Fix subtle memory profiler bug

[berleant]

This test fails sometimes: https://github.com/nipy/nipype/blob/master/nipype/interfaces/tests/test_runtime_profiler.py#L422

I'm 80% sure the problem is here, but I'd like to get confirmation from the original author @dclark87 :
https://github.com/nipy/nipype/blob/master/nipype/interfaces/base.py#L1368

When there are grandchildren processes and pyfunc=True, the above code will double-count the memory from the middle process.

I propose skipping the test (pr: #1664) for now until this issue is fixed--I can't get to it right now because I have other priorities, and I've already spent more time on this than I want to admit!

[chrisgorgo]

I'm afraid the original author switched his focus to other projects. I
would not hold back on waiting for his feedback.

On Mon, Sep 26, 2016 at 12:13 PM, Shoshana Berleant <
notifications@github.com> wrote:

This test fails sometimes: https://github.com/nipy/
nipype/blob/master/nipype/interfaces/tests/test_runtime_profiler.py#L422

I'm 80% sure the problem is here, but I'd like to get confirmation from
the original author ():
https://github.com/nipy/nipype/blob/master/nipype/interfaces/base.py#L1368

When there are grandchildren processes and pyfunc=True, the above code
will double-count the memory from the middle process.

I propose skipping the test (pr: ) for now until this issue is fixed.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#1663, or mute the thread
https://github.com/notifications/unsubscribe-auth/AAOkp9X5SY50X4Dv-GvSg31mS7338xLgks5quBlMgaJpZM4KG43u
.

[oesteban]

I am continuing with an old PR here #1643. I wonder if that PR would fix this issue.

[berleant]

@chrisfilo Noted!
@oesteban I don't think that PR would fix the problem, but I only skimmed it. You would know better than I