another mean-spirited controversy

[Paul Vixie Esq]

I'm porting and hacking /usr/lib/uucp/acucntrl to work on my Symmetric.
It seems to be an accepted tradition when working on UUCP code to be very
very sloppy, so I won't mention (except in passing) the vile feeling in
the pit of my stomach that grows by the hour as I look at this code...

Instead, I'd like to insert my entry into the UUCP/dialin/dialout war...
Here is a comment extracted from my version of settys() in acucntrl.c:

/* I don't know where in the Symmetric kernel to poke for this.  However,
 * there's a nifty little program called /etc/modemctl that does the trick
 * for us; we'll fork and exec that.  Have we created enough subprocesses
 * from UUCICO yet?
 *
 * the reason this program exists at all is because we can't open the
 * tty device for dialing if (modem control enabled) && (no carrier);
 * and if we set (no modem control), then a SIGHUP will never be sent
 * when the line is hung up (for dialup).
 *
 * the place to fix this problem cleanly is in the TTY DRIVER! yes!
 * create two tty files for each tty device -- a dialin and a dialout.
 * flip a bit in the minor number to tell which is which.  you could
 * even get cute (like HP-UX (SysV?)) and detect that the getty is
 * suspended in open(), waiting for carrier; you could then suspend
 * it even further, allowing the open of the dialout port to complete,
 * a full UUCP (or TIP) transaction to take place, and finally let
 * the poor getty back into its suspended open wait mode, NEVER
 * needing to send INIT any signals, because the getty always exists!
 * Too bad the people at CSRG don't have to use serial lines anymore,
 * or they might get around to this.
 */
-- 
Paul A. Vixie        {ptsfa, crash, winfree}!vixie!paul
329 Noe Street       dual!ptsfa!vixie!paul at ucbvax.Berkeley.EDU
San Francisco        
CA  94116            paul at vixie.UUCP     (415) 864-7013