Ok, here is the cut down version of what's being done:

my @actualData;
my @expectedData;

# $actualFile and $expectedFile are the names of the files 
# that will be diff'ed.
# At this point they are unmodified copies of the files I 
# want to manipulate and then compare
tie (@actualData, tie::File, $actualFile, mode => O_RDWR, autodefer =>
+ 0);
tie (@expectedData, tie::File, $expectedFile, mode => O_RDWR, autodefe
+r => 0);

# Here would do lots of stuff with @actualData 
# and @expectedData

untie @actualData;
untie @expectedData;

# ...and here would do the file comparison of $actualFile and $expecte
+dFile
[download]

I had a look at the caveats but couldn't see anything there that worried me (I'm not bothered about performance issues).