perl 6 hands-on tutorial

Perl 6 hands-on tutorial
A small introduction to a big language
Jonathan Worthington
c Edument AB
August 29, 2015

About the course
Not “all of Perl 6”, because we don’t have time.
Rather, enough to help you bootstrap an understanding of Perl 6,
so you will be able to go away and keep learning.
Discussion and questions encouraged!

Hands on: Try It And See
Every so often, we have try it and see slides: chances for you to
try things out for yourself in Perl 6.
They start out simply as one-liners in the REPL; later we move on
to writing scripts.
Hopefully you read the session abstract carefully and came
prepared. . . ;-)
Pairing encouraged! (Especially those of you who are prepared,
with those who are not!)

Numerics
We have integers. . .
1
And rationals (one integer divided by another):
0.25
And ﬂoating point numbers (the standard IEEE thing):
1.23e4

Values know their type
You can discover what type a value is using .WHAT. Therefore:
say 1.WHAT;
say 0.25.WHAT;
say 1.23e4.WHAT;
Tells us:
(Int)
(Rat)
(Num)

Basic calculations
We have the usual set of numeric operators available, and they are
deﬁned on all of the numeric types we’ve seen so far.
say 35 + 7; # 42
say 1.2 - 0.3; # 0.9
say 2.1e5 * 3.3e2; # 69300000
say 168 / 4; # 42
say 169 % 4; # 1
say 3 ** 3; # 27

Try it and see!
Your turn! Using the Perl 6 REPL, try and work out answers to
the following:
How big can integers get? (Hint: try raising to a large power!)
What type do you get when you divide two integers?
Try 0.1 + 0.2 == 0.3 in some other language, then try it in
Perl 6. What do you observe?

Strings
Perl 6 has lots of quoting constructs. The simplest are single and
double quotes.
say ’Gangnam style!’;
say "Much wow!";
The concatenation operator is the tilde. How to remember it?
Thing of it as stitching two things together. Also, it looks like a
piece of string. :-)
say "s" ~ "laughter"; # slaughter

Some common string operations
Here are some common operations on strings:
say chars("nyanCat"); # 7
say uc("nyanCat"); # NYANCAT
say lc("nyanCat"); # nyancat
say tclc("nyancat"); # Nyancat
Alternatively, you might prefer the method forms:
say "dogeCoin".chars; # 8
say "dogeCoin".uc; # DOGECOIN
say "dogeCoin".lc; # dogecoin
say "dogeCoin".tclc; # Dogecoin

What does .chars really tell you?
Unicode works at three levels:
Graphemes: Things that a human would think of as “a
character”
Codepoints: Things Unicode gives a number to (letters,
numbers, marks, diacritics, symbols)
Bytes: the way codepoints are represented on disk, on the
wire, etc.
The strings in diﬀerent programming languages work at diﬀerent
levels. A distinguishing feature of Perl 6 is strings at grapheme
level.
Therefore, chars gives the number of graphemes: the right answer
from a human perspective!

But Perl 6 does all the levels. . .
In Perl 6, we have diﬀerent types representing the diﬀerent levels
you may want to work at:
Str is an immutable string at grapheme level
Uni is an immutable bunch of Unicode codepoints, potentially
known to be normalized in some way
Buf is an bunch of bytes (and Blob is the same, but
immutable)
You can always go between Str and Uni. However, to go between
Buf and the higher levels, you must specify an encoding (UTF-8,
UTF-16, Latin-1, ASCII. . . )

Try it and see!
Take the string:
"Ac[COMBINING DOT ABOVE]c[COMBINING DOT BELOW]"
And:
Use .chars to get the number of graphemes, and .codes to
get the number of Unicode codepoints
Use .NFC and .NFD to view composed or decomposed
codepoints
Use .encode(’utf-8’) to see the UTF-8 bytes
If time allows, try using the uniname function to get the name of
NFC and NFD codepoints. Note that you are shown them in hex,
so you’ll need to do uniname(0x41), for example.

Booleans
Perl 6 provides True and False values of type Bool.
say False.WHAT; # (Bool)
say True.WHAT; # (Bool)
All types know how to “boolify” themselves. You can ask them to
with the ? operator, or ! to also negate that booliﬁcation.
say ?0; # False
say !0; # True
say so 0; # False (lower precedence form of ?)
say not 0; # True (lower precedence form of !)

