Batch download snippets from http://codesnippets.joyent.com and convert them to text files using man textutil (available on Mac OS X 10.4 or later).

Note: Old snippet versions will be automatically replaced by the downloaded snippets without a backup!

Author: jv
License: The MIT License, Copyright (c) 2008 jv

Usage:

# usage: bds [-p num] [-t tag] [-u user] tag
bds vim
bds -p 1280
bds -u jvs
bds -t plistbuddy
bds -t tar
bds -t ipfw -u jvs

#!/opt/local/bin/bash

# "batch download snippets" from http://codesnippets.joyent.com and
# convert them to text files using man textutil (available on Mac OS X 10.4 or later).
#
# Note: Old snippet versions will be automatically replaced by the downloaded snippets without a backup!
#      An alternative to man textutil is html2text, http://www.mbayer.de/html2text/ (which is available via MacPorts).
#
# Author: jv
# License: The MIT License, http://www.opensource.org/licenses/mit-license.php
# Copyright (c) 2008 jv
#
# cat /usr/local/bin/bds
#
# usage: bds [-p num] [-t tag] [-u user] tag


declare BaseURL='http://codesnippets.joyent.com'
declare download_dir="${HOME}/Desktop/Snippets"

# make sure there is no trailing slash
BaseURL="${BaseURL%/}"
download_dir="${download_dir%/}"

declare BasePostURL="${BaseURL}/posts/show"
declare BaseTagURL="${BaseURL}/tag"
declare BaseUserURL="${BaseURL}/user"

# make sure there is no trailing slash
BasePostURL="${BasePostURL%/}"
BaseTagURL="${BaseTagURL%/}"
BaseUserURL="${BaseUserURL%/}"


# man textutil
declare InputEncoding='utf-8'
declare OutputEncoding='utf-8'

export IFS=$' \t\n'


# function to download a single post specified by a post number: bds -p num
# cf. snippet, http://codesnippets.joyent.com/posts/show/1282

function snippet() {

   declare NL OPWD file outputfile postnum title url

   if [[ "${1//[[:digit:]]/}" != "" ]]; then echo "Argument error. No positive integer: ${1}"; return 1; fi

   postnum="${1}"
   url="${BasePostURL}/${postnum}"
   download_dir="${download_dir}/single-downloads"
   /bin/mkdir -p "${download_dir}"
   OPWD="${PWD}"
   cd "${download_dir}"
   /usr/bin/curl -L -O -s --max-time 25 "${url}" || exit 1    # download snippet web page
   file="${download_dir}/${url##*/}"
   trap '/bin/rm -f "${file}"; exit 0' 0 1 2 13 15

   # get title of downloaded web page
   #title="$(/usr/bin/sed -E -n -e '/<[tT][iI][tT][lL][eE]>/{s/^.*<[tT][iI][tT][lL][eE]>(.*)<\/[tT][iI][tT][lL][eE]>.*$/\1/p;q;}' "${file}" | \
   #         /usr/bin/sed -E -e 's/\[[^][:space:]]*\]//g')"    # delete [xxx] tag elements of title

   title="$(/usr/bin/egrep -m 1 -io '<title>.*</title>' "${file}" | /usr/bin/sed -E -e 's/^<title>[[:space:]]*|[[:space:]]*<\/title>$//g' \
             -e 's/\[[^][:space:]]*\]//g')"    # delete [xxx] tag elements of title


   title="${title//CodeSnippets:/}"
   title="${title//\//:}"
   title="${title// /_}"
   title="${title//[[:cntrl:]]/}"
   title="${title%"${title##*[!_]}"}"   # remove trailing underscores

   if [[ $title == '_CodeDrive_Snippets_courtesy_of_Peter_Coopers_handy_little_app' ]] || [[ -z "$title" ]]; then
      printf "\e[0K\e[31m%s\e[0m:  %s\n" "couldn't access" "${url}"
      /bin/rm "${file}"
      return 1
   fi

   outputfile="${download_dir}/${postnum}_${title}.txt"
   #outputfile="${download_dir}/${title}.txt"  # without post number prefix
   #outputfile="${outputfile//__/_}"  # uniq underscores

   printf "\n\e[0K\e[1;30m%s\e[0m:  %s\n\n" "saved as" "${outputfile}"

   /usr/bin/textutil -output "${outputfile}" -convert txt -inputencoding "${InputEncoding}" -encoding "${OutputEncoding}" "${file}"
   /bin/rm "${file}"

