Sun-Spots-Request@RICE.EDU (William LeFebvre) (08/25/88)
SUN-SPOTS DIGEST Tuesday, 23 August 1988 Volume 6 : Issue 195 Today's Topics: Re: Proxy arp Re: Manual Binders for the Sun Re: troff to postscript Announcement: Fig 1.4.FS and TransFig 1.4 Release 3 OS 4.0 lpd problem lpr -s /tmp/file fails on 4.0 clients (with fix) pwd: getwd: can't open .. Problem with biff and pseudo terminal ownership passwd matching across YP domains? printcap vs. XON/XOF (SunOS4.0)? troff leaders? 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: Thu, 18 Aug 88 13:25:11 EDT From: bob@allosaur.cis.ohio-state.edu (Bob Sutterfield) Subject: Re: Proxy arp Reference: v6n187 In article <1988.08.17.16.00.05.763.10323@rice.edu> Barry Shein <bzs@bu-cs.bu.edu> writes: >I wrote a proxy arp based upon Sun's rarpd server which has been in >production here at BU for over a year. and which we've been using for several months. Thanks! >...My Proxyd kind of got out there by accident when someone mentioned >I had given them a copy on one of the major lists. Oops - sorry (red face :-). I hope I'm not stealing Haavard Eidnes' thunder by mentioning this here, but since he did provide it: >... >B) Haavard Eidnes claims to have completely written a new Proxyd from >scratch based on my basic design (not the Sun sources) and I assume >will make all sources available... >You should probably wait for B) to become available and that will >solve all sorts of problems. Mr Eidnes asked us to provide a distribution point on this side of the pond. You can get his proxyd via: (a) anonymous FTP to tut.cis.ohio-state.edu in proxyarp/proxyarpd.shar.Z (b) anonymous UUCP to osu-cis in /u/public/proxyarp/proxyarp.sharZ Option (b) works like everything else we redistribute. If you don't have instructions, drop me a line. We haven't had a chance to try out this new Proxy ARP daemon. If you get it from us, and have problems with it, please send mail to the author, Haavard Eidnes (he@idt.unit.no). --- Bob Sutterfield, Department of Computer and Information Science The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277 bob@cis.ohio-state.edu or ...!{att,pyramid,killer}!cis.ohio-state.edu!bob ------------------------------ Date: 18 Aug 88 17:52:26 GMT From: ehrlich@blitz.cs.psu.edu (Dan Ehrlich) Subject: Re: Manual Binders for the Sun At Penn State we order them direct from the manufacturer. There may be more than one company that makes these, but here are the particulars on the one we use: Buchan Industries Clifton Heights, PA 19018 +1 215 622 3500 Call them up and ask for the current price list and catalogs. It seems to take two of the eleven inch racks to hold the entire set of SUN OS 3.5 manuals. Each rack was about $US 65. Haven't tried putting the SUN OS 4 manuals in yet. Dan Ehrlich <ehrlich@blitz.cs.psu.edu> The Pennsylvania State University Department of Computer Science University Park, PA 16802 ------------------------------ Date: Mon, 15 Aug 88 15:38:02 EDT From: suneast!rahthum!ppk@sun.com (Pravin Kumar) Subject: Re: troff to postscript >From: Paul Hudson <mcvax!moncam!paul@uunet.uu.net> > >I don't know of a PD filter, but a quick and dirty (And slow) one >shouldn't take too long. If there's any demand, I'll knock one up. Under SunOS 4.0, you can do "troff -t | lpr -t" to do troff to PostScript. Read the SunOS 4.0 Change Notes (page 29, Section 2.7 Utilities) for more information. pk [[ Although the Change Notes really do say something to that effect, I strongly suspect that you still need to have the "Transcript" software installed in order for it to work. You also need to have your default printer (or whatever printer is specified in your PRINTER environment variable) set up correctly in /etc/printcap. I think that the Change Notes were pointing out that "lpr -t" now worked with Transcript, making "ptroff" obsolete. I don't think they intended to imply that Transcript comes with 4.0. But, I have occasionally been wrong. --wnl ]] ------------------------------ Date: Thu, 18 Aug 88 10:21:18 EDT From: beck@cs.cornell.edu (Micah Beck) Subject: Announcement: Fig 1.4.FS and TransFig 1.4 Release 3 Announcing availability of two software components: Fig 1.4.FS and TransFig Version 1.4 Release 3: * Fig 1.4.FS is an enhanced version of Fig 1.4 Release 2. Some bugs are fixed, the user interface is enhanced, and some new graphics features are supported. * TransFig 1.4 Release 3 includes Fig code translators which implement the extension to Fig code needed by Fig 1.4.FS to implement new graphics features. New graphics features include setable text font and size, and area fill. Both Fig 1.4.FS and TransFig Release 3 use an extension to the definition of Fig code. This extension maintains compatibility with standard Fig code. Fig 1.4.FS was developed by Frank Schmuck at Cornell University. It is no sense an "official" version of Fig. The features added to Release 2 are summarized below: 1. New operation: zoom 2. New operation: change object properties 3. New modes: LaTeX line and vector 4. Search uses real mouse position rather than magnet-rounded coordinates 5. Centered and right justified text is drawn correctly. 6. Displays current file and directory on window label Frank Schmuck has left Cornell and I now support Fig 1.4.FS. It is available for anonymous FTP from svax.cs.cornell.edu as ~ftp/pub/fig/fig-fs.tar.Z. TransFig Release 3 is available as ~ftp/pub/fig/transfig.tar.Z. [[ Be sure to use "image" transfer type for compressed files (use the FTP command "binary" before transferring). --wnl ]] Some technical notes: * Fig 1.4.FS is different enough from Release 2 that it was not possible to automatically generate patch files for the bug fixes. Sorry. * In a recent posting, Ken Yap of Rochester U. suggested that Fig hacking be done on Xfig, his generalized Fig which runs under both Suntools and X11. Unfortunately, Fig 1.4.FS was developed in parallel to Xfig, and thus some work will be required to merge the two. * If you picked up a version of Fig 1.4.FS or TransFig Release 3 before reading this note, then you may have to apply this fix: line 121 of read.c should be tfx_flag = !strncmp(buf, "#FIG 1.4-TFX", 11); /* check for TFX */ Micah Beck beck@svax.cs.cornell.edu Cornell Dept of Computer Science ------------------------------ Date: Thu, 18 Aug 88 12:16:55 EDT From: ileaf!io!jaguar!dbjag@eddie.mit.edu (David Benjamin) Subject: OS 4.0 lpd problem We've run into a problem with the Sun OS 4.0 lpd (line printer daemon) in which some printing jobs queued up with the "lpr -s" command disappear for no apparent reason. The problem, apparently, is that lpd4.0 does not like "lpr -s" queueing when the file to be printed resides on an NFS filesystem. Our particular case was an OS4.0 system queueing files on a remote printer running OS3.2, but the problem also happens when the printer server is running OS3.4, as well as 4.0. In all cases the NFS mounted filesystem was from server machine running 3.4, I'm not sure if NFS mounting from an OS4.0 machine will produce the same effect. Presently (8/18/88) Sun considers the bug (#1007641) closed, but this may change since their only suggested workaround ("don't use lpr -s") does not offer a method to print files of over a megabyte in size. We've found that substituting the lpd4.0 with a copy of lpd3.4 seems to alleviate the problem, although we're not sure of what other side effects this may have. Also, there seems to be a new line in the /usr/spool/*/cf* file called the "S" line which contains two numerical arguments. I don't know what it means yet, but it may have something to do with the NFS-drop problem since the arguments turn strange on the files that get dropped. [[ The "S" line has always been in the control file specifications. It is used to hold "stat info" for symbolic link protection. The two numbers are the device and inode numbers of (I think) the original file. I believe this