ioutils.ReadAll()会清空对应的reader

最新推荐文章于 2024-09-08 08:57:54 发布

撒一片暖阳

最新推荐文章于 2024-09-08 08:57:54 发布

阅读量733

点赞数 1

分类专栏： golang

golang 专栏收录该内容

4 篇文章

订阅专栏

本文介绍在Golang中如何正确处理HTTP请求体的读取及重置问题，避免因重复读取而导致的数据丢失。通过示例展示了如何使用bytes.Buffer实现数据的“预览”功能。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

转载于:http://blog.youkuaiyun.com/pkueecser/article/details/50421562

解决方法

    reqBytes, err := ioutil.ReadAll(c.Request.Body)  
    if err != nil {  
        log.Error("fail to read requset data")  
        return 0, err  
    }  
    log.Infof("%s", reqBytes)  

    buf := bytes.NewBuffer(reqBytes)  
    c.Request.Body = ioutil.NopCloser(buf)

分析原因
http://stackoverflow.com/questions/30910487/why-an-io-reader-after-read-it-became-empty-in-golang

Reading from a bytes.Buffer drains or consumes the bytes that were read. This means if you try to read again those will not be returned.

Buffer.Bytes() returns the unread portion of the buffer so it is the expected result for you to see 0 length after everything has been read (this is exactly what ioutil.ReadAll() does).

What if you just want to "peek" and not really "read" bytes?

There is no "peek" functionality in bytes.Buffer. The easiest would be to get the bytes of the buffer, and construct another bytes.Buffer from it and read from the new buffer.

It could look something like this:

func peek(buf *bytes.Buffer, b []byte) (int, error) {
    buf2 := bytes.NewBuffer(buf.Bytes())
    return buf2.Read(b)
}