Nebula

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

/ Nebula / nebula.bin / Newsletters / GEnieUnixNews / unxnl-05.92 < prev next >

Wrap

Text File | 1993-01-27 | 43KB | 948 lines

_ _ _ _ _ _ // // //| // // \// N E W S //_// // |// // /\\ Vol 3, Issue 5 - May 1992 R o u n d T a b l e (tm) Items of interest to participants of the GEnie Unix RoundTable The RoundTable SysOps are: Andy Finkenstadt....ANDY Rick Mobley.........LRARK Gary Smith..........GARS Brian Riley.........DELPHI Mike Nolan..........M.NOLAN All Unix SysOps.....UNIXSYSOPS$ We strongly encourage you to contact any or all of us if you have -ANY- comments or suggestions. This is -YOUR- RoundTable. We are here to make your participation as pleasant and beneficial as possible. ED: editor notes - Previews of coming attractions (GARS) Gary Smith == Some of you have noticed the software libraries suddenly expanded from twenty-eight named libraries to thirty-one with an additional eight reserved libraries. Some even noticed the reference to SLA file sections. SLA, Stereolithographic Association, has elected to make the GEnie Unix RoundTable their home. This is a group that uses some very sophisticated equipment, frontended by a SGI Unix platform, to output solid engineering models in a very short time frame, thus reducing dramatically the time lapse from engineering design to prototype. It is exciting to see Unix blazing new paths in such a manner, and we look forward to their presence in our online family. There are other groups that will be joining us very shortly that will add spice to the online discussions and grant each of us an opportunity to gain insight through the activities of others. We have already seen the NeXT users demonstrate how a dedicated group, given an outlet like the GEnie Unix RoundTable can create a solid support and resource center that draws others to them. It excites me, as a SysOp and observer, to see limits removed and opportunities rise in their place. Stick around folks. The show has only begun. NUCON 22 in Atlanta, a report (M.NOLAN) Mike Nolan ============================= The 22nd annual NCR User's Conference was also the first since NCR's acquisition by AT&T. Although there were a few signs of AT&T's presence, for the most part it was business as usual. A reference to NCR as the 'Network Computer Resource of AT&T', which surfaced a few months ago, has apparently been dropped. There were a few NCR badges with New Jersey addresses, and some of them were ex-AT&T people, but an equal number were new hires. On the show floor, the most visible signs were the StarServer network products, buried along the back wall of the NCR booth, and the Safari notebook systems, now renumbered into the NCR 3100 series but some still in cases with the Safari logo on it, along with an NCR logo. The latest of these, the 3170, features a FingerPointMouse, sort of a built-in stickless joystick operated with the index or middle finger. Although it is primarily an MS-DOS machine, the 3170 is interesting because it comes with a flashcard port instead of a floppy drive, which makes it incredibly thin and light, even with a battery and an 80MB hard drive it was still under five pounds. An external floppy drive is available, as is an 'Expansion Manager' a docking station with a floppy drive, a CD ROM drive, and two expansion slots. Total weight is still only 13 pounds. NCR gave away a bag with a joint NCR/AT&T 'Putting it all together' theme, and most of the NCR employees had a pat 'this has been a good move' speech, but at the hospitality suite it was pretty easy to spot the former AT&T employees, they were the ones trying to figure out how much they should enjoy the party. NCR's reputation as a company that works hard and plays hard seems intact. (Whew!) NUCON does not usually get many significant product announcements, and 1992 was no exception there. However, a working model of the NCR 3600 massivly parallel system was on the floor, as were the recently released 3450 and 3550 systems. The 3550 features some innovative fault-tolerant features, including redundant power supplies and modularized peripherals. (Most peripherals can be shut down and serviced without bringing down the entire system.) The 3550, which holds 2-8 processors, uses SVR4-MP, NCR's multi- processor version of AT&T System V Release 4. NCR also showed production versions of their RAID (Redundant Array of Inexpensive Disks) systems, in which 9 drives appear as one. With one drive as a parity drive it is possible to lose one drive without losing any data, and