picomatch-browser/lib/scan.js

(temporary fork of picomatch) Blazing fast and accurate glob matcher written in JavaScript, with no dependencies and full support for standard and extended Bash glob features, including braces, extglobs, POSIX brackets, and regular expressions.

'use strict';

const utils = require('./utils');
const {
  CHAR_ASTERISK,             /* * */
  CHAR_AT,                   /* @ */
  CHAR_BACKWARD_SLASH,       /* \ */
  CHAR_COMMA,                /* , */
  CHAR_DOT,                  /* . */
  CHAR_EXCLAMATION_MARK,     /* ! */
  CHAR_FORWARD_SLASH,        /* / */
  CHAR_LEFT_CURLY_BRACE,     /* { */
  CHAR_LEFT_PARENTHESES,     /* ( */
  CHAR_LEFT_SQUARE_BRACKET,  /* [ */
  CHAR_PLUS,                 /* + */
  CHAR_QUESTION_MARK,        /* ? */
  CHAR_RIGHT_CURLY_BRACE,    /* } */
  CHAR_RIGHT_PARENTHESES,    /* ) */
  CHAR_RIGHT_SQUARE_BRACKET  /* ] */
} = require('./constants');

const isPathSeparator = code => {
  return code === CHAR_FORWARD_SLASH || code === CHAR_BACKWARD_SLASH;
};

const depth = token => {
  if (token.isPrefix !== true) {
    token.depth = token.isGlobstar ? Infinity : 1;
  }
};

/**
 * Quickly scans a glob pattern and returns an object with a handful of
 * useful properties, like `isGlob`, `path` (the leading non-glob, if it exists),
 * `glob` (the actual pattern), and `negated` (true if the path starts with `!`).
 *
 * ```js
 * const pm = require('picomatch');
 * console.log(pm.scan('foo/bar/*.js'));
 * { isGlob: true, input: 'foo/bar/*.js', base: 'foo/bar', glob: '*.js' }
 * ```
 * @param {String} `str`
 * @param {Object} `options`
 * @return {Object} Returns an object with tokens and regex source string.
 * @api public
 */

const scan = (input, options) => {
  const opts = options || {};

  const length = input.length - 1;
  const scanToEnd = opts.parts === true || opts.scanToEnd === true;
  const slashes = [];
  const tokens = [];
  const parts = [];

  let str = input;
  let index = -1;
  let start = 0;
  let lastIndex = 0;
  let isBrace = false;
  let isBracket = false;
  let isGlob = false;
  let isExtglob = false;
  let isGlobstar = false;
  let braceEscaped = false;
  let backslashes = false;
  let negated = false;
  let finished = false;
  let braces = 0;
  let prev;
  let code;
  let token = { value: '', depth: 0, isGlob: false };

  const eos = () => index >= length;
  const peek = () => str.charCodeAt(index + 1);
  const advance = () => {
    prev = code;
    return str.charCodeAt(++index);
  };

  while (index < length) {
    code = advance();
    let next;

    if (code === CHAR_BACKWARD_SLASH) {
      backslashes = token.backslashes = true;
      code = advance();

      if (code === CHAR_LEFT_CURLY_BRACE) {
        braceEscaped = true;
      }
      continue;
    }

    if (braceEscaped === true || code === CHAR_LEFT_CURLY_BRACE) {
      braces++;

      while (eos() !== true && (code = advance())) {
        if (code === CHAR_BACKWARD_SLASH) {
          backslashes = token.backslashes = true;
          advance();
          continue;
        }

        if (code === CHAR_LEFT_CURLY_BRACE) {
          braces++;
          continue;
        }

        if (braceEscaped !== true && code === CHAR_DOT && (code = advance()) === CHAR_DOT) {
          isBrace = token.isBrace = true;
          isGlob = token.isGlob = true;
          finished = true;

          if (scanToEnd === true) {
            continue;
          }

          break;
        }

        if (braceEscaped !== true && code === CHAR_COMMA) {
          isBrace = token.isBrace = true;
          isGlob = token.isGlob = true;
          finished = true;

          if (scanToEnd === true) {
            continue;
          }

          break;
        }

        if (code === CHAR_RIGHT_CURLY_BRACE) {
          braces--;

          if (braces === 0) {
            braceEscaped = false;
            isBrace = token.isBrace = true;
            finished = true;
            break;
          }
        }
      }

      if (scanToEnd === true) {
        continue;
      }

      break;
    }

    if (code === CHAR_FORWARD_SLASH) {
      slashes.push(index);
      tokens.push(token);
      token = { value: '', depth: 0, isGlob: false };

      if (finished === true) continue;
      if (prev === CHAR_DOT && index === (start + 1)) {
        start += 2;
        continue;
      }

      lastIndex = index + 1;
      continue;
    }

    if (opts.noext !== true) {
      const isExtglobChar = code === CHAR_PLUS
        || code === CHAR_AT
        || code === CHAR_ASTERISK
        || code === CHAR_QUESTION_MARK
        || code === CHAR_EXCLAMATION_MARK;

