Sun-Spots-Request@Rice.edu (William LeFebvre) (09/13/88)
SUN-SPOTS DIGEST Saturday, 10 September 1988 Volume 6 : Issue 224
Today's Topics:
Re: Notification on receiving new mail
Re: the "faces" program (new mail notification)
Lucid Lisp Editors
how 4.0 client default route is established
semantics of ioctl using TIOCCONS
Trouble porting AT&T C++ 1.2.1 to Sun 3
csh hanging under 4.0
4.0 linker : 'Wrong Machine Type' error
Bigger icons for frames?
UUCPD not included in SunOS 4.0?
NULLs in 4.0 file?
The .stabs assembler directive for debugging - any documentation?
Tool Recommendation?
LAT terminal servers?
Re: Notification on receiving new mail (source)
Send contributions to: sun-spots@rice.edu
Send subscription add/delete requests to: sun-spots-request@rice.edu
Bitnet readers can subscribe directly with the CMS command:
TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name
Recent backissues are available via anonymous FTP from "titan.rice.edu".
For volume X, issue Y, "get sun-spots/vXnY". They are also accessible
through the archive server: mail the request "send sun-spots vXnY" to
"archive-server@rice.edu" or mail the word "help" to the same address
for more information.
----------------------------------------------------------------------
Date: Fri, 9 Sep 88 17:24:12 BST
From: Andrew Findlay <Andrew.Findlay%brunel.ac.uk@nss.cs.ucl.ac.uk>
Subject: Re: Notification on receiving new mail
We now deliver mail into users' home directories, which achieves a similar
effect in most respects, but has other advantages. (In particular, mail
can then be subject to the user's file quota, and will be deleted along
with the rest of the files when the user leaves the system - very
important in places with a constantly changing user population)
Mail notification is handled by a new 'biff'-like datagram service called
'tell'. When mail is delivered, a 'tell' message is sent to all machines
where a user has recently logged in. If they are still there and have the
'biff bit' set on their TTY, the message is delivered.
The important thing here is only to send the message to those machines
which have a chance of delivering it. On a network with 50-100 machines
where mail messages arrive at the rate of 20-30 per hour, the load would
soon get silly. We handle that by having a file called '.whereami' in each
user's home directory. Whenever somebody logs in, their login script runs
'checkmail' which prints the familiar 'You have mail' message and also
enters the time and hostname into '.whereami'. The mail delivery program
(which delivers into '.mailbox') looks at .whereami to decide where to
send the 'tell' message.
I will supply source code to the delivery and notification programs if
anyone is interested. Given source to your mail delivery program, it
should be fairly easy to fit the 'tell' system in place of comsat.
>[.... it would be nice to have some sort of mail watcher tool that
>would notify you when /usr/spool/mail/<user> changed. To my knowledge,
>no such program exists. Although it wouldn't be hard.... --wnl ]]
The elm mail system comes with a couple of programs that do that -
'newmail' and 'wnewmail'. The main problem with them is that they do not
always die when the user logs out.
[[ I was thinking that it would be run as a tool under Suntools. In that
case, it would (hopefully) go away when you exited suntools. --wnl ]]
Andrew
>From Andrew Findlay at Brunel University, Uxbridge, UB8 3PH, UK
Andrew.Findlay@brunel.ac.uk phone: +44 895 74000 x2512
------------------------------
Date: 8 Sep 88 17:19:33 GMT
From: mkhaw@teknowledge-vaxc.arpa (Mike Khaw)
Subject: Re: the "faces" program (new mail notification)
Reference: v6n221
- Isn't the "faces" program in sun-sources just such a mail watcher?
-
- [[ I don't see a program called "faces" anywhere in the sun-source
- archives. Is it hiding as part of another package? --wnl ]]
Whoops, sorry. "faces" was posted to comp.sources.misc, in Vol. 3, #82.
It is apparently something developed by several people at Sun (but not a
Sun supported product).
Mike Khaw
--
internet: mkhaw@teknowledge.arpa
uucp: {uunet|sun|ucbvax|decwrl|uw-beaver}!mkhaw%teknowledge.arpa
hardcopy: Teknowledge Inc, 1850 Embarcadero Rd, POB 10119, Palo Alto, CA 94303
------------------------------
Date: Thu, 8 Sep 88 14:34:26 PDT
From: Ken Olum <kdo@lucid.com>
Subject: Lucid Lisp Editors
>SPE is a step in the right direction for bringing some of the power of the
>Symbolics' lisp development environment to the Sun. It has an improved
>editor (but face it, nearly _anything_ would have been an improvement over
>the Lucid sub-emacs editor :-( )....
