dmbcs/i-bash
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Imported from ../bash-2.0.tar.gz.

Browse source
This commit is contained in:
Jari Aalto 1996-12-23 17:02:34 +00:00
commit ccc6cda312
502 changed files with 91988 additions and 69123 deletions
Download patch file Download diff file Expand all files Collapse all files

59
examples/scripts.v2/PERMISSION Normal file
View file

@ -0,0 +1,59 @@
From spcecdt@armory.com Wed May 10 10:21:11 1995
Flags: 10
Return-Path: spcecdt@armory.com
Received: from po.cwru.edu (root@po.CWRU.Edu [129.22.4.2]) by odin.INS.CWRU.Edu with ESMTP (8.6.10+cwru/CWRU-2.1-ins)
id KAA22876; Wed, 10 May 1995 10:21:10 -0400 (from spcecdt@armory.com for <chet@odin.INS.CWRU.Edu>)
Received: from deepthought.armory.com (mmdf@deepthought.armory.com [192.122.209.42]) by po.cwru.edu with SMTP (8.6.10+cwru/CWRU-2.3)
id BAA16354; Wed, 10 May 1995 01:33:22 -0400 (from spcecdt@armory.com for <chet@po.cwru.edu>)
From: John DuBois <spcecdt@armory.com>
Date: Tue, 9 May 1995 22:33:12 -0700
In-Reply-To: Chet Ramey <chet@odin.ins.cwru.edu>
"ksh scripts" (May 9, 1:36pm)
X-Www: http://www.armory.com/~spcecdt/
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: chet@po.cwru.edu
Subject: Re: ksh scripts
Message-ID: <9505092233.aa13001@deepthought.armory.com>
Sure. The canonical versions are available on ftp.armory.com; you might
want to pick up the latest versions before modifying them.
John
On May 9, 1:36pm, Chet Ramey wrote:
} Subject: ksh scripts
} From odin.ins.cwru.edu!chet Tue May 9 10:39:51 1995
} Received: from odin.INS.CWRU.Edu by deepthought.armory.com id aa22336;
} 9 May 95 10:39 PDT
} Received: (chet@localhost) by odin.INS.CWRU.Edu (8.6.10+cwru/CWRU-2.1-ins)
} id NAA20487; Tue, 9 May 1995 13:39:24 -0400 (from chet)
} Date: Tue, 9 May 1995 13:36:54 -0400
} From: Chet Ramey <chet@odin.ins.cwru.edu>
} To: john@armory.com
} Subject: ksh scripts
} Cc: chet@odin.ins.cwru.edu
} Reply-To: chet@po.cwru.edu
} Message-ID: <9505091736.AA20411.SM@odin.INS.CWRU.Edu>
} Read-Receipt-To: chet@po.CWRU.Edu
} MIME-Version: 1.0
} Content-Type: text/plain; charset=us-ascii
} Status: OR
}
} Hi. I'm the maintainer of bash (the GNU `Bourne Again shell') for
} the FSF.
}
} I picked up a tar file of ksh scripts you wrote from an anon FTP site
} a while back. I'd like your permission to include modified versions
} of some of them in the next major bash distribution (with proper credit
} given, of course). Is it OK if I do that?
}
} Chet Ramey
}
} --
} ``The lyf so short, the craft so long to lerne.'' - Chaucer
}
} Chet Ramey, Case Western Reserve University Internet: chet@po.CWRU.Edu
}-- End of excerpt from Chet Ramey

33
examples/scripts.v2/README Normal file
View file

@ -0,0 +1,33 @@
This collection of scripts was originally written for ksh-88 by
John DuBois <spcecdt@armory.com>. The conversion to bash v2
syntax was done by Chet Ramey.
These scripts are as-is; there is no copyright associated with
any of them. They exist simply as examples of bash scripting.
Here's a description of what's in this directory:
arc2tarz Convert an "arc" archive to a compressed tar archive.
corename Tell what produced a core file.
fman Fast man replacement.
frcp Copy files using ftp but with rcp-type command line syntax.
lowercase Change filenames to lower case.
ncp A nicer front end for cp (has -i, etc.).
newext Change the extension of a group of files.
nmv A nicer front end for mv (has -i, etc.).
pages Print specified pages from files.
pf A pager front end that handles compressed files.
rename Change the names of files that match a pattern.
repeat Execute a command multiple times.
untar Unarchive a (possibly compressed) tarfile into a directory.
uudec Carefully uudecode multiple files.
uuenc uuencode multiple files.
vtree Print a visual display of a directory tree.
where Show where commands that match a pattern are.
The following scripts were written or converted by Chet Ramey:
bashrand Random number generator with upper and lower bounds and optional seed
cdhist cd replacement with a directory stack added
pmtop Poor man's `top' for SunOS 4.x and BSD/OS
shprof Line profiler for bash scripts

85
examples/scripts.v2/arc2tarz Normal file
View file

@ -0,0 +1,85 @@
#! /bin/bash
#
# original from:
# arc2tarz: convert arced file to tarred, compressed form.
# @(#) arc2tarz.ksh 1.0 92/02/16
# 91/03/28 john h. dubois iii (john@armory.com)
# 92/02/16 added -h option for help
#
# conversion to bash v2 syntax by Chet Ramey
unset ENV
Usage="Usage: $0 arcfile [-hcg] [ tarzfile ]"
phelp()
{
echo "$Usage
arcfile is the name of an arc file to convert to tarred, compressed form.
The file must have a .arc extension, but only the base name needs to be
given. If no output file name is given, it will be created in the current
directory with the name being the arcfile basename followed by .tar.EXT.
If the -c option is given, compress will be used, and EXT will be Z.
The default (also available with -g) is to use gzip, in which case EXT
is gz. If the basename is too long the extension may be truncated. All
uppercase letters in the names of files in the archive are moved to lowercase."
}
compress=gzip
ext=gz
while getopts "hcg" opt; do
case "$opt" in
h) phelp; exit 0;;
c) compress=compress; ext=Z;;
g) compress=gzip ; ext=gz ;;
*) echo "$Usage" 1>&2 ; exit 2;;
esac
done
shift $((OPTIND - 1))
if [ $# = 0 ]; then
phelp
exit 0
fi
[ -z "$TMP" ] && tmpdir=/tmp/arc2tarz.$$ || tmpdir=$TMP/arc2tarz.$$
case "$1" in
*.arc) arcfile=$1 ;;
*) arcfile=$1.arc ;;
esac
if [ ! -f $arcfile ] || [ ! -r $arcfile ]; then
echo "Could not open arc file \"$arcfile\"."
exit 1
fi
case "$arcfile" in
/*) ;;
*) arcfile=$PWD/$arcfile ;;
esac
basename=${arcfile%.arc}
basename=${basename##*/}
[ $# -lt 2 ] && tarzname=$PWD/$basename.tar.$ext || tarzname=$2
trap 'rm -rf $tmpdir $tarzname' 1 2 3 6 15
mkdir $tmpdir
cd $tmpdir
echo "unarcing files..."
arc -ie $arcfile
# lowercase
for f in *; do
new=$(echo $f | tr A-Z a-z)
if [ "$f" != "$new" ]; then
mv $f $new
fi
done
echo "tarring/compressing files..."
tar cf - * | $compress > $tarzname
cd -
rm -rf $tmpdir

76
examples/scripts.v2/bashrand Normal file
View file

@ -0,0 +1,76 @@
#! /bin/bash
# bashrand - generate a random number in a specified range with an
# optionally specified ``seed'' value.
#
# Original Author: Peter Turnbull, May 1993
usage()
{
echo "$PROG: usage: $PROG [-s seed] lower-limit upper-limit" >&2
}
PROG=${0##*/}
SEED=$$ # Initialize random-number seed value with PID
while getopts s: opt
do
case "$opt" in
s) SEED=$OPTARG ;;
*) usage ; exit 2 ;;
esac
done
shift $((OPTIND - 1))
# Process command-line arguments:
case $# in
2) Lower=$1; Upper=$2 ;;
*) usage ; exit 2 ;;
esac
# Check that specified values are integers:
expr "$Lower" + 0 >/dev/null 2>&1 || {
echo "$PROG: lower ($Lower) not an integer" >&2
exit 1
}
expr "$Upper" + 0 >/dev/null 2>&1 || {
echo "$PROG: upper ($Upper) not an integer" >&2
exit 1
}
expr "$SEED" + 0 >/dev/null 2>&1 || {
echo "$PROG: seed ($SEED) not an integer" >&2
exit 1
}
# Check that values are in the correct range:
(( $Lower < 0 )) || [ `expr "$Lower" : '.*'` -gt 5 ] && {
echo "$PROG: Lower limit ($Lower) out of range" >&2
exit 1
}
(( $Upper > 32767 )) || [ `expr "$Upper" : '.*'` -gt 5 ] && {
echo "$PROG: Upper limit ($Upper) out of range" >&2;
exit 1
}
(( $SEED < 0 )) || (( $SEED > 32767 )) || [ `expr "$SEED" : '.*'` -gt 5 ] && {
echo "$PROG: Seed value ($SEED) out of range (0 to 32767)" >&2
exit 1
}
(( $Upper <= $Lower )) && {
echo "$PROG: upper ($Upper) <= lower value ($Lower)" >&2
exit 1
}
# Seed the random-number generator:
RANDOM=$SEED
# Compute value, scaled within range:
let rand="$RANDOM % ($Upper - $Lower + 1) + $Lower"
# Report result:
echo $rand

