> Here's a bug that causes equivalent statements to give a slightly > different value to a double. Thanks for your bug report. I believe there is no bug in the compiler, see http://gcc.gnu.org/cgi-bin/fom.cgi?file=49 Regards, Martin