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

[Adrian]

Follow-up Comment #5, bug #53733 (project octave):

@Dimitri/comment#4: The ClassNotFoundException seems to indicate
that the class is not instantiated, maybe that is why you see no
segfault (running under gdb I suppose). Do you have Baz.class in
the pwd ?

@Mike/comment#3: Thanks for further narrowing things down. So if
I understand correctly loading the class alone already causes
some bad memory accesses, and the other instructions were just
necessary to reveal the memory corruption when run outside gdb ? That
would explain why this bug for a long time seemed almost non
reproducible to me, striking unexpectedly at different moments
when working interactively.

Just to be sure: you remove the octave-workspace before running
your minimal example ? I found that the presence of a workspace
resulting from a previous crash can change the point at which the
segmentation fault hits (my first attempt at a minimal example
did not crash anymore once I had removed the workspace).

Sorry for the confusion regarding the image package, somehow I
thought it provided imwrite, I should have checked.


    _______________________________________________________

Reply to this item at:

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

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/