176
examples/scripts.v2/cdhist.bash Normal file
View file

@ -0,0 +1,176 @@
#! /bin/bash
#
# cdhist - cd replacement with a directory stack like pushd/popd
#
# usage: cd [-l] [-n] [-] [dir]
#
# options:
# -l print the cd directory stack, one entry per line
# - equivalent to $OLDPWD
# -n cd to nth directory in cd directory stack
# -s cd to first directory in stack matching (substring) `s'
#
# arguments:
# dir cd to dir and push dir onto the cd directory stack
#
# If the new directory is a directory in the stack and the options selected
# it (-n, -s), the new working directory is printed
#
# If the variable CDHISTFILE is set, the cd directory stack is loaded from
# and written to $CDHISTFILE every time `cd' is executed.
#
# Note: I got this off the net somewhere; I don't know the original author
#
# Chet Ramey
# chet@po.cwru.edu
_cd_print()
{
echo -e "$@"
}
cd()
{
typeset -i cdlen i
typeset t
if [ $# -eq 0 ]
then
set -- $HOME
fi
if [ "$CDHISTFILE" -a -r "$CDHISTFILE" ] # if directory history exists
then
typeset CDHIST
i=-1
while read -r t # read directory history file
do
CDHIST[i=i+1]=$t
done <$CDHISTFILE
fi
if [ "${CDHIST[0]}" != "$PWD" -a "$PWD" != "" ]
then
_cdins # insert $PWD into cd history
fi
cdlen=${#CDHIST[*]} # number of elements in history
case "$@" in
-) # cd to new dir
if [ "$OLDPWD" = "" ] && ((cdlen>1))
then
'_cdprint' ${CDHIST[1]}
builtin cd ${CDHIST[1]}
pwd
else
builtin cd "$@"
# pwd
fi
;;
-l) # _cdprint directory list
((i=cdlen))
while (((i=i-1)>=0))
do
num=$i
'_cdprint' "$num ${CDHIST[i]}"
done
return
;;
-[0-9]|-[0-9][0-9]) # cd to dir in list
if (((i=${1#-})<cdlen))
then
'_cdprint' ${CDHIST[i]}
builtin cd ${CDHIST[i]}
pwd
else
builtin cd $@
# pwd
fi
;;
-*) # cd to matched dir in list
t=${1#-}
i=1
while ((i<cdlen))
do
case ${CDHIST[i]} in
*$t*)
'_cdprint' ${CDHIST[i]}
builtin cd ${CDHIST[i]}
pwd
break
;;
esac
((i=i+1))
done
if ((i>=cdlen))
then
builtin cd $@
# pwd
fi
;;
*) # cd to new dir
builtin cd $@
# pwd
;;
esac
_cdins # insert $PWD into cd history
if [ "$CDHISTFILE" ]
then
cdlen=${#CDHIST[*]} # number of elements in history
i=0
while ((i<cdlen))
do
echo ${CDHIST[i]} # update directory history
((i=i+1))
done >$CDHISTFILE
fi
}
_cdins() # insert $PWD into cd history
{ # meant to be called only by cd
typeset -i i
i=0
while (( i < ${#CDHIST[*]} )) # see if dir is already in list
do
if [ "${CDHIST[$i]}" = "$PWD" ]
then
break
fi
((i=i+1))
done
if (( i>22 )) # limit max size of list
then
i=22
fi
while (((i=i-1)>=0)) # bump old dirs in list
do
CDHIST[i+1]=${CDHIST[i]}
done
CDHIST[0]=$PWD # insert new directory in list
}
# examples
shopt -s expand_aliases
# go to known place before doing anything
cd /
echo CDHIST: "${CDHIST[@]}"
for dir in /tmp /bin - -2 -l
do
cd $dir
echo CDHIST: "${CDHIST[@]}"
echo PWD: $PWD
done
exit 0

43
examples/scripts.v2/corename Normal file
View file

@ -0,0 +1,43 @@
#! /bin/bash
#
# original from:
# @(#) corename.ksh 1.0 93/04/01
# 92/11/11 john h. dubois iii (john@armory.com)
# 92/02/16 Added help option.
# 92/02/22 Added cd to origdir to fix prob w/multiple relative paths.
# 93/04/01 Added check for whether file exists.
#
# conversion to bash v2 syntax done by Chet Ramey
# inspired by belal's equivalent utility
if [ "$1" = -h ]; then
echo \
"$0: print the names of executables that dumped core.
Usage: $0 [corename ...]
If no corename is given, \"core\" is assumed."
exit 0
fi
[ $# = 0 ] && set core
origdir=$PWD
for i; do
cd $origdir
file=${i##*/}
dir=${i%$file}
[ -z "$dir" ] && dir=$origdir/
if [ ! -f $dir$file ]; then
echo "$dir$file: No such file."
continue
fi
if [ ! -r $dir$file ]; then
echo "$dir$file: Cannot open."
continue
fi
cd $dir
# the adb output syntax is highly variable. this works on SunOS 4.x
set -- $(adb $file < /dev/null 2>&1 | sed 1q)
name=${7#??}
echo "$i: ${name%??}"
done

281
examples/scripts.v2/fman Normal file
View file

@ -0,0 +1,281 @@
#! /bin/bash
#
# original from:
# fman: new man program
# @(#) fman.ksh 1.5 94/04/16
# 91/07/03 john h. dubois iii (john@armory.com)
# 91/07/11 made it unpack man pages if neccessary
# 91/07/16 fixed test for whether man file pattern was expanded
# 92/01/21 made it read /etc/default/man to get section order,
# and only display the first section found.
# 92/02/06 changed name to fman
# 92/02/07 fixed bug in notfound
# 92/02/13 incorporated changes from DOS version
# 92/03/11 changed to use MANPATH from environment if set,
# and search all directories given in MANPATH
# 92/03/15 exec pager or man w/o forking
# 92/05/31 try using index if one exists
# 92/10/01 Added "See also <other sections>"
# 92/10/18 If PAGER is less, search for name of man page to make it easier
# to find information in man pages for multiple items
# 92/11/11 Make it work for compressed files not listed in index;
# deal with man pages listed in index that don't exist.
# 93/03/30 Fixed bug in MANPATH processing
# 93/06/17 Include paths in "See also:" message if they would be needed
# to get to a man page. Allow MANPATH spec on command line.
# 93/07/09 Added -h and -e options.
# 94/04/16 Added x option.
#
# conversion to bash v2 syntax done by Chet Ramey
istrue()
{
test 0 -ne "$1"
}
isfalse()
{
test 0 -eq "$1"
}
# Finds all sections that man page $1 is in and puts them in the the
# global array Sections[].
# The filename of each page is put in FileNames[] with the same index.
# Global vars used:
# patharr[] MANPATH directories.
FindSectionsInIndex ()
{
typeset index indexes section mpath page=$1
typeset -i i=0 NIndex=0
for mpath in "${patharr[@]}"; do
if [ -r $mpath/index ]; then
indexes="$indexes $mpath/index"
let NIndex+=1
fi
done
[ -z "$indexes" ] && return
# Make egrep give filename
[ NIndex -lt 2 ] && indexes="$indexes /dev/null"
# set positional parameters to
# indexfile:searchname pagename section ...
# e.g.
# /usr/man/index:FP_OFF Routines DOS
set -- `egrep "^$page[ ]" $indexes`
while [ $# -gt 2 ]; do
FileNames[i]=${1%%index*}cat$3/$2.$3
Sections[i]=$3
shift 3
let i+=1
done
}
# Finds all sections that man page $1 is in by searching each man directory
# in the order given in patharr[],
# and puts them in the the global array Sections[].
# The filename of each page is put in FileNames[] with the same index.
# Global vars used:
# patharr[] MANPATH directories.
FindSectionsInDirs ()
{
local page=$1 mpath AllPaths Path
typeset -i i
for mpath in "${patharr[@]}"; do
AllPaths="$AllPaths $mpath/cat[0-9]*/$page.* $mpath/man[0-9]*/$page.*"
done
i=0
for Path in $AllPaths; do
istrue $debug && echo Path = $Path
case "$Path" in
*\*) ;;
*)
# Remove compressed-file suffix to make FileNames be the same
# as it is when built by FindSectionsInIndex()
FileNames[i]=${Path%.[zZ]}
Path=${Path%/*}
Sections[i]=${Path##*/*.}
let i+=1 ;;
esac
done
}
# FindSection: display man page.
# Uses ordarr[] (built from $ORDER) to display the version of the man
# page that occurs first in $ORDER.
# Sections[] gives the sections that a man page was found in.
# If the global variable "exist" is set to 1, nothing is displayed;
# the function instead returns zero if a page is found, nonzero if not.
# The filename of each page is in FileNames[] with the same index.
# Global vars used:
# Sections[], FileNames[], ordarr[]
FindSection ()
{
typeset -i NumPages i foundsec
local section OtherSec filename NPAGER=$PAGER POpt page=$1 Pat
local PageFile
NumPages=${#Sections[*]} # Number of versions of man page found.
isfalse $NumPages && return 1
case "$PAGER" in
*less) Popt="-p$page" ;;
esac
# For each section in ORDER, determine if any man page was found in
# that section
for section in "${ordarr[@]}"; do
i=0
foundsec=0
while [ $i -lt $NumPages ]; do
if [ "${Sections[i]}" = $section ]; then
# Found a man page from this section of ORDER
filename=${FileNames[i]}
if [ -z "$PageFile" ]; then
PageFile=$filename
else
if istrue $foundsec; then
OtherSec="$OtherSec$page(${filename%/*/*} $section) "
else
OtherSec="$OtherSec$page($section) "
fi
fi
foundsec=1
istrue $exist && return
fi
let i+=1
done
done
# No pages with the specified section found.
[ -z "$PageFile" ] && return 1
# Return if all we want to know is whether the man page exists.
[ "$exist" = 1 ] && return 0
if [ -z "$OtherSec" ]; then
NPAGER="exec $PAGER"
fi
if [ -r $PageFile ]; then
$NPAGER $POpt $PageFile
elif [ -r $PageFile.z ]; then
pcat $PageFile.z | $NPAGER $POpt
elif [ -r $PageFile.Z ]; then
zcat $PageFile.Z | $NPAGER $POpt
elif [ -f $PageFile.gz ]; then
gzip -dc $PageFile.gz | $NPAGER $POpt
else
echo "$PageFile: cannot open." 1>&2
OtherSec=
unset Sections[i]
let i+=1
continue
fi
echo "See also $OtherSec"
exit 0
}
phelp()
{
echo "$name: print man pages.
$name locates and prints the specified manual pages from the online UNIX
documentation.
$Usage
Options:
-e: Determine whether the specified man page exists. Nothing is printed;
$0 exits with a zero status if the page exists and a nonzero status if
it does not.
-h: Print this help."
}
# main program
typeset -i exist=0 debug=0
name=${0##*/}
Usage="Usage: $name [-eh] [[manpath] section] command-name"
while getopts :hex opt; do
case $opt in
h) phelp; exit 0;;
e) exist=1 ;;
x) debug=1 ;;
+?) echo "$name: options should not be preceded by a '+'." 1>&2; exit 2;;
?)
echo "$name: $OPTARG: bad option. Use -h for help." 1>&2 ; exit 2 ;;
esac
done
# remove args that were options
shift $((OPTIND-1))
if [ $# -lt 1 ]; then
echo -e "$Usage\nUse -h for help." 1>&2
exit
fi
P=$PAGER
O=1:n:l:6:8:2:3:4:5:7:p:o
T=$TERM
M=${MANPATH:-/usr/local/man:/usr/man}
[ -f /etc/default/man ] && . /etc/default/man
[ -n "$P" ] && PAGER=$P
[ -n "$O" ] && ORDER=$O
[ -n "$T" ] && TERM=$T
[ -n "$M" ] && MANPATH=$M
case $# in
0) echo "No man page specified." ; exit 1;;
1) page=$1;;
2) ORDER=$(echo $1 | tr a-z A-Z) ; page=$2;;
3) MANPATH=$1
[ -n "$2" ] && ORDER=$(echo $2 | tr a-z A-Z)
page=$3;;
*) echo "Too many arguments."; exit 1;;
esac
aargs=("$@")
[ ! -t 0 ] && PAGER=cat
OIFS=$IFS
IFS=:
patharr=($MANPATH)
i=0
for d in $MANPATH; do
for sec in $ORDER; do
ordarr[i]=$d/cat${sec}
let i+=1
ordarr[i]=$d/man${sec}
let i+=1
done
done
IFS=$OIFS
istrue $debug && echo patharr = "${patharr[@]}"
# if less or more is being used, remove multiple blank lines
export LESS="-s $LESS"
export MORE="-s $MORE"
# Try using index
FindSectionsInIndex "$page"
# Exit 0 if a page was found and we're just testing for existence.
FindSection "$page" && exit 0
# Try searching directories
unset Sections[*]
FindSectionsInDirs "$page"
FindSection "$page" && exit 0
istrue $exist && exit 1
# Try using man
# If using more or less, make man run faster by letting more or less compress
# multiple blank lines instead of rmb
#case "$PAGER" in
#*more|*less) manopt=-b;;
#esac
#cmd=(man $manopt -p$PAGER "${aargs[@]}")
export PAGER
cmd=(man $manopt "${aargs[@]}")
istrue $debug && echo "$name: running ${cmd[*]}" 1>&2
exec "${cmd[@]}"