Coercions
In fact, ? and ! are just two coercion operators, which move
between types. There is also a ~ preﬁx operator for making
something stringy:
say (~4.2).WHAT; # Str
And a + preﬁx operator for making something numeric:
say (+"42").WHAT; # Int
Note that trying to numify a string that does not contain a valid
number gives a Failure (a lazy exception, which is thrown if the
value is used, but can be “defused” by testing it; more later!)

Try it and see!
Have a play with coercions. Try to ﬁgure out:
What string values are true and false? ""? "0"? "False"?
What types do you get for +"42", +"1.5", and +"4.2e4"?
(Hint: be careful with precedence; .WHAT binds tighter than a
preﬁx operator.)
What do True and False numify to?

Scalars
A scalar variable holds a single item. It has a $ sigil. In Perl 6, we
require explicit declaration of variables.
my $answer = 42;
A Scalar is actually a ﬁrst-class object in Perl 6. However, in
nearly all operations this is transparent.
say $answer.WHAT; # (Int), not (Scalar)
However, it really is there. The scope references the Scalar,
which in turn contains an Int.

Assignment vs. binding
Assignment to a Scalar is just putting a new value into it.
$answer = 2; # Store Int 2 in $answer Scalar
$answer = $answer + 19; # Add 19 to $answer and update it
$answer *= 2; # Short-hand; multiply $answer by 2
So far, no surprises. However, Perl 6 also provides binding, with
the := operator. This stores the bound thing directly in the scope;
my $can’t-assign-this := "hammer"; # Scope has Str
$can’t-assign-this = ’nope, no way’; # Dies

.VAR
You won’t use binding particularly often, but knowing it exists will
help you to understand other parts of the language.
If you’re curious, however, you can see whether you have a Scalar
or not using .VAR:
my $answer = 42;
say $answer.VAR.WHAT; # (Scalar)
my $can’t-assign-this := "hammer";
say $can’t-assign-this.VAR.WHAT; # (Int)
Note: .VAR and .WHAT are code smells. They’re useful to
understand the language and debug, but testing against what they
return is nearly always the wrong approach.

Try it and see!
Try to work out the following:
What does an unassigned Scalar start out containing? (hint:
.WHAT)
Can a variable change its type over time?
Then, try this code:
my $a = 42;
my $b := $a;
$b = 100;
say $a;
Try to understand what happened.

Arrays
An array is an ordered collection of Scalars. It can grow on
demand, and may also be used like a queue or a stack.
The easiest way to get an array is to declare an @ variable. When
you assign to such a thing, we will parse the right hand side of the
assignment at list precedence, meaning you can assign without
any parentheses:
my @countries = ’UK’, ’Slovakia’, ’Spain’, ’Sweden’;
Arrays are of type Array, and know their number of elements:
say @countries.WHAT; # (Array)
say @countries.elems; # 4

Array indexing
The [...] array indexer can be used with a single element.
Assigning to an element that does not yet exist will extend the
array.
say @countries[0]; # UK
@countries[4] = ’Czech Republic’;
say @countries.elems; # 5
You can also use multiple indexes. For example, we can swap the
order of a couple of elements like this:
@countries[1, 2] = @countries[2, 1];
say @countries[0, 1, 2]; # UK Spain Slovakia
say @countries[0..2]; # (the same)
say @countries[^3]; # (the same)

Arrays and binding
When you assign into an array, each element is a Scalar
container. You can therefore bind array elements.
my $first := @countries[0];
$first = ’England’;
say @countries[0]; # England
Again, you probably won’t do this much yourself. However, it is
the mechanism that is behind things like:
@values[$idx]++;

Quote words
When we initialized our array, we did it as follows:
my @countries = ’UK’, ’Slovakia’, ’Spain’, ’Sweden’;
However, all the quoting gets tedious. We could instead do:
my @countries = <UK Slovakia Spain Sweden>;
This splits on whitespace. But what if you want to have some
elements that contain whitespace? Then you can do this:
my @countries = << "United Kingdom" Slovakia Spain Sweden >>;

Stack/queue behaviour
You can push/pop at the end of an array:
my @values;
@values.push(35);
@values.push(7);
@values.push(@values.pop + @values.pop);
We could have used the push/pop functions also.
At the start of an array, we can use unshift and shift. Using
push and shift in combination, you get queue semantics.

