micromark/dist/tokenize/character-reference.mjs

small commonmark compliant markdown parser with positional info and concrete tokens

import decodeEntity from 'parse-entities/decode-entity.js'
import asciiAlphanumeric from '../character/ascii-alphanumeric.mjs'
import asciiDigit from '../character/ascii-digit.mjs'
import asciiHexDigit from '../character/ascii-hex-digit.mjs'

var characterReference = {
  name: 'characterReference',
  tokenize: tokenizeCharacterReference
}

function tokenizeCharacterReference(effects, ok, nok) {
  var self = this
  var size = 0
  var max
  var test
  return start

  function start(code) {
    effects.enter('characterReference')
    effects.enter('characterReferenceMarker')
    effects.consume(code)
    effects.exit('characterReferenceMarker')
    return open
  }

  function open(code) {
    if (code === 35) {
      effects.enter('characterReferenceMarkerNumeric')
      effects.consume(code)
      effects.exit('characterReferenceMarkerNumeric')
      return numeric
    }

    effects.enter('characterReferenceValue')
    max = 31
    test = asciiAlphanumeric
    return value(code)
  }

  function numeric(code) {
    if (code === 88 || code === 120) {
      effects.enter('characterReferenceMarkerHexadecimal')
      effects.consume(code)
      effects.exit('characterReferenceMarkerHexadecimal')
      effects.enter('characterReferenceValue')
      max = 6
      test = asciiHexDigit
      return value
    }

    effects.enter('characterReferenceValue')
    max = 7
    test = asciiDigit
    return value(code)
  }

  function value(code) {
    var token

    if (code === 59 && size) {
      token = effects.exit('characterReferenceValue')

      if (
        test === asciiAlphanumeric &&
        !decodeEntity(self.sliceSerialize(token))
      ) {
        return nok(code)
      }

      effects.enter('characterReferenceMarker')
      effects.consume(code)
      effects.exit('characterReferenceMarker')
      effects.exit('characterReference')
      return ok
    }

    if (test(code) && size++ < max) {
      effects.consume(code)
      return value
    }

    return nok(code)
  }
}

export default characterReference