xref: /plugin/asciidocjs/node_modules/nunjucks/src/lexer.js

'use strict';

var lib = require('./lib');
var whitespaceChars = " \n\t\r\xA0";
var delimChars = '()[]{}%*-+~/#,:|.<>=!';
var intChars = '0123456789';
var BLOCK_START = '{%';
var BLOCK_END = '%}';
var VARIABLE_START = '{{';
var VARIABLE_END = '}}';
var COMMENT_START = '{#';
var COMMENT_END = '#}';
var TOKEN_STRING = 'string';
var TOKEN_WHITESPACE = 'whitespace';
var TOKEN_DATA = 'data';
var TOKEN_BLOCK_START = 'block-start';
var TOKEN_BLOCK_END = 'block-end';
var TOKEN_VARIABLE_START = 'variable-start';
var TOKEN_VARIABLE_END = 'variable-end';
var TOKEN_COMMENT = 'comment';
var TOKEN_LEFT_PAREN = 'left-paren';
var TOKEN_RIGHT_PAREN = 'right-paren';
var TOKEN_LEFT_BRACKET = 'left-bracket';
var TOKEN_RIGHT_BRACKET = 'right-bracket';
var TOKEN_LEFT_CURLY = 'left-curly';
var TOKEN_RIGHT_CURLY = 'right-curly';
var TOKEN_OPERATOR = 'operator';
var TOKEN_COMMA = 'comma';
var TOKEN_COLON = 'colon';
var TOKEN_TILDE = 'tilde';
var TOKEN_PIPE = 'pipe';
var TOKEN_INT = 'int';
var TOKEN_FLOAT = 'float';
var TOKEN_BOOLEAN = 'boolean';
var TOKEN_NONE = 'none';
var TOKEN_SYMBOL = 'symbol';
var TOKEN_SPECIAL = 'special';
var TOKEN_REGEX = 'regex';
function token(type, value, lineno, colno) {
  return {
    type: type,
    value: value,
    lineno: lineno,
    colno: colno
  };
}
var Tokenizer = /*#__PURE__*/function () {
  function Tokenizer(str, opts) {
    this.str = str;
    this.index = 0;
    this.len = str.length;
    this.lineno = 0;
    this.colno = 0;
    this.in_code = false;
    opts = opts || {};
    var tags = opts.tags || {};
    this.tags = {
      BLOCK_START: tags.blockStart || BLOCK_START,
      BLOCK_END: tags.blockEnd || BLOCK_END,
      VARIABLE_START: tags.variableStart || VARIABLE_START,
      VARIABLE_END: tags.variableEnd || VARIABLE_END,
      COMMENT_START: tags.commentStart || COMMENT_START,
      COMMENT_END: tags.commentEnd || COMMENT_END
    };
    this.trimBlocks = !!opts.trimBlocks;
    this.lstripBlocks = !!opts.lstripBlocks;
  }
  var _proto = Tokenizer.prototype;
  _proto.nextToken = function nextToken() {
    var lineno = this.lineno;
    var colno = this.colno;
    var tok;
    if (this.in_code) {
      // Otherwise, if we are in a block parse it as code
      var cur = this.current();
      if (this.isFinished()) {
        // We have nothing else to parse
        return null;
      } else if (cur === '"' || cur === '\'') {
        // We've hit a string
        return token(TOKEN_STRING, this._parseString(cur), lineno, colno);
      } else if (tok = this._extract(whitespaceChars)) {
        // We hit some whitespace
        return token(TOKEN_WHITESPACE, tok, lineno, colno);
      } else if ((tok = this._extractString(this.tags.BLOCK_END)) || (tok = this._extractString('-' + this.tags.BLOCK_END))) {
        // Special check for the block end tag
        //
        // It is a requirement that start and end tags are composed of
        // delimiter characters (%{}[] etc), and our code always
        // breaks on delimiters so we can assume the token parsing
        // doesn't consume these elsewhere
        this.in_code = false;
        if (this.trimBlocks) {
          cur = this.current();
          if (cur === '\n') {
            // Skip newline
            this.forward();
          } else if (cur === '\r') {
            // Skip CRLF newline
            this.forward();
            cur = this.current();
            if (cur === '\n') {
              this.forward();
            } else {
              // Was not a CRLF, so go back
              this.back();
            }
          }
        }
        return token(TOKEN_BLOCK_END, tok, lineno, colno);
      } else if ((tok = this._extractString(this.tags.VARIABLE_END)) || (tok = this._extractString('-' + this.tags.VARIABLE_END))) {
        // Special check for variable end tag (see above)
        this.in_code = false;
        return token(TOKEN_VARIABLE_END, tok, lineno, colno);
      } else if (cur === 'r' && this.str.charAt(this.index + 1) === '/') {
        // Skip past 'r/'.
        this.forwardN(2);

