It's the same as @x = ("a", "b", c);

please note that c is not quoted, hence this shouldn't compile under strict

Now that you know what it does, here is what it probably should be:

my @x = qw( a b c );
[download]

or perhaps even:

my @x = 'a' .. 'c';
[download]

And remember, testing is fun! Cheers!

perl -MTest::More=no_plan -e'is_deeply [a=>b=>c],[qw(a b c)]'
perl -MTest::More=no_plan -e'is_deeply [a=>b=>c],["a".."c"]'
[download]

L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)

Correct, except => automatically quotes barewords on its left (=> also has the nickname "fat comma"). See Comma Operator in perlop and Re^5: Perl Moose syntax

Actually, to be more accurate: You are only correct if strict is off. But since everyone uses strict (right? ;-) ), @x = (a => b => c); is like @x = ("a", "b", c); (note the quotes), and the final "c" also needs to be quoted unless it's a function call.

Doesn't it need to be &c to be a function call?

Not if you declare/define it first:

no strict;
sub c { print "see?\n" }
@x = ( a => b => c );
[download]

or

no strict;
sub c;
@x = ( a => b => c );
sub c { print "see?\n" }
[download]

But not:

no strict;
@x = ( a => b => c );
sub c { print "see?\n" }
[download]

UPDATE: corrected 2nd example (used &c; which called the sub ... oops)

jeffa

L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)

Not at all.

Prepending & to a sub call causes the sub call to ignore the sub's prototype.

Prepending & to a sub call with no parameter list specified (not even ()) causes the sub call to ignore the sub's prototype, and forgoes the creation of a localized @_.

These are not thing you should ever need to do. Because it must be followed by a sub call, & has the side effect of forcing the following word to be treated as the name of a sub, but that's just a side effect.