rsalz@uunet.UU.NET (Rich Salz) (09/29/87)
Submitted-by: Kenneth Ingham <ingham@charon.unm.edu>
Posting-number: Volume 11, Issue 82
Archive-name: watcher/part01
[ Watcher is a system-monitor utililty, described at the Phoenix
Usenix. --r$ ]
#! /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 1 (of 2)."
# Contents: Docs Docs/Abstract Docs/Makefile Docs/README Docs/macros
# Ideas MANIFEST Makefile README Support Support/Daemons
# Support/README Support/UnmHosts Support/Watcherfile
# Support/crontab.entry Support/syswatch baderr.c check_item.c
# checkline.c clean_hist.c do_args.c doit.c externs.c find_of.c
# find_pre_cmd.c find_pre_val.c get_col_fld.c get_rel_fld.c
# getargv.c init.c init_sigs.c line_to_vec.c main.c open_cf.c
# open_hf.c pp.c pp_change.c pp_out.c read_hist.c save_key.c
# yyerror.c yylex.c
PATH=/bin:/usr/bin:/usr/ucb ; export PATH
if test ! -d Docs ; then
echo shar: Creating directory \"Docs\"
mkdir Docs
fi
if test -f Docs/Abstract -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Docs/Abstract\"
else
echo shar: Extracting \"Docs/Abstract\" \(3638 characters\)
sed "s/^X//" >Docs/Abstract <<'END_OF_Docs/Abstract'
X.sp 0.5i
X.ce 2
XKeeping watch over the flocks
Xat night (and day)
X.sp 0.3i
X.ce 8
XKenneth Ingham
XUniversity of New Mexico Computing Center
XDistributed Systems Group
X2701 Campus NE
XAlbuquerque, NM 87131
X(505) 277-8044
Xingham@charon.unm.edu
Xucbvax!unmvax!charon!ingham
X.sp 0.2i
X.ce
XTopic Areas: Applications, System management, Utilities
X.sp 0.5i
XThe computing facilities offered by the University of New Mexico
XComputing Center include three microvaxen, five large vaxen (780 or
Xbigger), and a Sequent B8000. In addition to these Unix/VMS machines,
Xthe UNMCC Distributed Systems Group (DSG) monitors a number of the
Xvarious microvaxen and sun workstations scattered across campus. This
Xduty falls to the DSG Programmer designated as "DOC", or "DSG On Call",
Xwho receives his beeper based on a monthly rotation schedule.
X.sp
XIn the past, shell scripts running every six hours reported various
Xsystem statistics to DOC, who then scanned the output for signs of
Xpossible trouble. As the number of machines and the number of
Xpotential problems grew, the mound of output that DOC had to process,
Xmost of which merely indicated normal system operation, became
Xoverwhelming. Now, with several machines to monitor and only one
Xperson acting in this capacity, DOC can often waste a tremendous amount
Xof time wading through system status reports, time which can be better
Xspent actually fixing system problems.
X.sp
XIn response to this situation, the author developed a tool which
Xintroduces some intelligence into the machine's self-reporting, letting
Xthe machine filter out messages indicating normal operation and
Xforwarding to DOC only those messages which point out trouble areas.
XThe result of these efforts is Watcher, a very general and extensible
Xsystem self-monitor. Running more often than the set of
Xshell scripts, Watcher keeps closer tabs on the system; since it
Xdelivers only a summary of potential problems, however, this extra
Xmonitoring produces \fIno\fR corresponding increase in the demand on
Xthe system manager. No problems slip by unnoticed in the more concise
Xoutput, leading to an improvement in overall system availability as well
Xas the more effective utilization of the system manager's time.
X.sp
XWatcher was designed to be almost as flexible as DOC in deciding what
Xconstitutes a problem with the system. Running at intervals specified
Xin crontab, Watcher issues a number of
Xuser-specified commands (each of which
Xdelivers its output in a different format), parsing all or part of the
Xoutput from either the left or the right. It compares this
Xto the last such output obtained, checking for indications
Xof a system abnormality. Such signs might take the form of a
Xtoo abrupt change in a certain value (e.g. a process which suddenly
Xbegins gobbling vast amounts of cpu time),
Xa value which exceeds the allowable maximum or minimum (such as a
Xan overly-full file system),
Xor an unacceptable change in a string value
X(e.g. when "up" changes to "down"). For commands such as
X"ps" whose output varies considerably with each run, specific
Xparts of the output can be designated as a key; successive runs of
XWatcher will home in on these key areas for their comparisons.
X.sp
XSince the user specifies not only the commands Watcher will execute and
Xthe time lapse between successive runs, but also the aforementioned
Xparameters which indicate system anomalies, Watcher can easily be seen
Xas a very flexible, general system monitor. Its use at UNM has provided
Xa marked increase in the productivity of the system manager, which has
Xled in turn to the increase in the reliability and availability of the
Xsystems at UNMCC.
END_OF_Docs/Abstract
if test 3638 -ne `wc -c <Docs/Abstract`; then
echo shar: \"Docs/Abstract\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Docs/Makefile -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Docs/Makefile\"
else
echo shar: Extracting \"Docs/Makefile\" \(73 characters\)
sed "s/^X//" >Docs/Makefile <<'END_OF_Docs/Makefile'
XTROFF=xroff
X
Xpaper:
X $(TROFF) macros Paper
X
Xabstract:
X $(TROFF) Abstract
END_OF_Docs/Makefile
if test 73 -ne `wc -c <Docs/Makefile`; then
echo shar: \"Docs/Makefile\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Docs/README -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Docs/README\"
else
echo shar: Extracting \"Docs/README\" \(315 characters\)
sed "s/^X//" >Docs/README <<'END_OF_Docs/README'
XIn this directory is all of the documenttion about watcher. It should
Xinclude a man page, the abstract sent in to usenix, and the paper
Xpublished in the proceedings of the 1987 Summer Usenix in Phoenix.
X
XThe Makefile shows how to print the abstract and the paper. The man
Xpage is the same as all other man pages.
END_OF_Docs/README
if test 315 -ne `wc -c <Docs/README`; then
echo shar: \"Docs/README\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Docs/macros -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Docs/macros\"
else
echo shar: Extracting \"Docs/macros\" \(712 characters\)
sed "s/^X//" >Docs/macros <<'END_OF_Docs/macros'
X\" macro definitions
X.de EX
X.nf
X.in +0.25i
X.sp
X..
X.de NX
X.in -0.25i
X.fi
X.ad
X.sp
X..
X.de SE
X.ps \\n(ps+1
X.ft B
X.sp 0.2i
X\\$1 \\$2 \\$3 \\$4 \\$5 \\$6 \\$7 \\$8 \\$9
X.ft
X.ps \\n(ps
X.sp 0.2i
X..
X.de NP
X'sp 0.75i
X.tl ''%''
X'bp
X'sp 1.0i
X..
X.de TI
X.ps \\n(ps+5
X.vs \\n(vs+5
X.ft B
X.ce
X\\$1 \\$2 \\$3 \\$4 \\$5 \\$6 \\$7 \\$8 \\$9
X.ps \\n(ps
X.vs \\n(vs
X.ft
X..
X.de AU
X.sp 0.5i
X.ps \\n(ps+3
X.ft I
X.ce \\$1
X.ps \\n(ps
X..
X.de AB
X.ll -0.5i
X.in +0.5i
X.sp 0.2i
X.ps \\n(ps+3
X.ft B
X.ce
XAbstract
X.ps \\n(ps
X.sp 0.2i
X.ft R
X..
