Re: Challenge Problem: Merging Network Addresses

[Dominus]

That's a great answer. Thanks a lot.

I can believe that Net::CIDR is the fastest way to do this in Perl. Unfortunately, it may be too slow. My test program took about 2.5 minutes to process test file nm5.in. The C program is about 120-150 times faster.

That's a bummer. I was hoping that there would be a Perl solution that would be fast enough to be useful.

--
Mark Dominus
Perl Paraphernalia

[tachyon]

I was hoping that there would be a Perl solution that would be fast enough to be useful

Of course if you patched the module with XS and your C code there would be :-) Just a random thought to increase your workload for the benefit of everyone (except perhaps you).

cheers

tachyon

s&&rsenoyhcatreve&&&s&n.+t&"$'$`$\"$\&"&ee&&y&srve&&d&&print

[Dominus]

Says tachyon:

Of course if you patched the module with XS and your C code there would be :-)

Not really. The C program is a complete program that solves this one specific problem. There's nowhere in the Net::CIDR interface to 'patch'; the module doesn't do what this program does, and this program doesn't implement any of the functions that Net::CIDR does.

Even if I did put it in, there wouldn't be any benefit. At present, you could compile my program and then use:

#!/usr/bin/perl
system("mjd-merge-blocks inputfile > outputfile");

Then you could call it a "Perl solution", but it wouldn't really be one. Hacking my code into Net::CIDR wouldn't be a "Perl solution" either.

--
Mark Dominus
Perl Paraphernalia