These scripts, while not fitting into the text of this document, do illustrate some interesting shell programming techniques. They are useful, too. Have fun analyzing and running them.
Example A-1. manview: Viewing formatted manpages
#!/bin/bash # manview.sh: Formats the source of a man page for viewing. # This script is useful when writing man page source. # It lets you look at the intermediate results on the fly #+ while working on it. E_WRONGARGS=65 if [ -z "$1" ] then echo "Usage: `basename $0` filename" exit $E_WRONGARGS fi # --------------------------- groff -Tascii -man $1 | less # From the man page for groff. # --------------------------- # If the man page includes tables and/or equations, #+ then the above code will barf. # The following line can handle such cases. # # gtbl < "$1" | geqn -Tlatin1 | groff -Tlatin1 -mtty-char -man # # Thanks, S.C. exit 0 |
Example A-2. mailformat: Formatting an e-mail message
#!/bin/bash
# mail-format.sh: Format e-mail messages.
# Gets rid of carets, tabs, also fold excessively long lines.
# =================================================================
# Standard Check for Script Argument(s)
ARGS=1
E_BADARGS=65
E_NOFILE=66
if [ $# -ne $ARGS ] # Correct number of arguments passed to script?
then
echo "Usage: `basename $0` filename"
exit $E_BADARGS
fi
if [ -f "$1" ] # Check if file exists.
then
file_name=$1
else
echo "File \"$1\" does not exist."
exit $E_NOFILE
fi
# =================================================================
MAXWIDTH=70 # Width to fold long lines to.
# Delete carets and tabs at beginning of lines,
#+ then fold lines to $MAXWIDTH characters.
sed '
s/^>//
s/^ *>//
s/^ *//
s/ *//
' $1 | fold -s --width=$MAXWIDTH
# -s option to "fold" breaks lines at whitespace, if possible.
# This script was inspired by an article in a well-known trade journal
#+ extolling a 164K Windows utility with similar functionality.
#
# An nice set of text processing utilities and an efficient
#+ scripting language provide an alternative to bloated executables.
exit 0 |
Example A-3. rn: A simple-minded file rename utility
This script is a modification of Example 12-18.
#! /bin/bash
#
# Very simpleminded filename "rename" utility (based on "lowercase.sh").
#
# The "ren" utility, by Vladimir Lanin (lanin@csd2.nyu.edu),
#+ does a much better job of this.
ARGS=2
E_BADARGS=65
ONE=1 # For getting singular/plural right (see below).
if [ $# -ne "$ARGS" ]
then
echo "Usage: `basename $0` old-pattern new-pattern"
# As in "rn gif jpg", which renames all gif files in working directory to jpg.
exit $E_BADARGS
fi
number=0 # Keeps track of how many files actually renamed.
for filename in *$1* #Traverse all matching files in directory.
do
if [ -f "$filename" ] # If finds match...
then
fname=`basename $filename` # Strip off path.
n=`echo $fname | sed -e "s/$1/$2/"` # Substitute new for old in filename.
mv $fname $n # Rename.
let "number += 1"
fi
done
if [ "$number" -eq "$ONE" ] # For correct grammar.
then
echo "$number file renamed."
else
echo "$number files renamed."
fi
exit 0
# Exercises:
# ---------
# What type of files will this not work on?
# How can this be fixed?
#
# Rewrite this script to process all the files in a directory
#+ containing spaces in their names, and to rename them,
#+ substituting an underscore for each space. |
Example A-4. blank-rename: renames filenames containing blanks
This is an even simpler-minded version of previous script.
#! /bin/bash
# blank-rename.sh
#
# Substitutes underscores for blanks in all the filenames in a directory.
ONE=1 # For getting singular/plural right (see below).
number=0 # Keeps track of how many files actually renamed.
FOUND=0 # Successful return value.
for filename in * #Traverse all files in directory.
do
echo "$filename" | grep -q " " # Check whether filename
if [ $? -eq $FOUND ] #+ contains space(s).
then
fname=$filename # Strip off path.
n=`echo $fname | sed -e "s/ /_/g"` # Substitute underscore for blank.
mv "$fname" "$n" # Do the actual renaming.
let "number += 1"
fi
done
if [ "$number" -eq "$ONE" ] # For correct grammar.
then
echo "$number file renamed."
else
echo "$number files renamed."
fi
exit 0 |
Example A-5. encryptedpw: Uploading to an ftp site, using a locally encrypted password
#!/bin/bash # Example "ex72.sh" modified to use encrypted password. # Note that this is still rather insecure, #+ since the decrypted password is sent in the clear. # Use something like "ssh" if this is a concern. E_BADARGS=65 if [ -z "$1" ] then echo "Usage: `basename $0` filename" exit $E_BADARGS fi Username=bozo # Change to suit. pword=/home/bozo/secret/password_encrypted.file # File containing encrypted password. Filename=`basename $1` # Strips pathname out of file name Server="XXX" Directory="YYY" # Change above to actual server name & directory. Password=`cruft <$pword` # Decrypt password. # Uses the author's own "cruft" file encryption package, #+ based on the classic "onetime pad" algorithm, #+ and obtainable from: #+ Primary-site: ftp://ibiblio.org/pub/Linux/utils/file #+ cruft-0.2.tar.gz [16k] ftp -n $Server <<End-Of-Session user $Username $Password binary bell cd $Directory put $Filename bye End-Of-Session # -n option to "ftp" disables auto-logon. # "bell" rings 'bell' after each file transfer. exit 0 |
Example A-6. copy-cd: Copying a data CD
#!/bin/bash
# copy-cd.sh: copying a data CD
CDROM=/dev/cdrom # CD ROM device
OF=/home/bozo/projects/cdimage.iso # output file
# /xxxx/xxxxxxx/ Change to suit your system.
BLOCKSIZE=2048
SPEED=2 # May use higher speed if supported.
echo; echo "Insert source CD, but do *not* mount it."
echo "Press ENTER when ready. "
read ready # Wait for input, $ready not used.
echo; echo "Copying the source CD to $OF."
echo "This may take a while. Please be patient."
dd if=$CDROM of=$OF bs=$BLOCKSIZE # Raw device copy.
echo; echo "Remove data CD."
echo "Insert blank CDR."
echo "Press ENTER when ready. "
read ready # Wait for input, $ready not used.
echo "Copying $OF to CDR."
cdrecord -v -isosize speed=$SPEED dev=0,0 $OF
# Uses Joerg Schilling's "cdrecord" package (see its docs).
# http://www.fokus.gmd.de/nthp/employees/schilling/cdrecord.html
echo; echo "Done copying $OF to CDR on device $CDROM."
echo "Do you want to erase the image file (y/n)? " # Probably a huge file.
read answer
case "$answer" in
[yY]) rm -f $OF
echo "$OF erased."
;;
*) echo "$OF not erased.";;
esac
echo
# Exercise:
# Change the above "case" statement to also accept "yes" and "Yes" as input.
exit 0 |
Example A-7. Collatz series
#!/bin/bash
# collatz.sh
# The notorious "hailstone" or Collatz series.
# -------------------------------------------
# 1) Get the integer "seed" from the command line.
# 2) NUMBER <--- seed
# 3) Print NUMBER.
# 4) If NUMBER is even, divide by 2, or
# 5)+ if odd, multiply by 3 and add 1.
# 6) NUMBER <--- result
# 7) Loop back to step 3 (for specified number of iterations).
#
# The theory is that every sequence,
#+ no matter how large the initial value,
#+ eventually settles down to repeating "4,2,1..." cycles,
#+ even after fluctuating through a wide range of values.
#
# This is an instance of an "iterate",
#+ an operation that feeds its output back into the input.
# Sometimes the result is a "chaotic" series.
MAX_ITERATIONS=200
# For large seed numbers (>32000), increase MAX_ITERATIONS.
h=${1:-$$} # Seed
# Use $PID as seed,
#+ if not specified as command-line arg.
echo
echo "C($h) --- $MAX_ITERATIONS Iterations"
echo
for ((i=1; i<=MAX_ITERATIONS; i++))
do
echo -n "$h "
# ^^^^^
# tab
let "remainder = h % 2"
if [ "$remainder" -eq 0 ] # Even?
then
let "h /= 2" # Divide by 2.
else
let "h = h*3 + 1" # Multiply by 3 and add 1.
fi
COLUMNS=10 # Output 10 values per line.
let "line_break = i % $COLUMNS"
if [ "$line_break" -eq 0 ]
then
echo
fi
done
echo
# For more information on this mathematical function,
#+ see "Computers, Pattern, Chaos, and Beauty", by Pickover, p. 185 ff.,
#+ as listed in the bibliography.
exit 0 |
Example A-8. days-between: Calculate number of days between two dates
#!/bin/bash
# days-between.sh: Number of days between two dates.
# Usage: ./days-between.sh [M]M/[D]D/YYYY [M]M/[D]D/YYYY
#
# Note: Script modified to account for changes in Bash 2.05b
#+ that closed the loophole permitting large negative
#+ integer return values.
ARGS=2 # Two command line parameters expected.
E_PARAM_ERR=65 # Param error.
REFYR=1600 # Reference year.
CENTURY=100
DIY=365
ADJ_DIY=367 # Adjusted for leap year + fraction.
MIY=12
DIM=31
LEAPCYCLE=4
MAXRETVAL=255 # Largest permissable
#+ positive return value from a function.
diff= # Declare global variable for date difference.
value= # Declare global variable for absolute value.
day= # Declare globals for day, month, year.
month=
year=
Param_Error () # Command line parameters wrong.
{
echo "Usage: `basename $0` [M]M/[D]D/YYYY [M]M/[D]D/YYYY"
echo " (date must be after 1/3/1600)"
exit $E_PARAM_ERR
}
Parse_Date () # Parse date from command line params.
{
month=${1%%/**}
dm=${1%/**} # Day and month.
day=${dm#*/}
let "year = `basename $1`" # Not a filename, but works just the same.
}
check_date () # Checks for invalid date(s) passed.
{
[ "$day" -gt "$DIM" ] || [ "$month" -gt "$MIY" ] || [ "$year" -lt "$REFYR" ] && Param_Error
# Exit script on bad value(s).
# Uses "or-list / and-list".
#
# Exercise: Implement more rigorous date checking.
}
strip_leading_zero () # Better to strip possible leading zero(s)
{ #+ from day and/or month
return ${1#0} #+ since otherwise Bash will interpret them
} #+ as octal values (POSIX.2, sect 2.9.2.1).
day_index () # Gauss' Formula:
{ # Days from Jan. 3, 1600 to date passed as param.
day=$1
month=$2
year=$3
let "month = $month - 2"
if [ "$month" -le 0 ]
then
let "month += 12"
let "year -= 1"
fi
let "year -= $REFYR"
let "indexyr = $year / $CENTURY"
let "Days = $DIY*$year + $year/$LEAPCYCLE - $indexyr + $indexyr/$LEAPCYCLE + $ADJ_DIY*$month/$MIY + $day - $DIM"
# For an in-depth explanation of this algorithm, see
#+ http://home.t-online.de/home/berndt.schwerdtfeger/cal.htm
echo $Days
}
calculate_difference () # Difference between to day indices.
{
let "diff = $1 - $2" # Global variable.
}
abs () # Absolute value
{ # Uses global "value" variable.
if [ "$1" -lt 0 ] # If negative
then #+ then
let "value = 0 - $1" #+ change sign,
else #+ else
let "value = $1" #+ leave it alone.
fi
}
if [ $# -ne "$ARGS" ] # Require two command line params.
then
Param_Error
fi
Parse_Date $1
check_date $day $month $year # See if valid date.
strip_leading_zero $day # Remove any leading zeroes
day=$? #+ on day and/or month.
strip_leading_zero $month
month=$?
let "date1 = `day_index $day $month $year`"
Parse_Date $2
check_date $day $month $year
strip_leading_zero $day
day=$?
strip_leading_zero $month
month=$?
date2=$(day_index $day $month $year) # Command substitution.
calculate_difference $date1 $date2
abs $diff # Make sure it's positive.
diff=$value
echo $diff
exit 0
# Compare this script with
#+ the implementation of Gauss' Formula in a C program at:
#+ http://buschencrew.hypermart.net/software/datedif |
Example A-9. Make a "dictionary"
#!/bin/bash
# makedict.sh [make dictionary]
# Modification of /usr/sbin/mkdict script.
# Original script copyright 1993, by Alec Muffett.
#
# This modified script included in this document in a manner
#+ consistent with the "LICENSE" document of the "Crack" package
#+ that the original script is a part of.
# This script processes text files to produce a sorted list
#+ of words found in the files.
# This may be useful for compiling dictionaries
#+ and for lexicographic research.
E_BADARGS=65
if [ ! -r "$1" ] # Need at least one
then #+ valid file argument.
echo "Usage: $0 files-to-process"
exit $E_BADARGS
fi
# SORT="sort" # No longer necessary to define options
#+ to sort. Changed from original script.
cat $* | # Contents of specified files to stdout.
tr A-Z a-z | # Convert to lowercase.
tr ' ' '\012' | # New: change spaces to newlines.
