре); 22:55
p6eval rakudo 496e1f: OUTPUT«error:imcc:syntax error, unexpected USTRINGC, expecting STRINGC ('unicode:"\u0434\u043e\u0431\u0440\u0435"')␤ in file 'EVAL_17' line 53␤error:imcc:syntax error, unexpected USTRINGC, expecting STRINGC ('unicode:"\u0434\u043e\u0431\u0440\u0435"')␤ in file 'EVAL_17' line
..56␤error:im...
jnthn Oh, too soon... 22:56
23:11 eternaleye joined 23:50 wknight8111 joined 23:55 japhb__ joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!