This is a port of the glibc gnu regex engine into perl.  There are few
reasons you would need this.  The few I can think of include:

0) You wish to use untrusted user expressions in such a way as to be
able to catch errors.  Example:  eval { alarm 2;  m/((){1024}){1024}/ }
is an instant uncatchable segmentation fault.  GNU's regexps will still
fail, but in a timeout way rather than an instant segfault way.

1) You wish to have POSIX compliance on ... something ...  Perl's
regexps are slightly different -- arguably better, but different.