# tr -cd '\012[a-z][0-9]' | # Get rid of everything non-alphanumeric
#+ (original script).
tr -c '\012a-z' '\012' | # Rather than deleting
#+ now change non-alpha to newlines.
sort | # $SORT options unnecessary now.
uniq | # Remove duplicates.
grep -v '^#' | # Delete lines beginning with a hashmark.
grep -v '^$' # Delete blank lines.
exit 0 |
Example A-10. Soundex conversion
#!/bin/bash
# soundex.sh: Calculate "soundex" code for names
# =======================================================
# Soundex script
# by
# Mendel Cooper
# thegrendel@theriver.com
# 23 January, 2002
#
# Placed in the Public Domain.
#
# A slightly different version of this script appeared in
#+ Ed Schaefer's July, 2002 "Shell Corner" column
#+ in "Unix Review" on-line,
#+ http://www.unixreview.com/documents/uni1026336632258/
# =======================================================
ARGCOUNT=1 # Need name as argument.
E_WRONGARGS=70
if [ $# -ne "$ARGCOUNT" ]
then
echo "Usage: `basename $0` name"
exit $E_WRONGARGS
fi
assign_value () # Assigns numerical value
{ #+ to letters of name.
val1=bfpv # 'b,f,p,v' = 1
val2=cgjkqsxz # 'c,g,j,k,q,s,x,z' = 2
val3=dt # etc.
val4=l
val5=mn
val6=r
# Exceptionally clever use of 'tr' follows.
# Try to figure out what is going on here.
value=$( echo "$1" \
| tr -d wh \
| tr $val1 1 | tr $val2 2 | tr $val3 3 \
| tr $val4 4 | tr $val5 5 | tr $val6 6 \
| tr -s 123456 \
| tr -d aeiouy )
# Assign letter values.
# Remove duplicate numbers, except when separated by vowels.
# Ignore vowels, except as separators, so delete them last.
# Ignore 'w' and 'h', even as separators, so delete them first.
#
# The above command substitution lays more pipe than a plumber <g>.
}
input_name="$1"
echo
echo "Name = $input_name"
# Change all characters of name input to lowercase.
# ------------------------------------------------
name=$( echo $input_name | tr A-Z a-z )
# ------------------------------------------------
# Just in case argument to script is mixed case.
# Prefix of soundex code: first letter of name.
# --------------------------------------------
char_pos=0 # Initialize character position.
prefix0=${name:$char_pos:1}
prefix=`echo $prefix0 | tr a-z A-Z`
# Uppercase 1st letter of soundex.
let "char_pos += 1" # Bump character position to 2nd letter of name.
name1=${name:$char_pos}
# ++++++++++++++++++++++++++ Exception Patch +++++++++++++++++++++++++++++++++
# Now, we run both the input name and the name shifted one char to the right
#+ through the value-assigning function.
# If we get the same value out, that means that the first two characters
#+ of the name have the same value assigned, and that one should cancel.
# However, we also need to test whether the first letter of the name
#+ is a vowel or 'w' or 'h', because otherwise this would bollix things up.
char1=`echo $prefix | tr A-Z a-z` # First letter of name, lowercased.
assign_value $name
s1=$value
assign_value $name1
s2=$value
assign_value $char1
s3=$value
s3=9$s3 # If first letter of name is a vowel
#+ or 'w' or 'h',
#+ then its "value" will be null (unset).
#+ Therefore, set it to 9, an otherwise
#+ unused value, which can be tested for.
if [[ "$s1" -ne "$s2" || "$s3" -eq 9 ]]
then
suffix=$s2
else
suffix=${s2:$char_pos}
fi
# ++++++++++++++++++++++ end Exception Patch +++++++++++++++++++++++++++++++++
padding=000 # Use at most 3 zeroes to pad.
soun=$prefix$suffix$padding # Pad with zeroes.
MAXLEN=4 # Truncate to maximum of 4 chars.
soundex=${soun:0:$MAXLEN}
echo "Soundex = $soundex"
echo
# The soundex code is a method of indexing and classifying names
#+ by grouping together the ones that sound alike.
# The soundex code for a given name is the first letter of the name,
#+ followed by a calculated three-number code.
# Similar sounding names should have almost the same soundex codes.
# Examples:
# Smith and Smythe both have a "S-530" soundex.
# Harrison = H-625
# Hargison = H-622
# Harriman = H-655
# This works out fairly well in practice, but there are numerous anomalies.
#
#
# The U.S. Census and certain other governmental agencies use soundex,
# as do genealogical researchers.
#
# For more information,
#+ see the "National Archives and Records Administration home page",
#+ http://www.nara.gov/genealogy/soundex/soundex.html
# Exercise:
# --------
# Simplify the "Exception Patch" section of this script.
exit 0 |
Example A-11. "Game of Life"
#!/bin/bash
# life.sh: "Life in the Slow Lane"
# ##################################################################### #
# This is the Bash script version of John Conway's "Game of Life". #
# "Life" is a simple implementation of cellular automata. #
# --------------------------------------------------------------------- #
# On a rectangular grid, let each "cell" be either "living" or "dead". #
# Designate a living cell with a dot, and a dead one with a blank space.#
# Begin with an arbitrarily drawn dot-and-blank grid, #
#+ and let this be the starting generation, "generation 0". #
# Determine each successive generation by the following rules: #
# 1) Each cell has 8 neighbors, the adjoining cells #
#+ left, right, top, bottom, and the 4 diagonals. #
# 123 #
# 4*5 #
# 678 #
# #
# 2) A living cell with either 2 or 3 living neighbors remains alive. #
# 3) A dead cell with 3 living neighbors becomes alive (a "birth"). #
SURVIVE=2 #
BIRTH=3 #
# 4) All other cases result in dead cells. #
# ##################################################################### #
startfile=gen0 # Read the starting generation from the file "gen0".
# Default, if no other file specified when invoking script.
#
if [ -n "$1" ] # Specify another "generation 0" file.
then
if [ -e "$1" ] # Check for existence.
then
startfile="$1"
fi
fi
ALIVE1=.
DEAD1=_
# Represent living and "dead" cells in the start-up file.
# This script uses a 10 x 10 grid (may be increased,
#+ but a large grid will will cause very slow execution).
ROWS=10
COLS=10
GENERATIONS=10 # How many generations to cycle through.
# Adjust this upwards,
#+ if you have time on your hands.
NONE_ALIVE=80 # Exit status on premature bailout,
#+ if no cells left alive.
TRUE=0
FALSE=1
ALIVE=0
DEAD=1
avar= # Global; holds current generation.
generation=0 # Initialize generation count.
# =================================================================
let "cells = $ROWS * $COLS"
# How many cells.
declare -a initial # Arrays containing "cells".
declare -a current
display ()
{
alive=0 # How many cells "alive".
# Initially zero.
declare -a arr
arr=( `echo "$1"` ) # Convert passed arg to array.
element_count=${#arr[*]}
local i
local rowcheck
for ((i=0; i<$element_count; i++))
do
# Insert newline at end of each row.
let "rowcheck = $i % ROWS"
if [ "$rowcheck" -eq 0 ]
then
echo # Newline.
echo -n " " # Indent.
fi
cell=${arr[i]}
if [ "$cell" = . ]
then
let "alive += 1"
fi
echo -n "$cell" | sed -e 's/_/ /g'
# Print out array and change underscores to spaces.
done
return
}
IsValid () # Test whether cell coordinate valid.
{
if [ -z "$1" -o -z "$2" ] # Mandatory arguments missing?
then
return $FALSE
fi
local row
local lower_limit=0 # Disallow negative coordinate.
local upper_limit
local left
local right
let "upper_limit = $ROWS * $COLS - 1" # Total number of cells.
if [ "$1" -lt "$lower_limit" -o "$1" -gt "$upper_limit" ]
then
return $FALSE # Out of array bounds.
fi
row=$2
let "left = $row * $ROWS" # Left limit.
let "right = $left + $COLS - 1" # Right limit.
if [ "$1" -lt "$left" -o "$1" -gt "$right" ]
then
return $FALSE # Beyond row boundary.
fi
return $TRUE # Valid coordinate.
}
IsAlive () # Test whether cell is alive.
# Takes array, cell number, state of cell as arguments.
{
GetCount "$1" $2 # Get alive cell count in neighborhood.
local nhbd=$?
if [ "$nhbd" -eq "$BIRTH" ] # Alive in any case.
then
return $ALIVE
fi
if [ "$3" = "." -a "$nhbd" -eq "$SURVIVE" ]
then # Alive only if previously alive.
return $ALIVE
fi
return $DEAD # Default.
}
GetCount () # Count live cells in passed cell's neighborhood.
# Two arguments needed:
# $1) variable holding array
# $2) cell number
{
local cell_number=$2
local array
local top
local center
local bottom
local r
local row
local i
local t_top
local t_cen
local t_bot
local count=0
local ROW_NHBD=3
array=( `echo "$1"` )
let "top = $cell_number - $COLS - 1" # Set up cell neighborhood.
let "center = $cell_number - 1"
let "bottom = $cell_number + $COLS - 1"
let "r = $cell_number / $ROWS"
for ((i=0; i<$ROW_NHBD; i++)) # Traverse from left to right.
do
let "t_top = $top + $i"
let "t_cen = $center + $i"
let "t_bot = $bottom + $i"
let "row = $r" # Count center row of neighborhood.
IsValid $t_cen $row # Valid cell position?
if [ $? -eq "$TRUE" ]
then
if [ ${array[$t_cen]} = "$ALIVE1" ] # Is it alive?
then # Yes?
let "count += 1" # Increment count.
fi
fi
let "row = $r - 1" # Count top row.
IsValid $t_top $row
if [ $? -eq "$TRUE" ]
then
if [ ${array[$t_top]} = "$ALIVE1" ]
then
let "count += 1"
fi
fi
let "row = $r + 1" # Count bottom row.
IsValid $t_bot $row
if [ $? -eq "$TRUE" ]
then
if [ ${array[$t_bot]} = "$ALIVE1" ]
then
let "count += 1"
fi
fi
done
if [ ${array[$cell_number]} = "$ALIVE1" ]
then
let "count -= 1" # Make sure value of tested cell itself
fi #+ is not counted.
return $count
}
next_gen () # Update generation array.
{
local array
local i=0
array=( `echo "$1"` ) # Convert passed arg to array.
while [ "$i" -lt "$cells" ]
do
IsAlive "$1" $i ${array[$i]} # Is cell alive?
if [ $? -eq "$ALIVE" ]
then # If alive, then
array[$i]=. #+ represent the cell as a period.
else
array[$i]="_" # Otherwise underscore
fi #+ (which will later be converted to space).
let "i += 1"
done
# let "generation += 1" # Increment generation count.
# Set variable to pass as parameter to "display" function.
avar=`echo ${array[@]}` # Convert array back to string variable.
display "$avar" # Display it.
echo; echo
echo "Generation $generation -- $alive alive"
if [ "$alive" -eq 0 ]
then
echo
echo "Premature exit: no more cells alive!"
exit $NONE_ALIVE # No point in continuing
fi #+ if no live cells.
}
# =========================================================
# main ()
# Load initial array with contents of startup file.
initial=( `cat "$startfile" | sed -e '/#/d' | tr -d '\n' |\
sed -e 's/\./\. /g' -e 's/_/_ /g'` )
# Delete lines containing '#' comment character.
# Remove linefeeds and insert space between elements.
clear # Clear screen.
echo # Title
echo "======================="
echo " $GENERATIONS generations"
echo " of"
echo "\"Life in the Slow Lane\""
echo "======================="
# -------- Display first generation. --------
Gen0=`echo ${initial[@]}`
display "$Gen0" # Display only.
echo; echo
echo "Generation $generation -- $alive alive"
# -------------------------------------------
let "generation += 1" # Increment generation count.
echo
# ------- Display second generation. -------
Cur=`echo ${initial[@]}`
next_gen "$Cur" # Update & display.
# ------------------------------------------
let "generation += 1" # Increment generation count.
# ------ Main loop for displaying subsequent generations ------
while [ "$generation" -le "$GENERATIONS" ]
do
Cur="$avar"
next_gen "$Cur"
let "generation += 1"
done
# ==============================================================
echo
exit 0
# --------------------------------------------------------------
# The grid in this script has a "boundary problem".
# The the top, bottom, and sides border on a void of dead cells.
# Exercise: Change the script to have the grid wrap around,
# + so that the left and right sides will "touch",
# + as will the top and bottom. |
Example A-12. Data file for "Game of Life"
# This is an example "generation 0" start-up file for "life.sh". # -------------------------------------------------------------- # The "gen0" file is a 10 x 10 grid using a period (.) for live cells, #+ and an underscore (_) for dead ones. We cannot simply use spaces #+ for dead cells in this file because of a peculiarity in Bash arrays. # [Exercise for the reader: explain this.] # # Lines beginning with a '#' are comments, and the script ignores them. __.__..___ ___._.____ ____.___.. _._______. ____._____ ..__...___ ____._____ ___...____ __.._..___ _..___..__ |
+++
The following two scripts are by Mark Moraes of the University of Toronto. See the enclosed file "Moraes-COPYRIGHT" for permissions and restrictions.
