Usenet 1994 January

home *** CD-ROM | disk | FTP | other *** search

/ Usenet 1994 January / usenetsourcesnewsgroupsinfomagicjanuary1994.iso / sources / unix / volume25 / ispin / part14 < prev next >

Wrap

Text File | 1992-02-01 | 56.3 KB | 1,278 lines

Newsgroups: comp.sources.unix From: sir-alan!ispin!lbartz@iuvax.cs.indiana.edu (Larry Bartz) Subject: v25i125: Indianapolis Standard Printer Interface for Networked printers, Part14/15 Sender: sources-moderator@pa.dec.com Approved: vixie@pa.dec.com Submitted-By: sir-alan!ispin!lbartz@iuvax.cs.indiana.edu (Larry Bartz) Posting-Number: Volume 25, Issue 125 Archive-Name: ispin/part14 #! /bin/sh # This is a shell archive. Remove anything before this line, then unpack # it by saving it into a file and typing "sh file". To overwrite existing # files, type "sh file -c". You can also feed this as standard input via # unshar, or by typing "sh <file", e.g.. If this archive is complete, you # will see the following message at the end: # "End of archive 14 (of 15)." # Contents: ISPIN/doc/OLD-DOCS/README.rel_2.0 # ISPIN/doc/OLD-DOCS/rel_2.0.fixes ISPIN/misc/ISPIT.dr/ISPIT.mail # ISPIN/misc/ISPIT.dr/IMPLEMENT ISPIN/misc/SUID.dr/mail.msg # ISPIN/misc/wrdprfct.doc # Wrapped by socrates@indy6 on Tue Jan 28 15:27:18 1992 PATH=/bin:/usr/bin:/usr/ucb ; export PATH if test -f 'ISPIN/doc/OLD-DOCS/README.rel_2.0' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'ISPIN/doc/OLD-DOCS/README.rel_2.0'\" else echo shar: Extracting \"'ISPIN/doc/OLD-DOCS/README.rel_2.0'\" $7661 characters$ sed "s/^X//" >'ISPIN/doc/OLD-DOCS/README.rel_2.0' <<'END_OF_FILE' X X X X date : August 9, 1990 X X to : X X from : Systems Administrator, Information Systems Division X Indianapolis District Office X X subject: Indianapolis Standard Printer Interface (for Network printers) X ISPIN X production release 2.0 X X X X *** *** *** *** *** *** INTRODUCTION *** *** *** *** *** *** *** *** X X This is the current version of the production release of ISPIN. X The following instructions should be followed by those sites which X have not previously installed ISPIN. If you already have ISPIN X release 1.1 or 1.2 installed on your system, you should follow the X instructions in the file "UPDATEto2.0". X X Several files in the ISPIN/doc directory document the design, devel- X opment, testing, and implementation history of ISPIN. X X ISPIN has been successfully installed and tested on: X X AT&T 3B1 X AT&T 3B2600G X CCI 6/32 MP (CCI SysV) X Compaq with Intel 80286 CPU X (running XENIX) X IBM PC/AT with Intel 80386 CPU X (running XENIX) X Prime EXL 50 X Pyramid 90x X Pyramid 9810 X Pyramid MIS-4 X Sequent Balance B8 (DYNIX) X Sequent Symmetry (DYNIX) X Zilog Model 31 X Zilog Model 32 X Zilog Model 130 (Rev J OS) X X CDN (IRS X.25 network with async interface) X Teltone/Tellabs Switch X Gandolph Switch X David Switch X Mitron Switch X Equinox Switch X Neac Switch X X X Except for conditional compile statements required to resolve the X differences between Zeus 3.21's nq/dqueuer and System V's lp and X between System V UNIX and DYNIX, the source code is the same for X all target environments. X X *** *** *** *** *** *** THIS TRANSMITTAL *** *** *** *** *** *** *** X X The tape upon which we have transmitted the application was created X in cpio format. cd to a directory in which you want this release to X be written, then execute the command: X cpio -iudmvc < your_tape_drive X X The result of this cpio will be the creation of a directory named X ISPIN.rel_2.0. ISPIN.rel_2.0 will contain two subdirectories, namely X ISPI and ISPIN. X X X X *** *** *** *** *** *** ISPI subdirectory *** *** *** *** *** *** *** X X The ISPI directory contains the freshest version of our ISPI X application. ISPI is an application which serves as a FRONTEND X for the two queuer/spoolers. The ISPI application is fully self- X contained, with its own source code, documentation, etc. We sent X ISPI along because we have found it to be an extremely useful tool X to support the native queuer/spoolers. If you have any version of X ISPI previous to this one, please use this one as a replacement. The X ISPI included with this transmittal supersedes the ISPI which X accompanied the ISPIN.rel_1.2 release. X X ISPIN and ISPI are TWO SEPARATE AND DISTINCT APPLICATIONS. X X Neither depends upon the other for functionality. You may use ISPIN X without ISPI. You may use ISPI without ISPIN. X X X X *** *** *** *** *** *** ISPIN subdirectory *** *** *** *** *** *** *** X X ISPIN contains the source code, install documents, and install scripts X for the ISPIN application. The ISPIN application serves as a BACKEND X for either of two native queuer/spoolers, lp or dqueuer. We have X organized the directory ISPIN and its contents into subdirectories X using standard names "doc" (for documentation files), "h" (for header X files), "install" (where install scripts exist, and from which they X must be executed), "obj" (where the executables will be placed as X they are compiled), and "src" (where the source code resides). X X X X *** *** *** *** *** *** ISPIN documentation *** *** *** *** *** *** X X The documentation for this application is informal in style, and the X code is overflowing with comments. If you don't have time and inclin- X ation to read everything in the ISPIN.rel_2.0/ISPIN/doc directory, X AT LEAST READ: LPinstall.doc, NQinstall.doc, README.beta.3, design, X rel_2.0.fixes, and rtab. X X X X X *** *** *** *** *** *** ISPIN installation *** *** *** *** *** *** *** X X If you are installing this application under ZEUS 3.21 and its X nq/xq/dqueuer queuer family, install per NQinstall.doc. X X If you are installing this application under System V UNIX and X the lp spooler, install per LPinstall.doc. X X X X *** *** *** *** *** *** FEEDBACK, please *** *** *** *** *** *** *** X X We are maintaining a list of sites to which we have sent ISPI/ISPIN X and of contact persons at those sites. Unless you tell