25573: Improves documentation and spell checking process

[howsohazard]

No description provided.

[domg123-hermes]

One thing I’d double-check in the new unparse / print descriptions: the @ emission part looks conditional on include_attributes, but the new wording reads unconditional.

Specifically, InterpretNode_ENT_UNPARSE defaults include_attributes to false and passes it through to Parser::Unparse(tree, pretty, include_attributes, ...). In Parser::Unparse, the @ marker for preevaluationNeeded is only appended inside the if(upd.emitAttributes) block. print also calls Parser::Unparse(cur, true, false, true), so it looks like it is safe for cyclic/graph structures but does not emit @ markers for parse-reconstructing those relationships by default.

So perhaps these should be softened to something like: cyclic/graph unparsing is safe, and parse-reconstructing @ output is available when attributes are included; print is safe but may not produce reconstructing @ output.