Example A-13. behead: Removing mail and news message headers
#! /bin/sh # Strips off the header from a mail/News message i.e. till the first # empty line # Mark Moraes, University of Toronto # ==> These comments added by author of this document. if [ $# -eq 0 ]; then # ==> If no command line args present, then works on file redirected to stdin. sed -e '1,/^$/d' -e '/^[ ]*$/d' # --> Delete empty lines and all lines until # --> first one beginning with white space. else # ==> If command line args present, then work on files named. for i do sed -e '1,/^$/d' -e '/^[ ]*$/d' $i # --> Ditto, as above. done fi # ==> Exercise: Add error checking and other options. # ==> # ==> Note that the small sed script repeats, except for the arg passed. # ==> Does it make sense to embed it in a function? Why or why not? |
Example A-14. ftpget: Downloading files via ftp
#! /bin/sh
# $Id: ftpget,v 1.2 91/05/07 21:15:43 moraes Exp $
# Script to perform batch anonymous ftp. Essentially converts a list of
# of command line arguments into input to ftp.
# Simple, and quick - written as a companion to ftplist
# -h specifies the remote host (default prep.ai.mit.edu)
# -d specifies the remote directory to cd to - you can provide a sequence
# of -d options - they will be cd'ed to in turn. If the paths are relative,
# make sure you get the sequence right. Be careful with relative paths -
# there are far too many symlinks nowadays.
# (default is the ftp login directory)
# -v turns on the verbose option of ftp, and shows all responses from the
# ftp server.
# -f remotefile[:localfile] gets the remote file into localfile
# -m pattern does an mget with the specified pattern. Remember to quote
# shell characters.
# -c does a local cd to the specified directory
# For example,
# ftpget -h expo.lcs.mit.edu -d contrib -f xplaces.shar:xplaces.sh \
# -d ../pub/R3/fixes -c ~/fixes -m 'fix*'
# will get xplaces.shar from ~ftp/contrib on expo.lcs.mit.edu, and put it in
# xplaces.sh in the current working directory, and get all fixes from
# ~ftp/pub/R3/fixes and put them in the ~/fixes directory.
# Obviously, the sequence of the options is important, since the equivalent
# commands are executed by ftp in corresponding order
#
# Mark Moraes (moraes@csri.toronto.edu), Feb 1, 1989
# ==> Angle brackets changed to parens, so Docbook won't get indigestion.
#
# ==> These comments added by author of this document.
# PATH=/local/bin:/usr/ucb:/usr/bin:/bin
# export PATH
# ==> Above 2 lines from original script probably superfluous.
TMPFILE=/tmp/ftp.$$
# ==> Creates temp file, using process id of script ($$)
# ==> to construct filename.
SITE=`domainname`.toronto.edu
# ==> 'domainname' similar to 'hostname'
# ==> May rewrite this to parameterize this for general use.
usage="Usage: $0 [-h remotehost] [-d remotedirectory]... [-f remfile:localfile]... \
[-c localdirectory] [-m filepattern] [-v]"
ftpflags="-i -n"
verbflag=
set -f # So we can use globbing in -m
set x `getopt vh:d:c:m:f: $*`
if [ $? != 0 ]; then
echo $usage
exit 65
fi
shift
trap 'rm -f ${TMPFILE} ; exit' 0 1 2 3 15
echo "user anonymous ${USER-gnu}@${SITE} > ${TMPFILE}"
# ==> Added quotes (recommended in complex echoes).
echo binary >> ${TMPFILE}
for i in $* # ==> Parse command line args.
do
case $i in
-v) verbflag=-v; echo hash >> ${TMPFILE}; shift;;
-h) remhost=$2; shift 2;;
-d) echo cd $2 >> ${TMPFILE};
if [ x${verbflag} != x ]; then
echo pwd >> ${TMPFILE};
fi;
shift 2;;
-c) echo lcd $2 >> ${TMPFILE}; shift 2;;
-m) echo mget "$2" >> ${TMPFILE}; shift 2;;
-f) f1=`expr "$2" : "\([^:]*\).*"`; f2=`expr "$2" : "[^:]*:\(.*\)"`;
echo get ${f1} ${f2} >> ${TMPFILE}; shift 2;;
--) shift; break;;
esac
done
if [ $# -ne 0 ]; then
echo $usage
exit 65 # ==> Changed from "exit 2" to conform with standard.
fi
if [ x${verbflag} != x ]; then
ftpflags="${ftpflags} -v"
fi
if [ x${remhost} = x ]; then
remhost=prep.ai.mit.edu
# ==> Rewrite to match your favorite ftp site.
fi
echo quit >> ${TMPFILE}
# ==> All commands saved in tempfile.
ftp ${ftpflags} ${remhost} < ${TMPFILE}
# ==> Now, tempfile batch processed by ftp.
rm -f ${TMPFILE}
# ==> Finally, tempfile deleted (you may wish to copy it to a logfile).
# ==> Exercises:
# ==> ---------
# ==> 1) Add error checking.
# ==> 2) Add bells & whistles. |
+
Antek Sawicki contributed the following script, which makes very clever use of the parameter substitution operators discussed in Section 9.3.
Example A-15. password: Generating random 8-character passwords
#!/bin/bash
# May need to be invoked with #!/bin/bash2 on older machines.
#
# Random password generator for bash 2.x by Antek Sawicki <tenox@tenox.tc>,
# who generously gave permission to the document author to use it here.
#
# ==> Comments added by document author ==>
MATRIX="0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
LENGTH="8"
# ==> May change 'LENGTH' for longer password, of course.
while [ "${n:=1}" -le "$LENGTH" ]
# ==> Recall that := is "default substitution" operator.
# ==> So, if 'n' has not been initialized, set it to 1.
do
PASS="$PASS${MATRIX:$(($RANDOM%${#MATRIX})):1}"
# ==> Very clever, but tricky.
# ==> Starting from the innermost nesting...
# ==> ${#MATRIX} returns length of array MATRIX.
# ==> $RANDOM%${#MATRIX} returns random number between 1
# ==> and length of MATRIX - 1.
# ==> ${MATRIX:$(($RANDOM%${#MATRIX})):1}
# ==> returns expansion of MATRIX at random position, by length 1.
# ==> See {var:pos:len} parameter substitution in Section 3.3.1
# ==> and following examples.
# ==> PASS=... simply pastes this result onto previous PASS (concatenation).
# ==> To visualize this more clearly, uncomment the following line
# ==> echo "$PASS"
# ==> to see PASS being built up,
# ==> one character at a time, each iteration of the loop.
let n+=1
# ==> Increment 'n' for next pass.
done
echo "$PASS" # ==> Or, redirect to file, as desired.
exit 0 |
+
James R. Van Zandt contributed this script, which uses named pipes and, in his words, "really exercises quoting and escaping".
Example A-16. fifo: Making daily backups, using named pipes
#!/bin/bash
# ==> Script by James R. Van Zandt, and used here with his permission.
# ==> Comments added by author of this document.
HERE=`uname -n` # ==> hostname
THERE=bilbo
echo "starting remote backup to $THERE at `date +%r`"
# ==> `date +%r` returns time in 12-hour format, i.e. "08:08:34 PM".
# make sure /pipe really is a pipe and not a plain file
rm -rf /pipe
mkfifo /pipe # ==> Create a "named pipe", named "/pipe".
# ==> 'su xyz' runs commands as user "xyz".
# ==> 'ssh' invokes secure shell (remote login client).
su xyz -c "ssh $THERE \"cat >/home/xyz/backup/${HERE}-daily.tar.gz\" < /pipe"&
cd /
tar -czf - bin boot dev etc home info lib man root sbin share usr var >/pipe
# ==> Uses named pipe, /pipe, to communicate between processes:
# ==> 'tar/gzip' writes to /pipe and 'ssh' reads from /pipe.
# ==> The end result is this backs up the main directories, from / on down.
# ==> What are the advantages of a "named pipe" in this situation,
# ==> as opposed to an "anonymous pipe", with |?
# ==> Will an anonymous pipe even work here?
exit 0 |
+
Stephane Chazelas contributed the following script to demonstrate that generating prime numbers does not require arrays.
Example A-17. Generating prime numbers using the modulo operator
#!/bin/bash
# primes.sh: Generate prime numbers, without using arrays.
# Script contributed by Stephane Chazelas.
# This does *not* use the classic "Sieve of Eratosthenes" algorithm,
#+ but instead uses the more intuitive method of testing each candidate number
#+ for factors (divisors), using the "%" modulo operator.
LIMIT=1000 # Primes 2 - 1000
Primes()
{
(( n = $1 + 1 )) # Bump to next integer.
shift # Next parameter in list.
# echo "_n=$n i=$i_"
if (( n == LIMIT ))
then echo $*
return
fi
for i; do # "i" gets set to "@", previous values of $n.
# echo "-n=$n i=$i-"
(( i * i > n )) && break # Optimization.
(( n % i )) && continue # Sift out non-primes using modulo operator.
Primes $n $@ # Recursion inside loop.
return
done
Primes $n $@ $n # Recursion outside loop.
# Successively accumulate positional parameters.
# "$@" is the accumulating list of primes.
}
Primes 1
exit 0
# Uncomment lines 16 and 24 to help figure out what is going on.
# Compare the speed of this algorithm for generating primes
#+ with the Sieve of Eratosthenes (ex68.sh).
# Exercise: Rewrite this script without recursion, for faster execution. |
+
Jordi Sanfeliu gave permission to use his elegant tree script.
Example A-18. tree: Displaying a directory tree
#!/bin/sh
# @(#) tree 1.1 30/11/95 by Jordi Sanfeliu
# email: mikaku@fiwix.org
#
# Initial version: 1.0 30/11/95
# Next version : 1.1 24/02/97 Now, with symbolic links
# Patch by : Ian Kjos, to support unsearchable dirs
# email: beth13@mail.utexas.edu
#
# Tree is a tool for view the directory tree (obvious :-) )
#
# ==> 'Tree' script used here with the permission of its author, Jordi Sanfeliu.
# ==> Comments added by the author of this document.
# ==> Argument quoting added.
search () {
for dir in `echo *`
# ==> `echo *` lists all the files in current working directory, without line breaks.
# ==> Similar effect to for dir in *
# ==> but "dir in `echo *`" will not handle filenames with blanks.
do
if [ -d "$dir" ] ; then # ==> If it is a directory (-d)...
zz=0 # ==> Temp variable, keeping track of directory level.
while [ $zz != $deep ] # Keep track of inner nested loop.
do
echo -n "| " # ==> Display vertical connector symbol,
# ==> with 2 spaces & no line feed in order to indent.
zz=`expr $zz + 1` # ==> Increment zz.
done
if [ -L "$dir" ] ; then # ==> If directory is a symbolic link...
echo "+---$dir" `ls -l $dir | sed 's/^.*'$dir' //'`
# ==> Display horiz. connector and list directory name, but...
# ==> delete date/time part of long listing.
else
echo "+---$dir" # ==> Display horizontal connector symbol...
# ==> and print directory name.
if cd "$dir" ; then # ==> If can move to subdirectory...
deep=`expr $deep + 1` # ==> Increment depth.
search # with recursivity ;-)
# ==> Function calls itself.
numdirs=`expr $numdirs + 1` # ==> Increment directory count.
fi
fi
fi
done
cd .. # ==> Up one directory level.
if [ "$deep" ] ; then # ==> If depth = 0 (returns TRUE)...
swfi=1 # ==> set flag showing that search is done.
fi
deep=`expr $deep - 1` # ==> Decrement depth.
}
# - Main -
if [ $# = 0 ] ; then
cd `pwd` # ==> No args to script, then use current working directory.
else
cd $1 # ==> Otherwise, move to indicated directory.
fi
echo "Initial directory = `pwd`"
swfi=0 # ==> Search finished flag.
deep=0 # ==> Depth of listing.
numdirs=0
zz=0
while [ "$swfi" != 1 ] # While flag not set...
do
search # ==> Call function after initializing variables.
done
echo "Total directories = $numdirs"
exit 0
# ==> Challenge: try to figure out exactly how this script works. |
Noah Friedman gave permission to use his string function script, which essentially reproduces some of the C-library string manipulation functions.