it can also be serviced while the remaining drives stay online. A utility program initializes and restores the data to the replaced disk. Sustained data transfer rates in excess of 20 megabytes per second are also possible on the SCSI-2 bus. Computer columnist and commentator John Dvorak was one of the keynote speakers. John seems to have done another turnabout and feels the PC industry is on the verge of the next major platform change. John downplayed the possible impact of Windows NT, saying that people won't trust Microsoft another time, and said that he feels Unix is 'cursed'. The new platform, John feels, will be based on OS/2. However, both of the last two platform changes (the Apple ][ and the IBM PC) were essentially driven by 'killer' applications (Visicalc and Lotus 1-2-3), and John was at a loss to predict the next killer app. Notably absent from John's commentary was any mention of Apple Computer. His disparaging remarks about Unix were ironic at a group which is becoming increasingly Unix oriented, as NCR is using Unix as its primary operating system on its new products. ---------- Lead Sysop Notes (ANDY) Andrew Finkenstadt ================ Chief Sysop Another month has gone by, and the Unix RoundTable is expanding still more. The Stereolithograpy Users Group is joining us with a bulletin board category for public discussion. Stereolithography is a 3D Modeling and Rapid Prototyping technique using Silicon Graphics machines, running a version of UNIX, of course. Even the most non-technical of us will find the messages here fascinating and informative. The month of April saw the status update on the GEnie Internet gateway. No further updates are available at this time, though I am anxiously awaiting the release of it. Stay tuned. The Unix Help Desk in the Conference Area and Chat Lines has helped a number of people with big and small problems. Recently a man located in Tennessee set up his company's network and lamented the long addresses required for each person from one end of his network to another. A quick pointer to 'smail' and its capabilities and he was all set to go. A tutorial on domain names and Internet Protocol Number assignments followed. He's now a satisfied man. Look for a notice from me about a Survey for the Unix RT participants. I want to offer you what you want and need, and this is one tool by which I can do this. Feel free to drop me an E-Mail (my address is ANDY, or andy@vistachrome.com, or uunet!rde!andy) and I'll read and respond to your feedback with great interest. Until next month, happy computing! -Andy Upload Contest Winners for April (ANDY) ================================ Congratulations go to Ken Ficara and Toshi Morita for their qualifying uploads in the Unix RoundTable Upload Contest. Each of them has won a free day of their choosing in the Unix RT. Ken uploaded the excellent guide to Internet folklore called {ZEN and the art of the Internet}, #3321 and Toshi uploaded some of the Linux free version of Unix for the 386 files, including the VERY popular Readme file, #3205. {ZEN} will be released on Monday, May 11, so don't fret if you can't find it right away. You too can qualify to win in the upload contest. Download file #2346 CONTEST.RULES and start uploading. usr/local: Items (scripts and news) snarfed from various sources ========= News for Unix Users and wanna-be's. (DELPHI) Brian Riley ----------------------------------- For those of you who can't keep up with the latest happenings in the Unix world, here are the highlights. New Telebit Modem: The WorldBlazer external modem has a maximum sustained throughput of 23000 bps using Turbo-PEP, a new propriatary protocol, over clean lines. The modem also supports V.32bis, V.42, V.42bis, UUCP, Kermit, XMODEM and YMODEM. Owners of T3000's will be able to upgrade to the WorldPort with a new chip set and faceplate. Fax capabilities are scheduled for release in September. Unix Tools for Dos & OS/2: Mortice Kern Systems has released version 3.1 of their popular LEXX and YACC utilities. These tools are useful for creating compilers, database languages or extension languages for Dos and OS/2 systems and are compatible with Unix tools of the same name. YACC 3.1 now includes YACC Tracker, a debugging tool for developers and there is online help for both tools. Also from MKS is version 5.2 of RCS, a source code control system wich now includes support for binary as well as text files and enhanced security features. MKS Toolkit is up to version 3.2 and emulates SVR4 utilities, adds enhancements to some commands