summaryrefslogtreecommitdiffhomepage
diff options
context:
space:
mode:
-rw-r--r--perl6.html.markdown915
1 files changed, 915 insertions, 0 deletions
diff --git a/perl6.html.markdown b/perl6.html.markdown
new file mode 100644
index 00000000..599f207c
--- /dev/null
+++ b/perl6.html.markdown
@@ -0,0 +1,915 @@
+---
+name: perl6
+category: language
+language: perl6
+filename: learnperl6.pl
+contributors:
+ - ["Nami-Doc", "http://github.com/Nami-Doc"]
+---
+
+Perl 6 is a highly capable, feature-rich programming language made for the upcoming hundred years.
+
+Perl 6 runs on [the Parrot VM](http://parrot.org/), the JVM and [the MoarVM](http://moarvm.com).
+
+Meta-note : the triple pound signs are here to denote headlines, double paragraphs, single notes.
+`#=>` represents the output of a command.
+
+```perl
+# Single line comment start with a pound
+
+#`(
+ Multiline comments use #` and a quoting construct. (), [], {}, 「」, etc, will work.
+)
+
+### Variables
+
+# In Perl 6, you declare a lexical variable using `my`
+
+# Perl 6 has 4 variable types :
+
+## - Scalars. They represent a single value. They start with a `$`
+
+my $str = 'String';
+my $str2 = "String"; # double quotes allow for interpolation
+
+# variable names can contain but not end with simple quotes and dashes, and can contain (and end with) underscores :
+# my $weird'variable-name_ = 5; # works !
+
+my $bool = True; # `True` and `False` are Perl 6's boolean
+my $inverse = !$bool; # You can invert a bool with the prefix `!` operator
+my $forced-bool = so $str; # And you can use the prefix `so` operator which turns its operand into a Bool
+
+## - Arrays. They represent multiple values. They start with `@`
+
+my @array = 1, 2, 3;
+my @array = 'a', 'b', 'c';
+# equivalent to :
+my @array = <a b c>; # array of words, delimited by space. similar to perl5's qw, or Ruby's %w
+
+say @array[2]; # Array indices start at 0 -- This is the third element
+
+say "Interpolate an array using [] : @array[]"; #=> Interpolate an array using [] : a b c
+
+## - Hashes. Key-Value Pairs.
+# Hashes are actually arrays of Pairs (`Key => Value`), "flattened" to remove duplicated keys.
+
+my %hash = 1 => 2,
+ 3 => 4;
+my %hash = autoquoted => "key", # keys are auto-quoted
+ "some other" => "value", # trailing commas are okay
+ ;
+my %hash = <key1 value1 key2 value2>; # you can also create a hash from an even-numbered array
+my %hash = key1 => 'value1', key2 => 'value2'; # same as this
+
+# You can also use the "colon pair" syntax: (especially handy for named parameters that you'll see later)
+my %hash = :w(1), # equivalent to `w => 1`
+ # this is useful for the `True` shortcut:
+ :truey, # equivalent to `:truey(True)`, or `truey => True`
+ # and for the `False` one:
+ :!falsey, # equivalent to `:falsey(False)`, or `falsey => False`
+ ;
+
+say %hash{'key1'}; # You can use {} to get the value from a key
+say %hash<key2>; # if it's a string, you can actually use <>
+
+## - Subs (subroutines, or functions in most other languages). Stored in variable, they use `&`
+sub say-hello { say "Hello, world" }
+
+sub say-hello-to(Str $name) { # you can provide the type of an argument
+ # and it'll be checked at compile-time
+
+ say "Hello, $name !";
+}
+
+# since you can omit parenthesis to call a function with no arguments, you need to use `&` also to capture `say-hello`
+my &s = &say-hello;
+my &other-s = sub { say "anonymous function !" }
+
+# A sub can have a "slurpy" parameter, or "doesn't-matter-how-many"
+sub as-many($head, *@rest) { # the `*@` slurpy will basically "take everything else".
+ # Note: you can have parameters *before* (like here) a slurpy one, but not *after*.
+ say @rest.join(' / ') ~ " !";
+}
+say as-many('Happy', 'Happy', 'Birthday'); #=> Happy Birthday !
+ # Note that the splat did not consume the parameter before.
+
+## You can call a function with an array using the "argument list flattening" operator `|`
+# (it's not actually the only feature of the operator, but it's one of them)
+sub concat3($a, $b, $c) {
+ say "$a, $b, $c";
+}
+concat3(|@array); #=> a, b, c
+ # `@array` got "flattened" as a part of the argument list
+
+## It can also have optional arguments:
+sub with-optional($arg?) { # the "?" marks the argument optional
+ say "I might return `(Any)` if I don't have an argument passed, or I'll return my argument";
+ $arg;
+}
+with-optional; # returns Any
+with-optional(); # returns Any
+with-optional(1); # returns 1
+
+## You can also give them a default value when they're not passed:
+sub hello-to($name = "World") {
+ say "Hello, $name !";
+}
+hello-to; #=> Hello, World !
