hc-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ortwin Gl├╝ck <ortwin.glu...@nose.ch>
Subject Re: [PATCH] handling close properly on response input streams
Date Fri, 06 Dec 2002 16:16:59 GMT
Eric Johnson wrote:
> Ortwin,
> Thanks for try out my patch.  Based on the stack traces you included, 
> I'm quite puzzled, and not sure how this failure is anything that my 
> patch could cause.  I was expecting a failure (if one were to occur due 
> to my changes, which I hope it wouldn't) on ConnectMethod.java line 119. 
> Instead, the failure indicated below occurs at line 115.  Only a 
> precious few changes in my patch, such as setting state variables, would 
> actually be executed by the time that the failure occurs.

Line numbers might not be in sync here with your local code. Line 115 
is:  int code = super.execute(state, conn);

> For the moment then, I'm standing by my patch, in that I believe 
> (perhaps wrongly!) that the problem lies somewhere else, and that if the 
> problem is in my patch, that I need more information.  Reading the code, 
> it looks like the communication is as follows:
> Client:
> Various headers: various values
> (wait for HTTP/1.1 200 OK response)
> (response's first line doesn't arrive, or arrives garbled, throw 
> recoverable exception)

Well, the problem does not occur in the current CVS HEAD. So it must be 
your patch. I am not saying that the error occurs in your code. But the 
changed code makes the error happen somehow. It looks a lot like a 
stream is not fully read before reuse or repeating the request or alike.

View raw message