and includes MKS Korn shell. Shrink wrapped UNIX: Unix System Laboratories announced the scheduled release of Desktop Unix, otherwise known as Unix Lite, for MIPS, SPARC and Intel platforms for the third quarter with source code versions available this summer. Five OEM's will be porting DeskTop Unix to the various platforms. This will reduce the number of Unix "flavors" so small vendors will be able to bundle Unix Lite with machines and still maintain a low profit margin. A new SUN rises: SPARCstation 3 is the latest offering from SUN MicroSystems in the battle for the top workstation of the future. The new SPARKstation is said to have a rated performance of 50 SPECmarks, a 40 Mhz processor and include an ISDN port on the motherboard. It will accomodate up to 4 processors with the base system price roughly equal to the SPARCstation 2. The price of the SPARCstation 2 is expected to come down around May 19 when the SPARCstation 3 will begin shipping. Brian T. Riley Tidbits (LRARK) Ricky Mobley ------- Article 187 of comp.bugs.misc: Path: lrark!ddi1!uunet!uunet!think.com!mips!mips! munnari.oz.au!deakin.OZ.AU!rand!cocamrd.mel.cocam.oz.au!rdb From: rdb@cocamrd.mel.cocam.oz.au (Rodney Brown) Newsgroups: comp.lang.misc,comp.bugs.misc,comp.unix.questions Subject: Limits to awk Summary: For NF > 99 or length($0) > 3071 use mawk, gawk or perl Message-ID: <1076@rand.mel.cocam.oz.au> Date: 25 Mar 92 22:23:22 GMT Sender: usenet@rand.mel.cocam.oz.au Followup-To: comp.bugs.misc Organization: Co-Cam Computer Group, Oz Lines: 167 Xref: lrark comp.lang.misc:473 comp.bugs.misc:187 comp.unix.questions:17886 Some years ago I was tripped up by the undocumented (as far as I could see) limits in line lengths & numbers of fields in awk. This summarizes the limits in the awk versions I tested with the script in the enclosed shar. Maximum Maximum Length Fields Machine & awk Size Date Uname 3072 100 /bin/awk 102364 880511 XENIX 2.3.3 SysV i80386 3072 100 /usr/bin/awk 89108 900612 SCO 3.2 2 i386 2561 99 /usr/bin/awk 155648 900118 HP-UX hp835 A.B7.00 D 9000/835 3072 99 /usr/bin/nawk 192512 900118 HP-UX hp835 A.B7.00 D 9000/835 2560 99 /bin/awk 86608 891031 AIX Self 2.1 2 20B0166C-IBM RT 3072 99 /bin/awk 194257 900521 AIX rs6000 1 3 000006073100 3072 99 /bin/awk 194769 901110 AIX rs6000 1 3 000060783100 2560 99 /usr/bin/awk 56072 910107 SysVR4v2.0 4.0 2.0 i386/i486 3072 199 /usr/bin/nawk 110356 910107 SysVR4v2.0 4.0 2.0 i386/i486 2560 99 /bin/awk 110592 910319 ULTRIX 4.2 0 RISC 19998 4000 /usr/bin/nawk 102400 910319 ULTRIX 4.2 0 RISC 3071 100 /usr/bin/nawk 163840 901012 SunOS 4.1.1 1 sun4c 2559 99 /usr/bin/awk 90112 901012 SunOS 4.1.1 1 sun4c >=4095 >=4096 Gnu Awk (gawk) 2.13, pl 2 SunOS 4.1.1 1 sun4c >2**17 32767 c.s.r/mawk/mawk1.1/mawk XENIX 2.3.3 SysV i80386 32767 32768 Gnu Awk (gawk) 2.13, pl 0 XENIX 2.3.3 SysV i80386 >2**17 65000 perl 4.0.1.6 pl 19 XENIX 2.3.3 SysV i80386 /bin/awk core dumps for lengths > 14778 XENIX 2.3.3 SysV i80386 /usr/bin/awk core dumps for lengths > 11419 SCO 3.2 2 i386 /bin/awk core dumps for lengths > 2564 AIX Self 2.1 2 20B0166C-IBM RT /bin/awk core dumps for lengths > 3120 AIX rs6000 1 3 000006073100 /bin/awk core dumps for lengths > 3120 AIX rs6000 1 3 000060783100 /usr/bin/nawk core dumps for lengths > 5384 HP-UX hp835 A.B7.00 D 9000/835 /usr/bin/awk core dumps for lengths > 2582 SysVR4v2.0 4.0 2.0 i386/i486 /usr/bin/nawk core dumps for lengths > 3098 SysVR4v2.0 4.0 2.0 i386/i486 /bin/awk core dumps for lengths > 9946 ULTRIX 4.2 0 RISC ULTRIX 4.2 0 RISC /usr/bin/nawk doesn't give an error at its maximum line length of 19999 (including the newline) instead it splits the input line into segments of that length. Both mawk 1.1 & gawk 2.13 seem to cope with very large records, gawk 2.13 however seems to be returning the length as a short which is why 32767 is the value shown in the table. This is the error gawk gives when presented with 32769 fields :- gawk: fatal error: set_field: nodes: can't allocate memory (Not enough core) and for mawk with 32768 fields :- mawk: program limit exceeded: maximum number of fields size=32767 The values for perl are from using a2p on the awk oneliners used. The 65000 is the split maximum introduced by a2p. #! /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 shell archive." # Contents: awkfld.sh # Wrapped by rdb@cocamrd on Wed Mar 25 23:20:07 1992 PATH=/bin:/usr/bin:/usr/ucb ; export PATH if test -f 'awkfld.sh' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'awkfld.sh'\" else echo shar: Extracting \"'awkfld.sh'\" $2152 characters$ sed "s/^X//" >'awkfld.sh' <<'END_OF_FILE' X: X# Script tries to find maximum record length for awk & maximum number X# of fields awk can cope with in a record. X# Also checks for core dumps on long records. X# It is using m4's eval(0,10,length) to generate output of given lengths X# - which probably requires some kind of SysV m4 - or GNU m4 1.0 X# Use length($0) instead of length for mawk 1.1 X# X# A maximum record length of 4096 is used X# Larger record lengths will require increasing the size of m4's push back X# buffer -- m4 -B140000 will cope with 2**17 X# The upper value to search for can be specified as an argument to the X# script. X# Note that large numbers of fields (32767) takes a long time (in mawk & gawk) . X# Xl=0; Xuname=`uname` Xif [ $uname = 'ULTRIX' ] ; then X M4=${M4-m4} X AWK=${AWK-awk} X awk_name=`which $AWK` X u=${1-4096}; Xelse X if [ $uname = 'SunOS' ] ; then X M4=${M4:=/usr/5bin/m4} X else X M4=${M4:=m4} X fi; X AWK=${AWK:=awk} X awk_name=`type $AWK | sed 's/^.*[ ]$[^ ]*$$/\1/'` X u=${1:-4096}; Xfi Xuname -a Xls -l $awk_name Xif [ `expr ${AWK} : 'gawk$'` != 0 ] ; then X ${AWK} -V Xfi Xif [ -f core ] ; then X rm -f core Xfi Xecho "eval(0,10,$u)" | $M4 \ X| $AWK '{print length($0)}' >> /dev/null 2>&1 ; Xif [ -f core ] ; Xthen X while [ `expr $l + 1` -ne $u ] ; do X m=`expr $l + $u`; X m=`expr $m / 2`; X if echo "eval(0,10,$m)" | $M4 \ X | $AWK '{print length($0)}' >>/dev/null 2>&1\ X || [ ! -f core ] ; X then X l=$m; X else X u=$m; X rm -f core X fi X done X echo $awk_name core dumps for record lengths greater than `expr $u - 1` Xfi Xl=0; X while [ `expr $l + 1` -ne $u ] ; do X m=`expr $l + $u`; X m=`expr $m / 2`; X x=`echo "eval(0,10,$m)" | $M4 | $AWK '{print length($0)}' 2>>/dev/null`; X if [ $? != 0 -o "$x" != $m ] ; X then X u=$m; X else X l=$m; X fi Xdone Xecho Longest record for $awk_name is `expr $u - 1` Xl=0; X# AWK=/user/AUUG/gnu/gawk-2.11/gawk Xwhile [ `expr $l + 1` -ne $u ] ; do X m=`expr $l + $u`; X m=`expr $m / 2`; X if echo "eval(0,10,$m)" | $M4 \ X | $AWK -F0 '{print NF}' >> /dev/null 2>&1 ; X then X l=$m; X else X u=$m; X fi Xdone Xecho Maximum fields for $awk_name is $u END_OF_FILE if test 2152 -ne `wc -c <'awkfld.sh'`; then echo shar: \"'awkfld.sh'\" unpacked with wrong size! fi chmod +x 'awkfld.sh' # end of 'awkfld.sh' fi echo shar: End of shell archive. exit 0 -- Rodney Brown, Co-Cam Computer Group, ACSNet: rdb@mel.cocam.oz.au -- rick@lrark.UUCP * bang path --> uunet!ddi1!lrark!rick * Ricky Mobley * lrark (501) 224-9454 * 1800 Sanford Dr. #4 * GEnie Unix RT SysOp address LRARK * Little Rock, AR 72207 * WB5FDP.AR.USA.NA wb5fdp.ampr.org * ---------- "POPULAR" Files Listing Andrew Finkenstadt ======================= GE Mail Address ANDY Over the past couple of months, many of you have visited the software libraries and downloaded some interesting files. In case you were wondering what you missed, here are 38 of our most popular files which will also be of interest. I would like to personally point out the files from the "Frequently Asked Questions" (FAQ) list for Unix systems in general. The questions answered here are ones that everyone will want to know. This is the April 1992 issue of the FAQ from the Usenet. File File Name Size Lib # Short Description ----- -------------------------------- ---------- ---- 3222 UNIX.FAQ1.492 13,568 5 3223 UNIX.FAQ2.492 29,568 5 3224 UNIX.FAQ3.492 28,288 5 3225 UNIX.FAQ4.492 23,296 5 And without further ado, here is the listing: 2466 ANON-FTPLIST.Z 45,056 16 File File Name Size Lib # Short Description ----- -------------------------------- ---------- ---- Anonymous FTP Site listing 3051 BSD-README 16,768 1 386BSD release 0.0 description 3132 CAWF2.ZIP 134,528 3 CAWF, an nroff clone for MS-DOS 2727 COMPRESS.ZIP 13,696 13 msdos/unix compress util 3167 CRACK41.TAR.Z 112,640 1 crack41.tar.Z, password cracker 921 DOMAIN.INF 7,560 15 Intro to paths and domain names 1525 DOSTAR.COM 22,680 13 This is a MS/DOS tar program 2376 E-MAIL.GUIDE 18,688 4 John Chew's Internetwork guide 2914 FPTED2.TAR.Z 34,688 11 FPTED2.tar.Z, User friendly editor 68 INTERNET.ARC 32,760 1 Hitchiker's Guide to Internet 2888 LINUX-ELVIS.TAR.Z 78,080 3 elvis (vi) for linux, binary 2758 LINUX-FAQ.TXT 32,768 5 Frequently Asked Questions-LINUX 2780 LINUX-FILES 15,488 10 Linux files available via FTP 2890 LINUX-GREP.TAR.Z 72,832 8 grep 1.5 binary for linux 2750 LINUX-INFO 27,904 10 Information on the LINUX unix clone 2891 LINUX-PMAKE.Z 72,320 10 Make for LINUX (binary) 3002 LINUX-RELNOTES-0.95 10,752 10 Release Notes, Linux Version 0.95 3121 LINUX-ROOTIMAGE-0.95A.Z 638,464 3 Linux 