[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference bulova::decw_jan-89_to_nov-90

Title:DECWINDOWS 26-JAN-89 to 29-NOV-90
Notice:See 1639.0 for VMS V5.3 kit; 2043.0 for 5.4 IFT kit
Moderator:STAR::VATNE
Created:Mon Oct 30 1989
Last Modified:Mon Dec 31 1990
Last Successful Update:Fri Jun 06 1997
Number of topics:3726
Total number of notes:19516

2389.0. ""Quit Session" kills remote application" by STREEP::MCCAUSLAND () Mon Mar 05 1990 06:49

	Can the effect of "Quit Session" on remote applications be controlled?
	Customer reports that remote clients get a fatal network error before
	his own handler manages to shutdown the display cleanly.

	Application is a large financial trading system where one major client
	controls many remote displays so clean a shutdown is very important.

	VMS V5.2/DECnet Transport

	Thanks,  Andrew
T.RTitleUserPersonal
Name		DateLines
2389.1STAR::MCLEMANJeff McLeman, VMS DevelopmentMon Mar 05 1990 09:474
    Quit Session causes the server to reset, therefore clearing all
    links. Call it a security feature.
2389.2STAR::ORGOVANVince OrgovanMon Mar 05 1990 20:397
    The Session Manager uses the Xlib function XKillClient to abort
    all connections to the server in VMS V5.2. In VMS V5.3, the 
    mechanism is different but the effect is the same: the session
    manager makes a special request (via an X extension) to the server
    asking it to abort all connections. 
    
    Is the customer using XSetIOErrorHandler?
2389.3VMS V5.3 may fix problem!STREEP::MCCAUSLANDWed Mar 07 1990 07:3813
	Re .2, Yes the application is using XSetIOErrorHandler.

	The handler is being called, but after the windows have been deleted
	and shortly before his client application gets killed.

	Customer will try VMS V5.3 as it sounds as if the client application
	may not be killed and connections to other displays may be unaffected.

	This behaviour should be OK though he would also like to close the
	display himself, however I guess this is the perogative of the window
	manager/user!

	Thanks, Andrew
2389.4Need ICCCMDECWIN::FISHERBurns Fisher 381-1466, ZKO3-4/W23Wed Mar 07 1990 20:3118
    I believe that the ICCCM defines a way to notify a client that it is
    about to die if it does not kill itself soon.  This is a relatively
    recent "standard", and we do not yet implement it.  We probably should,
    but I can't speak for anyone else as to whether it is on their work
    list.
    
    As to the error handler, there is another "misfeature" about this one.
    The io error handler must not return.  The inventor of Xlib (who shall
    remain nameless, since he has already issued plenty of mea culpas on
    this one) believe that once the connection died, the client was dead,
    not taking into account multiple connection clients.  Sigh.
    
    Thus, if you have multiple displays, and one of them shuts down, I'm
    afraid you are out of luck for now.  You should submit a QAR requesting
    that the ICCCM shutdown be implemented.
    
    Burns
2389.5Thanks, its #02700 in V5 QAR Database...STREEP::MCCAUSLANDFri Mar 09 1990 07:160