        // Extract until the end of the regex -- / ends it, \/ does not.
        var regexBody = '';
        while (!this.isFinished()) {
          if (this.current() === '/' && this.previous() !== '\\') {
            this.forward();
            break;
          } else {
            regexBody += this.current();
            this.forward();
          }
        }

        // Check for flags.
        // The possible flags are according to https://developer.mozilla.org/en/docs/Web/JavaScript/Reference/Global_Objects/RegExp)
        var POSSIBLE_FLAGS = ['g', 'i', 'm', 'y'];
        var regexFlags = '';
        while (!this.isFinished()) {
          var isCurrentAFlag = POSSIBLE_FLAGS.indexOf(this.current()) !== -1;
          if (isCurrentAFlag) {
            regexFlags += this.current();
            this.forward();
          } else {
            break;
          }
        }
        return token(TOKEN_REGEX, {
          body: regexBody,
          flags: regexFlags
        }, lineno, colno);
      } else if (delimChars.indexOf(cur) !== -1) {
        // We've hit a delimiter (a special char like a bracket)
        this.forward();
        var complexOps = ['==', '===', '!=', '!==', '<=', '>=', '//', '**'];
        var curComplex = cur + this.current();
        var type;
        if (lib.indexOf(complexOps, curComplex) !== -1) {
          this.forward();
          cur = curComplex;

          // See if this is a strict equality/inequality comparator
          if (lib.indexOf(complexOps, curComplex + this.current()) !== -1) {
            cur = curComplex + this.current();
            this.forward();
          }
        }
        switch (cur) {
          case '(':
            type = TOKEN_LEFT_PAREN;
            break;
          case ')':
            type = TOKEN_RIGHT_PAREN;
            break;
          case '[':
            type = TOKEN_LEFT_BRACKET;
            break;
          case ']':
            type = TOKEN_RIGHT_BRACKET;
            break;
          case '{':
            type = TOKEN_LEFT_CURLY;
            break;
          case '}':
            type = TOKEN_RIGHT_CURLY;
            break;
          case ',':
            type = TOKEN_COMMA;
            break;
          case ':':
            type = TOKEN_COLON;
            break;
          case '~':
            type = TOKEN_TILDE;
            break;
          case '|':
            type = TOKEN_PIPE;
            break;
          default:
            type = TOKEN_OPERATOR;
        }
        return token(type, cur, lineno, colno);
      } else {
        // We are not at whitespace or a delimiter, so extract the
        // text and parse it
        tok = this._extractUntil(whitespaceChars + delimChars);
        if (tok.match(/^[-+]?[0-9]+$/)) {
          if (this.current() === '.') {
            this.forward();
            var dec = this._extract(intChars);
            return token(TOKEN_FLOAT, tok + '.' + dec, lineno, colno);
          } else {
            return token(TOKEN_INT, tok, lineno, colno);
          }
        } else if (tok.match(/^(true|false)$/)) {
          return token(TOKEN_BOOLEAN, tok, lineno, colno);
        } else if (tok === 'none') {
          return token(TOKEN_NONE, tok, lineno, colno);
          /*
           * Added to make the test `null is null` evaluate truthily.
           * Otherwise, Nunjucks will look up null in the context and
           * return `undefined`, which is not what we want. This *may* have
           * consequences is someone is using null in their templates as a
           * variable.
           */
        } else if (tok === 'null') {
          return token(TOKEN_NONE, tok, lineno, colno);
        } else if (tok) {
          return token(TOKEN_SYMBOL, tok, lineno, colno);
        } else {
          throw new Error('Unexpected value while parsing: ' + tok);
        }
      }
    } else {
      // Parse out the template text, breaking on tag
      // delimiters because we need to look for block/variable start
      // tags (don't use the full delimChars for optimization)
      var beginChars = this.tags.BLOCK_START.charAt(0) + this.tags.VARIABLE_START.charAt(0) + this.tags.COMMENT_START.charAt(0) + this.tags.COMMENT_END.charAt(0);
      if (this.isFinished()) {
        return null;
      } else if ((tok = this._extractString(this.tags.BLOCK_START + '-')) || (tok = this._extractString(this.tags.BLOCK_START))) {
        this.in_code = true;
        return token(TOKEN_BLOCK_START, tok, lineno, colno);
      } else if ((tok = this._extractString(this.tags.VARIABLE_START + '-')) || (tok = this._extractString(this.tags.VARIABLE_START))) {
        this.in_code = true;
        return token(TOKEN_VARIABLE_START, tok, lineno, colno);
      } else {
        tok = '';
        var data;
        var inComment = false;
        if (this._matches(this.tags.COMMENT_START)) {
          inComment = true;
          tok = this._extractString(this.tags.COMMENT_START);
        }