+hello-to(); #=> Hello, World !
+hello-to('You'); #=> Hello, You !
+
+## You can also, by using a syntax akin to the one of hashes (yay unification !),
+## pass *named* arguments to a `sub`.
+sub with-named($normal-arg, :$named) {
+ say $normal-arg + $named;
+}
+with-named(1, named => 6); #=> 7
+with-named(2, :named(5)); #=> 7
+with-named(3, :4named); #=> 7
+ # (special colon pair syntax for numbers, mainly useful for `:2nd` etc)
+
+with-named(3); # warns, because we tried to use the undefined $named
+ # in a `+`: by default, named arguments are *optional*
+
+# To make a named argument mandatory, you can use `?`'s inverse, `!`
+sub with-mandatory-named(:$str!) {
+ say "$named !";
+}
+with-mandatory-named(str => "My String"); #=> My String
+with-mandatory-named; # run time error: "Required named parameter not passed"
+with-mandatory-named(3); # run time error: "Too many positional parameters passed"
+
+## If a sub takes a named boolean argument ...
+sub takes-a-bool($name, :$bool) {
+ say "$name takes $bool";
+}
+# ... you can use the same "short boolean" hash syntax:
+takes-a-bool('config', :bool); # config takes True
+takes-a-bool('config', :!bool); # config takes False
+# or you can use the "adverb" form:
+takes-a-bool('config'):bool; #=> config takes True
+takes-a-bool('config'):!bool; #=> config takes False
+# You'll learn to love (or maybe hate, eh) that syntax later.
+
+
+## You can also provide your named arguments with defaults:
+sub named-def(:$def = 5) {
+ say $def;
+}
+named-def; #=> 5
+named-def(:10def); #=> 10
+named-def(def => 15); #=> 15
+
+### Containers
+# In Perl 6, values are actually stored in "containers".
+# the assignment operator asks the container on the left to store the value on its right
+# When passed around, containers are marked as immutable. Which means that, in a function,
+# you'll get an error if you try to mutate one of your argument.
+# If you really need to, you can ask for a mutable container using `is rw` :
+sub mutate($n is rw) {
+ $n++;
+ say "\$n is now $n !";
+}
+
+# If what you want is a copy instead, use `is copy`.
+
+# A sub itself returns a container, which means it can be marked as rw :
+my $x = 42;
+sub mod() is rw { $x }
+mod() = 52; # in this case, the parentheses are mandatory
+say $x; #=> 52
+
+
+### Control Flow Structures
+
+# You don't need to put parenthesis around the condition, but that also means you always have to use brackets (`{ }`) for their body :
+
+## Conditionals
+
+# - `if`
+if True {
+ say "It's true !";
+}
+
+unless False {
+ say "It's not false !";
+}
+
+# You can also use their postfix versions, with the keyword after:
+say "Quite truthy" if True;
+
+# if (true) say; # This doesn't work !
+
+# - Ternary conditional, "?? !!" (like `x ? y : z` in some other languages)
+my $a = $condition ?? $value-if-true !! $value-if-false;
+
+# - `given`-`when` looks like other languages `switch`, but it's much more powerful thanks to smart matching.
+# given just puts its argument into `$_` (like a block), and `when` uses it using the "smart matching" operator.
+given "foo bar" {
+ when /foo/ { # you'll read about the smart-matching operator below -- just know `when` uses it
+ say "Yay !";
+ }
+ when $_.chars > 50 { # smart matching anything with True gives True, so you can also put "normal" conditionals
+ say "Quite a long string !";
+ }
+ default { # same as `when *` (using the Whatever Star)
+ say "Something else"
+ }
+}
+
+## Looping constructs
+
+# - `loop` is an infinite loop if you don't pass it arguments, but can also be a c-style `for` :
+loop {
+ say "This is an infinite loop !";
+ last; # last breaks out of the loop, like the `break` keyword in other languages
+}
+
+loop (my $i = 0; $i < 5; $i++) {
+ next if $i == 3; # `next` skips to the next iteration, like `continue` in other languages.
+ # Notice that you can also use postfix conditionals, loops, etc.
