remark-parse-no-trim/lib/tokenize/list.js

remark plugin to parse Markdown

'use strict'

var repeat = require('repeat-string')
var decimal = require('is-decimal')
var getIndent = require('../util/get-indentation')
var removeIndent = require('../util/remove-indentation')
var interrupt = require('../util/interrupt')

module.exports = list

var asterisk = '*'
var underscore = '_'
var plusSign = '+'
var dash = '-'
var dot = '.'
var space = ' '
var lineFeed = '\n'
var tab = '\t'
var rightParenthesis = ')'
var lowercaseX = 'x'

var tabSize = 4
var looseListItemExpression = /\n\n(?!\s*$)/
var taskItemExpression = /^\[([ X\tx])][ \t]/
var bulletExpression = /^([ \t]*)([*+-]|\d+[.)])( {1,4}(?! )| |\t|$|(?=\n))([^\n]*)/
var pedanticBulletExpression = /^([ \t]*)([*+-]|\d+[.)])([ \t]+)/
var initialIndentExpression = /^( {1,4}|\t)?/gm

function list(eat, value, silent) {
  var self = this
  var commonmark = self.options.commonmark
  var pedantic = self.options.pedantic
  var tokenizers = self.blockTokenizers
  var interuptors = self.interruptList
  var index = 0
  var length = value.length
  var start = null
  var size
  var queue
  var ordered
  var character
  var marker
  var nextIndex
  var startIndex
  var prefixed
  var currentMarker
  var content
  var line
  var previousEmpty
  var empty
  var items
  var allLines
  var emptyLines
  var item
  var enterTop
  var exitBlockquote
  var spread = false
  var node
  var now
  var end
  var indented

  while (index < length) {
    character = value.charAt(index)

    if (character !== tab && character !== space) {
      break
    }

    index++
  }

  character = value.charAt(index)

  if (character === asterisk || character === plusSign || character === dash) {
    marker = character
    ordered = false
  } else {
    ordered = true
    queue = ''

    while (index < length) {
      character = value.charAt(index)

      if (!decimal(character)) {
        break
      }

      queue += character
      index++
    }

    character = value.charAt(index)

    if (
      !queue ||
      !(character === dot || (commonmark && character === rightParenthesis))
    ) {
      return
    }

    /* Slightly abusing `silent` mode, whose goal is to make interrupting
     * paragraphs work.
     * Well, that’s exactly what we want to do here: don’t interrupt:
     * 2. here, because the “list” doesn’t start with `1`. */
    if (silent && queue !== '1') {
      return
    }

    start = parseInt(queue, 10)
    marker = character
  }

  character = value.charAt(++index)

  if (
    character !== space &&
    character !== tab &&
    (pedantic || (character !== lineFeed && character !== ''))
  ) {
    return
  }

  if (silent) {
    return true
  }

  index = 0
  items = []
  allLines = []
  emptyLines = []

  while (index < length) {
    nextIndex = value.indexOf(lineFeed, index)
    startIndex = index
    prefixed = false
    indented = false

    if (nextIndex === -1) {
      nextIndex = length
    }

    size = 0

    while (index < length) {
      character = value.charAt(index)

      if (character === tab) {
        size += tabSize - (size % tabSize)
      } else if (character === space) {
        size++
      } else {
        break
      }

      index++
    }

    if (item && size >= item.indent) {
      indented = true
    }

    character = value.charAt(index)
    currentMarker = null

    if (!indented) {
      if (
        character === asterisk ||
        character === plusSign ||
        character === dash
      ) {
        currentMarker = character
        index++
        size++
      } else {
        queue = ''

        while (index < length) {
          character = value.charAt(index)

          if (!decimal(character)) {
            break
          }

          queue += character
          index++
        }

        character = value.charAt(index)
        index++

        if (
          queue &&
          (character === dot || (commonmark && character === rightParenthesis))
        ) {
          currentMarker = character
          size += queue.length + 1
        }
      }

      if (currentMarker) {
        character = value.charAt(index)

        if (character === tab) {
          size += tabSize - (size % tabSize)
          index++
        } else if (character === space) {
          end = index + tabSize

          while (index < end) {
            if (value.charAt(index) !== space) {
              break
            }

            index++
            size++
          }

          if (index === end && value.charAt(index) === space) {
            index -= tabSize - 1
            size -= tabSize - 1
          }
        } else if (character !== lineFeed && character !== '') {
          currentMarker = null
        }
      }
    }

    if (currentMarker) {
      if (!pedantic && marker !== currentMarker) {
        break
      }

      prefixed = true
    } else {
      if (!commonmark && !indented && value.charAt(startIndex) === space) {
        indented = true
      } else if (commonmark && item) {
        indented = size >= item.indent || size > tabSize
      }

