Octave getting killed in an improper way? about oct2py HOT 6 CLOSED

Here is the whole traceback from glibc:

https://gist.githubusercontent.com/cel4/1148fe7b22f966fdb9f9/raw/f5ff0b8e238cf21e2db53faeebe006ad30e12c64/gistfile1.txt.

Probably not too interesting for you though, since the error is raised in octave and not in your python code.

Last two lines are part of the python stacktrace.

Edit: This bug was introduced by commit b2c14d4.

from oct2py.

I think I have it licked now, please try against the latest, 8650820.

from oct2py.

Unfortunately, it's not fixed yet.

I think the problem is that sending self.write('exit\n') and proc.terminate() afterwards can make octave processes non-responsive.

Using one of these two termination strategies works perfectly well, but the two together lead to the observed glibc crash and some surviving zombie processes.

I guess it would be the best to rely on proc.terminate()'s SIGTERM signal as sole termination strategy.

from oct2py.

We good now?

from oct2py.

Yes, looks good, thanks!

from oct2py.

Excellent, thanks again for the help!

from oct2py.