us, however, X we will have no idea whether or not you actually use these applications X in production. Our expectation is that you will, but we will not X presume it. Please confirm installation and usage of ISPI/ISPIN. We X prefer confirmation via mmdf mail sent to "ispin@indy6", but written X confirmation or a phone call will do. Confirmed users of ISPI/ISPIN X will receive all subsequent update and informational distributions. X We will notify users of updates via a mmdf electronic mail group, X "ispinusr@indy6". If possible, we would like your site to have an X user included in this mmdf group. An administrative mmdf address which X is aliased to an actual user at your end is preferred (i.e. X "sysadm@det1" instead of "jdoe@det1"). Please notify us of the mmdf X address at which you would like to receive ISPIN information. X X There is no bug-free software. Although the alpha site and beta site X testing has been extensive, we recognize the possibility of bugs which X have not yet been exposed. Please report them to us by telephone or by X sending mmdf mail to "ispin@indy6". We are committed to supporting this X application. X X If you install and use ISPI/ISPIN on hardware/software platforms which X are not listed in this memo, please let us know. This software is X intended to be portable, especially among machines which support a X System V UNIX environment. X X If you develop "rtab" entries which you feel others may find useful, X contact us. We'd like to be the clearinghouse for such information. X We have plans to implement a software distribution facility for X users of Indianapolis applications. The establishment of the mmdf X group "ispinusr@indy6" is the first step in the implementation of X this distribution facility. If you have any questions, please X feel free to contact us. X X X X X Larry Bartz X X END_OF_FILE if test 7661 -ne `wc -c <'ISPIN/doc/OLD-DOCS/README.rel_2.0'`; then echo shar: \"'ISPIN/doc/OLD-DOCS/README.rel_2.0'\" unpacked with wrong size! fi # end of 'ISPIN/doc/OLD-DOCS/README.rel_2.0' fi if test -f 'ISPIN/doc/OLD-DOCS/rel_2.0.fixes' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'ISPIN/doc/OLD-DOCS/rel_2.0.fixes'\" else echo shar: Extracting \"'ISPIN/doc/OLD-DOCS/rel_2.0.fixes'\" $8356 characters$ sed "s/^X//" >'ISPIN/doc/OLD-DOCS/rel_2.0.fixes' <<'END_OF_FILE' X X X ISPIN RELEASE 2.0 CHANGES X XHere, in no particular order of importance, is a listing of the differences Xin ISPIN.rel_2.0 X X XLOCK FILES - Take a look at the file ISPIN/doc/design. What I said there X about lock files still holds true. Lock files are a relatively X arcane and inefficient way for an application to control X access to resources. That's why previous releases of ISPIN X didn't employ lock files for its own purposes. Neither does X this release. The IQUEUER daemon still maintains its own X internal control of tty access among competing ISPIN processes. X X This release, however, implements lock files on top of, or X in addition to, IQUEUER's control. The purpose of these lock X files is to allow ISPIN to share tty resources with applications X such as cu(1), uucp(1), uugetty(M, on some machines), mmdf's X deliver, and others. Many of our customers' machines didn't X have enough tty ports to permit dedicating some to ISPIN and X other tty ports to other datacomm purposes. Hey, the customer X is king! You asked for it, you got it. You can get anything X you want at Alice's Restaurant (excepting Alice). Blah, blah, X blah, etc. (Really, this was a good idea all along. It just X took a few of you a while to beat it into my thick head.) X X ISPIN honors all "foreign" lock files, so if cu(1) or uucp(1) X is using the tty, ISPIN won't interfere. If more than one tty X is defined for ISPIN's use, ISPIN will move on to the next. X If the competitor has tied up the only tty which ISPIN knows, X we'll try again the next time the IQUEUER daemon loops around. X X The lock files ISPIN creates are intended to be unassailable X by competing "foreigners". Analysis and development of this X feature included a study of the lock file implementation on X five different UNIX hardware/software platforms. Guess what. X I saw five different lock file strategies. So much for any X attempt at compliance with the native schemes. We could have X complied with all five via liberal use of conditionally X compiled code, but what if a customer wants to run ISPIN on X a hardware/software combo which implements a sixth different X scheme? Can you say bye-bye to portability? I couldn't. X X ISPIN's lock files are implemented as directory files with X no contents and "zero" permissions all around. Alpha and X beta site testing has proven that this approach is universally X compatible with competing processes. This method is an X inelegant "brute force" technique, but it works. When we X all have POSIX-compliant boxes which implement a truly X universal lock file scheme, I'll do it the POSIX way. Until X then, we'll use the big hammer. X X XLP's "-m" - Here's the scenario: a user requests several files to be X printed via one lp(1) command, like so: X X lp -ddest -m file1 file2 file3 dir1/dir2/file4 file5 X X Now suppose that dir2's permissions were such that the user X "lp" could not get in there to read file4. If the user "lp" X can't read it, lp(1) can't print it. Since the user specified X the "-m" flag, he should receive mail from lp, reporting X that the print request failed. X X Under previous releases of ISPIN, user never got the message. X Now he will. X X XISPIN's log - All of ISPIN's log messages now show which tty was in use at X the time the log message was written. There are some phases in X the life of an ISPIN process in which the tty is undefined. If X the log message is written at such a point, there will be no X entry for "DEVICE". X X XRTAB flags - Two new flags for your rtab entries to further control the X behavior of the ISPIN process: X X -R X This flag takes no