rjf@utcsri.UUCP (Bob Francis) (03/04/86)
THE FOLLOWING POSTINGS ARE RELAYED FROM dw@utdw From dw Fri Dec 20 13:27:23 1985 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA09140; Fri, 20 Dec 85 13:27:18 est Date: Fri, 20 Dec 85 13:27:18 est From: Dave Wortman <dw> Message-Id: <8512201827.AA09140@utcsri.UUCP> To: bdown, rjf Subject: a bad tip Cc: dw Status: RO I've ran into some trouble when I copied /usr/src/usr.bin/tip/* to my 32016 system and tried to recompile it there. 1) contrary to the makefile, tip will not compile at all if ACULOG is not defined. dummy macros defined in tip.h conflict with the procedures defined in log.c. I think the solution is to wrap #ifdef ACULOG .. #endif around the procedure definitions in log.c 2) the makefile contains the instructions: cd aculib; make ${MFLAGS} by MFLAGS are nowhere defined in the makefile, so this step blows up. From dw@utdw.UUCP Sun Dec 22 18:53:11 1985 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA02672; Sun, 22 Dec 85 18:53:05 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00397; Sun, 22 Dec 85 18:50:26 est Date: Sun, 22 Dec 85 18:50:26 est From: Dave Wortman <dw@utdw.UUCP> To: utcsri!utpeter!peter@uthub Subject: vadic problems Cc: dw, dw@utcsri, rjf@utcsri Status: RO I've had some problems getting the 32016 software to work with my RV 3451 modem. Had to change dial request from D to d for both uucico and tip to make communication work. ALso noticed that standard tip dialer v3451.c wasn't hanging up phone properly. Fixed that by using longer disconnect sequence from local.c. I'd be happy to send you source for these changes. From dw@utdw.UUCP Tue Dec 24 20:05:37 1985 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA29875; Tue, 24 Dec 85 20:05:06 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00193; Tue, 24 Dec 85 19:59:45 est Date: Tue, 24 Dec 85 19:59:45 est From: Dave Wortman <dw@utdw.UUCP> To: rjf@utcsri Subject: uucp condevs.c revised source Status: RO THIS IS JUST THE COMMENT AT THE BEGINNING OF A LARGE SOURCE FILE. IF YOU WANT THE SOURCE CONTACT dw@utdw . (rjf) improved hangup sequence for v3451 modem. condevs.c =============================================================== #ifndef lint static char sccsid[] = "@(#)condevs.c 5.6 (Berkeley) 8/12/83"; #endif /* * Here are various dialers to establish the machine-machine connection. * conn.c/condevs.c was glued together by Mike Mitchell. * The dialers were supplied by many people, to whom we are grateful. * * --------------------------------------------------------------------- * NOTE: * There is a bug that occurs at least on PDP11s due to a limitation of * setjmp/longjmp. If the routine that does a setjmp is interrupted * and longjmp-ed to, it loses its register variables (on a pdp11). * What works is if the routine that does the setjmp * calls a routine and it is the *subroutine* that is interrupted. * * Anyway, in conclusion, condevs.c is plagued with register variables * that are used inside * if (setjmp(...)) { * .... * } * * THE FIX: In dnopn(), for example, delete the 'register' Devices *dev. * (That was causing a core dump; deleting register fixed it.) * Also for dnopn delete 'register' int dnf... . * In pkopn, delete 'register' flds... . * There may be others, especially mcm's version of hysopen. * You could just delete all references to register, that is safest. * This problem might not occur on 4.1bsd, I am not sure. * Tom Truscott */ From dw@utdw.UUCP Sat Jan 11 22:22:05 1986 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA28225; Sat, 11 Jan 86 22:22:01 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00096; Sat, 11 Jan 86 22:12:42 est Date: Sat, 11 Jan 86 22:12:42 est From: Dave Wortman <dw@utdw.UUCP> To: peter@uthub Subject: vc problems Cc: dw, drg@utcsri, rjf@utcsri