[Octave-bug-tracker] [bug #53733] Panic/segfault when calling Java method after forge:im[read|write]

[Rik]

Follow-up Comment #54, bug #53733 (project octave):

I added the following to my local .gdbinit


handle SIGSEGV nostop print pass


In this case, I get a printout about the segmentation fault which is used
internally by JVM for overflow detection, but it doesn't cause gdb to
immediately abort.  I am also using the pass option, so a true SIGSEGV would
be passed to enclosing program which created the JVM if it were real.  But,
this doesn't happen indicating that it is a spurious thing.

I ran both the tiny test program, and also Octave, and neither issues more
than one SIGSEGV (associated with the startup of the JVM).  The test code in
Octave I used was 


javaaddpath(".")
imwrite(zeros(1000), "foo.png");
graph = javaObject("Baz");
graph.addNodes(1:1000000);


Also, without gdb the code above runs fine multiple times using


run-octave -f crashtest.m


Don't know if it was the case that the original reporter was running with the
'-f' flag to prevent user customizations in .octaverc which might actually be
triggering this.


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?53733>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/