argument. If the -R flag is present, the tty through X which ISPIN communicates will be set so that it DOES NOT transmit the X "newline" (ascii "LF", or octal \012) as the newline/carriage return pair. X Note that the LF/LF-CR conversion is the default condition if -R is not X present. For most printers, DO NOT USE THIS FLAG. This flag is included X specifically for supporting the transmission of binary information (such X as would be the case with raster graphics). X X -T X This flag takes no argument. If the -T flag is present, tabs will be X expanded to spaces on output. If -R flag is present, -T has no effect. X X XIQ output - The utility iq (iqueuer status enquirer) is modified in some X subtle ways: X X TIME-IN - This column reflects the time the native spooler/ X queuer first invoked the ISPIN process. X X STATE-TIME - This column relects the time the ISPIN process X entered its current "STATE", or phase of execution. X X STATE - As before, this column shows the current phase of X the ISPIN process. A new state has been defined X and will be displayed when appropriate. X The state "QUITTING" will be displayed whenever X the ispin is in the process of an abnormal X termination. The state "DISCONNECTING" will be X displayed during normal termination and loop cycles. X X XISPI mods - As we've said many times before, ISPI is a companion application X to ISPIN. ISPI serves as a front-end menu for the native system's X spooler/queuer. ISPIN will work without ISPI, and vice versa. X Each, however, complements the native spooler/queuer, and the X trio (ISPI, native spooler/queuer, ISPIN) makes for a very X efficient and effective package. X X This latest release of ISPI incorporates yet another customer- X driven feature. Here is an excerpt form the latest ISPI.DOC: X X X As of 08/07/90 revision and later, the user may also set an environ- X mental variable, CONFIGDIR. The value of this variable will override the X compile-time-defined CONFIG_PATH. The purpose of CONFIGDIR is to allow X the Systems Administrator to set up various printer selection menus for X various groups of users. This may be necessary for reasons of: security (so X users of sensitive data may not inadvertently send jobs to insecure X printers [aren't they all?]); convenience (so users don't have to wade X through a menu filled with printers they never use); and practicality (only X so many printers will fit on one screen display page). X X Like so: X in C-shell: X set configdir=/z/bin/config/ex_configdir X setenv CONFIGDIR /z/bin/config/ex_configdir X X in Bourne shell: X CONFIGDIR=/z/bin/config/ex_configdir X EXPORT CONFIGDIR X X X X X X X/******************************************************************************/ XWell, that's the story on ISPIN.rel_2.0. I think the above-listed enhancements Xand fixes are sufficiently significant to warrant the minimal time you will Xspend upgrading from previous releases. If you haven't yet implemented ISPIN, Xthis release is what you've been waiting for. X XWith a lot of good customer involvement and a few extra keystrokes, we've made Xa good application even better. X X08/08/90 X X Internal Revenue Service X Indianapolis District Office X Information Systems Division X Operations Branch X X Larry Bartz END_OF_FILE if test 8356 -ne `wc -c <'ISPIN/doc/OLD-DOCS/rel_2.0.fixes'`; then echo shar: \"'ISPIN/doc/OLD-DOCS/rel_2.0.fixes'\" unpacked with wrong size! fi # end of 'ISPIN/doc/OLD-DOCS/rel_2.0.fixes' fi if test -f 'ISPIN/misc/ISPIT.dr/ISPIT.mail' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'ISPIN/misc/ISPIT.dr/ISPIT.mail'\" else echo shar: Extracting \"'ISPIN/misc/ISPIT.dr/ISPIT.mail'\" $19020 characters$ sed "s/^X//" >'ISPIN/misc/ISPIT.dr/ISPIT.mail' <<'END_OF_FILE' X XHave you seen the television commercial which says "Orange juice: It's not Xjust for breakfast anymore." ? X XTry this one. "ISPIN: It's not just for printing anymore." X XReally. Just as lp/lpsched is effectively a general purpose FIFO job scheduler, Xwhen you get right down to it, ISPIN is really a general purpose data transfer Xmechanism. It's just that up to now, we've only used it to transfer data to Xprinters. X XWhy couldn't we use ISPIN to transfer data to other systems or other media? X XWell, we can. I have. You can, too. Whether you want to (or not) is up Xto you. I'm not pushing this as any kind of world beater. X XThis message contains documentation, scripts, and programs to implement XIndianapolis Simple Protocol for Inter-system Transfers (ISPIT). ISPIT is Xa means of transferring data via the lp/ISPIN mechanism. ISPIT does not Xrequire modification to existing ISPIN code. X XI developed ISPIT for a particular purpose. The AT&T 3B1 which sits on my Xdesk has a 68MB hard disk but no decent magnetic media for backups. The Xcatridge tape drive is no longer in production, and putting that much data Xon floppies is beyond the scope of my patience. I wanted an automated full Xsystem backup, but I had no local hardware to do it. X XI tried several approaches via uucp but got no satisfactory result. Two Ximmutable factors of uucp data transfer got in my way. X XFirst is spooling. The uucico process on the receiving machine builds the Xincoming data in a temporary file in the spooling directory of your choice, Xthen copies that temporary file to the destination file you specify. The Xreceiving box must have ample disk capacity. Even if I specified the nine Xtrack drive on the receiving machine as the destination file (it works, by Xthe way), 35-40MB of free disk is pretty hard to come by on one of our Xproduction machines. (I'm sending my entire system as one cpio archive file Xwhich is compressed, then piped through a FIFO. I tried queueing up each Xseparate file, resulting in a mess of epic proportions). X XThe second limiting factor of uucp was the speed (lack thereof) of the Xuucp "G" protocol. My testing has shown that, in terms of raw throughput, XISPIT transfers