exists to handle the "-s" option and has nothing to do with NFS. --wnl ]] I hope that this info will save someone else from going through the hassles I went through before makeing the NFS-lpr connection. This way I can indirectly pay back SunSpots for all the information I've gleaned from this group! Thanks. Dave Benjamin Interleaf Cambridge, MA [[ To quote the lpr(1) manual page: "Note: the -s option only works on the local host (files sent to remote printer hosts are copied anyway), and only with named data files - it doesn't work if lpr is at the end of a pipeline." But there's also a problem with "lpr -s" and NFS. See the next message. --wnl ]] ------------------------------ Date: Thu, 18 Aug 88 17:19:52 CDT From: aydt%guitar.cs.uiuc.edu@a.cs.uiuc.edu (Ruth Aydt) Subject: lpr -s /tmp/file fails on 4.0 clients (with fix) There is a problem with lpr/lpd on 4.0 clients when you try to print a file with the -s option (use symlink to file rather than copying it into the spool area). lpr writes the device number as returned from stat to the control file with the S tag using printf %d format. However, this is a short and with the nfs-mounted filesystems it gets written as -32256 or some such number. When lpd then checks the S line in the cf file to make sure the device and inode numbers match those of the "current" file before printing the match fails and the job is rejected. The solution is to change lpr to use only relevant bits when building the cf file: (void) sprintf(buf, "%d %d", statb.st_dev&0xffff, statb.st_ino); For us this first turned up when some text-processing scripts submitted jobs that were rejected by lpd. The lpr -s was "hidden" within the scripts. Ruth Aydt Department of Computer Science University of Illinois at Urbana-Champaign ------------------------------ Date: 18 Aug 88 14:08:41 GMT From: oliveb!stpstn!aad@ames.arc.nasa.gov (Anthony A. Datri) Subject: pwd: getwd: can't open .. For the past week, I've had the following problem: <=>26<=>cd /usr/spool/uucp <=>27<=>pwd pwd: getwd: can't open .. It comes an goes. In the past, I've had problems with diskless clients doing that with NFS mounts, and it seemed to go away when I made the underlying mount point 777. We're running 3.2 on these machines. Any ideas? Almost everything works find -- cd works, but UNIPRESS Emacs complains when I try to run it out of there. I doubt that it's NFS related, especially since everything's up and it's still happening. -- Anthony A. Datri,SysAdmin,StepstoneCorporation,stpstn!aad [[ It is completely related to the permissions on the directories between the current one and the root. Try this experiment. "cd; mkdir foo; cd foo; mkdir bar; cd; chmod 111 foo; cd foo/bar; pwd" You will get "pwd: getwd: can't open ..". Why? Because "~/foo" is only scanable and not readable. You can cd to a subdirectory of "foo" because you explicitly named it ("x" permission on a directory means that you can access a file in that directory provided you know the name). So cd-ing there worked, but "pwd" needs to read all the directories back to the root. And it can't read "~/foo". Now do "cd; chmod 700 foo; rmdir foo/bar; rmdir foo" then go eat a grilled peanut butter sandwich. After you wash your hands and face, check the permissions on "/usr" and "/usr/spool" (and "/" too if you want). I don't know why it would "come and go" unless some evil daemon process keeps changing the permissions on you. Unipress Emacs has problems because it does the equivalent of a "pwd" (the library call "getwd") when it starts up. For more information on how "pwd" does its dirty work, look at v6n143 where it is discussed in the context of an NFS server crash. --wnl ]] ------------------------------ Date: Thu, 18 Aug 88 10:03:17 EDT From: ufnmr!ufnmr_1!gareth@bikini.cis.ufl.edu (Gareth J. Barker) Subject: Problem with biff and pseudo terminal ownership There's been a lot in Sun-Spots recently about ownership modes of pseudo terminals and