0.95a root filesystem image 2047 MK42BASE.TAR.Z 1,326,976 10 mk42base.tar.Z, Mach internals 3141 NETVIEWS-UNIX AND NT 78,208 27 Will Windows NT kill Unix? 3206 NIXLONG.04.92 52,480 15 NixPub long listing for April 1992 3106 NIXPUB.SHORT.LONG.0292 62,464 15 long and short public sites Feb 1992 2016 PCOMM12.TAR.Z 176,384 16 pcomm12.tar.Z, pcomm v 1.2.5 2728 PC_TAR.ZIP 25,984 13 tar utility for the pc 2892 PM_SUPP-TAR.Z 6,400 10 Supplemental files, Linux Pmake 722 PROTOCOL.TXT 46,620 17 Compression and Upload Protocol Docs 3205 README.95C 2,560 10 README file for boot95c (Linux 0.95c 2551 README.MAP 18,688 15 how to interpret the UUCPmap files 3066 RTC0310.92 17,664 2 Transcript of 03/10/92 RTC. 2485 RZSZ.SH.1208 176,512 16 ZMODEM C Source Code 3012 TREE.C 4,352 1 tree.c, directory tree browser 3040 UNIDOS11.ZIP 9,216 1 UNIX <--> DOS text file converter 3118 UNIV_ZIP.TAR.Z 194,816 13 univ_zip.tar.Z, universal zip/unzip 3222 UNIX.FAQ1.492 13,568 5 Frequent Unix Questions 4/92, part 1 3223 UNIX.FAQ2.492 29,568 5 Unix Frequent Questions 4/92, Part 2 3224 UNIX.FAQ3.492 28,288 5 Unix Frequent Questions 4/92, part 3 3225 UNIX.FAQ4.492 23,296 5 Unix Frequent Questions 4/92, part 4 69 UNIX.QUESTIONS.1089 47,880 5 Unix questions with answers 10/89. 1918 UNIX_NEW.TXT 30,592 5 unix_new.txt, answers to questions 2938 UNXNL-02.92 23,552 27 Unix RoundTable newsletter Feb 1992 3044 UNZIP 4.1 TAR.Z 112,256 13 UNIX utility to extract MS-DOS ZIPs 3179 VI.CKV.Z 3,840 3 vi.ckv.Z, C coding/log macros for vi 3020 VI.MACROS.Z 16,128 3 vi.macros.Z, tutorial - vi macro use 3187 XGIFTOOL.TAR.Z 21,760 18 Xgiftool.tar.Z, describe/view gif's 2504 XYZMODEM.ZOO 46,336 16 xyzmodem.zoo, comm C source ---------- FAST and NASTY, DOWN and DIRTY: quick fix scripts that do something ================================ Sub: The Colorado memory systems tape We have a Colorado Memory Systems tape backup at work. Yes, I knew you'd be thrilled to learn that. Well, it seems that thae the good people at CMS did not write jtape so that it would send out a return code when no tape's in the drive. I have resorted to a bit of a work around to deal with this situation. I thought I'd pass it along to you, and maybe it will turn out to be useful to someone. My script writing skills aren't too wonderful yet, but I believe you'll get the idea. SCRIPT: # This guy is set up for the Bourne shell. #Backup script to be run on a regular basis by cron if jtape retension | grep 'Error 85' # Error 85 = no tape then echo "NO BACKUP FOR YOU -- NO TAPE" >/dev/console exit 1 else jtape backup "/" -na -c -j -np -s -t "`date` backup" >somefile \ 2>somefile fi ---------- Tutorials ========= vi - part 3, advanced features (GARS) Gary Smith == ------ ----------------- A tutorial on the Unix visual text editor from the GEnie Unix RoundTable -- abstract -- Part 1 provided an introduction to vi and gave a quick start to vi usage. Part 2 finished the basic tools instruction for adding, inserting, changing and deleting text. Part 3 (this session) will be an introduction to macros and a summary of what we have learned. There will also be some explanation of vi's many hidden secrets. -- macro basics -- As we discovered in part 2, text can be yanked and put (cut and pasted) with vi(), and we even cleaned up our test file using these techniques. If you found yourself thinking that while that worked, more modern editors do the same job with more ease. You would be right, but you would also be making your judgement based on incomplete information. Vi has more tricks hidden in it than a carnival magician's coat. There is always another way to do things in vi, and many of them are single keystroke works of art. Others require some thought in advance, then perform incredible global edits. Macro commands can be very tricky, involving nested routines. If you really want to use one of the finest edit tools ever created you will take the time to explore each of these commands deeper than this tutorial can or should take you. -- buffers and macros -- "xyy Yanks the current line into a named buffer x, then "xP puts the text stored in buffer x before the cursor. x can be any of the twenty-six letters of the English alphabet. I usually just name mine a, b, c ... notice the reference to a 'buffer' area and remember an earlier reference to vi's use of buffers. Those buffers are a major key to vi's power. Another use of buffers many users aren't even aware of is what I call the 'oops' syndrome. Have you ever accidently deleted a word or line of text only to discover " 'oops', wrong one!"