X.de BD
X.in -0.5i
X.ll +0.5i
X.sp 0.3i
X.ps \\n(ps
X.ft R
X..
X\" end of macro definitions
X.wh -1.5i NP
X.in 0.25i
X.ll 6.25i
X.sp 0.5i \" for the first page
X.ps 11
X.vs 14
X.nr ps 11
X.nr vs 14
X\" end of macro package
END_OF_Docs/macros
if test 712 -ne `wc -c <Docs/macros`; then
echo shar: \"Docs/macros\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Ideas -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Ideas\"
else
echo shar: Extracting \"Ideas\" \(1347 characters\)
sed "s/^X//" >Ideas <<'END_OF_Ideas'
X Repace number with int or float. Currently watcher treats all
X numbers as floats. There is some code already there
X that could handle the different types. The %d for a
X float is counterintuitive. This would be fixed by
X making %d int and %f float like printf.
X
X Use yacc's union rather than YYSTYPE default. This would
X probably help portability.
X
X Print alias in pp output. Currently it isn't printed. No real
X reason for this.
X
X Multiple ok values for string fields. Sometimes there may be
X more than one string value which is considered "ok".
X Watcher should be able to handle this. Possible
X syntax:
X
X(/usr/ucb/ruptime | fgrep -f /usr/local/lib/watcher/UnmHosts) { ruptime }
X 2 status%s 1 machine%k 7 loadav%d:
X loadav 0 10;
X status "up", "slow", "dead".
X
X Positive and/or negative deltas make a difference. Currently
X watcher only notices things which increase. It would
X probably be useful to watch for things that drop also.
X Possible syntax:
X
X(df -i | /usr/ucb/tail +2 | grep -v tmp) { 'df no tmp' }
X 1-9 filesystem%k 41-42 spaceused%d 64-65 inodesused%d 1-9 device%k:
X spaceused -15%;
X spaceused 0 89;
X
X Right to left parsing as option instead of always l-r. It may
X be easier at times to specify that the useful info is
X located in the fifth from the right column rather than
X the eit`ghth from the left.
END_OF_Ideas
if test 1347 -ne `wc -c <Ideas`; then
echo shar: \"Ideas\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f MANIFEST -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"MANIFEST\"
else
echo shar: Extracting \"MANIFEST\" \(1494 characters\)
sed "s/^X//" >MANIFEST <<'END_OF_MANIFEST'
X File Name Archive # Description
X-----------------------------------------------------------
X Docs 1
X Docs/Abstract 1
X Docs/Makefile 1
X Docs/Paper 2
X Docs/README 1
X Docs/macros 1
X Docs/watcher.1 2
X Ideas 1
X MANIFEST 1 This shipping list
X Makefile 1
X README 1
X Support 1
X Support/Daemons 1
X Support/README 1
X Support/UnmHosts 1
X Support/Watcherfile 1
X Support/crontab.entry 1
X Support/syswatch 1
X baderr.c 1
X check_item.c 1
X checkline.c 1
X clean_hist.c 1
X control.y 2
X defs.h 2
X do_args.c 1
X doit.c 1
X externs.c 1
X find_of.c 1
X find_pre_cmd.c 1
X find_pre_val.c 1
X get_col_fld.c 1
X get_rel_fld.c 1
X getargv.c 1
X init.c 1
X init_sigs.c 1
X line_to_vec.c 1
X main.c 1
X open_cf.c 1
X open_hf.c 1
X pp.c 1
X pp_change.c 1
X pp_out.c 1
X read_hist.c 1
X save_key.c 1
X yyerror.c 1
X yylex.c 1
END_OF_MANIFEST
if test 1494 -ne `wc -c <MANIFEST`; then
echo shar: \"MANIFEST\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Makefile -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Makefile\"
else
echo shar: Extracting \"Makefile\" \(891 characters\)
sed "s/^X//" >Makefile <<'END_OF_Makefile'
X#
X# makefile for watcher
X#
X# SYS should be either BSD or SYSV
XSYS = BSD
X
XOBJS = baderr.o check_item.o checkline.o do_args.o doit.o externs.o\
X find_of.o find_pre_cmd.o find_pre_val.o get_col_fld.o\
X get_rel_fld.o getargv.o init.o init_sigs.o line_to_vec.o main.o\
X open_cf.o open_hf.o pp.o pp_change.o pp_out.o read_hist.o\
X save_key.o y.tab.o yyerror.o yylex.o
XCFLAGS = -O -D$(SYS)
X
XDIST = Makefile Docs Support Ideas
X
Xwatcher: $(OBJS)
X cc $(CFLAGS) -o watcher $(OBJS)
X
Xclean:
X rm -f a.out core watcher *.o y.tab.c y.tab.h y.output Make.out
X
Xlint:
X lint -x *.c > Lint.out
X
Xshar:
X shar README *.c *.h *.y $(DIST) > Watcher.shar
X
X# Two shar files are made because one is too large for the mailers with
X# a 64K limit.
Xdist:
X shar README *.c > Watcher.shar1
X shar *.h *.y $(DIST) > Watcher.shar2
X
Xy.tab.c y.tab.h: control.y
X yacc -d control.y
X
Xy.tab.o: y.tab.c
Xyylex.o: y.tab.h
X$(OBJS): defs.h
END_OF_Makefile
if test 891 -ne `wc -c <Makefile`; then
echo shar: \"Makefile\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f README -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"README\"
else
echo shar: Extracting \"README\" \(1471 characters\)
sed "s/^X//" >README <<'END_OF_README'
XWatcher has been sucessfully made and run on 4.3BSD on vaxen, Ultrix 1.1
X& 1.2 on vaxen (what else?), and a sun 3/160 version 3.0 of the sun os.
XThanks to Alan Silverstein for bringing up watcher on a system V.2 machine
Xand pointing out the berkeleyisms (which I hope are all #ifdef'd now).
X
XThere have been recent problems which I cannot reproduce which appear to
Xbe a pointer in a malloc'd area pointing off a bit. I have been unable
Xso far to track this problem down.
X
XA short description of all of the files and directories for watcher.
X
XIdeas - changes that I need to or would like to make for watcher.
Xknown bugs are listed here. Someday I will fix them. If you fix them
Xbefore I do, please send the fixes to me.
X
XDocs - here is all of the documentation for watcher. A man page and the
Xpaper that appeared in the proceedings of the summer usenix conference
Xin 1987.
X
XMakefile - you guessed it.
X
XREADME - this file
X
XSupport - directory of useful files to use with watcher. There were
Xtaken from actual use at UNM (although they were taken a while back and
Xmay not be what is in use now). They may help in seeing how at least
Xone site is running watcher.
X
XThe rest of the files should be the source.
X
XHappy Watching.
