------- Comment #14 from hjl dot tools at gmail dot com 2010-02-21 18:00 ------- That is true that the previous release didn't have proper excess precision semantics. But from the user perspective, the previous release handles
--
float f5(float x, int y)
{
return x * y;
}
--
correctly with "-Wconversion -std=c99". For this particular testcase,
gcc 4.5 has a user visible regression.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43128
