update dict key limits per 2.1 spec

[treyka]

In the course of testing the i18n code in the 2.1 branch, I noticed that while the language-content contents property allows 2 char dict keys for specifying languages (as shown in all of the examples and per RFC5646), the 2.1 branch still retained the 2.0 3 char restriction on dict key length.

QED, I split properties.py out v20 vs v21 as was done previously to accommodate other 2.1 changes.

Also, somewhat inexplicably the STIX2 dictionary property limits key length to 250 (not 256). While I was mucking about, I fixed this on for 2.1.

[codecov-io]

Codecov Report

Merging #198 into stix2.1 will decrease coverage by 1.27%.
The diff coverage is 90.23%.

@@             Coverage Diff             @@
##           stix2.1     #198      +/-   ##
===========================================
- Coverage    97.06%   95.79%   -1.28%     
===========================================
  Files           75       76       +1     
  Lines         7525     7745     +220     
===========================================
+ Hits          7304     7419     +115     
- Misses         221      326     +105

Impacted Files	Coverage Δ
stix2/base.py	99.47% <0%> (ø)	⬆️
stix2/test/test_language_content.py	100% <100%> (+21.42%)	⬆️
stix2/v20/observables.py	53.89% <100%> (-14.02%)	⬇️
stix2/v21/sdo.py	100% <100%> (ø)	⬆️
stix2/v21/sro.py	100% <100%> (ø)	⬆️
stix2/v21/observables.py	100% <100%> (ø)	⬆️
stix2/core.py	100% <100%> (ø)	⬆️
stix2/v20/properties.py	90.13% <100%> (ø)
stix2/v21/bundle.py	100% <100%> (ø)	⬆️
stix2/v20/bundle.py	39.47% <100%> (ø)	⬆️
... and 15 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 b852b91...59fdd30. Read the comment docs.

[emmanvg]

@treyka thanks for updating the properties and make LanguageContent work. I removed the mark.xfail() for the test to make sure it passes now. I saw that you created properties.py for both versions, but honestly most of the code is duplicated. We will revisit this in #193. I had to make a minimal amount of changes to the tests to make sure they pass. We will revisit this for #197 and #80.

[emmanvg]

There is more stuff to be done, but we can address in a different PR.

[gtback]

Thanks, @treyka !