TST: Reproduce paintEvent failures by GuillaumeFavelier · Pull Request #7294 · mne-tools/mne-python

GuillaumeFavelier · 2020-02-06T12:48:58Z

The goal is to fix the Azure failures encountered recently during testing in: #7290, #7235

The log is very scarce, I'm trying locally with virtualbox at the same time:

mne\viz\_3d.py s                                                         [ 91%]
mne\viz\circle.py s                                                      [ 91%]
mne\viz\misc.py .                                                        [ 91%]
mne\viz\utils.py s                                                       [ 91%]
mne\viz\_brain\tests\test_brain.py ......ssssssss.                       [ 92%]
mne\viz\backends\tests\test_renderer.py .sx..ss                          [ 92%]
mne\viz\backends\tests\test_utils.py ..                                  [ 93%]
Windows fatal exception: access violation

Thread 0x00000ca8 (most recent call first):
  File "d:\a\1\s\conda\lib\threading.py", line 296 in wait
  File "d:\a\1\s\conda\lib\queue.py", line 170 in get
  File "D:\a\1\s\mne\gui\_coreg_gui.py", line 1363 in worker
  File "d:\a\1\s\conda\lib\threading.py", line 870 in run
  File "d:\a\1\s\condaWindows fatal exception: \access violationl

ib\threading.py", line 926 in _bootstrap_inner
##[error]Cmd.exe exited with code '-1073741819'.

codecov · 2020-02-06T13:43:34Z

Codecov Report

Merging #7294 into master will increase coverage by 2.46%.
The diff coverage is 91.17%.

@@            Coverage Diff             @@
##           master    #7294      +/-   ##
==========================================
+ Coverage   87.60%   90.07%   +2.46%     
==========================================
  Files         454      454              
  Lines       82558    82391     -167     
  Branches    13050    13023      -27     
==========================================
+ Hits        72329    74216    +1887     
+ Misses       7426     5352    -2074     
- Partials     2803     2823      +20

GuillaumeFavelier · 2020-02-11T13:53:51Z

Closing now since this seems to be solved by #7295

GuillaumeFavelier · 2020-02-19T23:58:52Z

The failures still occur.

Reference:
#7342 (comment)

larsoner · 2020-02-22T21:32:30Z

Here is a nice readable traceback

https://dev.azure.com/mne-tools/mne-python/_build/results?buildId=5624&view=logs&jobId=20ec0733-a031-5a42-425a-ca7f9f76ad75&j=20ec0733-a031-5a42-425a-ca7f9f76ad75&t=3732c0a7-07d3-577b-8168-001d94589ebf

GuillaumeFavelier · 2020-02-23T08:16:36Z

I was able to read it when you did the comment but I forgot to save the log, now the build is not available anymore. I remember it was related somehow to QVTKRenderWindowInteractor and PaintEvent.

larsoner · 2020-02-24T01:54:25Z

Just go through the recent Azure builds, it usually only takes looking at a few to find a failure with a good traceback, like:

https://dev.azure.com/mne-tools/mne-python/_build/results?buildId=5634&view=logs&j=b195a6e3-7484-5943-e627-0b35c27605d6&t=6dcb3581-08d4-5e76-74f2-6edaf20ff912&l=257

GuillaumeFavelier · 2020-02-25T10:39:51Z

For reference, here is a log:

Details

mne\utils\tests\test_testing.py ....                                     [ 91%]
mne\viz\_3d.py s                                                         [ 91%]
mne\viz\circle.py s                                                      [ 91%]
mne\viz\misc.py .                                                        [ 91%]
mne\viz\utils.py s                                                       [ 92%]
mne\viz\_brain\tests\test_brain.py ..............                        [ 92%]
mne\viz\backends\tests\test_renderer.py ..x....                          [ 93%]
mne\viz\backends\tests\test_utils.py ..                                  [ 93%]
Windows fatal exception: access violation