+ say "This is a C-style for loop !";
+}
+
+# - `for` - Passes through an array
+for @array -> $variable {
+ say "I've found $variable !";
+}
+
+# default variable is $_ (like a block)
+for @array {
+ say "I've got $_";
+}
+
+for @array {
+ next if $_ == 3; # you can skip to the next iteration (like `continue` in C-like languages)
+ redo if $_ == 4; # you can re-do the iteration, keeping the same topic variable (`$_`)
+ last if $_ == 5; # you can also break out of a loop (like `break` in C-like languages)
+}
+
+# Note - the "lambda" `->` syntax isn't reserved to `for` :
+if long-computation() -> $result {
+ say "The result is $result";
+}
+
+### Operators
+
+## Since Perl languages are very much operator-based languages
+## Perl 6 operators are actually just funny-looking subroutines, in syntactic categories,
+## like infix:<+> (addition) or prefix:<!> (bool not)
+
+## The categories are :
+# - "prefix" : before (like `!` in `!True`).
+# - "postfix" : after (like `++` in `$a++`).
+# - "infix" : in between (like `*` in `4 * 3`).
+# - "circumfix" : around (like `[`-`]` in `[1, 2]`).
+# - "post-circumfix" : around, after another term (like `{`-`}` in `%hash{'key'}`)
+
+## The associativity and precedence list are explained below.
+
+# Alright, you're set to go !
+
+## * Equality Checking
+
+# - `==` is numeric comparison
+3 == 4; # False
+3 != 4; # True
+
+# - `eq` is string comparison
+'a' eq 'b';
+'a' ne 'b'; # not equal
+'a' !eq 'b'; # same as above
+
+# - `eqv` is canonical equivalence (or "deep equality")
+(1, 2) eqv (1, 3);
+
+# - `~~` is smart matching
+# for a complete combinations list, use this table : http://perlcabal.org/syn/S03.html#Smart_matching
+'a' ~~ /a/; # true if matches regexp
+'key' ~~ %hash; # true if key exists in hash
+$arg ~~ &bool-returning-function; # true if the function, passed `$arg` as an argument, returns True
+1 ~~ Int; # "is of type"
+1 ~~ True; # smart-matching against a boolean always returns that boolean (and will warn).
+
+# - `===` is value identity and uses `.WHICH` on the objects to compare them
+# - `=:=` is container identity and uses `VAR()` on the objects to compare them
+
+# You also, of course, have `<`, `<=`, `>`, `>=`.
+# Their string equivalent are also avaiable : `lt`, `le`, `gt`, `ge`.
+3 > 4;
+
+## * Range constructors
+3 .. 7; # 3 to 7, both included
+# `^` on either side them exclusive on that side :
+3 ^..^ 7; # 3 to 7, not included (basically `4 .. 6`)
+# this also works as a shortcut for `0..^N`
+^10; # means 0..^10
+
+# This also allows us to demonstrate that Perl 6 has lazy arrays, using the Whatever Star :
+my @array = 1..*; # 1 to Infinite !
+say @array[^10]; # you can pass arrays as subscripts and it'll return an array of results
+ # this will print "1 2 3 4 5 6 7 8 9 10" (and not run out of memory !)
+# Note : when reading an infinite list, Perl 6 will "reify" the elements it needs, then keep them in memory
+# They won't be calculated more than once.
+
+# Warning, though: if you try this example in the REPL and juste put `1..*`,
+# Perl 6 will be forced to try and evaluate the whole array (to print it),
+# so you'll end with an infinite loop.
+
+## * And, Or
+3 && 4; # True. Calls `.Bool` on `3`
+0 || False; # False. Calls `.Bool` on `0`
+
+## Short-circuit (and tight) versions of the above
+$a && $b && $c; # returns the first argument that evaluates to False, or the last argument
+$a || $b;
+
+### More on subs !
+
+## There's more to come, but we're going to end this paragraph with a really powerful feature:
+## Unpacking ! It's the ability to "extract" arrays and keys. It'll work in `my`s and parameters.
+my ($a, $b) = 1, 2;
+say $a; #=> 1
+my ($, $, $c) = 1, 2, 3; # keep the non-interesting anonymous
+say $c; #=> 3
+
+my ($head, *@tail) = 1, 2, 3; # Yes, it's the same as with "slurpy subs"
+my (*@small) = 1;
+
+sub foo(@array [$fst, $snd]) {
+ say "My first is $fst, my second is $snd ! All in all, I'm @array[]."; # (remember the `[]` to interpolate the array)
+}
+foo(@tail); #=> My first is 2, my second is 3 ! All in all, I'm 1 2
+
+
+# If you're not using the array itself, you can also keep it anonymous, much like a scalar:
+sub first-of-array(@ [$fst]) { $fst }
+first-of-array(@small); #=> 1
+first-of-array(@tail); # errors with "Too many positional parameters passed" (the array is too big)
+
+# You can also use a slurp ...
+sub slurp-in-array(@ [$fst, *@rest]) { # you could decide to keep `*@rest` anonymous
+ say $fst + @rest.elems;
+}
+slurp-in-array(@tail); #=> 3
+
+# You could even extract on a slurpy (but it's pretty useless ;-).)
