New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

One viz test still failing on mac os #758

Closed
arokem opened this Issue Nov 1, 2015 · 13 comments

Comments

Projects
None yet
2 participants
@arokem
Member

arokem commented Nov 1, 2015

At least on my laptop. Let's see what the buildbots have to say about that, though.

I will also try to debug this locally.

======================================================================
FAIL: dipy.viz.tests.test_fvtk_widgets.test_text_widget
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/Users/arokem/anaconda/envs/py2/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/Users/arokem/anaconda/envs/py2/lib/python2.7/site-packages/nose/util.py", line 620, in newfunc
    return func(*arg, **kw)
  File "/Users/arokem/source/dipy/dipy/testing/decorators.py", line 63, in test_with_xvfb
    my_test()
  File "/Users/arokem/anaconda/envs/py2/lib/python2.7/site-packages/numpy/testing/decorators.py", line 146, in skipper_func
    return f(*args, **kwargs)
  File "/Users/arokem/anaconda/envs/py2/lib/python2.7/site-packages/numpy/testing/decorators.py", line 146, in skipper_func
    return f(*args, **kwargs)
  File "/Users/arokem/source/dipy/dipy/viz/tests/test_fvtk_widgets.py", line 180, in test_text_widget
    npt.assert_equal(report.objects, 3)
  File "/Users/arokem/anaconda/envs/py2/lib/python2.7/site-packages/numpy/testing/utils.py", line 354, in assert_equal
    raise AssertionError(msg)
AssertionError: 
Items are not equal:
 ACTUAL: 492
 DESIRED: 3
@arokem

This comment has been minimized.

Member

arokem commented Nov 1, 2015

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

This is the one we know about correct? Some driver issue on Macosx with the screenshot.

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

Sorry I mean snapshot.

@arokem

This comment has been minimized.

Member

arokem commented Nov 1, 2015

Yes - that's the one. Just want to make sure we don't lost track.

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

These buildbots may not have VTK I believe. Those tests were skipped before. And they should stay this way. I think we need to be slowly progressing on enabling visual tests in all platforms. Snapshot was designed for start for Linux.

@arokem

This comment has been minimized.

Member

arokem commented Nov 1, 2015

Yeah - but users are going to run these tests. It's great to have tests that run on one platform, but it's truly invaluable to have tests that run robustly on all platforms.

@arokem

This comment has been minimized.

Member

arokem commented Nov 1, 2015

Agree that we need to figure this out locally, before we start running this on the bots, though.

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

Yes, but we don't know if that is feasible with visual tests in all platforms.

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

The idea with snapshot was that it would be completely stealthmode.

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

But what we see is that offscreen renderering is not supported by all OS with VTK

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

And there are also differences even in Linux with VTK 5 and 6. Differences that are difficult to monitor. The same with ordered transparency.

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Nov 1, 2015

What we could do is for the tests that fail in the Mac using record instead and analyze the png (rather than using snapshot). Record is a bit more robust and the problem we had before where sometimes the recorded png was completely black is now resolved using render.reset_clipping_plane. But now that I am thinking more about this. There is a new parameter with snapshot that disables offscreen rendering.
Can you try it? It may resolve your issue. Wild guess here :)

@arokem

This comment has been minimized.

Member

arokem commented Jan 27, 2016

This one's resolved. Closing.

@arokem arokem closed this Jan 27, 2016

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment