messenger / ZMQ session error upon "Matlab clear all" using 'magic'

[bs448c]

As outlined in #56 and #69 I manage to build pymatbridge-0.3-dev under WIndows 8.1 64 bit. I can start the matlabserver and pass back and forth variables between MATLAB and my iPython notebook (Anaconda 1.9.1, Python 2.7.6) using 'magic' (the x=x+1 and a=[1,2,3]

%%matlab -i a -o b,c
b = a + 3
c = b + 3

sort of examples)

However when trying to port a more complex example,

%%matlab -o x,xf

%Author: Osama Ullah Khan,
%        Phd Student, University of Michigan-Ann Arbor.
%        Email: oukhan@umich.edu
%        Version: 1.0
%
%This code demonstrate compressive sensing example. In this
%example the signal is sparse in frequency domain and random samples
%are taken in time domain.

close all;
clear all;

%setup path for the subdirectories of l1magic
path(path, 'C:\MATLAB\R2013a\l1magic\Optimization');
path(path, 'C:\MATLAB\R2013a\l1magic\Data');


%length of the signal
N=1024;

%Number of random observations to take
K=256;

%Discrete frequency of two sinusoids in the input signal
k1=29;
k2=59;
k3=109;

n=0:N-1;

%Sparse signal in frequency domain.
x=sin(2*pi*(k1/N)*n)+sin(2*pi*(k2/N)*n)+sin(2*pi*(k3/N)*n);

xf=abs(fft(x));

my matlabserver fails:

MATLAB is running in headless mode.  Figure windows will not be displayed.

To get started, type one of these: helpwin, helpdesk, or demo.
For product information, visit www.mathworks.com.

Socket created at: tcp://127.0.0.1:55555
Error using messenger
Error: ZMQ session not initialized

Error in matlabserver (line 30)
            messenger('respond', resp);

» 

I'd like to point out this exact snippet has been working flawlessly using pymatbridge-0.2 and the 'old' communication protocol.

I don't know if this is an issue of Windows or my build -- can pelase someone on a different OS try to replicate this behaviour?
Thanks!

EDIT: This seems to be due to the clear all; which seems to kill the matlabserver. I modified the issue title accordingly

[arokem]

I'm sorry that this has been so difficult to get up and running. I can
replicate this error, and I will start looking into it. It's not obvious to
me what exactly is going on, but I do have a 5-hour flight just ahead of
me, so I will take a look.

On Thu, May 15, 2014 at 3:15 AM, bs448c notifications@github.com wrote:

As outlined in #56https://github.com/arokem/python-matlab-bridge/pull/56and
#69 #69 I manage
to build pymatbridge-0.3-dev under WIndows 8.1. I can start the
matlabserver and pass back and forth variables between MATLAB and my
iPython notebook (Anaconda 1.9.1, Python 2.7.6) using 'magic' (the x=x+1and
a=[1,2,3]

%%matlab -i a -o b,c
b = a + 3
c = b + 3

sort of examples)

However when trying to port a more complex example,

%%matlab -o x,xf

%Author: Osama Ullah Khan,
% Phd Student, University of Michigan-Ann Arbor.
% Email: oukhan@umich.edu
% Version: 1.0
%
%This code demonstrate compressive sensing example. In this
%example the signal is sparse in frequency domain and random samples
%are taken in time domain.

close all;
clear all;

%setup path for the subdirectories of l1magic
path(path, 'C:\MATLAB\R2013a\l1magic\Optimization');
path(path, 'C:\MATLAB\R2013a\l1magic\Data');

%length of the signal
N=1024;

%Number of random observations to take
K=256;

%Discrete frequency of two sinusoids in the input signal
k1=29;
k2=59;
k3=109;

n=0:N-1;

%Sparse signal in frequency domain.
x=sin(2_pi_(k1/N)n)+sin(2_pi(k2/N)n)+sin(2_pi(k3/N)*n);

xf=abs(fft(x));

my matlabserver fails:

MATLAB is running in headless mode. Figure windows will not be displayed.

To get started, type one of these: helpwin, helpdesk, or demo.
For product information, visit www.mathworks.com.

Socket created at: tcp://127.0.0.1:55555
Error using messenger
Error: ZMQ session not initialized

Error in matlabserver (line 30)
messenger('respond', resp);

»

I'd like to point out this exact snippet has been working flawlessly using
pymatbridge-0.2 and the 'old' communication protocol.

I don't know if this is an issue of Windows or my build -- can pelase
someone on a different OS try to replicate this behaviour?
Thanks!

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/70
.

[bs448c]

glad that it's probably not due to my messenger build, then .. and thanks for looking into it. Again I don't think a sensible use case the clear all; is needed but still better to see what's going on

[arokem]

It's definitely that clear all that is causing the trouble. There's
probably a way to circumvent that, but it would require some kind of hack.
For now, I propose removing the call to clear all.

On Thu, May 15, 2014 at 8:58 AM, bs448c notifications@github.com wrote:

glad that it's probably not due to my messenger build, then .. and
thanks for looking into it. Again I don't think a sensible use case the clear
all; is needed but still better to see what's going on

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/70#issuecomment-43228635
.

[blink1073]

I think what's happening is clear all is removing variables used by matlabserver

[blink1073]

Specifically, it is clearing the initialized flag used here based on the error message.