[Issue 6505] New: Wrong code for expression involving 8 floats, only with -O

[d-bugmail at puremagic.com]

http://d.puremagic.com/issues/show_bug.cgi?id=6505

           Summary: Wrong code for expression involving 8 floats, only
                    with -O
           Product: D
           Version: D1 & D2
          Platform: Other
        OS/Version: Windows
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: DMD
        AssignedTo: nobody at puremagic.com
        ReportedBy: clugdbug at yahoo.com.au


--- Comment #0 from Don <clugdbug at yahoo.com.au> 2011-08-16 04:33:20 PDT ---
Found this when I removed a reference to std.random from the DMD test suite.
This bug is in test22.d. It was originally described in:
http://www.digitalmars.com/d/archives/digitalmars/D/bugs/4766.html
but wasn't reduced correctly -- it has never been fixed.
It only applies when compiling with -O: the result is -nan.
----
real randx()
{
    return 1.2;
}

void test1()
{
    float x10=randx();
    float x11=randx();
    float x20=randx();
    float x21=randx();
    float y10=randx();
    float y11=randx();
    float y20=randx();
    float y21=randx();

    float tmp=(
    x20*x21 + y10*y10 + y10*y11 + y11*y11 + 
    y11*y20 + y20*y20 + y10*y21 + y11*y21 +
    y21*y21);
    assert(tmp > 0);
}

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------