data from system to system at twice the rate of uucp "G", X100% faster. X XI'll also cite a figure I'll call "virtual throughput", or apparent Xthroughput. If you compress a file, uucp it to another system, then Xuncompress it on receipt, you've attained a faster file transfer. The com- Xpressed data is "data", so must travel through media (CDN, Tellabs, etc) Xwhich does not interpret software flow control. Fine for uucp. X XISPIN/ISPIT depends upon accurate software flow control, so compressed Xdata must be converted to printable ascii before transmission. See the Xdo_ispit.sh script, below. The btoa (binary to ascii) process expands input Xdata by roughly 30% on output. X XEven with this "handicap", ISPIT's virtual throughput is 70% faster than Xuucp "G". X XWhat of error checking, you say? Uucp "G" protocol may be slow, but at least Xit checks to make sure the data gets from end to end in one piece. X XI haven't noticed ISPIN dropping a lot of characters, even on big print jobs. XThere certainly isn't any error checking going on there. ISPIT's success X(if it has any) is heavily dependent on fast, modern I/O subsystems at Xboth ends (no Model 31 Zilogs, please), and a clean transport media which Xsupports a reliable cascading buffered software flow control (CDN, Tellabs, Xmaybe others). X X X(DISCLAIMER) XPlease don't get me wrong. I'm not proposing that we use this method as a Xreplacement for other valid data transfer mechanisms such as uucp, kermit, Xand others. But sometimes those others can't do what we need to do. Maybe Xthere are a few niches where ISPIT might fit. X XBelow is the output from "more * > ISPIT.doc" from within a directory which Xcontains all files necessary to implement ISPIT. X X:::::::::::::: XIMPLEMENT X:::::::::::::: XISPIT (Indianapolis Simple Protocol for Inter-system Transfers) X XIMPLEMENTATION NOTES X X XDON'T BOTHER IMPLEMENTING THIS ON 16-BIT MACHINES SUCH AS ZILOG MODEL 31 (or Xeven the so-called "32-bit Upgrade"). Their I/O systems are not adequate for Xthe reliable data flow ISPIT requires. X XOn dual universe machines, do everything from att universe. X XAlso, if any of the machines you use in this scenario do not support the Xcompress/uncompress/btoa/atob/zcat family of executables, you'll need 'em. XThe Pyramids and Sequents have these in the ucb universe. "Pure" System V's Xlike the AT&T boxes and Zilog 130's don't. I have source code which compiles Xand executes fine on our Zilog 130 and my AT&T 3B1. If you want it, ask me. X X XSOURCE MACHINE: X X First, the source machine must have ISPIN.rel_2.1 or later. X X do_ispit.sh - check this script for correct pathnames to X your executables such as compress, btoa, X cpio, and lp. On dual universe machines, the X Berzerkely commands will have to be preceded X by "ucb". X X - Put a copy of this script in a directory which X is in your path, renaming it as "ispit", such X as: X cp do_ispit.sh /usr/local/bin/ispit X X or X X cp do_ispit.sh /z/bin/ispit X X rtab.entry - Edit this rtab entry to make it valid for your X situation. You'll have to change: X X - the first field, for queue member name. X I suggest the node name of the destination X system. X X - the second field, the device(s). Use the X same one(s) as you do for ISPIN printing. X X - the connect chat fields. If you're going X straight through CDN to the remote system, X you'll only have to change the address. X If negotiating other switches, adjust X accordingly. X X Unlike uucp, ISPIT REQUIRES XON/XOFF X FLOW CONTROL! So make sure all the X network links you go through support it. X From the source machine, go out via ports X you use for printing. Log into the destina- X tion box through channels just like a X user (person). Don't use "no flow" uucp X file transfer channels. X X - Notice that this rtab entry results in a X user named "ispit" logging onto the remote X machine with password of "notvalid". Change X these to correspond with what you set up X on the destination system. X X - The rest of the rtab entry, including X "FILE?", "OK", etc MUST STAY AS IS! It X talks to the ispit.sh script on the dest- X ination box. X X - Add the rtab entry to your rtab file. X X lpadmin stuff - Add a queue member whose name matches the X first field of the rtab entry, using the X special new ispit interface script included X here. Such as: X X lpshut X X lpadmin -pINDY_5 -v/dev/null -iinterface X X accept INDY_5 X X enable INDY_5 X X lpsched X X X XDESTINATION MACHINE: X X Add a user to the system. Use the passwd.entry file included X here as a model. This user's logname and password must match X corresponding fields in the rtab entry you set up on the X source machine. X X Compile the program "eot_trap" from the source I've provided. X X cc -o eot_trap eot_trap.c X X Put the executable "eot_trap" and the script "ispit.sh" in X the new user's home directory. Make they are executable by X the new user. X X Test this new user by logging as him (OK, Kim, as her). X ispit.sh will prompt you for a file name. Give a name and X a carriage return. Type anything you want, ending with a X <ctrl>d. When you log in again as yourself, the file you X specified should consist of all you typed. X X XUSAGE: X X The script do_ispit.sh contains some internal documentation X on its usage. X X If you transfer data via the do_ispit.sh (or "ispit") script, X the file on the destination machine will be named as you X specify. The file's format is as a cpio archive which is X compressed, then converted by btoa. To disassemble this mess X on the destination machine, you can: X X cat filename|atob|zcat|cpio -iudmcv X X on a dual universe box, from att universe: X cat filename|ucb atob|ucb zcat|cpio -iudmcv X XAPPLICATION: X X I've even had good luck specifying the nine track drive as X the destination device (Zilog 130, not Sequent B8). Here's X how I send the entire contents of my AT&T 3B1's 68MB hard X drive (approx 64% full) to a nine track tape on a Zilog X Model 130 named INDY_5: X X #!