Would you be interested in an alternative patch that modifies lib/strtod.c so that it does not need 'pow'? This would fix a continuing porting hassle with that file.
The idea is to compute 10**exponent ourselves, instead of relying on pow. On IEEE floating point platforms this can be done without introducing any further rounding error. On other platforms there might be an additional rounding error but this isn't of much practical importance (as the function already has rounding errors, and these platforms are exceedingly rare these days anyway). We'd have to stop using AC_FUNC_STRTOD, but that's OK. AC_FUNC_STRTOD tests for bugs in parsing "NaN", but our replacement function doesn't parse "NaN" either. AC_FUNC_STRTOD also tests for bugs in parsing " +69", but is that bug of any practical importance any more? I doubt it, but if it is, then we should simply install a wrapper that strips leading " +". _______________________________________________ Bug-coreutils mailing list Bug-coreutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-coreutils
