/*!
 * encodeurl
 * Copyright(c) 2016 Douglas Christopher Wilson
 * MIT Licensed
 */

'use strict'

/**
 * Module exports.
 * @public
 */

module.exports = encodeUrl

/**
 * RegExp to match non-URL code points, *after* encoding (i.e. not including "%")
 * and including invalid escape sequences.
 * Group up to 1024 character and percent runs per replacement to reduce calls
 * to encodeURI while bounding RegExp stack use. Character runs stay intact so
 * surrogate pairs are never split between replacements.
 * @private
 */

var ENCODE_CHARS_REGEXP = /(?:[^\x21\x23-\x3B\x3D\x3F-\x5F\x61-\x7A\x7C\x7E]+|%+(?![0-9A-Fa-f]{2})){1,1024}/g

/**
 * RegExp to match surrogate pairs and unpaired surrogates.
 * @private
 */

var SURROGATE_REGEXP = /([\uD800-\uDBFF][\uDC00-\uDFFF])|[\uD800-\uDFFF]/g

/**
 * Encode a URL to a percent-encoded form, excluding already-encoded sequences.
 *
 * This function will take an already-encoded URL and encode all the non-URL
 * code points. This function will not encode the "%" character unless it is
 * not part of a valid sequence (`%20` will be left as-is, but `%foo` will
 * be encoded as `%25foo`).
 *
 * This encode is meant to be "safe" and does not throw errors. It will try as
 * hard as it can to properly encode the given URL, including replacing any raw,
 * unpaired surrogate pairs with the Unicode replacement character prior to
 * encoding.
 *
 * @param {string} url
 * @return {string}
 * @public
 */

function encodeUrl (url) {
  return String(url)
    .replace(SURROGATE_REGEXP, function (_, pair) {
      return pair || '\uFFFD'
    })
    .replace(ENCODE_CHARS_REGEXP, encodeURI)
}