Status: RO I finally tracked down the problems I was having with vc. There were 2 places in sc.c where getchar was called directly to read user input (one in checkit() and one in your decoding sequence for the e[vx<>] command). getchar was returning an 8-bit character and hence the following character compares were failing. The fix is to replace calls on getchar with the macro nmgetch (== getchar() & 0177). From dw@utdw.UUCP Mon Jan 13 21:24:37 1986 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA18558; Mon, 13 Jan 86 21:24:33 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00485; Mon, 13 Jan 86 21:24:18 est Date: Mon, 13 Jan 86 21:24:18 est From: Dave Wortman <dw@utdw.UUCP> To: rjf@utcsri, thomson@uthub Subject: 32016 cp problem Cc: dw Status: RO There may be a problem with cp on the 32016 port. If you ask it to copy a directory, e.g. cp /bin /tmp/bin It does the copy with no complaint, and leave the destination file with its "I am a directory" bit removed. This seems wrong to me, I thought cp dir1 dir2 should fail. From dw@utdw.UUCP Sun Feb 2 14:59:38 1986 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA00439; Sun, 2 Feb 86 14:59:33 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00253; Sun, 2 Feb 86 14:56:18 est Date: Sun, 2 Feb 86 14:56:18 est From: Dave Wortman <dw@utdw.UUCP> To: rjf@utcsri, peter@uthub Subject: -me Cc: dw Status: RO When I tried to do a document using the -me macros, I discovered that /usr/lib/me was missing from utdw. Other systems may have the same problem. From dw Sun Feb 2 20:44:24 1986 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA06225; Sun, 2 Feb 86 20:44:20 est Date: Sun, 2 Feb 86 20:44:20 est From: Dave Wortman <dw> Message-Id: <8602030144.AA06225@utcsri.UUCP> To: rjf Subject: 32016 addbib bug Cc: dw, dw@utdw.UUCP Status: RO It appears that the 32016 version of the addbib program has a bug that has been fixed on this system. Look at the rlog for addbib.c in /xusr/usr.bin.refer.RCS From dw@utdw.UUCP Sat Feb 8 17:40:47 1986 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA12649; Sat, 8 Feb 86 17:40:41 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00359; Sat, 8 Feb 86 17:36:53 est Date: Sat, 8 Feb 86 17:36:53 est From: Dave Wortman <dw@utdw.UUCP> To: rjf@utcsri, peter@uthub, stew@uthub Subject: /usr/lib/refer BOTCHED Cc: dw, dw@utcsri Status: RO /usr/lib/refer should be a directory containing the utility programs mkey, inv and hunt that are invoked by refer. On this system, /usr/lib/refer was a copy of the binary program hunt. I think someone goofed when refer was originally installed. From dw@utdw.UUCP Sat Feb 8 17:44:15 1986 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA12735; Sat, 8 Feb 86 17:44:10 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00398; Sat, 8 Feb 86 17:40:20 est Date: Sat, 8 Feb 86 17:40:20 est From: Dave Wortman <dw@utdw.UUCP> To: drg@utcsri, rjf@utcsri Subject: 32016 bugs Cc: dw, dw@utcsri, peter@uthub, stew@uthub Status: RO Is there a master list of 32016 software bugs that have been repaired somewhere? I would like to know what needs to be fixed on this system. From dw@utdw.UUCP Mon Feb 17 10:24:23 1986 Received: by utcsri.UUCP (4.24/5.1.UofT) id AA09939; Mon, 17 Feb 86 10:24:14 est Received: by utdw.UUCP (4.24/5.1.UofT) id AA00094; Mon, 17 Feb 86 10:18:40 est Date: Mon, 17 Feb 86 10:18:40 est From: Dave Wortman <dw@utdw.UUCP> To: rjf@utcsri Subject: /usr/lib/refer bug Cc: dw Status: RO I think the reason the the directory /usr/lib/refer was botched on this machine was that the makefile for refer is wrong. In the install step it does 3 installs on /usr/lib/refer without first checking that the directory exists. If it doesn't, then you'll get the problem I discovered of the last binary program being installed as /usr/lib/refer.