Try it and see!
What happens if you assign to slot 0 and 3 of an array? What
is .elems, and what do the slots 1 and 2 contain?
If you assign one array to another, do you get a fresh set of
scalars?
What happens if you bind an array to another array?
What happens if you bind a quote-words list directly to an
@variable? Can you assign to its elements?

Hashes
Hashes are good for storing mappings from one thing to another,
when you need to do quick lookups. The easiest way to get a hash
is to declare a variable with the % sigil. We typically assign pairs:
my %capitals = UK => ’London’, Slovakia => ’Bratislava’;
The variable is of type Hash, and the number of elements is the
number of key/value pairs:
say %capitals.WHAT; # (Hash)
say %capitals.elems; # 2

Indexing and slicing
Hashes can be indexed with {...}.
say %capitals{’UK’}; # London
say %capitals{’UK’, ’Slovakia’}; # London Bratislava
For literal keys, you can use the quote words syntax, which saves
some typing and looks neater:
say %capitals<UK>; # London
say %capitals<UK Slovakia>; # London Bratislava

Existence and deletion
There is no exists function or method in Perl 6. Instead, you use
the normal indexing syntax and add the :exists adverb:
say %capitals<UK>:exists; # True
say %capitals<Elbonia>:exists; # False
The same goes for deletion:
%capitals<England> = %capitals<UK>;
%capitals<Scotland> = ’Edinburgh’;
%capitals<UK>:delete;
Note that :delete returns the deleted elements - sometimes
handy!

Other useful Hash methods
You can get the keys and values in a hash:
say %capitals.keys; # UK Slovakia
say %capitals.values; # London Bratislava
You can also get a list of alternating key and value pairs, which -
as we’ll see later - turns out to be very useful for iteration.
say %capitals.kv; # UK London Slovakia Bratislava

Try it and see!
Are pairs ﬁrst-class objects in Perl 6? (Hint: (UK =>
’London’).WHAT)
Suppose you declare the hash in our examples, and have my
$a = ’UK’;. Think about what %capitals{$a} will do
compared to %capitals<$a>. Try it and see if you guessed
correctly!
What happens if you assign one hash to another? What if you
assign a hash to an array? How about an array to a hash?
What happens if you use .keys, .values, and .kv on an
array?

Interpolation basics
There are numerous advantages to variables having sigils. One of
them is easy interpolation of variables in strings.
First oﬀ, single quoted strings do not interpolate:
my $who = ’your mom’;
say ’$who ate the pie’; # $who ate the pie
In double-quoted strings, strings interpolate pretty much as you’d
expect:
say "$who ate the pie"; # your mom ate the pie

Interpolation of arrays and hashes
It’s possible to interpolate arrays and hashes, but you need to
make a little more eﬀort than with scalars. That way, we don’t
mangle email addresses.
say "Contact me@hotmale.com"; # Contact me@hotmale.com
Instead, you need to index. Note that the empty index, known as
the zen slice, allows you to get all elements.
my @winners = <David Ed Nicola>;
say "The winner is @winners[0]";
say "Top 3 were @winners[]";
The same rules apply for hashes.

Interpolating method calls
You can also interpolate method calls on any kind of variable:
my @winners = <David Ed Nicola>;
say "Top 3 are: @winners.join(’, ’)";
say "CONGRATULATIONS @winners[0].uc()!";
Note that this will not happen if you don’t put the parentheses on,
so we will not screw up things involving ﬁle extensions:
my $fh = open "$file.txt"; # Does not try and call .txt

Try it and see
You can embed code inside an interpolated string using curlies
{...}. Try doing a calculation inside of one - an addition of
two variables, say.
"<b>$name</b>" is the one common gotcha from the Perl
6 interpolation rules. What happens if you try it? Why does it
happen? Can you ﬁnd a solution?