Example A-19. string functions: C-like string functions
#!/bin/bash
# string.bash --- bash emulation of string(3) library routines
# Author: Noah Friedman <friedman@prep.ai.mit.edu>
# ==> Used with his kind permission in this document.
# Created: 1992-07-01
# Last modified: 1993-09-29
# Public domain
# Conversion to bash v2 syntax done by Chet Ramey
# Commentary:
# Code:
#:docstring strcat:
# Usage: strcat s1 s2
#
# Strcat appends the value of variable s2 to variable s1.
#
# Example:
# a="foo"
# b="bar"
# strcat a b
# echo $a
# => foobar
#
#:end docstring:
###;;;autoload ==> Autoloading of function commented out.
function strcat ()
{
local s1_val s2_val
s1_val=${!1} # indirect variable expansion
s2_val=${!2}
eval "$1"=\'"${s1_val}${s2_val}"\'
# ==> eval $1='${s1_val}${s2_val}' avoids problems,
# ==> if one of the variables contains a single quote.
}
#:docstring strncat:
# Usage: strncat s1 s2 $n
#
# Line strcat, but strncat appends a maximum of n characters from the value
# of variable s2. It copies fewer if the value of variabl s2 is shorter
# than n characters. Echoes result on stdout.
#
# Example:
# a=foo
# b=barbaz
# strncat a b 3
# echo $a
# => foobar
#
#:end docstring:
###;;;autoload
function strncat ()
{
local s1="$1"
local s2="$2"
local -i n="$3"
local s1_val s2_val
s1_val=${!s1} # ==> indirect variable expansion
s2_val=${!s2}
if [ ${#s2_val} -gt ${n} ]; then
s2_val=${s2_val:0:$n} # ==> substring extraction
fi
eval "$s1"=\'"${s1_val}${s2_val}"\'
# ==> eval $1='${s1_val}${s2_val}' avoids problems,
# ==> if one of the variables contains a single quote.
}
#:docstring strcmp:
# Usage: strcmp $s1 $s2
#
# Strcmp compares its arguments and returns an integer less than, equal to,
# or greater than zero, depending on whether string s1 is lexicographically
# less than, equal to, or greater than string s2.
#:end docstring:
###;;;autoload
function strcmp ()
{
[ "$1" = "$2" ] && return 0
[ "${1}" '<' "${2}" ] > /dev/null && return -1
return 1
}
#:docstring strncmp:
# Usage: strncmp $s1 $s2 $n
#
# Like strcmp, but makes the comparison by examining a maximum of n
# characters (n less than or equal to zero yields equality).
#:end docstring:
###;;;autoload
function strncmp ()
{
if [ -z "${3}" -o "${3}" -le "0" ]; then
return 0
fi
if [ ${3} -ge ${#1} -a ${3} -ge ${#2} ]; then
strcmp "$1" "$2"
return $?
else
s1=${1:0:$3}
s2=${2:0:$3}
strcmp $s1 $s2
return $?
fi
}
#:docstring strlen:
# Usage: strlen s
#
# Strlen returns the number of characters in string literal s.
#:end docstring:
###;;;autoload
function strlen ()
{
eval echo "\${#${1}}"
# ==> Returns the length of the value of the variable
# ==> whose name is passed as an argument.
}
#:docstring strspn:
# Usage: strspn $s1 $s2
#
# Strspn returns the length of the maximum initial segment of string s1,
# which consists entirely of characters from string s2.
#:end docstring:
###;;;autoload
function strspn ()
{
# Unsetting IFS allows whitespace to be handled as normal chars.
local IFS=
local result="${1%%[!${2}]*}"
echo ${#result}
}
#:docstring strcspn:
# Usage: strcspn $s1 $s2
#
# Strcspn returns the length of the maximum initial segment of string s1,
# which consists entirely of characters not from string s2.
#:end docstring:
###;;;autoload
function strcspn ()
{
# Unsetting IFS allows whitspace to be handled as normal chars.
local IFS=
local result="${1%%[${2}]*}"
echo ${#result}
}
#:docstring strstr:
# Usage: strstr s1 s2
#
# Strstr echoes a substring starting at the first occurrence of string s2 in
# string s1, or nothing if s2 does not occur in the string. If s2 points to
# a string of zero length, strstr echoes s1.
#:end docstring:
###;;;autoload
function strstr ()
{
# if s2 points to a string of zero length, strstr echoes s1
[ ${#2} -eq 0 ] && { echo "$1" ; return 0; }
# strstr echoes nothing if s2 does not occur in s1
case "$1" in
*$2*) ;;
*) return 1;;
esac
# use the pattern matching code to strip off the match and everything
# following it
first=${1/$2*/}
# then strip off the first unmatched portion of the string
echo "${1##$first}"
}
#:docstring strtok:
# Usage: strtok s1 s2
#
# Strtok considers the string s1 to consist of a sequence of zero or more
# text tokens separated by spans of one or more characters from the
# separator string s2. The first call (with a non-empty string s1
# specified) echoes a string consisting of the first token on stdout. The
# function keeps track of its position in the string s1 between separate
# calls, so that subsequent calls made with the first argument an empty
# string will work through the string immediately following that token. In
# this way subsequent calls will work through the string s1 until no tokens
# remain. The separator string s2 may be different from call to call.
# When no token remains in s1, an empty value is echoed on stdout.
#:end docstring:
###;;;autoload
function strtok ()
{
:
}
#:docstring strtrunc:
# Usage: strtrunc $n $s1 {$s2} {$...}
#
# Used by many functions like strncmp to truncate arguments for comparison.
# Echoes the first n characters of each string s1 s2 ... on stdout.
#:end docstring:
###;;;autoload
function strtrunc ()
{
n=$1 ; shift
for z; do
echo "${z:0:$n}"
done
}
# provide string
# string.bash ends here
# ========================================================================== #
# ==> Everything below here added by the document author.
# ==> Suggested use of this script is to delete everything below here,
# ==> and "source" this file into your own scripts.
# strcat
string0=one
string1=two
echo
echo "Testing \"strcat\" function:"
echo "Original \"string0\" = $string0"
echo "\"string1\" = $string1"
strcat string0 string1
echo "New \"string0\" = $string0"
echo
# strlen
echo
echo "Testing \"strlen\" function:"
str=123456789
echo "\"str\" = $str"
echo -n "Length of \"str\" = "
strlen str
echo
# Exercise:
# --------
# Add code to test all the other string functions above.
exit 0 |
Michael Zick's complex array example uses the md5sum check sum command to encode directory information.
Example A-20. Directory information
#! /bin/bash
# directory-info.sh
# Parses and lists directory information.
# NOTE: Change lines 273 and 353 per "README" file.
# Michael Zick is the author of this script.
# Used here with his permission.
# Controls
# If overridden by command arguments, they must be in the order:
# Arg1: "Descriptor Directory"
# Arg2: "Exclude Paths"
# Arg3: "Exclude Directories"
#
# Environment Settings override Defaults.
# Command arguments override Environment Settings.
# Default location for content addressed file descriptors.
MD5UCFS=${1:-${MD5UCFS:-'/tmpfs/ucfs'}}
# Directory paths never to list or enter
declare -a \
EXCLUDE_PATHS=${2:-${EXCLUDE_PATHS:-'(/proc /dev /devfs /tmpfs)'}}
# Directories never to list or enter
declare -a \
EXCLUDE_DIRS=${3:-${EXCLUDE_DIRS:-'(ucfs lost+found tmp wtmp)'}}
# Files never to list or enter
declare -a \
EXCLUDE_FILES=${3:-${EXCLUDE_FILES:-'(core "Name with Spaces")'}}
# Here document used as a comment block.
: << LSfieldsDoc
# # # # # List Filesystem Directory Information # # # # #
#
# ListDirectory "FileGlob" "Field-Array-Name"
# or
# ListDirectory -of "FileGlob" "Field-Array-Filename"
# '-of' meaning 'output to filename'
# # # # #
String format description based on: ls (GNU fileutils) version 4.0.36
Produces a line (or more) formatted:
inode permissions hard-links owner group ...
32736 -rw------- 1 mszick mszick
size day month date hh:mm:ss year path
2756608 Sun Apr 20 08:53:06 2003 /home/mszick/core
Unless it is formatted:
inode permissions hard-links owner group ...
266705 crw-rw---- 1 root uucp
major minor day month date hh:mm:ss year path
4, 68 Sun Apr 20 09:27:33 2003 /dev/ttyS4
NOTE: that pesky comma after the major number
NOTE: the 'path' may be multiple fields:
/home/mszick/core
/proc/982/fd/0 -> /dev/null
/proc/982/fd/1 -> /home/mszick/.xsession-errors
/proc/982/fd/13 -> /tmp/tmpfZVVOCs (deleted)
/proc/982/fd/7 -> /tmp/kde-mszick/ksycoca
/proc/982/fd/8 -> socket:[11586]
/proc/982/fd/9 -> pipe:[11588]
If that isn't enough to keep your parser guessing,
either or both of the path components may be relative:
../Built-Shared -> Built-Static
../linux-2.4.20.tar.bz2 -> ../../../SRCS/linux-2.4.20.tar.bz2
The first character of the 11 (10?) character permissions field:
's' Socket
'd' Directory
'b' Block device
'c' Character device
'l' Symbolic link
NOTE: Hard links not marked - test for identical inode numbers
on identical filesystems.
All information about hard linked files are shared, except
for the names and the name's location in the directory system.
NOTE: A "Hard link" is known as a "File Alias" on some systems.
'-' An undistingushed file
Followed by three groups of letters for: User, Group, Others
Character 1: '-' Not readable; 'r' Readable
Character 2: '-' Not writable; 'w' Writable
Character 3, User and Group: Combined execute and special
'-' Not Executable, Not Special
'x' Executable, Not Special
's' Executable, Special
'S' Not Executable, Special
Character 3, Others: Combined execute and sticky (tacky?)
'-' Not Executable, Not Tacky
'x' Executable, Not Tacky
't' Executable, Tacky
'T' Not Executable, Tacky
Followed by an access indicator
Haven't tested this one, it may be the eleventh character
or it may generate another field
' ' No alternate access
'+' Alternate access
LSfieldsDoc
ListDirectory()
{
local -a T
local -i of=0 # Default return in variable
# OLD_IFS=$IFS # Using BASH default ' \t\n'
case "$#" in
3) case "$1" in
-of) of=1 ; shift ;;
* ) return 1 ;;
esac ;;
2) : ;; # Poor man's "continue"
*) return 1 ;;
esac
# NOTE: the (ls) command is NOT quoted (")
T=( $(ls --inode --ignore-backups --almost-all --directory \
--full-time --color=none --time=status --sort=none \
--format=long $1) )
case $of in
# Assign T back to the array whose name was passed as $2
0) eval $2=\( \"\$\{T\[@\]\}\" \) ;;
# Write T into filename passed as $2
1) echo "${T[@]}" > "$2" ;;
esac
return 0
}
# # # # # Is that string a legal number? # # # # #
#
# IsNumber "Var"
# # # # # There has to be a better way, sigh...
IsNumber()
{
local -i int
if [ $# -eq 0 ]
then
return 1
else
(let int=$1) 2>/dev/null
return $? # Exit status of the let thread
fi
}
# # # # # Index Filesystem Directory Information # # # # #
#
# IndexList "Field-Array-Name" "Index-Array-Name"
# or
# IndexList -if Field-Array-Filename Index-Array-Name
# IndexList -of Field-Array-Name Index-Array-Filename
# IndexList -if -of Field-Array-Filename Index-Array-Filename
# # # # #
: << IndexListDoc
Walk an array of directory fields produced by ListDirectory
Having suppressed the line breaks in an otherwise line oriented
report, build an index to the array element which starts each line.
Each line gets two index entries, the first element of each line
(inode) and the element that holds the pathname of the file.
The first index entry pair (Line-Number==0) are informational:
Index-Array-Name[0] : Number of "Lines" indexed
Index-Array-Name[1] : "Current Line" pointer into Index-Array-Name
The following index pairs (if any) hold element indexes into
the Field-Array-Name per:
Index-Array-Name[Line-Number * 2] : The "inode" field element.
NOTE: This distance may be either +11 or +12 elements.
Index-Array-Name[(Line-Number * 2) + 1] : The "pathname" element.
NOTE: This distance may be a variable number of elements.
Next line index pair for Line-Number+1.
IndexListDoc
IndexList()
{
local -a LIST # Local of listname passed
local -a -i INDEX=( 0 0 ) # Local of index to return
local -i Lidx Lcnt
local -i if=0 of=0 # Default to variable names
case "$#" in # Simplistic option testing
0) return 1 ;;
1) return 1 ;;
2) : ;; # Poor man's continue
3) case "$1" in
-if) if=1 ;;
-of) of=1 ;;
* ) return 1 ;;
esac ; shift ;;
4) if=1 ; of=1 ; shift ; shift ;;
*) return 1
esac
# Make local copy of list
case "$if" in
0) eval LIST=\( \"\$\{$1\[@\]\}\" \) ;;
1) LIST=( $(cat $1) ) ;;
esac
# Grok (grope?) the array
Lcnt=${#LIST[@]}
Lidx=0
until (( Lidx >= Lcnt ))
do
if IsNumber ${LIST[$Lidx]}
then
local -i inode name
local ft
inode=Lidx
local m=${LIST[$Lidx+2]} # Hard Links field
ft=${LIST[$Lidx+1]:0:1} # Fast-Stat
case $ft in
b) ((Lidx+=12)) ;; # Block device
c) ((Lidx+=12)) ;; # Character device
*) ((Lidx+=11)) ;; # Anything else
esac
name=Lidx
case $ft in
-) ((Lidx+=1)) ;; # The easy one
b) ((Lidx+=1)) ;; # Block device
c) ((Lidx+=1)) ;; # Character device
d) ((Lidx+=1)) ;; # The other easy one
l) ((Lidx+=3)) ;; # At LEAST two more fields
# A little more elegance here would handle pipes,
#+ sockets, deleted files - later.
*) until IsNumber ${LIST[$Lidx]} || ((Lidx >= Lcnt))
do
((Lidx+=1))
done
;; # Not required
esac
INDEX[${#INDEX[*]}]=$inode
INDEX[${#INDEX[*]}]=$name
INDEX[0]=${INDEX[0]}+1 # One more "line" found
# echo "Line: ${INDEX[0]} Type: $ft Links: $m Inode: \
# ${LIST[$inode]} Name: ${LIST[$name]}"
else
((Lidx+=1))
fi
done
case "$of" in
0) eval $2=\( \"\$\{INDEX\[@\]\}\" \) ;;
1) echo "${INDEX[@]}" > "$2" ;;
esac
return 0 # What could go wrong?
}
# # # # # Content Identify File # # # # #
#
# DigestFile Input-Array-Name Digest-Array-Name
# or
# DigestFile -if Input-FileName Digest-Array-Name
# # # # #
# Here document used as a comment block.
: <<DigestFilesDoc
The key (no pun intended) to a Unified Content File System (UCFS)
is to distinguish the files in the system based on their content.
Distinguishing files by their name is just, so, 20th Century.
The content is distinguished by computing a checksum of that content.
