CODEWITHSUNDEEP
arraysperl
T G
T G

Reputation: 501

Perl Array contains itself?

I do not understand what is going on with this short segment of code.

I have put the output and comments about it in the middle, after the print statements which produce it.

my @regions = (
0,
1,
2,
[ 0, new Point(3,1), new Point(3,2), new Point(3,3) ],
4,
5 );

$reg1 = 3;
print "1: $regions[$reg1] \n";

@reg1 = @{regions[$reg1]};
print "2: $reg1[0]\n";
print "3: $reg1[0][1]\n";
print "4: ", Point::stringPoint($reg1[0][1]), "\n";

# HERE IS THE OUTPUT from the above print statements, with comments about my confusion appended
1: ARRAY(0xe8b0e0)      # ok - element 3 of @regions is an array, as expected.
2: ARRAY(0xe8b0e0)      # It appears to be an array of one element, which is itself. ???
3: Point=HASH(0xfda5e0) # We have go 'down' one more level to get what we want - that makes no sense
4: 3,1                  # Yes, there it is

package Point;
sub new {
   my $class = shift;
   my $self = {
      x => shift,
      y => shift
   };
   bless $self, $class;
   return $self;
}

sub stringPoint
{
 my $p = shift;
 return "$p->{x},$p->{y}";
}

" Code related to new question (with output) " ;

The real question I have is this:
How to work directly and conveniently with an array,
which is inside another array (not a copy of it) ?
Is the only way to do that by (always) de-referencing a reference?
Such as in the two non-working examples which follow.

Here is what I tried:

my $ref1 = \@{$regions[3]};

@{$ref1}[2] = new Point(4, 5);  # changes original array
print1Region(3, $ref1);
# OUTPUT = (3,1) (4,5) (3,3)

my @arr1 = @{$ref1};
$arr1[1] = new Point(2,6);  # does not
print1Region(3, $ref1);
# OUTPUT = (3,1) (4,5) (3,3)

$ref1[0] = new Point(1,4);  # does not
print1Region(3, $ref1);
# OUTPUT = (3,1) (4,5) (3,3)

Upvotes: 1

Views: 85

Answers (1)

ikegami
ikegami

Reputation: 385506

@{regions[$reg1]} is a weird and undocumented way of writing @regions[$reg1]. (It's a syntax documented for use in double-quoted string literals.)

@regions[$reg1] is an array slice with one element, which is a weird way of writing $regions[$reg1].

So you aren't getting the first element of the array referenced by $regions[$reg1] as you think you are; you are simply getting the first element of @regions.

Let's look at 

my $ref1 = \@{$regions[3]};
@{$ref1}[2] = new Point(4, 5);

Problem #1

\@ "cancel out", so 

my $ref1 = \@{$regions[3]};

is just a complicated way of writing

my $ref1 = $regions[3];

(Ok, that's not quite true because the former autovivifies, but that's not relevant here.)

Problem #2

Again you're using an array slice of one element. ALWAYS use use strict; use warnings; and heed the warnings!

@{$ref1}[2] = new Point(4, 5);

should be

${$ref1}[2] = new Point(4, 5);

which is more cleanly written using the "arrow notation".

$ref1->[2] = new Point(4, 5);

Problem #3

Finally, don't use indirect method calls. They cause issues.

$ref1->[2] = new Point(4, 5);

should be

$ref1->[2] = Point->new(4, 5);

Conclusion

my $ref1 = \@{$regions[3]};
@{$ref1}[2] = new Point(4, 5);

should be written as

my $ref1 = $regions[3];
$ref1->[2] = Point->new(4, 5);

Without the variable, this would be

$regions[3]->[2] = Point->new(4, 5);

or just

$regions[3][2] = Point->new(4, 5);

Upvotes: 4

Related Questions