The standard handles
Use say to write a line to standard output, and note to write a
line to standard error.
say "it again";
note "Uh-oh...";
A convenient way to write out a prompt and read a line from
standard input is the prompt function:
my $color = prompt "Name a color: ";

Working with ﬁles
We can read a ﬁle entirely into a scalar using slurp, or into an
array of lines by getting an IO handle and calling lines:
my $whole-file = slurp $filename;
my @lines = $filename.IO.lines;
The opposite of slurp is spurt:
spurt $filename, $stuff;
Naturally, you can do much more than this - but it’s enough for
now.

Comparisons
In Perl 6, we have separate operators for string and numeric
comparisons, so even if you don’t know the type of the data in the
variable, you understand the semantics of the operation.
Numeric String
Equal == eq
Not equal != ne
Less than < lt
Less than or equal <= le
Greater than > gt
Greater than or equal >= ge

Aside: general design rule
Perl has generally followed the view that diﬀerent semantics means
diﬀerent operator. This means that you know how the program
will behave even if you don’t know the exact types of the data.
There’s no question that == means numeric equality and eq means
string equality, or that + is addition and ~ is concatenation.
In Perl 6 we’ve de-confused a few things from Perl 5.
say ’edam’.flip; # made
say (1, 2, 3).reverse; # 3 2 1
say ’omg’ x 2; # (string ’omgomg’)
say ’omg’ xx 2; # (a list containing ’omg’ twice)

if/elsif/else
Basic conditonals are nothing especially surprising. Notice how you
don’t need any parentheses around the condition.
my $num = prompt "Enter a number: ";
if $num < 0 {
say "Negative";
}
elsif $num > 0 {
say "Positive";
}
else {
say "Zero";
}
There’s also unless (which has no elsif or else equivalent.)

given/when
A given/when is like a switch/case on steroids. We can use it
to get the $num variable out of our previous program.
given prompt "Enter a number: " {
when * < 0 {
say "Negative";
}
when * > 0 {
say "Positive";
}
default {
say "Zero";
}
}

Loops: loop
The loop statement is for writing an inﬁnite loop.
loop {
say "Fill all the screen: challenge accepted!";
}
You can also use it for C-style loops (but you usually can do
better):
loop (my $i = 10; $i > 0; $i--) {
say $i;
}
say "LIFT OFF!";

Loops: while/until
The while loop is rather unsurprising:
my @tasks;
while @tasks.elems < 5 {
@tasks.push(prompt("Enter a task: "))
}
Sometimes, until is more natural:
my @tasks;
until @tasks.elems == 5 {
@tasks.push(prompt("Enter a task: "))
}

Loops: while/until with pointy blocks
We can also test a value for truth and capture it into a variable
only visible inside the block loop:
my @tasks;
while prompt("Enter a task (blank if done): ") -> $t {
@tasks.push($t)
}
It works on if (and elsif) too.
if prompt("Enter file to write tasks (optional): ") -> $name {
# write file
}

Try it and see!
Write the classic “guess my number” game. Pick a random
number between 1 and 100 (hint: my $number =
(1..100).pick). Let the user make up to 7 guesses. If they get it
right, say they won and exit;. If not, state they are too high or
too low. If they don’t get it in 7 guesses, tell them they suck.
If you ﬁnish this, investigate using last instead of exit. After
this refactor, try to implement giving the option to play again
“yes/no”. If you’re still way ahead, make it multi-player, reading in
two player names and letting them take turns.

Iteration
The for loop is used to iterate (over arrays, hashes, and so forth).
The -> (pointy block) syntax is used again here, to scope the
iteration variable cleanly to the block.
for @tasks -> $task {
say $task.tclc;
}
Ranges are also common. For example, in the guessing game you
could have got the seven guesses using:
for ^7 {
# get/check guess...
}

Clarifying ranges
While we’re at it, let’s just clarify ranges a little more. First oﬀ,
they are a real object in Perl 6.
say (1..10).WHAT; # Range
There are inclusive/exclusive forms of range.
1 .. 5 # 1, 2, 3, 4, 5
1 ^.. 5 # 2, 3, 4, 5
1 ..^ 5 # 1, 2, 3, 4
1 ^..^ 5 # 2, 3, 4
^5 # 0, 1, 1, 2, 4 (short for 0..^5)

Two at a time
You can take two (or more) items at a time by adding extra
variables to the pointy block. This is especially useful on
%hash.kv:
for %capitals.kv -> $country, $city {
say "$city is the capital of $country";
}
You can do it on an array to loop over the indexes and values
together also:
for @tasks.kv -> $num, $task {
say "$num. $task";
}

Not just for .kv
You can take two (or three, or more) at a time from any array, not
just with .kv. For example, suppose we wanted to shuﬄe and
randomly pair up people to dance together, we could do:
my @dancers = <Jane Jill Jonny Jimmy Jenny Jack>;
for @dancers.pick(*) -> $first, $second {
say "$first dances with $second";
}
Note that this will fail if there are an odd number of dancers.
Soon, we’ll discover that the pointy block arguments are just a
special case of Perl 6 signatures, and you can make one optional or
set a default.

