my # token
$ # token
x # token
= # token
"1" # token
; # token
# statement detected!!
[download]

Perl ignores whitespace.

...except where it doesn't :)

Like for example in  < $fh >  vs.  <$fh>.  Or even in <$ fh> — to keep it on-topic.

Yeah, I suspect the perl toeknizer turns <$fh> into a single token... probably historical, since it always used to be <BLARG> and <> also does <stuff.*> ... so yeah, fine except where it doesn't work -- like way too much overloaded operators?

-Paul

Sure it's confusing. Without running the code, tell me what $i and $j are.

my $x = ['a', 'b', 'c', 'd', 'e'];

my $i = $ x -> [ 4 ];
my $j = "$ x -> [ 4 ]";

print "$i\n";
print "$j\n";
[download]

No :) some things aren't defined well, stick to what is defined :)

my $x = ['a', 'b', 'c', 'd', 'e'];

my $i = $ x -> [ 4 ];
my $j = "$ x -> [ 4 ]";
my $k = "$x->[4]";
warn $i;
warn $j;
warn $k;
__END__
e at - line 6.
ARRAY(0x225f04) -> [ 4 ] at - line 7.
e at - line 8.
[download]

Perhaps you could avoid unexpected behavior by being more deliberate about using interpolating quotes only when you want to interpolate. Otherwise, always use non-interpolating quotes when you do not want to interpolate. The following will set the $bar variable equal to a string containing a literal dollar sign, followed by a number of spaces, followed by the literal string 'foo' (if that is what you really want).

my $bar = '$    foo';
[download]

See Quote and Quote like Operators

Unless you 'wanted' to interpolate

my $bar = "$  foo and $cake";
[download]

Or a more likely example

my $bar = "$ $total_foo_dollars";
[download]

I had no idea that spaces in quotes could be abused this way and it just feels so wrong.

How about

my $foo = ['a', 'b', 'c'];
my $bar = "$   foo -> [2]";
[download]

vs

my $foo = ['a', 'b', 'c'];
my $bar = "$foo->[2]";
[download]

Before reading this thread I would have a clear answer for what the result of $bar would be in each case. After reading this thread I had a less clear different answer. And after actually testing the code I can see now that I was wrong in all my assumptions. Now I've got a nagging feeling that I should review all my code for unintended quoting results.

What do those print, and why?  (Without cheating and running the code! :)

#!/usr/bin/perl

$foo = "foo";

print "1: $ $ $ foo\n";
print "2: $ $$ foo\n";
print "3: $ $$foo\n";
print "4: $$$foo\n";
[download]

perl ignores whitespace

No it doesn't. If that where true, then there wouldn't be any difference between:

$foo =~ /bar/;
[download]

and

$ f o o = ~ / b a r /;
[download]

the latter not only does not compile, but the spaces inside // are significant, and while =~ is one token, = ~ are two. With a different meaning.

The reason that there is optional whitespace after the sigil is a left over from perl4. And probably has to do with the sigil and the variable name lexed as different tokens. And in almost all cases, Perl allows whitespace between tokens.

Now, in Perl6, the sigil is actually part of the variable name. Hence, '$foo' is a single token. And hence, no whitespace allowed between the sigil and the name.

See Is space supposed to be ignored between a sigil and its variable name? on stackoverflow.com.