X
XBug fixes and reports should be sent to:
X
XKenneth Ingham, UNM Computing Center, Albuquerque, NM 87131, 505-277-8044
X
XUsenet: {convex,ucbvax,gatech,csu-cs,anl-mcs}!unmvax!charon!ingham
XBITNET: ingham@unmb
XInternet: ingham@charon.UNM.EDU
X
END_OF_README
if test 1471 -ne `wc -c <README`; then
echo shar: \"README\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test ! -d Support ; then
echo shar: Creating directory \"Support\"
mkdir Support
fi
if test -f Support/Daemons -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Support/Daemons\"
else
echo shar: Extracting \"Support/Daemons\" \(91 characters\)
sed "s/^X//" >Support/Daemons <<'END_OF_Support/Daemons'
Xcron
Xrwhod
Xrouted
Xupdate
Xdaemon
Xinit
Xstatd
Xsendmail -bd
Xsyslogd
Xnamed
Xswapper
Xcensus
Xinetd
END_OF_Support/Daemons
if test 91 -ne `wc -c <Support/Daemons`; then
echo shar: \"Support/Daemons\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Support/README -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Support/README\"
else
echo shar: Extracting \"Support/README\" \(704 characters\)
sed "s/^X//" >Support/README <<'END_OF_Support/README'
XThese are files used at the University of New Mexico as of May 19, 1987
Xor so. Due to the evolving nature of what we determine needs to be
Xwatched and the various on call people coming up with new ways to watch
Xfor problems without generating more output, it is probably not what we
Xare now running. However it should give an idea of how watcher is used
Xat UNM.
X
XThe following files all live in /usr/local/lib/watcher (besides the
Xsource for watcher itself):
X Daemons
X UnmHosts
X Watcherfile
X
XThe following are misc. files which are also used with watcher.
X crontab.entry - what we have in our crontab
X syswatch - a shell script which actually runs watcher and mails
X the output to the on call person.
END_OF_Support/README
if test 704 -ne `wc -c <Support/README`; then
echo shar: \"Support/README\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Support/UnmHosts -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Support/UnmHosts\"
else
echo shar: Extracting \"Support/UnmHosts\" \(50 characters\)
sed "s/^X//" >Support/UnmHosts <<'END_OF_Support/UnmHosts'
Xariel
Xcharon
Xdeimos
Xeuropa
Xgeinah
Xhydra
Xizar
Xunmb
END_OF_Support/UnmHosts
if test 50 -ne `wc -c <Support/UnmHosts`; then
echo shar: \"Support/UnmHosts\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Support/Watcherfile -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Support/Watcherfile\"
else
echo shar: Extracting \"Support/Watcherfile\" \(936 characters\)
sed "s/^X//" >Support/Watcherfile <<'END_OF_Support/Watcherfile'
X(df -i | /usr/ucb/tail +2 | grep -v tmp) { 'df no tmp' }
X 1-9 filesystem%k 41-42 spaceused%d 64-65 inodesused%d 1-9 device%k:
X spaceused 15%;
X spaceused 0 89;
X inodesused 15%;
X inodesused 0 49.
X(df -i | /usr/ucb/tail +2 | grep tmp) { 'df tmp only' }
X 1-9 filesystem%k 41-42 spaceused%d 64-65 inodesused%d 1-9 device%k:
X spaceused 0 89;
X inodesused 0 49.
X(/usr/ucb/ruptime | fgrep -f /usr/local/lib/watcher/UnmHosts) { ruptime }
X 2 status%s 1 machine%k 7 loadav%d:
X loadav 0 10;
X status "up".
X(ps -aux | fgrep -v -f /usr/local/lib/watcher/Daemons | /usr/ucb/tail +2)
X { 'ps with no daemons' }
X 9-14 pid%k 16-19 percentcpu%d 42-45 cputime%d:
X cputime 0 10.
X(ps -ax | fgrep -f /usr/local/lib/watcher/Daemons | uniq -c) { 'daemon count' }
X 1 daemon_count%d 6 daemon_name%k:
X daemon_count 0 1.
X(ps -aux | fgrep -f /usr/local/lib/watcher/Daemons ) { 'daemon ps' }
X 9-14 pid%k 16-19 percentcpu%d 42-45 cputime%d:
X cputime 2.
END_OF_Support/Watcherfile
if test 936 -ne `wc -c <Support/Watcherfile`; then
echo shar: \"Support/Watcherfile\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Support/crontab.entry -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Support/crontab.entry\"
else
echo shar: Extracting \"Support/crontab.entry\" \(804 characters\)
sed "s/^X//" >Support/crontab.entry <<'END_OF_Support/crontab.entry'
X0 0 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 2 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 4 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 6 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 8 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 10 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 12 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 14 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 16 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 18 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 20 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
X0 22 * * * root /usr/local/lib/syswatch >> /usr/lib/cronlog
END_OF_Support/crontab.entry
if test 804 -ne `wc -c <Support/crontab.entry`; then
echo shar: \"Support/crontab.entry\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f Support/syswatch -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"Support/syswatch\"
else
echo shar: Extracting \"Support/syswatch\" \(286 characters\)
sed "s/^X//" >Support/syswatch <<'END_OF_Support/syswatch'
X#! /bin/csh -f
Xset FILE = /tmp/watch$$
Xset MAILTO = ( doc )
Xcd /usr/local/lib/watcher
Xwatcher > $FILE
Xif (-z $FILE) then
X echo " " | mail -s "`hostname` had no problems at `date`" $MAILTO
Xelse
X mail -s "System problem report for `hostname` at `date`" $MAILTO < $FILE
Xendif
X
Xrm -f $FILE
END_OF_Support/syswatch
if test 286 -ne `wc -c <Support/syswatch`; then
echo shar: \"Support/syswatch\" unpacked with wrong size!
fi
chmod +x Support/syswatch
# end of overwriting check
fi
if test -f baderr.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"baderr.c\"
else
echo shar: Extracting \"baderr.c\" \(853 characters\)
sed "s/^X//" >baderr.c <<'END_OF_baderr.c'
X/*
X baderr: A bad error has been caught. Print an error message explaining
X why we are dying, then exit.
X
X Assumptions:
X the signal number in sig is an actual signal number and has an
X entry in sys_siglist (BSD ONLY).
X
X Arguments:
X sig: the number of the signal which brought us here.
X
X Author:
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xbaderr(sig)
Xint sig;
X{
X extern char *sys_siglist[];
X
X printf(" \n");
X printf(" \n");
X#ifdef BSD
X printf(">> Unrecoverable error. %s Bye. <<\n",sys_siglist[sig]);
X#else
X printf(">> Unrecoverable error. Signal %d. Bye. <<\n", sig);
X#endif
X printf(" \n");
X printf(" \n");
X exit(1);
X}
END_OF_baderr.c
if test 853 -ne `wc -c <baderr.c`; then
echo shar: \"baderr.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f check_item.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"check_item.c\"
else
echo shar: Extracting \"check_item.c\" \(2976 characters\)
sed "s/^X//" >check_item.c <<'END_OF_check_item.c'
X/*
X check_item: given a value and a change format structure, make sure
X that the value is in range.
X
X Basically, this routine is a large switch statement on the type of
X change that grabs the necessary info, and checks to see if the item
X is worth mentioning.