Try it and see!
Create a file with various words in. Loop over the words with a
for loop (hint: you can put one word per line in the file and use
lines(’file’.IO)). Use a hash to count how many times you’ve
seen each word (use the ++ increment operator). Then loop over
the hash to show how many times each word was seen.
If you get done: make it case insensitive, and implement some
means to filter out very common words (‘a’, ‘the’, ‘is’, etc’).

Subs and signatures
Subs and signatures

Pointy blocks can go solo
We’ve already seen that “pointy blocks” can go on for/while
loops, if statements, and so forth. However, writing one alone
gives you a callable piece of code. You can store it in a variable:
my $greeter = -> $greeting, $name { say "$greeting, $name!" };
And later call it:
$greeter(’Hej’, ’Carl’); # Hej, Carl!
Of course, there’s a more convenient way. . .

Sub declarations
We can use the sub keyword to declare a subroutine:
sub truncate($text) {
$text.chars > 100
?? "$text.substr(0, 10)..."
!! $text
}
We can then call it in one of the following ways:
my $short = truncate($text); # Function call form
my $short = truncate $text; # List op form

Syntax rules for calls
There are a couple of consistent rules about how calls are parsed.
First, foo(...) is always a call. It even beats keywords.
sub if($cond, $code) { if $cond { $code() } }
if(True, -> { say "Wow, a sub called ’if’!" });
Second, whitespace after the sub name will always make it a list
op. This
say substr "craft", 1; # Correct, passes 2 args to substr
say substr ("craft", 1); # WRONG, passes 1 arg (a list) to substr

Returning results
Our truncate sub worked ﬁne because by default a sub returns
the value produced by its ﬁnal statement. We could have been
more explicit by using return:
sub truncate($text) {
return $text.chars > 100
?? "$text.substr(0, 10)..."
!! $text
}
In this case, it’s a matter of personal preference. However, in other
cases you may need return in order to return from a nested block.
Note that pointy blocks are transparent to return. You can
not return out of a pointy block, only out of a routine (that is, a
sub or a method, which we’ll see soon).

Passing arrays and hashes
Passing arrays and hashes works just like passing scalar values.
sub trim-to-shortest(@a, @b) {
if @a.elems > @b.elems {
@a[@b.elems .. *]:delete;
}
else {
@b[@a.elems .. *]:delete;
}
}
my @x = 1..10;
my @y = 5..9;
trim-to-shortest(@x, @y);
say @x; # 1 2 3 4 5

A note on terminology
In Perl 6, we quite consistently use the terms argument and
parameter.
We use argument to mean the things that you pass when you
call a sub (or other piece of code, such as a pointy block)
We call a set of zero or more arguments an argument list
We use parameter to mean the thing you declare that
receives the argument
We call a list of zero or more parameters a signature
We try to stick to these in error messages, documentation, and so
forth. So it’s worth getting clear on them.

Try it and see!
See how a couple of possible failure modes work out:
What happens if you call a sub with the wrong number of
arguments? Do you get an error? Is it compile time or
runtime?
Make a typo when calling a sub. Do we ﬁnd it at compile
time or runtime?
Next, refactor your word frequency program into two subs:
calculate-frequencies and display-frequencies.

