[MRG] Kit2fiff GUI: Test event settings

[christianbrodbeck]

Some enhancements to the KIT2FIFF GUI:

• button to preview the events: this is something I've wished for many times, makes it easier to make sure one has selected the right event settings
• button to plot raw (just uses raw.plot()). On my setup this works nicely with the traits gui.

[codecov-io]

Current coverage is 87.27% (diff: 60.83%)

Merging #3463 into master will decrease coverage by 0.02%

@@             master      #3463   diff @@
==========================================
  Files           341        341          
  Lines         60372      60452    +80   
  Methods           0          0          
  Messages          0          0          
  Branches       9229       9246    +17   
==========================================
+ Hits          52705      52760    +55   
- Misses         4915       4934    +19   
- Partials       2752       2758     +6   

Powered by Codecov. Last update 97861e9...12623b7

[christianbrodbeck]

Two errors that seem opaque to me since they don't seem to be relate to the changes. Do these look familiar to anyone?

  File "/home/ubuntu/miniconda/envs/circleenv/lib/python2.7/site-packages/pyface/ui/qt4/code_editor/pygments_highlighter.py", line 39, in get_tokens_unprocessed
    for item in action(self, m):
TypeError: 'NoneType' object is not callable

and

X Error of failed request:  BadDrawable (invalid Pixmap or Window parameter)
  Major opcode of failed request:  72 (X_PutImage)
  Resource id in failed request:  0x200013
  Serial number of failed request:  45
  Current serial number in output stream:  48

[larsoner]

Yes the first is a recurring problem (not ours, pyface) and so is the second (not ours, mayavi)

[christianbrodbeck]

In that case this is ready. The changes in non-GUI code are minimal, restricted to factoring out some functionality now required in the GUI.

[larsoner]

LGTM +1 for merge but you need a rebase

[christianbrodbeck]

Rebased. The Travis error is

X Error of failed request:  BadDrawable (invalid Pixmap or Window parameter)

[larsoner]

I get that error locally, too, sometimes. It's an unrelated mayavi problem (cc @agramfort this is what always prevents me from building doc)

[larsoner]

Any KIT users out there who could test this? @teonbrooks is that just you?

[teonbrooks]

@kingjr, free to test it?

[kingjr]

Not really

Le 23 août 2016 22:55, "Teon L Brooks" notifications@github.com a écrit :

@kingjr https://github.com/kingjr, free to test it?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#3463 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEp7DDXaOPb8wISbHbeVI072gBxggxcKks5qi15cgaJpZM4JWQ0I
.

[christianbrodbeck]

I've tested it with multiple people's data by now if that helps...

[larsoner]

Looks like Travis bombed with the VTK segfault of death. I'll restart and see if it works at least intermittently.

[larsoner]

It bombed again, I restarted it again. But really if it fails this often we need to fix it -- can't have the CIs fail that often. Did you add something in this PR that would cause it to render differently? VTK on Travis (and Ubuntu in general) is buggy, so we have to find a way to make it work.

[christianbrodbeck]

I don’t think there is anything in that PR that should cause it to fail more… What is the failure rate on other PRs?

On Sep 8, 2016, at 3:47 PM, Eric Larson notifications@github.com wrote:

It bombed again, I restarted it again. But really if it fails this often we need to fix it -- can't have the CIs fail that often. Did you add something in this PR that would cause it to render differently? VTK on Travis (and Ubuntu in general) is buggy, so we have to find a way to make it work.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub #3463 (comment), or mute the thread https://github.com/notifications/unsubscribe-auth/AAI5a4Tkdz4mF7h5qShRJ5M1qg9b3JHHks5qoGZQgaJpZM4JWQ0I.

[larsoner]

Near zero failures

[christianbrodbeck]

Is there a way to run it on the intermediate commits to see whether it starts failing at a specific point?

On Sep 8, 2016, at 6:15 PM, Eric Larson notifications@github.com wrote:

Near zero failures
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub #3463 (comment), or mute the thread https://github.com/notifications/unsubscribe-auth/AAI5a28h9HuWq8fIf_slVnDCl4P4MUlEks5qoIj8gaJpZM4JWQ0I.

[larsoner]

You can re-push them one by one (like via force-push for the first one) or enable Travis-CI on your own repository and push to some other branch there

[agramfort]

travis and appveyor are happy

+1 for merge

[christianbrodbeck]

Sorry, don't merge, I'm in the process of repushing individual commits to see whether the VTK error starts happening consistently at a specific point

[christianbrodbeck]

@Eric89GXL I've identified the problem, the Travis error occurred whenever the test of Kit2FiffModel opened a Progress Dialog. I've added a flag to the model to only show GUI elements when run from the actual GUI, not when run on its own.

[larsoner]

Great, thanks for taking the time to figure that out!