join on the fly while reading the files in line by line?

This is very easy to do using redo:

$ echo 'somedata data blah \
continuation of previous line \
another continuation of previous line \
stand alone line
stand alone line
multiple line \
continuation of multiple line' | perl -e'
while ( <> ) {
    if ( s/\\\n// ) {
        $_ .= <>;
        redo;
        }
    print;
    }
'
somedata data blah continuation of previous line another continuation 
+of previous line stand alone line
stand alone line
multiple line continuation of multiple line
[download]

If you want to 1)combine the multiple lines in advance, you could slurp the file with $/, use a regular expression to strip "\\\n" and then split on the remaining new lines.

If you want to 2)join on the fly while reading the files in line by line, you could use add buffer to your while loop and use next to do flow control if the line ends with backslash, like:

my $buffer = q{};
while(<F>) {
    $buffer .= $_;
    if (multiple line) {
        next;
    } else {
        #process;
        $buffer = q{};
    }
}
[download]

A simple approach would be to do it on the fly, concatenating lines if possible, otherwise proceeding with working on them.

The approach I use most often is a tip i found from one of merlyn's posts - use a buffer:

use strict;
use warnings;
use autodie qw/open close/;
## buffer for holding current text;
my $buffer = '';

open (my $fh, '/foo/bar.txt');

while(<$fh>) {
    ## clear trailing whitespace
    chomp;
    if (/\/$/) {
      ## use regex to see if line ends with a slash
      ## append to buffer
      $buffer .= $_;
    } else {
      ## it's a single line so just process
      process($buffer); ## sub elsewhere...
      ## reset
      $buffer = '';
    }
}

close $fh;
[download]

NB: Not tested! Just example...

Update - This is merlyn's node - Re: Reading multiple lines?. Maybe not quite the same thing, but it inspired this response, or at least was the first place i came across this kind of control structure... He also remembered to check for the 'end of file'...