   # escape backslashes
   # man bash 2>/dev/null | less -p 'Each command in a pipeline'
   #outputfile="$(printf "%q" "${outputfile}")"  # cf. help printf
   outputfile="${outputfile//\\/\\\\}"

   NL=$'\\\n'

cat <<EOF | /bin/ed -s "${outputfile}"
H
,g/Snippets is a public source code repository/1,/Snippets is a public source code repository/d
,g/You need to create an account or log in to post comments to this site//You need to create an account or log in to post comments to this site/,\$d
,g|(See related posts)$|s|.See related posts.|${NL}${NL}|
,g|^to.* by.* on .*[[:digit:]]$|s|^to\(.*\) by\(.*\) on \(.*[[:digit:]]\)$|${NL}${NL}Author:\2${NL}Date: \3${NL}URL: ${url}${NL}Tags:\1${NL}|
,g|^Comments on this post$|s|\(Comments on this post\)|${NL}\1:|
,g| posts on .* at |s|\(.* posts on .* at .*\)|${NL}\1:|
w
EOF

# additional ed commands
# delete line numbers
# ,g|^[[:space:]]*[[:digit:]]\{1,\}[[:space:]]\{1,3\}|s|^[[:space:]]*[[:digit:]]\{1,\}[[:space:]]\{1,3\}\(.*\)$|\1|
# delete range of lines
# 4,11d


   cd "${OPWD}"
   return 0

}



#----------------------------------------- end of function snippet



declare pflag tflag uflag
declare cnt count dir_name file no_posts_check NL OPWD outputfile postnum tagsite title url urls website 

if [[ $# -eq 0 ]]; then 
   printf "%s\n%s\n" 'No arguments given!' "Usage: ${0##*/} [-p num] [-t tag] [-u user] tag" 1>&2
   exit 1
fi


while getopts ":p:t:u:" option
do
  case $option in
    p) pflag="$OPTARG" ;;
    t) tflag="$OPTARG" ;;
    u) uflag="$OPTARG" ;;
    [?]) printf "%s\n%s\n" 'Argument error!' "Usage: ${0##*/} [-p num] [-t tag] [-u user] tag" 1>&2; exit 1;;
    *) ;;
  esac
done

shift $(($OPTIND - 1))


if [[ $# -eq 1 ]]; then

   dir_name="${1}"
   tagsite="${BaseTagURL}/${1}"

elif [[ $# -gt 1 ]]; then

   printf "%s\n%s\n" 'Too many arguments!' "Usage: ${0##*/} [-p num] [-t tag] [-u user] tag" 1>&2
   exit 1

elif [[ -n "${pflag}" ]]; then
   snippet "${pflag}"
   exit 0

elif [[ -n "${tflag}" ]] && [[ -n "${uflag}" ]]; then

   dir_name="${tflag}-${uflag}"
   tagsite="${BaseUserURL}/${uflag}/tag/${tflag}"

elif [[ -n "${tflag}" ]]; then

   dir_name="${tflag}"
   tagsite="${BaseTagURL}/${tflag}"

elif [[ -n "${uflag}" ]]; then

   dir_name="${uflag}"
   tagsite="${BaseUserURL}/${uflag}"

else

   printf "%s\n%s\n" 'Argument error!' "Usage: ${0##*/} [-p num] [-t tag] [-u user] tag" 1>&2
   exit 1

fi


tagsite="${tagsite%/}"

#echo $dir_name
#echo $tagsite

count=1
cnt=0
curl_max_time=20
website=''
no_posts_check=''
NL=$'\\\n'
download_dir="${download_dir}/${dir_name//\//:}"
download_dir="${download_dir%/}"
/bin/mkdir -p "${download_dir}"
OPWD="${PWD}"
cd "${download_dir}"

# print download directory
printf "\n\e[0K\e[1;30m%s\e[0m:  %s\n\n" "download directory" "${download_dir}"


while [[ -z "${no_posts_check}" ]]; do

   # download website of the form: 
   # http://somewebsite.com/tag/bash/1,
   # http://somewebsite.com/user/name/1 or 
   # http://somewebsite.com/user/name/tag/bash/1

   website="$(/usr/bin/curl -L -s --max-time $curl_max_time "${tagsite}/${count}" )"

   if [[ $? -ne 0 ]]; then 
      printf "\e[0K\e[31m%s\e[0m:  %s\n" "curl_max_time ${curl_max_time}" "${tagsite}/${count}"
      exit 1
   fi

