Known Bugs in Gmaj
The following anomalies in Gmaj's display and behavior have been
observed, but not yet resolved.  Some of these are intermittent,
and some may depend on your computer platform and/or version of
Java.
If you experience any of these (or similar) problems, it may help
to switch to a different version of Java.  For example, the first
bug listed below (phantom repeats) seems to have been introduced
with Sun's J2SE 1.4.x series, so you may be better off with 1.3.x
or 1.5.x.  You can also try switching to a different vendor's
implementation, but then you are on your own, as we develop and
test Gmaj using Sun's official API and SDK.  Currently Gmaj does
not rely on any features beyond Java 1.3, in order to preserve
version flexibility for our users.
Note that anomalies that depend on the Java version could be due
to bugs in Java itself, but could also be bugs in Gmaj that only
"matter" with certain implementations of Java.  In either case,
however, with further investigation it may be possible to work
around the problem in future releases of Gmaj.
- 	Phantom repeats. 
	After zooming in on a small region, the "repeats" panel
	sometimes shows a spurious extra bar superimposed on the
	legitimate features across the entire width of the window,
	or the features may disappear entirely.
 - 	Missing zoom box and crosshairs. 
	With some datasets (large ones especially) the lines for the
	zoom rectangle and crosshairs may not display.  It still
	zooms OK though, if you just pretend the lines are there.
	Update: This appears to have been fixed in Java 1.5.
 - 	Can't copy text from dialog messages in applet. 
	In applet mode, you may not be able to select and copy text
	from pop-up message boxes to the system clipboard for
	pasting elsewhere.  This is apparently due to a bug in Java;
	hopefully Sun will fix it eventually.
 - 	Patchy rendering of mark circle and/or red segments. 
	When using the Large Fonts option with Mark Color Varies,
	the red circle and plot segments may not appear as solid
	as they should.  This is because Mark Color Varies is
	implemented using Java's XOR drawing mode, and the thick
	lines are built up from several thin ones, so whenever two
	red pixels overlap they turn white again.  Fixing this will
	likely involve a custom implementation of Mark Color Varies
	that doesn't use XOR.  In the meantime, switching to Mark
	Always Red when using Large Fonts should solve the problem
	(except that it will be invisible against red underlays).
 - 	Memory leak when closing or rebuilding windows. 
	When windows are closed or rebuilt (e.g. when loading data
	via File - Open, switching to Large Fonts, or even just
	displaying dialog boxes), not all of the memory they were
	using is being reclaimed, so Gmaj will gradually run out
	and have to be restarted.  How quickly this happens will
	depend on how much memory you have allocated to Java with
	the 
-Xmx
	switch, and the size of your dataset.
 
If you encounter any bugs not listed here, please report them to
.
It would be helpful if you can include the version of Gmaj you
are using (i.e., the JarDate line from the Help - About message),
your computer platform and version of Java, copies of the data
files you were trying to view, and a description of exactly what
you were doing when the problem occurred.  This will help us to
reproduce the problem so we can track it down and fix it.
Non-bug suggestions and feedback are also welcome.
Thank you for using Gmaj, and helping us to make it better.
Cathy Riemer, June 2008