java - 为什么当我使用 HttpClient 时，分块流会意外结束？

Question

该问题似乎只发生在我尝试发布的“大”文件中。

我的代码如下所示:

PostMethod method = new PostMethod(url);

File input = new File(filePathname);
RequestEntity entity = new FileRequestEntity(input, "text/xml; charset=ISO-8859-1");

method.setRequestEntity(entity);
method.setRequestHeader("Content-Disposition", "attachment; filename=xyzzy")

HttpClient client = new HttpClient();
Credentials defaultcreds = new UsernamePasswordCredentials("userid", "pw");

client.getState().setCredentials(new AuthScope("hostname", port, AuthScope.ANY_REALM), defaultcreds);

 try {
    int statusCode = client.executeMethod(method);

    if (statusCode != HttpStatus.SC_OK) {
        throw new Exception("Method failed: " + method.getStatusLine());
    }

    // Read the response body.
    byte[] responseBody = method.getResponseBody();
    return new String(responseBody);
 }
 catch (HttpException e) {
    System.err.println("Fatal protocol violation: " + e.getMessage());
    throw e;

 }
 catch (IOException e) {
    System.err.println("Fatal transport error: " + e.getMessage());
    throw e;

 }
 finally {
     // Release the connection.
     method.releaseConnection();
 }

异常文本如下所示:

Fatal transport error: chunked stream ended unexpectedly
Exception in thread "main" java.io.IOException: chunked stream ended unexpectedly
    at org.apache.commons.httpclient.ChunkedInputStream.getChunkSizeFromInputStream(ChunkedInputStream.java:252)
    at org.apache.commons.httpclient.ChunkedInputStream.nextChunk(ChunkedInputStream.java:221)
    at org.apache.commons.httpclient.ChunkedInputStream.read(ChunkedInputStream.java:176)
    at java.io.FilterInputStream.read(FilterInputStream.java:127)
    at org.apache.commons.httpclient.AutoCloseInputStream.read(AutoCloseInputStream.java:108)
    at java.io.FilterInputStream.read(FilterInputStream.java:101)
    at org.apache.commons.httpclient.AutoCloseInputStream.read(AutoCloseInputStream.java:127)
    at org.apache.commons.httpclient.HttpMethodBase.getResponseBody(HttpMethodBase.java:690)

无论我使用 getResponseBody() 还是 getResponseBodyAsStream()，我都会遇到类似的异常。

我不应该得到太多数据，但我发布了超过 200mb 的数据。

Answer 1

我可以通过更改 PostMethod 的 requestHeader 中指定的文件名值的长度来解决此问题。我一直在请求 header 中包含完整文件路径名的编码版本。通过反复试验，我发现我“发布”的文件的成功或失败似乎取决于它所在的文件夹。长文件夹文件路径名不起作用，而短文件夹文件路径名则不起作用，尽管具有相同的文件。因此，我从请求 header 中删除了路径名，只开始包含文件名，并且我不再看到问题。