   #if [[ -n "$(printf "%s" "${website}" | /usr/bin/egrep -o 'Application error \(Apache\)')" ]]; then 
      #no_posts_check='Application error (Apache)'
      #printf "\e[0K\e[31m%s\e[0m:  %s\n" "no further posts" "${no_posts_check}"
   #fi

   if [[ -n "$(printf "%s" "${website}" | /usr/bin/egrep -o '>No posts<')" ]]; then 
      no_posts_check='>No posts<'
      #printf "\e[0K\e[31m%s\e[0m:  %s\n" "no further posts" "${no_posts_check}"
   fi

: <<-'COMMENT'

   # works for Bash 3.0 or later
   if [[ "${website}" =~ '>No posts<' ]]; then 
      no_posts_check="${BASH_REMATCH[0]}"
      #printf "\e[0K\e[31m%s\e[0m:  %s\n" "no further posts" "${no_posts_check}"
   fi

COMMENT


   if [[ -z "${no_posts_check}" ]]; then

      # extract relevant post URLs
      #urls=( $(printf "%s\n" "${website}" | /usr/bin/sed -E -n -e "s|^.* href=\"(/posts/show/[[:digit:]]+)\".*$|${BaseURL}\1|p;g") )
      urls=( $(printf "%s\n" "${website}" | /usr/bin/egrep -o 'href="/posts/show/[[:digit:]]+"' | /usr/bin/sed -E -n -e "s|href=\"(/posts/show/[[:digit:]]+)\"|${BaseURL}\1|p;g") )

      for ((i=0; i < "${#urls[@]}"; i++)); do

         url="${urls[${i}]}"

         postnum="${url##*/}"
         file="${download_dir}/${postnum}"
         trap '/bin/rm -f "${file}"; exit 0' 0 1 2 13 15

         /usr/bin/curl -L -O -s --max-time $curl_max_time "${url}"

         if [[ $? -ne 0 ]]; then 
            printf "\e[0K\e[31m%s\e[0m:  %s\n" "curl_max_time ${curl_max_time}" "${url}"
            continue
         fi
 

         # get title of downloaded web page
         #title="$(/usr/bin/sed -E -n -e '/<[tT][iI][tT][lL][eE]>/{s/^.*<[tT][iI][tT][lL][eE]>(.*)<\/[tT][iI][tT][lL][eE]>.*$/\1/p;q;}' "${file}" | \
         #    /usr/bin/sed -E -e 's/\[[^][:space:]]*\]//g')"    # delete [xxx] tag elements of title

         title="$(/usr/bin/egrep -m 1 -io '<title>.*</title>' "${file}" | /usr/bin/sed -E -e 's/^<title>[[:space:]]*|[[:space:]]*<\/title>$//g' \
                -e 's/\[[^][:space:]]*\]//g')"    # delete [xxx] tag elements of title


         title="${title//CodeSnippets:/}"
         title="${title//\//:}"
         title="${title// /_}"
         title="${title//[[:cntrl:]]/}"
         title="${title%"${title##*[!_]}"}"   # remove trailing underscores

         #printf "%s\n" "${title}"

         if [[ $title == '_CodeDrive_Snippets_courtesy_of_Peter_Coopers_handy_little_app' ]] || [[ -z "$title" ]]; then
            printf "\e[0K\e[31m%s\e[0m:  %s\n" "couldn't access" "${url}"
            /bin/rm "${file}"
            continue
         fi

         outputfile="${download_dir}/${postnum}_${title}.txt"
         #outputfile="${download_dir}/${title}.txt"  # without post number prefix
         #outputfile="${outputfile//__/_}"  # uniq underscores

         let cnt++
         printf "\e[0K\e[1;32m%-6s\e[0m  %s\n" "${cnt}" "${outputfile##*/}"

         /usr/bin/textutil -output "${outputfile}" -convert txt -inputencoding "${InputEncoding}" -encoding "${OutputEncoding}" "${file}"

         /bin/rm "${file}"


         # escape backslashes
         # man bash 2>/dev/null | less -p 'Each command in a pipeline'
         #outputfile="$(printf "%q" "${outputfile}")"  # cf. help printf
         outputfile="${outputfile//\\/\\\\}"

