c++/5060

[CHAILAN lionel]

Thanks,

I've just read in "What every computer scientist ..." from D. Golberg,
about extended precision, that such test could involve one double (x),
and
some more precise real (o_x+z) due to x86 floating point architecture.
I'll try some other compiler, but I don't think now the problem is
relevant yet.



http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=5060&database=gcc