+sub fst(*@ [$fst]) { # or simply : `sub fst($fst) { ... }`
+ say $fst;
+}
+fst(1); #=> 1
+fst(1, 2); # errors with "Too many positional parameters passed"
+
+# You can also destructure hashes (and classes, which you'll learn about later !)
+sub key-of(% (:value($val), :qua($qua))) {
+ say "Got val $val, $qua times.";
+}
+
+# Then call it with a hash: (you need to keep the brackets for it to be a hash)
+key-of({value => 1});
+#key-of(%hash); # the same (for an equivalent `%hash`)
+
+## The last expression of a sub is returned automatically (though you may use the `return` keyword, of course):
+sub next-index($n) {
+ $n + 1;
+}
+my $new-n = next-index(3); # $new-n is now 4
+# This is true for everything, except for the looping constructs (due to performance reasons):
+# there's no purpose in building a list if we're just going to discard all the results.
+# If you still want to build one, you can use the `do` prefix: (or the `gather` prefix, which we'll see later)
+sub list-of($n) {
+ do for ^$n { # note the use of the range-to prefix operator `^` (`0..^N`)
+ $_ # current loop iteration
+ }
+}
+my @list3 = list-of(3); #=> (0, 1, 2)
+
+## You can create a lambda with `-> {}` ("pointy block") or `{}` ("block")
+my &lambda = -> $argument { "The argument passed to this lambda is $argument" }
+# `-> {}` and `{}` are pretty much the same thing, except that the former can take arguments,
+# and that the latter can be mistaken as a hash by the parser.
+
+# We can, for example, add 3 to each value of an array using map:
+my @arrayplus3 = map({ $_ + 3 }, @array); # $_ is the implicit argument
+
+# a sub (`sub {}`) has different semantics than a block (`{}` or `-> {}`) :
+# a block doesn't have a function context (though it can have arguments), which means that if you
+# return from it, you're going to return from the parent function, compare:
+sub is-in(@array, $elem) {
+ # this will `return` out of `is-in` sub
+ # once the condition evaluated to True, the loop won't be run anymore
+ map({ return True if $_ == $elem }, @array);
+}
+sub truthy-array(@array) {
+ # this will produce an array of `True` and `False` :
+ # (you can also say `anon sub` for "anonymous subroutine")
+ map(sub { if $_ { return True } else { return False } }, @array); # returns the correct value, even in a `if`
+}
+
+# You can also use the "whatever star" to create an anonymous function
+# (it'll stop at the furthest operator in the current expression)
+my @arrayplus3 = map(*+3, @array); # `*+3` is the same as `{ $_ + 3 }`
+my @arrayplus3 = map(*+*+3, @array); # also works. Same as `-> $a, $b { $a + $b + 3 }`
+say (*/2)(4); #=> 2
+ # Immediatly execute the function Whatever created.
+say ((*+3)/5)(5); #=> 1.6
+ # works even in parens !
+
+# but if you need to have more than one argument (`$_`) in a block (without wanting to resort to `-> {}`),
+# you can also use the implicit argument syntax, `$^` :
+map({ $^a + $^b + 3 }, @array); # same as the above
+
+# Note : those are sorted lexicographically. `{ $^b / $^a }` is like `-> $a, b { $ b / $a }`
+
+## Multiple Dispatch
+# Perl 6 can decide which variant of a `sub` to call based on the type of the arguments,
+# or on arbitrary preconditions, like with a type or a `where`:
+
+# with types
+multi sub sayit(Int $n) { # note the `multi` keyword here
+ say "Number: $n";
+}
+multi sayit(Str $s) } # the `sub` is the default
+ say "String: $s";
+}
+sayit("foo"); # prints "String: foo"
+sayit(True); # fails at *compile time* with "calling 'sayit' will never work with arguments of types ..."
+
+# with arbitrary precondition:
+multi is-big(Int $n where * > 50) { "Yes !" } # using a closure
+multi is-big(Int $ where 10..50) { "Quite." } # this uses smart-matching (could use a regexp, etc)
+multi is-big(Int $) { "No" }
+
+# you can also name these checks, by creating "subsets":
+subset Even of Int where * %% 2;
+
+multi odd-or-even(Even) { "Even" } # the main case using the type. We don't name the argument
+multi odd-or-even($) { "Odd" } # "else"
+
+# You can even dispatch based on a positional's argument presence !
+multi with-or-without-you(:$with!) { # make it mandatory to be able to dispatch against it
+ say "I can live ! Actually, I can't.";
+}
+multi with-or-without-you {
+ say "Definitely can't live.";
+}
+# This is very, very useful for many purposes, like `MAIN` subs (covered later),
+# and even the language itself is using it in several places.
+# `is`, for example, is actually a `multi sub` named `trait_mod:<is>`, and it works off that.