/bin/sh X cd / X /usr/local/bin/ispit . /dev/rmiop0 INDY_5 X X When I want to check the correctness of my backup tape, X I login to INDY_5, mount my tape, and: X X cat /dev/rmiop0|atob|zcat|cpio -itcv > tape.list X X You could easily modify the ispit.sh script on the destina- X tion machine so that eot_trap sends its output through atob, X and zcat, and then to the specified output file. Then you'd X be dealing with a simple cpio archive at the destination. X X Come to think of it, you could modify any of this to do X anything you might dream up! X:::::::::::::: XREADME X:::::::::::::: X Xrtab.entry - see do_ispit.sh. Notice that this rtab entry requires that X this queue member be called by an lp(1) command which X must include a sane value for the "-o" flag. X The arg to the "-o" is used by ISPIN at the point in the X rtab where "\U" appears. X Note also the use of the "-R" flag for really raw output. X Xdo_ispit.sh - the shell script which is executed by the source system. X X The lp command line contains some interesting features. X X The string passed as the argument to the "-o" flag is X actually made up of two arguments which are delimited by X the "@" symbol. These two arguments are processed by X the special interface script on the source system. X X The first "argument" to lp's "-o" flag is processed X by ispin and inserted in a "sent" network negotiation X string where a "\U" appears. On the destination system, X ispit.sh picks this up as the name of the output file. X X Also notice that the second "argument" to the "-o" flag X is /tmp/$nowfile. This file is a named pipe, created "on X the fly". The lp/interface file (special for this purpose) X reads from the named pipe which was created and filled by X do_ispit.sh. X X The "filename" argument to the lp command is /etc/passwd. X This file is not actually transmitted by the ispit/ispin X process. We use /etc/passwd as a placeholder on the X command line since lp will not queue a request without a X valid non-null filename. X X do_ispit.sh requires three arguments: X X - First is the name of the source file to be transferred. X X - Second is the destination file name on the destination X system. This argument should be a full path name. All X elements of the path must exist. The destination directory X must be writable by the user "ispit" on the remote system. X If this argument is less than a fully qualified path, the X remote system will interpret the argument relative to X ispit's home directory. X X - Third is the (lp) queue member which will process the X request. X X On dual universe machines, the calls to "compress", "btoa", X "cut", and "lp" may have to be preceded by the appropriate X universe call. Path names to commands may also require X adjustment. X Xinterface - This shell script is the lp interface file for ispit on X the source system. X Xeot_trap.c * - source for a filter program which watches for an ascii X "EOT" character. eot_trap filters the input X on the receiving system. The source system's rtab entry X for the "ispit" queue member sends EOT as the first action X of the ispin quit sequence. eot_trap exits, ending the X login session, so ispit exits. This dependence upon X EOT makes it IMPERATIVE that all output be converted by X btoa(1). See do_ispit.sh. X Xispit.sh * - see passwd.entry, below. This is the script which is invoked X upon login by and for the ispit user. Notice that ispit.sh X expects (actually requires) that a variable X "OF" is assigned a sane value. This value is passed at login X time by the user "ispit". See rtab.entry for the "\U". X X NOTE: ispit.sh calls the executable "eot_trap", assuming X that the executable is located in ispit's home X directory. X Xpasswd.entry * - this is the /etc/passwd entry for the ispit user on the X destination machine. The last field of the entry is the full X path name of the ispit.sh script. Change the location of X ispit's home directory as you see fit. X X X X * - denotes this file, data, or program exists on the X destination machine. All else exists on the source X machine. X:::::::::::::: Xdo_ispit.sh X:::::::::::::: X#!/bin/sh X# do_ispit.sh 08/07/91 LSB X# When you install this in an accessible path, call it "ispit". This will X# assure compatibility with possible future executables. X# requires three args: X# $1 is file to be xferred: could be a directory, could be more than one X# file if all enclosed in double quote X# X# $2 is destination file name (full path on remote system, or relative to X# ispit's home directory on the remote system) X# X# $3 is the lp queue member to use (for destination system) X# X# so: X# ispit source_file destination_file destination_system_queue_member X# X# examples: X# X# send testfile to /tmp/testfile on queue indy6: X# ispit testfile /tmp/testfile indy6 X# X# send directory UUGETTY and all contents to /tmp/UUG.cp.Z.a on queue indy6: X# ispit UUGETTY /tmp/UUG.cp.Z.a indy6 X# X# send all files whose names contain "ind" to /tmp/ind.cp.Z.a on queue indy6: X# ispit "`ls *ind*`" /tmp/ind.cp.Z.a indy6 X# X# nowfile (not an argument) is the name of the named pipe in /tmp we'll use X# X# Xnowfile=`/bin/date +%m%d%y`.`/bin/date +%H%M%S` Xecho "ISPIT REQUEST" >> /tmp/ispit.log Xdate >> /tmp/ispit.log Xecho " user: "`id` >> /tmp/ispit.log Xecho " source: "$1 >> /tmp/ispit.log Xecho " destination: "$2 >> /tmp/ispit.log Xecho " queue: "$3 >> /tmp/ispit.log Xecho " pipe: /tmp/"$nowfile >> /tmp/ispit.log X/bin/rm -f /tmp/$nowfile X/etc/mknod /tmp/$nowfile p X/bin/chmod 666 /tmp/$nowfile Xfind $1 ! -type p -print|/bin/cpio -oc|/usr/local/bin/compress|/usr/local/bin/btoa > /tmp/$nowfile & X/usr/bin/lp -d$3 -o$2@/tmp/$nowfile /etc/passwd X# Yes, I do mean /etc/passwd! It's here as a necessary argument to fool lp. X:::::::::::::: Xeot_trap.c X:::::::::::::: X#include <stdio.h> X Xmain() X{ X int c; X X while ((c = getc(stdin)) != EOF) X { X if(c == 4) X { X exit(0); X } X X putc(c,stdout); X } X} X:::::::::::::: Xinterface X:::::::::::::: X#!