multipart原理简介

multipart的流是服务器临时文件流

multipart的流不是HTTPRequest的流，而是服务器的临时文件流

multipart/form-data是表单上传，接口可以从Mutipartfile对象获取输入流getInputStream，表面上看没产生临时文件
org.springframework.web.multipart.support.StandardMultipartHttpServletRequest的getInputStream

		@Overridepublic InputStream getInputStream() throws IOException {return this.part.getInputStream();}

这里面的流是从DiskFileItem（一眼磁盘）拿的，还告诉了你临时目录位置location

multipartfile生成临时文件到默认文件夹

从原理上看，在tomcat源码的org.apache.catalina.connector.Request类中，parseParts方法就将multipart格式的请求中文件生成到临时文件夹

			File location;String locationStr = mce.getLocation();if (locationStr == null || locationStr.length() == 0) {location = ((File) context.getServletContext().getAttribute(ServletContext.TEMPDIR));} else {// If relative, it is relative to TEMPDIRlocation = new File(locationStr);if (!location.isAbsolute()) {location = new File((File) context.getServletContext().getAttribute(ServletContext.TEMPDIR),locationStr).getAbsoluteFile();}}

临时文件的目录可配

这个location就是javax.servlet.MultipartConfigElement的配置，可以通过spring.servlet.multipart系列配置的location指定临时文件目录，这个临时文件用完，也会被自动清除

    private final String location;// = "";private final long maxFileSize;// = -1;private final long maxRequestSize;// = -1;private final int fileSizeThreshold;// = 0;

Mutipartfile的文件默认来自临时文件

Mutipartfile的注释中说明了，文件是stored in memory or temporarily on disk，存在内存或磁盘临时文件的，所以上传大文件时，会发现前端进度条走完了，代码才走进自己写得接口代码

/*** A representation of an uploaded file received in a multipart request.** <p>The file contents are either stored in memory or temporarily on disk.* In either case, the user is responsible for copying file contents to a* session-level or persistent store as and if desired. The temporary storage* will be cleared at the end of request processing.** @author Juergen Hoeller* @author Trevor D. Cook* @since 29.09.2003* @see org.springframework.web.multipart.MultipartHttpServletRequest* @see org.springframework.web.multipart.MultipartResolver*/
public interface MultipartFile extends InputStreamSource {

Mutipartfile产生临时文件的好处坏处

好处

帮开发封装了上传逻辑
可以在此基础实现切片上传到本地最后合并

坏处

对我来说
文件转储时间翻倍：我的文件存储服务器不是接收请求的服务所在服务器，我还得把文件转存到文件服务器，而服务到文件服务器还得再传一遍，时间基本double
前端进度不真实：前端显示的上传进度仅是浏览器给到服务的进度
占用资源：临时文件多少占点服务器磁盘空间

纯流式上传application/octet-stream

一段半伪代码，application/octet-stream参数体只能有文件，参数通过放headers里给就可以解决这个问题

    @PostMapping(value = "/upload", headers = "content-type=application/octet-stream;charset=utf-8")public Object upload(HttpServletRequest request,@RequestHeader("userId") String userId,@RequestHeader("fileName") String fileName,@RequestHeader("fileSize") String fileSize,@RequestHeader("dirId") String dirId) {try (InputStream in = request.getInputStream()) {fileName = URLDecoder.decode(fileName,"utf-8");uploadFileServer(in, fileName, fileSize, userId, dirId);return "success";} catch (Exception e) {log.error("上传报错", e);return "error";}}