matplotlib-users Mailing List for matplotlib

I have intermittent problems with the savefig('foo.eps') command when I am
using my native LaTeX interpreter to produce eps files.  The really
frustrating thing is that it is intermittent.  Sometimes I can save a file
that looks fine and sometimes I can't.

My box is running Fedora Core 4 and kernel 2.6.14.

I have the following rpms:
python-2.4.1-2
ipython-0.6.15-1.fc4
python-tools-2.4.1-2
python-matplotlib-0.84-1.fc4
python-numeric-23.7-2

Fedora has an out of date ghostscript 7.07, so I installed
ghostscript-8.51from source into /usr/local, but not the fonts, since
they seemed
unnecessary.

As long as I don't use

rc('text', usetex=3DTrue)

I'm ok, but whenever I turn it on, it doesn't work consistently.

Here's some sample code to demonstrate the problem.  If I comment out the r=
c
line, it works fine.

#!/usr/bin/env python

from matplotlib import rc
from matplotlib.numerix import arange, cos, pi
from pylab import *

forces =3D [1, 2, 3, 4]
pofn =3D [.2, .09, .01, .005]

rc('text', usetex=3DTrue)
figure(1)
semilogy(forces,pofn,'k')
xlabel(r'$f_n =3D F_n/\langle F_n \rangle$')
ylabel(r'$P(f_n)$')
axis([0,5,.001,1])

savefig('mtest1.eps')

However, if I run it as is, then I get the following:

[jwlandry@merrimac hopper]$ ./mat-test.py --verbose-helpful
matplotlib data path /usr/share/matplotlib
$HOME=3D/home/jwlandry
CONFIGDIR=3D/home/jwlandry/.matplotlib
loaded rc file /home/jwlandry/.matplotlib/matplotlibrc
matplotlib version 0.84
verbose.level helpful
interactive is False
platform is linux2
numerix Numeric 23.7
font search path ['/usr/share/matplotlib']
loaded ttfcache file /home/jwlandry/.matplotlib/ttffont.cache
backend GTK version 2.6.2

This is dvips(k) 5.95a Copyright 2005 Radical Eye Software (
www.radicaleye.com)
' TeX output 2005.11.25:1645' -> e0dab64d0b15f2821ec713a6e7903ff4.ps
<tex.pro><psfrag.pro><texps.pro><special.pro><color.pro>. <cmr8.pfb>
<cmsy8.pfb><cmsy10.pfb><cmmi8.pfb><cmmi10.pfb><cmr10.pfb>[1
<e0dab64d0b15f2821ec713a6e7903ff4.eps>]

When it doesn't work, a dialog box opens with this message

Save figure failure:
None: Interrupted system call


Help!

James Landry


--
James W. Landry
jwl...@gm...

Wasn't that simple... sigh. I'll keep looking.

On Fri, 25 Nov 2005 10:17:57 +0100
"phi...@ho..." <phi...@ho...> wrote:

> Hi Alan,
> 
> You'll see Python exceptions are obvious enough.
> 
> Here it looks like windows doesn't support the ".1".
> Try 0.1 instead.
> 
> Hope it will solve the problem. ;-)
> Regards,
> Philippe Collet
> 
> >Message: 6
> >Date: Thu, 24 Nov 2005 22:16:48 -0600
> >From: Alan Jackson <aja...@op...>
> >To: mat...@li...
> >Subject: [Matplotlib-users] Incompatibility between Linux and XP
> >
> >I'm a new python user, basically trying to learn python using matplotlib
> >for a little project.
> >
> >My code works fine on Linux, but it fails on Windows XP. I'm running
> >the latest version of Matplotlib and SciPy python, 2.3.3.
> >
> >Before I go to a lot of effort to build a small demo program to show the
> >failure, I thought I would ask to see of there are known incompatibilities
> >between Linux and XP that might be causing my problem.
> >
> >The line of code that seems to precipitate the failure is :
> >
> >self.Cax_c = figtext(0.84, .1, '000')
> >
> >that gives an error on Windows XP of :
> >
> >.....matplotlib\text.py", line 665 in set_text
> >TypeError: This doesn't look like a string: '0.0'
> >
> >Thanks!
> >
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users


-- 
-----------------------------------------------------------------------
| Alan K. Jackson            | To see a World in a Grain of Sand      |
| al...@aj...          | And a Heaven in a Wild Flower,         |
| www.ajackson.org           | Hold Infinity in the palm of your hand |
| Houston, Texas             | And Eternity in an hour. - Blake       |
-----------------------------------------------------------------------

Hi,

Here is a testcase :

a=range(10)
b=rand(10)
plot(a,b,'x')
axis([0.,1e-7,0.,1.])

Segmentation fault

This test have been done with last matplotlib version (0.85) and TkAgg
backend + python2.3.
I haven't had time to look at the code for now but it sounds like a
"divide by zero + malloc" in the code computing the symbol possition on
the image.

Ok, this is a wired corner case but I have a real testcase (too long to
be posted here) where this crash occurs.

I also have tested without symbols :
 
a=range(10)
b=rand(10)
plot(a,b)
axis([0.,1e-7,0.,1.])


and the result looks good and no crash occurs.

Thanks for your job  on matplotlib :)
Xavier.

 

Hi Alan,

You'll see Python exceptions are obvious enough.

Here it looks like windows doesn't support the ".1".
Try 0.1 instead.

Hope it will solve the problem. ;-)
Regards,
Philippe Collet

>Message: 6
>Date: Thu, 24 Nov 2005 22:16:48 -0600
>From: Alan Jackson <aja...@op...>
>To: mat...@li...
>Subject: [Matplotlib-users] Incompatibility between Linux and XP
>
>I'm a new python user, basically trying to learn python using matplotlib
>for a little project.
>
>My code works fine on Linux, but it fails on Windows XP. I'm running
>the latest version of Matplotlib and SciPy python, 2.3.3.
>
>Before I go to a lot of effort to build a small demo program to show the
>failure, I thought I would ask to see of there are known incompatibilities
>between Linux and XP that might be causing my problem.
>
>The line of code that seems to precipitate the failure is :
>
>self.Cax_c = figtext(0.84, .1, '000')
>
>that gives an error on Windows XP of :
>
>.....matplotlib\text.py", line 665 in set_text
>TypeError: This doesn't look like a string: '0.0'
>
>Thanks!
>

I'm a new python user, basically trying to learn python using matplotlib
for a little project.

My code works fine on Linux, but it fails on Windows XP. I'm running
the latest version of Matplotlib and SciPy python, 2.3.3.

Before I go to a lot of effort to build a small demo program to show the
failure, I thought I would ask to see of there are known incompatibilities
between Linux and XP that might be causing my problem.

The line of code that seems to precipitate the failure is :

self.Cax_c = figtext(0.84, .1, '000')

that gives an error on Windows XP of :

.....matplotlib\text.py", line 665 in set_text
TypeError: This doesn't look like a string: '0.0'

Thanks!

-- 
-----------------------------------------------------------------------
| Alan K. Jackson            | To see a World in a Grain of Sand      |
| al...@aj...          | And a Heaven in a Wild Flower,         |
| www.ajackson.org           | Hold Infinity in the palm of your hand |
| Houston, Texas             | And Eternity in an hour. - Blake       |
-----------------------------------------------------------------------