What is the difference between first-class entities and second-class entities in perl?

Question

What is defined by the saying "first-class entities" and how does it differ from "second-class entities"?

What does it mean when one says "regexes are first-class entities in modern perl when created with the qr// operator" (taken from Modern Perl: the book).

Answer 1

As MeNoMore correctly said, a first-class-entity is a data type of the language you can freely assign to variables etc. In Perl, these include:

• Scalars
• Arrays
• Hashes
• Coderefs (e.g. anonymous subroutines)
• IO
• Typeglobs (The symbol table is a hash of globs)
• Formats

Those can reside in the symbol table. The scalar slot can be occupied by various other types in addition:

• Signed integers
• Unsigned integers
• Floating point numbers
• Strings
• References
• Regexes

Some of these entities have built-in constructors into tha language: Number and String literals for scalars, list notation for arrays and hashes, [] and {} for anonymous array- and hashrefs, the sub keyword for code, the open function for IO objects, the format builtin for formats, the reference operator for references, and the qr{} operator for regexes.

There are language constructs in Perl that are not first-class entities and cannot be assigned to scalars or other first-class entities. For example, packages. This code doesn't work:

my $anonymous_package = package { ... };  # XXX

Shell commands have their own builtins, but are no data objects, so this won't work:

# don't execute `yes`, but store a handle to it in reference
my $shell_command = \qx{yes};

Instead, this statement should not terminate (and probably blow your memory).

Lists in Perl are language constructs, but no data types:

my $listref = \($x, $y, $z); # assigns reference to $z instead

The builtin types in Perl can have coercion rules:

• Numbers and Strings coerce back and forth.
• A single scalar in list context is a list of arity 1.
• An array in scalar context evaluates to the length of the array
• An (even valued) array can be assigned to a hash
• A Hash can be assigned to an array so that assigning this array to another hash would recreate the same hash
• A Hash in scalar context evaluates to (a) a false value if it is empty or (b) to a string indicating the number of filled and allocated buckets e.g. 1/8 or (c) to the number of keys in numerical context.
• Regexes in string context evaluate to a pattern string that behaves like the one they were specified with: qr(ab?c) eq "(?-xism:ab?c)", depending on the version of perl.

Objects can be overloaded to show similar coercion rules through overloading.

In the case of regex-refs, a scalar containing such a reference can be used interchangeably with a regex literal, e.g. in the pattern

$string =~ /ab?c/

the regex could be replaced with $regex if $regex is like above:

my $regex = qr/ab?c/;
$string =~ $regex ### no dereferencing syntax!
# $string =~ /$regex/ will work too, but may invoke string overloading first (?)

For example, coderefs require more biolerplate code:

sub foo {...}
foo();

versus

my $foo = sub {...};
$foo->();  # two possibilities
&$foo();