More upstream compatibility fixes

[kyleam]

Since the last round of compatibility fixes, more issues have sneaked in. This PR should address the remaining failures from datalad/datalad#3835.

Marking as a draft because the tip commit should be dropped before merging. edit: ~Hrm, that didn't seem to trigger a Travis build; perhaps that's a fluke or perhaps there has been (another) change in how Travis treats draft PRs. I'll take it out of draft mode to see if a build is triggered.~ Never mind, the draft did trigger a build on Travis, I'm just not seeing the marker (yellow dot) come up here. pending build

[kyleam]

The build with datalad master looks good. (The failing run is a py2 run, which is expected.) I'll drop the temporary tip commit now.

[codecov[bot]]

Codecov Report

Merging #62 into master will increase coverage by 0.01%. The diff coverage is 0%.

@@            Coverage Diff             @@
##           master      #62      +/-   ##
==========================================
+ Coverage   82.43%   82.45%   +0.01%     
==========================================
  Files          53       53              
  Lines        4231     4228       -3     
==========================================
- Hits         3488     3486       -2     
+ Misses        743      742       -1

Impacted Files	Coverage Δ
datalad_crawler/nodes/tests/test_annex.py	98.04% <ø> (-0.01%)	:arrow_down:
datalad_crawler/pipelines/tests/test_openfmri.py	44.19% <0%> (ø)	:arrow_up:
datalad_crawler/nodes/tests/test_misc.py	100% <0%> (ø)	:arrow_up:
datalad_crawler/pipelines/openfmri.py	86.44% <0%> (+1.44%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 31d32f3...5dda277. Read the comment docs.

[yarikoptic]

Thank you! I guess I will cut one more minor release today

[kyleam]

I guess I will cut one more minor release today

That'd be great. Then we can add the crawler back to datalad's build (datalad/datalad#3835) and at least we'd be aware that we broke something, though that of course doesn't necessarily avoid repeating this same cycle given the current testing setup.

And sorry for the inconvenience. I should've been better about checking for incompatibilities given the crawler build on datalad's side was disabled.