+# `is rw`, for example, is a dispatch to a function with this signature:
+# sub trait_mod:<is>(Routine $r, :$rw!) {}
+# (commented because running this would probably lead to some very surprising side-effects !)
+
+
+### Scoping
+# In Perl 6, contrarily to many scripting languages (Python, Ruby, PHP, for example),
+# you are to declare your variables before using them. You already saw it, with `my`.
+# (there are other declarator keywords, like `our`, `has` and `state`, but we'll talk about them later)
+# This is called "lexical scoping", where in inner blocks, you can access variables from outer blocks.
+my $foo = 'Foo';
+sub foo {
+ my $bar = 'Bar';
+ sub bar {
+ say "$foo $bar";
+ }
+ &bar; # return the function
+}
+foo()(); #=> 'Foo Bar'
+
+# As you can see, `$foo` and `$bar` were captured.
+# But if we were to try and use `$bar` outside of `foo`, the variable would be undefined.
+# (and you'd get a compile time error)
+
+# Perl 6 has another kind of scope : dynamic scope.
+# They use the twigil (composed sigil) `*` to mark dynamically-scoped variables:
+my $*a = 1;
+# Dyamically-scoped variables depend on the current call stack, instead of the current block stack.
+sub foo {
+ my $*foo = 1;
+ bar(); # call `bar` in-place
+}
+sub bar {
+ say $*foo; # Perl 6 will look into the call stack instead, and find `foo`'s `$*a`,
+ # even though the blocks aren't nested (they're call-nested).
+ #=> 1
+}
+
+### Object Model
+
+## Perl 6 has a quite comprehensive object model
+## You declare a class with the keyword `class`, fields with `has`, methods with `method`.
+## In Perl 6, every field is private, and named `$!attr`, but if you declare it with `$.`,
+## you get a public (immutable) accessor along with it.
+
+# (Perl 6's object model ("P6Model") is very flexible, and allows you to dynamically add methods,
+# change semantics, etc -- This will not be covered here, and you should refer to the Synopsis)
+
+class A {
+ has $.field; # `$.field` is immutable. Use `$!field` from inside the class to modify it.
+ has $.other-field is rw; # You can, however, mark a public field as being read/write.
+ has Int $!private-field = 10;
+
+ method get-value {
+ $.field + $!private-field + $n;
+ }
+
+ method set-value($n) {
+ # $.field = $n; # As stated before, you can't use the `$.` immutable version.
+ $!field = $n; # This works, because `$!` is always mutable.
+
+ $.other-field = 5; # This works, because `$.other-field` was declared `rw` (mutable).
+ }
+
+ method !private-method {
+ say "This method is private to the class !";
+ }
+};
+
+# Create a new instance of A with $.field set to 5 :
+# note : you can't set private-field from here (more later on)
+my $a = A.new(field => 5);
+$a.get-value; #=> 18
+#$a.field = 5; # This fails, because the `has $.field` is immutable
+$a.other-field = 10; # This, however, works, because the public field is mutable (`rw`).
+
+## Perl 6 also has inheritance (along with multiple inheritance ... Considered a misfeature by many)
+
+class A {
+ has $.val;
+
+ submethod not-inherited {
+ say "This method won't be available on B.";
+ say "This is most useful for BUILD, which we'll see later";
+ }
+
+ method bar { $.val * 5 }
+}
+class B is A { # inheritance uses `is`
+ method foo {
+ say $.val;
+ }
+
+ method bar { $.val * 10 } # this shadows A's `bar`
+}
+
+my B $b .= new(val => 5); # When you use `my T $var`, `$var` starts off with `T` itself in it, so you can call `new` on it
+ # (`.=` is just the compound operator composed of the dot-call and of the assignment operator)
+ #
+ # Also note that `BUILD` (the method called inside `new`) will set parent properties too,
+ # so you can pass `val => 5`
+# $b.not-inherited; # This won't work, for reasons explained above
+$b.foo; # prints 5
+$b.bar; #=> 50, since it calls B's `bar`
+
+## Roles are supported too (also called Mixins in other languages)
+role PrintableVal {
+ has $!counter = 0;
+ method print {
+ say $.val;
+ }
+}
+
+# you "use" a mixin with "does" :
+class Item does PrintableVal {
+ has $.val;
+
+ # When `does`-ed, a `role` literally "mixes in" the class :
+ # the methods and fields are put together, which means a class can access
+ # the private fields/methods of its roles (but not the inverse !) :
+ method access {
+ say $!counter++;
+ }
+
+ # However, this :
+ # method print {}
+ # is an error, since the compiler wouldn't know which `print` to use :
+ # contrarily to inheritance, methods mixed in can't be shadowed - they're put at the same "level"
+
+ # NOTE : You can use a role as a class (with `is ROLE`). In this case, methods will be shadowed,
+ # since the compiler will consider `ROLE` to be a class
+}
+
+### Exceptions
+# Exceptions are built on top of classes, usually in the package `X` (like `X::IO`).
