On Mon, 31 Jul 1995 21:33:18 GMT Steve Elmer said:
>Sorry about any vagueness, our understanding may not have been complete when
>the documentation was written...
>If you are in the shell, then the kill man page states:
>     -s signal_name
>          sends the signal signal_name to the process instead of the
>          SIGTERM signal.
 
And as someone else pointed out to me (I think by private e-mail) the signal
names are stated without the SIG prefix (e.g., kill -HUP, not kill -SIGHUP as
the man page sort of suggests).
 
>As long as the UID matches, you can send the signal.  I've done this many
>times from my test environment.  There is currently _no_ user on the system
>able to send a signal to a process it doesn't normally have access to.  The
>"privileged user" concept for signals is not implemented.
 
OK, I'll have to check this out.  When it failed for MANAGER.SYS, I never
thought to try from a non-privileged userID.
 
>The kill man page lists the available signals, I have frequently used SIGSTOP/
>SIGCONT to control processes "in the background" in my test environment.  The
>SIGUSR1 and SIGUSR2 signals are implemented, this should have been clearly
>documented; the file /usr/include/signals.h shows the names/values of all
>valid signals.
 
I had seen them (it was a programmatic kill I wanted anyway, specifically a
SIGUSR2 for a porting project).  If it works from the same UID I'll be elated
since I didn't want to run it as manager.sys in the first place :-)
 
Thanks for the info, will check it out tomorrow...
 
[\] Jeff Kell <[log in to unmask]>