浏览器上传大文件的尴尬

想上传超过1G的文件?

 


先来解读浏览器上传大文件的实际表现

 

1. IE 上传 1G 文件时,得到的Conent-Length 是不准确,上传3G,Content-Length 变为负数了。 囧

2. FireFox 添加后,点击提交没有反应。

 

 

再来看看标准协议  rfc1867.txt    里面有几点是关键内容      

 

1. 上传的表单编码 必须是 multipart/form-data

2. 上传多个文件时格式有2种,一种是

 

 Content-type: multipart/form-data, boundary=AaB03x

 --AaB03x
 content-disposition: form-data; name="field1"

 Joe Blow
--AaB03x
  content-disposition: form-data; name="pics"; filename="file1.txt"
        Content-Type: text/plain

... contents of file1.txt ...
--AaB03x--

 

另一种是

Content-type: multipart/form-data, boundary=AaB03x

        --AaB03x
        content-disposition: form-data; name="field1"

        Joe Blow
        --AaB03x
        content-disposition: form-data; name="pics"
        Content-type: multipart/mixed, boundary=BbC04y

        --BbC04y
        Content-disposition: attachment; filename="file1.txt"

 

IE 采用的是第一种,

 

 

 Java 实现注意的地方

 

1. Common-FileUpload  
是个专门用来处理上传的公共包,里面有个 MultipartStream  可以直接用来解析协议内容。

2. 注意的是,需要要解析大文件流,不能设置Conent-Length,将此值设置为 -1 即可。如

3. 得到上传的文件名如果乱码,是因为设置的编码不对,需要设置为GBK编码。

最后,贴一下 用 JBoss Netty 实现的上传

 

import java.io.IOException; import java.io.InputStream; import java.util.List; import org.apache.commons.fileupload.FileItem; import org.apache.commons.fileupload.FileItemFactory; import org.apache.commons.fileupload.FileItemIterator; import org.apache.commons.fileupload.FileUpload; import org.apache.commons.fileupload.FileUploadException; import org.jboss.netty.handler.codec.http.HttpMethod; import org.jboss.netty.handler.codec.http.HttpRequest; /** * 用Netty来实现上传 */ public class NettyUploader extends FileUpload { public static final boolean isMultipartContent(HttpRequest request) { if (HttpMethod.POST != request.getMethod()) { return false; } if (request.getHeaders("Content-Type") == null && request.getHeaders("Content-Type").size() == 0) { return false; } String contentType = request.getHeaders("Content-Type").get(0); if (contentType == null) { return false; } if (contentType.toLowerCase().startsWith("multipart/")) { return true; } return false; } /** * 上传的内容流 */ private InputStream inputStream; public NettyUploader(InputStream inputStream) { this.inputStream = inputStream; } public NettyUploader(FileItemFactory fileItemFactory) { super(fileItemFactory); } @SuppressWarnings("unchecked") public List<FileItem> parseRequest(String encoding, String contentType, int contentLength) throws FileUploadException { return parseRequest(new NettyRequestContext(encoding, contentType, contentLength, inputStream)); } public FileItemIterator getItemIterator(String encoding, String contentType, int contentLength) throws FileUploadException, IOException { this.setHeaderEncoding(encoding == null ? "GBK" : encoding); return super.getItemIterator(new NettyRequestContext(encoding, contentType, contentLength, inputStream)); } }

 

import java.io.IOException; import java.io.InputStream; import org.apache.commons.fileupload.RequestContext; /** * 实现FileUploader */ public class NettyRequestContext implements RequestContext { private String encoding; private String contentType; private long contentLength = -1; /** * 上传的内容流 */ private InputStream inputStream; public NettyRequestContext(String encoding, String contentType, long contentLength, InputStream inputStream) { this.encoding = encoding; this.contentType = contentType; this.contentLength = contentLength; this.inputStream = inputStream; } public String getCharacterEncoding() { return encoding; // return request.getHeader("Character-Encoding"); } public String getContentType() { return contentType; // return request.getHeader("Content-Type"); } public int getContentLength() { return (int) contentLength; // return (int) request.getContentLength(); } public InputStream getInputStream() throws IOException { // return new ChannelBufferInputStream(request.getContent()); // 不能直接用request的流,因为有HttpChunk return inputStream; } public String toString() { return "ContentLength=" + this.getContentLength() + ", ContentType=" + this.getContentType(); } }

 

NettyUploader uploader = new NettyUploader(uploadStream); // 不能读取头,因为大文件,IE的都不是正确的 FileItemIterator iterator = uploader.getItemIterator(encoding, contentType, -1); int counts = 0; while (iterator.hasNext()) { .............. }

 

你可能感兴趣的:(String,浏览器,null,iterator,import,encoding)