Destructuring Assignment Syntax for Perl
This code:
use assign::0;
my {$foo, bar => [ $first, @rest, $last ]} = $self->data;
works the same as this code:
my $_temp1 = $self->data;
my $foo = $_temp1->{foo};
my $_temp2 = $_temp1->{bar};
my $first = $_temp2->[0];
my @rest = @$_temp2[1..(@$_temp2-2)];
my $last = $_temp2->[@$_temp2-1] if @$_temp2 > 1;
This module enables a destructuring assignment syntax for Perl.
Instead of assigning a value to a variable, you can now assign to a data structure (array-ref or hash-ref) that specifies how to unpack a data structure into all the variables you need.
When you use this module you can put array-refs or hash-refs in places where you would put a variable that will be assigned to. The contents of these refs act as instructions for what variables to create and which parts of the target to take the assignment values from.
There are many useful combinations, which are documented below. The destructuring refs can be nested, allowing you to extract all the values you want from a complex data structure in a single statement.
This module is very new and experimental. The hope is that this may one day become a pragma module or possibly even part of the Perl language's syntax.
To preserve backwards compatability for early adopters, the module currently
requires you to use assign::0;
.
When the module becomes stable and vetted, it will become simply use assign;
.
Simply add a use assign::0;
line to your program, and then you can use any of
the assignment forms described below.
You must use
the module before the first line where you use one of the
assignment forms.
If you need to turn off assign
after turning it on for some reason, you can
use the line: no assign
.
To see how the assign
module turns the new style assignment forms into plain
old Perl code in you program, you can use the assign::0->debug()
method.
require assign::0;
print assign::0->debug(<perl-file-name>);
print assign::0->debug(<perl-code-string-as-scalar-ref>);
This will perform the assign
transformations on the Perl code and print the
result.
You can do that as a perl one-liner like so:
perl -e 'require assign::0; print assign::0->debug("program.pl")'
Remember to use require assign::0;
, not use assign::0;
.
The best way to explain this module is by example.
All the forms below have been implemented and should work as described for this version of the module.
Here we go...
-
my [$foo, $bar] = $array_ref;
Define 2
my
variables and assign the array-ref values in order.If there are less values on the RHS, the variables will be undefined. If there are more, they will be ignored.
The RHS can be a variable or any expression as long as its value is an array-ref.
-
my [$foo, $bar] = (111, 222);
ERRORIf the LHS is an array-ref the RHS must also be an array-ref. If the LHS is a hash-ref the RHS must also be a hash-ref.
-
our [$foo, $bar] = $array_ref;
Define
our
variables and assign the array-ref values. -
local [$foo, $bar] = $array_ref;
Define
local
variables and assign the array-ref values. -
my $foo; our $bar; [$foo, $bar] = $array_ref;
Assign the array-ref values to 2 pre-defined variables.
-
my {$foo, $bar} = {bar => 111, foo => 222};
Define 2
my
variables and use the variable name as the key to extract from the hash-ref on the RHS.
-
my [ $foo, _, _, $bar ] = $array_ref;
You can skip array values with the
_
symbol. -
my [ $foo, $_, $bar ] = $array_ref;
Unpack into global
$_
, not a my variable. The$foo
and$bar
here are stillmy
lexical vars. -
my [ 7, $foo, 42, $bar ] = $array_ref;
You can skip any number of array values by using a positive integer.
-
my [ $a, $b=42, $c="1\n2", $d=$a ] = $array_ref;
You can define default values for variables. Currently the default must be a single token (string, number, variable). Also there must not be whitespace on either side of the
=
. -
my [ $a, $b, @rest ] = $array_ref
Slurp remaining elements into an array variable.
-
my [ $a, $b, @$rest ] = $array_ref
Slurp remaining elements into an array reference variable (
$rest
).
There are many more forms that are intended to work, and many more cases we haven't thought of yet. Pretty much any place in Perl where a variable can be assigned to should also allow a structure to be assigned to.
Here's the current list of things intended to be added soon:
my [ $x1, $xs->[5], $xs[10] ] = $d;
my [ $a, @middle, $z ] = $d; # Set all but first and last into array
my [ $first, @-, $last ] = $d; # Ignore middle
my [ $x1, $x2, @-25 ] = $d; # Take -27 and -26
my [ @a => @b => @c ] = $d; # Evenly distribute values over multiple arrays
my [ @a => @- => @- ] = $d; # Take every third element (0, 3, 6, ...)
my [ @a => @19 ] = $d; # Take every 19th element
my [ @a, @b, @c ] = $d; # Split into thirds
my [ @a, @, @ ] = $d; # Get first third
# Hash destructuring:
my { $k1, $k2 } = $d; # Unpack a hash ref
my { $k1, $k2 } = %d; # Unpack a hash
my { k1 => $x, $k2 } = $d; # Use a var name different than key
my { k1 => $x=111, $k2=222 } = $d; # Set default values
my { $key => $val } = $d; # Unpack a single pair hash
my { @key => @val } = $d; # Unpack all keys and values (unzip)
my [ @key => @val ] = $d; # Unpack all keys and values *sorted*
my { @keys } = $d; # Key array of all keys
my [ @keys ] = $d; # Key array of all keys *sorted* (RHS must be hashref)
my { @keys => _ } = $d; # Same as above
my [ @keys => _ ] = $d; # Same as above but *sorted*
my { _ => @vals } = $d; # Get array of all values
my { 'a-key', 'b.key' } = $d; # Short for { 'a-key' => $a_key, 'b.key' => $b_key }
my { 'foo bar' } = $d; # Short for { "foo bar" => $foo_bar }
# Nested destructuring:
my { k1 => { $k2, $k3 }} = $d; # Unpack nested hash (no $k1)
my { $k1 => { $k2, $k3 }} = $d; # Unpack nested hash w/ $k1 set to inner hash ref
my { k1 => [ $x1, $x2 ]} = $d; # Unpack array ref nested in hash (no nesting depth limit)
# Operator assignments:
[ $a, $b ] //= $d; # Only assign undefined variables
[ $a, $b ] .= $d; # Append string to every var
[ $a, $b ] += $d; # Add number to every var
# In for loops:
for my { $k => [ $x1, $x2 ]} (@list) { # Unpack each collection from a list
for my { $k => $v } (%hash) { $d; # Unpack each key/val pair from a hash
# In signatures:
sub foo( $a, {$k1, [ $x1, $x2 ]} ) { … }
sub foo({
$name = "Fred",
number => $num = 42,
}) { … }
# Regex:
my [ $match, $cap1, $cap2 ] = $str =~ /…/;
my [ $match, $cap1, $cap2 ] = /…/; # Match using $_
# Inline list expressions:
my [ $a, @l{reverse}, $y, $z ] = $d;
my [ $a, @l{map ($_ + 1), grep ($_ > 10)}, $z ] = $d;
my [ $a, @{join '-'} => $s, $z ] = $d;
All of the new assignment forms introduced here would cause Perl syntax or
runtime errors without using assign.pm
.
Currently this module is a working prototype that uses Filter::Simple (Perl source code filtering) and PPI.pm (to parse and restructure Perl code).
Each destructuring form is removed and replaced with the appropriate Perl code to do the intended actions.
When all the syntax forms have been implemented and fully tested and the module has become stable, it will be rewritten as an XS module.
Note: Code transformations adjust the line numbers with #line <num>
statements so that warnings and errors report line numbers that make sense.
Destructuring assignment is available in many common languages.
The assign
module got many ideas from these.
- Ingy döt Net
<[email protected]>
- Kang-min Liu
<[email protected]>
Copyright 2023 by Ingy döt Net
This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.