ENH: Stop printing false positive differences when logging cached nodes

[oesteban]

The function calculating changed input fields did not preempt the problem of comparing tuples vs. lists in most of the cases. This PR simplifies the logic of the dict_diff accessory function, and fixes the issue - resulting in way shorter indices of changed inputs (esp. when an interface has several inputs and some of them contain long lists).

This issue translated in practice into showing long lists of unchanged values as if they had changed, when often, the changed input invalidating the cache would be some other field.

EDIT: Beyond the more robust implementation (former was kind of brittle), this PR is useful, e.g., when re-running fMRIPrep and large lists of files have changed (e.g., the inputs to an fslmerge node with a list of 800 input files).

[oesteban]

I don't think the failing test is related, right?

[effigies]

@oesteban Heads up that I rebased.

[codecov]

Codecov Report

Merging #3376 (1beec59) into master (d8dbc6f) will increase coverage by 0.07%.
The diff coverage is 90.90%.

@@ Coverage Diff @@ ## master #3376 +/- ## ========================================== + Coverage 64.88% 64.95% +0.07%  ========================================== Files 307 307 Lines 40365 40365 Branches 5337 5339 +2 ========================================== + Hits 26189 26220 +31  + Misses 13081 13049 -32  - Partials 1095 1096 +1 

Flag	Coverage Δ
unittests	64.95% <90.90%> (+0.07%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
nipype/utils/misc.py	78.98% <90.90%> (+19.74%)	⬆️

---

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 d8dbc6f...1beec59. Read the comment docs.