        // Continually consume text, breaking on the tag delimiter
        // characters and checking to see if it's a start tag.
        //
        // We could hit the end of the template in the middle of
        // our looping, so check for the null return value from
        // _extractUntil
        while ((data = this._extractUntil(beginChars)) !== null) {
          tok += data;
          if ((this._matches(this.tags.BLOCK_START) || this._matches(this.tags.VARIABLE_START) || this._matches(this.tags.COMMENT_START)) && !inComment) {
            if (this.lstripBlocks && this._matches(this.tags.BLOCK_START) && this.colno > 0 && this.colno <= tok.length) {
              var lastLine = tok.slice(-this.colno);
              if (/^\s+$/.test(lastLine)) {
                // Remove block leading whitespace from beginning of the string
                tok = tok.slice(0, -this.colno);
                if (!tok.length) {
                  // All data removed, collapse to avoid unnecessary nodes
                  // by returning next token (block start)
                  return this.nextToken();
                }
              }
            }
            // If it is a start tag, stop looping
            break;
          } else if (this._matches(this.tags.COMMENT_END)) {
            if (!inComment) {
              throw new Error('unexpected end of comment');
            }
            tok += this._extractString(this.tags.COMMENT_END);
            break;
          } else {
            // It does not match any tag, so add the character and
            // carry on
            tok += this.current();
            this.forward();
          }
        }
        if (data === null && inComment) {
          throw new Error('expected end of comment, got end of file');
        }
        return token(inComment ? TOKEN_COMMENT : TOKEN_DATA, tok, lineno, colno);
      }
    }
  };
  _proto._parseString = function _parseString(delimiter) {
    this.forward();
    var str = '';
    while (!this.isFinished() && this.current() !== delimiter) {
      var cur = this.current();
      if (cur === '\\') {
        this.forward();
        switch (this.current()) {
          case 'n':
            str += '\n';
            break;
          case 't':
            str += '\t';
            break;
          case 'r':
            str += '\r';
            break;
          default:
            str += this.current();
        }
        this.forward();
      } else {
        str += cur;
        this.forward();
      }
    }
    this.forward();
    return str;
  };
  _proto._matches = function _matches(str) {
    if (this.index + str.length > this.len) {
      return null;
    }
    var m = this.str.slice(this.index, this.index + str.length);
    return m === str;
  };
  _proto._extractString = function _extractString(str) {
    if (this._matches(str)) {
      this.forwardN(str.length);
      return str;
    }
    return null;
  };
  _proto._extractUntil = function _extractUntil(charString) {
    // Extract all non-matching chars, with the default matching set
    // to everything
    return this._extractMatching(true, charString || '');
  };
  _proto._extract = function _extract(charString) {
    // Extract all matching chars (no default, so charString must be
    // explicit)
    return this._extractMatching(false, charString);
  };
  _proto._extractMatching = function _extractMatching(breakOnMatch, charString) {
    // Pull out characters until a breaking char is hit.
    // If breakOnMatch is false, a non-matching char stops it.
    // If breakOnMatch is true, a matching char stops it.