You have clearly been using the 2.1 version of the Lucid Editor. I'll be
the first to admit that it wasn't very good. For version 3.0 of Sun/Lucid
Common Lisp we completely redid the editor. It is better documented and
more mature, although it still lacks some of the features of GNU Emacs.
The mouse interface is much better (and is user-extensible), but there are
still lots of improvements that could be made. You can save keyboard
macros to files. You can get scrollbars. You can get arglists and
function-documentation. You also get a real input-editor similar to that
of the Lisp Machine, which you don't get with GNU Emacs.
Sun Common Lisp 3.0 is currently in Beta test and will be available from
Sun very shortly.
Ken Olum
Lucid, Inc.
------------------------------
Date: Thu, 08 Sep 88 13:54:35 -0500
From: abe@mace.cc.purdue.edu (Vic Abell)
Subject: how 4.0 client default route is established
I previously reported that booting 4.0 on Sun clients appeared to produce
a bogus default route in their routing tables. I have managed to solve
this puzzle.
The questionable default route appears when a Sun client is booted from a
Sun server via an Ethernet on which an HP server (9000/350) also resides.
The Sun client ends up with a default route that points to the HP server,
not the Sun server as I had expected.
The following comment from the code for bootparamd explains what is
happening. There is one additional important fact - the HP server does
not run gated or routed, so its routing needs are met with static route
entries in its routing tables and those of the Sun server.
/*
* get_route paws through the kernel's routing table looking for a route
* via a gateway that is on the same network and subnet as the client.
* Any gateway will do because the selected gateway will return ICMP redirect
* messages to the client if it can not route the traffic itself.
*/
The result of the code that follows this comment is that bootparamd
discovers a routing table entry in the Sun server for the HP server's
diskless client net. Because the HP server is on the same network and
subnet as the Sun client, bootparamd then establishes the HP server's
address as the default route for the Sun client.
If it weren't for the fact that I have been unable to induce the HP server
to ICMP redirect or IP route, I may have never discovered this oddity. ;-(
Vic Abell
------------------------------
Date: Thu, 8 Sep 88 12:19:47 EDT
From: biggers@dg-rtp.dg.com (Mark Biggers)
Subject: semantics of ioctl using TIOCCONS
using this code on the slave side of a pty, one can get console i/o
redirection:
#include <sgtty.h>
#include <sys/file.h>
main()
{
int on = 1;
int tty;
if ( (tty = open("/dev/tty", O_RDWR, 0)) != -1 )
{
ioctl (tty, TIOCCONS, (char *) &on);
}
}
My question is, what, if anything, happens when the value of "on" is
changed (i.e., what is the semantics of TIOCCONS - I have read console(4)
and tty(4))? I have changed it to -1 and 0, and ran this program a second
time in a (X) window, and the console i/o still goes to that window. As
far as I can tell, TIOCCONS is a SunOS-ism, not in 4.3bsd or SysV.
However, it *is* useful! Any comments?
thanks, mark
biggers@dg-rtp.dg.com
------------------------------
Date: 8 Sep 88 17:22:35 GMT
From: acornrc!bob@ames.arc.nasa.gov (Bob Weissman)
Subject: Trouble porting AT&T C++ 1.2.1 to Sun 3
I'm having a devil of a time getting C++ version 1.2.1 to build on my Sun
3/60 running SunOS 3.5. I've had to edit makefiles and header files, but
that's no problem.
The problem is that when I try to compile the HelloWorld program, cfront
goes into an infinite loop.
I'd appreciate hearing from anyone who has successfully built AT&T C++
version 1.2.1 on a Sun 3 running SunOS 3.5 and who might have any hints
for me.
I had expected this procedure to be an "installation," not a "port."
