matplotlib-users Mailing List for matplotlib (Page 2)

To close this thread: the PyEval_RestoreThread crash is a known bug in
the tkagg backend on Windows. See line 375 of backend_tkagg.py:

    def show(self):
        """
        this function doesn't segfault but causes the
        PyEval_RestoreThread: NULL state bug on win32
        """

-- Christoph

On 2/9/2010 10:26 AM, Christoph Gohlke wrote:
> Change the last line to pylab.show() and it should work.
> 
> Anyway, this example should not crash the interpreter. I can reproduce
> the crash on Python 2.4, 2.5, and 2.6 (32 and 64 bit) on Windows with
> mpl 0.99.1 but not on Ubuntu 9.1 with mpl 0.99.0.
> 
> The shortest example that crashes is:
> 
> python -c "import pylab;pylab.subplot(111).figure.show()"
> 
> or on the interactive prompt:
> 
>>>> import pylab
>>>> pylab.subplot(111).figure.show()
>>>> exit()
> Fatal Python error: PyEval_RestoreThread: NULL tstate
> 
> This application has requested the Runtime to terminate it in an unusual
> way.
> Please contact the application's support team for more information.
> 
> 
> -- Christoph
> 
> On 2/9/2010 7:57 AM, Lee Boger wrote:
>> > 
>> > Windows XP Professional with Python 2.5 installed (pywin32 build 210) -
>> > came with dSPACE software package
>> > 
>> > Downloaded and installed matplotlib-0.99.1.win32-py2.5.exe from
>> > sourceforge.net
>> > 
>> > Downloaded and installed numpy-1.4.0-win32-superpack-python2.5.exe from
>> > sourceforge.net
>> > 
>> > Executing the following simple "log plot" script within PythonWin:
>> > 
>> > *from* matplotlib *import* pylab
>> > 
>> > # Create some artificial data.
>> > test_frequency = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
>> > 17, 18, 19, 20]
>> > test_results = [-0.2, -0.7, -1.0, -1.5, -2.0, -2.5, -3.0, -3.5, -4, -5,
>> > -6, -7.1, -8, -9, -10, -11, -12, -15, -20, -25]
>> > spec_frequency = [6, 8]
>> > spec_results = [-3.0, -3.0]
>> > 
>> > # Plot
>> > figure = pylab.subplot(111)
>> > figure.semilogx()
>> > figure.scatter(test_frequency, test_results, s=20, c='b', marker='s',
>> > edgecolors='none')
>> > figure.scatter(spec_frequency, spec_results, s=40, c='g', marker='s',
>> > edgecolors='none')
>> > figure.grid(True)
>> > figure.set_xlabel(r"Frequency (Hz)", fontsize = 12)
>> > figure.set_ylabel(r"Actuator Response (db)", fontsize = 12)
>> > 
>> > figure.figure.savefig('log_plot')
>> > figure.figure.show()
>> > 
>> > 
>> > Plots a figure on the screen that looks correct, then the following
>> > error (when I click OK, PythonWin closes)
>> > 
>> > 
>> > 
>> > Any advice would be appreciated. Maybe it's an installation or setup
>> > issue, but I'm pretty knew to Python programming and don't know how to
>> > debug this.
>> > 
>> > Lee Boger

<sie...@sb...> wrote:
> Subject is the question.
>
> As I see it, it's useful to know MatLab. A simple query with matplotlib
> tutorial shows a number of hits. The first, reference to v0.99.a
> documentation barely qualifies. Examples galore and a pretty minimal
> introduction. In the first 10 or so hits ther's a blog and mention of a
> video. The blog may appeal to some, but it seems unelementary. The video
> basically asks to sign in. Who knows where that goes? I've seen a few
> videos for MPL, but they all look tied into $$.
>
> I've made some reasonable progress on MPL, but am still far short of
> being confident of using it.  Too much try this and see.
>
> I know of exactly one book on MPL ( for scientists. sounds interesting).
> It was published recently by a foreign author. It is not yet widely
> distributed.
>
> Your turn. Comments?

Have you read the manual itself?  I found that quite helpful in getting started:

http://matplotlib.sourceforge.net/users/index.html

I've been satisfied with learning a bit here and there as I go.  It's
a big project, and I don't expect to understand it all at once.  But,
that said, even with a partial understanding I'm pleased at what it
can do for my needs.  What I find is first you should determine what
is the desired outcome you want, then Google around for mentions of
this in old threads or on the matplotlib site, then maybe try asking
for assistance on the list.

Che

Subject is the question.

As I see it, it's useful to know MatLab. A simple query with matplotlib 
tutorial shows a number of hits. The first, reference to v0.99.a 
documentation barely qualifies. Examples galore and a pretty minimal 
introduction. In the first 10 or so hits ther's a blog and mention of a 
video. The blog may appeal to some, but it seems unelementary. The video 
basically asks to sign in. Who knows where that goes? I've seen a few 
videos for MPL, but they all look tied into $$.

I've made some reasonable progress on MPL, but am still far short of 
being confident of using it.  Too much try this and see.

I know of exactly one book on MPL ( for scientists. sounds interesting). 
It was published recently by a foreign author. It is not yet widely 
distributed.

Your turn. Comments?
-- 
"Crime is way down. War is declining. And that's far from the good 
news." -- Steven Pinker (and other sources) Why is this true, but yet 
the media says otherwise? The media knows very well how to manipulate us 
(see limbic, emotion, $$). -- WTW

Hi List,

I made a script to draw very simple (single-direction single-input 
single-sided single-everything) Sankey diagrams (attached). I think I 
could share, if it can be of any use...

Cheers,

Yannick