This version uses the md5sum program to generate a 128 bit checksum
representative of the file's contents.
There is a chance that two files having different content might
generate the same checksum using md5sum (or any checksum). Should
that become a problem, then the use of md5sum can be replace by a
cyrptographic signature. But until then...
The md5sum program is documented as outputting three fields (and it
does), but when read it appears as two fields (array elements). This
is caused by the lack of whitespace between the second and third field.
So this function gropes the md5sum output and returns:
[0] 32 character checksum in hexidecimal (UCFS filename)
[1] Single character: ' ' text file, '*' binary file
[2] Filesystem (20th Century Style) name
Note: That name may be the character '-' indicating STDIN read.
DigestFilesDoc
DigestFile()
{
local if=0 # Default, variable name
local -a T1 T2
case "$#" in
3) case "$1" in
-if) if=1 ; shift ;;
* ) return 1 ;;
esac ;;
2) : ;; # Poor man's "continue"
*) return 1 ;;
esac
case $if in
0) eval T1=\( \"\$\{$1\[@\]\}\" \)
T2=( $(echo ${T1[@]} | md5sum -) )
;;
1) T2=( $(md5sum $1) )
;;
esac
case ${#T2[@]} in
0) return 1 ;;
1) return 1 ;;
2) case ${T2[1]:0:1} in # SanScrit-2.0.5
\*) T2[${#T2[@]}]=${T2[1]:1}
T2[1]=\*
;;
*) T2[${#T2[@]}]=${T2[1]}
T2[1]=" "
;;
esac
;;
3) : ;; # Assume it worked
*) return 1 ;;
esac
local -i len=${#T2[0]}
if [ $len -ne 32 ] ; then return 1 ; fi
eval $2=\( \"\$\{T2\[@\]\}\" \)
}
# # # # # Locate File # # # # #
#
# LocateFile [-l] FileName Location-Array-Name
# or
# LocateFile [-l] -of FileName Location-Array-FileName
# # # # #
# A file location is Filesystem-id and inode-number
# Here document used as a comment block.
: <<StatFieldsDoc
Based on stat, version 2.2
stat -t and stat -lt fields
[0] name
[1] Total size
File - number of bytes
Symbolic link - string length of pathname
[2] Number of (512 byte) blocks allocated
[3] File type and Access rights (hex)
[4] User ID of owner
[5] Group ID of owner
[6] Device number
[7] Inode number
[8] Number of hard links
[9] Device type (if inode device) Major
[10] Device type (if inode device) Minor
[11] Time of last access
May be disabled in 'mount' with noatime
atime of files changed by exec, read, pipe, utime, mknod (mmap?)
atime of directories changed by addition/deletion of files
[12] Time of last modification
mtime of files changed by write, truncate, utime, mknod
mtime of directories changed by addtition/deletion of files
[13] Time of last change
ctime reflects time of changed inode information (owner, group
permissions, link count
-*-*- Per:
Return code: 0
Size of array: 14
Contents of array
Element 0: /home/mszick
Element 1: 4096
Element 2: 8
Element 3: 41e8
Element 4: 500
Element 5: 500
Element 6: 303
Element 7: 32385
Element 8: 22
Element 9: 0
Element 10: 0
Element 11: 1051221030
Element 12: 1051214068
Element 13: 1051214068
For a link in the form of linkname -> realname
stat -t linkname returns the linkname (link) information
stat -lt linkname returns the realname information
stat -tf and stat -ltf fields
[0] name
[1] ID-0? # Maybe someday, but Linux stat structure
[2] ID-0? # does not have either LABEL nor UUID
# fields, currently information must come
# from file-system specific utilities
These will be munged into:
[1] UUID if possible
[2] Volume Label if possible
Note: 'mount -l' does return the label and could return the UUID
[3] Maximum length of filenames
[4] Filesystem type
[5] Total blocks in the filesystem
[6] Free blocks
[7] Free blocks for non-root user(s)
[8] Block size of the filesystem
[9] Total inodes
[10] Free inodes
-*-*- Per:
Return code: 0
Size of array: 11
Contents of array
Element 0: /home/mszick
Element 1: 0
Element 2: 0
Element 3: 255
Element 4: ef53
Element 5: 2581445
Element 6: 2277180
Element 7: 2146050
Element 8: 4096
Element 9: 1311552
Element 10: 1276425
StatFieldsDoc
# LocateFile [-l] FileName Location-Array-Name
# LocateFile [-l] -of FileName Location-Array-FileName
LocateFile()
{
local -a LOC LOC1 LOC2
local lk="" of=0
case "$#" in
0) return 1 ;;
1) return 1 ;;
2) : ;;
*) while (( "$#" > 2 ))
do
case "$1" in
-l) lk=-1 ;;
-of) of=1 ;;
*) return 1 ;;
esac
shift
done ;;
esac
# More Sanscrit-2.0.5
# LOC1=( $(stat -t $lk $1) )
# LOC2=( $(stat -tf $lk $1) )
# Uncomment above two lines if system has "stat" command installed.
LOC=( ${LOC1[@]:0:1} ${LOC1[@]:3:11}
${LOC2[@]:1:2} ${LOC2[@]:4:1} )
case "$of" in
0) eval $2=\( \"\$\{LOC\[@\]\}\" \) ;;
1) echo "${LOC[@]}" > "$2" ;;
esac
return 0
# Which yields (if you are lucky, and have "stat" installed)
# -*-*- Location Discriptor -*-*-
# Return code: 0
# Size of array: 15
# Contents of array
# Element 0: /home/mszick 20th Century name
# Element 1: 41e8 Type and Permissions
# Element 2: 500 User
# Element 3: 500 Group
# Element 4: 303 Device
# Element 5: 32385 inode
# Element 6: 22 Link count
# Element 7: 0 Device Major
# Element 8: 0 Device Minor
# Element 9: 1051224608 Last Access
# Element 10: 1051214068 Last Modify
# Element 11: 1051214068 Last Status
# Element 12: 0 UUID (to be)
# Element 13: 0 Volume Label (to be)
# Element 14: ef53 Filesystem type
}
# And then there was some test code
ListArray() # ListArray Name
{
local -a Ta
eval Ta=\( \"\$\{$1\[@\]\}\" \)
echo
echo "-*-*- List of Array -*-*-"
echo "Size of array $1: ${#Ta[*]}"
echo "Contents of array $1:"
for (( i=0 ; i<${#Ta[*]} ; i++ ))
do
echo -e "\tElement $i: ${Ta[$i]}"
done
return 0
}
declare -a CUR_DIR
# For small arrays
ListDirectory "${PWD}" CUR_DIR
ListArray CUR_DIR
declare -a DIR_DIG
DigestFile CUR_DIR DIR_DIG
echo "The new \"name\" (checksum) for ${CUR_DIR[9]} is ${DIR_DIG[0]}"
declare -a DIR_ENT
# BIG_DIR # For really big arrays - use a temporary file in ramdisk
# BIG-DIR # ListDirectory -of "${CUR_DIR[11]}/*" "/tmpfs/junk2"
ListDirectory "${CUR_DIR[11]}/*" DIR_ENT
declare -a DIR_IDX
# BIG-DIR # IndexList -if "/tmpfs/junk2" DIR_IDX
IndexList DIR_ENT DIR_IDX
declare -a IDX_DIG
# BIG-DIR # DIR_ENT=( $(cat /tmpfs/junk2) )
# BIG-DIR # DigestFile -if /tmpfs/junk2 IDX_DIG
DigestFile DIR_ENT IDX_DIG
# Small (should) be able to parallize IndexList & DigestFile
# Large (should) be able to parallize IndexList & DigestFile & the assignment
echo "The \"name\" (checksum) for the contents of ${PWD} is ${IDX_DIG[0]}"
declare -a FILE_LOC
LocateFile ${PWD} FILE_LOC
ListArray FILE_LOC
exit 0 |
Stephane Chazelas demonstrates object-oriented programming in a Bash script.
Example A-21. Object-oriented database
#!/bin/bash
# obj-oriented.sh: Object-oriented programming in a shell script.
# Script by Stephane Chazelas.
person.new() # Looks almost like a class declaration in C++.
{
local obj_name=$1 name=$2 firstname=$3 birthdate=$4
eval "$obj_name.set_name() {
eval \"$obj_name.get_name() {
echo \$1
}\"
}"
eval "$obj_name.set_firstname() {
eval \"$obj_name.get_firstname() {
echo \$1
}\"
}"
eval "$obj_name.set_birthdate() {
eval \"$obj_name.get_birthdate() {
echo \$1
}\"
eval \"$obj_name.show_birthdate() {
echo \$(date -d \"1/1/1970 0:0:\$1 GMT\")
}\"
eval \"$obj_name.get_age() {
echo \$(( (\$(date +%s) - \$1) / 3600 / 24 / 365 ))
}\"
}"
$obj_name.set_name $name
$obj_name.set_firstname $firstname
$obj_name.set_birthdate $birthdate
}
echo
person.new self Bozeman Bozo 101272413
# Create an instance of "person.new" (actually passing args to the function).
self.get_firstname # Bozo
self.get_name # Bozeman
self.get_age # 28
self.get_birthdate # 101272413
self.show_birthdate # Sat Mar 17 20:13:33 MST 1973
echo
# typeset -f
# to see the created functions (careful, it scrolls off the page).
exit 0 |
Now for a script that does something useful: installing and mounting those cute USB keychain solid-state "hard drives."
Example A-22. Mounting USB keychain storage devices
#!/bin/bash
# ==> usb.sh
# ==> Script for mounting and installing pen/keychain USB storage devices.
# ==> Runs as root at system startup (see below).
# This code is free software covered by GNU GPL license version 2 or above.
# Please refer to http://www.gnu.org/ for the full license text.
#
# Some code lifted from usb-mount by Michael Hamilton's usb-mount (LGPL)
#+ see http://users.actrix.co.nz/michael/usbmount.html
#
# INSTALL
# -------
# Put this in /etc/hotplug/usb/diskonkey.
# Then look in /etc/hotplug/usb.distmap, and copy all usb-storage entries
#+ into /etc/hotplug/usb.usermap, substituting "usb-storage" for "diskonkey".
# Otherwise this code is only run during the kernel module invocation/removal
#+ (at least in my tests), which defeats the purpose.
#
# TODO
# ----
# Handle more than one diskonkey device at one time (e.g. /dev/diskonkey1
#+ and /mnt/diskonkey1), etc. The biggest problem here is the handling in
#+ devlabel, which I haven't yet tried.
#
# AUTHOR and SUPPORT
# ------------------
# Konstantin Riabitsev, <icon linux duke edu>.
# Send any problem reports to my email address at the moment.
#
# ==> Comments added by ABS Guide author.
SYMLINKDEV=/dev/diskonkey
MOUNTPOINT=/mnt/diskonkey
DEVLABEL=/sbin/devlabel
DEVLABELCONFIG=/etc/sysconfig/devlabel
IAM=$0
##
# Functions lifted near-verbatim from usb-mount code.
#
function allAttachedScsiUsb {
find /proc/scsi/ -path '/proc/scsi/usb-storage*' -type f | xargs grep -l 'Attached: Yes'
}
function scsiDevFromScsiUsb {
echo $1 | awk -F"[-/]" '{ n=$(NF-1); print "/dev/sd" substr("abcdefghijklmnopqrstuvwxyz", n+1,
1) }'
}
if [ "${ACTION}" = "add" ] && [ -f "${DEVICE}" ]; then
##
# lifted from usbcam code.
#
if [ -f /var/run/console.lock ]; then
CONSOLEOWNER=`cat /var/run/console.lock`
elif [ -f /var/lock/console.lock ]; then
CONSOLEOWNER=`cat /var/lock/console.lock`
else
CONSOLEOWNER=
fi
for procEntry in $(allAttachedScsiUsb); do
scsiDev=$(scsiDevFromScsiUsb $procEntry)
# Some bug with usb-storage?
# Partitions are not in /proc/partitions until they are accessed
#+ somehow.
/sbin/fdisk -l $scsiDev >/dev/null
##
# Most devices have partitioning info, so the data would be on
#+ /dev/sd?1. However, some stupider ones don't have any partitioning
#+ and use the entire device for data storage. This tries to
#+ guess semi-intelligently if we have a /dev/sd?1 and if not, then
#+ it uses the entire device and hopes for the better.
#
if grep -q `basename $scsiDev`1 /proc/partitions; then
part="$scsiDev""1"
else
part=$scsiDev
fi
##
# Change ownership of the partition to the console user so they can
#+ mount it.
#
if [ ! -z "$CONSOLEOWNER" ]; then
chown $CONSOLEOWNER:disk $part
fi
##
# This checks if we already have this UUID defined with devlabel.
# If not, it then adds the device to the list.
#
prodid=`$DEVLABEL printid -d $part`
if ! grep -q $prodid $DEVLABELCONFIG; then
# cross our fingers and hope it works
$DEVLABEL add -d $part -s $SYMLINKDEV 2>/dev/null
fi
##
# Check if the mount point exists and create if it doesn't.
#
if [ ! -e $MOUNTPOINT ]; then
mkdir -p $MOUNTPOINT
fi
##
# Take care of /etc/fstab so mounting is easy.
#
if ! grep -q "^$SYMLINKDEV" /etc/fstab; then
# Add an fstab entry
echo -e \
"$SYMLINKDEV\t\t$MOUNTPOINT\t\tauto\tnoauto,owner,kudzu 0 0" \
>> /etc/fstab
fi
done
if [ ! -z "$REMOVER" ]; then
##
# Make sure this script is triggered on device removal.
#
mkdir -p `dirname $REMOVER`
ln -s $IAM $REMOVER
fi
elif [ "${ACTION}" = "remove" ]; then
##
# If the device is mounted, unmount it cleanly.
#
if grep -q "$MOUNTPOINT" /etc/mtab; then
# unmount cleanly
umount -l $MOUNTPOINT
fi
##
# Remove it from /etc/fstab if it's there.
#
if grep -q "^$SYMLINKDEV" /etc/fstab; then
grep -v "^$SYMLINKDEV" /etc/fstab > /etc/.fstab.new
mv -f /etc/.fstab.new /etc/fstab
fi
fi
exit 0 |
Here is something to warm the hearts of webmasters and mistresses everywhere: a script that saves weblogs.
