Carsten Hey wrote:
> +sub read_file_and_hashify {
> + my $file = shift;
> +
> + my ($a_ref, $h_ref) = ([], {});
> + filemap $file, sub {
> + push @$a_ref, $_;
> + $h_ref->{$_} = 1;
> + };
> + return ($a_ref, $h_ref);
> +}
This could be written using filemap.
> sub compare_or {
> my ($file1, $file2) = @_;
>
> @@ -105,8 +116,11 @@ sub compare_or {
> sub compare_xor {
> my ($file1, $file2) = @_;
>
> - compare_not($file1, $file2);
> - compare_not($file2, $file1);
> + my ($a1, $h1) = read_file_and_hashify $file1;
> + my ($a2, $h2) = read_file_and_hashify $file2;
> +
> + foreach (@$a1) { print $_, "\n" unless exists $h2->{$_} }
> + foreach (@$a2) { print $_, "\n" unless exists $h1->{$_} }
This uses something like 4x the memory that it used before. It would be
easy to use only 3x[1]. With a good data structure, it should be
possible to only use 2x. Using an ordered hash would be one way. I have
no experience with Tie::IxHash, but it looks promising.
--
see shy jo
[1] get both lists,
make second hash,
iterate first list against second hash,
free second hash,
make first hash,
drop first list,
iterate second list against first hash
signature.asc
Description: Digital signature
