33.7. Assorted Tips
You have a problem that you want to solve by writing a Bash script. Unfortunately, you don't know quite where to start. One method is to plunge right in and code those parts of the script that come easily, and write the hard parts as pseudo-code.
1 #!/bin/bash 2 3 ARGCOUNT=1 # Need name as argument. 4 E_WRONGARGS=65 5 6 if [ number-of-arguments is-not-equal-to "$ARGCOUNT" ] 7 # ^^^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^ 8 # Can't figure out how to code this . . . 9 #+ . . . so write it in pseudo-code. 10 11 then 12 echo "Usage: name-of-script name" 13 # ^^^^^^^^^^^^^^ More pseudo-code. 14 exit $E_WRONGARGS 15 fi 16 17 . . . 18 19 exit 0 20 21 22 # Later on, substitute working code for the pseudo-code. 23 24 # Line 6 becomes: 25 if [ $# -ne "$ARGCOUNT" ] 26 27 # Line 12 becomes: 28 echo "Usage: `basename $0` name"
For an example of using pseudo-code, see the Square Root exercise.
To keep a record of which user scripts have run during a particular session or over a number of sessions, add the following lines to each script you want to keep track of. This will keep a continuing file record of the script names and invocation times.
1 # Append (>>) following to end of each script tracked. 2 3 whoami>> $SAVE_FILE # User invoking the script. 4 echo $0>> $SAVE_FILE # Script name. 5 date>> $SAVE_FILE # Date and time. 6 echo>> $SAVE_FILE # Blank line as separator. 7 8 # Of course, SAVE_FILE defined and exported as environmental variable in ~/.bashrc 9 #+ (something like ~/.scripts-run)
The >> operator appends lines to a file. What if you wish to prepend a line to an existing file, that is, to paste it in at the beginning?
1 file=data.txt 2 title="***This is the title line of data text file***" 3 4 echo $title | cat - $file >$file.new 5 # "cat -" concatenates stdout to $file. 6 # End result is 7 #+ to write a new file with $title appended at *beginning*.
This is a simplified variant of the Example 18-13 script given earlier. And, of course, sed can also do this.
A shell script may act as an embedded command inside another shell script, a Tcl or wish script, or even a Makefile. It can be invoked as an external shell command in a C program using the system() call, i.e., system("script_name");.
Setting a variable to the contents of an embedded sed or awk script increases the readability of the surrounding shell wrapper. See Example A-1 and Example 14-20.
Put together files containing your favorite and most useful definitions and functions. As necessary, "include" one or more of these "library files" in scripts with either the dot (.) or source command.
1 # SCRIPT LIBRARY 2 # ------ ------- 3 4 # Note: 5 # No "#!" here. 6 # No "live code" either. 7 8 9 # Useful variable definitions 10 11 ROOT_UID=0 # Root has $UID 0. 12 E_NOTROOT=101 # Not root user error. 13 MAXRETVAL=255 # Maximum (positive) return value of a function. 14 SUCCESS=0 15 FAILURE=-1 16 17 18 19 # Functions 20 21 Usage () # "Usage:" message. 22 { 23 if [ -z "$1" ] # No arg passed. 24 then 25 msg=filename 26 else 27 msg=$@ 28 fi 29 30 echo "Usage: `basename $0` "$msg"" 31 } 32 33 34 Check_if_root () # Check if root running script. 35 { # From "ex39.sh" example. 36 if [ "$UID" -ne "$ROOT_UID" ] 37 then 38 echo "Must be root to run this script." 39 exit $E_NOTROOT 40 fi 41 } 42 43 44 CreateTempfileName () # Creates a "unique" temp filename. 45 { # From "ex51.sh" example. 