Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

#3 FAIL: gcc.c-torture/execute/conversion.c execution

gcc
open
nobody
FAIL (33)
5
2015-02-14
2008-04-21
NightStrike
No

FAIL: gcc.c-torture/execute/conversion.c execution, -O0
Executing on host: /var/tmp/build/gcc-svn/build-x86_64-pc-linux/gcc/xgcc -B/var/tmp/build/gcc-svn/build-x86_64-pc-linux/gcc/ /var/tmp/build/gcc-svn/gcc/gcc/
testsuite/gcc.c-torture/execute/conversion.c -w -O1 -fno-show-column -lm -o /var/tmp/build/gcc-svn/build-x86_64-pc-linux/gcc/testsuite/gcc/conversion.
x1 (timeout = 300)
PASS: gcc.c-torture/execute/conversion.c compilation, -O1
Executing on mingw: chmod +x /tmp/conversion.x1.2433 (timeout = 300)
XYZ0ZYX
Executing on mingw: /tmp/conversion.x1.2433 (timeout = 300)

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
XYZ3ZYX
Executing on mingw: rm -f /tmp/conversion.x1.2433.o /tmp/conversion.x1.2433 (timeout = 300)
XYZ0ZYX

conversion.c
------------

/* Test front-end conversions, optimizer conversions, and run-time
conversions between different arithmetic types.

Constants are specified in a non-obvious way to make them work for
any word size. Their value on a 32-bit machine is indicated in the
comments.

Note that this code is NOT intended for testing of accuracy of fp
conversions. */

float
u2f(u)
unsigned int u;
{
return u;
}

double
u2d(u)
unsigned int u;
{
return u;
}

long double
u2ld(u)
unsigned int u;
{
return u;
}

float
s2f(s)
int s;
{
return s;
}

double
s2d(s)
int s;
{
return s;
}

long double
s2ld(s)
int s;
{
return s;
}

int
fnear (float x, float y)
{
float t = x - y;
return t == 0 || x / t > 1000000.0;
}

int
dnear (double x, double y)
{
double t = x - y;
return t == 0 || x / t > 100000000000000.0;
}

int
ldnear (long double x, long double y)
{
long double t = x - y;
return t == 0 || x / t > 100000000000000000000000000000000.0;
}

test_integer_to_float()
{
if (u2f(0U) != (float) 0U) /* 0 */
abort();
if (!fnear (u2f(~0U), (float) ~0U)) /* 0xffffffff */
abort();
if (!fnear (u2f((~0U) >> 1), (float) ((~0U) >> 1))) /* 0x7fffffff */
abort();
if (u2f(~((~0U) >> 1)) != (float) ~((~0U) >> 1)) /* 0x80000000 */
abort();

if (u2d(0U) != (double) 0U) /* 0 */
abort();
if (!dnear (u2d(~0U), (double) ~0U)) /* 0xffffffff */
abort();
if (!dnear (u2d((~0U) >> 1),(double) ((~0U) >> 1))) /* 0x7fffffff */
abort();
if (u2d(~((~0U) >> 1)) != (double) ~((~0U) >> 1)) /* 0x80000000 */
abort();

if (u2ld(0U) != (long double) 0U) /* 0 */
abort();
if (!ldnear (u2ld(~0U), (long double) ~0U)) /* 0xffffffff */
abort();
if (!ldnear (u2ld((~0U) >> 1),(long double) ((~0U) >> 1))) /* 0x7fffffff */
abort();
if (u2ld(~((~0U) >> 1)) != (long double) ~((~0U) >> 1)) /* 0x80000000 */
abort();

if (s2f(0) != (float) 0) /* 0 */
abort();
if (!fnear (s2f(~0), (float) ~0)) /* 0xffffffff */
abort();
if (!fnear (s2f((int)((~0U) >> 1)), (float)(int)((~0U) >> 1))) /* 0x7fffffff */
abort();
if (s2f((int)(~((~0U) >> 1))) != (float)(int)~((~0U) >> 1)) /* 0x80000000 */
abort();