+# Unlike many other languages, in Perl 6, you put the `CATCH` block *within* the block to `try`.
+# By default, a `try` has a `CATCH` block that catches any exception (`CATCH { default {} }`).
+# You can redefine it using `when`s (and `default`) to handle the exceptions you want:
+try {
+ open 'foo';
+ CATCH {
+ when X::AdHoc { say "unable to open file !" }
+ # any other exception will be re-raised, since we don't have a `default`
+ }
+}
+
+# You can throw an exception using `die`:
+die X::AdHoc.new(payload => 'Error !');
+# TODO warn
+# TODO fail
+# TODO CONTROL
+
+### Packages
+# Packages are a way to reuse code. Packages are like "namespaces", and any element of the six model
+# (`module`, `role`, `class`, `grammar`, `subset` and `enum`) are actually packages.
+# (you can say that packages are the lowest common denomitor between them)
+# Packages play a big part in a language, as Perl is well-known for CPAN,
+# the Comprehensive Perl Archive Network.
+# You usually don't use packages directly : you use `class Package::Name::Here;`, or if you
+# only want to export variables/subs, you can use `module`:
+module Hello::World { # bracketed form
+ # if `Hello` doesn't exist yet, it'll just be created as an "empty package stub"
+ # that can be redeclared as something else later.
+ # declarations here
+}
+module Parse::Text; # file-scoped form
+grammar Parse::Text::Grammar { # A grammar is a fine package, which you could `use`
+}
+
+# NOTE for Perl 5 users: even though the `package` keyword exists,
+# the braceless form is invalid (to catch a "perl5ism"). This will error out:
+# package Foo; # because Perl 6 will think the entire file is Perl 5
+# Just use `module` or the brace version of `package`.
+
+# You can use a module (bring its declarations into scope) with `use`
+use JSON::Tiny; # if you installed Rakudo* or Panda, you'll have this module
+say from-json('[1]').perl; #=> [1]
+
+# As said before, any part of the six model is also a package.
+# Since `JSON::Tiny` uses (its own) `JSON::Tiny::Actions` class, you can use it:
+my $actions = JSON::Tiny::Actions.new;
+
+# We'll see how to export variables and subs in the next part:
+
+### Declarators
+# In Perl 6, you get different behaviors based on how you declare a variable.
+# You've already seen `my` and `has`, we'll now explore the others.
+
+## * `our` (happens at `INIT` time -- see "Phasers" below)
+# Along with `my`, there are several others declarators you can use.
+# The first one you'll want for the previous part is `our`.
+# (All packagish things (`class`, `role`, etc) are `our` by default)
+# it's like `my`, but it also creates a package variable:
+module Foo::Bar {
+ our $n = 1; # note: you can't put a type constraint on an `our` variable
+ our sub inc {
+ our sub available { # if you try to make scoped `sub`s `our` ... Better know what you're doing (Don't !).
+ say "Don't do that. Seriously. You'd get burned.";
+ }
+ my sub unavailable { # `my sub` is the default
+ say "Can't access me from outside, I'm my !";
+ }
+ }
+
+ say ++$n; # lexically-scoped variables are still available
+}
+say $Foo::Bar::n; #=> 1
+Foo::Bar::inc; #=> 2
+Foo::Bar::inc; #=> 3
+
+## * `constant` (happens at `BEGIN` time)
+# You can use the `constant` keyword to declare a compile-time variable/symbol:
+constant Pi = 3.14;
+constant $var = 1;
+
+## * `state` (happens at run time, but only once)
+# State variables are only executed one time
+# (they exist in other langages such as C as `static`)
+sub fixed-rand {
+ state $val = rand;
+ say $rand;
+}
+fixed-rand for ^10; # will print the same number 10 times
+
+# Note, however, that they exist separately in different enclosing contexts.
+# If you declare a function with a `state` within a loop, it'll re-create the variable
+# for each iteration of loop. See:
+for ^5 -> $a {
+ sub foo {
+ state $val = rand; # This will be a different value for every value of `$a`
+ }
+ for ^5 -> $b {
+ say foo; # This will print the same value 5 times, but only 5. Next iteration will re-run `rand`
+ }
+}
+
+
+
+### Phasers
+# Phasers in Perl 6 are blocks that happen at determined points of time in your program
+# When the program is compiled, when a for loop runs, when you leave a block, when
+# an exception gets thrown ... (`CATCH` is actually a phaser !)
+# Some of them can be used for their return values, some of them can't
+# (those that can have a "[*]" in the beginning of their explanation text).
+# Let's have a look !