Optional parameters
When we put a ? on a parameter, we make it optional. If no
argument is passed for it, then it will contain Any. This
“undeﬁned” value is false in boolean context, so we can do this:
my @dancers = <Jane Jill Jonny Jimmy Jenny>;
for @dancers.pick(*) -> $first, $second? {
if $second {
}
else {
say "$first dances alone";
}
}

Default values
Another way to make a parameter optional is to supply a default:
my @dancers = <Jane Jill Jonny Jimmy Jenny>;
for @dancers.pick(*) -> $first, $second = ’the teacher’ {
}
Defaults are evaluated per call. Further, since we bind left to right,
it is possible to default to a previous parameter:
sub log-event($event, $occurred = now, $logged = $occurred) {
# ...
}

Named arguments/parameters
It’s also possible to use named arguments and parameters. These
are typically used for extra conﬁguration and options, and so they
are optional by default.
sub truncate($text, :$limit = 100, :$trailer = ’...’) {
$text.chars > $limit
?? "$text.substr(0, $limit)$trailer"
!! $text
}
say truncate("Drink a beer", limit => 11); # Drink a bee...

Other ways to pass named arguments
We have a number of diﬀerent syntaxes for named arguments.
limit => 11 # fat arrow syntax
:limit(11) # colon pair syntax
The colon pair syntax has some special cases for booleans:
:global # same as :global(True)
:!global # same as :global(False)
And also one for passing on a variable with the same name as the
parameter:
:$global # same as :global($global)

Slurpy parameters
Also we have slurpy parameters, when we want to accept any
number of arguments. There are a few forms of these. For
example, we can take all positionals:
sub join-and-truncate($joiner, $limit, *@values) {
truncate(@values.join($joiner), limit => $limit);
}
Or perhaps also all the named arguments:
sub join-and-truncate($joiner, *@values, *%options) {
# ... but how to pass the options on to truncate?
}

Flattening arguments
The “opposite” of slurping is ﬂattening an array or hash into an
argument list. This is done with the | preﬁx operator (which is only
meaningful inside of an argument list).
sub join-and-truncate($joiner, *@values, *%options) {
truncate(@values.join($joiner), |%options);
}

Slurpy positionals and ﬂattening
It’s sometimes important to remember that *@pos will ﬂatten
incoming arrays into a single array. Therefore, if you want to write
this:
sub check-array-lengths(*@arrays) {
for @arrays -> @arr {
die "Oh no it’s too long!" if @arr.elems > 5;
}
}
my @a = 1..4;
my @b = 1..3;
check-array-lengths(@a, @b);
You’ll end up in a mess, with @arrays containing 7 Ints!

Non-ﬂattening slurpy positional
You can prevent this steam-rollering by using **@arrays instead:
sub check-array-lengths(**@arrays) {
for @arrays -> @arr {
die "Oh no it’s too long!" if @arr.elems > 5;
}
}
my @a = 1..4;
my @b = 1..3;
check-array-lengths(@a, @b);

Code arguments
To pass along a subroutine as an argument to another subroutine,
pass it by putting the & before its name. On the parameter side,
you could happily use a Scalar, but using the & sigil there too
oﬀers convenient calling.
sub run-without-banned-options(&run, %options, @banned) {
my %filtered = %options;
%filtered{@banned}:delete;
run(|%filtered);
}
sub operate(:$force, :$strip, :$verbose) {
say $force, $strip, $verbose;
}
my %opts = force => 1, verbose => 1;
my @banned = ’force’, ’strip’;
run-without-banned-options(&operate, %opts, @banned);

is copy
Note how we made a copy of the %options hash on the previous
slide. We did this so as to avoid modifying the existing one that
was passed to us. We can achieve this more easily with is copy:
sub run-without-banned-options(&run, %options is copy, @banned) {
%options{@banned}:delete;
run(|%options);
}
sub operate(:$force, :$strip, :$verbose) {
say $force, $strip, $verbose;
}
my %opts = force => 1, verbose => 1;
my @banned = ’force’, ’strip’;
run-without-banned-options(&operate, %opts, @banned);

Try it and see
First, try to write a sub inc that, when passed a variable as an
argument, adds one to it in place (so my $a = 41; inc($a);
say $a; gives 42). What happens? Try using is copy and see
how it changes (but does not ﬁx) the problem. Then try out is
rw.
Next, make your word frequency ﬁlter take the list of blocked
words as a named parameter. Also add a threshold named
parameter (defaulted to zero), and don’t return any words with
frequencies below the threshold.