288
examples/scripts.v2/frcp Executable file
View file

@ -0,0 +1,288 @@
#! /bin/bash
#
# original from:
#
# @(#) frcp.ksh 2.2 93/11/14
# 92/06/29 john h. dubois iii (john@armory.com)
# 92/10/14 Cleaned up, improved, added -d and -r options
# 92/11/11 Made work with a dest of '.'
# 93/07/09 Added -l and -n options, & login as anonymous if no .netrc entry
# 93/11/14 Use either passwd or password in .netrc, since ftp does.
#
# conversion to bash v2 syntax by Chet Ramey
#
# frcp: ftp front end with rcp-like syntax.
# Note: requires any machine names given to be listed with
# user and password in .netrc. If not, anonymous FTP is
# done.
#
# full path to ftp binary
if [ -x /usr/bin/ftp ]; then
FTP=/usr/bin/ftp;
elif [ -x /usr/ucb/ftp ]; then
FTP=/usr/ucb/ftp
else
FTP=ftp
fi
istrue()
{
test 0 -ne "$1"
}
isfalse()
{
test 0 -eq "$1"
}
# For each filename given, put the filename in filename[n]
# and the machine it is on in machine[n].
function SplitNames {
typeset file
typeset -i i=1
unset filename[*] machine[*]
for file; do
case "$file" in
*:*) machine[i]=${file%%:*} ;;
*) machine[i]=$LocalMach ;;
esac
filename[i]=${file#*:}
let i+=1
done
}
function verboseprint {
echo "$@"
echo "$@" 1>&2
}
function MakeDir {
OFS=$IFS
local IFS=/ dir component
case "$1" in
/*) ;;
*) dir=.
esac
set -- $1
IFS=$OFS
for component; do
dir=$dir/$component
if [ ! -d "$dir" ]; then
if mkdir "$dir"; then :; else
echo "Could not make directory $dir." >&2
return 1
fi
fi
done
return 0
}
lastisdot ()
{
case "$1" in
*/.|*/..) return 0;;
*) return 1;;
esac
}
# CopyFiles: issue ftp(TC) commands to copy files.
# Usage: CopyFiles [sourcemachine:]sourcepath ... [destmachine:]destpath
# Global vars:
# Uses LocalMach (should be name of local machine)
# Sets global arrs machine[]/filename[]
function CopyFiles {
unset machine[*] filename[*]
SplitNames "$@" # split names into filename[1..n] and machine[1..n]
local DestMach=${machine[$#]} # Machine to copy files to
local DestPath=${filename[$#]} # Destination file/dir
unset machine[$#] filename[$#]
[ -z "$DestPath" ] && DestPath=. # dest was given as machine:
# Try to determine if destination should be a directory
# so that it can be forced to be a directory.
case "$DestPath" in
*/) ;; # don't add / if trailing / already present
*) if [ $# -gt 2 ] || # if more than two args given, last must be a dir
# If dest in on local machine, check whether it is a directory
[ $DestMach = $LocalMach -a -d $DestPath ] ||
# If dest ends with . or .., it is a directory
lastisdot "$DestPath"
then
DestPath=$DestPath/
fi ;;
esac
# If one of the above tests made us think dest is a directory,
# but it isn't, complain
case "$DestPath" in
*/) if [ "$DestMach" = "$LocalMach" ] && [ ! -d "$DestPath" ]; then
echo "Destination is not a directory." 1>&2
exit 1
fi ;;
esac
DoCopy "$DestMach" "$DestPath"
}
# Usage: OpenMachine machine-name
# Emits login sequence or doesn't, depending on .netrc file and global
# variables anon and noanon
OpenMachine ()
{
local machine=$1 netrc=$HOME/.netrc user= password=
if isfalse $anon && [ -r $netrc ]; then
set -- $(gawk '
/machine (.* )?'"$machine"'($| )/,/^ *$/ {
Fields[$1] = $2
if ("passwd" in Fields)
Fields["password"] = Fields["passwd"]
if ("login" in Fields && "password" in Fields) {
print Fields["login"] " " Fields["password"]
exit
}
}
' $netrc )
user=$1
password=$2
fi
if [ -z "$password" ]; then
if istrue $noanon; then
echo "No .netrc entry for machine $machine" 1>&2
exit 1
fi
user=anonymous
password=$USER@$LocalMach
fi
verboseprint open $machine
echo user $user "*******" 1>&2
echo user $user $password
}
# Usage: DoCopy destination-machine destination-path
# Copies the files in global arrs machine[]/filename[] to the given dest
# Global vars:
# Uses machine[], filename[], LocalMach, check
DoCopy ()
{
local DestMach=$1
local DestPath=$2
local OpenMach # Machine that connection is currently open to
local OWD=$PWD SourceMach SourceFile
local FileName
typeset -i i=1
while [ $i -le ${#machine[*]} ]; do
istrue $check && verboseprint "runique"
SourceMach=${machine[i]}
SourceFile=${filename[i]}
DestFile=$DestPath
# if DestPath is a dir,
# add source filename to it without source path
case "$DestFile" in
*/) DestFile=$DestFile${SourceFile##*/} ;;
esac
if [ $SourceMach = $LocalMach ]; then
if [ $DestMach != "$OpenMach" ]; then
OpenMachine $DestMach
OpenMach=$DestMach
fi
verboseprint put $SourceFile $DestFile
elif [ $DestMach = $LocalMach ]; then
if istrue $check && [ -f "$DestFile" ]; then
echo "$DestFile already exists." 1>&2
continue
fi
# If destination is on local machine,
# the dest will be a full dir/filename
if istrue $createdirs; then
MakeDir "${DestFile%/*}" || continue
fi
if [ $SourceMach != "$OpenMach" ]; then
OpenMachine $SourceMach
OpenMach=$SourceMach
fi
# If source filename has wildcards ([, ], *, ?) do an mget
case "$SourceFile" in
\[*\]|*\**|*\?*)
verboseprint lcd "$DestFile"
verboseprint mget "$SourceFile"
verboseprint lcd $OWD ;;
*) verboseprint get "$SourceFile" "$DestFile" ;;
esac
else
echo "Neither source machine \"$SourceMach\" "\
"nor destination machine \"$DestMach\" is local." 1>&2
fi
let i+=1
done
}
# Start of main program
name=${0##*/}
if [ "$1" = -h ]; then
echo \
"$name: do ftp transfers using rcp-style parameters.
Usage: $name <source> <destpath> or $name <source> [<source> ...] <destdir>
At least one of <source> and <destpath> must be the local system.
A remote filename is given as machinename:filename
If remote filenames contain wildcards, they will be globbed on the remote
machine. Make sure they are quoted when $name is invoked.
If the invoking user's .netrc file (see ftp(TC)) contains an entry for the
remote system with a login and password supplied, $name will log in using
the given login and password. If not, $name will login in as user
anonymous and with the user@localsystem as the password.
Options:
-c: check: do not overwrite files.
-d: create directories as needed.
-f: force: overwrite files (default).
-h: print this help.
-l: fail if there is no entry with login and password for the remote system,
instead of logging in as anonymous.
-n: log in as anonymous even if there is an entry for the remote system in
the user's .netrc file.
-r: read source/dest filename pairs from the standard input,
one pair per line, and copy files accordingly."
exit 0
fi
typeset -i check=0 createdirs=0 readinput=0 anon=0 noanon=0
while getopts :cdflnr Option
do
case "$Option" in
c) check=1;;
d) createdirs=1;;
f) check=0;;
l) noanon=1;;
n) anon=1;;
r) readinput=1;;
\?) echo "$OPTARG: invalid option."; exit 1;;
esac
done
shift $((OPTIND-1))
LocalMach=`hostname`
if istrue $readinput; then
while read line; do
CopyFiles $line
done | $FTP -nv
else
if [ $# -lt 2 ]; then
echo "$name: Not enough arguments. Use -h for help." 1>&2
exit
fi
CopyFiles "$@" | $FTP -nv
fi