? Of course, you have. We all have. Vi buffers to the rescue. Vi maintains the last nine (9) deletes in buffers. If you know which delete (say 2 have occured since then, your 'oops' is number 3) it was all you have to do is position your cursor and type "np where n is the number of the delete. So, you have nine delete buffers, twenty-six named buffers and your entire edit file. That's right, vi is editing your file in a temporary buffer. Not only that, but if you lose power your file will be preserved in a .preserve directory. When the system powers back up, /tmp contains ex files. A utility called /usr/lib/ex3.7preserve is run from etc/rc which sees /tmp and converts it to a preserve file. As soon as you log back in, you will get a mail memo telling you there is a preserved editor file (vi or ex) you can recover. If you type ex -r or vi -r at the system prompt, you will be presented a list of recovered files. The last edit you did will probably not be saved, but the file as it existed before your last edit will be entirely intact. If you had suffered a system crash while editing our test file, you would simply enter: vi -r test Yes, you can use this preserve buffer even without a system crash by using the ex command :pre. Now, let's say you are editing a group of files into one cohesive report, but you can't remember all their names. An example might be a series of dated travel reports you have to assemble as one trip report, and the days are not sequential. You could write down or screen print the file list before you start your edit and refer to your list. You could also let another macro, map do the work for you. Map lets you create an escape macro that will do a job for you. In our case, it is going to create a single keystroke command to list the files in our directory and return us to vi and our edit file. The syntax is simple, and uses the ex map command. :map l :!ls <enter> Now, each time you want to list the files in the directory, make sure you are in the command mode and hit l. You will get a directory list immediately with a massage to hit enter to return to vi. You hit enter, and type :r filename <enter> to reaad in the next file who's name you just verified. If your terminal has function keys you can make this even more fun. If you want function key 3 to list your directory the syntax would be: :map #3 :!ls <enter> What if you discovered you had screwed up by mapping a key you want to use in its normal capacity - like assigning l, when we are using it for cursor control. If you realize it immediately use your u (undo) command. If you discover it later in your edit type: :unmap l <enter> You can chain mapped macros in one map statement. It is not necessary to create one macro function per map statement. Macros are simply pre-defined sequences of keystrokes. There is nothing to prevent you from calling other predefined keystrokes. CAUTION!!! DO NOT create a recursive macro. vi will forbid you from creating a macro that calls itself at the end of its sequence, but it will not prevent you from embedding such a call. DO NOT do it. The results are disastrous. A recursive macro will expand very rapidly, and can loop indefinitely, wipe out your text or suck memory like a black hole. Should you accidently create such a monster remember that most macros can be interrupted. Try hitting DELETE, RESET/BREAK and <CNTL> C. Two very important macros have been reserved for the next session. These are autoindentation and abbreviations. We will also examine some limitations of macros. -- vi/ex archive -- The GEnie Unix RoundTable has been authorized to mirror the vi/ex archive maintained in Bergen Norway by Ove Ruben R. Olsen. The authorization letter is conatined in file #2774. There are MANY tutorials, fine help and usage files, and refernce charts for vi() and ex() in our library as a result. You can find these files by using vi. (with a dot) and ex. as your search keywords. I want to encourage you to capture these files and learn from the masters how to turn vi and ex into the power user's tools they can and should be. One of those files is vi.ref.Z, and I am going to copy it here to summarize the ex/vi commands you should be comfortable with by now. If you see a command and don't know how it works, load your test file and find out by using it. I repeat from sessions 1, "Using computers, and especially a complicated editor like vi() is a participation, not a spectator, activity." Unix vi Reference Card arrow keys move cursor ^n,^p,<SP>,<BS> same as arrows <ESC> escape from text entry ^g tell what is going