Differences between revisions 35 and 45 (spanning 10 versions)
Revision 35 as of 2011-10-08 15:56:56
Size: 10333
Editor: dslb-092-076-251-070
Comment:
Revision 45 as of 2012-02-01 12:36:56
Size: 8709
Editor: a91-155-176-245
Comment: Trim down example and leave only option handling
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
#pragma section-numbers 3
Line 2: Line 3:
Line 3: Line 5:
Well, that depends a great deal on what you want to do with them.  There are several approaches, each with its strengths and weaknesses.
Well, that depends a great deal on what you want to do with them. There are several approaches, each with its strengths and weaknesses.

<<TableOfContents>>
Line 6: Line 11:
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.

Here's an example that will handle a combination of short (`-h`) and long (`--help`) options.

This approach handles any arbitrary set of options, because you're writing the parser yourself. For 90% of programs, it may suffice.

Here's an example that will handle a combination of short (`-v`, `-h`) and long (`--verbose`, `--help`) options; and also style `--verbose=LEVEL`.

Main ()
{
    help="" # Reset all option variables that might be set later
    verbose=""
    verbose_level=""

    while :
    do
        case "$1" in
            -h | --help)
                help="help"
                shift # Remove from argument list
  # Now, e.g. call Help() function
                ;;
            --verbose=*)
  # If you want to support --option=VALUE style
                verbose="verbose"
  verbose_level=${1#*=} # Delete everyting up till "="
  shift
  ;;
            -v | --verbose)
                verbose="verbose"
  verbose_level=1
  shift

  # Support "--verbose LEVEL" by reading next argument.

  case "$1" in
      [0-9]) verbose_level=$1
      shift
      ;;
  esac
                ;;
            --) # End of all options
                shift
                break
                ;;
            -*)
                Warn "WARN: Unknown option: $1"
                shift
                ;;
            *) # no more options. Stop while loop
                break
                ;;
        esac
    done

    # Suppose, some options are required. Check that we got them

    [ "$verbose" ] || Die "ERROR: option --verbose not given. See --help"

    # <Rest of your code here>
}

Main "$@"

# End of file
}}}

What is ''not possible'' to do with pure shell approach, is to try to
parse separate options concatenaed together. Like like `-xvf` which
would understoos as `-x -v -f`. This could be possible with lot of
effort, but in practise it wouldn't be worth it.

You may be interested in knowing that some Bash programmers like
to write this at the beginning of their scripts to guard against
unused variables:
Line 11: Line 85:
# 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
            exit 1
          fi ;;
      --) shift; break;;
      -*) echo "invalid option: $1" 1>&2; show_help; exit 1;;
    set -u # or, set -o nounset
}}}

=== 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:

{{{#!highlight bash
#!/bin/sh

# A POSIX variable
OPTIND=1 # Reset in case 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
     ;;
Line 26: Line 115:
}}}
Now all of the remaining arguments are the filenames which followed the optional switches. You can process those with `for i` or `"$@"`.

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
}}}

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:

{{{
# Bash (with set -u)
while [[ ${1+defined} && $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
done
}}}

Of course, a simpler fix would be ''not to use'' `set -u` in the first place; or at least to use it only after the option processing is finished.

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`.

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 in case 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
Line 99: Line 117:
if [ "$1" = -- ]; then shift; fi
[ "$1" = "--" ] && shift
Line 101: Line 121:
}}}

The disadvantage of `getopts` is that it can only handle short options (`-h`) without trickery. It handles `-vf filename` in the expected Unix way, automatically. `getopts` is a good candidate because it is portable and e.g. also works in dash.

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. This can be avoided by using callback functions, but this approach kind of defeats the purpose of using getopts at all.

Here is an example which claims to parse long options with `getopts`. The basic idea is quite simple: just put "-:" into the optstring.  This trick requires a shell which permits the option-argument (i.e. the filename in "-f filename") to be concatenated to the option (as in "-ffilename").  The [[http://pubs.opengroup.org/onlinepubs/9699919799/utilities/getopts.html|POSIX standard]] says there must be a space between them; bash and dash permit the "-ffilename" variant, but one should not rely on this leniency if attempting to write a portable script.
 ''I do not quite understand the point, can you give an example? The example script below accepts --loglevel 5 and --loglevel=5. Support for the second form may be removed. Maybe the usage message is not quite correctly formatted? Anyway there is no concatenation necessary nor supported, neither does the script affect the treatment of short options in any way, does it?'' -- the author of the example

# End of file
}}}

The disadvantage of `getopts` is that it can only handle short options
(`-h`) without trickery. It handles `-vf filename` in the expected
Unix way, automatically. `getopts` is a good candidate because it is
portable and e.g. also works in dash.