44
examples/scripts.v2/lowercase Normal file
View file

@ -0,0 +1,44 @@
#! /bin/bash
#
# original from
# @(#) lowercase.ksh 1.0 92/10/08
# 92/10/08 john h. dubois iii (john@armory.com)
#
# conversion to bash v2 syntax done by Chet Ramey
Usage="Usage: $name file ..."
phelp()
{
echo "$name: change filenames to lower case.
$Usage
Each file is moved to a name with the same directory component, if any,
and with a filename component that is the same as the original but with
any upper case letters changed to lower case."
}
name=${0##*/}
while getopts "h" opt; do
case "$opt" in
h) phelp; exit 0;;
*) echo "$Usage" 1>&2; exit 2;;
esac
done
shift $((OPTIND - 1))
for file; do
filename=${file##*/}
case "$file" in
*/*) dirname=${file%/*} ;;
*) dirname=. ;;
esac
nf=$(echo $filename | tr A-Z a-z)
newname="${dirname}/${nf}"
if [ "$nf" != "$filename" ]; then
mv "$file" "$newname"
echo "$0: $file -> $newname"
else
echo "$0: $file not changed."
fi
done

187
examples/scripts.v2/ncp Normal file
View file

@ -0,0 +1,187 @@
#! /bin/bash
#
# original from:
# @(#) ncp.ksh,nmv.ksh 1.1 94/07/23
# 92/01/18 john h. dubois iii (john@armory.com)
# 92/01/31 added check for no args left after shifts
# 92/02/17 added help
# 92/02/25 remove path component from filename before tacking it onto dest.
# 92/03/15 exec mv or cp
# 93/07/13 Added -i
# 93/09/29 Made abort if file exists optional.
# 93/11/19 Exit before invoking mv if no files to move
# 94/01/03 Added o option
# 94/04/13 Added x option.
# Fixed appending of source filename, broken by earlier change.
# 94/07/23 Append only the filename part of the source path.
#
# conversion to bash v2 syntax done by Chet Ramey
false()
{
return 1
}
true()
{
return 0
}
phelp()
{
echo "$name: do a $cmd with extra checking and options.
$Usage
$name is used as a front end for $cmd to get the [icfo] options, and so
that a trailing / will force the last component of the path to be
interpreted as a directory, so that $name foo bar/ will fail if bar is
not an existing directory, instead of changing the name of foo to bar.
Effectively, $name foo bar/ is short for $name foo bar/foo
Options:
-h prints this help.
-c checks first for the existence of each file, and fails if it exists.
-i is like -c except that if the file exists and stdin and stdout are a
tty, a query is printed and a reply is read; a file is overwritten only
if the reply begins with 'y'.
-f unsets -c and -i (in case $cmd is aliased to $name).
-o (overwrite only) checks that the named file(s) exist and fails for any
that do not. It is the complement of the -c option.
Whichever of [cifo] comes later on the command line determines the behaviour.
Any of these options must come before any standard $cmd options."
}
# interactive: Attempt to overwrite file should result in interactive
# query rather than automatic failure.
# noover: Do not overwrite files (if interactive is true, query, else fail)
# overwrite: Only overwriting is allowed, not creation of new files.
# debug: Print debugging info.
typeset interactive=false noover=false overwrite=false debug=false
name=${0##*/}
case "$name" in
ncp|nmv) cmd=/bin/${name#?} ;;
*) echo "$name: Must be invoked as ncp or nmv." 1>&2 ; exit 2;;
esac
Usage="Usage: $name [-cfhio] $cmd-cmd-line"
while getopts :cfhiox opt; do
case $opt in
h) phelp; exit 0;;
x) debug=true ;;
c) noover=true ;;
i) noover=true ; interactive=true ;;
f) noover=false ; interactive=false ;;
o) overwrite=true ; noover=false ; interactive=false;;
+?) echo "$name: options should not be preceded by a '+'." 1>&2; exit 2;;
?) echo "$name: $OPTARG: bad option. Use -h for help." 1>&2 ; exit 2;;
esac
done
# remove args that were options
shift $((OPTIND - 1))
if [ $# -lt 2 ]; then
echo -e "$Usage\nUse -h for help."
exit
fi
Check()
{
if [ ! -f "$1" ] && $overwrite; then
echo "$name: $1: File does not exist." 1>&2
return 1
elif [ -f "$1" ] && $noover; then
if [ $interactive = false ] || [ ! -t 0 ] || [ ! -t 1 ]; then
echo "$name: $1: File exists." 1>&2
return 1
else
while :; do
echo -n \
"$name: $1: File exists. Overwrite? (y)es/(n)o/(a)bort/(Y)es for all: " 1>&2
read reply
case "$reply" in
y*)
echo "$name: Overwriting $1."
return 0
;;
Y*)
echo "$name: Overwriting $1."
interactive=false
noover=false
return 0
;;
[nN]*)
echo "$name: Skipping $2."
return 1
;;
[aA]*)
echo "$name: Aborting."
exit 1
;;
*)
echo "$name: Invalid response." 1>&2
;;
esac
done
fi
else
return 0
fi
}
# i is the index of the filename being examined
# lastarg is the index of the last filename before the dest directory name
typeset -i i=0 lastarg=$(($#-1))
# Sets argv[0..$#-1]
argv=("$@")
$debug && echo argv = "${argv[@]}" 1>&2
dest=${argv[lastarg]}
if $debug; then
echo \
"interactive=$interactive noover=$noover overwrite=$overwrite debug=$debug
lastarg=$lastarg dest=$dest name=$name cmd=$cmd
files=$*" 1>&2
fi
if $noover || $overwrite; then
$debug && echo "checking for existance of directories..." 1>&2
# If the destination is not intended to be a directory...
if [ $# -eq 2 ] && [ ! -d "$dest" ]; then
Check "$dest" "$1" || exit 0 # No files to copy
else
while [ $i -lt $lastarg ]; do
Check "$dest/${argv[i]##*/}" "${argv[i]}" || unset argv[i]
let i+=1
done
fi
fi
[ ${#argv[@]} -lt 2 ] && exit 0
# If only 2 args are given, mv/cp will not insist that the destination
# be a directory, which we want if the destination ends in "/" or if
# the original number of args was >2.
# $# is still the original number of args.
# Tack the file name onto the destination to force this behaviour.
lastisslash()
{
case "$1" in
*/) return 0;;
*) return 1;;
esac
}
if [ ${#argv[@]} = 2 ] && { lastisslash "$2" || [ $# -gt 2 ]; }; then
$debug && echo "Appending filename." 1>&2
# Don't know which element of argv[] holds the source filename,
# since may have started with more than 1 source file & had some unset.
# So, compact args to make it easy to find the set one.
argv=("${argv[@]}")
argv[1]="${argv[1]}/${argv[0]##*/}"
fi
$debug && echo "Executing command: $cmd ${argv[@]}" 1>&2
exec $cmd "${argv[@]}"

