bugunderware - Bugs: bug #6117, client crash or hangs when quitting

 
 
Show feedback again

bug #6117: client crash or hangs when quitting

Submitted by:  Loic Dachary <loic>
Submitted on:  Fri 09 Jun 2006 09:16:24 AM UTC  
 
Category: NoneEffort: 0.00
Priority: 5 - NormalSeverity: 5 - Blocker
Status: ConfirmedPrivacy: Public
Assigned to: Igor Kravtchenko <izidor79>Open/Closed: Open

Add a New Comment (Rich MarkupRich Markup):
   

You are not logged in

Please log in, so followups can be emailed to you.

 

Wed 14 Jun 2006 03:21:06 PM UTC, comment #3:

> Although TCP is reliable for packet transmission by design, is it
> also designed to be reliable at disconnection ? Or is it possible
> that a disconnection is unoticed if packet retransmissions are
> too high / takes too long


Is it not supposed to be "time-out" dependant using the setsockopt()?

It's then supposed to be closed "gracefully" and recv() returns
an error message. To be "unoticed" of a disconnection after a
time-out value seems to me not possible.

Igor Kravtchenko <izidor79>
Project MemberIn charge of this item.
Wed 14 Jun 2006 03:13:16 PM UTC, comment #2:

Tried 3 times on the Roland's PC.
It never said that the client was already connected.
I was able each time to reconnect properly and to continue playing.

Igor Kravtchenko <izidor79>
Project MemberIn charge of this item.
Fri 09 Jun 2006 06:02:11 PM UTC, comment #1:

The client was run from a laptop with wifi and the wifi environment was hostile (many AP present, daily timeouts). It may be associated with the error. In any case it's worth trying to reproduce the bug with wifi instead of cable connection. If the wifi connection is suspended while the clients quits, it may be the cause of the error. Although TCP is reliable for packet transmission by design, is it also designed to be reliable at disconnection ? Or is it possible that a disconnection is unoticed if packet retransmissions are too high / takes too long ?

Loic Dachary <loic>
Project Administrator
Fri 09 Jun 2006 09:16:24 AM UTC, original submission:

Here is how to reproduce the problem
- run the client on Roland machine
- login
- connect to a table
- quit
- run the client again
- it says the client is already connected

From the server side it shows that the client is still connected. Which means that either the client did not really quit and did not close the socket (hang) but was dead enough for another instance to be run without troubles. Or that the client crashes and the corresponding sockets are not closed immediately by the system (windows XP).

After a while (2 minutes) the server timesout on the client and decides that it is disconnected (keepalive safeguard plays its role).

It may be quite difficult to reproduce the bug but please add context description in this bug entry if it ever shows up again.

Loic Dachary <loic>
Project Administrator

 

(Note: upload size limit is set to 1024 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -unavailable- added by cpinson (Updated the item)
  • -unavailable- added by izidor79 (Posted a comment)
  • -unavailable- added by loic
  • -unavailable- added by loic (Submitted the item)
  •  

    Do you think this task is very important?
    If so, you can click here to add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.

     

    Please enter the title of George Orwell's famous dystopian book (it's a date):

     

     

    Follow 2 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Tue 10 Oct 2006 01:09:47 PM UTCcpinsonPriority9 - Immediate=>5 - Normal
    Fri 09 Jun 2006 09:16:24 AM UTCloicCarbon-Copy-=>Added -unavailable-
    Show feedback again

    Back to the top


    Powered by Savane 3.1-cleanup