+
+## * Compile-time phasers
+BEGIN { say "[*] Runs at compile time, as soon as possible, only once" }
+CHECK { say "[*] Runs at compile time, instead as late as possible, only once" }
+
+## * Run-time phasers
+INIT { say "[*] Runs at run time, as soon as possible, only once" }
+END { say "Runs at run time, as late as possible, only once" }
+
+## * Block phasers
+ENTER { say "[*] Runs everytime you enter a block, repeats on loop blocks" }
+LEAVE { say "Runs everytime you leave a block, even when an exception happened. Repeats on loop blocks." }
+
+PRE { say "Asserts a precondition at every block entry, before ENTER (especially useful for loops)" }
+POST { say "Asserts a postcondition at every block exit, after LEAVE (especially useful for loops)" }
+
+## * Block/exceptions phasers
+sub {
+ KEEP { say "Runs when you exit a block successfully (without throwing an exception)" }
+ UNDO { say "Runs when you exit a block unsuccessfully (by throwing an exception)" }
+}
+
+## * Loop phasers
+for ^5 {
+ FIRST { say "[*] The first time the loop is run, before ENTER" }
+ NEXT { say "At loop continuation time, before LEAVE" }
+ LAST { say "At loop termination time, after LEAVE" }
+}
+
+## * Role/class phasers
+COMPOSE { "When a role is composed into a class. /!\ NOT YET IMPLEMENTED" }
+
+# They allow for cute trick or clever code ...:
+say "This code took " ~ (time - CHECK time) ~ "s to run";
+
+# ... or clever organization:
+sub do-db-stuff {
+ ENTER $db.start-transaction; # create a new transaction everytime we enter the sub
+ KEEP $db.commit; # commit the transaction if all went well
+ UNDO $db.rollback; # or rollback if all hell broke loose
+}
+
+
+### More operators thingies !
+
+## Everybody loves operators ! Let's get more of them
+
+## The precedence list can be found here : http://perlcabal.org/syn/S03.html#Operator_precedence
+## But first, we need a little explanation about associativity :
+
+# - Binary operators:
+$a ! $b ! $c; # with a left-associative `!`, this is `($a ! $b) ! $c`
+$a ! $b ! $c; # with a right-associative `!`, this is `$a ! ($b ! $c)`
+$a ! $b ! $c; # with a non-associative `!`, this is illegal
+$a ! $b ! $c; # with a chain-associative `!`, this is `($a ! $b) and ($b ! $c)`
+$a ! $b ! $c; # with a list-associative `!`, this is `infix:<>`
+
+# - Unary operators:
+!$a! # with left-associative `!`, this is `(!$a)!`
+!$a! # with right-associative `!`, this is `!($a!)`
+!$a! # with non-associative `!`, this is illegal
+
+## Create your own operators !
+# Okay, you've been reading all of that, so I guess I should try to show you something exciting.
+# I'll tell you a little secret (actually not): In Perl 6, all operators are actually just funny-looking subroutines.
+# You can declare an operator just like you declare a sub:
+sub prefix:<win>($winner) { # refer to the operator categories
+ # (yes, it's the "words operator" `<>`)
+ say "$winner Won !";
+}
+win "The King"; #=> The King Won !
+ # (prefix is before)
+
+# you can still call the sub with its "full name"
+say prefix:<!>(True); #=> False
+
+sub postfix:<!>(Int $n) {
+ [*] 2..$n; # using the reduce meta-operator ... See below ;-) !
+}
+say 5!; #=> 120
+ # (postfix is after)
+
+
+sub infix:<times>(Int $n, Block $r) { # infix in the middle
+ for ^$n {
+ $r(); # needs the parentheses because it's a scalar
+ }
+}
+3 times -> { say "hello" }; #=> hello
+ #=> hello
+ #=> hello
+
+# For circumfix and post-circumfix ones
+sub circumfix:<[ ]>(Int $n) {
+ $n ** $n
+}
+say [5]; #=> 3125
+ # circumfix is around
+
+sub postcircumfix:<{ }>(Str $s, Int $idx) { # post-circumfix is "after a term, around something"
+ $s.substr($idx, 1);
+}
+say "abc"{1}; #=> b
+ # after the term `"abc"`, and around the index (1)
+
+# This really means a lot -- because everything in Perl 6 uses this.
+# For example, to delete a key from a hash, you use the `:delete` adverb (named argument)
+%h{$key}:delete;
+# equivalent to:
+postcircumfix:<{ }>(%h, $key, :delete);
+# It's *all* using the same building blocks! Syntactic categories (prefix infix ...),
+# named arguments (adverbs), ..., used to build the language are available to you.
+
+# (you are, obviously, recommended against making an operator out of *everything* --
+# with great power comes great responsibility)
+
+## Meta operators !
