parser_inline.js

/** internal
 * class ParserInline
 *
 * Tokenizes paragraph content.
 **/
'use strict';


var Ruler           = require('./ruler');


////////////////////////////////////////////////////////////////////////////////
// Parser rules

var _rules = [
  [ 'text',            require('./rules_inline/text') ],
  [ 'linkify',         require('./rules_inline/linkify') ],
  [ 'newline',         require('./rules_inline/newline') ],
  [ 'escape',          require('./rules_inline/escape') ],
  [ 'backticks',       require('./rules_inline/backticks') ],
  [ 'strikethrough',   require('./rules_inline/strikethrough').tokenize ],
  [ 'emphasis',        require('./rules_inline/emphasis').tokenize ],
  [ 'link',            require('./rules_inline/link') ],
  [ 'image',           require('./rules_inline/image') ],
  [ 'autolink',        require('./rules_inline/autolink') ],
  [ 'html_inline',     require('./rules_inline/html_inline') ],
  [ 'entity',          require('./rules_inline/entity') ]
];

// `rule2` ruleset was created specifically for emphasis/strikethrough
// post-processing and may be changed in the future.
//
// Don't use this for anything except pairs (plugins working with `balance_pairs`).
//
var _rules2 = [
  [ 'balance_pairs',   require('./rules_inline/balance_pairs') ],
  [ 'strikethrough',   require('./rules_inline/strikethrough').postProcess ],
  [ 'emphasis',        require('./rules_inline/emphasis').postProcess ],
  // rules for pairs separate '**' into its own text tokens, which may be left unused,
  // rule below merges unused segments back with the rest of the text
  [ 'fragments_join',  require('./rules_inline/fragments_join') ]
];


/**
 * new ParserInline()
 **/
function ParserInline() {
  var i;

  /**
   * ParserInline#ruler -> Ruler
   *
   * [[Ruler]] instance. Keep configuration of inline rules.
   **/
  this.ruler = new Ruler();

  for (i = 0; i < _rules.length; i++) {
    this.ruler.push(_rules[i][0], _rules[i][1]);
  }

  /**
   * ParserInline#ruler2 -> Ruler
   *
   * [[Ruler]] instance. Second ruler used for post-processing
   * (e.g. in emphasis-like rules).
   **/
  this.ruler2 = new Ruler();

  for (i = 0; i < _rules2.length; i++) {
    this.ruler2.push(_rules2[i][0], _rules2[i][1]);
  }
}


// Skip single token by running all rules in validation mode;
// returns `true` if any rule reported success
//
ParserInline.prototype.skipToken = function (state) {
  var ok, i, pos = state.pos,
      rules = this.ruler.getRules(''),
      len = rules.length,
      maxNesting = state.md.options.maxNesting,
      cache = state.cache;


  if (typeof cache[pos] !== 'undefined') {
    state.pos = cache[pos];
    return;
  }

  if (state.level < maxNesting) {
    for (i = 0; i < len; i++) {
      // Increment state.level and decrement it later to limit recursion.
      // It's harmless to do here, because no tokens are created. But ideally,
      // we'd need a separate private state variable for this purpose.
      //
      state.level++;
      ok = rules[i](state, true);
      state.level--;

      if (ok) {
        if (pos >= state.pos) { throw new Error("inline rule didn't increment state.pos"); }
        break;
      }
    }
  } else {
    // Too much nesting, just skip until the end of the paragraph.
    //
    // NOTE: this will cause links to behave incorrectly in the following case,
    //       when an amount of `[` is exactly equal to `maxNesting + 1`:
    //
    //       [[[[[[[[[[[[[[[[[[[[[foo]()
    //
    // TODO: remove this workaround when CM standard will allow nested links
    //       (we can replace it by preventing links from being parsed in
    //       validation mode)
    //
    state.pos = state.posMax;
  }

  if (!ok) { state.pos++; }
  cache[pos] = state.pos;
};


// Generate tokens for input range
//
ParserInline.prototype.tokenize = function (state) {
  var ok, i, prevPos,
      rules = this.ruler.getRules(''),
      len = rules.length,
      end = state.posMax,
      maxNesting = state.md.options.maxNesting;

  while (state.pos < end) {
    // Try all possible rules.
    // On success, rule should:
    //
    // - update `state.pos`
    // - update `state.tokens`
    // - return true
    prevPos = state.pos;

    if (state.level < maxNesting) {
      for (i = 0; i < len; i++) {
        ok = rules[i](state, false);
        if (ok) {
          if (prevPos >= state.pos) { throw new Error("inline rule didn't increment state.pos"); }
          break;
        }
      }
    }

    if (ok) {
      if (state.pos >= end) { break; }
      continue;
    }

    state.pending += state.src[state.pos++];
  }

  if (state.pending) {
    state.pushPending();
  }
};


/**
 * ParserInline.parse(str, md, env, outTokens)
 *
 * Process input string and push inline tokens into `outTokens`
 **/
ParserInline.prototype.parse = function (str, md, env, outTokens) {
  var i, rules, len;
  var state = new this.State(str, md, env, outTokens);

  this.tokenize(state);

  rules = this.ruler2.getRules('');
  len = rules.length;

  for (i = 0; i < len; i++) {
    rules[i](state);
  }
};


ParserInline.prototype.State = require('./rules_inline/state_inline');


module.exports = ParserInline;