64
examples/scripts.v2/newext Normal file
View file

@ -0,0 +1,64 @@
#! /bin/bash
#
# original from:
# newext: change filename extension
# @(#) newext.sh 1.1 93/04/13
# 90/06/06 john h. dubois iii (john@armory.com)
# 90/11/14 changed ksh-specific code to hybrid: if running under Bourne,
# uses expr instead of ksh builtin ops. Removed SYSV specific code.
# 91/08/06 added -t option
# 92/11/06 made earlier code actually work!
# 93/04/13 If no filenames given, act on files in current dir
#
# conversion to bash v2 syntax by Chet Ramey
usage="Usage: newext [-th] <oldext> <newext> [filename ...]"
phelp()
{
echo "$usage
Rename all given files that end in oldext with newext replacing oldext.
If no filenames are given, all files in the current directory that end
in oldext are acted on (no filename is equivalent to '*').
Options:
-h: Print this help.
-t: Test: No action is taken except to print the mv commands that would
be executed if -t was not given."
}
while getopts "th" opt; do
case "$opt" in
t) echo=echo;;
h) phelp; exit 0;;
*) echo "$usage" 1>&2; exit 2;;
esac
done
shift $((OPTIND - 1))
oldext=$1
newext=$2
case $# in
[01]) echo -e "$usage\nUse -h for help." 1>&2; exit 2;;
2) shift ; shift; set -- *;;
*) shift ; shift;;
esac
found=
for file
do
case "$file" in
*$oldext)
newname="${file%$oldext}$newext"
$echo mv "$file" "$newname"
found=true;;
esac
done
if [ -z "$found" ]; then
echo "No files ending in \"$oldext\"."
exit 1
fi
exit 0

187
examples/scripts.v2/nmv Normal file
View file

@ -0,0 +1,187 @@
#! /bin/bash
#
# original from:
# @(#) ncp.ksh,nmv.ksh 1.1 94/07/23
# 92/01/18 john h. dubois iii (john@armory.com)
# 92/01/31 added check for no args left after shifts
# 92/02/17 added help
# 92/02/25 remove path component from filename before tacking it onto dest.
# 92/03/15 exec mv or cp
# 93/07/13 Added -i
# 93/09/29 Made abort if file exists optional.
# 93/11/19 Exit before invoking mv if no files to move
# 94/01/03 Added o option
# 94/04/13 Added x option.
# Fixed appending of source filename, broken by earlier change.
# 94/07/23 Append only the filename part of the source path.
#
# conversion to bash v2 syntax done by Chet Ramey
false()
{
return 1
}
true()
{
return 0
}
phelp()
{
echo "$name: do a $cmd with extra checking and options.
$Usage
$name is used as a front end for $cmd to get the [icfo] options, and so
that a trailing / will force the last component of the path to be
interpreted as a directory, so that $name foo bar/ will fail if bar is
not an existing directory, instead of changing the name of foo to bar.
Effectively, $name foo bar/ is short for $name foo bar/foo
Options:
-h prints this help.
-c checks first for the existence of each file, and fails if it exists.
-i is like -c except that if the file exists and stdin and stdout are a
tty, a query is printed and a reply is read; a file is overwritten only
if the reply begins with 'y'.
-f unsets -c and -i (in case $cmd is aliased to $name).
-o (overwrite only) checks that the named file(s) exist and fails for any
that do not. It is the complement of the -c option.
Whichever of [cifo] comes later on the command line determines the behaviour.
Any of these options must come before any standard $cmd options."
}
# interactive: Attempt to overwrite file should result in interactive
# query rather than automatic failure.
# noover: Do not overwrite files (if interactive is true, query, else fail)
# overwrite: Only overwriting is allowed, not creation of new files.
# debug: Print debugging info.
typeset interactive=false noover=false overwrite=false debug=false
name=${0##*/}
case "$name" in
ncp|nmv) cmd=/bin/${name#?} ;;
*) echo "$name: Must be invoked as ncp or nmv." 1>&2 ; exit 2;;
esac
Usage="Usage: $name [-cfhio] $cmd-cmd-line"
while getopts :cfhiox opt; do
case $opt in
h) phelp; exit 0;;
x) debug=true ;;
c) noover=true ;;
i) noover=true ; interactive=true ;;
f) noover=false ; interactive=false ;;
o) overwrite=true ; noover=false ; interactive=false;;
+?) echo "$name: options should not be preceded by a '+'." 1>&2; exit 2;;
?) echo "$name: $OPTARG: bad option. Use -h for help." 1>&2 ; exit 2;;
esac
done
# remove args that were options
shift $((OPTIND - 1))
if [ $# -lt 2 ]; then
echo -e "$Usage\nUse -h for help."
exit
fi
Check()
{
if [ ! -f "$1" ] && $overwrite; then
echo "$name: $1: File does not exist." 1>&2
return 1
elif [ -f "$1" ] && $noover; then
if [ $interactive = false ] || [ ! -t 0 ] || [ ! -t 1 ]; then
echo "$name: $1: File exists." 1>&2
return 1
else
while :; do
echo -n \
"$name: $1: File exists. Overwrite? (y)es/(n)o/(a)bort/(Y)es for all: " 1>&2
read reply
case "$reply" in
y*)
echo "$name: Overwriting $1."
return 0
;;
Y*)
echo "$name: Overwriting $1."
interactive=false
noover=false
return 0
;;
[nN]*)
echo "$name: Skipping $2."
return 1
;;
[aA]*)
echo "$name: Aborting."
exit 1
;;
*)
echo "$name: Invalid response." 1>&2
;;
esac
done
fi
else
return 0
fi
}
# i is the index of the filename being examined
# lastarg is the index of the last filename before the dest directory name
typeset -i i=0 lastarg=$(($#-1))
# Sets argv[0..$#-1]
argv=("$@")
$debug && echo argv = "${argv[@]}" 1>&2
dest=${argv[lastarg]}
if $debug; then
echo \
"interactive=$interactive noover=$noover overwrite=$overwrite debug=$debug
lastarg=$lastarg dest=$dest name=$name cmd=$cmd
files=$*" 1>&2
fi
if $noover || $overwrite; then
$debug && echo "checking for existance of directories..." 1>&2
# If the destination is not intended to be a directory...
if [ $# -eq 2 ] && [ ! -d "$dest" ]; then
Check "$dest" "$1" || exit 0 # No files to copy
else
while [ $i -lt $lastarg ]; do
Check "$dest/${argv[i]##*/}" "${argv[i]}" || unset argv[i]
let i+=1
done
fi
fi
[ ${#argv[@]} -lt 2 ] && exit 0
# If only 2 args are given, mv/cp will not insist that the destination
# be a directory, which we want if the destination ends in "/" or if
# the original number of args was >2.
# $# is still the original number of args.
# Tack the file name onto the destination to force this behaviour.
lastisslash()
{
case "$1" in
*/) return 0;;
*) return 1;;
esac
}
if [ ${#argv[@]} = 2 ] && { lastisslash "$2" || [ $# -gt 2 ]; }; then
$debug && echo "Appending filename." 1>&2
# Don't know which element of argv[] holds the source filename,
# since may have started with more than 1 source file & had some unset.
# So, compact args to make it easy to find the set one.
argv=("${argv[@]}")
argv[1]="${argv[1]}/${argv[0]##*/}"
fi
$debug && echo "Executing command: $cmd ${argv[@]}" 1>&2
exec $cmd "${argv[@]}"