      prefixed = false
      index = startIndex
    }

    line = value.slice(startIndex, nextIndex)
    content = startIndex === index ? line : value.slice(index, nextIndex)

    if (
      currentMarker === asterisk ||
      currentMarker === underscore ||
      currentMarker === dash
    ) {
      if (tokenizers.thematicBreak.call(self, eat, line, true)) {
        break
      }
    }

    previousEmpty = empty
    empty = !prefixed && !content.trim().length

    if (indented && item) {
      item.value = item.value.concat(emptyLines, line)
      allLines = allLines.concat(emptyLines, line)
      emptyLines = []
    } else if (prefixed) {
      if (emptyLines.length !== 0) {
        spread = true
        item.value.push('')
        item.trail = emptyLines.concat()
      }

      item = {
        value: [line],
        indent: size,
        trail: []
      }

      items.push(item)
      allLines = allLines.concat(emptyLines, line)
      emptyLines = []
    } else if (empty) {
      if (previousEmpty && !commonmark) {
        break
      }

      emptyLines.push(line)
    } else {
      if (previousEmpty) {
        break
      }

      if (interrupt(interuptors, tokenizers, self, [eat, line, true])) {
        break
      }

      item.value = item.value.concat(emptyLines, line)
      allLines = allLines.concat(emptyLines, line)
      emptyLines = []
    }

    index = nextIndex + 1
  }

  node = eat(allLines.join(lineFeed)).reset({
    type: 'list',
    ordered: ordered,
    start: start,
    spread: spread,
    children: []
  })

  enterTop = self.enterList()
  exitBlockquote = self.enterBlock()
  index = -1
  length = items.length

  while (++index < length) {
    item = items[index].value.join(lineFeed)
    now = eat.now()

    eat(item)(listItem(self, item, now), node)

    item = items[index].trail.join(lineFeed)

    if (index !== length - 1) {
      item += lineFeed
    }

    eat(item)
  }

  enterTop()
  exitBlockquote()

  return node
}

function listItem(ctx, value, position) {
  var offsets = ctx.offset
  var fn = ctx.options.pedantic ? pedanticListItem : normalListItem
  var checked = null
  var task
  var indent

  value = fn.apply(null, arguments)

  if (ctx.options.gfm) {
    task = value.match(taskItemExpression)

    if (task) {
      indent = task[0].length
      checked = task[1].toLowerCase() === lowercaseX
      offsets[position.line] += indent
      value = value.slice(indent)
    }
  }

  return {
    type: 'listItem',
    spread: looseListItemExpression.test(value),
    checked: checked,
    children: ctx.tokenizeBlock(value, position)
  }
}

// Create a list-item using overly simple mechanics.
function pedanticListItem(ctx, value, position) {
  var offsets = ctx.offset
  var line = position.line

  // Remove the list-item’s bullet.
  value = value.replace(pedanticBulletExpression, replacer)

  // The initial line was also matched by the below, so we reset the `line`.
  line = position.line

  return value.replace(initialIndentExpression, replacer)

  // A simple replacer which removed all matches, and adds their length to
  // `offset`.
  function replacer($0) {
    offsets[line] = (offsets[line] || 0) + $0.length
    line++

    return ''
  }
}

// Create a list-item using sane mechanics.
function normalListItem(ctx, value, position) {
  var offsets = ctx.offset
  var line = position.line
  var max
  var bullet
  var rest
  var lines
  var trimmedLines
  var index
  var length

  // Remove the list-item’s bullet.
  value = value.replace(bulletExpression, replacer)

  lines = value.split(lineFeed)

  trimmedLines = removeIndent(value, getIndent(max).indent).split(lineFeed)

  // We replaced the initial bullet with something else above, which was used
  // to trick `removeIndentation` into removing some more characters when
  // possible.  However, that could result in the initial line to be stripped
  // more than it should be.
  trimmedLines[0] = rest

  offsets[line] = (offsets[line] || 0) + bullet.length
  line++

  index = 0
  length = lines.length

  while (++index < length) {
    offsets[line] =
      (offsets[line] || 0) + lines[index].length - trimmedLines[index].length
    line++
  }

  return trimmedLines.join(lineFeed)

  /* eslint-disable-next-line max-params */
  function replacer($0, $1, $2, $3, $4) {
    bullet = $1 + $2 + $3
    rest = $4

    // Make sure that the first nine numbered list items can indent with an
    // extra space.  That is, when the bullet did not receive an extra final
    // space.
    if (Number($2) < 10 && bullet.length % 2 === 1) {
      $2 = space + $2
    }

    max = $1 + repeat(space, $2.length) + $3

    return max + rest
  }
}