[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8},"actionData","errors","root",{"_11":-5},"routes/user.$name.replies",{"_9":10,"_11":12,"_13":14,"_15":16,"_17":18,"_19":16},"replies",[35],"user",{"_20":21,"_13":14,"_22":23,"_24":25,"_26":27,"_28":29,"_30":31,"_32":16,"_33":34},"loginname","peidehua","page",1,"limit",50,"total","id","2328","avatar_url","http://www.gravatar.com/avatar/3665068087993e18580ee70f0fb4cdd0?size=48","githubUsername","","create_at","2012-02-02 06:08:39.393","score",5,"topic_count",0,"reply_count","is_star",false,{"_20":36,"_37":38,"_39":-5,"_40":41,"_42":43,"_44":45,"_46":34,"_47":34,"_32":48,"_49":50,"_26":51,"_52":53},"21","author_id","19","tab","content","这篇文章由于从wordpress升级过来，格式完全不能看了，请访问这里查看吧，内容一样的。\r\n

\r\n\r\n\r\n\r\n\r\n最近遇见一个从前没有遇见的陷阱，就是data里的chunk拼接。 \r\n由于本人身为前端工程师，对buffer的概念实在是认识不足。这次的场景是我要通过http.get去抓取远端的网页文件，很不小心的是对方的文件编码是gbk（估计是老年代Java环境下的解决方案），而我本地的代码是utf8的编码，最终我需要将两部分代码合并之后输出到客户端，所以我需要将接受到的部分进行转码，转码则需要通过iconv实现。 \r\n在这之前我需要将接受到的chunk进行组装。下面是我最原始的组装方式，因为在我的概念中都把他们当做string给组装了。\r\n\r\n var data = \"\"; \r\n res.on('data', function (chunk) { \r\n data += chunk; \r\n }) \r\n .on(\"end\", function () { \r\n //对data转码 \r\n }); \r\n\r\n很遗憾，我调用:\r\n\r\n var iconv = new Iconv('GBK', 'UTF-8'); \r\n iconv.convert(data).toString(); \r\n\r\nEILSEQ异常被抛出。 \r\n其原因是两个chunk（Buffer对象）的拼接并不正常，相当于进行了buffer.toString() + buffer.toString()。如果buffer不是完整的，则toString出来后的string是存在问题的（比如一个中文字被截断）。这样出来的string就无法被iconv正常转码。 \r\n那么正确的拼接该是怎样呢，在大神兼好基友[@Python](/user/Python)发烧友的帮助指点下，以下代码才是正确的： \r\n\r\n var chunks = []; \r\n var size = 0; \r\n res.on('data', function (chunk) { \r\n chunks.push(chunk); \r\n size += chunk.length; \r\n }); \r\n res.on('end', function () { \r\n var data = null; \r\n switch(chunks.length) { \r\n case 0: data = new Buffer(0); \r\n break; \r\n case 1: data = chunks[0]; \r\n break; \r\n default: \r\n data = new Buffer(size); \r\n for (var i = 0, pos = 0, l = chunks.length; i < l; i++) { \r\n var chunk = chunks[i]; \r\n chunk.copy(data, pos); \r\n pos += chunk.length; \r\n } \r\n break; \r\n } \r\n }); \r\n\r\n这时候的data才是一个正确的buffer对象。\r\n\r\n但是，对于接收数据而言，这样的场景应当是一个十分常见的场景才对，每次都要写这样一大堆的代码，实在是很费事的。那么我们封装重构吧：\r\n\r\n var BufferHelper = function () {\r\n this.buffers = [];\r\n this.size = 0;\r\n this._status = \"changed\";\r\n };\r\n\r\n BufferHelper.prototype.concat = function (buffer) {\r\n for (var i = 0, l = arguments.length; i < l; i++) {\r\n this._concat(arguments[i]);\r\n }\r\n return this;\r\n };\r\n\r\n BufferHelper.prototype._concat = function (buffer) {\r\n this.buffers.push(buffer);\r\n this.size = this.size + buffer.length;\r\n this._status = \"changed\";\r\n return this;\r\n };\r\n\r\n BufferHelper.prototype._toBuffer = function () {\r\n var data = null;\r\n var buffers = this.buffers;\r\n switch(buffers.length) {\r\n case 0:\r\n data = new Buffer(0);\r\n break;\r\n case 1:\r\n data = buffers[0];\r\n break;\r\n default:\r\n data = new Buffer(this.size);\r\n for (var i = 0, pos = 0, l = buffers.length; i < l; i++) {\r\n var buffer = buffers[i];\r\n buffer.copy(data, pos);\r\n pos += buffer.length;\r\n }\r\n break;\r\n }\r\n // Cache the computed result\r\n this._status = \"computed\";\r\n this.buffer = data;\r\n return data;\r\n };\r\n\r\n BufferHelper.prototype.toBuffer = function () {\r\n return this._status === \"computed\" ? this.buffer : this._toBuffer();\r\n };\r\n\r\n BufferHelper.prototype.toString = function () {\r\n return Buffer.prototype.toString.apply(this.toBuffer(), arguments);\r\n };\r\n\r\n module.exports = BufferHelper;\r\n\r\n这里有两个私有方法，_concat和_toBuffer。其目的是保证每个方法的职责单一，还在toBuffer里做了一下状态设置，使得不浪费CPU。接下来的调用就非常之简单了。\r\n\r\n```\r\nvar http = require('http');\r\nvar BufferHelper = require('bufferhelper');\r\n\r\nhttp.createServer(function (request, response) {\r\n var bufferHelper = new BufferHelper();\r\n\r\n request.on(\"data\", function (chunk) {\r\n bufferHelper.concat(chunk);\r\n });\r\n\r\n request.on('end', function () {\r\n var html = bufferHelper.toBuffer().toString();\r\n response.writeHead(200);\r\n response.end(html);\r\n });\r\n}).listen(8001);\r\n```\r\n\r\n可以看到代码量少了很多，跟第一种方法的使用相差无几。嗯，不虚心的自夸一下，很干净利落:)。 \r\n另外可以看到上面的代码是直接require的：\r\n\r\n```\r\nvar BufferHelper = require('bufferhelper');\r\n```\r\n其原因是我已经将其发布到NPM中，可以通过npm install bufferhelper直接搞定。项目地址在github上：\r\n\r\n\r\n最后这个lib还没写单元测试，和做压测，之后会添加上。 \r\n最后谢谢基友[@Python](/user/Python)发烧友。 \r\n最后，其实node-iconv的作者还提供了一个工具集（），是有部分通过c/c++完成的，不过我的需求没那么复杂，只要一个最简单的concat就可以满足了。相信这个bufferhelper对于中文环境下的同学是非常有用的~ ","title","小心data事件里的chunk拼接","last_reply_at","2013-10-21 11:54:48.673","good","top",14,"visit_count",24719,"2012-01-14 16:21:37","author",{"_13":54,"_22":55},"JacksonTian","https://avatars.githubusercontent.com/u/327019?v=4"]