jGRASP An Integrated Development Environment with Visualizations for Improving Software Comprehensibility


Home
Download
Contact Us
Team Members
Resources
Archive
Privacy Policy
Support jGRASP

Documentation

jGRASP Help
On-line Papers
Tips
FAQ
Known Bugs
Version History
Future Plans
License

Intro Videos

Getting Started
Interactions

Tutorials (PDF)

Overview
Installation
Getting Started
Objects First
Interactions
CSD
Debugger
Projects
UML
Workbench
Viewers
All (zipped)

Sponsors
NSF

     Known Bugs

Click here for the bug-fix history.



Our Known Bugs

Interactions is missing some Java language features: static import, uninitialized variables (interactions locals are now initialized as if they were fields), generics.

Interactions does not allow the minimum long and integer constants to be used.

When running under 64-bit Java, file deletion from jGRASP will not move files to the recycle bin (a "permanent delete" warning is given), and J2ME installations are not automatically detected.

In the Java debugger and workbench, if a field exists in an implemented interface of the declared type of an object, and a field with the same name exists in the superclass but is "not inherited" according to JLS terminology (is declared private or is inaccessible from the child), the field declared in the interface will be shown as "not visible" (it will have a red bar through it), although it should be visible (jGRASP definition of "visible" is "available without casting"). This is due to Sun bug 6655791, but we may work around it in a future jGRASP release.

In the Java debugger and workbench, elements with replaced generic type are displayed with the replaced type (when known), but for "invoke method" and other operations, they are treated as the actual run-time type. Also, generic type parameters supplied when creating an instance are not thoroughly checked for validity, and arguments for "invoke method" or "create instance" are not checked for compile-time validity with respect to generic replacements or generic type parameter bounds. These problems don't limit the use of the debugger or workbench in any way, and they will be fixed in a future release.

On Windows, if documentation is viewed "outside of jGRASP", operations that should display javadoc documentation for fields or methods will only display the corresponding file (won't scroll to the field or method) for documentation on the local file system (Java library documentation on the web will work correctly). The built-in HTML viewer does not have this problem.

On Windows NT, a 16 bit Cygwin process that is run with the %CY or %CYD flag (console mode) can not be killed until the virtual DOS machine initializes and starts the process (which can take several seconds). An attempt to kill such a process will fail, but the process can be killed by closing the console, which will initially be iconified. This problem is not fatal and will rarely be encountered. It will not be fixed.



Bugs Caused by Java, Swing, or Window Manager Problems

These are difficult or impossible to work around. The eventual "fix" for them is (or will be) to get a newer version of the JDK. Many of them apply only to old versions of the JDK.

File chooser details view does not resize rows when the font size is changed under some JDKs (all file choosers).

On some JDK versions, the Help window will sometimes not come up, and a stack dump will be generated, the first time help is used. If this happens, help will not work until you shut down jGRASP and restart it.

Performance is bad and there is much unnecessary repainting with the Mac Look-And-Feel and Nimbus Look-And-Feel.

The help browser is not suitable for browsing the web. HTML display in Swing is limited and very buggy. The jGRASP help pages are constructed so as to work within the limitations and work around the bugs. This is somewhat better with newer versions of Java.

On some UNIX/Linux VMs, running jGRASP through the single-instance shell will cause the JVM to crash at shutdown. This seems to be harmless.

Memory leaks - due to Swing bugs, there are some small memory leaks. These have been minimized as much as was possible. This should not be a problem unless you run jGRASP for weeks without restarting. There may be larger memory leaks that we are unaware of.


           


Copyright © 2011 Auburn University
Valid HTML 4.0