on : give ex command ZZ write and quit <CR> start of next line - start of previous line ^b go back one page ^u scroll back 1/2 page ^f go forward one page ^d scroll forward 1/2 page or end indent ^e scroll down one line ^y scroll up one line H home screen line L last screen line x delete character r replace char with next typed a append after cursor A append after line i insert before cursor I insert before line p put before cursor P put before line u undo changes U undo changes in this line only c obj change object to following text Object defining characters w forward one word W forward, ignore punctuation b backward one word B backward, ignore punc. e end of current word <CR>,- as above Characters using buffers d obj delete object dd delete line y obj yanks object yy yank line p place what is in buffer " select buffer NOTE: There are 26 buffers, each corresponding to a letter of the alphabet. To refer to a specific buffer type "<letter> e.g. "a refers to buffer a. Therefore "add would delete the line that cursor is on and place it into buffer a. To place the deleted line elsewhere, position cursor and type "ap. In order to delete or yank a section of text, you must place a mark to where the deletion/yank is to stop. This is done by typing: m<letter> (do not use the same buffer to place a mark and store text). To return to that location, type `<letter>. When deleting or yanking to a mark, say b, position the cursor at the beginning of the desired text, and type "ad`b. This deletes the text from cursor until mark b, and places it in buffer a. "ad'b will delete until the end of the line in which mark b is (note difference in ` and '). Search /text<CR> search forward for text ?text<CR> search backward n next occurance N next occurance in opposite direction Miscellaneous J adjoin current line to line above !! place results of command into file i.e. !!cat it.p will place the list of it.p in your file. Unix ex Reference Card n set pointer to line n n,n specify a range of lines for next command $ used to indicate last line in file . print current line or end insertion p print specified lines p# print specified lines with line numbers a append lines after this line; end insertion with "." i insert lines after this line; end insertion with "." d delete specified lines u undo last deletion or substitution c change (i.e. replace) specified lines to new text ya yank specified lines into buffer pu lines from buffer before this one w write current text to file q quit edit q! forces quit without write wq write and quit vi enter vi editor se display settings; refer to manual for details /text/ search for text // search for next occurrence s/oldtext/newtext/ substitute newtext for oldtext in specified lines n1,n2copy3 copy lines n1-n2 to after n3 global/pattern/cmds search file for pattern and do cmds on those lines. set nu insert line numbers set nonu remove line numbers -- "...nothing kills that does not know ye." -Meg Davis, `The Elf Glade' UUCP: ihnp4!gargoyle!sphinx!d757 d757@sphinx.uchicago.edu Lawrence Lerner University of Chicago Computation Center __ _ Gary Smith * ... uunet!ddi1!lrark!glsrk!gars * / _' _ _ (_' P. O. Drawer 7680 * nuucp%ddi1@uunet.UU.NET * /__/ (_|_/ '._) Little Rock,AR 72217 * GEnie Forth RT & Unix RT SysOp * --------------- - U. S. A. - * ph:501-227-7817,fax:501-228-9374 * (M-F 0800-1700 CST) **** End of article ---------- M4 (1) Tutorial, Part 2 Andrew Finkenstadt ======================== GE Mail address ANDY In February 1992 we began part 1 of this tutorial (file #2938) by discussing the uses of m4(1) as a filter for programming and documentation, and then used it as part of a database design project using SQL, Structured Query Language. This month we discuss the various command line options that are given to m4 when running it, and why we would want to use these options. Under Interactive Unix System 5 Release 3.2 the m4 processor has just 8 command line options to consider. These are: -e "Operate interactively. Interrupts are ignored and the output is unbuffered." That is to say, when you specify -e on the command line, you run m4 as if you are typing directly to it. The reason why interrupts are ignored is to compensate for a noisy phone line. -e is the easiest way to test out some m4 commands before you commit them to disk. -s "Enable line sync output for the C preprocessor." The C preprocessor (typically systems now call this /lib/cpp or cc -E) uses this to enable it to track where it is within the original source code program.c - it causes m4 to emit placemarkers after