problems with incorrect modes. I recently discovered that the 'biff' mail notification program doesn't work correctly under Suntools - typing 'biff' alone gives the response 'is y' or 'is n' (as it should), but trying to set biff (with 'biff y', for example) produces '/dev/ttyp3: Not owner' or similar. An 'ls' shows that /dev/ttyp3 is 'crw-rw-rw- 1 root reebok' - is this correct, and why does 'biff' need to know who owns the terminal anyway? This is not a big problem, as setting biff in '.login' makes everything work as expected (with mail notification going to the 'console' window under Suntools) and the shell mail/MAIL variables can be used if notification to other windows is needed, but I'd like to know what's going on. Thanks for any information, [[ This last paragraph was sent in a little later: --wnl ]] A correction to my last note - I'm not sure that setting the mail/MAIL variable in a Suntools window has any effect. A quick test has just shown no obvious response when I send mail either to myself or to 'root' (just an idea, seeing as root owns the window ...). Gareth J. Barker, University of Florida, Department of Radiology. BITNET : GJBARKER@UFFSC.BITNET INTERNET : ufnmr!gareth@BIKINI.CIS.UFL.EDU UUCP : ...gatech!uflorida!ufnmr!gareth [[ Didn't I explain this before? Oh well. When mail arrives for user X, the mail daemon checks all the terminals to see where X is logged in (I'm not sure if it checks /etc/utmp or just tty ownerships). If the owner execute bit is set (that is, mode 722 or "crwx..."), then notification is written to that terminal. Biff does its job by changing the permission bits. But since biff is not setuid, it can't change the permission bits if the user running biff doesn't own the terminal. This is the case for the pseudo ttys used by shelltools and cmdtools. Try using "mesg" to "turn messages off" in one of your windows. Same result for similar reasons. As for the cshell "mail" variable: it is more of a periodic synchrnous check which is completely separate from the asynchronous mailer daemon notification. That just tells the cshell to check for mail before it prompts you for another command. Read the csh(1) manual page for more information about that. --wnl ]] ------------------------------ Date: Thu, 18 Aug 88 18:52 GMT From: Royd Whittington <GEW%DLGM.DARESBURY.AC.UK@cunyvm.cuny.edu> Subject: passwd matching across YP domains? I want to have an entry in my /etc/passwd file that refers to a netgroup. The thing is, I want that netgroup to have a different domainname entry from my machine's domain in order to pick up the passwd entry from the other domain. At the moment it just ignores the entry in /etc/passwd. For example, /etc/passwd /etc/netgroup . . +@test_group | ----------> test_group (,,different_domain) Has anyone made this work? Thanks for any help! Royd Whittington, SERC Daresbury Laboratory, Warrington, UK. (Running SunOS 3.5) ------------------------------ Date: Thu, 18 Aug 88 07:49:05 HST From: Bob Cunningham <bob@kahala.hig.hawaii.edu> Subject: printcap vs. XON/XOF (SunOS4.0)? Could somebody please show me an example /etc/printcap entry for a simple serial port printer that actually works using XON/XOFF under SunOS4.0? The obvious DOESN'T seem to work; this drops the ends of even moderately-sized files: oki|local okidata printer:\ :lp=/dev/ttya:sd=/var/spool/oki:\ :br#9600:\ :ms=evenp,ixon,-echo:\ :lf=/var/adm/lpd-errs: ------------------------------ Date: Thu, 18 Aug 88 09:45:37 EDT From: ufnmr!ufnmr_1!mike@bikini.cis.ufl.edu (Michael D. Cockman) Subject: troff leaders? Could anyone tell us if the leader function ( \a ) in troff is functioning as shown in the 'Tabs, Leaders, and Fields' section of the troff manual ? If there is no software bug in troff, could someone show us how to use the leader function correctly ?? Reply to: Username: ufnmr!mike Node: bikini.cis.ufl.edu ------------------------------ End of SUN-Spots Digest ***********************