# edit $outputfile in-place with man ed
# first delete lines at the beginning & end,
# then remove the string 'See related posts' and add some newlines with $NL,
# then convert the line 'to...by...on' to line 'Author:...', line 'Date:...', line 'URL:...' and line 'Tags:...'
# and finally the last two ed commands insert two further newlines with $NL

cat <<EOF | /bin/ed -s "${outputfile}"
H
,g/Snippets is a public source code repository/1,/Snippets is a public source code repository/d
,g/You need to create an account or log in to post comments to this site//You need to create an account or log in to post comments to this site/,\$d
,g|(See related posts)$|s|.See related posts.|${NL}${NL}|
,g|^to.* by.* on .*[[:digit:]]$|s|^to\(.*\) by\(.*\) on \(.*[[:digit:]]\)$|${NL}${NL}Author:\2${NL}Date: \3${NL}URL: ${url}${NL}Tags:\1${NL}|
,g|^Comments on this post$|s|\(Comments on this post\)|${NL}\1:|
,g| posts on .* at |s|\(.* posts on .* at .*\)|${NL}\1:|
w
EOF

# additional ed commands
# delete line numbers
# ,g|^[[:space:]]*[[:digit:]]\{1,\}[[:space:]]\{1,3\}|s|^[[:space:]]*[[:digit:]]\{1,\}[[:space:]]\{1,3\}\(.*\)$|\1|
# delete range of lines
# 4,11d


      done  # for

      let count++

   fi

done   # while


   cd "${OPWD}"


exit 0

Author: jv
License: The MIT License, Copyright (c) 2008 jv
Description: a basic regex-based command line parser for use in bash scripts (Mac OS X); an alternative to the builtin getopts command (cf. help getopts); use at your own risk
Usage: /path/to/script_with_cmdparser -a -b -c -f file
Related links: Process positional parameters non-destructively in Bash and ws - search the web from the command line (an example of using cmdparser)

#!/bin/bash

export PATH=/usr/bin:/bin:/usr/sbin:/sbin
export IFS=$' \t\n'

# create a fake command line
set -- -abcc -c -zz -flag1="" -flag2=arg -flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' -flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces  / * + ` \ ! ' -flag9 ~/Desktop/*.txt filename1 filename2 filename3

#set -- -abcc -c -zz -flag1="" -flag2=arg -flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' -flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces / * + ` \ ! ' -flag9 '~/Desktop/*.txt' filename1 filename2 filename3


printf "%s\n" "$@" | nl
#printf "%s" "$@"$'\n' | nl
#printf "%s" "${@/%/ }" | nl


: <<-'COMMENT'

# copy & paste examples for the command line

echo "filename1" "filename2" "filename3" | ~/Desktop/cmdparser.txt -abcc -c -zz -flag1 arg -flag2=arg --flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' --flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces / * + ` \ ! ' -flag9 ~/Desktop/*.txt -

echo "filename1" "filename2" "filename3" | ~/Desktop/cmdparser.txt -abcc -c -zz -flag1 arg -flag2=arg --flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' --flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces / * + ` \ ! ' -flag9 '~/Desktop/*.txt' -

COMMENT



# cmdparser

usage="usage: $(/usr/bin/basename "$0") [-a] [-b] [-c] [-cc] [-zz] [-flag1 arg] [-flag2 'arg1 arg2 ...'] [-flag3=arg] [-flag4=\"arg1 arg2\"] ..."


# define the names of flags as a regular expression
# flags are command line options that require arguments

flags="(flag1|flag2|flag3|flag4|flag5|flag6|flag7|flag8|flag9)"


# define the names of switches as a regular expression
# Switches are command line options that do not take arguments.
# Make sure multi-char switches precede single-char switches in the regular expression.
# Note that the regular expression contains neither the special read-from-stdin switch "-" 
# nor the special end-of-options switch "--".

switches="(cc|zz|a|b|c)"  


declare flag1 flag2 flag3 flag4 flag5 flag6 flag7 flag8 flag9                # flags
declare -i a=0 b=0 c=0 cc=0 zz=0                                            # switches
                         
declare argstr argvar argvar_escaped char flagvar optstr piped pipedstr       # script variables
declare -i optid pipedvar

# piped="piped" will be used for variable creation 
# example: piped="piped"; pipedstr="piped arg"; eval $piped='"$(echo "$pipedstr")"'; echo "$piped"

piped="piped"