46 prefix=temp 47 suffix=`eval date +%s` 48 Tempfilename=$prefix.$suffix 49 } 50 51 52 isalpha2 () # Tests whether *entire string* is alphabetic. 53 { # From "isalpha.sh" example. 54 [ $# -eq 1 ] || return $FAILURE 55 56 case $1 in 57 *[!a-zA-Z]*|"") return $FAILURE;; 58 *) return $SUCCESS;; 59 esac # Thanks, S.C. 60 } 61 62 63 abs () # Absolute value. 64 { # Caution: Max return value = 255. 65 E_ARGERR=-999999 66 67 if [ -z "$1" ] # Need arg passed. 68 then 69 return $E_ARGERR # Obvious error value returned. 70 fi 71 72 if [ "$1" -ge 0 ] # If non-negative, 73 then # 74 absval=$1 # stays as-is. 75 else # Otherwise, 76 let "absval = (( 0 - $1 ))" # change sign. 77 fi 78 79 return $absval 80 } 81 82 83 tolower () # Converts string(s) passed as argument(s) 84 { #+ to lowercase. 85 86 if [ -z "$1" ] # If no argument(s) passed, 87 then #+ send error message 88 echo "(null)" #+ (C-style void-pointer error message) 89 return #+ and return from function. 90 fi 91 92 echo "$@" | tr A-Z a-z 93 # Translate all passed arguments ($@). 94 95 return 96 97 # Use command substitution to set a variable to function output. 98 # For example: 99 # oldvar="A seT of miXed-caSe LEtTerS" 100 # newvar=`tolower "$oldvar"` 101 # echo "$newvar" # a set of mixed-case letters 102 # 103 # Exercise: Rewrite this function to change lowercase passed argument(s) 104 # to uppercase ... toupper() [easy]. 105 }Use special-purpose comment headers to increase clarity and legibility in scripts.
1 ## Caution. 2 rm -rf *.zzy ## The "-rf" options to "rm" are very dangerous, 3 ##+ especially with wildcards. 4 5 #+ Line continuation. 6 # This is line 1 7 #+ of a multi-line comment, 8 #+ and this is the final line. 9 10 #* Note. 11 12 #o List item. 13 14 #> Another point of view. 15 while [ "$var1" != "end" ] #> while test "$var1" != "end"
A particularly clever use of if-test constructs is commenting out blocks of code.
1 #!/bin/bash 2 3 COMMENT_BLOCK= 4 # Try setting the above variable to some value 5 #+ for an unpleasant surprise. 6 7 if [ $COMMENT_BLOCK ]; then 8 9 Comment block -- 10 ================================= 11 This is a comment line. 12 This is another comment line. 13 This is yet another comment line. 14 ================================= 15 16 echo "This will not echo." 17 18 Comment blocks are error-free! Whee! 19 20 fi 21 22 echo "No more comments, please." 23 24 exit 0
Compare this with using here documents to comment out code blocks.
Using the $? exit status variable, a script may test if a parameter contains only digits, so it can be treated as an integer.
1 #!/bin/bash 2 3 SUCCESS=0 4 E_BADINPUT=65 5 6 test "$1" -ne 0 -o "$1" -eq 0 2>/dev/null 7 # An integer is either equal to 0 or not equal to 0. 8 # 2>/dev/null suppresses error message. 9 10 if [ $? -ne "$SUCCESS" ] 11 then 12 echo "Usage: `basename $0` integer-input" 13 exit $E_BADINPUT 14 fi 15 16 let "sum = $1 + 25" # Would give error if $1 not integer. 17 echo "Sum = $sum" 18 19 # Any variable, not just a command line parameter, can be tested this way. 20 21 exit 0
The 0 - 255 range for function return values is a severe limitation. Global variables and other workarounds are often problematic. An alternative method for a function to communicate a value back to the main body of the script is to have the function write to stdout (usually with echo) the "return value," and assign this to a variable. This is actually a variant of command substitution.
