Split index into multiple parts and add automated document check

[jjhursey]

• Multiple indices will make it easier to find APIs vs attributes vs ...
• Add an automated document cross-check against OpenPMIx
  • Checks PMIx Standard items not included in OpenPMIx
  • Checks OpenPMIx items not included in PMIx Standard

Closes #149

[jjhursey]

Ref Issue #149 - this will generate .idx files that can be used to access the attributes/apis/macros/constants/structs from the standard. That will be useful when building the document checker.

Keeping this a WIP for now as I work on the document checker script.

[jjhursey]

PDF rendering of the current state of the PR (so you can see the indexing)
pmix-standard-pr-241-v2.pdf

[jjhursey]

I plan to discuss the indexing changes in the PMIx Standard meeting on Friday, June 11. So I won't plan to merge before then.

This is missing one optional feature of (we don't need it right now so we can wait on it)

• Whitelisting of items that are ok to be missing on either side (e.g., something in OpenPMIx that is not meant to be part of the PMIx Standard or vise versa)

[SteVwonder]

Very nice @jjhursey! I wonder if we can leverage this work to improve the coverage analysis and cleanup some of the inconsistencies that @rhc54 pointed out.

[jjhursey]

Per teleconf June 12, 2020

• Can we remove the Definition line, and instead have a bolded page number at the front of the reference list? That would compress the index a bit, and still preserve meaning.
• Generally like the breakdown in the current index. Generally ok with the mix of double and single-column format.

[jjhursey]

pmix-standard.pdf

This version uses bold and underlining to denote the Definition in the index (instead of a separate Definition line). I moved all the indices to single column as that helped scanning down the list for a given item.

I think this is ready for review for inclusion in v4.

[rhc54]

Looks good! Thanks!