perl grep

[RandomWalk]

What does

grep /regex/ => @_

mean? I don't see how it follows the format "EXPR, LIST". Thanks.

[!1]

=> is a glorified comma. So now we actually have:

grep /regex/, @_;
#    EXPR   , LIST

The regex is an expression. @_ can provide data like a list. And thus it is grep EXPR, LIST.

[Corion]

The "fat arrow" => is often also called the "fat comma", because it functions the same as the comma. So the following two pieces of code are the same, functionally:

  grep /regex/ => @_;
  
  # and

  grep /regex/, @_;

I think that perldoc perlop discusses the peculiarities of the fat arrow, but for many cases, you can simply replace the arrow by a comma.

[RandomWalk]

Ouch! I knew => was a comma in hash definition from the Llama but I didn't make the connection, sorry. But why would an author be using this form

unless (grep /regexp/ => @_) {...}

? Is it just easier to read than it would be using a simple comma? Or does the special property of interpreting the word to the left of => as a string have a beneficial effect? Thank you for your kind help.

[Corion]

I think the fat comma is used here for aestethical reasons, as /regexp/ is applied to @_. I am also guilty of that sometimes. The notation makes sense sometimes, originally for key-value pairs in hashes :

  my %hash = ( foo => bar );

You can also use it for the parameters if you have a subroutine that acts as a pipeline :

  sub transform {
    my ($source, $target) = @_;
    ...
  };

  transform( 'input/start.file' => 'output/' );

In that case, the "arrow" could be interpreted also in the mapping sense.

[tcf22]

It does the equivalent of grep /regex/, @_. => has most of the same properties of a comma.

For example

%hash = (this => 'that', now => 'then');

is equivalent to

%hash = ('this', 'that', 'now', 'then');

You can see how it differs from the comma in the quoting of the keys being optional in most cases. But for most normal uses, its just a funny looking comma.