      if (isExtglobChar === true && peek() === CHAR_LEFT_PARENTHESES) {
        isGlob = token.isGlob = true;
        isExtglob = token.isExtglob = true;
        finished = true;

        if (scanToEnd === true) {
          while (eos() !== true && (code = advance())) {
            if (code === CHAR_BACKWARD_SLASH) {
              backslashes = token.backslashes = true;
              code = advance();
              continue;
            }

            if (code === CHAR_RIGHT_PARENTHESES) {
              isGlob = token.isGlob = true;
              finished = true;
              break;
            }
          }
          continue;
        }
        break;
      }
    }

    if (code === CHAR_ASTERISK) {
      if (prev === CHAR_ASTERISK) isGlobstar = token.isGlobstar = true;
      isGlob = token.isGlob = true;
      finished = true;

      if (scanToEnd === true) {
        continue;
      }
      break;
    }

    if (code === CHAR_QUESTION_MARK) {
      isGlob = token.isGlob = true;
      finished = true;

      if (scanToEnd === true) {
        continue;
      }
      break;
    }

    if (code === CHAR_LEFT_SQUARE_BRACKET) {
      while (eos() !== true && (next = advance())) {
        if (next === CHAR_BACKWARD_SLASH) {
          backslashes = token.backslashes = true;
          advance();
          continue;
        }

        if (next === CHAR_RIGHT_SQUARE_BRACKET) {
          isBracket = token.isBracket = true;
          isGlob = token.isGlob = true;
          finished = true;

          if (scanToEnd === true) {
            continue;
          }
          break;
        }
      }
    }

    if (opts.nonegate !== true && code === CHAR_EXCLAMATION_MARK && index === start) {
      negated = token.negated = true;
      start++;
      continue;
    }

    if (opts.noparen !== true && code === CHAR_LEFT_PARENTHESES) {
      isGlob = token.isGlob = true;

      if (scanToEnd === true) {
        while (eos() !== true && (code = advance())) {
          if (code === CHAR_LEFT_PARENTHESES) {
            backslashes = token.backslashes = true;
            code = advance();
            continue;
          }

          if (code === CHAR_RIGHT_PARENTHESES) {
            finished = true;
            break;
          }
        }
        continue;
      }
      break;
    }

    if (isGlob === true) {
      finished = true;

      if (scanToEnd === true) {
        continue;
      }

      break;
    }
  }

  if (opts.noext === true) {
    isExtglob = false;
    isGlob = false;
  }

  let base = str;
  let prefix = '';
  let glob = '';

  if (start > 0) {
    prefix = str.slice(0, start);
    str = str.slice(start);
    lastIndex -= start;
  }

  if (base && isGlob === true && lastIndex > 0) {
    base = str.slice(0, lastIndex);
    glob = str.slice(lastIndex);
  } else if (isGlob === true) {
    base = '';
    glob = str;
  } else {
    base = str;
  }

  if (base && base !== '' && base !== '/' && base !== str) {
    if (isPathSeparator(base.charCodeAt(base.length - 1))) {
      base = base.slice(0, -1);
    }
  }

  if (opts.unescape === true) {
    if (glob) glob = utils.removeBackslashes(glob);

    if (base && backslashes === true) {
      base = utils.removeBackslashes(base);
    }
  }

  const state = {
    prefix,
    input,
    start,
    base,
    glob,
    isBrace,
    isBracket,
    isGlob,
    isExtglob,
    isGlobstar,
    negated
  };

  if (opts.tokens === true) {
    state.maxDepth = 0;
    if (!isPathSeparator(code)) {
      tokens.push(token);
    }
    state.tokens = tokens;
  }

  if (opts.parts === true || opts.tokens === true) {
    let prevIndex;

    for (let idx = 0; idx < slashes.length; idx++) {
      const n = prevIndex ? prevIndex + 1 : start;
      const i = slashes[idx];
      const value = input.slice(n, i);
      if (opts.tokens) {
        if (idx === 0 && start !== 0) {
          tokens[idx].isPrefix = true;
          tokens[idx].value = prefix;
        } else {
          tokens[idx].value = value;
        }
        depth(tokens[idx]);
        state.maxDepth += tokens[idx].depth;
      }
      if (idx !== 0 || value !== '') {
        parts.push(value);
      }
      prevIndex = i;
    }

    if (prevIndex && prevIndex + 1 < input.length) {
      const value = input.slice(prevIndex + 1);
      parts.push(value);

      if (opts.tokens) {
        tokens[tokens.length - 1].value = value;
        depth(tokens[tokens.length - 1]);
        state.maxDepth += tokens[tokens.length - 1].depth;
      }
    }

    state.slashes = slashes;
    state.parts = parts;
  }

  return state;
};

module.exports = scan;