feat: implement scanSync

[markandrus]

This PR exposes pg_query_scan as scanSync, copying the examples from https://github.com/pganalyze/libpg_query/blob/16-latest/examples/scan.c and the existing splitWithScannerSync function. I also added Token and KeywordKind types.

import { scanSync } from '@pg-nano/pg-parser'

const input = 'SELECT xyz WHERE xyz = 123'

const tokens = scanSync(inputs)

console.log(tokens)

/*

[
  { kind: 'SELECT', start: 0, end: 6, keyword: 'RESERVED_KEYWORD' },
  { kind: 'IDENT', start: 7, end: 10, keyword: 'NO_KEYWORD' },
  { kind: 'WHERE', start: 11, end: 16, keyword: 'RESERVED_KEYWORD' },
  { kind: 'IDENT', start: 17, end: 20, keyword: 'NO_KEYWORD' },
  { kind: 'ASCII_61', start: 21, end: 22, keyword: 'NO_KEYWORD' },
  { kind: 'ICONST', start: 23, end: 26, keyword: 'NO_KEYWORD' }
]

*/

for (const token of tokens) {
  console.log(input.slice(token.start, token.end))
}

/*

SELECT
xyz
WHERE
xyz
=
123

*/

My hope is to use the tokens' start and end values in my application to annotate the parse tree from parseQuerySync with more complete location information.

EDIT: This function is working well for me. I blogged about @pg-nano/pg-parser and using it here!

[aleclarson]

Ooh that's really sleek! Nice find. I'll give this a review after I'm done fixing #3. 👍