if (s2d(0) != (double) 0) /* 0 */
abort();
if (!dnear (s2d(~0), (double) ~0)) /* 0xffffffff */
abort();
if (!dnear (s2d((int)((~0U) >> 1)), (double)(int)((~0U) >> 1))) /* 0x7fffffff */
abort();
if (s2d((int)~((~0U) >> 1)) != (double)(int)~((~0U) >> 1)) /* 0x80000000 */
abort();

if (s2ld(0) != (long double) 0) /* 0 */
abort();
if (!ldnear (s2ld(~0), (long double) ~0)) /* 0xffffffff */
abort();
if (!ldnear (s2ld((int)((~0U) >> 1)), (long double)(int)((~0U) >> 1))) /* 0x7fffffff */
abort();
if (s2ld((int)~((~0U) >> 1)) != (long double)(int)~((~0U) >> 1)) /* 0x80000000 */
abort();
}

#if __GNUC__
float
ull2f(u)
unsigned long long int u;
{
return u;
}

double
ull2d(u)
unsigned long long int u;
{
return u;
}

long double
ull2ld(u)
unsigned long long int u;
{
return u;
}

float
sll2f(s)
long long int s;
{
return s;
}

double
sll2d(s)
long long int s;
{
return s;
}

long double
sll2ld(s)
long long int s;
{
return s;
}

test_longlong_integer_to_float()
{
if (ull2f(0ULL) != (float) 0ULL) /* 0 */
abort();
if (ull2f(~0ULL) != (float) ~0ULL) /* 0xffffffff */
abort();
if (ull2f((~0ULL) >> 1) != (float) ((~0ULL) >> 1)) /* 0x7fffffff */
abort();
if (ull2f(~((~0ULL) >> 1)) != (float) ~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (ull2d(0ULL) != (double) 0ULL) /* 0 */
abort();
#if __HAVE_68881__
/* Some 68881 targets return values in fp0, with excess precision.
But the compile-time conversion to double works correctly. */
if (! dnear (ull2d(~0ULL), (double) ~0ULL)) /* 0xffffffff */
abort();
if (! dnear (ull2d((~0ULL) >> 1), (double) ((~0ULL) >> 1))) /* 0x7fffffff */
abort();
#else
if (ull2d(~0ULL) != (double) ~0ULL) /* 0xffffffff */
abort();
if (ull2d((~0ULL) >> 1) != (double) ((~0ULL) >> 1)) /* 0x7fffffff */
abort();
#endif
if (ull2d(~((~0ULL) >> 1)) != (double) ~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (ull2ld(0ULL) != (long double) 0ULL) /* 0 */
abort();
if (ull2ld(~0ULL) != (long double) ~0ULL) /* 0xffffffff */
abort();
if (ull2ld((~0ULL) >> 1) != (long double) ((~0ULL) >> 1)) /* 0x7fffffff */
abort();
if (ull2ld(~((~0ULL) >> 1)) != (long double) ~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (sll2f(0LL) != (float) 0LL) /* 0 */
abort();
if (sll2f(~0LL) != (float) ~0LL) /* 0xffffffff */
abort();
if (! fnear (sll2f((long long int)((~0ULL) >> 1)), (float)(long long int)((~0ULL) >> 1))) /* 0x7fffffff */
abort();
if (sll2f((long long int)(~((~0ULL) >> 1))) != (float)(long long int)~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (sll2d(0LL) != (double) 0LL) /* 0 */
abort();
if (sll2d(~0LL) != (double) ~0LL) /* 0xffffffff */
abort();
if (!dnear (sll2d((long long int)((~0ULL) >> 1)), (double)(long long int)((~0ULL) >> 1))) /* 0x7fffffff */
abort();
if (! dnear (sll2d((long long int)~((~0ULL) >> 1)), (double)(long long int)~((~0ULL) >> 1))) /* 0x80000000 */
abort();

if (sll2ld(0LL) != (long double) 0LL) /* 0 */
abort();
if (sll2ld(~0LL) != (long double) ~0LL) /* 0xffffffff */
abort();
if (!ldnear (sll2ld((long long int)((~0ULL) >> 1)), (long double)(long long int)((~0ULL) >> 1))) /* 0x7fffffff */
abort();
if (! ldnear (sll2ld((long long int)~((~0ULL) >> 1)), (long double)(long long int)~((~0ULL) >> 1))) /* 0x80000000 */
abort();
}
#endif