/bin/sh X# 08/07/91 LSB X# This interface script usable for ISPIT file transfer protocol Xecho "ISPIT FILE TRANSFER" >> /tmp/ispit.log Xdate >> /tmp/ispit.log Xecho " lp job id : "$1 >> /tmp/ispit.log Xecho "requesting user : "$2 >> /tmp/ispit.log Xecho " destination : "`echo $5|/usr/bin/cut -d@ -f1` >> /tmp/ispit.log Xecho " pipe : "`echo $5|/usr/bin/cut -d@ -f2` >> /tmp/ispit.log Xexec /usr/spool/lp/ISPIN/ispin `/bin/basename $0|/usr/bin/tr -d "[\012*]"` $1 $2 "" 1 `echo $5|/usr/bin/cut -d@ -f1` `echo $5|/usr/bin/cut -d@ -f2` X:::::::::::::: Xispit.sh X:::::::::::::: X#!/bin/sh X# ispit.sh 080791 LSB X# written to run as the user ispit's default shell (last field in passwd entry) X# stty call is of "att" type, so must be run under att on dual universe boxes Xecho "OUTPUT_FILE?" Xread OF Xecho "OK" Xstty raw ixoff -echo hupcl brkint Xecho "ISPIT START" >> ispit.log Xdate >> ispit.log Xecho $OF >> ispit.log X./eot_trap > $OF Xdate >> ispit.log Xstty sane Xexit X:::::::::::::: Xpasswd.entry X:::::::::::::: Xispit:9CKY7K.On.kVs:437:4::/zb/ispit:/zb/ispit/ispit.sh X:::::::::::::: Xrtab.entry X:::::::::::::: X# This rtab entry is usable only for the ISPIT file transfer protocol. XINDY_5;/dev/tty000;19200;-L;-R;-Bcongestion;-Bremote dte;-Q\004\L\L\K\dclr\r\d;-DConnection cleared;-Dad>:;-Dogin:;ad>-\p\K\pclr\r\d\020\pclr\r\d-ad>;clr\r\r\d;ad>-\p\K\pclr\r\d\020\pclr\r\d-ad>;set recall=none\r\pset break-handling=send-break+recall\r\pset pad-flow=on\r\pset terminal-flow=on\r\p;ad>-\r\r\d-ad>;\pconnect\s00000099999999\r\d;Opened;\r\r\d;ogin:-\r\d-ogin:;ispit\r\d;assword:;notvalid\r\d\d;FILE?;\U\r\d\d;OK;\w;;; X#ispit to INDY_5 X# END_OF_FILE if test 19020 -ne `wc -c <'ISPIN/misc/ISPIT.dr/ISPIT.mail'`; then echo shar: \"'ISPIN/misc/ISPIT.dr/ISPIT.mail'\" unpacked with wrong size! fi # end of 'ISPIN/misc/ISPIT.dr/ISPIT.mail' fi if test -f 'ISPIN/misc/ISPIT.dr/IMPLEMENT' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'ISPIN/misc/ISPIT.dr/IMPLEMENT'\" else echo shar: Extracting \"'ISPIN/misc/ISPIT.dr/IMPLEMENT'\" $6949 characters$ sed "s/^X//" >'ISPIN/misc/ISPIT.dr/IMPLEMENT' <<'END_OF_FILE' XISPIT (Indianapolis Simple Protocol for Inter-system Transfers) X XIMPLEMENTATION NOTES X X XDON'T BOTHER IMPLEMENTING THIS ON 16-BIT MACHINES SUCH AS ZILOG MODEL 31 (or Xeven the so-called "32-bit Upgrade"). Their I/O systems are not adequate for Xthe reliable data flow ISPIT requires. X XOn dual universe machines, do everything from att universe. X XAlso, if any of the machines you use in this scenario do not support the Xcompress/uncompress/btoa/atob/zcat family of executables, you'll need 'em. XThe Pyramids and Sequents have these in the ucb universe. "Pure" System V's Xlike the AT&T boxes and Zilog 130's don't. I have source code which compiles Xand executes fine on our Zilog 130 and my AT&T 3B1. If you want it, ask me. X X XSOURCE MACHINE: X X First, the source machine must have ISPIN.rel_2.1 or later. X X do_ispit.sh - check this script for correct pathnames to X your executables such as compress, btoa, X cpio, and lp. On dual universe machines, the X Berzerkely commands will have to be preceded X by "ucb". X X - Put a copy of this script in a directory which X is in your path, renaming it as "ispit", such X as: X cp do_ispit.sh /usr/local/bin/ispit X X or X X cp do_ispit.sh /z/bin/ispit X X rtab.entry - Edit this rtab entry to make it valid for your X situation. You'll have to change: X X - the first field, for queue member name. X I suggest the node name of the destination X system. X X - the second field, the device(s). Use the X same one(s) as you do for ISPIN printing. X X - the connect chat fields. If you're going X straight through CDN to the remote system, X you'll only have to change the address. X If negotiating other switches, adjust X accordingly. X X Unlike uucp, ISPIT REQUIRES XON/XOFF X FLOW CONTROL! So make sure all the X network links you go through support it. X From the source machine, go out via ports X you use for printing. Log into the destina- X tion box through channels just like a X user (person). Don't use "no flow" uucp X file transfer channels. X X - Notice that this rtab entry results in a X user named "ispit" logging onto the remote X machine with password of "notvalid". Change X these to correspond with what you set up X on the destination system. X X - The rest of the rtab entry, including X "FILE?", "OK", etc MUST STAY AS IS! It X talks to the ispit.sh script on the dest- X ination box. X X - Add the rtab entry to your rtab file. X X lpadmin stuff - Add a queue member whose name matches the X first field of the rtab entry, using the X special new ispit interface script included X here. Such as: X X lpshut X X lpadmin -pINDY_5 -v/dev/null -iinterface X X accept INDY_5 X X enable INDY_5 X X lpsched X X X XDESTINATION MACHINE: X X Add a user to the system. Use the passwd.entry file included X here as a model. This user's logname and password must match X corresponding fields in the rtab entry you set up on the X source machine. X X Compile the program "eot_trap" from the source I've provided. X X cc -o eot_trap eot_trap.c X X Put the executable "eot_trap" and the script "ispit.sh" in X the new user's home directory. Make they are executable by X the new user. X X Test this new user by logging as him (OK, Kim, as her). X ispit.sh will prompt you for a file name. Give a name and X a carriage return. Type anything you want, ending with a X <ctrl>d. When you log in again as yourself, the file you X specified should consist of all you typed. X X XUSAGE: X X The script do_ispit.sh contains some internal documentation X on its usage. X X If you transfer data via the do_ispit.sh (or "ispit") script, X the file on the destination machine will be named as you X specify. The file's format is as a cpio archive which is X compressed, then converted by btoa. To disassemble this mess X on the destination machine, you can: X X cat filename|atob|zcat|cpio -iudmcv X X on a dual universe box, from att universe: X cat filename|ucb atob|ucb zcat|cpio -iudmcv X XAPPLICATION: X X I've even had good luck specifying the nine track drive as X the destination device (Zilog 130, not Sequent B8). Here's X how I send the entire contents of my AT&T 3B1's 68MB hard X drive (approx 64% full) to a nine track tape on a Zilog X Model 130 named INDY_5: X X #!