processing each line. (C compilers typically tell you that 'such and such error' occured on line 44. It knows this via the #line comments from m4.) #line 1 the results of line 1 #line 2 the results of line 2 -B### "Change the size of the push-back and argument collection buffers from the default of 4,096." m4 only allocates a mediocre amount of memory to macro expansion and substitutable parameters, namely 4K. In general this is sufficient for just about any project, but if you end up going macro happy or have a very large expansion, you may run out of push-back space and receive a cryptic error message - or (gasp!) output that is just plain wrong. If you encounter problems with macros seemingly not working right, and can't find any other causes, try increasing the push-back size by 4K or more. (IE: try -B8192 or -B16384 and see if the problem goes away.) -H### "Change the size of the symbol table hash array from the default of 199. The size should be prime." m4 uses a hash table to increase the speed with which definitions and macros are found. A typical hash function will cause the definition to be found on the first or second try. If many definitions are used then better performance may result by increasing the number of elements in the symbol table array. The number had better be prime, else Bad Things Happen. -S### "Change the size of the call stack from the default of 100 slots. Macros take three slots, and non-macro arguments take one." m4 uses a fixed-size stack to contain macro expansions and arguments as they are evaluated. (A stack can be likened to the stack of plates at a restaurant: you put a plate on the top of the stack, and then when you need a plate, you take the top plate from the stack.) As larger macros are used, with greater complexity, it is likely that the stack array will need to be increased. It is good to try increasing this value slowly, perhaps to 150 and then to 200. -T### "Change the size of the token buffer from the default of 512 bytes." A token is the largest "thing" that m4 will recognize. If you were to set T to -T32, then any "token" (you can think of a token as a word between spaces or punctuation) longer than 32 characters would be either ignored, rejected, or shortened. If 512 bytes is not enough, perhaps you should teach others how you type that fast. :) All of the preceeding command line options MUST be specified before any files or other options. There is a reason for this: memory allocation is done at the beginning, and must be performed before the first files or definitions are created. -Dname=val -Uname These two command line options allow you to communicate with your m4 program and change its behaviour or perhaps alter its output from the command line. Within your m4 program you could test for the existence of the definition for UNIX and if it were there, you could choose to include a paragraph about making sure your local System Administrator has sufficient coffee to drink while on the job, where as if UNIX is not defined, but DOS is, then it could include a blurb about regular backups. This also allows you to override certain settings. For example, some C compilers do not grok the "void" type of functions. You might define void to be int for these old-style compilers: m4 -Dvoid=int .... That about does it for command line options to m4. Next time we look at more functions beyond define(). Until then, have fun! -Andy *** End of article --------------- REMINDER - This newsletter is being sent to you 'by request'. If you do not wish to keep receiving it, e-mail a stop notice to GARS. On the other hand, we would very much appreciate it if you would pass the word that we do distribute this item near the tenth (10th) of the month of issue to any- one on GEnie who requests it. P L E A S E also remember contributions are most welcome. Please e-mail items and/or suggestions to GARS. (EOF) Trademark and Copyright notices: Unix is a Trademark of UNIX System Laboratories, Inc.; GEnie, LiveWire, and RoundTable are Trademarks of General Electric Information Services Company; Xenix and ms-dos are Trademarks of Microsoft Corporation; NeXT and NeXTstep are Trademarks of NeXT Computer Systems, Inc., Coherent is a Trademark of Mark Williams Company, Sun is a Trademark of Sun Microsystems, Inc., Colorado Memory Systems is a Trademark of Colorado Memory Systems. The contents of this newsletter are copyright(c) 1992 and may be copied whole or in part only if original credit is included. The GEnie UNIX RoundTable is not affiliated with AT&T or UNIX System Laboratories, Inc.