https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=2592
Abhik Sarkar <sarkar.abhik@xxxxxxxxx> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |sarkar.abhik@xxxxxxxxx
--- Comment #1 from Abhik Sarkar <sarkar.abhik@xxxxxxxxx> 2008-06-19 13:54:45 PDT ---
Looking at dumpcap.c, I don't think the problem could happen on non-Win32
platforms(I have only Win32, so can't be sure). However, on Windows it seems to
be by design because of some restrictions... if we look at capture_cleanup (for
windows):
<quote>
If we're not running as a capture child, we might be running as
a service; ignore CTRL_LOGOFF_EVENT, so we keep running after the
user logs out. (XXX - can we explicitly check whether we're
running as a service?) */
...
/* Keep capture running if we're a service and a user logs off */
if (capture_child || (dwCtrlType != CTRL_LOGOFF_EVENT)) {
capture_loop_stop();
return TRUE;
} else {
return FALSE;
}
</quote>
If "dumpcap -S" is started directly in a terminal, capture_child will be false,
and the signal handler will hence always return a FALSE... resulting in the
program not exiting on a ctrl-C.
I can see only two solutions:
1- We make it possible to determine if dumpcap is indeed running as a service
or not. Possible solutions:
- Pass a unique command line argument only when starting as a service
- Perhaps this link has the answer:
http://marc.info/?l=openssl-dev&m=104401851331452&w=2
2- Perhaps a quick hack for now (unless this will have any side effects), have
print_statistics as a static variable and change
if (capture_child || (dwCtrlType != CTRL_LOGOFF_EVENT)) {
to
if (capture_child || (dwCtrlType != CTRL_LOGOFF_EVENT) || print_statistics) {
--
Configure bugmail: https://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
