Problem at removing/set header of HttpMessage

["Trustin Lee (이희승)"]

Here's the latest build that includes the fix: http://is.gd/bSZKk

Trustin Lee (이희승) wrote:
> Nevermind.  I succeeded to reproduce the problem easily and checked in
> the fix with a proper test case.  Thanks again for reporting the bug!
> 
> Trustin
> 
> Trustin Lee (이희승) wrote:
>> Hi K,
>>
>> Thanks for reporting the bug.  You can post the bug here or file a JIRA
>> issue at: http://jira.jboss.org/
>>
>> Before applying the suggested fix, I'd like to make sure I can reproduce
>> the bug.  Could you please post some simple test code that reproduces
>> the problem?  For example:
>>
>>         HttpMessage m = new DefaultHttpRequest(...);
>>         m.setHeader("SESSION", "A");
>>         m.removeHeader("SESSION");
>>         System.out.println(m.getHeader("SESSION"));
>>         System.out.println(m.getHeaders());
>>
>> The expected output is 'null' and '[]' in this case.  You could
>> add/set/remove some headers to show the case where a certain header is
>> not removed.
>>
>> Thanks a lot,
>> Trustin
>>
>> krislo wrote:
>>> Hi,
>>>
>>>
>>> I detected some strange behavior on manipulating Http req/resp header (eq.
>>> removing or setting the new value to 'SESSION' header of RTSP message).
>>> Header line is not removed or is duplicated with the new set value.
>>>
>>> I have checked a little and focused on HttpHeader class and when
>>> removeHeader0(...) is called, Entry is just not detached from before/after
>>> in the linked list. I have changed the first for-loop to remove the links
>>> with e.remove() :
>>>
>>> ...
>>> for (;;) {
>>>             if (e.hash == h && eq(name, e.key)) {
>>>                 Entry next = e.next;
>>>                 e.remove();
>>>                 if (next != null) {
>>>                     entries[i] = next;
>>>                     e = next;
>>>                 } else {
>>>                     entries[i] = null;
>>>                     return;
>>>                 }
>>>             } else {
>>>                 break;
>>>             }
>>>         }
>>>
>>>
>>> By doing this it seems function behaves better but I haven't check deeper,
>>> so just FYI that issue is present.
>>>
>>>
>>>
>>> For further issue reporting (if I detect some other...) where it can be
>>> reported? 
>>> Is it ok to post to this forum?
>>>
>>> tnx & brg,
>>> K
>>>
>>>
> 

-- 
what we call human nature in actuality is human habit
http://gleamynode.net/


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 260 bytes
Desc: OpenPGP digital signature
Url : http://lists.jboss.org/pipermail/netty-users/attachments/20100504/ff533890/attachment-0001.bin