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. This can be
avoided by using callback functions, but this approach kind of defeats
the purpose of using getopts at all.

Here is an example which parses long options with `getopts`. The basic
idea is quite simple: just put "-:" into the optstring. This trick
requires a shell which permits the option-argument (i.e. the filename
in "-f filename") to be concatenated to the option (as in
"-ffilename"). The
[[http://pubs.opengroup.org/onlinepubs/9699919799/utilities/getopts.html|POSIX
standard]] says there must be a space between them; bash and dash
permit the "-ffilename" variant, but one should not rely on this
leniency if attempting to write a portable script.
Line 115: Line 157:
Line 116: Line 159:
while getopts "$optspec" optchar; do
  case "${optchar}" in
    -)
  
case "${OPTARG}" in
        loglevel)
          val="${!
OPTIND}"; OPTIND=$(( $OPTIND + 1 ))
         echo "Parsing option: '--${OPTARG}', value: '${val}'" >&2;
          ;;
  
loglevel=*)
         val=${OPTARG#*=}
         opt=${OPTARG%=$val}
         echo "Parsing option: '--${opt}', value: '${val}'" >&2
          ;;
      esac;;
    h)
  
echo "usage: $0 [--loglevel[=]<value>]" >&2
   exit 2
      ;;
  esac

while getopts "$optspec" optchar
do
    case "${optchar}" in
 -)
 
case "${OPTARG}" in
       loglevel)
    eval val="\$${
OPTIND}"; OPTIND=$(( $OPTIND + 1 ))
   echo "Parsing option: '--${OPTARG}', value: '${val}'" >&2
    ;;
 
loglevel=*)
   val=${OPTARG#*=}
   opt=${OPTARG%=$val}
   echo "Parsing option: '--${opt}', value: '${val}'" >&2
    ;;
     esac
     ;;
 h)
 
echo "usage: $0 [--loglevel[=]<value>]" >&2
  exit 2
     ;;
    esac
Line 137: Line 183:
 ''Even if we ignore the portability issue, are you sure this is an ''improvement'' over the manual loop in the first example? The manual loop is much simpler. Also, why is it checking `OPTERR` which is something the programmer sets, not something set by `getopts`?'' - GreyCat

 ''Thre is nore than one way to do a thing :-) Since this is section is titled "getopts", I guess it fits here. Why not let people decide what suits their needs, without bias. The example works in the more stricly POSIX-compliant Debian Almquist shell (apart from the variable indirection ${!OPTIND}), so I guessed that portability would be ok. The OPTERR-conditional behavior was just non-essential sugar, I removed it for clarity of the example.'' -- the author of the example
Line 143: Line 186:
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
Another approach is to check options with `if` statements "on demand".
A function like this one may be useful:

{{{#!highlight bash
#!/bin/bash

HaveOpt ()
{
   local needle=$1
Line 156: Line 196:
  done
  return 1

    while [[ $1 == -* ]]
    do
 # By convention, "--" means end of options.
 case "$1" in
     --) return 1 ;;
     $needle) return 0 ;;
 esac

 shift
    done

    return 1
Line 159: Line 210:
if HaveOpt --quick "$@"; then echo "Option quick is set"; fi
HaveOpt --quick "$@" && echo "Option quick is set"

# End of
file
Line 172: Line 226:
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.
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.
Line 208: Line 267:
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.
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.
Line 211: Line 275:
 wooledg@wooledg:~/process-getopt-1.6$ set -- one 'rm -rf /' 'foo;bar' "'"
 wooledg@wooledg:~/process-getopt-1.6$ call_getopt "$@"
 ~/process-getopt-1.6$ set -- one 'rm -rf /' 'foo;bar' "'"
 ~/process-getopt-1.6$ call_getopt "$@"
Line 214: Line 278:
 }}}
 . ''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]]
}}}
 . ''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]] ''
Line 220: Line 284:
CategoryShell '' CategoryShell ''

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.

Manual loop

This approach handles any arbitrary set of options, because you're writing the parser yourself. For 90% of programs, it may suffice.

Here's an example that will handle a combination of short (-v, -h) and long (--verbose, --help) options; and also style --verbose=LEVEL.

