[Octave-bug-tracker] [bug #63971] Invoking Octave's Java subsystem crashes Octave

[Philip Nienhuis]

Update of bug #63971 (group octave):

                 Summary: [octave forge] (io) package crashes Octave =>
Invoking Octave's Java subsystem crashes Octave

    _______________________________________________________

Follow-up Comment #14:

Thanks for reporting back.
The result of a simple javachk function call makes it clear that it's Octave's
Java subsystem that somehow got crippled and the io package isn't the cause
(using it just uncovers the bug).
I'll change the bug title accordingly.

BTW please try the command
>> system ("java --version")
rather with one hyphen, i.e.,
>> system ("java -version")
and report back.

The exact Java version doesn't make much difference, on one of my boxes I have
a much older version running fine with Octave.
I tend to suspect that your IT dept has done something (policies or the like)
that inhibits Octave from properly invoking the Java subsystem. After all the
command
>> system ("java --version")
returns an error message emitted by Java (here Octave just invokes a Windows
command outside its own interpreter) which means that Java itself is working.

Then:
You didn't report back the results of this Octave command:

io_testscript ("oct", "io-test.xlsx")

(after having disabled the  __io_init__  function as per comment #3.
Would you please try that command?
If that works w/o the crash, it (= using the native "OCT" spreadsheet I/O
interface) would be a workaround for you.

Another thing that might help is to fiddle with the Java virtual machine (JVM)
settings. We have several related bug reports on Java and IIRC one or more
pointed to a problem with resetting the JVM memory to some unworkable setting.
But I'll have to dig this up.



    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?63971>

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