# default value is set to "no pipe"
pipedvar=0
pipedstr=""

# if /dev/stdin has a size greater than zero ...
if [[ -s /dev/stdin ]]; then pipedstr="$(</dev/stdin)"; fi 

if [[ $# -eq 0 ]] && [[ -z "$pipedstr" ]]; then
  printf "\n%s\n\n%s\n\n" 'No arguments specified!' "$usage" 1>&2
  exit 1
fi 

if [[ $# -eq 0 ]] && [[ -n "$pipedstr" ]]; then
  eval $piped='"${pipedstr}"'  
  pipedvar=1
fi 

# if there are command line arguments ...
# Note that $pipedvar may still be set to 1 below if the special read-from-stdin switch "-" is given.

if [[ $pipedvar -eq 0 ]]; then

   optstr=" "  
   optid=0

   while [[ -n "$optstr" ]]; do     

      # try to extract valid flags or switches from positional parameter $1
      # $1 gets shifted afterwards (cf. help shift)

      optstr="$(printf "%s" "$1" | /usr/bin/egrep -e "^--?${flags}$")"

      if [[ -n "$optstr" ]]; then optid=1; fi
      if [[ -z "$optstr" ]]; then optid=2; optstr="$(printf "%s" "$1" | /usr/bin/egrep -e "^--?${switches}$")"; fi
      if [[ -z "$optstr" ]]; then optid=3; optstr="$(printf "%s" "$1" | /usr/bin/egrep -e "^--?${switches}+$")"; fi
      if [[ -z "$optstr" ]]; then optid=4; optstr="$(printf "%s" "$1" | /usr/bin/egrep -e "^--?(${flags}=.*|${flags}[^[:space:]]+)$")"; fi

      if [[ -z "$optstr" ]]; then  
         if [[ "$1" = "-" ]] && [[ "$@" = "-" ]]; then  
            optid=5
            optstr="-" 
         elif [[ -n "$(printf "%s" "${@/%/ }" | /usr/bin/egrep -e "[[:space:]]--?(${flags}|${switches})")" ]]; then 
            # append a space to each command line argument
            argstr="$(printf "%s" "${@/%/ }")"
            printf "\n%s\x21\n\n%s\n\n%s\n\n" "Undefined non-option string: ${1} is followed by a legal flag or switch" "${argstr}" "$usage" 1>&2
            exit 1
         fi
      fi

      if [[ "$1" = "--" ]]; then shift; break; fi     # -- marks end of options

      if [[ -z "$optstr" ]]; then break; fi     # no further flags or switches to process


      # flag followed by space (example: -f file)
      if [[ $optid -eq 1 ]]; then 

         if [[ -z "$2" ]]; then
            printf "%s\n%s\n" "no argument given to flag: ${1}" "$usage" 1>&2
            exit 1
         fi 

         flagvar="${1##*-}"     # remove leading - or --
         argvar="$2"
         eval $flagvar='"${argvar}"'
         shift 2     # shift positional parameters $1 & $2 (that is, a flag plus its argument)
         continue

      # single switch (example: -a)
      elif [[ $optid -eq 2 ]]; then
         flagvar="${1##*-}"
         eval $flagvar='"1"'
         shift
         continue
  
      # combined switch (example: -abcc)
      elif [[ $optid -eq 3 ]]; then
         flagvar="${1##*-}"
         while [[ -n "$flagvar" ]]; do
            char="$(printf "%s" "$flagvar" | /usr/bin/sed -E "s/^${switches}.*$/\1/")"
            eval $char='"1"'
            flagvar="$(printf "%s" "$flagvar" | /usr/bin/sed -E "s/^${switches}//")"
         done
         shift
         continue

      # flag without following space (example: -ffile)
      elif [[ $optid -eq 4 ]]; then 

: <<-'COMMENT'

         argvar="$(printf "%s" "$1" | /usr/bin/sed -E "s/^\-\-?${flags}\=?//")"

         argvar2="${argvar//\\\\/\\\\}"       # escape \  (for Bash version 2.05b.0(1)-release)
         #argvar2="${argvar//\\/\\\\}"          # escape \  

         flagvar="${1%${argvar2}}"          # remove escaped $argvar string
         flagvar="${flagvar%=}"             # remove trailing =
         flagvar="${flagvar##*-}"           # remove leading - or --
         eval $flagvar='"${argvar}"'
         shift
         continue

