The security things can be a real consideration in the wild.  Some 
XML-RPC servers in use today expect custom User-Agent headers, Cookies, 
HTTP Basic Authentication, and/or SSL.   Not everyone is writing both 
the client and the server portions here.

Ryan Hoegg

Paul Libbrecht wrote:

> Yaxiong,
>
>
> If this is still the weird wrong end-of-header detection you 
> mentionned, than clearly looks to be bound to the headers that the 
> URL.getConnection does send. A simple test for you would be to 
> implement the dumbest http server, something that listens to the given 
> port and writes anything it receives to the system.out then close the 
> connection.
>
> You should then be able to see the differences between HttpClient and 
> HttpClientLite.
>
> Except for security things, I can't have a single clue why you would 
> not use HttpClientLite.
>
> Paul
>
>
> On Lundi, mars 18, 2002, at 05:39 , Lin, Yaxiong wrote:
>
>>
>> Todd:
>>
>> Thanks for the information.  I am using XmlRpcClient.  I did switch to
>> XmlRpcClientLite and ran it on NT  and that seems to have fixed my 
>> problem
>> also. But I have a problem with XmlRpcClientLite as it does not work 
>> on IBM
>> webSphere on zOS which is going to be our production environment.  So 
>> I have
>> to stay with XmlRpcClient.
>>
>> I was wondering if you or anyone know why XmlRpcClient causes 
>> webserver to
>> hang while XmlRpcClientLite does not and if there is any hope to 
>> debug and
>> fix this problem?
>>
>> Thanks.
>>
>> - Yaxiong
>>
>


Reply via email to