I have taken as rule to use awk only for trivial tasks and to switch to perl as soon as the syntax is slightly beyond my usual use cases. In perl, I would do:
There are many AWK implementations that are obscenely portable. A case in point is busybox AWK. Another might be http://unxutils.sourceforge.net
A complex AWK script is very, very easy to move to a platform that lacks any AWK parsers. It can easily be done on Windows, without administrative rights, by the placement of a single .EXE - Perl can do many things, but that is not one of them (to the best of my knowledge).
If you have array @foo in perl, $#foo is the index of the last element of the array, which is just the size of the array -1. So if @foo is undefined, $#foo is -1.
Using a variable instead of 'foo' is a symbolic reference, so this is effectively using the symbol table as the associative array. This means that this solution also gets it wrong if your file contains a line that matches the name of a built-in variable in perl. That would be tough to debug!
If your file contains
This is the first line of the file
then during execution of
++$#$_
the result is the same as if you had written
++$#{This is the first line of the file}
So the variable @{This is the first line of the file} goes from undefined to an array of length 1, turning $#{This is the first line of the file} to 0.
Incidentally, this is why the snippet fails to work for a line repeated more than once: for each occurrence of the expression, the value returned is in the sequence -1, 0, 1, 2, 3, ... so it is only false for the second occurrence.
Using preincrement instead of postincrement means the values returned are 0, 1, 2, 3, ... which means that inverting the test makes it false for every occurrence after the first.
