use IO::Uncompress::Gunzip qw(:all);

    # assume compressed HTML is in $compressedHTML 
    my $uncompresedHTML;

    gunzip \$compressedHTML => \$uncompresedHTML;
[download]

To expand on the above post from pmqs. If you are generating the HTML, you can use the code below to compress/uncompress the HTML. I find I can compress a 100K HTML to about 9K with these tools, and then save the variable in a DB for later reuse. Excellent!

use IO::Compress::Gzip qw( gzip $GzipError );
use IO::Uncompress::Gunzip qw(gunzip $GunzipError );

my ( $uncompresedHTML, $compressedHTML );
. . .
gzip \$uncompressedHTML => \$compresedHTML;
. . .
gunzip \$compressedHTML => \$uncompresedHTML;
[download]

If your getting the zipped variable from a file, there are many ways to compress/uncompress data, so you may have to use 'qx/ /;' to use a system application and get the results back.

Good Luck

As an aside, it can be quite useful to pre-compress even “ordinary” HTML (of any significant size) into gzip files so that the web server does not have to fully compress the stream each time.   You can do the work ahead of time (or “on the fly, but once”) and then simply send the stream with the appropriate HTML headers.   Natcherly, this qualifies as “something that has already been well-done,” such that there are a plethora of plug-ins for various platforms that will do it (more, or less) automagically for you.   I didn't fully realize just how much time a web-server could be spending on the repetitious task of compressing things for transmission.