sub bar {
         my $str = "foo";
         return scalar reverse $str;
 }
 
 print chop bar();
 
captvanhalen: $ perl foo
 Can't modify non-lvalue subroutine call in chop at foo line 6, near "
+);"
 Execution of foo aborted due to compilation errors.
[download]

How often have you wanted to say:

reverse $foo = "merlyn";
[download]

in sensible code? That's why it doesn't return an lvalue!

-- Randal L. Schwartz, Perl hacker

Definite agreement... reverse $foo = "merlyn"; makes little to no sense. However, chop reverse $foo; does seem relatively reasonable... (though I see how it quickly goes to hell in a handbasket.) I remember when in my Data Structures in C++ class they tried explaining the reason to overload some operators as lvalues, and some not... it always seemed like you could *find* a scenario where you may want that functionality (though I doubt 2 + 2 = $foo; would ever be a good one :) Well... thanks to everyone for the answers! I am going to agree that it really makes very little sense to let reverse return an lvalue... though I think the obfuscation "value" alone may warrant it :)

For the same reason 2 + 2 doesn't return an lvalue.
Besides, there's no reason for having reverse return an lvalue. TMTOWTDI, and reverse chop was never supposed to be one of them. Still, there are a lot of other WTDI.


2;0 juerd@ouranos:~$ perl -e'undef christmas'
Segmentation fault
2;139 juerd@ouranos:~$
[download]