[Mono-bugs] [Bug 478479] New: New test_0_long_to_double_conversion is bogus

bugzilla_noreply at novell.com bugzilla_noreply at novell.com
Sat Feb 21 18:03:12 EST 2009


           Summary: New test_0_long_to_double_conversion is bogus
    Classification: Mono
           Product: Mono: Runtime
           Version: SVN
          Platform: PowerPC-64
        OS/Version: openSUSE 10.3
            Status: NEW
          Severity: Normal
          Priority: P5 - None
         Component: misc
        AssignedTo: mono-bugs at lists.ximian.com
        ReportedBy: munroesj at us.ibm.com
         QAContact: mono-bugs at lists.ximian.com
          Found By: ---

User-Agent:       Mozilla/5.0 (X11; U; Linux ppc64; en-US; rv:
Gecko/20081213 SUSE/ Firefox/

In mini/basic-float.cs the test:

    static int test_0_long_to_double_conversion ()
        long l = 9223372036854775807L;
        long conv = (long)((double)l);
        if (conv != -9223372036854775808l)
            return 1;

        return 0;

is invalid because long value 9223372036854775807 requires 63-bits but a
IEEE754 long only carries 53 bits precision. So the conversion will round to
9223372036854775808. When thins is converted back the result can not be
represented in a long and the result is implementation dependent.

>From the PowerISA-2.05 description of fctid[z]

If the operand in FRB is greater than 2**63 - 1, then FRT
is set to 0x7FFF_FFFF_FFFF_FFFF. If the operand in
FRB is less than -263, then FRT is set to

Net on conversion overflow; the instruction returns the largest representable
64-bit integer for the sign.

This is allowed in IEEE754.

This test assumes that conversion returns the largest negative value regardless
of the original sign.

Reproducible: Always

Steps to Reproduce:
1. autogen
2. make 
3. make check
Actual Results:  
Test run:
Test 'test_0_long_to_double_conversion' failed result (got 1, expected 0).
Results: total tests: 32, failed: 1, cfailed: 0 (pass: 96.88%)
Elapsed time: 0.003475 secs (0.001053, 0.002422), Code size: 9312

Expected Results:  
Test run:
Results: total tests: 32, all pass.

Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.

More information about the mono-bugs mailing list