fix windows utf8 encoding part 3/3

[phaustin]

see jupyter-book/jupyter-book#719

[codecov]

Codecov Report

Merging #203 into master will not change coverage.
The diff coverage is 83.33%.

@@           Coverage Diff           @@
##           master     #203   +/-   ##
=======================================
  Coverage   85.49%   85.49%           
=======================================
  Files           9        9           
  Lines         834      834           
=======================================
  Hits          713      713           
  Misses        121      121           

Flag	Coverage Δ
#pytests	85.49% <83.33%> (ø)

Impacted Files	Coverage Δ
myst_nb/nb_glue/utils.py	73.01% <0.00%> (ø)
myst_nb/cache.py	78.44% <100.00%> (ø)
myst_nb/converter.py	77.69% <100.00%> (ø)
myst_nb/nb_glue/domain.py	90.64% <100.00%> (ø)

---

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 962d535...72e5f47. Read the comment docs.

[phaustin]

@choldgraf -- I reverted one change to get the file_regression test to pass -- the open statement
in the notebook artifact should also specify encoding="utf8":
i.e. this commit

[choldgraf]

@phaustin re: the file regression, I'm not quite sure what you mean by reverted. do you mean there's one part in this codebase that doesn't use utf-8 now?

[phaustin]

I mean that if I add encoding="utf8" to the ipynb version of this file
pytest fails with a "file regression" error.

[choldgraf]

could you post the error?

[phaustin]

reran with the same command line arguments as the CI, and all tests pass locally on my mac, I think it was a timeout issue, so I think this is ready to merge. On my local windows there are still about half a dozen pytest failures, but several of those are related to the windows path problem. I've isolated that issue to app.env.doc2path in cache.py (and also in jupyter-book/toc.py) and we can retest once I've pushed those PRs.

[choldgraf]

ok cool! let's merge this in and see how things go :-) thanks for the improvement!