Example A-23. Preserving weblogs
#!/bin/bash
# archiveweblogs.sh v1.0
# Troy Engel <tengel@fluid.com>
# Slightly modified by document author.
# Used with permission.
#
# This script will preserve the normally rotated and
#+ thrown away weblogs from a default RedHat/Apache installation.
# It will save the files with a date/time stamp in the filename,
#+ bzipped, to a given directory.
#
# Run this from crontab nightly at an off hour,
#+ as bzip2 can suck up some serious CPU on huge logs:
# 0 2 * * * /opt/sbin/archiveweblogs.sh
PROBLEM=66
# Set this to your backup dir.
BKP_DIR=/opt/backups/weblogs
# Default Apache/RedHat stuff
LOG_DAYS="4 3 2 1"
LOG_DIR=/var/log/httpd
LOG_FILES="access_log error_log"
# Default RedHat program locations
LS=/bin/ls
MV=/bin/mv
ID=/usr/bin/id
CUT=/bin/cut
COL=/usr/bin/column
BZ2=/usr/bin/bzip2
# Are we root?
USER=`$ID -u`
if [ "X$USER" != "X0" ]; then
echo "PANIC: Only root can run this script!"
exit $PROBLEM
fi
# Backup dir exists/writable?
if [ ! -x $BKP_DIR ]; then
echo "PANIC: $BKP_DIR doesn't exist or isn't writable!"
exit $PROBLEM
fi
# Move, rename and bzip2 the logs
for logday in $LOG_DAYS; do
for logfile in $LOG_FILES; do
MYFILE="$LOG_DIR/$logfile.$logday"
if [ -w $MYFILE ]; then
DTS=`$LS -lgo --time-style=+%Y%m%d $MYFILE | $COL -t | $CUT -d ' ' -f7`
$MV $MYFILE $BKP_DIR/$logfile.$DTS
$BZ2 $BKP_DIR/$logfile.$DTS
else
# Only spew an error if the file exits (ergo non-writable).
if [ -f $MYFILE ]; then
echo "ERROR: $MYFILE not writable. Skipping."
fi
fi
done
done
exit 0 |
How do you keep the shell from expanding and reinterpreting strings?
Example A-24. Protecting literal strings
#! /bin/bash
# protect_literal.sh
# set -vx
:<<-'_Protect_Literal_String_Doc'
Copyright (c) Michael S. Zick, 2003; All Rights Reserved
License: Unrestricted reuse in any form, for any purpose.
Warranty: None
Revision: $ID$
Documentation redirected to the Bash no-operation.
Bash will '/dev/null' this block when the script is first read.
(Uncomment the above set command to see this action.)
Remove the first (Sha-Bang) line when sourcing this as a library
procedure. Also comment out the example use code in the two
places where shown.
Usage:
_protect_literal_str 'Whatever string meets your ${fancy}'
Just echos the argument to standard out, hard quotes
restored.
$(_protect_literal_str 'Whatever string meets your ${fancy}')
as the right-hand-side of an assignment statement.
Does:
As the right-hand-side of an assignment, preserves the
hard quotes protecting the contents of the literal during
assignment.
Notes:
The strange names (_*) are used to avoid trampling on
the user's chosen names when this is sourced as a
library.
_Protect_Literal_String_Doc
# The 'for illustration' function form
_protect_literal_str() {
# Pick an un-used, non-printing character as local IFS.
# Not required, but shows that we are ignoring it.
local IFS=$'\x1B' # \ESC character
# Enclose the All-Elements-Of in hard quotes during assignment.
local tmp=$'\x27'$@$'\x27'
# local tmp=$'\''$@$'\'' # Even uglier.
local len=${#tmp} # Info only.
echo $tmp is $len long. # Output AND information.
}
# This is the short-named version.
_pls() {
local IFS=$'x1B' # \ESC character (not required)
echo $'\x27'$@$'\x27' # Hard quoted parameter glob
}
# :<<-'_Protect_Literal_String_Test'
# # # Remove the above "# " to disable this code. # # #
# See how that looks when printed.
echo
echo "- - Test One - -"
_protect_literal_str 'Hello $user'
_protect_literal_str 'Hello "${username}"'
echo
# Which yields:
# - - Test One - -
# 'Hello $user' is 13 long.
# 'Hello "${username}"' is 21 long.
# Looks as expected, but why all of the trouble?
# The difference is hidden inside the Bash internal order
#+ of operations.
# Which shows when you use it on the RHS of an assignment.
# Declare an array for test values.
declare -a arrayZ
# Assign elements with various types of quotes and escapes.
arrayZ=( zero "$(_pls 'Hello ${Me}')" 'Hello ${You}' "\'Pass: ${pw}\'" )
# Now list that array and see what is there.
echo "- - Test Two - -"
for (( i=0 ; i<${#arrayZ[*]} ; i++ ))
do
echo Element $i: ${arrayZ[$i]} is: ${#arrayZ[$i]} long.
done
echo
# Which yields:
# - - Test Two - -
# Element 0: zero is: 4 long. # Our marker element
# Element 1: 'Hello ${Me}' is: 13 long. # Our "$(_pls '...' )"
# Element 2: Hello ${You} is: 12 long. # Quotes are missing
# Element 3: \'Pass: \' is: 10 long. # ${pw} expanded to nothing
# Now make an assignment with that result.
declare -a array2=( ${arrayZ[@]} )
# And print what happened.
echo "- - Test Three - -"
for (( i=0 ; i<${#array2[*]} ; i++ ))
do
echo Element $i: ${array2[$i]} is: ${#array2[$i]} long.
done
echo
# Which yields:
# - - Test Three - -
# Element 0: zero is: 4 long. # Our marker element.
# Element 1: Hello ${Me} is: 11 long. # Intended result.
# Element 2: Hello is: 5 long. # ${You} expanded to nothing.
# Element 3: 'Pass: is: 6 long. # Split on the whitespace.
# Element 4: ' is: 1 long. # The end quote is here now.
# Our Element 1 has had its leading and trailing hard quotes stripped.
# Although not shown, leading and trailing whitespace is also stripped.
# Now that the string contents are set, Bash will always, internally,
#+ hard quote the contents as required during its operations.
# Why?
# Considering our "$(_pls 'Hello ${Me}')" construction:
# " ... " -> Expansion required, strip the quotes.
# $( ... ) -> Replace with the result of..., strip this.
# _pls ' ... ' -> called with literal arguments, strip the quotes.
# The result returned includes hard quotes; BUT the above processing
#+ has already been done, so they become part of the value assigned.
#
# Similarly, during further usage of the string variable, the ${Me}
#+ is part of the contents (result) and survives any operations
# (Until explicitly told to evaluate the string).
# Hint: See what happens when the hard quotes ($'\x27') are replaced
#+ with soft quotes ($'\x22') in the above procedures.
# Interesting also is to remove the addition of any quoting.
# _Protect_Literal_String_Test
# # # Remove the above "# " to disable this code. # # #
exit 0 |
What if you want the shell to expand and reinterpret strings?
Example A-25. Unprotecting literal strings
#! /bin/bash
# unprotect_literal.sh
# set -vx
:<<-'_UnProtect_Literal_String_Doc'
Copyright (c) Michael S. Zick, 2003; All Rights Reserved
License: Unrestricted reuse in any form, for any purpose.
Warranty: None
Revision: $ID$
Documentation redirected to the Bash no-operation. Bash will
'/dev/null' this block when the script is first read.
(Uncomment the above set command to see this action.)
Remove the first (Sha-Bang) line when sourcing this as a library
procedure. Also comment out the example use code in the two
places where shown.
Usage:
Complement of the "$(_pls 'Literal String')" function.
(See the protect_literal.sh example.)
StringVar=$(_upls ProtectedSringVariable)
Does:
When used on the right-hand-side of an assignment statement;
makes the substitions embedded in the protected string.
Notes:
The strange names (_*) are used to avoid trampling on
the user's chosen names when this is sourced as a
library.
_UnProtect_Literal_String_Doc
_upls() {
local IFS=$'x1B' # \ESC character (not required)
eval echo $@ # Substitution on the glob.
}
# :<<-'_UnProtect_Literal_String_Test'
# # # Remove the above "# " to disable this code. # # #
_pls() {
local IFS=$'x1B' # \ESC character (not required)
echo $'\x27'$@$'\x27' # Hard quoted parameter glob
}
# Declare an array for test values.
declare -a arrayZ
# Assign elements with various types of quotes and escapes.
arrayZ=( zero "$(_pls 'Hello ${Me}')" 'Hello ${You}' "\'Pass: ${pw}\'" )
# Now make an assignment with that result.
declare -a array2=( ${arrayZ[@]} )
# Which yielded:
# - - Test Three - -
# Element 0: zero is: 4 long # Our marker element.
# Element 1: Hello ${Me} is: 11 long # Intended result.
# Element 2: Hello is: 5 long # ${You} expanded to nothing.
# Element 3: 'Pass: is: 6 long # Split on the whitespace.
# Element 4: ' is: 1 long # The end quote is here now.
# set -vx
# Initialize 'Me' to something for the embedded ${Me} substitution.
# This needs to be done ONLY just prior to evaluating the
#+ protected string.
# (This is why it was protected to begin with.)
Me="to the array guy."
# Set a string variable destination to the result.
newVar=$(_upls ${array2[1]})
# Show what the contents are.
echo $newVar
# Do we really need a function to do this?
newerVar=$(eval echo ${array2[1]})
echo $newerVar
# I guess not, but the _upls function gives us a place to hang
#+ the documentation on.
# This helps when we forget what a # construction like:
#+ $(eval echo ... ) means.
# What if Me isn't set when the protected string is evaluated?
unset Me
newestVar=$(_upls ${array2[1]})
echo $newestVar
# Just gone, no hints, no runs, no errors.
# Why in the world?
# Setting the contents of a string variable containing character
#+ sequences that have a meaning to Bash is a general problem in
#+ script programming.
#
# This problem is now solved in eight lines of code
#+ (and four pages of description).
# Where is all this going?
# Dynamic content Web pages as an array of Bash strings.
# Content set per request by a Bash 'eval' command
#+ on the stored page template.
# Not intended to replace PHP, just an interesting thing to do.
###
# Don't have a webserver application?
# No problem, check the example directory of the Bash source;
#+ there is a Bash script for that also.
# _UnProtect_Literal_String_Test
# # # Remove the above "# " to disable this code. # # #
exit 0 |
To end this section, a review of the basics . . . and more.
Example A-26. Basics Reviewed
#!/bin/bash
# basics-reviewed.bash
# File extension == *.bash == specific to Bash
# Copyright (c) Michael S. Zick, 2003; All rights reserved.
# License: Use in any form, for any purpose.
# Revision: $ID$
#
# Edited for layout by M.C.
# (author of the "Advanced Bash Scripting Guide")
# This script tested under Bash versions 2.04, 2.05a and 2.05b.
# It may not work with earlier versions.
# This demonstration script generates one --intentional--
#+ "command not found" error message. See line 394.
# The current Bash maintainer, Chet Ramey, has fixed the items noted
#+ for an upcoming version of Bash.
###-------------------------------------------###
### Pipe the output of this script to 'more' ###
###+ else it will scroll off the page. ###
### ###
### You may also redirect its output ###
###+ to a file for examination. ###
###-------------------------------------------###
# Most of the following points are described at length in
#+ the text of the foregoing "Advanced Bash Scripting Guide."
# This demonstration script is mostly just a reorganized presentation.
# -- msz
# Variables are not typed unless otherwise specified.
# Variables are named. Names must contain a non-digit.
# File descriptor names (as in, for example: 2>&1)
#+ contain ONLY digits.
# Parameters and Bash array elements are numbered.
# (Parameters are very similar to Bash arrays.)
# A variable name may be undefined (null reference).
unset VarNull
# A variable name may be defined but empty (null contents).
VarEmpty='' # Two, adjacent, single quotes.
# A variable name my be defined and non-empty
VarSomething='Literal'
# A variable may contain:
# * A whole number as a signed 32-bit (or larger) integer
# * A string
# A variable may also be an array.
# A string may contain embedded blanks and may be treated
#+ as if it where a function name with optional arguments.
# The names of variables and the names of functions
#+ are in different namespaces.
# A variable may be defined as a Bash array either explicitly or
#+ implicitly by the syntax of the assignment statement.
# Explicit:
declare -a ArrayVar
# The echo command is a built-in.
echo $VarSomething
# The printf command is a built-in.
# Translate %s as: String-Format
printf %s $VarSomething # No linebreak specified, none output.
echo # Default, only linebreak output.