COMMENT

         # alternative: no string escaping necessary
         #argvar="$(printf "%s" "$1" | /usr/bin/sed -E "s/^\-\-?${flags}\=?//")"
         #flagvar="$(printf "%s" "$1" | /usr/bin/sed -E -n -e "s/^\-\-?${flags}\=?.*$/\\1/p")"

         argvar="$(printf "%s" "${1##*-}" | /usr/bin/sed -E "s/^${flags}\=?//")"
         flagvar="$(printf "%s" "${1##*-}" | /usr/bin/sed -E -n -e "s/^${flags}\=?.*$/\\1/p")"

         eval $flagvar='"${argvar}"'
         shift
         continue


      # the special read-from-stdin switch "-"
      elif [[ $optid -eq 5 ]]; then 
         pipedvar=1
         eval $piped='"${pipedstr}"'
         shift
         break

      fi

      # remove positional parameter $1 from "$@"
      shift

   done

fi   # if [[$pipedvar -eq 0 ]]; then ...


echo 

printf "%s\t%s\n" "a:" "${a}"
printf "%s\t%s\n" "b:" "${b}"
printf "%s\t%s\n" "c:" "${c}"
printf "%s\t%s\n" "cc:" "${cc}"
printf "%s\t%s\n" "zz:" "${zz}"
printf "%s\t%s\n" "flag1:" "${flag1}"
printf "%s\t%s\n" "flag2:" "${flag2}"
printf "%s\t%s\n" "flag3:" "${flag3}"
printf "%s\t%s\n" "flag4:" "${flag4}"
printf "%s\t%s\n" "flag5:" "${flag5}"
printf "%s\t%s\n" "flag6:" "${flag6}"
printf "%s\t%s\n" "flag7:" "${flag7}"
printf "%s\t%s\n" "flag8:" "${flag8}"
printf "%s\t%s\n" "flag9:" "${flag9}"

echo


if [[ $pipedvar -eq 1 ]] && [[ -z "$@" ]]; then 
   echo "remaining string-piped: ${piped}"
else 
   echo "remaining string: ${@}"
fi

echo

if [[ $flag9 == '~/Desktop/*.txt' ]]; then printf "%s\n" ~/Desktop/*.txt | nl; fi

echo

exit 0

The non-destructive version of cmdparser does no modify (the number of) command line arguments ($# and $@):

#!/bin/bash

# create a fake command line
#set -- -abcc -c -zz -flag1="" -flag2=arg$'\n'plus_newline -flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' -flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces / * + ` \ !' -flag9 ~/Desktop/*.txt filename1 filename2 filename3

#set -- -abcc -c -zz -flag1="" -flag2=arg$'\n'plus_newline -flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' -flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces / * + ` \ !' -flag9 '~/Desktop/*.txt' filename1 filename2 filename3


printf "%s\n" "$@" | nl
#printf "%s" "$@"$'\n' | nl
#printf "%s" "${@/%/ }" | nl


: <<-'COMMENT'

# copy & paste examples

echo "filename1" "filename2" "filename3" | ~/Downloads/Mac-OS-X-bash-scripts/bash-cmdparser/cmdparser-non-destructive-1.txt -abcc -c -zz -flag1 arg -flag2=arg$'\n'plus_newline --flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' --flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces / * + ` \ !' -flag9 ~/Desktop/*.txt -

echo "filename1" "filename2" "filename3" | ~/Downloads/Mac-OS-X-bash-scripts/bash-cmdparser/cmdparser-non-destructive-1.txt -abcc -c -zz -flag1 arg -flag2=arg$'\n'plus_newline --flag3="arg" -flag4='arg1=*,arg2=?,arg3=!' -flag5 '(arg1|arg2|arg3)' -flag6 'arg1=ag,arg2=bg,arg3=cg' --flag7 An\ argument\ with\ spaces\! -flag8='Yet another argument with spaces / * + ` \ !' -flag9 '~/Desktop/*.txt' -

COMMENT


echo

echo "Number of positional parameters: ${#}"

echo


# cmdparser

export PATH=/usr/bin:/bin:/usr/sbin:/sbin
export IFS=$' \t\n'


# non-builtin commands used
# cf. man builtin
declare basename=/usr/bin/basename egrep=/usr/bin/egrep sed=/usr/bin/sed


# define the names of flags as a regular expression
# flags are command line options that require arguments

flags="(flag1|flag2|flag3|flag4|flag5|flag6|flag7|flag8|flag9)"


