You are viewing a plain text version of this content. The canonical link for it is here.
Posted to httpclient-users@hc.apache.org by Robert Bowen <sy...@yahoo.com> on 2006/07/24 17:35:14 UTC

Bug in 3.0.1 HttpParser.readRawLine?

Hello all.

I looked in the bug database for this but couldn't find anything. 

We are calling objHttp.executeMethod() with method type Post. A strange is happening the first time we change the info being sent in out HTML form -- it fails with a "server failed to respond" error. The second time we send that same form data, it works, always.

I debugged it until I found that in HttpParser.readRawLine(), the following while:

public static byte[] readRawLine(InputStream inputStream)
  ...  

   while ((ch = inputStream.read()) >= 0) {
      ...
   }

fails after the first iteration, and ch is equal to '-1', which causes HttpClient to throw an Exception. As I said, it only happens the first time you send the form, and it happens 99.9% of the time on the first try.

We just did a "what the heck" test with version 3.1 and it seems to be fixed, but we are hesitant to install an alpha.jar on the client's machine.

Anyone a) seen this error b) know for sure if it's fixed in 3.1 c) know of any other fix and d) when 3.1 will be final? (on the JIRA page it says it IS final already ...)

Many thanks in advance!
Bob

Re: Bug in 3.0.1 HttpParser.readRawLine?

Posted by Oleg Kalnichevski <ol...@apache.org>.
On Mon, 2006-07-24 at 08:35 -0700, Robert Bowen wrote:
> Hello all.
> 
> I looked in the bug database for this but couldn't find anything. 
> 
> We are calling objHttp.executeMethod() with method type Post. A strange is happening the first time we change the info being sent in out HTML form -- it fails with a "server failed to respond" error. The second time we send that same form data, it works, always.
> 
> I debugged it until I found that in HttpParser.readRawLine(), the following while:
> 
> public static byte[] readRawLine(InputStream inputStream)
>   ...  
> 
>    while ((ch = inputStream.read()) >= 0) {
>       ...
>    }
> 
> fails after the first iteration, and ch is equal to '-1', which causes HttpClient to throw an Exception. As I said, it only happens the first time you send the form, and it happens 99.9% of the time on the first try.
> 

This is a server side problem. The server is not meant to drop the
connection without returning a valid HTTP response


> We just did a "what the heck" test with version 3.1 and it seems to be fixed, but we are hesitant to install an alpha.jar on the client's machine.

There's been no code changes in HttpParser class between 3.0 and
3.1-alpha1

> 
> Anyone a) seen this error b) know for sure if it's fixed in 3.1 c) know of any other fix and d) when 3.1 will be final? (on the JIRA page it says it IS final already ...)

The only thing that makes HttpClient 3.1 ALPHA is a new cookie policy
(RFC2965), whose API we cannot be frozen at this point. The rest of the
code base is at least as stable as HttpClient 3.0.1

Oleg

> 
> Many thanks in advance!
> Bob


---------------------------------------------------------------------
To unsubscribe, e-mail: httpclient-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: httpclient-user-help@jakarta.apache.org