X
X Note that what we print out depends on whether or not something else
X has been found wrong on this line.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xcheck_item(cf, value, cmd, line, prev_val)
Xchar *value, *cmd, *line;
Xstruct change_fmt_st *cf;
Xunion all_u *prev_val;
X{
X extern int line_ok, cmd_ok;
X double v, pct_chg, abs_chg;
X int len;
X
X v = atof(value);
X len = strlen(value);
X
X switch(cf->type) {
X case PERCENT:
X if (prev_val == NULL) /* nothing to compare with */
X return;
X if (v == 0) /* avoid divide by 0 */
X return;
X pct_chg = (v - prev_val->floatval) / v;
X if (pct_chg > cf->fmt.percent) {
X if (line_ok) {
X printf("%s had ", cmd);
X printf("%s change by more than %.2f percent.\n",
X cf->name, cf->fmt.percent*100);
X printf("%s\n",line);
X }
X else {
X printf("Also, it had ");
X printf("%s change by more than %.2f percent.\n",
X cf->name, cf->fmt.percent*100);
X }
X printf("Previous value %.2f; ",
X prev_val->floatval);
X printf("current value %.2f.\n", v);
X line_ok = False;
X }
X break;
X case ABSOLUTE:
X if (prev_val == NULL) /* nothing to compare with */
X return;
X abs_chg = v - prev_val->floatval;
X if (abs_chg > cf->fmt.abs_amount) {
X if (line_ok) {
X printf("%s had ", cmd);
X printf("%s change by more than %.2f.\n",
X cf->name, cf->fmt.abs_amount);
X printf("%s\n",line);
X }
X else {
X printf("Also, it had ");
X printf("%s change by more than %.2f.\n",
X cf->name, cf->fmt.abs_amount);
X }
X printf("Previous value %.2f; ",
X prev_val->floatval);
X printf("current value %.2f.\n", v);
X line_ok = False;
X }
X break;
X case MAX_MIN:
X if (v > cf->fmt.max_min.max || v < cf->fmt.max_min.min) {
X if (line_ok) {
X printf("%s has a ", cmd);
X printf("max/min value out of range:\n");
X printf("%s\n",line);
X }
X else {
X printf("Also, it has a ");
X printf("max/min value out of range:\n");
X }
X printf("where %s = %.2f; ", cf->name, v);
X printf("valid range %.2f to %.2f.\n",
X cf->fmt.max_min.min, cf->fmt.max_min.max);
X line_ok = False;
X }
X break;
X case STRING:
X if (strncmp(cf->fmt.str_value, value, len) != 0) {
X if (line_ok) {
X printf("%s has a string ", cmd);
X printf("value which is not valid:\n");
X printf("%s\n",line);
X }
X else {
X printf("Also, it has a string");
X printf("value which is not valid:\n");
X }
X printf("where %s = '%s'; Should be '%s'\n",
X cf->name, value, cf->fmt.str_value);
X line_ok = False;
X }
X break;
X default:
X printf("check_item: impossible condition\n");
X break;
X }
X cmd_ok = line_ok;
X}
END_OF_check_item.c
if test 2976 -ne `wc -c <check_item.c`; then
echo shar: \"check_item.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f checkline.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"checkline.c\"
else
echo shar: Extracting \"checkline.c\" \(1962 characters\)
sed "s/^X//" >checkline.c <<'END_OF_checkline.c'
X/*
X checkline: check the input line just read in against what we expect
X to find and report any problems. Actually, most of the work is done
X by check_item. We just identify what needs to be checked.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xcheckline(cmd, line, prev_res)
Xstruct cmd_st *cmd;
Xchar *line;
Xstruct old_cmd_st *prev_res;
X{
X extern int line_ok;
X extern FILE *hf;
X struct change_fmt_st *cf;
X union out_fmt_u of;
X union all_u *prev_value;
X char value[MAX_STR];
X char key_val[MAX_STR];
X char *cmd_name;
X
X cmd_name = (cmd->alias != NULL ? cmd->alias : cmd->pipeline);
X
X save_key(cmd, line, key_val); /* side effect: return key value */
X /* for each change format item */
X line_ok = True;
X for (cf=cmd->change_fmt; cf; cf=cf->next) {
X /* find the output format entry for this item */
X if (!find_of(cmd->out_type, cf->name, cmd->out_fmt, &of)) {
X fprintf(stderr, "Warning: %s appears in change list ",
X cf->name);
X fprintf(stderr, "but not in output format for %s\n",
X cmd_name);
X continue;
X }
X
X /*
X find the part of the line corresponding to check.
X Also find the previous results corresponding to the
X key for this line (if any).
X */
X switch (cmd->out_type) {
X case RELATIVE:
X find_prev_value(prev_res, of.rel_fmt->name,
X key_val, &prev_value);
X if (get_rel_field(line, of.rel_fmt->field, value) !=
X NULL)
X check_item(cf, value, cmd_name, line,
X prev_value);
X break;
X case COLUMN:
X find_prev_value(prev_res, of.rel_fmt->name,
X key_val, &prev_value);
X if (get_col_field(line, of.col_fmt->start, of.col_fmt->end, value) != NULL)
X check_item(cf, value, cmd_name, line,
X prev_value);
X break;
X }
X
X /*
X save the value in the history file for future
X comparisons.
X */
X if (cmd->key.rel_fmt != NULL)
X fprintf(hf, "\t\t%s %c %s\n", cf->name,
X TCHAR(cf->type), value);
X }
X if (!line_ok)
X printf("---------\n");
X}
END_OF_checkline.c
if test 1962 -ne `wc -c <checkline.c`; then
echo shar: \"checkline.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f clean_hist.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"clean_hist.c\"
else
echo shar: Extracting \"clean_hist.c\" \(983 characters\)
sed "s/^X//" >clean_hist.c <<'END_OF_clean_hist.c'
X/*
X clean_hist: free up all memory used by the history structure. The
X main purpose of this is to use the range checking of malloc to check
X for pointer problems. If you don't have source, this is of dubious
X value.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xclean_hist()
X{
X extern struct old_cmd_st *chead;
X struct old_cmd_st *bcp;
X struct val_st *vp, *bvp;
X struct key_st *kp, *bkp;
X
X while (chead != NULL) {
X printf("freeing '%s'\n",chead->pipeline);
X free(chead->pipeline);
X kp = chead->keys;
X while (kp != NULL) {
X printf("\tfreeing '%s'\n",kp->key_value);
X free(kp->key_value);
X vp = kp->vals;
X while (vp != NULL) {
X printf("\tfreeing '%s'\n",kp->vals->name);
X free(vp->name);
X bvp = vp;
X vp = vp->next;
X free(bvp);
X }
X bkp = kp;
X kp = kp->next;
X free(bkp);
X }
X bcp = chead;
X chead = chead->next;
X free(bcp);
X }
X printf("\nThe world is now free again! Aren't you glad?\n");
X}
END_OF_clean_hist.c
if test 983 -ne `wc -c <clean_hist.c`; then
echo shar: \"clean_hist.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f do_args.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"do_args.c\"
else
echo shar: Extracting \"do_args.c\" \(1260 characters\)
sed "s/^X//" >do_args.c <<'END_OF_do_args.c'
X/*
X do_args: parse the comand line arguments and set variables related to
X them.
X
X Copied from main:
X watcher [-p] [-v] [-h histfile] [-f controlfile]
X
X -p : pretty print control file as a verification of parse
X (default no pretty print). This option prevents
X processing of control file.
X -v : be verbose.
X -h : file in which to save output for future compare (default
X ./watcher.history).
X -f : controlfile to use (default ./DEF_CONTROL{,2}).