# define the names of switches as a regular expression
# Switches are command line options that do not take arguments.
# Make sure multi-char switches precede single-char switches in the regular expression.
# Note that the regular expression contains neither the special read-from-stdin switch "-" 
# nor the special end-of-options switch "--".

switches="(cc|zz|a|b|c)"  


usage="usage: $(${basename} "$0") [-a] [-b] [-c] [-cc] [-zz] [-flag1 arg] [-flag2 'arg1 arg2 ...'] [-flag3=arg] [-flag4=\"arg1 arg2\"] ..."

declare flag1 flag2 flag3 flag4 flag5 flag6 flag7 flag8 flag9                # flags
declare -i a=0 b=0 c=0 cc=0 zz=0                                            # switches
                         
declare argn argstr argvar argvar_escaped char flagvar optstr piped pipedstr       # script variables
declare -i optid pipedvar

# piped="piped" will be used for variable creation 
# example: piped="piped"; pipedstr="piped arg"; eval $piped='"$(echo "$pipedstr")"'; echo "$piped"

piped="piped"

# default value is set to "no pipe"
pipedvar=0
pipedstr=""

# if /dev/stdin has a size greater than zero ...
if [[ -s /dev/stdin ]]; then pipedstr="$(</dev/stdin)"; fi 

if [[ $# -eq 0 ]] && [[ -z "$pipedstr" ]]; then
  printf "\n%s\n\n%s\n\n" 'No arguments specified!' "$usage" 1>&2
  exit 1
fi 

if [[ $# -eq 0 ]] && [[ -n "$pipedstr" ]]; then
  eval $piped='"${pipedstr}"'  
  pipedvar=1
fi 

# if there are command line arguments ...
# Note that $pipedvar may still be set to 1 below if the special read-from-stdin switch "-" is given

if [[ $pipedvar -eq 0 ]]; then

   optstr=" "  
   optid=0


   # processing one positional parameter at a time without modifying $# or $@
   # Process positional parameters non-destructively in Bash, http://codesnippets.joyent.com/posts/show/1706

   for (( i=1; i <= $#; i++ )); do 


      argn="${@:${i}:1}"     # current positional parameter
                             # "${@:(${i}+1):1}": the positional parameter following the current one
                             # "${@:${i}}": all positional parameters starting with the current one


      if [[ ${argn:0:1} != '-' ]]; then break; fi   # every flag or switch has to have a leading -

      optstr="$(printf "%s" "${argn}" | ${egrep} -e "^--?${flags}$")"

      if [[ -n "$optstr" ]]; then optid=1; fi
      if [[ -z "$optstr" ]]; then optid=2; optstr="$(printf "%s" "${argn}" | ${egrep} -e "^--?${switches}$")"; fi
      if [[ -z "$optstr" ]]; then optid=3; optstr="$(printf "%s" "${argn}" | ${egrep} -e "^--?${switches}+$")"; fi
      if [[ -z "$optstr" ]]; then optid=4; optstr="$(printf "%s" "${argn}" | ${egrep} -e "^--?(${flags}=.*|${flags}[^[:space:]]+)$")"; fi

      if [[ -z "$optstr" ]]; then  
         if [[ "${argn}" = "-" ]] && [[ "${@:${i}}" = "-" ]]; then  
            optid=5
            optstr="-" 

         elif [[ -n "$(printf "%s" "${@:${i}/%/ }" | ${egrep} -e "[[:space:]]--?(${flags}|${switches})")" ]]; then 
            # create argstr by appending a space to each command line argument
            argstr="$(printf "%s" "${@:${i}/%/ }" )"
            printf "\n%s\x21\n\n%s\n\n%s\n\n" "Undefined non-option string: ${argn} is followed by a legal flag or switch" "${argstr}" "$usage" 1>&2
            exit 1
         fi
      fi

      if [[ "${argn}" = "--" ]]; then break; fi     # -- marks end of options

      if [[ -z "$optstr" ]]; then break; fi     # no further flags or switches to process


      # flag followed by space (example: -f file)
      if [[ $optid -eq 1 ]]; then 

         if [[ -z "${@:(${i}+1):1}" ]]; then
            printf "%s\n%s\n" "no argument given to flag: ${argn}" "$usage" 1>&2
            exit 1
         fi 

         flagvar="${argn##*-}"     # remove leading dashes
         argvar="${@:(${i}+1):1}"
         eval $flagvar='"${argvar}"'
         let "i += 1"     # skip argument of current flag in next for loop
         continue

