Make the rest of flatteners functional in anticipation of #381

[tmcw]

cc @jfirebaugh the singleUse property of these functions will power the lint check in #381

[tmcw]

This now hits the following lint errors:

• When you use a tag that's "single use only" multiple times (any of the tags that flatten flattens to a non-array value)
• When you use a combination of tags that result in something getting overwritten:
• A "kind" tag plus explicit @kind
• A "kind" tag with type plus explicit @type
• A "kind" tag with name plus explicit @name or @alias

cc @jfirebaugh for the review

[jfirebaugh]

I think this would be better handled in each individual flattener function, rather than in a separate lint step. Handling in the flattener function keeps the logic for a given tag in a single place, and can more easily handle cases like where both @class and @kind class appear (whichever one comes later should get linted as a conflict to the first one).

[tmcw]

I think it'll be harder to do it that way: an important part of these messages is that they say

• warning tag private can only be used once but is used multiple times
• warning tag access can only be used once, but is already declared by its synonym private

They can differentiate between tags declared multiple times and those declared by a synonym. We can do that cleanly in the linting step, but not really as cleanly in the individual flatteners that only have awareness of a single tag, by design. Otherwise we'd need to do extra bookkeeping to track already-defined tags - something we can do, but I'm betting is less clean.

[jfirebaugh]

Yes -- I have ideas in that vein.

Can we please hold on this for now? It's going to cause conflicts with #388 and some other parsing changes I have queued up.

[tmcw]

Sure, will let this PR sit as-is.

[tmcw]

Closing - I trust @jfirebaugh has a better plan for this.