ENH Add a switch to override existing global variables

[pierreglaser]

Following #214.
This PR implements a switch called override_existing_globals, available from CloudPickler.__init__, cloudpickle.dumps and cloudpickle.dump.

• If set to True, the globals from a function originally created in the __main__ module, dynamic modules or in a nested scope override the globals of the function's module at unpickling time.
• If set to False, the existing variables in the function's module overrides the globals of those functions at unpickling time.

[codecov-io]

Codecov Report

Merging #216 into master will increase coverage by 0.07%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #216 +/- ##
========================================
+ Coverage 84.92% 85% +0.07% 
========================================
Files 1 1 Lines 577 580 +3 Branches 114 115 +1 ========================================
+ Hits 490 493 +3 
Misses 63 63 Partials 24 24

Impacted Files	Coverage Δ
cloudpickle/cloudpickle.py	85% <100%> (+0.07%)	⬆️

---

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 7fd15c2...3f32518. Read the comment docs.

[tomMoral]

Overall, this looks good.

I agree that override_existing_globals should be passed through the state to ensure compatibility.
Also, some nitpicks on the tests and comments.

[pierreglaser]

@suquarkoverride_existing_globals is now in state.
@tomMoral I now loop over override_existing_globals in the test. Because of that, the diff looks messy. You can refer to this commit for a cleaner diff.

[tomMoral]

LGTM. just a typo.

[ogrisel]

+1 on my side besides the following comments.

I think we need a changelog entry targeting the 0.7 version.

[ogrisel]

I would re-add the original comment for this block of the condition:

# Only set global variables that do not exist.

[ogrisel]

LGTM, please add an entry to the change log (for 0.7).

[mitar]

@suquark Does this address fully #214?

[suquark]

@mitar It only partly fixes #214. Fixing another part is tricky. It is related to recursive functions and only fails in CI tests. We still don't know how it happens; it is about how Python deals with global variables of a created function.

[ogrisel]

@suquark could you please provide more details? Do you have a standalone example of one such problematic recursive function?

[ogrisel]

Closing in favor of #240.