Hi,
 
when experimenting with a larger number of real life articles, I observed the client to stop getting
headers at a given point. Debugging into this I found out that these articles contained either
newline or tab characters in their titles.
 
Looking at RFC2980: 2.8 XOVER I found:
>> Note that any tab and end-of-line characters in any header data that
>>    is returned will be converted to a space character.
 
After filtering these characters from the response the client received the headers without further complains.
 
The attached patch does this filtering. Also it uses a StringBuffer and a single call to prt.println()
instead many calls to prt.print(). This allowed to trace the communication between server and client and finally
to find the problem.
 
 
Rgs,
 
Christian
 

Attachment: NNTPArticleImpl.xover.diff
Description: Binary data

--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>


Reply via email to