DEBUG: Revert to pre-error trees and enable AppVeyor debugging

[effigies]

No description provided.

[yarikoptic]

Restart a few times?

[effigies]

https://ci.appveyor.com/project/nipy/nibabel/builds/19485598/job/cxp38t30ee2aylci

[yarikoptic]

Yeap... So should we try one more used to be ok commit a month before that one? If that one fails than most likely it is not nibabel fault, and python and/or numpy or some library

[effigies]

Restarted, because there were files added during these intervals that don't get removed with git checkout <githash> .. That needs to be followed by a git checkout -p <githash> ..

Two targets are dfd113c (Jun 4) and 68cf671 (Apr 24).

4 build attempts on each.

[effigies]

Ah, this may be relevant:

I think next step should be to try a range of Numpy versions, starting from 1.14.2 or so.

[effigies]

Error from dfd113c reversion: https://ci.appveyor.com/project/nipy/nibabel/builds/19518274/job/24o3p60j85fu6g7j

[nibotmi]

☔ The latest upstream changes (presumably 1e9c4d4) made this pull request unmergeable. Please resolve the merge conflicts.

[effigies]

Yes! Failure at last! So it's not our fault. I'll iterate over numpy releases next.

cc @yarikoptic

[effigies]

Problem reproduces in numpy 1.14.2: https://ci.appveyor.com/project/nipy/nibabel/builds/19582043/job/8iuss25ktq30hqde?fullLog=true

Will push back a few versions, just to be sure.

[yarikoptic]

whoohoo -- what a fishing expedition ;-) "love it"!

[yarikoptic]

grr... only 1.12 fails similarly ... so either it is even before then, or not numpy fault

[effigies]

Looks like probably not numpy's fault.

[nibotmi]

☔ The latest upstream changes (presumably #700) made this pull request unmergeable. Please resolve the merge conflicts.

[effigies]

This effort is dead. The bug seems to be happening less lately, so hopefully whatever was going on over at AppVeyor is resolved. I think if this recurs, we should think about moving to Azure Pipelines for Windows tests.