JTF > bugs

This page reports the status of all bugs in the ACM Java Libraries that have been reported to the Java Task Force. To report a new bug, please send mail to jtf@cs.stanford.edu.

Bugs are listed chronologically in order of the date on which they were reported. Clicking on the triangle to the left of the bug entry displays the details.

•  JTFBug 2006-001: Incorrect handling of mouse events  [Fixed]
  Reported by: Chris Nevison on 26-Jul-06
  Fixed by: Eric Roberts on 31-Jul-06
  Affected modules: GCanvas, GCompound
  Description:

  This bug appeared in code designed to streamline the handling of mouse events. As an efficiency enhancement, the implementation maintains a list of items for which mouse listeners have been enabled. That list, however, was not being properly maintained, which meant that mouse events were sometimes lost.

•  JTFBug 2006-002: Precompiled images were not being cached  [Fixed]
  Fixed by: Eric Roberts on 30-Sep-06
  Affected modules: MediaTools
  Description:

  Unlike images read from files, precompiled images generated using the MediaTools facility were not being cached when the setCachingEnabled flag was set to true. Students were unlikely to be affected by this bug, because caching is disabled by default.

•  JTFBug 2006-003: setImage did not reset image size  [Fixed]
  Fixed by: Eric Roberts on 21-Dec-06
  Affected modules: GImage
  Description:

  The implementation of the GImage class continued to believe the old size information for an image, even after setImage was used to reset the size. The revised code invalidates the cached size on any call to setImage and then repaints the canvas.

•  JTFBug 2007-001: decodeFont failed to copy existing properties  [Fixed]
  Fixed by: Eric Roberts on 2-Jan-07
  Affected modules: JTFTools
  Description:

  According to its documentation, the decodeFont method in JTFTools was supposed to retain any properties of the existing font that were not specified in the font string. The original implementation failed to do so in many cases, but the code was easy to fix.

•  JTFBug 2007-002: Programs failed to work in a headless environment  [Fixed]
  Reported by: Leland Beck on 10-Jan-07
  Fixed by: Eric Roberts on 30-Jan-07
  Affected modules: Program
  Description:

  The Program class is a subclass of JApplet and can therefore operate only in a Java environment with a graphics context. Java programs that use the command line, however, often do not have access to Java’s AWT (such programs are called headless in Java). To allow students to run simple programs in a headless environment, the startup code for the Program class now checks to see if the program is operating in such an environment and, if so, substitutes an equivalent CommandLineProgram, which does not require a graphics environment.

•  JTFBug 2007-004: Formatting problems in numeric fields  [Fixed]
  Fixed by: Eric Roberts on 2-Mar-07
  Affected modules: DoubleField, IntField
  Description:

  The original code for the DoubleField and IntField classes failed to ensure that the values would fit in the available space, which meant that the displayed value was sometimes clipped by the boundaries of the input field.

•  JTFBug 2007-005: Cross-file references caused compilation errors  [Fixed]
  Reported by: Leland Beck on 10-Apr-07
  Fixed by: Eric Roberts on 26-Apr-07
  Affected modules: GArc, GCompound, GOval, GRoundRect, GTurtle, GPen
  Description:

  The standard javac tool was unable to compile some applications because it could not determine that particular package-private classes were required. The problem arose only when the application was compiled from library sources and did not occur when compiling with the acm.jar file.

•  JTFBug 2007-006: Localization not performed for numeric fields  [Fixed]
  Fixed by: Eric Roberts on 6-May-07
  Affected modules: DoubleField, IntField
  Description:

  The original code for the DoubleField and IntField classes did not pay attention to the current locale and therefore could not interpret or display numbers in many countries.

•  JTFBug 2007-007: JDK 1.6 introduced incompatibilities with the Program class  [Fixed]
  Fixed by: Eric Roberts on 8-May-07
  Affected modules: Program
  Description:

  The release of JDK 1.6 (now called Java 6) created a serious incompatibility in the implementation of the Program class, which made it impossible to embed the actual JApplet object in the program frame. The current implementation transfers the content pane into the program frame instead.

•  JTFBug 2007-008: The Print Console menu item failed under JDK 1.6  [Fixed]
  Fixed by: Eric Roberts on 8-May-07
  Affected modules: IOConsole
  Description:

  The release of JDK 1.6 (now called Java 6) changed the model for printing so that the Print Console menu item in the File menu no longer worked correctly. The code was rewritten so that it was compatible with the new model.

•  JTFBug 2007-009: Rendered arcs were improperly synchronized  [Fixed]
  Fixed by: Eric Roberts on 28-May-07
  Affected modules: GArc
  Description:

  The code for the package-private ArcRenderer class did not correctly synchronize the rendering operation because of misplaced synchronize statements. The bug was uncovered in a code review, and it is not clear whether it actually caused problems in practice.

•  JTFBug 2007-010: Centering of the GPolygon class was incorrect  [Fixed]
  Reported by: Steve Wolfman on 10-Jun-07
  Fixed by: Eric Roberts on 6-Jul-07
  Affected modules: GPolygon
  Description:

  The code for recentering a GPolygon incorrectly calculated the new center because the calculation used an x coordinate value that should have been the corresponding y coordinate.

•  JTFBug 2007-011: Incorrect javadoc for the trig functions in GMath  [Fixed]
  Reported by: William Slough on 15-Aug-07
  Fixed by: Eric Roberts on 15-Aug-07
  Affected modules: GMath
  Description:

  The javadoc comments for the trigonometric functions sinDegrees, cosDegrees, and tanDegrees all indicated that the function calculated the sine of the angle instead of the appropriate function.

•  JTFBug 2007-012: Potential deadlock in applet thread handling  [Fixed]
  Fixed by: Eric Roberts on 10-Sep-07
  Affected modules: Program
  Description:

  A change to the security manager handling of applet threads introduced in JDK 1.6 (now Java 6) introduced a potential deadlock in the creation of the thread used for the run method.

•  JTFBug 2007-013: Change listeners were incorrectly handled in Animator  [Fixed]
  Fixed by: Eric Roberts on 12-Sep-07
  Affected modules: Animator
  Description:

  The implementation of the ChangeListener handler contained a bug that kept it from working correctly. This code, fortunately, was not used much by clients and was therefore unlikely to be noticed.

•  JTFBug 2008-001: TableLayout sometimes incorrectly computed cell dimensions  [Fixed]
  Fixed by: Eric Roberts on 8-Jan-08
  Affected modules: TableLayout
  Description:

  The implementation of TableLayout did not correctly take into account the gridwidth and gridheight options when computing cell dimensions.

•  JTFBug 2008-002: isAppletMode returned wrong answer during startup  [Fixed]
  Fixed by: Eric Roberts on 21-May-08
  Affected modules: Program
  Description:

  The isAppletMode method in the Program class returned the wrong answer if it was called from the constructor. The implementation now determines whether the code is in a browser by examining the call stack, which gives the correct answer even before the initialization code is complete.

•  JTFBug 2008-003: Ambiguous overrides of getWidth and getHeight  [Fixed]
  Fixed by: Eric Roberts on 10-Jun-08
  Affected modules: Program
  Description:

  The original implementation of the Program class overrode the definitions of getWidth and getHeight so that they returned the dimensions of the central region rather than the program frame as a whole. While this interpretation made it easier for students to write graphics code, it depended on the fact that the implementation of the Swing library did not itself call these methods. That assumption was true in Java 5 but not in Java 6. To avoid requiring changes to existing student code, the implementation now checks to see whether these methods have been called from inside the java and javax package hierarchy and, if so, restores their traditional semantics.