7764
Comment: Reset OPTIND before using getopts
|
8761
getopt from util-linux supports different styles of long options and options with mandatory and optional arguments. I have used it in my linux-purge script that is meant for Ubuntu Linux.
|
Deletions are marked like this. | Additions are marked like this. |
Line 2: | Line 2: |
== How can I handle command-line arguments (options) to my script easily? == Well, that depends a great deal on what you want to do with them. There are several approaches, each with its strengths and weaknesses. |
== How can I handle command-line options and arguments in my script easily? == Well, that depends a great deal on what you want to do with them. There are two standard approaches, each with its strengths and weaknesses. <<TableOfContents>> === Overview === A Unix command generally has an argument syntax like this: {{{ tar -x -f archive.tar -v -- file1 file2 file3 }}} Please note the conventions and the ordering here, because they are ''important''. They actually matter. This command has some ''arguments'' (file1, file2, file3), and some ''options'' (-x -f archive.tar -v), as well as the traditional ''end of options indicator'' "--". The options appear ''before'' the non-option arguments. They do not appear afterward. They do not appear at just any old random place in the command. Some options (-x, -v) are standalones. They are either present, or not. Some options (-f) take an ''argument'' of their own. In all cases, option processing involves writing a loop. Ideally, this loop will make ''one pass'' over the argument list, examining each argument in turn, and setting appropriate shell variables so that the script remembers which options are in effect. Ultimately, it will ''discard'' all of the options, so that the argument list is left holding only the ''non-option arguments'' (file1 file2 file3). The rest of the script, then, can simply begin processing those, referring as needed to the variables that were set up by the option processor. The option processor recognizes the end of options when it finds a -- argument, or when it finds an argument that doesn't start with a hyphen. (The ''option argument'' archive.tar does not signal the end of options, because it is processed along with the -f option.) There are two basic approaches to writing an option processing loop: either write the loop yourself from scratch (we'll call this a "manual loop"), or use the shell's `getopts` command to assist with option splitting. We'll cover both of these cases. '''Advisory about'''''' getopt(1):''' Traditional versions of `getopt` cannot handle empty argument strings, or arguments with embedded whitespace. There is a version of `getopt` in util-linux which supports these, but you '''should not use it unless the script is only used in systems that has util-linux'''. Why not? Because: 1. You would need to write special safety-checking code to ensure that you've actually got the util-linux version of `getopt`, and then you would ''still'' need to write a fallback option processor for when you ''don't'' have it. So you're doing twice as much work. |
Line 6: | Line 31: |
This approach handles any arbitrary set of options, because you're writing the parser yourself. For 90% of programs, this turns out to be the simplest and most direct approach, since very few scripts need complicated option processing. | Manually parsing options is the most flexible approach. It is the ''best'' way, really, because it allows you to do anything you like: you can handle both single-letter and long options, with or without option arguments. That's why we're showing it first. |
Line 8: | Line 33: |
Here's an example that will handle a combination of short (`-h`) and long (`--help`) options. | If you want to handle GNU-style `--long-options` or Tcl-style `-longopts`, a manual loop is your only choice. `getopts` does not support these. |
Line 10: | Line 35: |
{{{ # Bash while [[ $1 == -* ]]; do case "$1" in -h|--help|-\?) show_help; exit 0;; -v|--verbose) verbose=1; shift;; -f) if (($# > 1)); then output_file=$2; shift 2 else echo "-f requires an argument" 1>&2 |