Thread 0x00001198 (most recent call first):
  File "d:\a\1\s\conda\lib\site-packages\tvtk\pyface\ui\qt4\QVTKRenderWindowInteractor.py", line 370 in paintEvent
  File "d:\a\1\s\conda\lib\site-packages\tvtk\pyface\ui\qt4\scene.py", line 69 in paintEvent
  File "d:\a\1\s\conda\lib\site-packages\pyface\ui\qt4\gui.py", line 85 in process_events
  File "d:\a\1\s\conda\lib\site-packages\surfer\viz.py", line 183 in _force_render
  File "d:\a\1\s\conda\lib\site-packages\surfer\viz.py", line 550 in _toggle_render
  File "d:\a\1\s\conda\lib\site-packages\surfer\viz.py", line 487 in __init__
  File "D:\a\1\s\mne\viz\_3d.py", line 2385 in plot_vector_source_estimates
  File "D:\a\1\s\mne\utils\_logging.py", line 89 in wrapper
  File "<D:\a\1\s\mne\externals\decorator.py:decorator-gen-131>", line 2 in plot_vector_source_estimates
  File "D:\a\1\s\mne\source_estimate.py", line 2197 in plot
  File "D:\a\1\s\mne\viz\tests\test_3d.py", line 651 in test_plot_vector_source_estimates
  File "D:\a\1\s\mne\utils\_testing.py", line 259 in dec
  File "d:\a\1\s\conda\lib\site-packages\_pytest\python.py", line 167 in pytest_pyfunc_call
  File "d:\a\1\s\conda\lib\site-packages\pluggy\callers.py", line 187 in _multicall
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 87 in <lambda>
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 93 in _hookexec
  File "d:\a\1\s\conda\lib\site-packages\pluggy\hooks.py", line 286 in __call__
  File "d:\a\1\s\conda\lib\site-packages\_pytest\python.py", line 1445 in runtest
  File "d:\a\1\s\conda\lib\site-packages\_pytest\runner.py", line 134 in pytest_runtest_call
  File "d:\a\1\s\conda\lib\site-packages\pluggy\callers.py", line 187 in _multicall
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 87 in <lambda>
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 93 in _hookexec
  File "d:\a\1\s\conda\lib\site-packages\pluggy\hooks.py", line 286 in __call__
  File "d:\a\1\s\conda\lib\site-packages\_pytest\runner.py", line 210 in <lambda>
  File "d:\a\1\s\conda\lib\site-packages\_pytest\runner.py", line 237 in from_call
  File "d:\a\1\s\conda\lib\site-packages\_pytest\runner.py", line 210 in call_runtest_hook
  File "d:\a\1\s\conda\lib\site-packages\_pytest\runner.py", line 185 in call_and_report
  File "d:\a\1\s\conda\lib\site-packages\_pytest\runner.py", line 99 in runtestprotocol
  File "d:\a\1\s\conda\lib\site-packages\_pytest\runner.py", line 84 in pytest_runtest_protocol
  File "d:\a\1\s\conda\lib\site-packages\pluggy\callers.py", line 187 in _multicall
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 87 in <lambda>
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 93 in _hookexec
  File "d:\a\1\s\conda\lib\site-packages\pluggy\hooks.py", line 286 in __call__
  File "d:\a\1\s\conda\lib\site-packages\_pytest\main.py", line 271 in pytest_runtestloop
  File "d:\a\1\s\conda\lib\site-packages\pluggy\callers.py", line 187 in _multicall
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 87 in <lambda>
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 93 in _hookexec
  File "d:\a\1\s\conda\lib\site-packages\pluggy\hooks.py", line 286 in __call__
  File "d:\a\1\s\conda\lib\site-packages\_pytest\main.py", line 247 in _main
  File "d:\a\1\s\conda\lib\site-packages\_pytest\main.py", line 197 in wrap_session
  File "d:\a\1\s\conda\lib\site-packages\_pytest\main.py", line 240 in pytest_cmdline_main
  File "d:\a\1\s\conda\lib\site-packages\pluggy\callers.py", line 187 in _multicall
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 87 in <lambda>
  File "d:\a\1\s\conda\lib\site-packages\pluggy\manager.py", line 93 in _hookexec
  File "d:\a\1\s\conda\lib\site-packages\pluggy\hooks.py", line 286 in __call__
  File "d:\a\1\s\conda\lib\site-packages\_pytest\config\__init__.py", line 93 in main
  File "d:\a\1\s\conda\Scripts\pytest-script.py", line 10 in <module>

larsoner · 2020-02-26T05:09:30Z

Travis even hit one:

https://travis-ci.org/mne-tools/mne-python/jobs/655188026#L2938-L2944

GuillaumeFavelier · 2020-03-12T09:25:12Z

The latest commit 499f917 on #7380 seems to reproduce this behaviour on Travis consistently. I restarted the failed job 3 times.

GuillaumeFavelier · 2020-03-12T14:49:46Z

Reproduced in here, it appears to be consistent.

GuillaumeFavelier · 2020-03-12T14:51:25Z

My first idea is to use the renderer fixture in test_mixed_sources_plot_surface

GuillaumeFavelier · 2020-03-12T16:45:25Z

Test 1) is green
Test 2) is green
Test 3) is green
Test 4) is green
Test 5) ~~failed because of #7437, will be restarted once this is fixed~~ is green
Test 6) is green
Test 7) is green
Test 8) is green
Test 9) is green
Test 10) is green

mne/viz/tests/test_3d.py

GuillaumeFavelier · 2020-03-13T14:52:54Z

I ported 532aba9 in #7380

This reverts commit 7f6120b.

