Removed guessing of best optimizing CFLAGS.

With old versions of GCC "-O1 -fomit-frame-pointer" produced the best
code for x86 systems, but with GCC 4.x this is no longer the case.

With GCC 4.1.0, building with "-O1 -fomit-frame-pointer" makes TRE
about 15% slower than when built with "-O2 -g" (the default).
Removing the "-g" has a negligible positive effect, so it's better to
leave it in so the result is actually debuggable.

darcs-hash:20060722115947-ced27-23242246b62f51ef08efd91e501e7a9756a88411.gz

configure.ac

@@ -42,9 +42,6 @@ if test "$tre_debug" = "yes"; then
     [ Define if you want TRE to print debug messages to stdout. ])
 else
   AM_CONDITIONAL(TRE_DEBUG, false)
-  if test "$tre_profile" != "yes"; then
-    TRE_PROG_CC_OPTIMIZATIONS
-  fi
   AC_DEFINE(NDEBUG, 1, [ Define if you want to disable debug assertions. ])
 fi
 

m4/Makefile.am

@@ -1,7 +1,6 @@
 EXTRA_DIST = \
 	ac_libtool_tags.m4 \
 	ax_check_funcs_comp.m4 \
-	tre_prog_cc_optimizations.m4 \
 	ax_check_sign.m4 \
 	ax_decl_wchar_max.m4 \
 	vl_prog_cc_warnings.m4