# The Bash parser word breaks on whitespace.
# Whitespace, or the lack of it is significant.
# (This holds true in general; there are, of course, exceptions.)
# Translate the DOLLAR_SIGN character as: Content-Of.
# Extended-Syntax way of writing Content-Of:
echo ${VarSomething}
# The ${ ... } Extended-Syntax allows more than just the variable
#+ name to be specified.
# In general, $VarSomething can always be written as: ${VarSomething}.
# Call this script with arguments to see the following in action.
# Outside of double-quotes, the special characters @ and *
#+ specify identical behavior.
# May be pronounced as: All-Elements-Of.
# Without specification of a name, they refer to the
#+ pre-defined parameter Bash-Array.
# Glob-Pattern references
echo $* # All parameters to script or function
echo ${*} # Same
# Bash disables filename expansion for Glob-Patterns.
# Only character matching is active.
# All-Elements-Of references
echo $@ # Same as above
echo ${@} # Same as above
# Within double-quotes, the behavior of Glob-Pattern references
#+ depends on the setting of IFS (Input Field Separator).
# Within double-quotes, All-Elements-Of references behave the same.
# Specifying only the name of a variable holding a string refers
#+ to all elements (characters) of a string.
# To specify an element (character) of a string,
#+ the Extended-Syntax reference notation (see below) MAY be used.
# Specifying only the name of a Bash array references
#+ the subscript zero element,
#+ NOT the FIRST DEFINED nor the FIRST WITH CONTENTS element.
# Additional qualification is needed to reference other elements,
#+ which means that the reference MUST be written in Extended-Syntax.
# The general form is: ${name[subscript]}.
# The string forms may also be used: ${name:subscript}
#+ for Bash-Arrays when referencing the subscript zero element.
# Bash-Arrays are implemented internally as linked lists,
#+ not as a fixed area of storage as in some programming languages.
# Characteristics of Bash arrays (Bash-Arrays):
# --------------------------------------------
# If not otherwise specified, Bash-Array subscripts begin with
#+ subscript number zero. Literally: [0]
# This is called zero-based indexing.
###
# If not otherwise specified, Bash-Arrays are subscript packed
#+ (sequential subscripts without subscript gaps).
###
# Negative subscripts are not allowed.
###
# Elements of a Bash-Array need not all be of the same type.
###
# Elements of a Bash-Array may be undefined (null reference).
# That is, a Bash-Array my be "subscript sparse."
###
# Elements of a Bash-Array may be defined and empty (null contents).
###
# Elements of a Bash-Array may contain:
# * A whole number as a signed 32-bit (or larger) integer
# * A string
# * A string formated so that it appears to be a function name
# + with optional arguments
###
# Defined elements of a Bash-Array may be undefined (unset).
# That is, a subscript packed Bash-Array may be changed
# + into a subscript sparse Bash-Array.
###
# Elements may be added to a Bash-Array by defining an element
#+ not previously defined.
###
# For these reasons, I have been calling them "Bash-Arrays".
# I'll return to the generic term "array" from now on.
# -- msz
# Demo time -- initialize the previously declared ArrayVar as a
#+ sparse array.
# (The 'unset ... ' is just documentation here.)
unset ArrayVar[0] # Just for the record
ArrayVar[1]=one # Unquoted literal
ArrayVar[2]='' # Defined, and empty
unset ArrayVar[3] # Just for the record
ArrayVar[4]='four' # Quoted literal
# Translate the %q format as: Quoted-Respecting-IFS-Rules.
echo
echo '- - Outside of double-quotes - -'
###
printf %q ${ArrayVar[*]} # Glob-Pattern All-Elements-Of
echo
echo 'echo command:'${ArrayVar[*]}
###
printf %q ${ArrayVar[@]} # All-Elements-Of
echo
echo 'echo command:'${ArrayVar[@]}
# The use of double-quotes may be translated as: Enable-Substitution.
# There are five cases recognized for the IFS setting.
echo
echo '- - Within double-quotes - Default IFS of space-tab-newline - -'
IFS=$'\x20'$'\x09'$'\x0A' # These three bytes,
#+ in exactly this order.
printf %q "${ArrayVar[*]}" # Glob-Pattern All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[*]}"
###
printf %q "${ArrayVar[@]}" # All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[@]}"
echo
echo '- - Within double-quotes - First character of IFS is ^ - -'
# Any printing, non-whitespace character should do the same.
IFS='^'$IFS # ^ + space tab newline
###
printf %q "${ArrayVar[*]}" # Glob-Pattern All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[*]}"
###
printf %q "${ArrayVar[@]}" # All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[@]}"
echo
echo '- - Within double-quotes - Without whitespace in IFS - -'
IFS='^:%!'
###
printf %q "${ArrayVar[*]}" # Glob-Pattern All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[*]}"
###
printf %q "${ArrayVar[@]}" # All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[@]}"
echo
echo '- - Within double-quotes - IFS set and empty - -'
IFS=''
###
printf %q "${ArrayVar[*]}" # Glob-Pattern All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[*]}"
###
printf %q "${ArrayVar[@]}" # All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[@]}"
echo
echo '- - Within double-quotes - IFS undefined - -'
unset IFS
###
printf %q "${ArrayVar[*]}" # Glob-Pattern All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[*]}"
###
printf %q "${ArrayVar[@]}" # All-Elements-Of
echo
echo 'echo command:'"${ArrayVar[@]}"
# Put IFS back to the default.
# Default is exactly these three bytes.
IFS=$'\x20'$'\x09'$'\x0A' # In exactly this order.
# Interpretation of the above outputs:
# A Glob-Pattern is I/O; the setting of IFS matters.
###
# An All-Elements-Of does not consider IFS settings.
###
# Note the different output using the echo command and the
#+ quoted format operator of the printf command.
# Recall:
# Parameters are similar to arrays and have the similar behaviors.
###
# The above examples demonstrate the possible variations.
# To retain the shape of a sparse array, additional script
#+ programming is required.
###
# The source code of Bash has a routine to output the
#+ [subscript]=value array assignment format.
# As of version 2.05b, that routine is not used,
#+ but that might change in future releases.
# The length of a string, measured in non-null elements (characters):
echo
echo '- - Non-quoted references - -'
echo 'Non-Null character count: '${#VarSomething}' characters.'
# test='Lit'$'\x00''eral' # $'\x00' is a null character.
# echo ${#test} # See that?
# The length of an array, measured in defined elements,
#+ including null content elements.
echo
echo 'Defined content count: '${#ArrayVar[@]}' elements.'
# That is NOT the maximum subscript (4).
# That is NOT the range of the subscripts (1 . . 4 inclusive).
# It IS the length of the linked list.
###
# Both the maximum subscript and the range of the subscripts may
#+ be found with additional script programming.
# The length of a string, measured in non-null elements (characters):
echo
echo '- - Quoted, Glob-Pattern references - -'
echo 'Non-Null character count: '"${#VarSomething}"' characters.'
# The length of an array, measured in defined elements,
#+ including null-content elements.
echo
echo 'Defined element count: '"${#ArrayVar[*]}"' elements.'
# Interpretation: Substitution does not effect the ${# ... } operation.
# Suggestion:
# Always use the All-Elements-Of character
#+ if that is what is intended (independence from IFS).
# Define a simple function.
# I include an underscore in the name
#+ to make it distinctive in the examples below.
###
# Bash separates variable names and function names
#+ in different namespaces.
# The Mark-One eyeball isn't that advanced.
###
_simple() {
echo -n 'SimpleFunc'$@ # Newlines are swallowed in
} #+ result returned in any case.
# The ( ... ) notation invokes a command or function.
# The $( ... ) notation is pronounced: Result-Of.
# Invoke the function _simple
echo
echo '- - Output of function _simple - -'
_simple # Try passing arguments.
echo
# or
(_simple) # Try passing arguments.
echo
echo '- Is there a variable of that name? -'
echo $_simple not defined # No variable by that name.
# Invoke the result of function _simple (Error msg intended)
###
$(_simple) # Gives an error message:
# line 394: SimpleFunc: command not found
# ---------------------------------------
echo
###
# The first word of the result of function _simple
#+ is neither a valid Bash command nor the name of a defined function.
###
# This demonstrates that the output of _simple is subject to evaluation.
###
# Interpretation:
# A function can be used to generate in-line Bash commands.
# A simple function where the first word of result IS a bash command:
###
_print() {
echo -n 'printf %q '$@
}
echo '- - Outputs of function _print - -'
_print parm1 parm2 # An Output NOT A Command.
echo
$(_print parm1 parm2) # Executes: printf %q parm1 parm2
# See above IFS examples for the
#+ various possibilities.
echo
$(_print $VarSomething) # The predictable result.
echo
# Function variables
# ------------------
echo
echo '- - Function variables - -'
# A variable may represent a signed integer, a string or an array.
# A string may be used like a function name with optional arguments.
# set -vx # Enable if desired
declare -f funcVar #+ in namespace of functions
funcVar=_print # Contains name of function.
$funcVar parm1 # Same as _print at this point.
echo
funcVar=$(_print ) # Contains result of function.
$funcVar # No input, No output.
$funcVar $VarSomething # The predictable result.
echo
funcVar=$(_print $VarSomething) # $VarSomething replaced HERE.
$funcVar # The expansion is part of the
echo #+ variable contents.
funcVar="$(_print $VarSomething)" # $VarSomething replaced HERE.
$funcVar # The expansion is part of the
echo #+ variable contents.
# The difference between the unquoted and the double-quoted versions
#+ above can be seen in the "protect_literal.sh" example.
# The first case above is processed as two, unquoted, Bash-Words.
# The second case above is processed as one, quoted, Bash-Word.
# Delayed replacement
# -------------------
echo
echo '- - Delayed replacement - -'
funcVar="$(_print '$VarSomething')" # No replacement, single Bash-Word.
eval $funcVar # $VarSomething replaced HERE.
echo
VarSomething='NewThing'
eval $funcVar # $VarSomething replaced HERE.
echo
# Restore the original setting trashed above.
VarSomething=Literal
# There are a pair of functions demonstrated in the
#+ "protect_literal.sh" and "unprotect_literal.sh" examples.
# These are general purpose functions for delayed replacement literals
#+ containing variables.
# REVIEW:
# ------
# A string can be considered a Classic-Array of elements (characters).
# A string operation applies to all elements (characters) of the string
#+ (in concept, anyway).
###
# The notation: ${array_name[@]} represents all elements of the
#+ Bash-Array: array_name.
###
# The Extended-Syntax string operations can be applied to all
#+ elements of an array.
###
# This may be thought of as a For-Each operation on a vector of strings.
###
# Parameters are similar to an array.
# The initialization of a parameter array for a script
#+ and a parameter array for a function only differ
#+ in the initialization of ${0}, which never changes its setting.
###
# Subscript zero of the script's parameter array contains
#+ the name of the script.
###
# Subscript zero of a function's parameter array DOES NOT contain
#+ the name of the function.
# The name of the current function is accessed by the $FUNCNAME variable.
###
# A quick, review list follows (quick, not short).
echo
echo '- - Test (but not change) - -'
echo '- null reference -'
echo -n ${VarNull-'NotSet'}' ' # NotSet
echo ${VarNull} # NewLine only
echo -n ${VarNull:-'NotSet'}' ' # NotSet
echo ${VarNull} # Newline only
echo '- null contents -'
echo -n ${VarEmpty-'Empty'}' ' # Only the space
echo ${VarEmpty} # Newline only
echo -n ${VarEmpty:-'Empty'}' ' # Empty
echo ${VarEmpty} # Newline only
echo '- contents -'
echo ${VarSomething-'Content'} # Literal
echo ${VarSomething:-'Content'} # Literal
echo '- Sparse Array -'
echo ${ArrayVar[@]-'not set'}
# ASCII-Art time
# State Y==yes, N==no
# - :-
# Unset Y Y ${# ... } == 0
# Empty N Y ${# ... } == 0
# Contents N N ${# ... } > 0
# Either the first and/or the second part of the tests
#+ may be a command or a function invocation string.
echo
echo '- - Test 1 for undefined - -'
declare -i t
_decT() {
t=$t-1
}
# Null reference, set: t == -1
t=${#VarNull} # Results in zero.
${VarNull- _decT } # Function executes, t now -1.
echo $t
# Null contents, set: t == 0
t=${#VarEmpty} # Results in zero.
${VarEmpty- _decT } # _decT function NOT executed.
echo $t
# Contents, set: t == number of non-null characters
VarSomething='_simple' # Set to valid function name.
t=${#VarSomething} # non-zero length
${VarSomething- _decT } # Function _simple executed.
echo $t # Note the Append-To action.
# Exercise: clean up that example.
unset t
unset _decT
VarSomething=Literal
echo
echo '- - Test and Change - -'
echo '- Assignment if null reference -'
echo -n ${VarNull='NotSet'}' ' # NotSet NotSet
echo ${VarNull}
unset VarNull
echo '- Assignment if null reference -'
echo -n ${VarNull:='NotSet'}' ' # NotSet NotSet
echo ${VarNull}
unset VarNull
echo '- No assignment if null contents -'
echo -n ${VarEmpty='Empty'}' ' # Space only
echo ${VarEmpty}
VarEmpty=''
echo '- Assignment if null contents -'
echo -n ${VarEmpty:='Empty'}' ' # Empty Empty
echo ${VarEmpty}
VarEmpty=''
echo '- No change if already has contents -'
echo ${VarSomething='Content'} # Literal
echo ${VarSomething:='Content'} # Literal
# "Subscript sparse" Bash-Arrays
###
# Bash-Arrays are subscript packed, beginning with
#+ subscript zero unless otherwise specified.
