| Summary: | cimagf gives incorrect results | ||
|---|---|---|---|
| Product: | glibc | Reporter: | Ismail "cartman" Donmez <null> |
| Component: | math | Assignee: | Andreas Jaeger <aj> |
| Status: | RESOLVED INVALID | ||
| Severity: | normal | CC: | glibc-bugs |
| Priority: | P2 | Flags: | fweimer:
security-
|
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Host: | i686-gnu-linux | Target: | |
| Build: | Last reconfirmed: | ||
Possibly a related problem:
#include <complex.h>
#include <stdio.h>
int main() {
float complex volatile z;
z = 1.2 - 4.0*I;
z = clogf(z);
printf("%lf\n", cimagf(z));
return 0;
}
prints -0.000029 instead of -1.2793396 .
-> Assign to math component math regression tests fail hence this is likely a gcc 4.3 problem instead. Closing as Invalid. |
This is a stripped down testcase from gfortran testcases converted to C : #include <complex.h> #include <stdio.h> int main() { float complex volatile z, one; z = 5.0; one = 1.0; z = cpowf (z, one); printf("%f\n", (double)cimagf(z)); return 0; } prints -0.000141 with libc 2.7 where as it prins 0.000000 with libc 2.3.6 and libc 2.4.11 on i686 host.