/bin/sh X cd / X /usr/local/bin/ispit . /dev/rmiop0 INDY_5 X X When I want to check the correctness of my backup tape, X I login to INDY_5, mount my tape, and: X X cat /dev/rmiop0|atob|zcat|cpio -itcv > tape.list X X You could easily modify the ispit.sh script on the destina- X tion machine so that eot_trap sends its output through atob, X and zcat, and then to the specified output file. Then you'd X be dealing with a simple cpio archive at the destination. X X Come to think of it, you could modify any of this to do X anything you might dream up! END_OF_FILE if test 6949 -ne `wc -c <'ISPIN/misc/ISPIT.dr/IMPLEMENT'`; then echo shar: \"'ISPIN/misc/ISPIT.dr/IMPLEMENT'\" unpacked with wrong size! fi # end of 'ISPIN/misc/ISPIT.dr/IMPLEMENT' fi if test -f 'ISPIN/misc/SUID.dr/mail.msg' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'ISPIN/misc/SUID.dr/mail.msg'\" else echo shar: Extracting \"'ISPIN/misc/SUID.dr/mail.msg'\" $5839 characters$ sed "s/^X//" >'ISPIN/misc/SUID.dr/mail.msg' <<'END_OF_FILE' XThe UNIX System V, release 2 lp subsystem incorporates a peculiar security Xmechanism. Unless the administrative user "lp" can read a file, it cannot Xbe printed. This effectively requires that any file which is to be printed Xmust have read permissions for the world. X XFor many sites/users/systems administrators this is an unacceptable situation. XFiles and directories are created with restrictive permissions for valid Xsecurity reasons. Even temporarily opening perms on a file to permit printing Xrepresents a potential security breach. X XIn 1987, Centel's Perry Mathis devised a Bourne shell script called "lpc". XIt was designed to be run in place of /usr/bin/lp (after the "real" lp had Xbeen moved to another name). lpc allowed any file the calling user could Xread to be printed (by catting the file to a pipe to the "real" lp command). X XMany sites use the lpc script and find it acceptable. It has a side-effect Xwhich many find unacceptable. When the "real" lp is asked to print the Xstandard input (from the "cat" command), lp MUST make a temporary copy of Xthe input so that it can be printed when lpsched allows. This copy is created Xin the /usr/spool/lp/request/printername directory. This process eats cpu Xand disk access time. Even worse, if the file(s) is(are) REALLY BIG, the Xcopying activity can run the /usr filesystem down to zero free blocks! X XCatch-22. X XUntil now. X XI've recently had the pleasure of working with a box which runs XUNIX SysV release 3. The lp subsystem under release 3 employs a more logical, Xmore flexible security approach. Under Sys V release 3, the daemon process Xlpsched runs as setuid to root. Depending upon the particular flavor of XSysVrel3, some of the other related process also run as superuser. At the Xend of the print request/spooling/scheduling/printing process, the program Xwhich actually reads and prints the file is executed under the identity Xof the user who requested the job. X XThe net effect is that the user is allowed to print any file which they Xcan read, but no extra disk copy need be created. This flexibility is Xenabled while still retaining the security of not allowing users to print Xfiles which they cannot read. X XI've devised a means by which you can make your System V release 2 lp Xsubsystem emulate release 3's behavior. Two files are attached below. X XFirst is a Bourne shell script which changes lp's uid to 0 and sets correct Xfile permissions on all lp subsystem executables. X XSecond is a code fragment with which you can patch your XISPIN.rel_2.1/ISPIN/src/ISPIN.c. This patch is good only for a 2.1 ISPIN! X XDO NOT USE THE FIRST WITHOUT THE SECOND! If you do, you'll totally subvert Xall security relative to printing. X X--- cut here --- cut here --- cut here --- cut here --- cut here --- cut here X#!/bin/sh X# lp_perms.sh 05/13/91 LSB IRS Indianapolis X# X# This routine sets up a System V, release 2 lp subsystem to behave X# similar to a System V, release 3 lp subsystem with respect to file X# access. It all runs as setuid to superuser. Under this setup, it X# is CRITICALLY IMPORTANT to have an interface program which will X# setgid and setuid back to the user who requested the job, thus X# maintaining security. X# X# After you run this routine, MAKE SURE that the user "lp" does have X# a password which is unknown to all. Anyone who logs on as lp will X# be superuser. X# X# After implementation of this routine and the setgid/setuid interface, X# users will be able to print any file which they themselves can read. X# X# This is the code fragment for ispin which does the trick: X# #ifdef LP X# /* If this process is running as setuid to root (uid == 0), X# then setuid to the requesting user. 05/09/91 LSB */ X# if(geteuid() == 0) X# { X# setgid(pass->pw_gid); X# setuid(pass->pw_uid); X# } X# #endif X# ISPIN.rel_2.1/src/ISPIN.c can be patched with this code fragment after line X# 1269. All ISPIN releases after ISPIN.rel_2.1 will contain this code. X# X# change lp's uid to 0 Xset UID Xset LYNE XUID=`grep lp\: /etc/passwd|cut -d\: -f3` XLYNE=`grep -n lp\: /etc/passwd|cut -d\: -f1` Xsed $LYNE\ s/\:$UID\:/\:0\:/g /etc/passwd > /tmp/passwd.TMP Xcp /etc/passwd /tmp/passwd.sav Xcp /tmp/passwd.TMP /etc/passwd X# X# lp administrative stuff, superuser only Xchown lp /usr/lib/lpsched Xchmod 700 /usr/lib/lpsched Xchown lp /usr/lib/lpshut Xchmod 700 /usr/lib/lpshut Xchown lp /usr/lib/lpmove Xchmod 700 /usr/lib/lpmove Xchown lp /usr/lib/accept Xchmod 700 /usr/lib/accept