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xdo_args(argc, argv)
Xint argc;
Xchar *argv[];
X{
X extern int pflag, cflag, vflag;
X extern char controlname[], histfilename[];
X
X register int i;
X
X /* defaults */
X pflag = False;
X cflag = False;
X vflag = False;
X (void) sprintf(histfilename, "%s", DEF_HISTFILE);
X
X for (i=1; i<argc; i++) {
X if (argv[i][0] == '-') {
X switch(argv[i][1]) {
X case 'v':
X vflag = True;
X break;
X case 'p':
X pflag = True;
X break;
X case 'h':
X i = getargv(histfilename, argv, i,
X "history file name");
X break;
X case 'f':
X i = getargv(controlname, argv, i,
X "controlfile name");
X cflag = True;
X break;
X default:
X fprintf(stderr, "Unknown flag '%s'\n", argv[i]);
X exit(1);
X }
X }
X }
X}
END_OF_do_args.c
if test 1260 -ne `wc -c <do_args.c`; then
echo shar: \"do_args.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f doit.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"doit.c\"
else
echo shar: Extracting \"doit.c\" \(1048 characters\)
sed "s/^X//" >doit.c <<'END_OF_doit.c'
X/*
X doit: here is where the real purpose of the program is actually
X carried out.
X
X for each command, run it and look for problems.
X write results from this run back out.
X
X need to change from popen to ps_open.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xdoit()
X{
X extern struct cmd_st *clist;
X extern int vflag;
X extern int cmd_ok;
X extern FILE *hf;
X
X char line[MAX_STR];
X struct cmd_st *p;
X struct old_cmd_st *prev_results, *find_prev_cmd();
X FILE *ps, *popen();
X
X /* run commands */
X for (p=clist; p != NULL; p=p->next) {
X cmd_ok = True;
X if (vflag)
X printf("Executing: '%s'\n\n", p->pipeline);
X if (p->key.rel_fmt != NULL)
X fprintf(hf, "%s\n", p->pipeline);
X /* get prev results for comparison */
X prev_results = find_prev_cmd(p->pipeline);
X ps = popen(p->pipeline, "r");
X while (fgets(line, MAX_STR, ps) != NULL) {
X line[strlen(line)-1] = '\0';
X if (vflag)
X printf(" Read: '%s'\n",line);
X checkline(p, line, prev_results);
X }
X if (!cmd_ok)
X printf("\n");
X }
X}
END_OF_doit.c
if test 1048 -ne `wc -c <doit.c`; then
echo shar: \"doit.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f externs.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"externs.c\"
else
echo shar: Extracting \"externs.c\" \(429 characters\)
sed "s/^X//" >externs.c <<'END_OF_externs.c'
X/*
X externs: external variable declarations.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
XFILE *cf, *hf;
X
Xint intval, ointval;
Xchar *strval, ostrval[MAX_STR];
Xchar pipeline[MAX_STR];
X
Xstruct cmd_st *clist = NULL;
Xstruct old_cmd_st *chead = NULL;
X
Xint parse_error = False;
X
Xint pflag, cflag, vflag;
Xchar histfilename[MAX_STR];
Xchar controlname[MAX_STR];
Xint line_ok;
Xint cmd_ok;
END_OF_externs.c
if test 429 -ne `wc -c <externs.c`; then
echo shar: \"externs.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f find_of.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"find_of.c\"
else
echo shar: Extracting \"find_of.c\" \(749 characters\)
sed "s/^X//" >find_of.c <<'END_OF_find_of.c'
X/*
X find_of: find the output format corresponding for the field whose
X name we are given.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xfind_of(type, name, of_head, of)
Xint type;
Xunion out_fmt_u of_head, *of;
Xchar *name;
X{
X struct rel_out_st *rf;
X struct col_out_st *cf;
X
X switch(type) {
X case RELATIVE:
X for (rf=of_head.rel_fmt; rf; rf=rf->next)
X if (strcmp(name, rf->name) == 0) {
X (*of).rel_fmt = rf;
X return True;
X }
X break;
X case COLUMN:
X for (cf=of_head.col_fmt; cf; cf=cf->next)
X if (strcmp(name, cf->name) == 0) {
X (*of).col_fmt = cf;
X return True;
X }
X break;
X default:
X fprintf(stderr,"internal error; unknown type in find_of\n");
X exit(1);
X }
X
X return False;
X}
END_OF_find_of.c
if test 749 -ne `wc -c <find_of.c`; then
echo shar: \"find_of.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f find_pre_cmd.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"find_pre_cmd.c\"
else
echo shar: Extracting \"find_pre_cmd.c\" \(441 characters\)
sed "s/^X//" >find_pre_cmd.c <<'END_OF_find_pre_cmd.c'
X/*
X find_prev_cmd: find the previous results for this command (if any). If
X there is no match, then we return NULL.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xstruct old_cmd_st *
Xfind_prev_cmd(pipeline)
Xchar *pipeline;
X{
X extern struct old_cmd_st *chead;
X struct old_cmd_st *cp;
X
X cp = chead;
X while (cp != NULL && strcmp(pipeline, cp->pipeline) != 0)
X cp = cp->next;
X
X return cp;
X}
END_OF_find_pre_cmd.c
if test 441 -ne `wc -c <find_pre_cmd.c`; then
echo shar: \"find_pre_cmd.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f find_pre_val.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"find_pre_val.c\"
else
echo shar: Extracting \"find_pre_val.c\" \(904 characters\)
sed "s/^X//" >find_pre_val.c <<'END_OF_find_pre_val.c'
X/*
X find_prev_value: given an old command structure, look through the
X prior output for the output for name of type type. Return it in the
X union pointed to by value or NULL if not found.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xfind_prev_value(cmd, field_name, key_val, value)
Xstruct old_cmd_st *cmd;
Xchar *field_name, *key_val;
Xunion all_u **value;
X{
X struct val_st *vp;
X struct key_st *kp;
X
X *value = NULL;
X
X if (cmd == NULL)
X return;
X if (cmd->keys == NULL)
X return;
X
X /* find the correct keyword */
X for (kp=cmd->keys; kp != NULL; kp=kp->next) {
X if (strcmp(kp->key_value, key_val) == 0)
X break;
X }
X
X if (kp == NULL)
X return;
X if (kp->vals == NULL)
X return;
X
X /* find the value under the keyword */
X for (vp=kp->vals; vp != NULL; vp=vp->next) {
X if (strcmp(vp->name, field_name) == 0) {
X *value = &(vp->val);
X return;
X }
X }
X}
END_OF_find_pre_val.c
if test 904 -ne `wc -c <find_pre_val.c`; then
echo shar: \"find_pre_val.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f get_col_fld.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"get_col_fld.c\"
else
echo shar: Extracting \"get_col_fld.c\" \(787 characters\)
sed "s/^X//" >get_col_fld.c <<'END_OF_get_col_fld.c'
X/*
X get_col_field: we have a column output format and want a certain
X part of it. Place the it in 'value' if it exists and return a
X pointer to the string. If there are problems place a null terminated
X zero length string in 'value' and return NULL as an error condition.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xchar *
Xget_col_field(line, start, end, value)
Xchar *line, *value;
Xint start, end;
X{
X int len;
X
X value[0] = '\0';
X len = strlen(line);
X
X /* error checking */
X if (start > len || end > len) {
X fprintf(stderr,"'%s' has %d columns. Specified were %d to %d\n",
X line, len, start, start);
X return NULL;
X }
X len = end - start + 1;
X (void) strncpy(value, &line[start-1], len);
X value[len] = '\0';
X
X return value;
X}
END_OF_get_col_fld.c
if test 787 -ne `wc -c <get_col_fld.c`; then