Thanks for any help,
--
Bob Weissman
Internet: bob@acornrc.uucp
UUCP: ...!{ ames | decwrl | oliveb | pyramid }!acornrc!bob
Arpanet: bob%acornrc.uucp@ames.arc.nasa.gov
------------------------------
Date: Thu, 8 Sep 88 16:11:28 EDT
From: schwartz@gondor.cs.psu.edu (Scott Schwartz)
Subject: csh hanging under 4.0
On several occasions (at least once a day) I have the following happen to
me: info is scrolling past on the screen, I maybe hit ctrl-Q, and in any
case hit ctrl-C. The program dutifully exits, but the shell is left
hanging. "ps l" says WCHAN == Sysbase, STAT == I. What exactly is
Sysbase? To make matters stranger, if the hanging shell is the child of
another shell, killing the hanging shell causes the parent to inherit the
comatose condition.
I can't reproduce this on demand, which makes it hard to describe in
detail, unfortunately.
I am connected to the machine (A 4/260) via an Annex terminal
concentrator, if that has any bearing.
Anyone else experience anything similar?
-- Scott Schwartz
<schwartz@gondor.cs.psu.edu>
------------------------------
Date: Thu, 8 Sep 88 15:40:17 BST
From: <@nss.cs.ucl.ac.uk,@cvaxa.sussex.ac.uk:philipt@csta.uucp (Philip Taylor)>
Subject: 4.0 linker : 'Wrong Machine Type' error
When I try and link in an object file (called First) under SunOS 4.0 I get
the following message:
ld: ./.lnk/First: wrong machine type
Which is, apparently, telling me that the code is for an unsupported
architecture. However, using 'file' on ./.lnk/First tells me :
./.lnk/First: old sun-2 executable not stripped
I do not get the ld error under SunOS 3.4. Any ideas why there is a
difference (and how to get round the problem?).
Philip Taylor, School of Cognitive and Computing Sciences, University
of Sussex, Falmer, Sussex, UK. BN1 9QH.
JANET: philipt%csta@uk.ac.sussex.cvaxa
UUCP: ...mcvax!ukc!cvaxa!philipt
ARPA: philipt%csta%uk.ac.sussex.cvaxa@nss.cs.ucl.ac.uk
------------------------------
Date: Thu, 8 Sep 88 11:13:41 PDT
From: sytek!syteka!sridhar@hplabs.hp.com (Sridhar Acharya)
Subject: Bigger icons for frames?
I was trying to assign bigger icons to frames, like two 64 x 64 icons
joined together. I tried concatenating the two images I got from iconedit
and increasing the ICON_HEIGHT/ICON_WIDTH in "icon_create" but it did not
work.
Thanx for any pointers.
Sridhar Acharya
sun!sytek!syteka!sridhar
------------------------------
Date: 8 Sep 88 19:17:10 GMT
From: ehrlich@blitz.cs.psu.edu (Dan Ehrlich)
Subject: UUCPD not included in SunOS 4.0?
Does anyone know why Sun decided not to include the UUCPD in SunOS4.0? It
seems a reaonable thing for people who don't like hacking sendmail config
files to be able to send to local hosts vip UUCP over TCP.
Dan Ehrlich <ehrlich@blitz.cs.psu.edu>
The Pennsylvania State University
Department of Computer Science
University Park, PA 16802
------------------------------
Date: Thu, 08 Sep 88 14:08:34 -0500
From: abe@mace.cc.purdue.edu (Vic Abell)
Subject: NULLs in 4.0 file?
Has anyone seen any problems with characters of a file being NULLed under
SunOS 4.0? I have a situation in which a printer log file occasionally
has a legitimate entry converted to NULLs for no apparent reason. The
NULLs never seem to appear at the end of the file, but occur prior to the
last entry.
Only one program writes the log file, using stream I/O with an fopen()
mode of "a" and fwrite(). Immediately after the entry is written, it
appears in the file correctly. Some time later it is NULLed.
It almost appears as though the next appendage to the file leads the file
system code to believe that the end of file is at the beginning (not end)
of the previous entry, but the file is positioned past that entry. Thus,
when the write occurs, the file system code believes that there is a hole
to be filled with NULLs and does so.
The file system on which this happens is shared among the clients and the
server, but only a server process writes in the file. The problem occurs
several times a day, but so far I've been unable to discover a sequence of
events that makes it happen.
Any ideas, anyone?
Vic Abell
Purdue University Computing Center
------------------------------
Date: 8 Sep 88 12:44:12 GMT
From: mcvax!daimi!pederch@uunet.uu.net (Peder Chr. N|rgaard)
Subject: The .stabs assembler directive for debugging - any documentation?
We are developing a compiler for a novel language, and would like to be
able to debug the stuff without developing an entire new debugger. There-
fore I ask, whether anybody know the syntax for the arguments for the
.stabs assembler directive, so we can generate code which can be debugged
be adb or dbx.