Xchown lp /usr/lib/reject Xchmod 700 /usr/lib/reject X# X# lp user stuff Xchown lp /usr/bin/lp Xchmod 4711 /usr/bin/lp Xchown lp /usr/bin/lpstat Xchmod 4711 /usr/bin/lpstat Xchown lp /usr/bin/cancel X# The following lets any user cancel any job, same as before. X# If you don't like that, change 4711 to 700. Then only superuser can cancel. X# Same goes for enable and disable. Xchmod 4711 /usr/bin/cancel Xchown lp /usr/bin/enable Xchmod 4711 /usr/bin/enable Xchown lp /usr/bin/disable Xchmod 4711 /usr/bin/disable X# end of lp_perms.sh X--- cut here --- cut here --- cut here --- cut here --- cut here --- cut here X XThe following code module is to be inserted after line 1269 of the file XISPIN.rel_2.1/ISPIN/src/ISPIN.c X XAfter you have inserted this code fragment, cd to X X ISPIN.rel_2.1/ISPIN/install X Xthen X X make ../obj/ispin X Xthen (when there are no active print jobs) X X cp ../obj/ispin /usr/spool/lp/ISPIN/ispin X X X--- cut here --- cut here --- cut here --- cut here --- cut here --- cut here X#ifdef LP X /* If this process is running as setuid to root (uid == 0), X then setuid to the requesting user. 05/09/91 LSB */ X if(geteuid() == 0) X { X setgid(pass->pw_gid); X setuid(pass->pw_uid); X } X#endif X--- cut here --- cut here --- cut here --- cut here --- cut here --- cut here END_OF_FILE if test 5839 -ne `wc -c <'ISPIN/misc/SUID.dr/mail.msg'`; then echo shar: \"'ISPIN/misc/SUID.dr/mail.msg'\" unpacked with wrong size! fi # end of 'ISPIN/misc/SUID.dr/mail.msg' fi if test -f 'ISPIN/misc/wrdprfct.doc' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'ISPIN/misc/wrdprfct.doc'\" else echo shar: Extracting \"'ISPIN/misc/wrdprfct.doc'\" $4179 characters$ sed "s/^X//" >'ISPIN/misc/wrdprfct.doc' <<'END_OF_FILE' XFrom uucp Mon Feb 4 23:51 EST 1991 X>From lbartz Mon Feb 4 23:35:53 1991 remote from localhost XReceived: by localhost.UUCP (4.12/4.7) X id AA00203; Mon, 4 Feb 91 23:35:53 est XMessage-Id: <9102050435.AA00203@localhost.UUCP> XReceived: from cr07 by indy6.indy6.irs.gov id ab00186; 4 Feb 91 23:35 EST XReceived: from tis by cr07.cr07.irs.gov id ab24006; 4 Feb 91 22:14 EST XReceived: from wr by no1.irs id ac21732; 4 Feb 91 22:18 EST XReceived: from sea by wr1.irs id aa00381; 4 Feb 91 9:58 PST XDate: 4 Feb 91 9:45:27-EST (Mon) XFrom: localhost!"Mwilson at sea"@wr1.ARPA XTo: ispinusr@indy6.ARPA XStatus: RO X X X X X X X INTERFACING ISPIN WITH WORDPERFECT 5.0 AND SCO UNIX V/386 X X X Larry "The Supreme Print Commander" Bartz asked me to pass this X along in the hopes that it may be of some help to the folks X who are running WordPerfect 4.2 on their Sequents. X X X The problem we encountered was that Wordperfect assumes that X the "interface" program for each printer is a Bourne shell X script. When you use "wpport" to let WP know about the printer X it tries to add a few lines of its own to the script in X order to format the document for the printer type. If the X interface just happens to be "ispintrfce" (a C executable) X good ole WP can't do its thing. After a few phone calls to X the "Bartzman" we've come up with the following. X X 1. De-install the printer that you want to use. X X 2. Create a file that will hold the formatted WP output X file for each printer. X X touch /usr/tmp/yourprinter.tmp X X or any other directory that suits you. Remember, references X to "yourprinter" means the name you give your printer. X X 3. Put this pseudo interface file in your lp "interface" X directory. We call it "wpface" X X---------------------------CUT HERE----------------------------------------- X# wpface Xfor option in $5 Xdo X if [ $option = WP50 ] X then X ### Make sure to change PRINTER.tmp to your "queuename".tmp X # and have correct directory references X cat /dev/null > /usr/tmp/PRINTER.tmp X /u1/WP50/sbin/wpp50 < $6 X ### Make sure to change PRINTER.tmp to your "queuename".tmp X exec /usr/spool/lp/ISPIN/ispin `/bin/basename $0|/usr/bin/tr -d "[\012*]"` $1 $2 "" $4 option /usr/tmp/PRINTER.tmp X fi Xdone Xif [ "$option" = "" ] X then X exec /usr/spool/lp/ISPIN/ispin `/bin/basename $0|/usr/bin/tr -d "[\012*]"` $1 $2 "" $4 "" $6 X fi X---------------------------CUT HERE------------------------------------------ X X X Don't worry about the contents of this just yet. WP will stomp on it X and put its junk in when using "wpport" X X 4. Time to install your printer to the "lp" system X "cd" to your interface directory X replace the "lpadmin" command options in your ispin X documentation with the following: X "/usr/lib/lpadmin -pyourprinter -iwpface -v/usr/tmp/yourprinter.tmp X X Remember that file you created in step 2. That's your "-v" X flag option X Start your scheduler and do your "accepts" and "enables" X X 5. Install your printer for WP using "wpport". We've found X that WP will try to set the line settings using "stty" X which it can't do because the printer isn't connected X to particular tty. You defeat this by defining the line X type as parallel (Yes it works!!!!!!!!!) X X 6. Go back to your lp "interface" directory. Copy "wpface" X over "yourprinter" interface file created in step 4. X Edit the new "yourprinter" file replacing the references X to "PRINTER" with "yourprinter" name. X X 7. That's all. Hope this helps out. I'll be glad to pass X along any enhancements or changes that we come across. X X Mike Wilson X S.A. Sys Supp Br. X Seattle DO ISD X X END_OF_FILE if test 4179 -ne `wc -c <'ISPIN/misc/wrdprfct.doc'`; then echo shar: \"'ISPIN/misc/wrdprfct.doc'\" unpacked with wrong size! fi # end of 'ISPIN/misc/wrdprfct.doc' fi echo shar: End of archive 14 $of 15$. cp /dev/null ark14isdone MISSING="" for I in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ; do if test ! -f ark${I}isdone ; then MISSING="${MISSING} ${I}" fi done if test "${MISSING}" = "" ; then echo You have unpacked all 15 archives. rm -f ark[1-9]isdone ark[1-9][0-9]isdone else echo You still need to unpack the following archives: echo " " ${MISSING} fi ## End of shell archive. exit 0