On Thu, 21 Nov 2002, Christopher Sawtell wrote:
...
> Is the lock file appearing and disappearing correctly, and are both
> mgetty and sendfax both configured to use the same file?

Thank you, this indeed is a very good question. Let's take a look.
When I send a fax I have this one:

-rw-r--r--    1 root     root           11 Nov 22 02:14 LCK..ttyS0

which relates perfectly to the line

fax-devices ttyS0

in sendfax.config. Now if I have mgetty wait for an incoming fax, I
get no lockfile at all. But I think that is ok, because there is no
reason to lock the modem as long as there are no transmissions.
However, mgetty will probably check if the modem line is locked before
respawning. My inittab entry used mgetty /dev/modem. All right, change
sendfax.config to use "modem", too; now we have

-rw-r--r--    1 root     root           11 Nov 22 02:27 LCK..modem

for sendfax, and when mgetty really takes the incoming call, we have

-rw-r--r--    1 root     root           11 Nov 22 02:30 LCK..modem

but only as long as the modem is OH. Now the trial with everything set
in inittab, result: It is working, the original mgetty surprisingly
even stays until sendfax has completed, but after that it is
terminated, and a new mgetty is started (respawn)! I shortly got a
zombie cron process after sendfax started, but functionally everything
seems to be all right now.

So, while /dev/modem is a symbolic link to /dev/ttyS0, the lock files
are named according to whatever naming is used in the commands
starting either sendfax or mgetty. And mgetty, of course, has no means
to know it is all the same device.

Thank you for pointing me in the right direction, Christopher!

Cheers,

Helmut.

+----------------+
| Helmut Walle   |
| [EMAIL PROTECTED] |
+----------------+


Reply via email to