#!/usr/bin/perl -w
use strict;
my $file = "test.txt"; # Text file to open
my @newtext; # array to hold target data
my $char = chr(186); # ACSII character º

# Open the file
open(FILE, "<$file") || die "Couldn't open $file: $!\n";
# If the contents of each line matches the º character at the beginnin
+g and end of the line,
# push it onto an array.
while(<FILE>) {
     if(/$char(.+)$char/) {
          push(@newtext, $1);
     }
}
close FILE;
# Print out the results
foreach(@newtext) {print "$_\n";}
[download]

# clean lines with words in
$text =~ s/.*?             # anything
           (?:             # don't remember me
            (\w.*\w)       # everything from first to last word charac
+ter
                           # on the line - remember this
            |              # or
            [^\w\s]{6,}    # at least 6 characters that aren't words o
+r
                           # spaces (in a row) - forget me
           )
           .*              # the rest of the line
           /$1/gx;         # replace whole line with (\w.*\w) match,
                           # if made.
[download]

Now, if your text contains puntuation, you might want to amend the (\w.*\w) match to take account of commas, etc that may appear at the end of a line:

(\w.*\w) => ([\w\.,].*[\w\.,])
[download]

cLive ;-)

ps - IANAL (I am not a lecythis)

 s/\361//g 

metadoktor

"The doktor is in."

You open the file read and write. You might be truncating the file before you can read from it.

You are having trouble getting past the first line, might also indicate premature end of file that can happen when reading binary files on a Windoze OS. You'll want to use binmode function to read binary files. See 'perldoc -f binmode'.

Also editing the file in place is prone to error. Some code demonstating how you are doing this would be helpful.

YuckFoo