unsigned int
f2u(float f)
{
return (unsigned) f;
}

unsigned int
d2u(double d)
{
return (unsigned) d;
}

unsigned int
ld2u(long double d)
{
return (unsigned) d;
}

int
f2s(float f)
{
return (int) f;
}

int
d2s(double d)
{
return (int) d;
}

int
ld2s(long double d)
{
return (int) d;
}

test_float_to_integer()
{
if (f2u(0.0) != 0)
abort();
if (f2u(0.999) != 0)
abort();
if (f2u(1.0) != 1)
abort();
if (f2u(1.99) != 1)
abort();
if (f2u((float) ((~0U) >> 1)) != (~0U) >> 1 && /* 0x7fffffff */
f2u((float) ((~0U) >> 1)) != ((~0U) >> 1) + 1)
abort();
if (f2u((float) ~((~0U) >> 1)) != ~((~0U) >> 1)) /* 0x80000000 */
abort();

/* These tests require double precision, so for hosts that don't offer
that much precision, just ignore these test. */
if (sizeof (double) >= 8) {
if (d2u(0.0) != 0)
abort();
if (d2u(0.999) != 0)
abort();
if (d2u(1.0) != 1)
abort();
if (d2u(1.99) != 1)
abort();
if (d2u((double) (~0U)) != ~0U) /* 0xffffffff */
abort();
if (d2u((double) ((~0U) >> 1)) != (~0U) >> 1) /* 0x7fffffff */
abort();
if (d2u((double) ~((~0U) >> 1)) != ~((~0U) >> 1)) /* 0x80000000 */
abort();
}

/* These tests require long double precision, so for hosts that don't offer
that much precision, just ignore these test. */
if (sizeof (long double) >= 8) {
if (ld2u(0.0) != 0)
abort();
if (ld2u(0.999) != 0)
abort();
if (ld2u(1.0) != 1)
abort();
if (ld2u(1.99) != 1)
abort();
if (ld2u((long double) (~0U)) != ~0U) /* 0xffffffff */
abort();
if (ld2u((long double) ((~0U) >> 1)) != (~0U) >> 1) /* 0x7fffffff */
abort();
if (ld2u((long double) ~((~0U) >> 1)) != ~((~0U) >> 1)) /* 0x80000000 */
abort();
}

if (f2s(0.0) != 0)
abort();
if (f2s(0.999) != 0)
abort();
if (f2s(1.0) != 1)
abort();
if (f2s(1.99) != 1)
abort();
if (f2s(-0.999) != 0)
abort();
if (f2s(-1.0) != -1)
abort();
if (f2s(-1.99) != -1)
abort();
if (f2s((float)(int)~((~0U) >> 1)) != (int)~((~0U) >> 1)) /* 0x80000000 */
abort();

/* These tests require double precision, so for hosts that don't offer
that much precision, just ignore these test. */
if (sizeof (double) >= 8) {
if (d2s(0.0) != 0)
abort();
if (d2s(0.999) != 0)
abort();
if (d2s(1.0) != 1)
abort();
if (d2s(1.99) != 1)
abort();
if (d2s(-0.999) != 0)
abort();
if (d2s(-1.0) != -1)
abort();
if (d2s(-1.99) != -1)
abort();
if (d2s((double) ((~0U) >> 1)) != (~0U) >> 1) /* 0x7fffffff */
abort();
if (d2s((double)(int)~((~0U) >> 1)) != (int)~((~0U) >> 1)) /* 0x80000000 */
abort();
}

/* These tests require long double precision, so for hosts that don't offer
that much precision, just ignore these test. */
if (sizeof (long double) >= 8) {
if (ld2s(0.0) != 0)
abort();
if (ld2s(0.999) != 0)
abort();
if (ld2s(1.0) != 1)
abort();
if (ld2s(1.99) != 1)
abort();
if (ld2s(-0.999) != 0)
abort();
if (ld2s(-1.0) != -1)
abort();
if (ld2s(-1.99) != -1)
abort();
if (ld2s((long double) ((~0U) >> 1)) != (~0U) >> 1) /* 0x7fffffff */
abort();
if (ld2s((long double)(int)~((~0U) >> 1)) != (int)~((~0U) >> 1)) /* 0x80000000 */
abort();
}
}