Main () {

  • help="" # Reset all option variables that might be set later verbose="" verbose_level="" while : do
    • case "$1" in
      • -h | --help)
        • help="help" shift # Remove from argument list # Now, e.g. call Help() function ;;
        --verbose=*)
        • # If you want to support --option=VALUE style verbose="verbose" verbose_level=${1#*=} # Delete everyting up till "=" shift ;;
        -v | --verbose)
        • verbose="verbose" verbose_level=1 shift # Support "--verbose LEVEL" by reading next argument. case "$1" in
          • [0-9]) verbose_level=$1
            • shift ;;
          esac ;;
        --) # End of all options
        • shift break ;;
        -*)
        • Warn "WARN: Unknown option: $1" shift ;;
      • ) # no more options. Stop while loop
        • break ;;
      esac
    done # Suppose, some options are required. Check that we got them

    [ "$verbose" ] || Die "ERROR: option --verbose not given. See --help"

    # <Rest of your code here>

}

Main "$@"

# End of file }}}

What is not possible to do with pure shell approach, is to try to parse separate options concatenaed together. Like like -xvf which would understoos as -x -v -f. This could be possible with lot of effort, but in practise it wouldn't be worth it.

You may be interested in knowing that some Bash programmers like to write this at the beginning of their scripts to guard against unused variables:

    set -u     # or, set -o nounset

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:

   1 #!/bin/sh
   2 
   3 # A POSIX variable
   4 OPTIND=1         # Reset in case getopts has been used previously in the shell.
   5 
   6 while getopts "h?vf:" opt; do
   7     case "$opt" in
   8         h|\?)
   9             show_help
  10             exit 0
  11             ;;
  12         v) verbose=1
  13             ;;
  14         f) output_file=$OPTARG
  15             ;;
  16     esac
  17 done
  18 
  19 shift $((OPTIND-1))
  20 
  21 [ "$1" = "--" ] && shift
  22 
  23 echo "verbose=$verbose, output_file='$output_file', Leftovers: $@"
  24 
  25 # End of file

The disadvantage of getopts is that it can only handle short options (-h) without trickery. It handles -vf filename in the expected Unix way, automatically. getopts is a good candidate because it is portable and e.g. also works in dash.

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. This can be avoided by using callback functions, but this approach kind of defeats the purpose of using getopts at all.

Here is an example which parses long options with getopts. The basic idea is quite simple: just put "-:" into the optstring. This trick requires a shell which permits the option-argument (i.e. the filename in "-f filename") to be concatenated to the option (as in "-ffilename"). The [[http://pubs.opengroup.org/onlinepubs/9699919799/utilities/getopts.html|POSIX standard]] says there must be a space between them; bash and dash permit the "-ffilename" variant, but one should not rely on this leniency if attempting to write a portable script.

   1 #!/bin/bash
   2 # Uses bash extensions.  Not portable as written.
   3 
   4 optspec=":h-:"
   5 
   6 while getopts "$optspec" optchar
   7 do
   8     case "${optchar}" in
   9         -)
  10             case "${OPTARG}" in
  11               loglevel)
  12                   eval val="\$${OPTIND}"; OPTIND=$(( $OPTIND + 1 ))
  13                   echo "Parsing option: '--${OPTARG}', value: '${val}'" >&2
  14                   ;;
  15               loglevel=*)
  16                   val=${OPTARG#*=}
  17                   opt=${OPTARG%=$val}
  18                   echo "Parsing option: '--${opt}', value: '${val}'" >&2
  19                   ;;
  20             esac
  21             ;;
  22         h)
  23             echo "usage: $0 [--loglevel[=]<value>]" >&2
  24             exit 2
  25             ;;
  26     esac
  27 done

Silly repeated brute-force scanning

Another approach is to check options with if statements "on demand". A function like this one may be useful:

   1 #!/bin/bash
   2 
   3 HaveOpt ()
   4 {
   5     local needle=$1
   6     shift
   7 
   8     while [[ $1 == -* ]]
   9     do
  10         # By convention, "--" means end of options.
  11         case "$1" in
  12             --)      return 1 ;;
  13             $needle) return 0 ;;
  14         esac
  15 
  16         shift
  17     done
  18 
  19     return 1
  20 }
  21 
  22 HaveOpt --quick "$@" && echo "Option quick is set"
  23 
  24 # End of file

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

bhepple suggests the use of 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:

     ~/process-getopt-1.6$ set -- one 'rm -rf /' 'foo;bar' "'"
     ~/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 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. -bhepple


CategoryShell

BashFAQ/035 (last edited 2024-02-26 07:51:38 by larryv)