在Android中使用HttpUrlConnection编码问题

时间:2014-02-19 09:31:01

标签: android encoding httpclient httpurlconnection

我想通过HTTP post从我的Android Mobile应用程序向服务器发送XML消息。

我尝试使用HttpUrlConnection,遵循以下步骤:

URL url = new URL(vURL);

HttpUrlConnection conn = (HttpURLConnection) url.openConnection();

conn.setDoInput(true);
conn.setDoOutput(true);

// Adding headers (code removed)

conn.setRequestProperty("Content-Type", "text/xml; charset=utf-16");
OutputStream out = new BufferedOutputStream(conn.getOutputStream());

// Adding XML message to the connection output stream 
// I have removed exception handling to improve readability for posting it here
out.write(pReq.getBytes()); // here pReq is the XML message in String
out.close();

conn.connect();

获得响应后,流读取部分就是以这种方式完成的:

BufferedReader in = null;
StringBuffer sb;
String result = null;

    try {

        InputStreamReader   isr = new InputStreamReader(is);
            // Just in case, I've also tried:
            // new InputStreamReader(is, "UTF-16");
            // new InputStreamReader(is, "UTF-16LE");
            // new InputStreamReader(is, "UTF-16BE");
            // new InputStreamReader(is, "UTF-8");

        in = new BufferedReader(isr);

        sb = new StringBuffer("");
        String line = "";

        while ((line = in.readLine()) != null)
            sb.append(line);

        in.close();

        result = sb.toString();

    } catch (Exception e) {

        e.printStackTrace();
    }

现在我得到的结果字符串是一些不可读的格式/编码。

当我使用HttpClient尝试相同的操作时,它可以正常工作。一旦我在HttpClient.execute调用之后获得HttpResponse,这是流式读取部分:

BufferedReader in = null;
    InputStream         is;
    StringBuffer sb;
    String decompbuff = null;

    try {

        is = pResponse.getEntity().getContent();
        InputStreamReader   isr = new InputStreamReader(is);
        in = new BufferedReader(isr);

        // Prepare the String buffer
        sb = new StringBuffer("");

        String line = "";

        while ((line = in.readLine()) != null)
            sb.append(line);

        in.close();

         // gZip decompression of response. Note: message was compressed before
         // posting it via HttpClient (Posting code is not mentioned here)
        decompbuff = Decompress(sb.toString());

    } catch (Exception e) {

        e.printStackTrace();
    }

    return decompbuff;

在理解问题时,我们对一些帮助表示赞赏。

1 个答案:

答案 0 :(得分:1)

一个(严重)问题可能是您忽略了输入和输出的编码。

<强>输入

conn.setRequestProperty("Content-Type", "text/xml; charset=utf-16");
OutputStream out = new BufferedOutputStream(conn.getOutputStream());

// Adding XML message to the connection output stream 
// I have removed exception handling to improve readability for posting it here
out.write(pReq.getBytes()); // <-- you use standard platform encoding
out.close();

更好:

out.write(pReq.getBytes("UTF-16"));

<强>输出

您可能忽略了压缩,最好看起来像这样(取自DavidWebb):

static InputStream wrapStream(String contentEncoding, InputStream inputStream) 
        throws IOException {
    if (contentEncoding == null || "identity".equalsIgnoreCase(contentEncoding)) {
        return inputStream;
    }
    if ("gzip".equalsIgnoreCase(contentEncoding)) {
        return new GZIPInputStream(inputStream);
    }
    if ("deflate".equalsIgnoreCase(contentEncoding)) {
        return new InflaterInputStream(inputStream, new Inflater(false), 512);
    }
    throw new RuntimeException("unsupported content-encoding: " + contentEncoding);
}

// ...

    InputStream is = wrapStream(conn.getContentEncoding(), is);
    InputStreamReader isr = new InputStreamReader(is, "UTF-16");

    in = new BufferedReader(isr);

    sb = new StringBuffer("");
    String line = "";

    while ((line = in.readLine()) != null)
        sb.append(line); // <-- you're swallowing linefeeds!

    in.close();
    result = sb.toString();

最好让XML-Parser直接使用您的InputStream。不要创建JAVA字符串,而是让解析器扫描字节。它会自动检测XML的编码。

通常可能仍存在问题,因为我们不知道您使用的是什么类型的UTF-16。可以是BigEndian或LittleEndian。这就是为什么我问,如果你真的需要UTF-16。如果您不需要使用某些亚洲语言,UTF-8应该更高效,更易于使用。

所以我给你的“解决方案”不能保证工作 - 你必须摆弄UTF-16 BE / LE,祝你好运和耐心。

另一个评论:在上面的示例中,您首先构造String,然后Decompress。这是错误的顺序。流被压缩(gzip,deflate)并且必须首先解压缩。然后你得到String。