extreemly slow regexp when using large strings

[David Bishop]

Hello, I am trying to port an inhouse app from Solaris to GNU/Linux, and have 
run into a slight problem.  Our app uses advance (from regexp.h) from the gnu 
libc, to parse really long strings using regexps.  When I say really long, we 
slurp entire files into a single string and then parse it at once.  This 
seemed to work fine on Solaris (using gcc 2.95.2 and libc 2.10.0) but takes 
incredibly longer (i.e., 5 minutes per function call vs. sub-1 second) 
using egcs-2.91.66 and glibc 2.1.3 (redhat 6.2).  Now, I am reluctant to 
force an upgrade on my linux machines if it will do no good, so I was 
wondering if you have heard of such a problem, and if so what the workaround 
is (or, should I just bite the bullet and upgrade).  Any enlightenment shed 
our way would be mightly appreciated, and further information can be given if 
needed.  Thanks very much for a great compiler and libc!

D.A.Bishop
Programmer/Webmaster
Micron Tech