Javascript 通过URL编码传递正则表达式

我正在尝试使用Angular JS客户端和Nancy FX服务器端将正则表达式从客户端传递到服务器。我对“+”字符有一个特殊的问题。我当前将“+”显式编码为“%2B”客户端：

    this.getMatchingPoints = function (stringMatch) {
        console.log("Application Points Service: getPointsMatching('" + stringMatch + "')");
        /* we have a particular problem with the character '+' in regular expressions, since
         * encodeURIComponent ignores it, and urldecoders treat it as a space. So it must be
         * manually url encoded after passing the rest of the string to encodeURIcomponent */
        var encodedPattern = encodeURIComponent(stringMatch).replace('+', '%2B');
        return $http.get('/Json/PointsMatching/' + encodedPattern);
    };

…但它仍然作为空间服务器端接收

从某种意义上说，这并不重要，因为正则表达式中的任何实例，例如，“[a-Z]+”都可以替换为“[a-Z][a-Z]*”；或者，我也可以用自己的专用解码器服务器端为“+”客户端编写自己的专用编码器

---

但是我想知道以前是否有人遇到过这个问题，如果是的话，他们是如何解决的？

更多的调查表明，这是Nancy FX中一个深层次的、难以解决的错误，开发者都知道；看

我的解决方法是，假设在我工作的特定上下文中，空格不是有效字符，用pluses服务器端替换所有空格：

        /* <summary>
         *  Return a JSON formatted document comprising a list of all those application points whose 
         *  names match the regular expression.
         *  </summary>
         *  <remarks>
         *  <para>This entry point is expected to be invoked with the pattern as the final element in the path
         *  (i.e. <code>/Json/PointsMatching/pattern_here</code>). This variant cannot cope with either slashes
         *  or backslashes in the pattern, even if they're URL encoded.</para>
         *  </remarks>
         *  <param name="pattern">A regular expression.</param>
         *  <returns>A JSON formatted document comprising a list of all those application points whose 
         *  names match the regular expression.</returns>
         */
        this.Get["/PointsMatching/{pattern}"] = _ =>
        {
            string pattern = (string)_.pattern;
            /* THIS IS A HACK! The URL decoder used by Nancy decodes both '+' and '%2B' as ' ', which
             * rather defeats the point of url encoding. However, we don't normally use spaces in
             * application point names (although it would be valid to do so), so it's fairly save to
             * replace all spaces in patterns with pluses. */
            pattern = pattern.Replace(" ", "+");

            return this.WithValidUser<string>(pattern, InterestRegistry.Instance.GetPointsMatching);
        };

---

这显然很难看，不是一个通用的解决方案，但它目前还有效。

好吧，我诽谤了encodeURIComponent的作者；它确实正确地将加号（“+”）替换为“%2B”。问题出在世界的南希一边。这很奇怪，因为C#HttpUtility.UrlDecode库函数正确地将“%2B”解码为“+”。您是否尝试过使用OWIN而不是旧的Nancy自宿主进行自宿主？据我记忆所及，OWIN way不应该遭受双重编码。