echo shar: \"get_col_fld.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f get_rel_fld.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"get_rel_fld.c\"
else
echo shar: Extracting \"get_rel_fld.c\" \(945 characters\)
sed "s/^X//" >get_rel_fld.c <<'END_OF_get_rel_fld.c'
X/*
X get_rel_field: we have a relative output format and want a certain
X field from it. Place the field in 'value' if it exists and return a
X pointer to the string. If there are problems place a null terminated
X zero length string in 'value' and return NULL as an error condition.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xchar *
Xget_rel_field(line, field, value)
Xchar *line, *value;
Xint field;
X{
X int i;
X int nfields;
X char *vec[MAX_VEC];
X
X value[0] = '\0';
X
X /* break up the line */
X nfields = line_to_vec(line, vec, " \t");
X
X /* error checking */
X if (nfields <= 0) {
X fprintf(stderr, "'%s' didn't break up into fields.\n",line);
X return NULL;
X }
X if (nfields < field) {
X fprintf(stderr, "Warning: '%s' has only %d fields, not %d\n",
X line, nfields, field);
X return NULL;
X }
X
X (void) strcpy(value, vec[field-1]);
X
X for (i=0; i<nfields; i++)
X free(vec[i]);
X
X return value;
X}
END_OF_get_rel_fld.c
if test 945 -ne `wc -c <get_rel_fld.c`; then
echo shar: \"get_rel_fld.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f getargv.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"getargv.c\"
else
echo shar: Extracting \"getargv.c\" \(495 characters\)
sed "s/^X//" >getargv.c <<'END_OF_getargv.c'
X/*
X getargv: get a value from argv, whether it is immediately following
X the flag or is the next argument. Complain if not found and exit.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xgetargv(what, argv, i, err)
Xchar *what, *argv[], *err;
Xint i;
X{
X if (argv[i][2])
X (void) strcpy(what, &argv[i][2]);
X else {
X if (argv[++i])
X (void) strcpy(what, argv[i]);
X else {
X fprintf(stderr,"Missing %s!\n", err);
X exit(1);
X }
X }
X return i;
X}
END_OF_getargv.c
if test 495 -ne `wc -c <getargv.c`; then
echo shar: \"getargv.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f init.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"init.c\"
else
echo shar: Extracting \"init.c\" \(296 characters\)
sed "s/^X//" >init.c <<'END_OF_init.c'
X/*
X init: initialize any variables needing values, do any other
X initialization needed.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xinit()
X{
X extern char ostrval[], *strval;
X
X ostrval[0] = '\0';
X strval = ostrval;
X init_sigs();
X open_cf();
X}
END_OF_init.c
if test 296 -ne `wc -c <init.c`; then
echo shar: \"init.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f init_sigs.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"init_sigs.c\"
else
echo shar: Extracting \"init_sigs.c\" \(507 characters\)
sed "s/^X//" >init_sigs.c <<'END_OF_init_sigs.c'
X/*
X init_sigs: take care of setting up the signal handling.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xinit_sigs()
X{
X int exit(), baderr();
X
X (void) signal(SIGINT, exit);
X (void) signal(SIGHUP, exit);
X (void) signal(SIGQUIT, baderr);
X (void) signal(SIGSEGV, baderr);
X (void) signal(SIGBUS, baderr);
X (void) signal(SIGFPE, baderr);
X (void) signal(SIGILL, baderr);
X#ifdef BSD
X (void) signal(SIGTTIN, baderr);
X (void) signal(SIGTTOU, baderr);
X#endif
X}
END_OF_init_sigs.c
if test 507 -ne `wc -c <init_sigs.c`; then
echo shar: \"init_sigs.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f line_to_vec.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"line_to_vec.c\"
else
echo shar: Extracting \"line_to_vec.c\" \(1705 characters\)
sed "s/^X//" >line_to_vec.c <<'END_OF_line_to_vec.c'
X/*
X line_to_vec: take a string and separate it into a vector of strings,
X splitting it at characters which are supplied in 'splits'.
X Ignore any 'splits' at the beginning. Multiple 'splits' are
X condensed into one. Splits are discarded.
X
X Assumptions:
X line is null terminated.
X no single word is longer than MAX_STR.
X
X Arguments:
X line: line to split.
X vec: split line.
X splits: array of characters on which to split. Null terminated.
X
X Global data used:
X none.
X
X Local variables:
X len: length of 'word' so far.
X name: current entry in the vector we are building.
X word: pointer into name.
X
X Returns:
X The number of vectors created; -1 if malloc fails.
X The argument vec is left with a NULL pointer after the last word.
X
X Author:
X Kenneth Ingham
X
X Date:
X Thu Sep 5 13:59:21 MDT 1985
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xline_to_vec(line, vec, splits)
Xchar *line, *vec[], *splits;
X{
X register int i, v, j;
X register unsigned len;
X int n;
X char word[MAX_STR];
X
X if (line == NULL || line[0] == '\0')
X return 0;
X
X /* skip any splits in the beginning */
X for (i=0; line[i] && index(splits, line[i]) != 0; i++)
X ;
X
X j = 0;
X len = 0;
X v = 0;
X n = 0;
X while (line[i]) {
X if (index(splits, line[i]) != 0) {
X word[j] = '\0';
X vec[v] = malloc(len+1);
X if (vec[v] == NULL)
X return -1;
X strcpy(vec[v], word);
X j = 0;
X len = 0;
X v++;
X n++;
X i++;
X for ( ; line[i] && index(splits, line[i]) != 0; i++)
X ;
X }
X else {
X word[j++] = line[i++];
X len++;
X }
X }
X
X if (index(splits, line[i]) != 0) {
X word[j] = '\0';
X vec[v] = malloc(len+1);
X if (vec[v] == NULL)
X return -1;
X strcpy(vec[v], word);
X n++;
X }
X
X return n;
X}
END_OF_line_to_vec.c
if test 1705 -ne `wc -c <line_to_vec.c`; then
echo shar: \"line_to_vec.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f main.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"main.c\"
else
echo shar: Extracting \"main.c\" \(2114 characters\)
sed "s/^X//" >main.c <<'END_OF_main.c'
X/*
X main: main routine for the watcher program.
X
X read from a file describing commands (pipelines) to execute, formats of
X the output, max changes allowed (% or absolute), and max & min
X values for various fields.
X problems noticed are reported.
X as a side effect, be able to pretty print the description file
X (originally use to verify parsing).
X
X format:
X (command)\tformat :
X \tfield\tchange\tmax\tmin
X .
X .
X .
X
X See yacc file for complete grammar description of control file.
X
X outline of program:
X parse control file and build data structures.
X run each pipeline and compare output to previous output (only
X save relevant fields; save directory is either default
X or command line specified; no previous file or format
X changed (ie we are watching different or new fields) we
X create new file and next time we do compare).
X differences that are not allowable are reported.
X
X Usage of program:
X watcher [-p] [-v] [-h histfile] [-f controlfile]
X
X -p : pretty print control file as a verification of parse
X (default no pretty print). This option prevents
X processing of control file.
X -v : be verbose when doing work; useful for debugging.
X -h : file in which to save output for future compare (default
X ./watcher.history).
X -f : controlfile to use (default ./watcherfile or ./Watcherfile).
