failure 55:

Requirement failed: {fdiml.08.01.01} Function executed successfully with EOK error code, while `isOverflow_Unifloat( model_res ) == true` means that ERANGE SHALL be set

location
trace	/var/opt/lsb/test/olver-core/2013-07-25_08-37-08/math_real_fdim_overflow_scenario.utz (math_real_fdim_overflow_scenario.utt), line 134
occurence
scenario	math_real_fdim_overflow_scenario
specification function	fdim_spec()
parameter value	CallContext context = [pid=22161,thr=b752d700]
parameter value	Unifloat * @x = (long double) 1.1897314953572317650212638530309702e+4932 (0x.FFFFFFFFFFFFFFFFP16384L)
parameter value	Unifloat * x = (long double) 1.1897314953572317650212638530309702e+4932 (0x.FFFFFFFFFFFFFFFFP16384L)
parameter value	Unifloat * @y = (long double) -1.1897314953572317650212638530309702e+4932 (-0x.FFFFFFFFFFFFFFFFP16384L)
parameter value	Unifloat * y = (long double) -1.1897314953572317650212638530309702e+4932 (-0x.FFFFFFFFFFFFFFFFP16384L)
parameter value	ErrorCode * @errno = EOK
parameter value	ErrorCode * errno = EOK
return value	(Unifloat *) (long double) Infinity
coverage & branch	C x and y - finite values
properties
REQ failed	fdiml.08.01.01
expected value	(long double) nan (0x.FFFFFFFFFFFFFFFFP16385L)
kind	POSTCONDITION_FAILED

similar known bug(s)
The fdim(x, y) family of functions shall determine the positive difference between their arguments. According to LSB, if the result overflows then fdim() shall fail and shall set ERANGE error code. But on the target machine, the functions set error code EOK instead of ERANGE when overflow occurs.