The machine is running FC5test1 updated to 2006-01-04. Running standalone: cagney@tomago$ /notnfs/cagney/native/frysk-gui/frysk/gui/FryskGui could not create widget: java.lang.NullPointerException Exception in thread "main" java.lang.NullPointerException at frysk.gui.monitor.MainWindow.<init> (MainWindow.java:57) at frysk.gui.monitor.WindowManager.initWindows (WindowManager.java:75) at frysk.gui.FryskGui.<init> (FryskGui.java:118) at frysk.gui.FryskGui.mainGui (FryskGui.java:148) at frysk.gui.FryskGui.main (FryskGui.java:233) under gdb Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1208723792 (LWP 18942)] 0x0782114f in org.gnu.gtk.Widget.removePackageName(java.lang.Class) ( klass=@79af9c0) at src/java/org/gnu/gtk/Widget.java:1249 1249 int length = klass.getPackage().getName().length(); java.lang.Class.getPackage is returning <<NULL>> which, as far as I know, can't happen. -- The package "org.gnu.gtk" is no longer being registered when loaded (?) and consequently a search for it fails. Under RHEL-4 and FC-4, the org.gnu.gtk package was being registering itself so the above search worked. Tom's come up with a testcase, that I'll attach. It, however, also fails on both FC-4/RHEL-4 and FC-5. This suggests that there are two potential problems: the issue identified by the testcase; and the apparent regression.
Created attachment 122770 [details] Tom's test case
The fix is on svn head. I'll be checking it into the 4.1 branch shortly. I haven't looked at 4.0.x yet.
warren / jkeating asked for it to be added to FC5 blocker.
Created attachment 122775 [details] Regression test case (worked on 4.0.x broken in 4.1.x) This version shows the regression between 4.0.x and 4.1.x (which is now fixed)
Should be fixed in gcc-4.1.0-0.13 and later.