+# Oh boy, get ready. Get ready, because we're dwelving deep into the rabbit's hole, and you probably won't want
+# to go back to other languages after reading that (I'm sure you don't want to already at that point).
+
+# - Reduce meta-operator
+
+## End of the operator list:
+
+
+## Sequence operator
+# The sequence operator is one of Perl 6's most powerful features:
+# it's composed of first, on the left, the list you want Perl 6 to deduce from (and might include a closure),
+# and on the right, a value or the predicate for when to stop, or even Whatever for a lazy infinite list.
+my @list = 1, 2, 3 ... 10; # basic deducing
+#my @list = 1, 3, 6 ... 10; # this throws you into an infinite loop, because Perl 6 can't figure out the end
+my @list = 1, 2, 3 ...^ 10; # as with ranges, you can exclude the last element (when the predicate matches)
+my @list = 1, 3, 9 ... * > 30; # you can use a predicate (with the Whatever Star, here)
+my @list = 1, 3, 9 ... { $_ > 30 }; # (equivalent to the above)
+my @fib = 1, 1, *+* ... *; # lazy infinite list of prime numbers, computed using a closure !
+my @fib = 1, 1, -> $a, $b { $a + $b } ... *; # (equivalent to the above)
+say @fib[^10]; #=> 1 1 2 3 5 8 13 21 34 55
+ # (using a range as the index)
+# Note : as for ranges, once reified, elements aren't re-calculated.
+# That's why `@primes[^100]` will take a long time the first time you print it, then be instant
+
+
+## * Sort comparison
+# They return one value of the `Order` enum : `Less`, `Same` and `More` (which numerify to -1, 0 or +1).
+1 <=> 4; # sort comparison for numerics
+'a' leg 'b'; # sort comparison for string
+$obj eqv $obj2; # sort comparison using eqv semantics
+
+## * Generic ordering
+3 before 4; # True
+'b' after 'a'; # True
+
+## * Short-circuit default operator
+# Like `or` and `||`, but instead returns the first *defined* value :
+say Any // Nil // 0 // 5; #=> 5
+
+## * Short-circuit exclusive or (XOR)
+# Returns `True` if one (and only one) of its arguments is true
+say True ^^ False; #=> True
+
+## * Flip Flop
+# The flip flop operators (`ff` and `fff`, equivalent to Perl 5/Ruby's `..` and `...`).
+# are operators that take two predicates to test:
+# They are `False` until their left side returns `True`, then are `True` until their right side returns `True`.
+# Like for ranges, you can exclude the iteration when it became `True`/`False` by using `^` on either side.
+# Let's start with an example :
+for <well met young hero we shall meet later> {
+ # by default, `ff`/`fff` smart-match (`~~`) against `$_`:
+ if 'met' ^ff 'meet' { # won't enter the if for "met" (explained in details below).
+ .say
+ }
+
+ if rand == 0 ff rand == 1 { # compare variables other than `$_`
+ say "This ... probably will never run ...";
+ }
+}
+# This will print "young hero we shall meet" (excluding "met"):
+# the flip-flop will start returning `True` when it first encounters "met"
+# (but will still return `False` for "met" itself, due to the leading `^` on `ff`),
+# until it sees "meet", which is when it'll start returning `False`.
+
+# The difference between `ff` (flip-flop) and `fff` (flip-flop) is that
+# `ff` will test its right side just as its left side changes to `True`,
+# and can get back to `False` right away (*except* it'll be `True` for the iteration that matched)
+# while `fff` will wait for the next iteration to try its right side, once its left side changed:
+.say if 'B' ff 'B' for <A B C B A>; #=> B B
+ # because the right-hand-side was tested directly (and returned `True`).
+ # "B"s are still printed since it matched that time
+ # (it just went back to `False` right away)
+.say if 'B' fff 'B' for <A B C B A>; #=> B C B
+ # because the right-hand-side wasn't tested until `$_` became "C"
+ # (and thus did not match directly).
+
+# A flip-flop can change state as many times as needed:
+for <test start print this stop you stopped printing start printing again stop not anymore> {
+ .say if $_ eq 'start' ^ff^ $_ eq 'stop'; # exclude both "start" and "stop",
+ #=> "print this printing again"
+}
+
+# you might also use a Whatever Star, which is equivalent to `True` for the left side or `False` for the right :
+for (1, 3, 60, 3, 40, 60) {
+ .say if $_ > 50 ff *; # Once the flip-flop reached a number greater than 50, it'll never go back to `False`
+ #=> 60 3 40 60
+}
+
+# You can also use this property to create an `If` that'll not execute the first time :
+for <a b c> {
+ .say if * ^ff *; # the flip-flop is `True` and never goes back to `False`,
+ # but the `^` makes it *not run* on the first iteration
+ #=> b c
+}
+```