#if __GNUC__
unsigned long long int
f2ull(float f)
{
return (unsigned long long int) f;
}

unsigned long long int
d2ull(double d)
{
return (unsigned long long int) d;
}

unsigned long long int
ld2ull(long double d)
{
return (unsigned long long int) d;
}

long long int
f2sll(float f)
{
return (long long int) f;
}

long long int
d2sll(double d)
{
return (long long int) d;
}

long long int
ld2sll(long double d)
{
return (long long int) d;
}

test_float_to_longlong_integer()
{
if (f2ull(0.0) != 0LL)
abort();
if (f2ull(0.999) != 0LL)
abort();
if (f2ull(1.0) != 1LL)
abort();
if (f2ull(1.99) != 1LL)
abort();
if (f2ull((float) ((~0ULL) >> 1)) != (~0ULL) >> 1 && /* 0x7fffffff */
f2ull((float) ((~0ULL) >> 1)) != ((~0ULL) >> 1) + 1)
abort();
if (f2ull((float) ~((~0ULL) >> 1)) != ~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (d2ull(0.0) != 0LL)
abort();
if (d2ull(0.999) != 0LL)
abort();
if (d2ull(1.0) != 1LL)
abort();
if (d2ull(1.99) != 1LL)
abort();
if (d2ull((double) ((~0ULL) >> 1)) != (~0ULL) >> 1 && /* 0x7fffffff */
d2ull((double) ((~0ULL) >> 1)) != ((~0ULL) >> 1) + 1)
abort();
if (d2ull((double) ~((~0ULL) >> 1)) != ~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (ld2ull(0.0) != 0LL)
abort();
if (ld2ull(0.999) != 0LL)
abort();
if (ld2ull(1.0) != 1LL)
abort();
if (ld2ull(1.99) != 1LL)
abort();
if (ld2ull((long double) ((~0ULL) >> 1)) != (~0ULL) >> 1 && /* 0x7fffffff */
ld2ull((long double) ((~0ULL) >> 1)) != ((~0ULL) >> 1) + 1)
abort();
if (ld2ull((long double) ~((~0ULL) >> 1)) != ~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (f2sll(0.0) != 0LL)
abort();
if (f2sll(0.999) != 0LL)
abort();
if (f2sll(1.0) != 1LL)
abort();
if (f2sll(1.99) != 1LL)
abort();
if (f2sll(-0.999) != 0LL)
abort();
if (f2sll(-1.0) != -1LL)
abort();
if (f2sll(-1.99) != -1LL)
abort();
if (f2sll((float)(long long int)~((~0ULL) >> 1)) != (long long int)~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (d2sll(0.0) != 0LL)
abort();
if (d2sll(0.999) != 0LL)
abort();
if (d2sll(1.0) != 1LL)
abort();
if (d2sll(1.99) != 1LL)
abort();
if (d2sll(-0.999) != 0LL)
abort();
if (d2sll(-1.0) != -1LL)
abort();
if (d2sll(-1.99) != -1LL)
abort();
if (d2sll((double)(long long int)~((~0ULL) >> 1)) != (long long int)~((~0ULL) >> 1)) /* 0x80000000 */
abort();

if (ld2sll(0.0) != 0LL)
abort();
if (ld2sll(0.999) != 0LL)
abort();
if (ld2sll(1.0) != 1LL)
abort();
if (ld2sll(1.99) != 1LL)
abort();
if (ld2sll(-0.999) != 0LL)
abort();
if (ld2sll(-1.0) != -1LL)
abort();
if (ld2sll(-1.99) != -1LL)
abort();
if (ld2sll((long double)(long long int)~((~0ULL) >> 1)) != (long long int)~((~0ULL) >> 1)) /* 0x80000000 */
abort();
}
#endif

main()
{
test_integer_to_float();
test_float_to_integer();
#if __GNUC__
test_longlong_integer_to_float();
test_float_to_longlong_integer();
#endif
exit(0);
}

Discussion

  • Kai Tietz
    Kai Tietz
    2008-09-23

    This one seems to be solved by side-effect.