matplotlib-users Mailing List for matplotlib

2013/8/31 Dino Bektešević <lj...@gm...>:
> Hello,
>
> After a little mishap from ubuntu 12.04 after which I reinstalled the
> OS, on this fresh install I did:
>
>>sudo apt-get install python-numpy python-scipy python-matplotlib ipython ipython-notebook python-pandas python-sympy python-nose
>
> as per scipy stack installation instructions and
> everything went more or less as it should have no errors reported
> during installation that I saw. Keep in mind the entire install like
> this had ~500MB or so and I wasn't always paying attention.
> I ran python, and did numpy.test(), returned:
>>Ran 3161 tests in 50.667s
>>OK (KNOWNFAIL=3, SKIP=4)
>><nose.result.TextTestResult run=3161 errors=0 failures=0>
>
> did scipy.test(), returned:
>>Ran 3780 tests in 74.809s
>>FAILED (KNOWNFAIL=11, SKIP=13, failures=2)
>><nose.result.TextTestResult run=3780 errors=0 failures=2>
>
> I send a mail to scipy mailing list couple of days ago, but still no answer,
> if someone knows how "bad" those 2 failures are please share and then
> did matplotlib.test() which was disasterous:
>>Ran 1065 tests in 284.956s
>>FAILED (KNOWNFAIL=267, errors=772)

With mpl 1.3.0 (packaged for Raring by Thomas Kluyver):

Ran 1465 tests in 402.499s
FAILED (KNOWNFAIL=1, SKIP=5, errors=1331)

But matplotlib itself is working pretty well. The output is full with
error messages like:

IOError: Baseline image
'/home/goyo/result_images/test_triangulation/tripcolor1-expected.svg'
does not exist.

It maybe that distro packages do not ship with baseline images. Looks
sensible to me since there must be an awful lot of them and most users
do not need them.

Goyo

Hello,

After a little mishap from ubuntu 12.04 after which I reinstalled the
OS, on this fresh install I did:

>sudo apt-get install python-numpy python-scipy python-matplotlib ipython ipython-notebook python-pandas python-sympy python-nose

as per scipy stack installation instructions and
everything went more or less as it should have no errors reported
during installation that I saw. Keep in mind the entire install like
this had ~500MB or so and I wasn't always paying attention.
I ran python, and did numpy.test(), returned:
>Ran 3161 tests in 50.667s
>OK (KNOWNFAIL=3, SKIP=4)
><nose.result.TextTestResult run=3161 errors=0 failures=0>

did scipy.test(), returned:
>Ran 3780 tests in 74.809s
>FAILED (KNOWNFAIL=11, SKIP=13, failures=2)
><nose.result.TextTestResult run=3780 errors=0 failures=2>

I send a mail to scipy mailing list couple of days ago, but still no answer,
if someone knows how "bad" those 2 failures are please share and then
did matplotlib.test() which was disasterous:
>Ran 1065 tests in 284.956s
>FAILED (KNOWNFAIL=267, errors=772)

I've also pasted the entire thing online and it should be availible
for 2 weeks or so:
>http://pastebin.com/rAvnnuCc
>http://pastebin.com/4aPTKJ3V

and for the scipy if anyone is willing:
>http://pastebin.com/FutWAR56

Any help is greatly appreciated,
Dino Bektešević.

I just installed matplotlib-1.3.0.win32-py2.7.exe from
http://matplotlib.org/downloads.html
and I now get the old error:
   ImportError: matplotlib requires dateutil

Is this a policy change?  The installation instructions at
   http://matplotlib.org/users/installing.html
still say "Windows users only need the first two (python and numpy)
since the others are built into the matplotlib Windows installers
available for download at the download page.  Of course, that points to
    https://github.com/matplotlib/matplotlib/downloads
where 1.3 is not yet available ...

Thanks,
Alan Isaac