X
X Note that the basic data structures are all linear linked lists, with
X many items in the list being heads of other lists. When problems
X occur, get out the pencil and paper and start drawing the lists.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xmain(argc, argv)
Xint argc;
Xchar *argv[];
X{
X extern int parse_error;
X extern struct cmd_st *clist;
X extern int pflag;
X
X do_args(argc, argv);
X init();
X
X if (yyparse() == 1 || parse_error) {
X fprintf(stderr, "%s: parse error in control file.\n", NAME);
X exit(1);
X }
X
X if (clist == NULL) {
X fprintf(stderr, "No command list to execute!\n");
X exit(1);
X }
X
X if (pflag)
X pp(clist);
X else {
X read_hist();
X open_hf(); /* for writing our history */
X doit();
X }
X}
END_OF_main.c
if test 2114 -ne `wc -c <main.c`; then
echo shar: \"main.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f open_cf.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"open_cf.c\"
else
echo shar: Extracting \"open_cf.c\" \(897 characters\)
sed "s/^X//" >open_cf.c <<'END_OF_open_cf.c'
X/*
X open_cf: open the control file. It is either specified on the
X command line or DEF_CONTROL or DEF_CONTROL2.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xopen_cf()
X{
X extern char controlname[];
X extern FILE *cf;
X extern int cflag;
X
X if (cflag) { /* specified control file */
X cf = fopen(controlname, "r");
X if (cf == NULL) {
X fprintf(stderr, "Unable to open '%s'\n",
X controlname);
X exit(1);
X }
X }
X else { /* try defaults */
X if ((cf = fopen(DEF_CONTROL, "r")) == NULL) { /* #1 */
X if ((cf = fopen(DEF_CONTROL2, "r")) == NULL) { /*#2*/
X fprintf(stderr, "Unable to open %s or %s\n",
X DEF_CONTROL, DEF_CONTROL2);
X exit(1);
X }
X /* if we're here #2 must have worked */
X (void) strcpy(controlname, DEF_CONTROL2);
X }
X else
X /* if we're here #1 must have worked */
X (void) strcpy(controlname, DEF_CONTROL);
X }
X}
END_OF_open_cf.c
if test 897 -ne `wc -c <open_cf.c`; then
echo shar: \"open_cf.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f open_hf.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"open_hf.c\"
else
echo shar: Extracting \"open_hf.c\" \(373 characters\)
sed "s/^X//" >open_hf.c <<'END_OF_open_hf.c'
X/*
X open_hf: open the history file for writing out log of what we looked
X at this run.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xopen_hf()
X{
X extern char histfilename[];
X extern FILE *hf;
X
X hf = fopen(histfilename, "w");
X if (hf == NULL) {
X fprintf(stderr, "Unable to open '%s'\n",
X histfilename);
X exit(1);
X }
X}
END_OF_open_hf.c
if test 373 -ne `wc -c <open_hf.c`; then
echo shar: \"open_hf.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f pp.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"pp.c\"
else
echo shar: Extracting \"pp.c\" \(534 characters\)
sed "s/^X//" >pp.c <<'END_OF_pp.c'
X/*
X pp: pretty print the command structure.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xpp(clist)
Xstruct cmd_st *clist;
X{
X extern char histfilename[];
X extern char controlname[];
X
X printf("History file name: %s\n", histfilename);
X printf("Control file name: %s\n", controlname);
X printf("\n\n");
X
X while (clist != NULL) {
X printf("( %s )\n", clist->pipeline);
X pp_out(clist->out_type, clist->out_fmt);
X printf(" :\n");
X pp_change(clist->change_fmt);
X clist = clist->next;
X }
X}
END_OF_pp.c
if test 534 -ne `wc -c <pp.c`; then
echo shar: \"pp.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f pp_change.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"pp_change.c\"
else
echo shar: Extracting \"pp_change.c\" \(813 characters\)
sed "s/^X//" >pp_change.c <<'END_OF_pp_change.c'
X/*
X pp_change: pretty print the change format.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xpp_change(cf)
Xstruct change_fmt_st *cf;
X{
X while (cf != NULL) {
X switch(cf->type) {
X case PERCENT:
X printf("\t\t%s %5.2f %%", cf->name,
X cf->fmt.percent*100);
X break;
X case ABSOLUTE:
X printf("\t\t%s %6.2f", cf->name,
X cf->fmt.abs_amount);
X break;
X case MAX_MIN:
X printf("\t\t%s %6.2f %6.2f", cf->name,
X cf->fmt.max_min.min,
X cf->fmt.max_min.max);
X break;
X case STRING:
X printf("\t\t%s \"%s\"", cf->name,
X cf->fmt.str_value);
X break;
X default:
X printf("Impossible change format type: %d\n",
X cf->type);
X break;
X }
X if (cf->next != NULL)
X printf(" ;\n");
X else
X printf(" .\n");
X cf = cf->next;
X }
X}
END_OF_pp_change.c
if test 813 -ne `wc -c <pp_change.c`; then
echo shar: \"pp_change.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f pp_out.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"pp_out.c\"
else
echo shar: Extracting \"pp_out.c\" \(696 characters\)
sed "s/^X//" >pp_out.c <<'END_OF_pp_out.c'
X/*
X pp_out: pretty print the output of command structure.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xpp_out(type,of)
Xint type;
Xunion out_fmt_u of;
X{
X struct rel_out_st *rp;
X struct col_out_st *cp;
X
X switch(type) {
X case RELATIVE:
X rp = of.rel_fmt;
X while (rp != NULL) {
X printf(" %d %s %% %c", rp->field, rp->name,
X TCHAR(rp->type));
X rp = rp->next;
X }
X break;
X case COLUMN:
X cp = of.col_fmt;
X while (cp != NULL) {
X printf(" %d - %d %s %% %c", cp->start,
X cp->end, cp->name, TCHAR(cp->type));
X cp = cp->next;
X }
X break;
X default:
X printf("Impossible value for outfmt type: %d\n", type);
X break;
X }
X}
END_OF_pp_out.c
if test 696 -ne `wc -c <pp_out.c`; then
echo shar: \"pp_out.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f read_hist.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"read_hist.c\"
else
echo shar: Extracting \"read_hist.c\" \(3404 characters\)
sed "s/^X//" >read_hist.c <<'END_OF_read_hist.c'
X/*
X read_hist: open the file containing the results of our last run. If
X it is not there then we assume that this is a first run and set the
X head of the list to NULL. Otherwise, we read the results of the previous
X run into a mess of a data structure for later use in comparisons.
X
X Assumed format of history file:
X
X command
X key
X name type value
X
X witht the following definitions:
X command: pipeline that was executed
X key: value of key on line
X name: output field name
X type: field type (same as defined in output format)
X value: what was in the field.