Example 33-15. Return value trickery
1 #!/bin/bash 2 # multiplication.sh 3 4 multiply () # Multiplies params passed. 5 { # Will accept a variable number of args. 6 7 local product=1 8 9 until [ -z "$1" ] # Until uses up arguments passed... 10 do 11 let "product *= $1" 12 shift 13 done 14 15 echo $product # Will not echo to stdout, 16 } #+ since this will be assigned to a variable. 17 18 mult1=15383; mult2=25211 19 val1=`multiply $mult1 $mult2` 20 echo "$mult1 X $mult2 = $val1" 21 # 387820813 22 23 mult1=25; mult2=5; mult3=20 24 val2=`multiply $mult1 $mult2 $mult3` 25 echo "$mult1 X $mult2 X $mult3 = $val2" 26 # 2500 27 28 mult1=188; mult2=37; mult3=25; mult4=47 29 val3=`multiply $mult1 $mult2 $mult3 $mult4` 30 echo "$mult1 X $mult2 X $mult3 X $mult4 = $val3" 31 # 8173300 32 33 exit 0The same technique also works for alphanumeric strings. This means that a function can "return" a non-numeric value.
1 capitalize_ichar () # Capitalizes initial character 2 { #+ of argument string(s) passed. 3 4 string0="$@" # Accepts multiple arguments. 5 6 firstchar=${string0:0:1} # First character. 7 string1=${string0:1} # Rest of string(s). 8 9 FirstChar=`echo "$firstchar" | tr a-z A-Z` 10 # Capitalize first character. 11 12 echo "$FirstChar$string1" # Output to stdout. 13 14 } 15 16 newstring=`capitalize_ichar "every sentence should start with a capital letter."` 17 echo "$newstring" # Every sentence should start with a capital letter.It is even possible for a function to "return" multiple values with this method.
Example 33-16. Even more return value trickery
1 #!/bin/bash 2 # sum-product.sh 3 # A function may "return" more than one value. 4 5 sum_and_product () # Calculates both sum and product of passed args. 6 { 7 echo $(( $1 + $2 )) $(( $1 * $2 )) 8 # Echoes to stdout each calculated value, separated by space. 9 } 10 11 echo 12 echo "Enter first number " 13 read first 14 15 echo 16 echo "Enter second number " 17 read second 18 echo 19 20 retval=`sum_and_product $first $second` # Assigns output of function. 21 sum=`echo "$retval" | awk '{print $1}'` # Assigns first field. 22 product=`echo "$retval" | awk '{print $2}'` # Assigns second field. 23 24 echo "$first + $second = $sum" 25 echo "$first * $second = $product" 26 echo 27 28 exit 0Next in our bag of trick are techniques for passing an array to a function, then "returning" an array back to the main body of the script.
Passing an array involves loading the space-separated elements of the array into a variable with command substitution. Getting an array back as the "return value" from a function uses the previously mentioned strategem of echoing the array in the function, then invoking command substitution and the ( ... ) operator to assign it to an array.
Example 33-17. Passing and returning arrays
1 #!/bin/bash 2 # array-function.sh: Passing an array to a function and... 3 # "returning" an array from a function 4 5 6 Pass_Array () 7 { 8 local passed_array # Local variable. 9 passed_array=( `echo "$1"` ) 10 echo "${passed_array[@]}" 11 # List all the elements of the new array 12 #+ declared and set within the function. 13 } 14 15 16 original_array=( element1 element2 element3 element4 element5 ) 17 18 echo 19 echo "original_array = ${original_array[@]}" 20 # List all elements of original array. 21 22 23 # This is the trick that permits passing an array to a function. 24 # ********************************** 25 argument=`echo ${original_array[@]}` 26 # ********************************** 27 # Pack a variable 28 #+ with all the space-separated elements of the original array. 29 # 30 # Note that attempting to just pass the array itself will not work. 31 32 33 # This is the trick that allows grabbing an array as a "return value". 34 # ***************************************** 35 returned_array=( `Pass_Array "$argument"` ) 36 # ***************************************** 37 # Assign 'echoed' output of function to array variable. 38 39 echo "returned_array = ${returned_array[@]}" 40 41 echo "=============================================================" 42 43 # Now, try it again, 44 #+ attempting to access (list) the array from outside the function. 45 Pass_Array "$argument" 46 47 # The function itself lists the array, but... 48 #+ accessing the array from outside the function is forbidden. 49 echo "Passed array (within function) = ${passed_array[@]}" 50 # NULL VALUE since this is a variable local to the function. 51 52 echo 53 54 exit 0For a more elaborate example of passing arrays to functions, see Example A-10.