GuillaumeFavelier · 2020-03-17T12:38:03Z

Travis jobs are up in here, waiting for the good news 🤞

agramfort · 2020-03-17T13:21:50Z

thx @GuillaumeFavelier !

agramfort · 2020-03-17T13:21:51Z

thx @GuillaumeFavelier !

larsoner · 2020-03-17T19:09:16Z

mne/viz/_brain/_brain.py

+            times, array, self.interp_kind, axis=1,
+            assume_sorted=True)(time_idx)
+        ifunc = interp1d(times, self._data['time'])
+        act_time = ifunc(time_idx)


This PR undid (again) some of the refactoring. The way my commits worked:

the interpolators were set up once and reused (efficiency)

a helper _safe_interp1d (avoids bugs now back in master)

And this commit appears to have un-done these. Accordingly we have lost the efficiency and the singleton regression is back in master:

https://circleci.com/gh/mne-tools/mne-python/18710?utm_campaign=workflow-failed&utm_medium=email&utm_source=notification

I think the real (only) bug that probably needed to be fixed here had to do with calling the interpolation functions incorrectly, right? If so @GuillaumeFavelier can you make a revert commit for this PR and try making the minimal changes required to get CIs happy again? I'd rather not sacrifice the _safe_interp1d and funcs efficiency gains...

Ahh it's even more extensive -- I see this started out by completely reverting #7257. Are we planning on putting that back in somehow?

I updated #7162 and delayed thoses features to 0.21 for now

GuillaumeFavelier self-assigned this Feb 6, 2020

GuillaumeFavelier closed this Feb 11, 2020

GuillaumeFavelier reopened this Feb 19, 2020

This was referenced Feb 26, 2020

MRG, FIX: Remove force render nipy/PySurfer#289

Merged

MAINT: Better test fixtures #7358

Merged

GuillaumeFavelier changed the title ~~TST: Reproduce failures in Azure~~ TST: Reproduce paintEvent failures Mar 12, 2020

GuillaumeFavelier mentioned this pull request Mar 12, 2020

pytest-sugar is not working with pytest v5.4.0 (released on March 12th 2020) #7437

Closed

larsoner reviewed Mar 12, 2020

View reviewed changes

mne/viz/tests/test_3d.py Outdated Show resolved Hide resolved

GuillaumeFavelier closed this Mar 13, 2020

GuillaumeFavelier mentioned this pull request Mar 16, 2020

MRG: Use the renderer fixture in test_mixed_sources_plot_surface #7454

Merged

GuillaumeFavelier reopened this Mar 17, 2020

Revert "MRG: Add toggle to save_movie in _TimeViewer (mne-tools#7257)"

fe70356

This reverts commit 7f6120b.

GuillaumeFavelier force-pushed the test_azure branch from e26d41d to fe70356 Compare March 17, 2020 10:00

Revert _interpolate_data signature

17dd265

agramfort marked this pull request as ready for review March 17, 2020 13:21

agramfort merged commit 95ad0ee into mne-tools:master Mar 17, 2020

larsoner reviewed Mar 17, 2020

View reviewed changes

GuillaumeFavelier deleted the test_azure branch June 11, 2020 09:49

Uh oh!

Conversation

GuillaumeFavelier commented Feb 6, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov bot commented Feb 6, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

GuillaumeFavelier commented Feb 11, 2020

Uh oh!

GuillaumeFavelier commented Feb 19, 2020

Uh oh!

larsoner commented Feb 22, 2020

Uh oh!

GuillaumeFavelier commented Feb 23, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

larsoner commented Feb 24, 2020

Uh oh!

GuillaumeFavelier commented Feb 25, 2020

Uh oh!

larsoner commented Feb 26, 2020

Uh oh!

GuillaumeFavelier commented Mar 12, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

GuillaumeFavelier commented Mar 12, 2020

Uh oh!

GuillaumeFavelier commented Mar 12, 2020

Uh oh!

GuillaumeFavelier commented Mar 12, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

GuillaumeFavelier commented Mar 13, 2020

Uh oh!

GuillaumeFavelier commented Mar 17, 2020

Uh oh!

agramfort commented Mar 17, 2020

Uh oh!

agramfort commented Mar 17, 2020

Uh oh!

larsoner Mar 17, 2020

Choose a reason for hiding this comment

Uh oh!

larsoner Mar 17, 2020

Choose a reason for hiding this comment

Uh oh!

GuillaumeFavelier Mar 18, 2020

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

GuillaumeFavelier commented Feb 6, 2020 •

edited

Loading

codecov bot commented Feb 6, 2020 •

edited

Loading

GuillaumeFavelier commented Feb 23, 2020 •

edited

Loading

GuillaumeFavelier commented Mar 12, 2020 •

edited

Loading

GuillaumeFavelier commented Mar 12, 2020 •

edited

Loading