In this example, notice how both `--file FILE` and `--file=FILE` are handled. {{{#!highlight bash #!/bin/sh # POSIX die() { printf '%s\n' "$1" >&2 exit 1 } # Initialize all the option variables. # This ensures we are not contaminated by variables from the environment. file= verbose=0 while :; do case $1 in -h|-\?|--help) show_help # Display a usage synopsis. exit ;; -f|--file) # Takes an option argument; ensure it has been specified. if [ "$2" ]; then file=$2 shift else die 'ERROR: "--file" requires a non-empty option argument.' fi ;; --file=?*) file=${1#*=} # Delete everything up to "=" and assign the remainder. ;; --file=) # Handle the case of an empty --file= die 'ERROR: "--file" requires a non-empty option argument.' ;; -v|--verbose) verbose=$((verbose + 1)) # Each -v adds 1 to verbosity. ;; --) # End of all options. shift break ;; -?*) printf 'WARN: Unknown option (ignored): %s\n' "$1" >&2 ;; *) # Default case: No more options, so break out of the loop. break esac shift done # if --file was provided, open it for writing, else duplicate stdout if [ "$file" ]; then exec 3> "$file" else exec 3>&1 fi # Rest of the program here. # If there are input files (for example) that follow the options, they # will remain in the "$@" positional parameters. }}} This parser does not handle single-letter options concatenated together (like `-xvf` being understood as `-x -v -f`). This could be added with effort, but it's left as an exercise for the reader. In practice, it's exceptionally rare for shell scripts that handle long options to handle single-letter option splitting as well. It's simply not worth the effort. For the most part, shell scripts that you write will not need to worry about single-letter option splitting, because you are the only person using them. Fancy option processing is only desirable if you are releasing the program for general use, and that is almost ''never'' going to be the case in real life. Single-letter option combining also precludes the use of Tcl-style long arguments (`-foo`), which some commands like [[https://pubs.opengroup.org/onlinepubs/9699919799/utilities/find.html|find(1)]], [[https://linux.die.net/man/1/gcc|gcc(1)]] and [[https://linux.die.net/man/1/star|star(1)]] use. === getopts === The main benefit of `getopts` is to allow single-letter option splitting (`-xvf` handled as `-x -v -f`). The trade-off for this is that you cannot use long arguments of any kind (GNU-style `--foo` or Tcl-style `-foo`), or options with an optional argument (like `mysql`'s `-p[password]` option). `getopts` is suitable for simple scripts. The more complex your option parsing needs are, the less likely it is that you'll be able to make use of `getopts`. Here is a `getopts` example: {{{#!highlight bash #!/bin/sh # Usage info show_help() { cat << EOF Usage: ${0##*/} [-hv] [-f OUTFILE] [FILE]... Do stuff with FILE and write the result to standard output. With no FILE or when FILE is -, read standard input. -h display this help and exit -f OUTFILE write the result to OUTFILE instead of standard output. -v verbose mode. Can be used multiple times for increased verbosity. EOF } # Initialize our own variables: output_file="" verbose=0 OPTIND=1 # Resetting OPTIND is necessary if getopts was used previously in the script. # It is a good idea to make OPTIND local if you process options in a function. while getopts hvf: opt; do case $opt in h) show_help exit 0 ;; v) verbose=$((verbose+1)) ;; f) output_file=$OPTARG ;; *) show_help >&2 |
Line 21: | Line 148: |
fi ;; --) shift; break;; -*) echo "invalid option: $1" 1>&2; show_help; exit 1;; |
;; |
Line 26: | Line 151: |
shift "$((OPTIND-1))" # Discard the options and sentinel -- # Everything that's left in "$@" is a non-option. In our case, a FILE to process. printf 'verbose=<%d>\noutput_file=<%s>\nLeftovers:\n' "$verbose" "$output_file" printf '<%s>\n' "$@" # End of file |
|
Line 27: | Line 159: |
Now all of the remaining arguments are the filenames which followed the optional switches. You can process those with `for i` or `"$@"`. | There is a [[http://wiki.bash-hackers.org/howto/getopts_tutorial|getopts tutorial]] which explains what all of the syntax and variables mean. In bash, there is also `help getopts`. |
Line 29: | Line 161: |
A POSIX version of that same code: {{{ # POSIX while true; do case "$1" in -h|--help|-\?) show_help; exit 0;; -v|--verbose) verbose=1; shift;; -f) if [ $# -gt 1 ]; then output_file=$2; shift 2 else echo "-f requires an argument" 1>&2 exit 1 fi ;; --) shift; break;; -*) echo "invalid option: $1" 1>&2; show_help; exit 1;; *) break;; esac done }}} |
The advantages of `getopts` over a manual loop: |
Line 49: | Line 163: |
Some Bash programmers write this at the beginning of their scripts: {{{ set -u # or, set -o nounset }}} This way Bash stops if it's forced to work with the value of an unset variable. If you use `set -o nounset`, the Bash version of the "manual loop" shown above may break, if there are no additional non-option arguments. It can be fixed thus: |
1. It can handle things like `-xvf filename` in the expected Unix way, automatically. 1. It makes sure options are parsed like any standard command (lowest common denominator), avoiding surprises. 1. With some implementations, the error messages will be localised in the language of the user. |
Line 56: | Line 167: |
{{{ # Bash (with set -u) while [[ $1 == -* ]]; do case "$1" in -h|--help|-\?) show_help; exit 0;; -v|--verbose) verbose=1; shift;; -f) if (($# > 1)); then output_file=$2; shift 2 else echo "-f requires an argument" 1>&2 exit 1 fi ;; --) shift; break;; -*) echo "invalid option: $1" 1>&2; show_help; exit 1;; esac if ! test "${1+defined}"; then break fi done }}} |
The disadvantages of `getopts`: |
Line 77: | Line 169: |
What these examples ''do not'' handle are: * You want things like `-xvf` to be handled as three separate flags (equivalent to `-x -v -f`). * You want to parse arguments out of `--file=bar`. |
1. (Except for ksh93) it can only handle short options (`-h`, not `--help`). 1. It cannot handle options with optional arguments like `mysql`'s `-p[password]`. 1. It doesn't exist in the Bourne shell. 1. It ''only'' allows options to be parsed in the "standard way" (lowest common denominator). 1. Options are coded in at least 2, probably 3 places -- in the call to `getopts`, in the case statement that processes them, and in the help/usage message that documents them. |
Line 81: | Line 175: |
It's certainly possible to do those things by changing the code, but at least in the first case, there's another approach that handles that automatically. === getopts === '''Never use getopt(1).''' `getopt` cannot handle empty arguments strings, or arguments with embedded whitespace. Please forget that it ever existed. The POSIX shell (and others) offer `getopts` which is safe to use instead. Here is a simplistic `getopts` example: {{{ # POSIX OPTIND=1 # Reset if getopts has been used previously in the shell. while getopts "h?vf:" opt; do case "$opt" in h|\?) show_help; exit 0;; v) verbose=1;; f) output_file=$OPTARG;; esac done shift $((OPTIND-1)) if [ "$1" = -- ]; then shift; fi echo "verbose=$verbose, output_file='$output_file', Leftovers: $@" }}} The disadvantage of `getopts` is that it can only handle short options (`-h`), not long options. But it handles `-vf filename` in the expected Unix way, automatically. So, if one wishes to sacrifice `--help` to get `-vf filename`, then `getops` is a good candidate. There is a [[http://wiki.bash-hackers.org/howto/getopts_tutorial|getopts tutorial]] which explains what all of the syntax and variables mean. In bash, there is also `help getopts`, which might be informative. There is also still the disadvantage that options are coded in at least 2, probably 3 places - in the call to `getopts`, in the case statement that processes them and presumably in the help message that you are going to get around to writing one of these days. This is a classic opportunity for errors to creep in as the code is written and maintained - often not discovered till much, much later. === Silly repeated brute-force scanning === Another approach is to check options with `if` statements "on demand". A function like this one may be useful: {{{ # Bash HaveOpt() { local needle=$1 shift while [[ $1 == -* ]]; do case "$1" in --) return 1;; # by convention, -- is end of options $needle) return 0;; esac shift done return 1 } if HaveOpt --quick "$@"; then echo "Option quick is set"; fi }}} and it will work if script is run as: * YES: ./script --quick * YES: ./script -other --quick but will stop on first argument with no "-" in front (or on --): * NO: ./script -bar foo --quick * NO: ./script -bar -- --quick Of course, this approach (iterating over the argument list every time you want to check for one) is far less efficient than just iterating once and setting flag variables. It also spreads the options throughout the program. The literal option `--quick` may appear a hundred lines down inside the main body of the program, nowhere near any other option name. This is a nightmare for maintenance. === Complex nonstandard add-on utilities === [[http://bhepple.freeshell.org/oddmuse/wiki.cgi/process-getopt|bhepple]] suggests the use of [[http://sourceforge.net/projects/process-getopt/|process-getopt]] (GPL licensed) and offers this example code: {{{ PROG=$(basename $0) VERSION='1.2' USAGE="A tiny example using process-getopt(1)" # call process-getopt functions to define some options: source process-getopt SLOT="" SLOT_func() { [ "${1:-""}" ] && SLOT="yes"; } # callback for SLOT option add_opt SLOT "boolean option" s "" slot TOKEN="" TOKEN_func() { [ "${1:-""}" ] && TOKEN="$2"; } # callback for TOKEN option add_opt TOKEN "this option takes a value" t n token number add_std_opts # define the standard options --help etc: TEMP=$(call_getopt "$@") || exit 1 eval set -- "$TEMP" # just as with getopt(1) # remove the options from the command line process_opts "$@" || shift "$?" echo "SLOT=$SLOT" echo "TOKEN=$TOKEN" echo "args=$@" }}} Here, all information about each option is defined in one place making for much easier authoring and maintenance. A lot of the dirty work is handled automatically and standards are obeyed as in getopt(1) - because it calls getopt for you. . ''Actually, what the author forgot to say was that it's actually using `getopts` semantics, rather than `getopt`. I ran this test:'' {{{ wooledg@wooledg:~/process-getopt-1.6$ set -- one 'rm -rf /' 'foo;bar' "'" wooledg@wooledg:~/process-getopt-1.6$ call_getopt "$@" -- 'rm -rf /' 'foo;bar' ''\''' }}} . ''It appears to be intelligent enough to handle null options, whitespace-containing options, and single-quote-containing options in a manner that makes the [[BashFAQ/048|eval]] not blow up in your face. But this is not an endorsement of the process-getopt software overall; I don't know it well enough. -GreyCat It's written and tested on Linux where getopt(1) supports long options. For portability, it tests the local getopt(1) at runtime and if it finds an non-GNU one (ie one that does not return 4 for {{{getopt --test}}}) it only processes short options. It does not use the bash builtin getopts(1) command. -[[http://bhepple.freeshell.org/oddmuse/wiki.cgi/process-getopt|bhepple]] |
For other, more complicated ways of option parsing, see ComplexOptionParsing. |
How can I handle command-line options and arguments in my script easily?
Well, that depends a great deal on what you want to do with them. There are two standard approaches, each with its strengths and weaknesses.
Contents
Overview
A Unix command generally has an argument syntax like this:
tar -x -f archive.tar -v -- file1 file2 file3
Please note the conventions and the ordering here, because they are important. They actually matter. This command has some arguments (file1, file2, file3), and some options (-x -f archive.tar -v), as well as the traditional end of options indicator "--".
The options appear before the non-option arguments. They do not appear afterward. They do not appear at just any old random place in the command.
Some options (-x, -v) are standalones. They are either present, or not. Some options (-f) take an argument of their own.
In all cases, option processing involves writing a loop. Ideally, this loop will make one pass over the argument list, examining each argument in turn, and setting appropriate shell variables so that the script remembers which options are in effect. Ultimately, it will discard all of the options, so that the argument list is left holding only the non-option arguments (file1 file2 file3). The rest of the script, then, can simply begin processing those, referring as needed to the variables that were set up by the option processor.
The option processor recognizes the end of options when it finds a -- argument, or when it finds an argument that doesn't start with a hyphen. (The option argument archive.tar does not signal the end of options, because it is processed along with the -f option.)
There are two basic approaches to writing an option processing loop: either write the loop yourself from scratch (we'll call this a "manual loop"), or use the shell's getopts command to assist with option splitting. We'll cover both of these cases.
Advisory about getopt(1): Traditional versions of getopt cannot handle empty argument strings, or arguments with embedded whitespace. There is a version of getopt in util-linux which supports these, but you should not use it unless the script is only used in systems that has util-linux. Why not? Because:
You would need to write special safety-checking code to ensure that you've actually got the util-linux version of getopt, and then you would still need to write a fallback option processor for when you don't have it. So you're doing twice as much work.
Manual loop
Manually parsing options is the most flexible approach. It is the best way, really, because it allows you to do anything you like: you can handle both single-letter and long options, with or without option arguments. That's why we're showing it first.
If you want to handle GNU-style --long-options or Tcl-style -longopts, a manual loop is your only choice. getopts does not support these.
In this example, notice how both --file FILE and --file=FILE are handled.
1 #!/bin/sh
2 # POSIX
3
4 die() {
5 printf '%s\n' "$1" >&2
6 exit 1
7 }
8
9 # Initialize all the option variables.
10 # This ensures we are not contaminated by variables from the environment.
11 file=
12 verbose=0
13
14 while :; do
15 case $1 in
16 -h|-\?|--help)
17 show_help # Display a usage synopsis.
18 exit
19 ;;
20 -f|--file) # Takes an option argument; ensure it has been specified.
21 if [ "$2" ]; then
22 file=$2
23 shift
24 else
25 die 'ERROR: "--file" requires a non-empty option argument.'
26 fi
27 ;;
28 --file=?*)
29 file=${1#*=} # Delete everything up to "=" and assign the remainder.
30 ;;
31 --file=) # Handle the case of an empty --file=
32 die 'ERROR: "--file" requires a non-empty option argument.'
33 ;;
34 -v|--verbose)
35 verbose=$((verbose + 1)) # Each -v adds 1 to verbosity.
36 ;;
37 --) # End of all options.
38 shift
39 break
40 ;;
41 -?*)
42 printf 'WARN: Unknown option (ignored): %s\n' "$1" >&2
43 ;;
44 *) # Default case: No more options, so break out of the loop.
45 break
46 esac
47
48 shift
49 done
50
51 # if --file was provided, open it for writing, else duplicate stdout
52 if [ "$file" ]; then
53 exec 3> "$file"
54 else
55 exec 3>&1
56 fi
57
58 # Rest of the program here.
59 # If there are input files (for example) that follow the options, they
60 # will remain in the "$@" positional parameters.
This parser does not handle single-letter options concatenated together (like -xvf being understood as -x -v -f). This could be added with effort, but it's left as an exercise for the reader. In practice, it's exceptionally rare for shell scripts that handle long options to handle single-letter option splitting as well. It's simply not worth the effort.
For the most part, shell scripts that you write will not need to worry about single-letter option splitting, because you are the only person using them. Fancy option processing is only desirable if you are releasing the program for general use, and that is almost never going to be the case in real life. Single-letter option combining also precludes the use of Tcl-style long arguments (-foo), which some commands like find(1), gcc(1) and star(1) use.
getopts
The main benefit of getopts is to allow single-letter option splitting (-xvf handled as -x -v -f). The trade-off for this is that you cannot use long arguments of any kind (GNU-style --foo or Tcl-style -foo), or options with an optional argument (like mysql's -p[password] option).
getopts is suitable for simple scripts. The more complex your option parsing needs are, the less likely it is that you'll be able to make use of getopts.
Here is a getopts example:
1 #!/bin/sh
2
3 # Usage info
4 show_help() {
5 cat << EOF
6 Usage: ${0##*/} [-hv] [-f OUTFILE] [FILE]...
7 Do stuff with FILE and write the result to standard output. With no FILE
8 or when FILE is -, read standard input.
9
10 -h display this help and exit
11 -f OUTFILE write the result to OUTFILE instead of standard output.
12 -v verbose mode. Can be used multiple times for increased
13 verbosity.
14 EOF
15 }
16
17 # Initialize our own variables:
18 output_file=""
19 verbose=0
20
21 OPTIND=1
22 # Resetting OPTIND is necessary if getopts was used previously in the script.
23 # It is a good idea to make OPTIND local if you process options in a function.
24
25 while getopts hvf: opt; do
26 case $opt in
27 h)
28 show_help
29 exit 0
30 ;;
31 v) verbose=$((verbose+1))
32 ;;
33 f) output_file=$OPTARG
34 ;;
35 *)
36 show_help >&2
37 exit 1
38 ;;
39 esac
40 done
41 shift "$((OPTIND-1))" # Discard the options and sentinel --
42
43 # Everything that's left in "$@" is a non-option. In our case, a FILE to process.
44 printf 'verbose=<%d>\noutput_file=<%s>\nLeftovers:\n' "$verbose" "$output_file"
45 printf '<%s>\n' "$@"
46
47 # End of file
There is a getopts tutorial which explains what all of the syntax and variables mean. In bash, there is also help getopts.
The advantages of getopts over a manual loop:
It can handle things like -xvf filename in the expected Unix way, automatically.
- It makes sure options are parsed like any standard command (lowest common denominator), avoiding surprises.
- With some implementations, the error messages will be localised in the language of the user.
The disadvantages of getopts:
(Except for ksh93) it can only handle short options (-h, not --help).
It cannot handle options with optional arguments like mysql's -p[password].
- It doesn't exist in the Bourne shell.
It only allows options to be parsed in the "standard way" (lowest common denominator).
Options are coded in at least 2, probably 3 places -- in the call to getopts, in the case statement that processes them, and in the help/usage message that documents them.
For other, more complicated ways of option parsing, see ComplexOptionParsing.