NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
Hi Don, > We are having some problems with the latest release of > visad.jar (Wed Aug 9 12:57:31 CDT 2000) on Linux. If > we click in any VisAD display window, the program crashes > with a sigseg violation. The specifics of our installation > are: > > jdk1.2.2 from Blackdown > java3d1_1_3 > > If we use a previous visad.jar (Fri Jul 21 10:38:54 CDT 2000) > we do not have this problem. This is undoubtedly linked to the code that I added to visad/MouseHelper.java for you recently, so that translating the display with the mouse would be precise. Specifically, the block of code: VisADRay start_ray = behavior.findRay(start_x, start_y); VisADRay start_ray_x = behavior.findRay(start_x + 100, start_y); VisADRay start_ray_y = behavior.findRay(start_x, start_y + 100); xmul = 0.01 * (start_ray_x.position[0] - start_ray.position[0]); ymul = 0.01 * (start_ray_y.position[1] - start_ray.position[1]); tstart = proj.getMatrix(); double[] rot = new double[3]; double[] scale = new double[1]; double[] trans = new double[3]; behavior.instance_unmake_matrix(rot, scale, trans, tstart); xmul = xmul * scale[0]; ymul = ymul * scale[0]; added after:" if (mousePressed1 || mouseCombo1) { and after: else if (mousePressed2 || mouseCombo2) { is probably triggering some sort of bug in the Linux Java implementation (Java code isn't supposed to cause segmentation faults, as you know). So might try rearranging this code I added, or even testing to see if its dividing by zero or something like that. If you come up with a fix, feel free to update the system. Cheers, Bill ---------------------------------------------------------- Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI 53706 hibbard@xxxxxxxxxxxxxxxxx 608-263-4427 fax: 608-263-6738 http://www.ssec.wisc.edu/~billh/vis.html
visad
archives: