[Bug AWT/16825] New: Bad results for QuadCurve2D.solveQuadratic

[Bug AWT/16825] New: Bad results for QuadCurve2D.solveQuadratic

[mark at gcc dot gnu dot org]

The method java.awt.geom.QuadCurve2D.solveQuadratic sometimes does not return
the right results for the following equation:

(x^2)/10 + 20x + 1000 = 0

Expected result: -100

This equation is tested by Mauve; see gnu.testlet.java.awt.geom.QuadCurve2D.

The problem occurs with various versions of gcj, compiling for IA-32. However,
it does NOT occur when executing the same Mauve testlet on jamvm, using the same
implementation for java.awt.geom.QuadCurve2D.

It turned out that a small Java test program for computing the equation's
discriminant gives a very small negative number when run on gcj/IA-32, but
exactly zero when run on jamvm or the Sun J2SE 1.4.1_01.

This bug report comes from GNU Classpath:
https://savannah.gnu.org/bugs/index.php?func=detailitem&item_id=7123

(Please see the very large audit trail there for more information.)

-- 
           Summary: Bad results for QuadCurve2D.solveQuadratic
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: AWT
        AssignedTo: fitzsim at redhat dot com
        ReportedBy: mark at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org,java-prs at gcc dot gnu
                    dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16825

[Bug AWT/16825] Bad results for QuadCurve2D.solveQuadratic

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-07-29 21:16 -------
Confirmed, I have seen the discussion of this bug before somewhere but I had forgot where, I think the 
java-patches or the java list.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-07-29 21:16:25
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16825

[Bug AWT/16825] Bad results for QuadCurve2D.solveQuadratic

[fitzsim at redhat dot com]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|fitzsim at redhat dot com   |graydon at redhat dot com
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16825

[Bug AWT/16825] Bad results for QuadCurve2D.solveQuadratic

[sven at physto dot se]

------- Additional Comments From sven at physto dot se  2004-11-09 23:22 -------
This isn't a Awt or libgcj bug itself, but rather an effect of the extended-mode
of FPU being the default leading to inconsistent results with 64-bit doubles. 

A suggestion for a 'fix' is to set the FPU to 64-bit double rounding for Java
programs, since Java does not have a 'long double' type. This would both
stricter and give more consistent results with other platforms.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16825

[Bug AWT/16825] Bad results for QuadCurve2D.solveQuadratic

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-11-09 23:32 -------
Note the java "standard" does not require the FPU set to double-precision except when working in strict 
mode (this was changed after 1.0 IIRC).

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16825

[Bug awt/16825] Bad results for QuadCurve2D.solveQuadratic

[fitzsim at redhat dot com]

------- Additional Comments From fitzsim at redhat dot com  2005-08-24 19:43 -------
*** Bug 22724 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |from-classpath at savannah
                   |                            |dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16825

[Bug awt/16825] need to reimplement QuadCurve2D.solveQuadratic

[fitzsim at redhat dot com]

------- Additional Comments From fitzsim at redhat dot com  2005-08-24 23:49 -------
We need to reimplement QuadCurve2D.solveQuadratic, possibly using the algorithm
described here: http://www.library.cornell.edu/nr/bookcpdf/c5-6.pdf

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Bad results for             |need to reimplement
                   |QuadCurve2D.solveQuadratic  |QuadCurve2D.solveQuadratic


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16825