Revision of perl script for updating mirror page on website

Revision of perl script which pulls in the entries from include/tor-mirrors.csv and put the entries on https://www.torproject.org/getinvolved/mirrors

In legacy/trac#22150 (moved) arma wrote:

"A) We should think about ways to automate the script. Maybe we shouldn't spend too much energy on this one until we've thought about how the mirror pages will fit into the future website redo."

Script: https://gitweb.torproject.org/project/web/webwml.git/tree/update-mirrors.pl

Topics after a first review of the existing script:

1. I found several small perl bugs (or better, "not a nice coding style", not really bugs) in the existing script
2. The current way how the script is checking the mirror sites, isn't the best (it is looking for existing .xpi, .dmg, .exe, .tar.gz files)
3. The script has to be run manually at the moment. Can this be done automatically in the future (without creating security issues)? Have also attention here on the existing mailing list (https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-mirrors)