###
# The initialization of ArrayVar was one way
#+ to "otherwise specify". Here is the other way:
###
echo
declare -a ArraySparse
ArraySparse=( [1]=one [2]='' [4]='four' )
# [0]=null reference, [2]=null content, [3]=null reference
echo '- - Array-Sparse List - -'
# Within double-quotes, default IFS, Glob-Pattern
IFS=$'\x20'$'\x09'$'\x0A'
printf %q "${ArraySparse[*]}"
echo
# Note that the output does not distinguish between "null content"
#+ and "null reference".
# Both print as escaped whitespace.
###
# Note also that the output does NOT contain escaped whitespace
#+ for the "null reference(s)" prior to the first defined element.
###
# This behavior of 2.04, 2.05a and 2.05b has been reported
#+ and may change in a future version of Bash.
# To output a sparse array and maintain the [subscript]=value
#+ relationship without change requires a bit of programming.
# One possible code fragment:
###
# local l=${#ArraySparse[@]} # Count of defined elements
# local f=0 # Count of found subscripts
# local i=0 # Subscript to test
( # Anonymous in-line function
for (( l=${#ArraySparse[@]}, f = 0, i = 0 ; f < l ; i++ ))
do
# 'if defined then...'
${ArraySparse[$i]+ eval echo '\ ['$i']='${ArraySparse[$i]} ; (( f++ )) }
done
)
# The reader coming upon the above code fragment cold
#+ might want to review "command lists" and "multiple commands on a line"
#+ in the text of the foregoing "Advanced Bash Scripting Guide."
###
# Note:
# The "read -a array_name" version of the "read" command
#+ begins filling array_name at subscript zero.
# ArraySparse does not define a value at subscript zero.
###
# The user needing to read/write a sparse array to either
#+ external storage or a communications socket must invent
#+ a read/write code pair suitable for their purpose.
###
# Exercise: clean it up.
unset ArraySparse
echo
echo '- - Conditional alternate (But not change)- -'
echo '- No alternate if null reference -'
echo -n ${VarNull+'NotSet'}' '
echo ${VarNull}
unset VarNull
echo '- No alternate if null reference -'
echo -n ${VarNull:+'NotSet'}' '
echo ${VarNull}
unset VarNull
echo '- Alternate if null contents -'
echo -n ${VarEmpty+'Empty'}' ' # Empty
echo ${VarEmpty}
VarEmpty=''
echo '- No alternate if null contents -'
echo -n ${VarEmpty:+'Empty'}' ' # Space only
echo ${VarEmpty}
VarEmpty=''
echo '- Alternate if already has contents -'
# Alternate literal
echo -n ${VarSomething+'Content'}' ' # Content Literal
echo ${VarSomething}
# Invoke function
echo -n ${VarSomething:+ $(_simple) }' ' # SimpleFunc Literal
echo ${VarSomething}
echo
echo '- - Sparse Array - -'
echo ${ArrayVar[@]+'Empty'} # An array of 'Empty'(ies)
echo
echo '- - Test 2 for undefined - -'
declare -i t
_incT() {
t=$t+1
}
# Note:
# This is the same test used in the sparse array
#+ listing code fragment.
# Null reference, set: t == -1
t=${#VarNull}-1 # Results in minus-one.
${VarNull+ _incT } # Does not execute.
echo $t' Null reference'
# Null contents, set: t == 0
t=${#VarEmpty}-1 # Results in minus-one.
${VarEmpty+ _incT } # Executes.
echo $t' Null content'
# Contents, set: t == (number of non-null characters)
t=${#VarSomething}-1 # non-null length minus-one
${VarSomething+ _incT } # Executes.
echo $t' Contents'
# Exercise: clean up that example.
unset t
unset _incT
# ${name?err_msg} ${name:?err_msg}
# These follow the same rules but always exit afterwards
#+ if an action is specified following the question mark.
# The action following the question mark may be a literal
#+ or a function result.
###
# ${name?} ${name:?} are test-only, the return can be tested.
# Element operations
# ------------------
echo
echo '- - Trailing sub-element selection - -'
# Strings, Arrays and Positional parameters
# Call this script with multiple arguments
#+ to see the parameter selections.
echo '- All -'
echo ${VarSomething:0} # all non-null characters
echo ${ArrayVar[@]:0} # all elements with content
echo ${@:0} # all parameters with content;
# ignoring parameter[0]
echo
echo '- All after -'
echo ${VarSomething:1} # all non-null after character[0]
echo ${ArrayVar[@]:1} # all after element[0] with content
echo ${@:2} # all after param[1] with content
echo
echo '- Range after -'
echo ${VarSomething:4:3} # ral
# Three characters after
# character[3]
echo '- Sparse array gotch -'
echo ${ArrayVar[@]:1:2} # four - The only element with content.
# Two elements after (if that many exist).
# the FIRST WITH CONTENTS
#+ (the FIRST WITH CONTENTS is being
#+ considered as if it
#+ were subscript zero).
# Executed as if Bash considers ONLY array elements with CONTENT
# printf %q "${ArrayVar[@]:0:3}" # Try this one
# In versions 2.04, 2.05a and 2.05b,
#+ Bash does not handle sparse arrays as expected using this notation.
#
# The current Bash maintainer, Chet Ramey, has corrected this
#+ for an upcoming version of Bash.
echo '- Non-sparse array -'
echo ${@:2:2} # Two parameters following parameter[1]
# New victims for string vector examples:
stringZ=abcABC123ABCabc
arrayZ=( abcabc ABCABC 123123 ABCABC abcabc )
sparseZ=( [1]='abcabc' [3]='ABCABC' [4]='' [5]='123123' )
echo
echo ' - - Victim string - -'$stringZ'- - '
echo ' - - Victim array - -'${arrayZ[@]}'- - '
echo ' - - Sparse array - -'${sparseZ[@]}'- - '
echo ' - [0]==null ref, [2]==null ref, [4]==null content - '
echo ' - [1]=abcabc [3]=ABCABC [5]=123123 - '
echo ' - non-null-reference count: '${#sparseZ[@]}' elements'
echo
echo '- - Prefix sub-element removal - -'
echo '- - Glob-Pattern match must include the first character. - -'
echo '- - Glob-Pattern may be a literal or a function result. - -'
echo
# Function returning a simple, Literal, Glob-Pattern
_abc() {
echo -n 'abc'
}
echo '- Shortest prefix -'
echo ${stringZ#123} # Unchanged (not a prefix).
echo ${stringZ#$(_abc)} # ABC123ABCabc
echo ${arrayZ[@]#abc} # Applied to each element.
# Fixed by Chet Ramey for an upcoming version of Bash.
# echo ${sparseZ[@]#abc} # Version-2.05b core dumps.
# The -it would be nice- First-Subscript-Of
# echo ${#sparseZ[@]#*} # This is NOT valid Bash.
echo
echo '- Longest prefix -'
echo ${stringZ##1*3} # Unchanged (not a prefix)
echo ${stringZ##a*C} # abc
echo ${arrayZ[@]##a*c} # ABCABC 123123 ABCABC
# Fixed by Chet Ramey for an upcoming version of Bash
# echo ${sparseZ[@]##a*c} # Version-2.05b core dumps.
echo
echo '- - Suffix sub-element removal - -'
echo '- - Glob-Pattern match must include the last character. - -'
echo '- - Glob-Pattern may be a literal or a function result. - -'
echo
echo '- Shortest suffix -'
echo ${stringZ%1*3} # Unchanged (not a suffix).
echo ${stringZ%$(_abc)} # abcABC123ABC
echo ${arrayZ[@]%abc} # Applied to each element.
# Fixed by Chet Ramey for an upcoming version of Bash.
# echo ${sparseZ[@]%abc} # Version-2.05b core dumps.
# The -it would be nice- Last-Subscript-Of
# echo ${#sparseZ[@]%*} # This is NOT valid Bash.
echo
echo '- Longest suffix -'
echo ${stringZ%%1*3} # Unchanged (not a suffix)
echo ${stringZ%%b*c} # a
echo ${arrayZ[@]%%b*c} # a ABCABC 123123 ABCABC a
# Fixed by Chet Ramey for an upcoming version of Bash.
# echo ${sparseZ[@]%%b*c} # Version-2.05b core dumps.
echo
echo '- - Sub-element replacement - -'
echo '- - Sub-element at any location in string. - -'
echo '- - First specification is a Glob-Pattern - -'
echo '- - Glob-Pattern may be a literal or Glob-Pattern function result. - -'
echo '- - Second specification may be a literal or function result. - -'
echo '- - Second specification may be unspecified. Pronounce that'
echo ' as: Replace-With-Nothing (Delete) - -'
echo
# Function returning a simple, Literal, Glob-Pattern
_123() {
echo -n '123'
}
echo '- Replace first occurrence -'
echo ${stringZ/$(_123)/999} # Changed (123 is a component).
echo ${stringZ/ABC/xyz} # xyzABC123ABCabc
echo ${arrayZ[@]/ABC/xyz} # Applied to each element.
echo ${sparseZ[@]/ABC/xyz} # Works as expected.
echo
echo '- Delete first occurrence -'
echo ${stringZ/$(_123)/}
echo ${stringZ/ABC/}
echo ${arrayZ[@]/ABC/}
echo ${sparseZ[@]/ABC/}
# The replacement need not be a literal,
#+ since the result of a function invocation is allowed.
# This is general to all forms of replacement.
echo
echo '- Replace first occurrence with Result-Of -'
echo ${stringZ/$(_123)/$(_simple)} # Works as expected.
echo ${arrayZ[@]/ca/$(_simple)} # Applied to each element.
echo ${sparseZ[@]/ca/$(_simple)} # Works as expected.
echo
echo '- Replace all occurrences -'
echo ${stringZ//[b2]/X} # X-out b's and 2's
echo ${stringZ//abc/xyz} # xyzABC123ABCxyz
echo ${arrayZ[@]//abc/xyz} # Applied to each element.
echo ${sparseZ[@]//abc/xyz} # Works as expected.
echo
echo '- Delete all occurrences -'
echo ${stringZ//[b2]/}
echo ${stringZ//abc/}
echo ${arrayZ[@]//abc/}
echo ${sparseZ[@]//abc/}
echo
echo '- - Prefix sub-element replacement - -'
echo '- - Match must include the first character. - -'
echo
echo '- Replace prefix occurrences -'
echo ${stringZ/#[b2]/X} # Unchanged (neither is a prefix).
echo ${stringZ/#$(_abc)/XYZ} # XYZABC123ABCabc
echo ${arrayZ[@]/#abc/XYZ} # Applied to each element.
echo ${sparseZ[@]/#abc/XYZ} # Works as expected.
echo
echo '- Delete prefix occurrences -'
echo ${stringZ/#[b2]/}
echo ${stringZ/#$(_abc)/}
echo ${arrayZ[@]/#abc/}
echo ${sparseZ[@]/#abc/}
echo
echo '- - Suffix sub-element replacement - -'
echo '- - Match must include the last character. - -'
echo
echo '- Replace suffix occurrences -'
echo ${stringZ/%[b2]/X} # Unchanged (neither is a suffix).
echo ${stringZ/%$(_abc)/XYZ} # abcABC123ABCXYZ
echo ${arrayZ[@]/%abc/XYZ} # Applied to each element.
echo ${sparseZ[@]/%abc/XYZ} # Works as expected.
echo
echo '- Delete suffix occurrences -'
echo ${stringZ/%[b2]/}
echo ${stringZ/%$(_abc)/}
echo ${arrayZ[@]/%abc/}
echo ${sparseZ[@]/%abc/}
echo
echo '- - Special cases of null Glob-Pattern - -'
echo
echo '- Prefix all -'
# null substring pattern means 'prefix'
echo ${stringZ/#/NEW} # NEWabcABC123ABCabc
echo ${arrayZ[@]/#/NEW} # Applied to each element.
echo ${sparseZ[@]/#/NEW} # Applied to null-content also.
# That seems reasonable.
echo
echo '- Suffix all -'
# null substring pattern means 'suffix'
echo ${stringZ/%/NEW} # abcABC123ABCabcNEW
echo ${arrayZ[@]/%/NEW} # Applied to each element.
echo ${sparseZ[@]/%/NEW} # Applied to null-content also.
# That seems reasonable.
echo
echo '- - Special case For-Each Glob-Pattern - -'
echo '- - - - This is a nice-to-have dream - - - -'
echo
_GenFunc() {
echo -n ${0} # Illustration only.
# Actually, that would be an arbitrary computation.
}
# All occurrences, matching the AnyThing pattern.
# Currently //*/ does not match null-content nor null-reference.
# /#/ and /%/ does match null-content but not null-reference.
echo ${sparseZ[@]//*/$(_GenFunc)}
# A possible syntax would be to make
#+ the parameter notation used within this construct mean:
# ${1} - The full element
# ${2} - The prefix, if any, to the matched sub-element
# ${3} - The matched sub-element
# ${4} - The suffix, if any, to the matched sub-element
#
# echo ${sparseZ[@]//*/$(_GenFunc ${3})} # Same as ${1} here.
# Perhaps it will be implemented in a future version of Bash.
exit 0 |