Evaluate Deprecating Syntactic Content Assist Support

[bd82]

• https://sap.github.io/chevrotain/docs/features/syntactic_content_assist.html

The issues are:

1. This feature is not orthogonal with other features (e.g error recovery/gates)
   as it is implemented as a separate interpreter directly on the GAST structure.
   Basically it is a different flow to the regular parsing flow.

2. This feature is slow (due to interpreter approach).

3. Such a feature should probably be implemented on the CST structure post parsing.
   See: https://github.com/SAP/xml-tools/tree/master/packages/content-assist

4. General need to reduce API surface area and complexity of Chevrotain.

[bd82]

I've started looking into this and it does not seem that removing this feature will actually git rid of significant amount of productive code.
(~150 LOC). Maybe it is best to change the API instead of removing it completely?

• Expose it as a utility NextPossibleTokenAfter
• Document it could be used for trivial content assist and its limitations (small input size, e.g when each line known to be a different statement).
• Provide a more complex content assist example that builds upon error recovery and CST visitor (like in XML-Tools)
  - Such an example can also implement semantic content assist.
  - Perhaps SQL like syntax in the tutorial would be a good place to expand upon.

One thing to consider is the value of such NextPossibleTokenAfter utility?
Does it have a legitimate purpose?