187
examples/scripts.v2/pages Normal file
View file

@ -0,0 +1,187 @@
#! /bin/bash
#
# original from:
# @(#) pages.sh 1.0 92/09/26
# 92/09/05 John H. DuBois III (jhdiii@armory.com)
# 92/09/26 Added help
#
# conversion to bash v2 syntax by Chet Ramey
Usage="$0 [-h] [-n lines/page] page-ranges [file ...]"
usage()
{
echo "$Usage" 1>&2
}
phelp()
{
echo "$0: print selected pages.
Usage: $Usage
If no file names are given, the standard input is read.
The input is grouped into pages and a selected subset of them is printed.
Formfeeds are acted on correctly.
If the output device does automatic line wrap, lines that longer than
the width of the output device will result in incorrect output.
The first non-option argument is a list of pages to print.
Pages are given as a list of ranges separated by commas.
A range is either one number, two numbers separted by a dash,
or one number followed by a dash. A range consisting of one
number followed by a dash extends to the end of the document.
Options:
-n sets the number of lines per page to n. The default is 66."
}
while getopts "n:h" opt; do
case "$opt" in
n) LinesPerPage=$OPTARG;;
h) phelp; exit 0;;
*) usage; exit 2;;
esac
done
shift $(($OPTIND - 1))
if [ $# -eq 0 ]; then
echo $0: no page ranges given. 1>&2
usage
exit 1
fi
PageList=$1
shift
gawk "
BEGIN {
PageList = \"$PageList\"; LinesPerPage = \"$LinesPerPage\""'
if (LinesPerPage == "")
LinesPerPage = 66
else
if (LinesPerPage !~ "[1-9][0-9]*")
ErrExit("Bad value for lines per page: " LinesPerPage)
LinesPerPage += 0
NumRanges = split(PageList,Ranges,",")
for (i = 1; i <= NumRanges; i++) {
if ((StartRange = EndRange = Ranges[i]) !~ "^[0-9]+(-([0-9]+)?)?$")
ErrExit("Bad range \"" StartRange "\"")
sub("-.*","",StartRange)
sub(".*-","",EndRange)
if (EndRange == "")
EndRange = 2 ^ 30
# Force StartRange and EndRange to be numeric values
if ((StartRange += 0) == 0 || (EndRange += 0) == 0)
ErrExit("Invalid page number \"0\" in range " Ranges[i])
if (StartRange > EndRange)
ErrExit("Start page comes after end page in range " Ranges[i])
TmpRangeStarts[i] = StartRange
TmpRangeEnds[i] = EndRange
}
# Sort ranges
qsort(TmpRangeStarts,k)
RangeEnds[0] = 0
for (i = 1; i <= NumRanges; i++) {
RangeEnds[i] = TmpRangeEnds[k[i]]
if ((RangeStarts[i] = TmpRangeStarts[k[i]]) <= RangeEnds[i - 1])
ErrExit("Overlapping ranges: " Ranges[k[i]] "," Ranges[k[i - 1]])
}
RangeNum = LineNum = PageNum = 1
InRange = In(PageNum,RangeStarts[RangeNum],RangeEnds[RangeNum])
FS = "\014"
}
{
if (LineNum > LinesPerPage)
NewPage()
if (InRange)
printf "%s",$1
# Deal with formfeeds
for (i = 2; i <= NF; i++) {
if (InRange)
printf "\014"
NewPage()
if (InRange)
printf "%s",$i
}
if (InRange)
print ""
LineNum++
}
function NewPage() {
PageNum++
LineNum = 1
# At the start of each page, check whether we are in a print range
WereInRange = InRange
InRange = In(PageNum,RangeStarts[RangeNum],RangeEnds[RangeNum])
# If last page was in range and we no longer are, move to next range
if (WereInRange && !InRange && ++RangeNum > NumRanges)
exit
}
function In(a,Min,Max) {
return (Min <= a && a <= Max)
}
function ErrExit(S) {
print S > "/dev/stderr"
Err = 1
exit 1
}
# Arr is an array of values with arbitrary indices.
# Array k is returned with numeric indices 1..n.
# The values in k are the indices of array arr,
# ordered so that if array arr is stepped through
# in the order arr[k[1]] .. arr[k[n]], it will be stepped
# through in order of the values of its elements.
# The return value is the number of elements in the array (n).
function qsort(arr,k, ArrInd,end) {
end = 0
for (ArrInd in arr)
k[++end] = ArrInd;
qsortseg(arr,k,1,end);
return end
}
function qsortseg(arr,k,start,end, left,right,sepval,tmp,tmpe,tmps) {
# handle two-element case explicitely for a tiny speedup
if ((end - start) == 1) {
if (arr[tmps = k[start]] > arr[tmpe = k[end]]) {
k[start] = tmpe
k[end] = tmps
}
return
}
left = start;
right = end;
sepval = arr[k[int((left + right) / 2)]]
# Make every element <= sepval be to the left of every element > sepval
while (left < right) {
while (arr[k[left]] < sepval)
left++
while (arr[k[right]] > sepval)
right--
if (left < right) {
tmp = k[left]
k[left++] = k[right]
k[right--] = tmp
}
}
if (left == right)
if (arr[k[left]] < sepval)
left++
else
right--
if (start < right)
qsortseg(arr,k,start,right)
if (left < end)
qsortseg(arr,k,left,end)
}
' "$@"

127
examples/scripts.v2/pf Normal file
View file

@ -0,0 +1,127 @@
#! /bin/bash
#
# original from:
#
# @(#) p.ksh 1.1 93/11/09
# p: page compressed & plain files in the order given
# 92/01/23 john h. dubois iii (john@armory.com)
# 92/02/14 changed incorrect zpack to pcat
# 92/02/16 added help
# 92/10/11 search for file.Z and file.z if file not found
# 92/10/18 pass options to pager
# 93/11/09 Understand gzipped files too
# Wait after printing message about unreadable files
# Make less prompt include name of file being uncompressed
#
# conversion to bash v2 by Chet Ramey; renamed to pf
#
DefPager=/local/bin/less
istrue()
{
test 0 -ne "$1"
}
warn()
{
echo "$@" 1>&2
}
if [ "$1" = -h ]; then
echo \
"$0: page a file.
Usage: $0 [pager-option ...] [filename ...]
Files are paged by the program specified in the user's PAGER
environment variable, or by $DefPager if PAGER is not set.
If no filename is given, text to page is read from the standard input.
If filenames are given, they are either paged directly, or unpacked/
uncompressed and then paged. Files are assumed to be in packed, compressed,
or gzipped format if the filename ends in .Z, .z, or .gz respectively.
If a filename that does not end in .Z, .z, or .gz is not found, it is
searched for with one of those extensions attached.
Each group of plain files is paged by a single instance of the pager.
Each packed or compressed file is paged by a separate instance of the
pager.
Initial arguments beginning with + or - are taken to be pager options and
are passed to each instance of the pager.
If a pager option takes a value it should be given with the option as a
single argument (with no space between the option and the value)."
exit 0
fi
# Get pager options
while [ $# -gt 0 ]; do
case "$1" in
-*|+*) Opts="$Opts $1" ; shift;;
*) break;;
esac
done
[ -z "$PAGER" ] && PAGER=$DefPager
# Read from stdin
[ $# = 0 ] && exec $PAGER $Opts
typeset -i filenum=0 badfile=0
for file; do
if [ ! -r "$file" ]; then
case "$file" in
*.[Zz]|*.gz)
# Check if user specified a compressed file without giving its extension
for ext in Z z gz; do
if [ -r "$file.$ext" ]; then
file="$file.$ext"
break
fi
done;;
esac
fi
if [ ! -r "$file" ]; then
warn "$file: cannot read."
badfile=1
else
files[filenum]=$file
let filenum+=1
fi
done
if istrue $badfile && [ $filenum -gt 0 ]; then
echo -n "Press return to continue..." 1>&2
read
fi
unset plain
for file in "${files[@]}"; do
case "$file" in
*.[zZ]|*.gz)
set -- Z zcat z pcat gz gzcat
# Find correct uncompression program
while [ $# -gt 0 ]; do
case "$file" in
*.$1)
# Page any uncompressed files so that they will be read
# in the correct order
[ ${#plain[@]} -gt 0 ] && $PAGER $Opts "${plain[@]}"
unset plain[*]
# If page is less, set the prompt to include the name of
# the file being uncompressed. Escape the . in the extension
# because less treats is specially in prompts (other dots
# in filenames will still be mucked with).
case "$PAGER" in
*less) Prompt="-P[${file%.$1}\\.$1] (%pb\\%)" ;;
*) unset Prompt ;;
esac
$2 "$file" | $PAGER "$Prompt" $Opts
break
esac
shift 2
done
;;
*) plain[${#plain[@]}]=$file;;
esac
done
# Page any uncompressed files that haven't been paged yet
[ ${#plain[@]} -gt 0 ] && exec $PAGER $Opts "${plain[@]}"

25
examples/scripts.v2/pmtop Normal file
View file

@ -0,0 +1,25 @@
#! /bin/bash
#
# pmtop - poor man's `top' for SunOS 4.x
#
CLEAR=clear # could also be 'tput clear'
HEADER="USER PID %CPU %MEM SZ RSS TT STAT START TIME COMMAND"
if [ -n "$LINES" ]; then
SS=$(( $LINES - 2 ))
else
SS=20
fi
while :
do
$CLEAR
echo "$HEADER"
ps -aux | sort -nr +2 | sed ${SS}q
sleep 5
done
exit 0

122
examples/scripts.v2/rename Normal file
View file

@ -0,0 +1,122 @@
#! /bin/bash
#
# original from:
# @(#) rename.ksh 1.1 94/05/10
# 90/06/01 John DuBois (spcecdt@armory.com)
# 91/02/25 Improved help info
# 92/06/07 remove quotes from around shell pattern as required by new ksh
# 94/05/10 Exit if no globbing chars given.
#
# conversion to bash v2 syntax by Chet Ramey
phelp()
{
echo "$usage
All files that match oldpattern will be renamed with the
filename components that match the constant parts of oldpattern
changed to the corresponding constant parts of newpattern.
The components of the filename that match variable parts of
oldpattern will be preserved. Variable parts in oldpattern
must occur in the same order in newpattern. Variables parts
can be '?' and '*'.
Example:
rename \"/tmp/foo*.ba.?\" \"/tmp/new*x?\"
All files in /tmp that match foo*.ba.? will have the \"foo\" part
replaced by \"new\" and the \".ba.\" part replaced by \"x\"."
}
usage="usage: $name [-htv] oldpattern newpattern"
name=${0##/}
while getopts "htv" opt; do
case "$opt" in
t) tell=true;;
v) verbose=true;;
h) phelp; exit 0;;
*) echo "$name: $usage" 1>&2; exit 2;;
esac
done
shift $((OPTIND - 1))
if [ $# -lt 2 ]; then
phelp
exit 2
fi
oldpat=$1
newpat=$2
set $1
if [ ! -a "$1" ]; then
echo "$name: no files match $oldpat."
exit 1
fi
typeset -i i=1 j
# Example oldpat: foo*.a
# Example newpat: bar*.b
# Examples given for first iteration (in the example, the only interation)
while :; do
case "$oldpat" in
*[\*\?]*) ;;
*) break;;
esac
# Get leftmost globbing pattern in oldpat
pat=${oldpat#*[\*\?]} # pat=.a
pat=${oldpat%%"$pat"} # pat=foo*
pat=${pat##*[!\?\*]} # pat=*
# Find parts before & after pattern
oldpre[i]=${oldpat%%"$pat"*} # oldpre[1]=foo
oldsuf[i]=${oldpat#*"$pat"} # oldsuf[1]=.a
newpre[i]=${newpat%%"$pat"*} # newpre[1]=bar
# Get rid of processed part of patterns
oldpat=${oldpat#${oldpre[i]}"$pat"} # oldpat=.a
newpat=${newpat#${newpre[i]}"$pat"} # newpat=.b
let i=i+1
done
if [ $i -eq 1 ]; then
print -u2 "No globbing chars in pattern."
exit 1
fi
oldpre[i]=${oldpat%%"$pat"*} # oldpre[2]=.a
oldsuf[i]=${oldpat#*"$pat"} # oldsuf[2]=.a
newpre[i]=${newpat%%"$pat"*} # newpre[2]=.b
if [ -n "$verbose" ]; then
j=1
while let "j < i"; do
echo \
"Old prefix: ${oldpre[j]} Old suffix: ${oldsuf[j]} New prefix: ${newpre[j]}"
let j=j+1
done
fi
# Example file: foox.a
for file; do
j=1
origname=$file # origname=foox.a
newfile=
while let "j <= i"; do
# Peel off a prefix interation 1 2
file=${file#${oldpre[j]}} # file=x.a file=
# Save the part of this prefix that is to be retained
const=${file%${oldsuf[j]}} # const=x const=
newfile=$newfile${newpre[j]}$const # newfile=barx newfile=barx.b
file=${file#$const} # file=.a file=.a
let j=j+1
done
if [ -n "$tell" ]; then
echo "Would move \"$origname\" to \"$newfile\"."
else
if [ -n "$verbose" ]; then
echo "Moving \"$origname\" to \"$newfile\"."
fi
mv $origname $newfile
fi
done

