rfc3986.js : RFC3986 に準拠した URI チェッカー

rfc3986.js

/**
 * rfc3986.js
 * Uniform Resource Identifier (URI): Generic Syntax (RFC3986)
 *
 * @version 1.0
 * @author think49
 * @url https://gist.github.com/770087
 * @see <a href="http://www.ietf.org/rfc/rfc3986.txt">www.ietf.org/rfc/rfc3986.txt</a>
 * @see <a href="http://www.studyinghttp.net/rfc_ja/rfc3986">Uniform Resource Identifier (URI): &#x4e00;&#x822c;&#x7684;&#x69cb;&#x6587;</a>
 * @license http://www.opensource.org/licenses/mit-license.php (The MIT License)
 */

'use strict';
function RFC3986 () {

  var

      // === RFC2234

      // 6.1  Core Rules
      ALPHA  = '[\u0041-\u005A\u0061-\u007A]', // [A-Za-z]
      DIGIT  = '[\u0030-\u0039]',              // [0-9]
      HEXDIG = '(?:' + DIGIT + '|[A-Z])',


      // === RFC3986

      // 2.1.  Percent-Encoding
      pct_encoded = '%' + HEXDIG + HEXDIG,

      // 2.2.  Reserved Characters
      gen_delims  = '[:/?#[\\]@]',
      sub_delims  = '[!$&\u0027()*+,;=]',
      reserved    = '(?:' + gen_delims + '|' + sub_delims + ')',

      // 2.3.  Unreserved Characters
      unreserved  = '(?:' + ALPHA + '|' + DIGIT + '|[\\-._~])',

      // 3.1.  Scheme
      scheme      = ALPHA + '(?:' + ALPHA + '|' + DIGIT + '|[+\\-.])*',

      // 3.2.1.  User Information
      userinfo    = '(?:' + unreserved + '|' + pct_encoded + '|' + sub_delims + '|:)*',

      // 3.2.2.  Host
      dec_octet   = '(?:' + DIGIT + '|[\u0031-\u0039]' + DIGIT + '|1(?:' + DIGIT + '){2}|2[\u0030-\u0034]' + DIGIT + '|25[\u0030-\u0035])',
      IPv4address = dec_octet + '\\.' + dec_octet + '\\.' + dec_octet + '\\.' + dec_octet,
      h16         = '(?:' + HEXDIG + '){1,4}',
      ls32        = '(?:' + h16 + ':' + h16 + '|' + IPv4address + ')',
      IPv6address = ['(?:' + h16 + ':){6}' + ls32,
                     '::' +  '(?:' + h16 + ':){5}' + ls32,
                     '(?:' + h16 + ')?::' +  '(?:' + h16 + ':){4}' + ls32,
                     '(?:(?:' + h16 + ':)+' + h16 + ')?::' + '(?:' + h16 + ':){3}' + ls32,
                     '(?:(?:' + h16 + ':){0,2}' + h16 + ')?::' + '(?:' + h16 + ':){2}' + ls32,
                     '(?:(?:' + h16 + ':){0,3}' + h16 + ')?::' + h16 + ':' + ls32,
                     '(?:(?:' + h16 + ':){0,4}' + h16 + ')?::' + ls32,
                     '(?:(?:' + h16 + ':){0,5}' + h16 + ')?::' + h16,
                     '(?:(?:' + h16 + ':){0,6}' + h16 + ')?::' + h16
                    ].join('|'),
      IPvFuture   = 'v(?:' + HEXDIG + ')+\\.(?:' + unreserved + '|' + sub_delims + '|:)+',
      IP_literal  = '(?:' + IPv6address + '|' + IPvFuture + ')?',
      reg_name    = '(?:' + unreserved + '|' + pct_encoded + '|' + sub_delims + ')*',
      host        = '(?:' + IP_literal + '|' + IPv4address + '|' + reg_name + ')',

      // 3.2.3.  Port
      port        = '(?:' + DIGIT + ')*',

      // 3.2.  Authority
      authority   = '(?:' + userinfo + '@)?' + host + '(?::' + port + ')?',

      // 3.3.  Path
      pchar         = '(?:' + unreserved + '|' + pct_encoded + '|' + sub_delims + '|[:@])',
      segment       = '(?:' + pchar + ')*',
      segment_nz    = '(?:' + pchar + ')+',
      segment_nz_nc = '(?:' + unreserved + '|' + pct_encoded + '|' + sub_delims + '|@)+',
      path_abempty  = '(?:/' + segment + ')*',
      path_absolute = '/(?:' + segment_nz + '(?:/' + segment + ')*)?',
      path_noscheme = segment_nz_nc + '(?:/' + segment + ')*',
      path_rootless = segment_nz + '(?:/' + segment + ')*',
      path_empty    = '(?:' + pchar + '){0}',
      path          = '(?:' + path_abempty + '|' + path_absolute + '|' + path_noscheme + '|' + path_rootless + '|' + path_empty + ')',

      // 3.4.  Query
      query       = '(?:' + pchar + '|[/?])*',

      // 3.5.  Fragment
      fragment    = '(?:' + pchar + '|[/?])*',

      // 3.  Syntax Components
      hier_part   = '(?://' + authority + path_abempty + '|' + path_absolute + '|' + path_rootless + '|' + path_empty + ')',
      URI         = scheme + ':' + hier_part + '(?:\\?' + query + ')?(?:#' + fragment + ')?',

      // 4.2.  Relative Reference
      relative_part = '(?://' + authority + path_abempty + '|' + path_absolute + '|' + path_noscheme + '|' + path_empty + ')',
      relative_ref  = relative_part + '(?:\\?' + query + ')?(?:#' + fragment + ')?',

      // 4.3.  Absolute URI
      absolute_URI  = scheme + ':' + hier_part + '(?:\\?' + query + ')?',

      // 4.1.  URI Reference
      URI_reference = '(?:' + URI + '|' + relative_ref + ')';


      this.absolute_URI   = absolute_URI;
      this.relative_ref   = relative_ref;
      this.URI           = URI;
      this.URI_reference = URI_reference;
}

RFC3986.prototype.isAbsoluteURI = function (arg) {
  if ('string' !== typeof arg) throw new TypeError(arg + ' is not a string');
  return new RegExp('^' + this.absolute_URI + '$').test(arg);
};

RFC3986.prototype.isRelativeRef = function (arg) {
  if ('string' !== typeof arg) throw new TypeError(arg + ' is not a string');
  return new RegExp('^' + this.relative_ref + '$').test(arg);
};

RFC3986.prototype.isURI = function (arg) {
  if ('string' !== typeof arg) throw new TypeError(arg + ' is not a string');
  return new RegExp('^' + this.URI + '$').test(arg);
};

RFC3986.prototype.isURIReference = function (arg) {
  if ('string' !== typeof arg) throw new TypeError(arg + ' is not a string');
  return new RegExp('^' + this.URI_reference + '$').test(arg);
};

参考リンク。

• rfc3986.js