Re: Bring Pytave into core

[Jesús Zazueta]

Hi. I'm also using the C++ api to control the Octave interpreter within the JVM, and I also get a seg fault if I don't call do_octave_at_exit().

However, if I do, the process does shut down Octave, but it also brings down the JVM as well.

Is there a method we could call to reset/do cleanup on the engine without terminating the host process?

Thanks for your time and help!

On 14 Apr 2015 10:19 am, "Mike Miller" <address@hidden> wrote:

On Thu, Apr 9, 2015 at 09:10:48 -0400, Jordi Gutiérrez Hermoso wrote:
> I started looking at building Pytave. It's very crufty, and much of it
> can be thankfully thrown out. I ran into problems with both Pytave and
> Octave having their own config.h file. I expect there will be further
> problems with the deprecation of the octave_map class, but these are
> not insurmountable problems.
>
> This weekend is US Pycon. I'll be meeting with Mike Miller, and maybe
> we can have an impromptu hacking session on Pytave, an appropriate
> endeavour for Pycon. :-)

Following up on this thread now that PyCon is over and we have some
initial progress to share. We did meet and talk about pytave briefly,
but there was not enough time for a hacking session.

Jordi was kind enough to set up a repository for us to start
collaborating on pytave, see http://hg.octave.org/pytave. I am also
hosting my own repository at https://hg.mtmxr.com/pytave.

I have pushed a series of (draft) changesets that allow pytave to
compile against the latest Octave. It works, but segfaults when the
python interpreter exits for now. I will make some notes about next
steps and share that soon.

Looking forward to making more progress soon,

--
mike