MATLAB R2016b segmentation-faults when started from emacs

MATLAB R2016b segmentation-faults when started from emacs

[Peter Mao]

[-- Attachment #1: Type: text/plain, Size: 843 bytes --]

Apologies if this is not the right place to post this issue, but I couldn't
find a better place!

This problem was originally posted to the matlab.el project on
SourceForge.  In summary, the latest version of Matlab crashes when started
from "shell" or "term" within emacs (I haven't tried e-shell yet, but I can
check it tomorrow morning).  The program starts up without any issues
either from the X11 terminal or OSX's terminal.  In my case, it's always
using /bin/bash.

The problem appears to be some sort of Java or JRE issue, but I can't
imagine how running /bin/bash from within emacs would trigger such an error.

O/P details here:
https://sourceforge.net/p/matlab-emacs/mailman/message/35402696/

I can provide more details as requested.  I don't know if this problem
extends to GNU/Linus or MS operating systems.

thanks,

Peter Mao

[-- Attachment #2: Type: text/html, Size: 1145 bytes --]

Re: MATLAB R2016b segmentation-faults when started from emacs

[Paul Eggert]

In reading the comments this appears to be a Matlab bug; more precisely, 
a bug in the JVM shipped with Matlab. I saw reports that the problem is 
fixed by using a free JVM rather than Matlab's bundled one, so perhaps 
you can suggest this to other Matlab users. It could be difficult to 
solve the underlying problem (whatever it is) without more cooperation 
from the Matlab developers.

Re: MATLAB R2016b segmentation-faults when started from emacs

[Peter Mao]

[-- Attachment #1: Type: text/plain, Size: 761 bytes --]

Hi Paul,

thanks for your input -- do you have a link to such reports?

Looks like Eric Ludlam in on the case at Mathworks, so there's some hope
for a solution.

Until we know more, the great mystery is why the shell would behave any
differently inside of emacs vs outside of it.

Peter

On Tue, Nov 15, 2016 at 10:02 AM, Paul Eggert <eggert@cs.ucla.edu> wrote:

> In reading the comments this appears to be a Matlab bug; more precisely, a
> bug in the JVM shipped with Matlab. I saw reports that the problem is fixed
> by using a free JVM rather than Matlab's bundled one, so perhaps you can
> suggest this to other Matlab users. It could be difficult to solve the
> underlying problem (whatever it is) without more cooperation from the
> Matlab developers.
>

[-- Attachment #2: Type: text/html, Size: 1136 bytes --]

Re: MATLAB R2016b segmentation-faults when started from emacs

[Paul Eggert]

On 11/15/2016 10:26 AM, Peter Mao wrote:
> thanks for your input -- do you have a link to such reports?

I was merely reading here:

https://sourceforge.net/p/matlab-emacs/mailman/matlab-emacs-discuss/

which talks about running /usr/lib/jvm/java-7-openjdk-amd64/jre/ to work 
around the problem. Not knowing Matlab I don't know how well that would 
work in practice. If Mathworks is on the case I assume they will let us 
know if it's actually an Emacs bug.