Using the double parentheses construct, it is possible to use C-like syntax for setting and incrementing variables and in for and while loops. See Example 10-12 and Example 10-17.
Setting the path and umask at the beginning of a script makes it more "portable" -- more likely to run on a "foreign" machine whose user may have bollixed up the $PATH and umask.
1 #!/bin/bash 2 PATH=/bin:/usr/bin:/usr/local/bin ; export PATH 3 umask 022 # Files that the script creates will have 755 permission. 4 5 # Thanks to Ian D. Allen, for this tip.
A useful scripting technique is to repeatedly feed the output of a filter (by piping) back to the same filter, but with a different set of arguments and/or options. Especially suitable for this are tr and grep.
1 # From "wstrings.sh" example. 2 3 wlist=`strings "$1" | tr A-Z a-z | tr '[:space:]' Z | \ 4 tr -cs '[:alpha:]' Z | tr -s '\173-\377' Z | tr Z ' '`
Example 33-18. Fun with anagrams
1 #!/bin/bash 2 # agram.sh: Playing games with anagrams. 3 4 # Find anagrams of... 5 LETTERSET=etaoinshrdlu 6 FILTER='.......' # How many letters minimum? 7 # 1234567 8 9 anagram "$LETTERSET" | # Find all anagrams of the letterset... 10 grep "$FILTER" | # With at least 7 letters, 11 grep '^is' | # starting with 'is' 12 grep -v 's$' | # no plurals 13 grep -v 'ed$' # no past tense verbs 14 # Possible to add many combinations of conditions and filters. 15 16 # Uses "anagram" utility 17 #+ that is part of the author's "yawl" word list package. 18 # http://ibiblio.org/pub/Linux/libs/yawl-0.3.2.tar.gz 19 # http://personal.riverusers.com/~thegrendel/yawl-0.3.2.tar.gz 20 21 exit 0 # End of code. 22 23 24 bash$ sh agram.sh 25 islander 26 isolate 27 isolead 28 isotheral 29 30 31 32 # Exercises: 33 # --------- 34 # Modify this script to take the LETTERSET as a command-line parameter. 35 # Parameterize the filters in lines 11 - 13 (as with $FILTER), 36 #+ so that they can be specified by passing arguments to a function. 37 38 # For a slightly different approach to anagramming, 39 #+ see the agram2.sh script.
See also Example 27-3, Example 15-23, and Example A-9.
Use "anonymous here documents" to comment out blocks of code, to save having to individually comment out each line with a #. See Example 18-11.
Running a script on a machine that relies on a command that might not be installed is dangerous. Use whatis to avoid potential problems with this.
1 CMD=command1 # First choice. 2 PlanB=command2 # Fallback option. 3 4 command_test=$(whatis "$CMD" | grep 'nothing appropriate') 5 # If 'command1' not found on system , 'whatis' will return 6 #+ "command1: nothing appropriate." 7 # 8 # A safer alternative is: 9 # command_test=$(whereis "$CMD" | grep \/) 10 # But then the sense of the following test would have to be reversed, 11 #+ since the $command_test variable holds content only if 12 #+ the $CMD exists on the system. 13 # (Thanks, bojster.) 14 15 16 if [[ -z "$command_test" ]] # Check whether command present. 17 then 18 $CMD option1 option2 # Run command1 with options. 19 else # Otherwise, 20 $PlanB #+ run command2. 21 fi
An if-grep test may not return expected results in an error case, when text is output to stderr, rather that stdout.
