停止SAX分析器后，为什么InputStream会继续下载文件？

Question

I'm working on an RSS feed parser in Android and I've implemented a SAX parser which works perfectly under most situations.

However, I've run into problems on a few of my test feeds. After I've parsed a specified number of feed items, I throw a SAXException to stop the parser, which AFAIK is the right way to do that. On most feeds, this stops the parsing and my catch block (see below) handles and logs the StopParsingException.

On SOME feeds, however, the parser stops parsing, but there is a long delay between the exception being thrown and my catch block being run, during which no parsing is done, but just enough time passes to download the entire file (which is what I suspect is happening).

Here's my setup and error handling code:

public boolean parse(){
        SAXParserFactory factory = SAXParserFactory.newInstance();
        try {
            SAXParser parser = factory.newSAXParser();
            URL u = new URL(mUrl);
            URLConnection UC = u.openConnection();
            UC.setConnectTimeout(CONNECT_TIMEOUT);
            UC.setReadTimeout(CONNECT_TIMEOUT);
            InputStreamReader r = new InputStreamReader(UC.getInputStream());
            parser.parse(new InputSource(r), this);     
        }catch(SAXException sax)
        {
            Exception ex = sax.getException();
            if(ex != null)
            {
                if(ex instanceof StopParsingException)
                {
                    //Feed was intentionally stopped (i.e. reached episode limit)
                    DebugLog.w(TAG, "Feed update stopped for: " + mUrl, ex);
                    return true;
                }else
                {
                    //Something went wrong, non-standard error
                    DebugLog.e(TAG, "Feed update failed for: " + mUrl, ex);
                    return false;
                }
            }else{
                //Something went wrong, non-standard error
                DebugLog.e(TAG, "Feed update failed fatally for: " + mUrl, sax);
                return false;
            }

        }
        catch(Exception e){
            DebugLog.e(TAG, "Unknown parse error on feed: "+mUrl, e);
            return false;
        }
        DebugLog.i(TAG, "Entire Feed Parsed successfully: "+mUrl);
        return true;
    }

When one of my conditions is met, I use this code:

throw (new SAXException(new StopParsingException("Max Items reached")));

for example to stop the parser.

My guess is that when I throw the exception, the SAXParser stops working, but the InputSteamReader continues to download the rss feed from the server, since that is almost exactly the timing that my logs reveal.

Is there something wrong with my connection setup that makes only some servers not cooperate with me?

Alternatively is there a way to directly stop that InputStream safely before throwing my SAXException so that I don't have this problem?

Answer 1

It is just possible that the SAX parser is trying to "recover" in some way and reading your input stream to (for example) close matching tags. If this was occurring, you could probably prevent it by closing the input stream before throwing your exception.

Another alternative would be to register an error handler . In fact, the javadoc suggests that this could be the more correct approach.