Add multi context

[drew-512]

Hi @sbinet

So this PR has all the py.Context stuff wrapped into it minus:

• examples/embedded
• examples/multi-context
• py/util.go (consumed by examples/embedded)
• README.md edits

On the sync.CloseSync issue you raised in the prev thread, the intention is to ensure that multiple holders of a py.Context can all call Close() without any (future) consequences. IMO, this is congruent with multiple holders of a py.Context (and how it's annoying when ya have to ensure that only one Close() is called). People see that close code and they can feel comforted it's an interface ready for the real world. And it basically comes at no cost or allocations.

The last commit on this PR should prevent any conflicts (assuming you merge PR 156 and 157 first)

Otherwise, lmk what you'd like to add/remove to this, otherwise I'll wait for this merge and then I'll immediately submit a PR with the above items rolled into a final PR or lmk what I should separate, etc. Thanks.

[codecov]

Codecov Report

Merging #158 (87d5755) into master (3068cb5) will decrease coverage by 0.17%.
The diff coverage is 66.12%.

@@            Coverage Diff             @@
##           master     #158      +/-   ##
==========================================
- Coverage   74.55%   74.38%   -0.18%     
==========================================
  Files          64       66       +2     
  Lines       11187    11395     +208     
==========================================
+ Hits         8341     8476     +135     
- Misses       2297     2352      +55     
- Partials      549      567      +18     

Impacted Files	Coverage Δ
py/int.go	77.13% <0.00%> (+0.26%)	⬆️
py/traceback.go	22.22% <ø> (-2.11%)	⬇️
py/import.go	15.62% <45.16%> (-0.86%)	⬇️
modules/runtime.go	45.86% <45.86%> (ø)
py/run.go	60.46% <60.46%> (ø)
py/module.go	71.69% <72.50%> (+11.69%)	⬆️
pytest/pytest.go	49.09% <81.81%> (+1.92%)	⬆️
py/list.go	64.13% <83.33%> (+0.03%)	⬆️
sys/sys.go	60.75% <88.23%> (+1.60%)	⬆️
builtin/builtin.go	80.06% <88.88%> (+0.03%)	⬆️
... and 13 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 3068cb5...87d5755. Read the comment docs.

[sbinet]

LGTM