Thank you in advance.
--
Peder Chr. Norgaard / pederch@daimi.dk (..uunet!mcvax!dkuug!daimi!pederch)
Computer Science Department, Aarhus University
Ny Munkegade 116, DK-8000 Aarhus C, DENMARK
phone: +45 6 12 71 88 / telefax: +45 6 13 99 19 / telex: 64767 aausci dk
------------------------------
Date: Thu Sep 8 11:07:21 1988
From: portal!cup.portal.com!DMasterson@sun.com
Subject: Tool Recommendation?
Are there any recommendations for tools for the SunOS. In particular, I'm
looking for a tool that would allow the development frames, windows,
canvases, text items, popups, etc. interactively and then would output a C
program skeleton that contains the SunView structures to make use of those
windows. Anyone know of such a beast?
David Masterson
DMasterson@cup.portal.com
------------------------------
Date: Thu, 8 Sep 88 17:19 EST
From: Peter W. Resnick <RESNICK@BINGVAXA.BITNET>
Subject: LAT terminal servers?
Howdy.
Any of you folk out there know if SUN's can deal with LAT ethernet
protocol terminal servers (i.e. DEC Servers) or if someone makes a good
piece of software or the like that will allow them to do so. Any or all
info would be good.
Pete
SUNY Binghamton
------------------------------
Date: Thu, 08 Sep 88 17:53:19 PDT
From: Craig Leres <leres@helios.ee.lbl.gov>
Subject: Re: Notification on receiving new mail (source)
Another option is to modify /bin/mail so that it sends biff packets to the
desired list of hosts. Appened is my replacement for mail's notifybiff()
routine. The list of hosts is maintained by yp in the network group entry
<domain-name>-biff. For example, some of my Suns are in the domain "rtsg"
so my /etc/netgroup file the following line:
rtsg-biff (ace,-,rtsg) (helios,-,rtsg) (hot,-,rtsg) (surf,-,rtsg)
If the net group entry doesn't exist, then the biff packets go to
localhost, just like before.
Craig
__________
#include <sys/socket.h>
#include <netinet/in.h>
#include <netdb.h>
#define NEW(t) (t *)malloc(sizeof(t))
notifybiff(msg)
char *msg;
{
char domain[64];
char *mp, *up, *dp;
int didany;
struct sockaddr_in addr;
struct hostlist {
struct hostlist *h_next;
struct sockaddr_in h_addr;
};
static struct hostlist *hl = 0;
register struct hostlist *hlp;
struct hostlist hjunk;
/* Build list of hostnames and inet addrs */
if (hl == 0 && getdomainname(domain, sizeof(domain)) >= 0) {
strcat(domain, "-biff");
setnetgrent(domain);
hlp = &hjunk;
while (getnetgrent(&mp, &up, &dp))
if (mp && nametoaddr(mp, &addr)) {
hlp->h_next = NEW(struct hostlist);
hlp = hlp->h_next;
hlp->h_addr = addr;
}
hlp->h_next = 0;
hl = hjunk.h_next;
endnetgrent();
}
/* Feep hosts */
didany = 0;
for (hlp = hl; hlp; hlp = hlp->h_next)
if (notifyone(&hlp->h_addr, msg))
didany++;
/* If we didn't do any, try localhost */
if (!didany && nametoaddr("localhost", &addr))
(void) notifyone(&addr, msg);
}
nametoaddr(host, ap)
char *host;
struct sockaddr_in *ap;
{
static u_short port = 0;
struct servent *sp;
struct hostent *hp;
if (port == 0) {
sp = getservbyname("biff", "udp");
if (sp == 0)
return(0);
port = sp->s_port;
}
hp = gethostbyname(host);
if (hp == 0)
return(0);
ap->sin_family = hp->h_addrtype;
ap->sin_port = port;
bcopy(hp->h_addr, &ap->sin_addr, hp->h_length);
return(1);
}
notifyone(ap, msg)
struct sockaddr_in *ap;
char *msg;
{
static int f = -1;
if (f < 0)
f = socket(AF_INET, SOCK_DGRAM, 0);
if (f >= 0)
sendto(f, msg, strlen(msg)+1, 0, ap, sizeof(*ap));
return(1);
}
------------------------------
End of SUN-Spots Digest
***********************