X
X We create a linked list of linked lists of linked lists. Improvement
X would be to change to a tree of some sort to speed up searches.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xread_hist()
X{
X extern char histfilename[];
X extern FILE *hf;
X extern int vflag;
X extern struct old_cmd_st *chead;
X
X char line[MAX_STR];
X struct old_cmd_st *cp;
X struct val_st *vp;
X struct key_st *kp;
X int len;
X char *sp;
X
X if (vflag)
X printf("Using %s for historyfile\n", histfilename);
X
X hf = fopen(histfilename, "r");
X if (hf == NULL) {
X if (vflag)
X printf("This is a first run.\n");
X chead = NULL;
X return;
X }
X
X chead = NULL; cp = NULL; kp = NULL;
X while (fgets(line, MAX_STR, hf) != NULL) {
X line[strlen(line)-1] = '\0'; /* kill trailing cr */
X if (line[0] != '\t') { /* command */
X if (chead == NULL) {
X cp = (struct old_cmd_st *)malloc(sizeof(struct old_cmd_st));
X chead = cp;
X }
X else {
X cp->next = (struct old_cmd_st *)malloc(sizeof(struct old_cmd_st));
X cp = cp->next;
X }
X
X cp->pipeline = malloc((unsigned)strlen(line)+1);
X (void) strcpy(cp->pipeline, line);
X cp->next = NULL;
X }
X else if (line[0] == '\t' && line[1] != '\t') { /* key */
X if (cp == NULL) {
X printf("Bad history file: keyword found ");
X printf("before pipeline. Ignoring history ");
X printf("file.\n");
X chead = NULL;
X return;
X }
X if (cp->keys == NULL) {
X cp->keys = (struct key_st *)malloc(sizeof(struct key_st));
X kp = cp->keys;
X }
X else {
X kp->next = (struct key_st *)malloc(sizeof(struct key_st));
X kp = kp->next;
X }
X kp->next = NULL;
X kp->key_value = malloc((unsigned)strlen(line)+1);
X (void) strcpy(kp->key_value, &line[1]);
X }
X else if (line[0] == '\t' && line[1] == '\t') { /* vals */
X if (kp == NULL) {
X printf("Bad history file: value found ");
X printf("before keyword. Ignoring history ");
X printf("file.\n");
X chead = NULL;
X return;
X }
X if (kp->vals == NULL) {
X kp->vals = (struct val_st *)malloc(sizeof(struct val_st));
X vp = kp->vals;
X }
X else {
X vp->next = (struct val_st *)malloc(sizeof(struct val_st));
X vp = vp->next;
X }
X vp->next = NULL;
X sp = index(&line[2], ' ');
X len = sp - &line[2];
X vp->name = malloc((unsigned)len+1);
X (void) strncpy(vp->name, &line[2], len);
X vp->name[len] = '\0';
X sp++;
X len = strlen(line) - len;
X switch (*sp) {
X case 's':
X vp->val.strval = malloc((unsigned)len+1);
X (void) strcpy(vp->val.strval, sp+2);
X vp->type = STRING;
X break;
X case 'd':
X vp->val.floatval = atof(sp+2);
X vp->type = FLOAT;
X break;
X default:
X /* bad condition */
X printf("Unknown data type in history file.\n");
X printf("Offending line: %s\n",line);
X exit(1);
X break;
X }
X }
X }
X (void) fclose(hf);
X}
END_OF_read_hist.c
if test 3404 -ne `wc -c <read_hist.c`; then
echo shar: \"read_hist.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f save_key.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"save_key.c\"
else
echo shar: Extracting \"save_key.c\" \(1239 characters\)
sed "s/^X//" >save_key.c <<'END_OF_save_key.c'
X/*
X save_key: save the key for this line in the history file so that
X we can use it next run.
X
X Assume that the keyory file is opened and that the pipeline has
X already been placed in the file. We place the data in the file in
X the following format:
X
X pipeline
X key_value
X name value
X .
X .
X .
X .
X .
X .
X pipeline
X .
X .
X .
X
X
X Side effect: return the actual key value for other parts of the
X program to use.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xsave_key(cmd, line, key_val)
Xstruct cmd_st *cmd;
Xchar *line, *key_val;
X{
X extern int vflag;
X extern FILE *hf; /* assumed to be already open */
X
X if (cmd->key.col_fmt == NULL) { /* no key; no reason to save. */
X if (vflag)
X printf("%s has no key field.\n",cmd->pipeline);
X return;
X }
X
X switch (cmd->out_type) {
X case RELATIVE:
X (void) get_rel_field(line, cmd->key.rel_fmt->field,
X key_val);
X break;
X case COLUMN:
X (void) get_col_field(line, cmd->key.col_fmt->start,
X cmd->key.col_fmt->end, key_val);
X break;
X }
X
X if (!key_val[0]) {
X if (vflag)
X printf("the key field for %s is empty\n",cmd->pipeline);
X return;
X }
X
X fprintf(hf, "\t%s\n",key_val);
X}
END_OF_save_key.c
if test 1239 -ne `wc -c <save_key.c`; then
echo shar: \"save_key.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f yyerror.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"yyerror.c\"
else
echo shar: Extracting \"yyerror.c\" \(138 characters\)
sed "s/^X//" >yyerror.c <<'END_OF_yyerror.c'
X/*
X wow.
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xyyerror(s)
Xchar *s;
X{
X fprintf(stderr,"%s\n", s);
X}
END_OF_yyerror.c
if test 138 -ne `wc -c <yyerror.c`; then
echo shar: \"yyerror.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
if test -f yylex.c -a "${1}" != "-c" ; then
echo shar: Will not over-write existing file \"yylex.c\"
else
echo shar: Extracting \"yylex.c\" \(1804 characters\)
sed "s/^X//" >yylex.c <<'END_OF_yylex.c'
X/*
X yylex for watcher: this is a simple routine looking for numbers,
X special characters and strings. The special chars are stored in
X 'words' and represent tokens by themselves. In y.tab.h are the
X values to return for the various tokens which are not listed in
X 'words'.
X
X Kenneth Ingham
X
X Copyright (C) 1987 The University of New Mexico
X*/
X
X#include "defs.h"
X
Xchar words[] = "\".*|;:%@$-{}";
X
Xyylex()
X{
X extern int yylval, intval, ointval;
X extern char *strval, ostrval[], pipeline[];
X extern FILE *cf;
X int c, value, i;
X static char str[MAX_STR];
X
X while (isspace(c = getc(cf)))
X ;
X
X if (c == EOF)
X return EOF;
X
X if (c == '(') { /* aha, pipeline */
X c = getc(cf);
X for (i=0; c != EOF && c != ')'; i++) {
X str[i] = c;
X c = getc(cf);
X }
X str[i] = '\0';
X if (c == EOF) {
X fprintf(stderr, "Missing ')' to end pipeline.\n");
X return EOF;
X }
X (void) strcpy(pipeline, str);
X return PIPELINE;
X }
X
X if (c == '#') { /* comment to end of line */
X while (c != '\n' && c != EOF)
X c = getc(cf);
X if (c == EOF)
X return EOF;
X return yylex();
X }
X
X if (index(words, c) != 0) {
X yylval = c;
X return c;
X }
X
X if (isdigit(c)) { /* a number */
X value = c - '0';
X while (isdigit(c = getc(cf)))
X value = value * 10 + c - '0';
X ointval = intval;
X intval = value;
X (void) ungetc(c, cf);
X return NUMBER;
X }
X
X (void) strcpy(ostrval, strval);
X
X if (c == '\'') { /* literal string */
X c = getc(cf);
X for (i=0; c != EOF && c != '\''; i++) {
X str[i] = c;
X c = getc(cf);
X }
X str[i] = '\0';
X strval = str;
X return STRING;
X }
X
X /* nothing else matched. Must be plain string (whitespace sep) */
X for (i=1, str[0]=c; c != EOF && !isspace(c) && !index(words,c); i++) {
X c = getc(cf);
X str[i] = c;
X }
X (void) ungetc(c, cf);
X str[i-1] = '\0';
X strval = str;
X return STRING;
X}
END_OF_yylex.c
if test 1804 -ne `wc -c <yylex.c`; then
echo shar: \"yylex.c\" unpacked with wrong size!
fi
# end of overwriting check
fi
echo shar: End of archive 1 \(of 2\).
cp /dev/null ark1isdone
MISSING=""
for I in 1 2 ; do
if test ! -f ark${I}isdone ; then
MISSING="${MISSING} ${I}"
fi
done
if test "${MISSING}" = "" ; then
echo You have unpacked both archives.
rm -f ark[1-9]isdone
else
echo You still need to unpack the following archives:
echo " " ${MISSING}
fi
## End of shell archive.
emenmen