8000 Improving documentation: Tests failing on a osx setup · Issue #7234 · matplotlib/matplotlib · GitHub
[go: up one dir, main page]

Skip to content
Improving documentation: Tests failing on a osx setup #7234
Closed
@NelleV

Description

@NelleV

We've ran into some problems getting the matplotlib environment working fine on linux, and I am not sure this is documented anywhere.

We had a massive number of tests failing, as the tests where not running on the correct freetype issue. The fix was easy:

  1. edit the setup.cfg file and set the test to True, and local_freetype to true.
  2. Delete the old build - matplotlib caches the results and will not rebuild with the new options despite they changed.
  3. Rebuild.

Our documentation need to be updated to reflect this issue.

Original issue.

On system python2.7.8, on latest master version, with both the macosx and agg backend, @trpham is having a lot of tests failing on images comparisons:

======================================================================
FAIL: mpl_toolkits.tests.test_axes_grid1.test_inset_locator(0, u'inset_locator', u'png')
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/Library/Python/2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/Users/TruongPham/BIDS/matplotlib/lib/matplotlib/testing/nose/decorators.py", line 60, in failer
    result = f(*args, **kwargs)
  File "/Users/TruongPham/BIDS/matplotlib/lib/matplotlib/testing/decorators.py", line 297, in compare
    raise_on_image_difference(expected_fname, actu
634B
al_fname, self.tol)
  File "/Users/TruongPham/BIDS/matplotlib/lib/matplotlib/testing/decorators.py", line 216, in raise_on_image_difference
    '(RMS %(rms).3f)' % err)
ImageComparisonFailure: images not close: /Users/TruongPham/BIDS/matplotlib/result_images/test_axes_grid1/inset_locator.png vs. /Users/TruongPham/BIDS/matplotlib/result_images/test_axes_grid1/inset_locator-expected.png (RMS 0.001)

Around 900 tests are failing, but the RMS is always extremely small. Anyone guesses what is going on?

Apart from this, matplotlib seems to be running fine. Any help would be appreciated.

Metadata

Metadata

Assignees

Labels

Difficulty: Easyhttps://matplotlib.org/devdocs/devel/contribute.html#good-first-issuesDocumentation

Type

No type

Projects

No projects

Milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions

    0