    if (this.isFinished()) {
      return null;
    }
    var first = charString.indexOf(this.current());

    // Only proceed if the first character doesn't meet our condition
    if (breakOnMatch && first === -1 || !breakOnMatch && first !== -1) {
      var t = this.current();
      this.forward();

      // And pull out all the chars one at a time until we hit a
      // breaking char
      var idx = charString.indexOf(this.current());
      while ((breakOnMatch && idx === -1 || !breakOnMatch && idx !== -1) && !this.isFinished()) {
        t += this.current();
        this.forward();
        idx = charString.indexOf(this.current());
      }
      return t;
    }
    return '';
  };
  _proto._extractRegex = function _extractRegex(regex) {
    var matches = this.currentStr().match(regex);
    if (!matches) {
      return null;
    }

    // Move forward whatever was matched
    this.forwardN(matches[0].length);
    return matches;
  };
  _proto.isFinished = function isFinished() {
    return this.index >= this.len;
  };
  _proto.forwardN = function forwardN(n) {
    for (var i = 0; i < n; i++) {
      this.forward();
    }
  };
  _proto.forward = function forward() {
    this.index++;
    if (this.previous() === '\n') {
      this.lineno++;
      this.colno = 0;
    } else {
      this.colno++;
    }
  };
  _proto.backN = function backN(n) {
    for (var i = 0; i < n; i++) {
      this.back();
    }
  };
  _proto.back = function back() {
    this.index--;
    if (this.current() === '\n') {
      this.lineno--;
      var idx = this.src.lastIndexOf('\n', this.index - 1);
      if (idx === -1) {
        this.colno = this.index;
      } else {
        this.colno = this.index - idx;
      }
    } else {
      this.colno--;
    }
  }

  // current returns current character
  ;
  _proto.current = function current() {
    if (!this.isFinished()) {
      return this.str.charAt(this.index);
    }
    return '';
  }

  // currentStr returns what's left of the unparsed string
  ;
  _proto.currentStr = function currentStr() {
    if (!this.isFinished()) {
      return this.str.substr(this.index);
    }
    return '';
  };
  _proto.previous = function previous() {
    return this.str.charAt(this.index - 1);
  };
  return Tokenizer;
}();
module.exports = {
  lex: function lex(src, opts) {
    return new Tokenizer(src, opts);
  },
  TOKEN_STRING: TOKEN_STRING,
  TOKEN_WHITESPACE: TOKEN_WHITESPACE,
  TOKEN_DATA: TOKEN_DATA,
  TOKEN_BLOCK_START: TOKEN_BLOCK_START,
  TOKEN_BLOCK_END: TOKEN_BLOCK_END,
  TOKEN_VARIABLE_START: TOKEN_VARIABLE_START,
  TOKEN_VARIABLE_END: TOKEN_VARIABLE_END,
  TOKEN_COMMENT: TOKEN_COMMENT,
  TOKEN_LEFT_PAREN: TOKEN_LEFT_PAREN,
  TOKEN_RIGHT_PAREN: TOKEN_RIGHT_PAREN,
  TOKEN_LEFT_BRACKET: TOKEN_LEFT_BRACKET,
  TOKEN_RIGHT_BRACKET: TOKEN_RIGHT_BRACKET,
  TOKEN_LEFT_CURLY: TOKEN_LEFT_CURLY,
  TOKEN_RIGHT_CURLY: TOKEN_RIGHT_CURLY,
  TOKEN_OPERATOR: TOKEN_OPERATOR,
  TOKEN_COMMA: TOKEN_COMMA,
  TOKEN_COLON: TOKEN_COLON,
  TOKEN_TILDE: TOKEN_TILDE,
  TOKEN_PIPE: TOKEN_PIPE,
  TOKEN_INT: TOKEN_INT,
  TOKEN_FLOAT: TOKEN_FLOAT,
  TOKEN_BOOLEAN: TOKEN_BOOLEAN,
  TOKEN_NONE: TOKEN_NONE,
  TOKEN_SYMBOL: TOKEN_SYMBOL,
  TOKEN_SPECIAL: TOKEN_SPECIAL,
  TOKEN_REGEX: TOKEN_REGEX
};