[BEEPbuilders] [beepcore] Caught NPE on SessionEvent.toString() when receivingsession close event

Dann Martens dann at tomoton.com
Mon Jun 20 14:55:27 PDT 2005


Hi,

Sure, sounds like a start of something to me 8)

My concerns regarding the current implementation:
 - pursuing a clear separation between:
  (a) the BEEP multi-channel functionality (more than one 
input/outputstream) as suggested by the the SCTP on TCP mapping 
contained in the spec; should culminate in a package wich would allow a 
developer to use this function separately (using typical Java I/O 
primitives).
  (b) the BEEP protocol framework (-> stuff with Profiles)
 - implementing as many Profiles in existence as humanly possible (very 
keen to see the Tunnel Profile, not at least to see if this thing 
actually works)
 - fixing SASL Profile issues
 - fixing inband messaging for Tuning
 - getting hold of Huston's last coding effort (would be related to (a) 
mostly)

I'll start posting my patches ASAP, together with the rationale for each 
change or fix.

Cya,
 Dann

P.S. Does everybody have the O'Reilly BEEP book, for reference ?

Daan Hoogland wrote:

>Dann Martens wrote:
>
>  
>
>>Daan, Christian,
>>
>>We have several options to host a branch of the project. If we agree
>>on the name for this branch, we could take it from there. Ping-ponging
>>patches would work of course, given enough dedication and hard work...
>>but I suppose we'd better find some version control shelter ;)
>>
>>Okay, here's me taking a stab at it:
>>'reBEEP' ?
>>
>>Same BSD-style license ?
>>    
>>
>
>You've got my vote on the name and the license, and my voice on the
>fork. I don't see a plan in the project you are proposing and I think it
>would need the founders somehow. Let's just start with considering
>patches here and consider a hosting scheme once there is some agreement
>on improvements.
>
>
>i.o.w. I opt for the patches if it is only three of us.
>
>ok?
>
>  
>
>>Cya,
>>Dann
>>
>>
>>
>>Daan Hoogland wrote:
>>
>>    
>>
>>>Dann Martens wrote:
>>>
>>> 
>>>
>>>      
>>>
>>>>Hi,
>>>>
>>>>In order for this to happen, either the original founding fathers of
>>>>the beepcore project need to overcome any personal reservations
>>>>regarding passing the torch, or someone should take the initiative to
>>>>branch the project and give it a new home as well.
>>>>  
>>>>        
>>>>
>>>Or we could use this list to share considderations and patches.
>>>
>>> 
>>>
>>>      
>>>
>>>>The new Java 5.0 features are compelling enough to consider a branch
>>>>(built-in SASL, ...). In addition, lessons should be learned from the
>>>>success/failure of the project up to now; some soul searching is in
>>>>order.
>>>>  
>>>>        
>>>>
>>>Hereby,... Are there any more people on this list actually using it in a
>>>project?
>>>
>>> 
>>>
>>>      
>>>
>>>>BEEP provides functionality at different levels, in a package deal.
>>>>The original authors have tried to spin it around already several
>>>>times in their attempts to huckster the technology, to no avail. I can
>>>>sum up many uses for BEEP, but none of them seem to require the entire
>>>>package at the same time.
>>>>
>>>>Clearly, BEEP appeals in many ways to a lot of us, but without any
>>>>means to commit your effort spent to improve the current
>>>>implementation, we would be wasting our time...
>>>>
>>>>Cya,
>>>>Dann   Daan Hoogland wrote:
>>>>
>>>>  
>>>>
>>>>        
>>>>
>>>>>H,
>>>>>
>>>>>nice to hear someone does something with this.
>>>>>
>>>>>I saw something maybe related with a tls profile. I tested the
>>>>>stability
>>>>>by removing the ip connection from under the c/s connection. It was
>>>>>the
>>>>>client that couldn't recover. Ididn't look into it to much as the
>>>>>connection is generally very stable. Also I thought it was JSSE
>>>>>related,
>>>>>but your scenario rings a bell somehow. I'm working with the latest
>>>>>version as well:-(
>>>>>
>>>>>Maybe if you get to do a fix you could post it here? The project could
>>>>>get alive the old fashioned way again:-)
>>>>>
>>>>>Christian Möller wrote:
>>>>>
>>>>>
>>>>>
>>>>>    
>>>>>
>>>>>          
>>>>>
>>>>>>Hi,
>>>>>>
>>>>>>currently I'm writing a prototype BEEP client and server testing the
>>>>>>protocol, some implementations of it - and my knowledge.
>>>>>>
>>>>>>In the course of my investigation I've written an ugly behaving
>>>>>>client
>>>>>>not sending any channel or session close requests; he only shuts down
>>>>>>the TCP connection when finished. My
>>>>>>org.beepcore.beep.core.event.SessionListener tries to log the
>>>>>>event he
>>>>>>receives with a call of sessionClosed(final SessionEvent event) - and
>>>>>>fails like this:
>>>>>>
>>>>>>java.lang.NullPointerException
>>>>>>   at
>>>>>>org.beepcore.beep.transport.tcp.TCPSession.toString(TCPSession.java:230)
>>>>>>
>>>>>>
>>>>>>   at java.lang.String.valueOf(String.java:2131)
>>>>>>   at java.lang.StringBuffer.append(StringBuffer.java:370)
>>>>>>   at java.util.EventObject.toString(EventObject.java:54)
>>>>>>   ...
>>>>>>
>>>>>>No heavy bug, but it shouldn't be much work to fix it. I'm working
>>>>>>with
>>>>>>version 0.9.08.
>>>>>>
>>>>>>Thanks and regards.
>>>>>>
>>>>>>Christian
>>>>>>_______________________________________________
>>>>>>BEEPbuilders mailing list
>>>>>>BEEPbuilders at lists.beepcore.org
>>>>>>http://drakken.dbc.mtview.ca.us/mailman/listinfo/beepbuilders
>>>>>>
>>>>>>
>>>>>> 
>>>>>>      
>>>>>>            
>>>>>>
>>>>>
>>>>>
>>>>>    
>>>>>          
>>>>>
>>> 
>>>
>>>      
>>>
>
>
>  
>



More information about the BEEPbuilders mailing list