1 if ls -l nonexistent_filename | grep -q 'No such file or directory' 2 then echo "File \"nonexistent_filename\" does not exist." 3 fi
Redirecting stderr to stdout fixes this.
1 if ls -l nonexistent_filename 2>&1 | grep -q 'No such file or directory' 2 # ^^^^ 3 then echo "File \"nonexistent_filename\" does not exist." 4 fi 5 6 # Thanks, Chris Martin, for pointing this out.
The run-parts command is handy for running a set of command scripts in sequence, particularly in combination with cron or at.
It would be nice to be able to invoke X-Windows widgets from a shell script. There happen to exist several packages that purport to do so, namely Xscript, Xmenu, and widtools. The first two of these no longer seem to be maintained. Fortunately, it is still possible to obtain widtools here.
The widtools (widget tools) package requires the XForms library to be installed. Additionally, the Makefile needs some judicious editing before the package will build on a typical Linux system. Finally, three of the six widgets offered do not work (and, in fact, segfault).
The dialog family of tools offers a method of calling "dialog" widgets from a shell script. The original dialog utility works in a text console, but its successors, gdialog, Xdialog, and kdialog use X-Windows-based widget sets.
Example 33-19. Widgets invoked from a shell script
1 #!/bin/bash 2 # dialog.sh: Using 'gdialog' widgets. 3 # Must have 'gdialog' installed on your system to run this script. 4 # Version 1.1 (corrected 04/05/05) 5 6 # This script was inspired by the following article. 7 # "Scripting for X Productivity," by Marco Fioretti, 8 # LINUX JOURNAL, Issue 113, September 2003, pp. 86-9. 9 # Thank you, all you good people at LJ. 10 11 12 # Input error in dialog box. 13 E_INPUT=65 14 # Dimensions of display, input widgets. 15 HEIGHT=50 16 WIDTH=60 17 18 # Output file name (constructed out of script name). 19 OUTFILE=$0.output 20 21 # Display this script in a text widget. 22 gdialog --title "Displaying: $0" --textbox $0 $HEIGHT $WIDTH 23 24 25 26 # Now, we'll try saving input in a file. 27 echo -n "VARIABLE=" > $OUTFILE 28 gdialog --title "User Input" --inputbox "Enter variable, please:" \ 29 $HEIGHT $WIDTH 2>> $OUTFILE 30 31 32 if [ "$?" -eq 0 ] 33 # It's good practice to check exit status. 34 then 35 echo "Executed \"dialog box\" without errors." 36 else 37 echo "Error(s) in \"dialog box\" execution." 38 # Or, clicked on "Cancel", instead of "OK" button. 39 rm $OUTFILE 40 exit $E_INPUT 41 fi 42 43 44 45 # Now, we'll retrieve and display the saved variable. 46 . $OUTFILE # 'Source' the saved file. 47 echo "The variable input in the \"input box\" was: "$VARIABLE"" 48 49 50 rm $OUTFILE # Clean up by removing the temp file. 51 # Some applications may need to retain this file. 52 53 exit $?
For other methods of scripting with widgets, try Tk or wish (Tcl derivatives), PerlTk (Perl with Tk extensions), tksh (ksh with Tk extensions), XForms4Perl (Perl with XForms extensions), Gtk-Perl (Perl with Gtk extensions), or PyQt (Python with Qt extensions).
For doing multiple revisions on a complex script, use the rcs Revision Control System package.
Among other benefits of this is automatically updated ID header tags. The co command in rcs does a parameter replacement of certain reserved key words, for example, replacing #$Id$ in a script with something like:
1 #$Id: hello-world.sh,v 1.1 2004/10/16 02:43:05 bozo Exp $