Thank you Nico (and Jay) for the help. I looked at the processes running as you suggested Nico and there doesn't seem to be any competition for /dev/ttyS1 that I can see (I have appended the output of ps -A at the end the this mail).
I've been trying everything I can think of. I tried to look at the what is going on with the "sh -x `which faxaddmodem`" and here is the pertinent output:
"+ prompt Serial port that modem is connected to []?
+ echo -n Serial port that modem is connected to []?
Serial port that modem is connected to []? + read TTY
ttyS1
+ [ -z ttyS1 -o ! -c /dev/ttyS1 ]
+ JUNK=/tmp/addmodem997
+ trap /bin/rm -f $JUNK; exit 1 0 1 2 15
+ [ ! -d /var/lock ]
++ ttyPort ttyS1
++ expr ttyS1 : tty\(.*\)
+ PORT=S1
++ ttyLocks ttyS1
++ echo /var/lock/LCK..ttyS1
+ LOCKX=/var/lock/LCK..ttyS1
++ ttyAliases ttyS1
++ echo /dev/ttyS1
+ DEVS=/dev/ttyS1
++ ttyDev ttyS1
++ echo /dev/ttyS1
+ tdev=/dev/ttyS1
++ echo ttyS1
++ tr / _
+ DEVID=ttyS1
+ CONFIG=/var/spool/fax/etc/config.ttyS1
+ [ -f /var/lock/LCK..ttyS1 ]
+ OLDCONFIG=
+ OLDFIFO=
+ checkPort ttyS1
+ return
+ JUNK=/tmp/addmodem997 /var/lock/LCK..ttyS1
++ expr 997 : .*\(..........\)
+ LOCKSTR= 997
+ echo 997
+ fuser -k /dev/ttyS1
+ cat
Hmm, there does not appear to be an fuser command on your machine.
This means that I am unable to insure that all processes using the
modem have been killed. I will keep going, but beware that you may
have competition for the modem.
+ cat
And when I Ctrl-C out I get this (irrelevant code cut out).
"+++ /bin/rm -f /tmp/addmodem997 /var/lock/LCK..ttyS1 /tmp/faxpr997
+++ exit 1
+ /bin/rm -f /tmp/addmodem997 /var/lock/LCK..ttyS1 /tmp/faxpr997
+ exit 1
"
I can see that the program uses fuse -k to kill the process, but I am not knowledgeable enough to know what else is going on (yet).
By the way, I get this problem on two machines, both UART 16450 (but diff manufacturer for the cards), both 486's, and both RH 6.0. The only consistency is the modem. But it seems to work fine with minicom, and statserial looks good.
thanks again for your help,
Jeff
PS. the ps list below
>> On Mon, Aug 02, 1999 at 11:17:39AM -0400, Nico Kadel-Garcia wrote:
>> > > I get the error message:
>> > > "Hmmm, there does not appear to be an fuser command on your machine. This
>> > > means that I an unable to insure that all processes using the modem have
>> > > been killed. I will keep going but be aware that you may have competition
>> > > for the modem"
>> >
>> > Jeff, look for other jobs that have things running on /dev/modem or
>> > /dev/ttyS1. mgetty (if you have that installed!), minicom, PPP, etc.
>> > "fuser" should *not* fail.
>
>
>> Forgive me, Nico, but doesn't that message mean that fuser is
>> _missing_, rather than merely not running?
>
>Actually, no. This is an unfortunate aspect of many error messages
>in systems maintained by people without a manager to tell them "make
>the messages correct for every possibility": it wasn't able to successfully
>*RUN* fuser, and assumes you actually don't have one. It's not always
>right about it. The odds of a modern OS not having fuser is quite
>small.
>
>If you run "sh -x `which faxaddmodem`", you should be able to see
>exactly what it does with the fuser that generates this message.