      # single switch (example: -a)
      elif [[ $optid -eq 2 ]]; then
         flagvar="${argn##*-}"
         eval $flagvar='"1"'
         continue
  
      # combined switch (example: -abcc)
      elif [[ $optid -eq 3 ]]; then
         flagvar="${argn##*-}"
         while [[ -n "$flagvar" ]]; do
            char="$(printf "%s" "$flagvar" | ${sed} -E "s/^${switches}.*$/\1/")"
            eval $char='"1"'
            flagvar="$(printf "%s" "$flagvar" | ${sed} -E "s/^${switches}//")"
         done
         continue

      # flag without following space (example: -ffile)
      elif [[ $optid -eq 4 ]]; then 

: <<-'COMMENT'

         argvar="$(printf "%s" "${argn}" | ${sed} -E "s/^\-\-?${flags}\=?//")"

         argvar2="${argvar//\\\\/\\\\}"       # escape \  (for Bash version 2.05b.0(1)-release)
         #argvar2="${argvar//\\/\\\\}"          # escape \  

         flagvar="${argn%${argvar2}}"         # remove escaped $argvar string
         flagvar="${flagvar%=}"               # remove trailing =
         flagvar="${flagvar##*-}"             # remove leading - or --
         eval $flagvar='"${argvar}"'
         continue

COMMENT

        # alternative: no string escaping required
         #argvar="$(printf "%s" "${argn}" | ${sed} -E "s/^\-\-?${flags}\=?//")"
         #flagvar="$(printf "%s" "${argn}" | ${sed} -E -n -e "s/^\-\-?${flags}\=?.*$/\\1/p")"

         argvar="$(printf "%s" "${argn##*-}" | ${sed} -E "s/^${flags}\=?//")"
         flagvar="$(printf "%s" "${argn##*-}" | ${sed} -E -n -e "s/^${flags}\=?.*$/\\1/p")"

         eval $flagvar='"${argvar}"'
         continue


      # the special read-from-stdin switch "-"
      elif [[ $optid -eq 5 ]]; then 
         pipedvar=1
         eval $piped='"${pipedstr}"'
         break

      fi

   done   # for loop

fi   # if [[$pipedvar -eq 0 ]]; then ...


echo 

printf "%s\t%s\n" "a:" "${a}"
printf "%s\t%s\n" "b:" "${b}"
printf "%s\t%s\n" "c:" "${c}"
printf "%s\t%s\n" "cc:" "${cc}"
printf "%s\t%s\n" "zz:" "${zz}"
printf "%s\t%s\n" "flag1:" "${flag1}"
printf "%s\t%s\n" "flag2:" "${flag2}"
printf "%s\t%s\n" "flag3:" "${flag3}"
printf "%s\t%s\n" "flag4:" "${flag4}"
printf "%s\t%s\n" "flag5:" "${flag5}"
printf "%s\t%s\n" "flag6:" "${flag6}"
printf "%s\t%s\n" "flag7:" "${flag7}"
printf "%s\t%s\n" "flag8:" "${flag8}"
printf "%s\t%s\n" "flag9:" "${flag9}"

echo


if [[ $pipedvar -eq 1 ]] && [[ -z "$@" ]]; then 
   echo "remaining string-piped: ${piped}"
else 
   echo "remaining string: ${@}"
fi

echo

echo "Number of positional parameters: ${#}"

echo

if [[ $flag9 == '~/Desktop/*.txt' ]]; then printf "%s\n" ~/Desktop/*.txt | nl; fi

echo

exit 0

Further information:

- Command-line argument
- In the Beginning... was the Command Line
- Handling Command Line Arguments
- Utility Conventions - Utility Argument Syntax (POSIX)
- Utility Conventions - Utility Syntax Guidelines (POSIX)
- GNU coding standards: 4.7 Standards for Command Line Interfaces
- bash-getopts
- Bash Shell my_getopts
- Parsing arguments for your shell script
- Bash: parsing arguments with 'getopts'
- More Power with Bash Getopts
- Getopt and getopts
- Option-ize your shell scripts
- Emulating getopt
- Positional Parameters
- Parsing Command Line Options in Shell Scripts
- Command Line Processing in Cocoa
- ddcli: An Objective-C Command Line Helper
- Arg_parser