Markdown bug injecting ‘klzzwxh’ into pages, need a fix and tests to ensure it stays fix

[danbri]

@RichardWallis suggested making a "test to try all Term pages on http://schema.org looking for ‘klzzwxh’"

+1

Immediate fixes and more diagnosis also needed, we've repaired health-lifesci.schema.org but the markdown library seems prone to this so unit tests are essential to avoid future recurrence of the problem.

[AymenLoukil]

Search page http://schema.org/docs/search_results.html#q=klzzwxh is showing klzzwxh

[danbri]

See also:

• https://sourceforge.net/p/python-markdown/mailman/message/24612191/
• Nested inline pattern always returns klzzwxh:0002 Python-Markdown/markdown#354
• Links break if they contain underscores Python-Markdown/markdown#32

[danbri]

We decided to switch to a different markdown library, plus a custom function for parsing wikilink-style link syntax.

• Moved to Markdown2 Python libray and use our own Wikilinks parsing #1287 merged.

[danbri]

Deployed! Cleanup tasks -

• expunge old library (and file a bug for them, although we can't reproduce it reliably)
• updated scripts/*py to use markdown2
• given the library switch do we still need to integrate tests for klzzwxh or is that string tied to the oddities of the first library? (if in doubt, test, I guess...)
• anything else @RichardWallis ?

[RichardWallis]

This problem was specific to the Markdown library being used.

This has been replaced by another (Markdown2) library, implemented in an entirely different way so 'this' problem will not reappear in this form.

Hense closing.

[danbri]

@RichardWallis can you confirm everything in scripts/ uses the new lib?

[RichardWallis]

Yes - no code including scripts and tests make reference to the old Markdown library.

Commit 661088b removes references to it in SOFTWARE_README.md