Accept only identifiers on LHS of :=

[adonovan]

This PR changes short_var_declaration to accept only a list of identifiers on the left side. This prevents an (expression_list (identifier)) query from spuriously matching an identifier in a declaration.

Also:

• we no longer treat the identifiers nil, true, false, iota, new, and make specially. All are legal identifiers, readily reused as field or method names without ambiguity, and the last three are not infrequently redefined as local variables.
• fold special cases for new(T) and make(T, len) into call_expression, which now allows a type for its first argument.
• added test case for built-in identifiers.
• in tests, replace 'expr' with 'var _ = expr' in cases where expr is not a valid standalone statement.
• remove iota/true/false/nil cases from highlighting now that they are not special identifiers.

I notice there is a tension here between what highlighting wants (mostly correct heuristics for things like new and iota) and what Stack Graphs wants (a minimal, regular, and semantically factored tree---since the program that converts TS trees to SG graphs can't express top-down recursive propagation of modal concepts like def vs. use).

Also, unrelated:

• reformatting of corpus. Not sure why this happens.
• clean up broken test case with ERROR expectations accidentally committed in a previous PR (sorry).
• merge dec_statement into inc_statement.
• commute field(..) and commaSep(..) operations in three productions so we don't name the commas.
• document build/test in README.

Checklist:

• [ ] All tests pass in CI.
• [ ] There are sufficient tests for the new fix/feature.
• [ ] Grammar rules have not been renamed unless absolutely necessary.
• [ ] The conflicts section hasn't grown too much.
• [ ] The parser size hasn't grown too much (check the value of STATE_COUNT in src/parser.c).

[maxbrunsfeld]

Makes sense!

[dominikh]

Are there any plans to move forward with this PR? In particular

we no longer treat the identifiers nil, true, false, iota, new, and make specially. All are legal identifiers, readily reused as field or method names without ambiguity, and the last three are not infrequently redefined as local variables.

would be nice to get in.

[adonovan]

I have no plans to do anything with this PR. Feel free to take it over.