MNT: Use 'datalad push' instead of deprecated 'datalad publish'

[kyleam]

'datalad publish' was marked as deprecated in 0.14 in favor of 'datalad push', which was added in 0.13 (ReproMan's current minimum version). For the purpose of prepare_remote(), the main differences are 1) the '' value for --since must be spelled as '^', and 2) push() must be called with data=auto to retain the behavior of the publish() call (though going forward it's worth considering whether data=nothing should be used instead).

[codecov[bot]]

Codecov Report

Merging #575 (5f982f2) into master (6c49cfe) will decrease coverage by 0.45%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #575      +/-   ##
==========================================
- Coverage   89.22%   88.77%   -0.46%     
==========================================
  Files         149      149              
  Lines       13054    12762     -292     
==========================================
- Hits        11647    11329     -318     
- Misses       1407     1433      +26     

Impacted Files	Coverage Δ
reproman/support/jobs/orchestrators.py	94.03% <100.00%> (-0.17%)	:arrow_down:
reproman/distributions/tests/test_venv.py	85.71% <0.00%> (-10.02%)	:arrow_down:
reproman/tests/skip.py	93.25% <0.00%> (-4.52%)	:arrow_down:
...eproman/interface/tests/test_backend_parameters.py	95.65% <0.00%> (-4.35%)	:arrow_down:
reproman/formats/base.py	59.57% <0.00%> (-2.43%)	:arrow_down:
reproman/support/protocol.py	85.08% <0.00%> (-2.42%)	:arrow_down:
reproman/distributions/venv.py	86.04% <0.00%> (-2.33%)	:arrow_down:
reproman/distributions/tests/test_singularity.py	33.33% <0.00%> (-2.16%)	:arrow_down:
reproman/support/jobs/submitters.py	67.56% <0.00%> (-1.80%)	:arrow_down:
reproman/interface/test.py	85.71% <0.00%> (-1.79%)	:arrow_down:
... and 58 more

---

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 6c49cfe...5f982f2. Read the comment docs.

[yarikoptic]

Looks good, thank you!

though going forward it's worth considering whether data=nothing should be used instead

right -- probably might as well be nothing for that invocation since we just created a sibling and it would have no wanted set anyways... and if it is an existing one, having it as auto leaves possibility to trigger transmission of data which is not pertinent at that stage (of just establishing the remote, not when we transferring the inputs), so indeed "nothing" might be better.