could not run flopy3_MT3DMS_examples.ipynb

[mwtoews]

While running autotest_notebooks.py on Python 3.7, a new issue is found. See traceback starting here:
https://travis-ci.org/modflowpy/flopy/jobs/499596952#L1137

[christianlangevin]

This is an intermittent issue that I haven't been able to figure out. For some reason, flopy is having a hard time reading the concentration file, but only some times. I can reproduce it locally sometimes. Strange.

[jtwhite79]

I've got this problem now too on the feature_drt branch I've been working in with py3.7 and the MT3DMS notebook on travis. I fired up a py37 notebook on ubuntu and can run to completion (using a fresh mt3dms builds with gcc7). It is possible to install a travis docker image but I don't have the bandwidth at the mo. Any thoughts?

[mwtoews]

This issue has nothing to do with flopy or Python. It's with gfortran-compiled mt3dms; see MODFLOW-ORG/mt3d-usgs#46 for a possible fix.

[emorway-usgs]

@mwtoews offered this pull request on the MT3D-USGS repo which has been adopted in the latest release of MT3D-USGS that is now up on the web. Hopefully the initialization of the velocity array in MT3D-USGS fixes this issue. Does that mean that a new MT3D-USGS executable needs to be stored somewhere for the flopy testing to use?

[christianlangevin]

We'll need to update the path to the distribution file. I'll take a look at it.