Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site sequel.UUCP
Path: utzoo!linus!security!genrad!decvax!microsoft!uw-beaver!tektronix!ogcvax!sequel!phil
From: p...@sequel.UUCP
Newsgroups: net.bugs.4bsd
Subject: 4.2 BSD bugs in tip
Message-ID: <295@sequel.UUCP>
Date: Mon, 31-Oct-83 11:54:09 EST
Article-I.D.: sequel.295
Posted: Mon Oct 31 11:54:09 1983
Date-Received: Thu, 3-Nov-83 23:37:18 EST
Organization: Sequent Computer Systems, Portland
Lines: 90
4.2 BSD tip program has bugs in how it supports ventel modems. The version
of tip with 4.1c assumed the modem was set to "echo" while the 4.2 release
assumes "no echo". In addition, the code that checks for legal phone numbers
would not allow you to use either an '&' or '%' which are needed for the ventel.
I have included my diffs to tip/acu.c and tip/aculib/ventel.c that make it
work with the ventel (with echo on) and allow '&' and '%' in phone numbers.
Also I rearranged the code somewhat to make dialing faster (and used <..> mode).
Cheers..
% diff old.acu.c acu.c
66c66
< for (phnum = cp; any(*cp, "0123456789-*=K"); cp++)
---
> for (phnum = cp; any(*cp, "0123456789-*=K&%"); cp++)
101c101
< for (phnum = cp; any(*cp, "0123456789-*="); cp++)
---
> for (phnum = cp; any(*cp, "0123456789-*=&%"); cp++)
% diff old.ventel.c ventel.c
7c7
< * The Ventel is expected to be strapped for "no echo".
---
> * The Ventel is expected to be strapped for "echo".
16a17,25
> #define uucpdelay(num,denom) busyloop(CPUSPEED*num/denom)
> #define CPUSPEED 1000000 /* VAX 780 is 1MIPS */
> #define DELAY(n) { register long N = (n); while (--N > 0); }
>
> busyloop(n)
> {
> DELAY(n);
> }
>
40,45c49,53
< echo("#k$\r$\n$D$I$A$L$:$ ");
< for (cp = num; *cp; cp++) {
< sleep(1);
< write(FD, cp, 1);
< }
< echo("\r$\n");
---
> write(FD, "<K", 2);
> write(FD, num, strlen(num));
> write(FD, ">", 1);
> if (gobble('\n') && gobble('\n') && gobble('\007'))
> echo("\r$\n");
125,128d132
< #ifdef notdef
< if (boolean(value(VERBOSE)))
< putchar(c);
< #endif
146,156d149
< /*
< * Toggle DTR to force anyone off that might have left
< * the modem connected, and insure a consistent state
< * to start from.
< *
< * If you don't have the ioctl calls to diddle directly
< * with DTR, you can always try setting the baud rate to 0.
< */
< ioctl(FD, TIOCCDTR, 0);
< sleep(2);
< ioctl(FD, TIOCSDTR, 0);
164c157
< sleep(1);
---
> uucpdelay(1,10); /* sleep .1 sec */
166c159,161
< sleep(3);
---
> uucpdelay(1,10); /* sleep .1 sec */
> ioctl(fd, TIOCFLUSH, (struct sgttyb *)0);
> uucpdelay(1,10); /* sleep .1 sec */
177a173,184
> /*
> * Toggle DTR to force anyone off that might have left
> * the modem connected, and insure a consistent state.
> *
> * If you don't have the ioctl calls to diddle directly
> * with DTR, you can always try setting the baud rate to 0.
> */
> if (already == 0) {
> ioctl(FD, TIOCCDTR, 0);
> sleep(2);
> ioctl(FD, TIOCSDTR, 0);
> }
--
Phil Hochstetler (503) 627-9811
Sequent Computer Systems
...!sequent!phil
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site sequel.UUCP
Path: utzoo!linus!security!genrad!decvax!microsoft!uw-beaver!tektronix!ogcvax!sequel!phil
From: p...@sequel.UUCP
Newsgroups: net.bugs.4bsd
Subject: 4.2 BSD bug in getty
Message-ID: <294@sequel.UUCP>
Date: Sat, 5-Nov-83 00:14:43 EST
Article-I.D.: sequel.294
Posted: Sat Nov 5 00:14:43 1983
Date-Received: Thu, 3-Nov-83 23:15:28 EST
Organization: Sequent Computer Systems, Portland
Lines: 26
Bug in /usr/src/etc/getty/subr.c:
routine setchars():
[lines]
108 for (i = 0; charnames[i]; i++) {
109 p = *charnames[i];
110 if (p && *p)
111 *charvars[i] = *p;
112 else
113 *charvars[i] = '\0377'; <- ** constant wrong **
114 }
Constant '\0377' should be '\377'. VAX compiler actually generates:
cvtwb $7991,(r0)
which ends up being decimal 55 (the digit '7'). The constant is declared
incorrectly. K&R states an octal constant is a '\' followed by 1, 2, or
3 digits (not 4). Anybody got any guesses how it got 7991 anyway?
The intent was to set disable a value by setting it to (char)-1 as
tty(4) documents but instead sets it to a funny value.
** Why is not (char) -1 used instead of '\377' anyway? **
--
Phil Hochstetler (503) 627-9811
Sequent Computer Systems
...!sequel!phil (on November 1, sequel becomes sequent)
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site sequent.UUCP
Path: utzoo!linus!decvax!tektronix!ogcvax!sequent!phil
From: p...@sequent.UUCP
Newsgroups: net.bugs.4bsd
Subject: 4.2 BSD bug (/usr/dict/words -> /usr/lib/w2006)
Message-ID: <300@sequent.UUCP>
Date: Sun, 6-Nov-83 19:26:38 EST
Article-I.D.: sequent.300
Posted: Sun Nov 6 19:26:38 1983
Date-Received: Fri, 11-Nov-83 04:44:43 EST
Organization: Sequent Computer Systems, Portland
Lines: 27
Here is an interesting bug in the 4.2 BSD distribution.
% diff /usr/dict/words /usr/lib/w2006
4955c4955
< consanguine
---
> consanguiNe
4970,4971c4970,4971
< conservatism
< conservative
---
> conservadmsm
> conser6ative
6389c6389
< diamagnetism
---
> diamegnetism
8208d8207
< famish
You will notice that some words are trashed in /usr/lib/w2006 and
famish is deleted. I don't understand why the two aren't just
links. Does anything reference /usr/lib/w2006?
--
Phil Hochstetler (503) 627-9811
Sequent Computer Systems
...!sequent!phil
|