119
examples/scripts.v2/repeat Normal file
View file

@ -0,0 +1,119 @@
#! /bin/bash
#
# original from:
# repeat: repeat a command.
# @(#) repeat.ksh 1.1 93/06/03
# 90/05 john h. dubois iii (john@armory.com)
# 90/11 added help
# 93/06/03 Added s, h, p, and v options
#
# conversion to bash v2 syntax done by Chet Ramey
istrue()
{
test 0 -ne "$1"
}
isfalse()
{
test 0 -eq "$1"
}
phelp()
{
echo "$name: repeatedly execute a command line.
$Usage
commandline is executed once for each integer from startcount through endcount
inclusive. The default for startcount is 1 if a positive endcount or no
endcount is given, and -1 if a negative endcount is given. A count
parameter consisting of a single number is taken to be an endcount. If
only an endcount is given and it is positive, commandline is executed
endcount times. endcount may be less than startcount. If no endcount is
given (e.g. a count parameter of \"10-\"), commandline execution repeats
indefinitely with the iteration variable incrementing in a positive
direction. A count parameter of consisting of \"-\" will repeat
indefinitely starting with 1.
Note that quoting and variables in commandline are interpreted twice, once
when it is passed to the repeat command, and once when it is actually executed.
The iteration variable is \"count\". If \$count is used in commandline, make
sure it is quoted with ' or \.
Options:
-h: Print this help.
-p: Print value of iteration variable on stderr before each iteration.
-s <sec>: sleep for <sec> seconds after each iteration except the last.
-v: Print start and end values before beginning."
}
name=${0##*/}
Usage="Usage: repeat [-hpv] [-s <sec>] [[startcount]-][endcount] command [arg ...]"
typeset -i count=1 forever=0 sleep=0 print=0 verbose=0
while getopts :0123456789hpvs: opt; do
case $opt in
h) phelp; exit 0;;
s) sleep=$OPTARG || exit 1;;
p) print=1;;
v)verbose=1;;
[0-9]) break;;
+?) echo "$name: options should not be preceded by a '+'." 1>&2; exit 2;;
?) echo "$name: $OPTARG: bad option. Use -h for help." 1>&2; exit 2;;
esac
done
# remove args that were options
shift $((OPTIND-1))
if [ $# -lt 2 ]; then
echo -e "$Usage\nUse -h for help." 1>&2
exit 2
fi
case "$1" in
-[0-9]*-|[0-9]*-)
# Start value only
count=${1%-}
forever=1
;;
-[0-9]*-[0-9]*|[0-9]*-[0-9]*)
# Start and end value
s=${1%-}
end=${s##[0-9]*-}
count=${s%-$end}
;;
-[0-9]*|[0-9]*)
end=$1
case "$end" in
-\*) count=-1;;
esac
;;
-)
forever=1
;;
*)
echo "$name: bad count parameter: $1" 1>&2
exit 1
;;
esac
shift
[ -z "$end" -o $count -le "$end" ] && increment=1 || increment=-1
istrue $verbose && echo "start=$count end=$end" 1>&2
# Need to do this here so that up to this point, -0 will keep the leading -
# and end will not be 0 if no value assigned
typeset -i end
let end+=increment # make loop inclusive of original endcount
while istrue $forever || [ $count -ne $end ]; do
istrue $print && echo $count 1>&2
eval "$@"
istrue $sleep && sleep $sleep
let count+=increment
done

66
examples/scripts.v2/shprof Normal file
View file

@ -0,0 +1,66 @@
#! /bin/bash
#
# shprof - a line profiler for shell scripts
#
# adapted from a similar program included in `The New KornShell' by
# Bolsky and Korn and posted to usenet by bsh20858@challenger.fhda.edu
#
# converted to bash v2 syntax by Chet Ramey
#
TMPFILE=${TMP:-/tmp}/shprof$$
trap 'rm -f $TMPFILE' EXIT
errexit()
{
echo $0: "$@" >&2
exit 1
}
# create script with profiling enabled
cat > $TMPFILE <<- \_EOF_
declare -a _line
_profend()
{
case "$1" in
/*|./*) file="$1" ;;
*) file=$(type -path "$1") ;;
esac
echo "*** line profile for $file ***"
i=1;
while read -r && [ $i -le $NLINE ]; do
count=${_line[$i]}
if [ "$count" -gt 0 ]; then
echo "[$count] $i: $REPLY"
fi
i=$((i + 1))
done <$file
_EOF_
# make the profiling script remove itself after printing line stats
echo "rm -f $TMPFILE" >> $TMPFILE
cat >> $TMPFILE <<- \_EOF_
}
_command=$1
shift
i=1
NLINE=$(wc -l < "$_command")
while [ $i -le $NLINE ]; do
_line[$i]=0
i=$((i + 1))
done
unset i
trap "_profend ${_command}" EXIT
trap '_line[$LINENO]=$((${_line[$LINENO]} + 1))' DEBUG
LINENO=0
_EOF_
case "$1" in
/*|./*) file=$1 ;;
*) file=$((type -path "$1")) ;;
esac
cat "${file-$1}" >> $TMPFILE || errexit "${1}: cannot open"
chmod +x $TMPFILE
exec -a "$file" $TMPFILE "$@"

80
examples/scripts.v2/untar Normal file
View file

@ -0,0 +1,80 @@
#! /bin/bash
#
# original from:
# @(#) untar.ksh 1.0 93/11/10
# 92/10/08 john h. dubois iii (john@armory.com)
# 92/10/31 make it actually work if archive isn't in current dir!
# 93/11/10 Added pack and gzip archive support
#
# conversion to bash v2 syntax done by Chet Ramey
phelp()
{
echo \
"$name: extract tar archives into directories, uncompressing if neccessary.
Usage: $name archive[.tar[.[Z|gz]]] ..
If an archive name given does not end in .tar, .tar.Z, or .tar.gz, it is
searched for first with .tar added, then .tar.Z, and then .tar.gz added.
The real filename must end in either .tar, .tar.Z, or .tar.gz. A
directory with the name of the archive is created in the current directory
(not necessarily the directory that the archive is in) if it does not
exist, and the the contents of the archive are extracted into it.
Absolute pathnames in tarfiles are suppressed."
}
if [ $# -eq 0 ]; then
phelp
exit 1
fi
name=${0##/}
OWD=$PWD
for file; do
cd $OWD
case "$file" in
*.tar.Z) ArchiveName=${file%%.tar.Z} zcat=zcat;;
*.tar.z) ArchiveName=${file%%.tar.z} zcat=pcat;;
*.tar.gz) ArchiveName=${file%%.tar.gz} zcat=gzcat;;
*) ArchiveName=$file
for ext in "" .Z .z .gz; do
if [ -f "$file.tar$ext" ]; then
file="$file.tar$ext"
break
fi
done
if [ ! -f "$file" ]; then
echo "$file: cannot find archive." 1>&2
continue
fi
;;
esac
if [ ! -r "$file" ]; then
echo "$file: cannot read." >&2
continue
fi
DirName=${ArchiveName##*/}
[ -d "$DirName" ] || {
mkdir "$DirName" || {
echo "$DirName: could not make archive directory." 1>&2
continue
}
}
cd $DirName || {
echo "$name: cannot cd to $DirName" 1>&2
continue
}
case "$file" in
/*) ;;
*) file=$OWD/$file ;;
esac
echo "Extracting archive $file into directory $DirName..."
case "$file" in
*.tar.Z|*.tar.z|*.tar.gz) $zcat $file | tar xvf -;;
*.tar) tar xvf $file;;
esac
echo "Done extracting archive $file into directory $DirName."
done

45
examples/scripts.v2/uudec Normal file
View file

@ -0,0 +1,45 @@
:
# @(#) uudec.sh 1.0 93/11/22
# 92/08/04 john@armory.com (John H. DuBois III)
# 93/11/22 Added help.
isfalse()
{
test 0 -eq "$1"
}
phelp()
{
"$name: process uuencoded files.
Usage: uudec [-h] filename ...
Options:
-h: Print this help."
}
name=${0##*/}
typeset -i force=0
while getopts "hf" opt; do
case "$opt" in
h) phelp; exit 0;;
f) force=1;;
*) echo "$Usage" 1>&2; exit 2;;
esac
done
shift $((OPTIND - 1))
for file; do
echo "$file"
while read b mode filename && [ "$b" != begin ]; do :; done < "$file"
if [ "$b" = begin ]; then
if [ -f "$filename" ] && isfalse $force; then
echo "Output file \"$filename\" exists. Not written."
else
uudecode "$file"
fi
else
echo "No begin line."
fi
done