Classes, attributes, and methods
Classes, attributes, and
methods

What are objects?
“I’m sorry that I long ago coined the term ‘objects’ for this topic
because it gets many people to focus on the lesser idea. The big
idea is ‘messaging’. . . ” - Alan Kay
Objects are autonomous components with internal state - which
we aren’t meant to care about from the outside. Instead, we
communicate with an object entirely by sending it messages.
In Perl 6 parlance, as in many other languages, we call message
sends “method calls”. But we get better OO designs if we think of
them as sending a message to some independent agent.

Declaring a class with attributes
We’ll build a simple score keeper for a game where players earn
points. Write class to declare a class, and has to introduce state
private to it. Note the ! twigil means “this is private to the class”.
class ScoreKeeper {
has %!player-points;
}
There are various other forms of attribute declaration, but the real
name of an attribute is always $!foo/@!foo/%!foo/&!foo.

Adding methods
When players earn points, we want to add them. We also want a
method to get the standings. Methods look very much like subs,
except you declare them with method.
method score($player, $points) {
%!player-points{$player} += $points;
}
method ranking() {
%!player-points.pairs.sort({ -.value })
}

BUILD
Finally, we need to initialize all players to zero points. (Yes, Perl 6
does have auto-viviﬁcation, but if we rely on that then our ranking
will lack players who never earn points).
We write a submethod (a method only callable directly on the
class, not inherited) with the special name BUILD to do this:
submethod BUILD(:@players) {
for @players -> $player {
%!player-points{$player} = 0;
}
}

Creating an instance
We can now create a new instance by calling the new method,
which we always inherit. Any named arguments we pass will make
it to BUILD:
my $game = ScoreKeeper.new(players => <jnthn masak lizmat>);
We can then use the object as follows:
$game.score(’jnthn’, 100);
$game.score(’lizmat’, 150);
for $game.ranking -> $s {
say "$s.key(): $s.value() points";
}

Try it and see!
Write a DaySchedule class for scheduling hourly appointments. It
should work as follows:
my $sched = DaySchedule.new(opening => 9, closing => 17);
$sched.add-appointment(10, ’Fred’); # Works fine
$sched.add-appointment(8, ’Brad’); # Dies: too early
$sched.add-appointment(17, ’Anna’); # Dies: too late
$sched.add-appointment(9, ’Lena’); # Works fine
$sched.add-appointment(10, ’Adam’); # Dies: conflict
If you’re free time, add a method that describes the schedule.

Roles
A role is a collection of methods, and perhaps attributes, that can
be incorporated into a class.
Suppose we wanted to handle point-earning games like those we
have so far, along with games where the winner is the ﬁrst to reach
a goal. We can get some re-use by factoring the scoring
functionality out into a role.
role Scoring {
has %!player-points;
method !zero-scores(@players) {
for @players -> $player {
%!player-points{$player} = 0;
}
}
method score($player, $points) { ... }
method ranking() { ... }
}

Doing a role
The does keyword is used to incorporate a role into a class.
class SimpleScoreKeeper does Scoring {
submethod BUILD(:@players) {
self!zero-scores(@players);
}
}
You can incorporate many roles into a class. Any conﬂicts - where
two roles try to provide a method of the same name - will be
identiﬁed at compile time.

Re-using Scoring
Here’s how we can implement the “ﬁrst to meet the goal” score
class.
class FirstToGoalScoreKeeper does Scoring {
has $!goal;
submethod BUILD(:@players, :$!goal) {
self!zero-scores(@players);
}
method winners() {
%!player-points.grep({ .value >= $!goal }).map({ .key })
}
}

If time allows. . .
(whatever topics there’s interest in)

Where to learn more
The primary source for Perl 6 documentation is doc.perl6.org.
There are many examples at examples.perl6.org.
There are more links at perl6.org/documentation.

Where to ask questions
The #perl6 channel in irc.freenode.org is a great place to get
help. Many Perl 6 contributors and users are there, and there is a
friendly atmosphere (please help keep it that way!)
If you prefer mailing lists, then the perl6-users list is a good place
to ask questions also. Find out more at perl6.org/community/.

Thank you for coming!
Thank you for coming!

perl 6 hands-on tutorial

Recommended

More Related Content

What's hot (19)

Viewers also liked (14)

Similar to perl 6 hands-on tutorial (20)

More from mustafa sarac (20)

Recently uploaded (20)

perl 6 hands-on tutorial