bug: tcp RST instead of FIN if child exits after parent closes path

[Brian White]

There have been feelings that the source of this problem is with the Windows
socket interface, but it doesn't seem to fit with the descriptions others
people have provided.  My original report is attached.

Max Bowsher <maxb@ukf.net> believes it's has to do with reference counting,
but I don't understand what he means by it.  There must already be some
kind of reference counting because the socket does not close until the
the last process with an open path to the socket either exits or explicitly
closes that path.

The problem is not dependant upon data pending in the TCP send window as
the example program I created sends no data and yet still demonstrates
the problem.

Clewis@mobilecom.com suggested the problem might be that the last process to
close the socket is not the same process that opened it.

I don't mean to harp on this issue; it's just that it is a fairly significant
problem that does not seem to have been adequately explained.

                                          Brian
                                 ( bcwhite@precidia.com )

-------------------------------------------------------------------------------
 the difference between theory and practice is less in theory than in practice
-------------- next part --------------
An embedded message was scrubbed...
From: Brian White <bcwhite@precidia.com>
Subject: bug: tcp RST instead of FIN if child exists after parent closes path 
Date: Tue, 15 Apr 2003 12:46:06 -0400
Size: 8211
URL: <http://cygwin.com/pipermail/cygwin/attachments/20030421/a31ad8c7/attachment.eml>
-------------- next part --------------
--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/