Modulo operation on float/double type creates incorrect c-code
double d = 23.0;
d %= 23.0;
expected output with 0.42.x:
gdouble d = 0.0;
gdouble _tmp2_;
gdouble _tmp3_;
d = 23.0;
_tmp2_ = d;
d = fmod (_tmp2_, 23.0);
broken output with 0.43.x:
gdouble d = 0.0;
d = 23.0;
d %= 23.0;