69
examples/scripts.v2/uuenc Normal file
View file

@ -0,0 +1,69 @@
#! /bin/bash
#
# original from:
# @(#) uuenc.ksh 1.0 93/09/18
# 93/09/18 john h. dubois iii (john@armory.com)
#
# conversion to bash v2 syntax by Chet Ramey
istrue()
{
test 0 -ne "$1"
}
isfalse()
{
test 0 -eq "$1"
}
phelp()
{
echo "$name: uuencode files.
$Usage
For each filename given, $name uuencodes the file, using the final
component of the file's path as the stored filename in the uuencoded
archive and, with a .${SUF} appended, as the name to store the archive in.
Example:
$name /tmp/foo
The file /tmp/foo is uuencoded, with \"foo\" stored as the name to uudecode
the file into, and the output is stored in a file in the current directory
with the name \"foo.${SUF}\".
Options:
-f: Normally, if the file the output would be stored in already exists,
it is not overwritten and an error message is printed. If -f (force)
is given, it is silently overwritten.
-h: Print this help."
}
name=${0##*/}
Usage="Usage: $name [-hf] <filename> ..."
typeset -i force=0
SUF=uu
while getopts :hf opt; do
case $opt in
h) phelp; exit 0;;
f) force=1;;
+?) echo "$name: options should not be preceded by a '+'." 1>&2 ; exit 2;;
?) echo "$name: $OPTARG: bad option. Use -h for help." 1>&2 ; exit 2;;
esac
done
# remove args that were options
shift $((OPTIND - 1))
if [ $# -lt 1 ]; then
echo "$Usage\nUse -h for help." 1>&2
exit
fi
for file; do
tail=${file##*/}
out="$tail.${SUF}"
if isfalse $force && [ -a "$out" ]; then
echo "$name: $out: file exists. Use -f to overwrite." 1>&2
else
uuencode $file $tail > $out
fi
done

137
examples/scripts.v2/vtree Normal file
View file

@ -0,0 +1,137 @@
#! /bin/bash
#
# original from:
# vtree: visual directory tree
# @(#) vtree.sh 1.1 91/07/01
# 90/04 john h. dubois iii (john@armory.com)
# 91/07/01 fixed bug that caused problems when dir given on command line,
# added some info to help, changed to 4-space indenting
#
# conversion to bash v2 syntax done by Chet Ramey
#
help=\
"Syntax: vtree [startdir] [namelen=#] [linelen=#]
If startdir is not specified, tree will start at current dir.
namelen specifies the minimum number of characters of a directory name that
are guaranteed to be printed.
This is a tradeoff between the number of tree levels that can fit on a
screen line and the number of chars of each dir name that can be printed.
In most cases it will be possible to print more than namelen characters of
the name (a name up to namelen+1 chars will always be printed in full),
but in some cases truncation down to namelen chars will occur.
If truncation occurs, a '>' is printed at the end of the name.
namelen=8 (the default) typically causes about 5 dirs/1000 to be truncated.
namelen=7 typically causes about 10 dirs/1000 to be truncated.
namelen=8 will allow 6 full length dirs to be printed in 79 columns.
namelen=7 will allow 7 full length dirs to be printed in 79 columns;
linelen specifies the maximum number of characters to print on one screen
line. All characters beyond this are truncated. The default is 1024.
To avoid line wrap on an 80 column terminal with autowrap, use linelen=79.
"
for i in "$@"; do
case $i in
-h) echo "$help"; exit;;
*=*)
vars="$vars $i"
;;
*)
if [ ! -x $i -o ! -d $i ]; then # arg must be a dir and executable
echo "$i: directory not accessible."
exit
fi
cd $i
;;
esac
shift
done
pwd # print path of root of tree
# find all directories depth first; ignore permission errors
find . -type d -print 2> /dev/null | \
gawk -F/ '
# Do this block for NR == 1 instead of BEGIN because command line var
# assignments are not done until after BEGIN block is executed.
NR == 1 {
if (namelen)
MaxLen = namelen;
else
MaxLen = 8;
if (!linelen)
linelen = 1024
HSpace = substr(" ",1,MaxLen); # used to indent tree
n = 0; # number of dirs found on one major branch
}
$0 != "." { # do for every line produced by find except tree root dir
if (NF == 2 && n > 0) # print major branch whenever a new one starts
list();
Depth[n] = NF - 1; # record depth and name of dir
Name[n++] = $NF;
}
END {
list() # print last major branch
}
function list() {
Line = Name[0]; # initialize first line of branch to be branch base
for (i = 1; i < n; i++) { # for each name in major branch
if (Depth[i] == Depth[i-1] + 1)
AddHLink(); # if moving deeper into branch, use same line
else {
print substr(Line,1,linelen); # last line is done; print it
Line = ""; # start new line
# print indentation, vert links, and vert/horiz links
for (d = 1; d < Depth[i] - 1; d++) # for each level of indentation
# if a vert. link has been established for this level
if (VLink[d])
Line = Line HSpace " | ";
else # print empty indentation
Line = Line HSpace " ";
# Print last part of vert. link
if (VLink[d] == i) {
VLink[d] = 0; # mark level for no vert link
Line = Line HSpace " \\--";
}
else
Line = Line HSpace " |--";
}
Line = Line Name[i]; # Add dir name to line
}
print substr(Line,1,linelen); # print last line of major branch
n = 0; # reset name counter
}
function AddHLink() {
NDepth = Depth[i]; # Depth of this name
VLink[NDepth - 1] = 0;
# search until a name found at a level less than this one
for (j = i + 1; j < n && Depth[j] >= NDepth; j++)
# keep track of last name that VLink should connect to
if (Depth[j] == NDepth)
VLink[NDepth - 1] = j;
if (VLink[NDepth - 1]) {
NLine = substr(Line,1,(NDepth - 2) * (MaxLen + 4) + MaxLen + 1);
if (length(NLine) < length(Line))
Line = substr(NLine,1,length(NLine) - 1) ">"
else
Line = NLine;
Line = Line substr("--------------+--",
18 - ((NDepth - 1) * (MaxLen + 4) - length(Line)));
}
else {
NLine = substr(Line,1,(NDepth - 2) * (MaxLen + 4) + MaxLen + 3);
if (length(NLine) < length(Line))
Line = substr(NLine,1,length(NLine) - 1) ">"
else
Line = NLine;
Line = Line substr("-----------------",
1,(NDepth - 1) * (MaxLen + 4) - length(Line));
}
}
' $vars

111
examples/scripts.v2/where Normal file
View file

@ -0,0 +1,111 @@
#! /bin/bash
#
# original from:
# @(#) where.ksh 1.1 94/07/11
# 91/01/12 john h. dubois iii (john@armory.com)
# 92/08/10 Only print executable *files*.
# 92/10/06 Print err msg if no match found.
# 92/11/27 Added implicit *
# 93/07/23 Print help only if -h is given.
# 94/01/01 Added -x option
# 94/07/11 Don't bother with eval
#
# conversion to bash v2 syntax done by Chet Ramey
name=${0##*/}
Usage="Usage: $name [-hx] 'pattern' ..."
typeset -i exact=0
phelp()
{
echo "$name: find executable files in PATH that match patterns.
$Usage
$name searches each directory specified in the PATH environment variable
for executable files that match the specified patterns. Patterns are
given as Korn shell filename patterns. They are surrounded by implicit
'*' characters, so that \"foo\" will match any executble file whose name
contains contains \"foo\". This can be overridden by using '^' and '$' to
force a match to start at the beginning and end at the end of a filename
respectively. Characters that are special to the shell must generally
be protected from the shell by surrounding them with quotes.
Examples:
$name foo
lists all executable files in PATH that contain foo.
$name '^b*sh$'
lists all executable files in PATH that start with b and end with sh.
An error message is printed if a no matching file is found for a pattern.
Options:
-h: Print this help.
-x: Find exact matches only; equivalent to putting ^ and $ at the start
and end of each pattern."
}
istrue()
{
test 0 -ne "$1"
}
isfalse()
{
test 0 -eq "$1"
}
while getopts "xh" opt; do
case "$opt" in
x) exact=1;;
h) phelp ; exit 0;;
*) echo -e "$Usage\nUse -h for help." 1>&2; exit 2;;
esac
done
shift $((OPTIND-1))
set +f # make sure filename globbing is on
Args=("$@") # save args
OIFS=$IFS
IFS=: # Make PATH be split on :
Paths=($PATH)
IFS=$OIFS
for arg in "${Args[@]}"; do
# get rid of leading ^
if istrue $exact; then
arg=${arg}
else
case "$arg" in
^*) arg=${arg#?};;
*) arg="*$arg" ;; # Pattern is not anchored at start
esac
fi
# get rid of trailing $
if istrue $exact; then
arg="$arg"
else
case "$arg" in
*\$) arg=${arg%?} ;;
*) arg="$arg*" ;;
esac
fi
found=0 # Pattern not found yet
Patterns=
# Make a pattern for each element of PATH
for PathElem in "${Paths[@]}"; do
[ -z "$PathElem" ] && PathElem=.
Patterns="$Patterns $PathElem/$arg"
done
# Find all pattern matches that are executable regular files.
for file in $Patterns; do
if [ -x "$file" ] && [ -f "$file" ]; then
echo "$file"
found=1
fi
done
if [ $found = 0 ]; then
echo "$arg: not found." 1>&2
fi
done