#!/bin/sh

teststr=$1

echo "$teststr" | awk -v dummy="'" '
function url_encode(rawURL) {
    cleanURL=""
    nonURLPos=match(rawURL,/[^[:alnum:]]/)
    while( nonURLPos > 0 ) {
        rawChar=substr(rawURL,nonURLPos,1)
        replaceChar=sprintf("%02x", dummy rawChar)
        cleanURL = cleanURL substr(rawURL,1,nonURLPos-1) "%" replaceChar
        rawURL = substr(rawURL,nonURLPos+1)
        nonURLPos=match(rawURL,/[^[:alnum:]]/)
    }
    cleanURL = cleanURL rawURL
    return cleanURL
}
{ print(url_encode($0))}
'

BEGIN {
    for (i = 0 ; i <= 255 ; i++) {
            t = sprintf("%c", i)
            _ord_[t] = sprintf("%x", i)
        }
}
function ord(str,    c)
{
    c = substr(str, 1, 1)
    return _ord_[c]
}

BEGIN {
    for (i = 0 ; i <= 255 ; i++) {
            t = sprintf("%c", i)
            _ord_[t] = sprintf("%x", i)
        }
}

function ord(str,    c)
{
    c = substr(str, 1, 1)
    return _ord_[c]
}

function url_encode(rawURL) {
    cleanURL=""
    do {
        nonURLPos = match(rawURL,/[^a-zA-Z0-9_.\:\/]/)
        if (nonURLPos > 0) {
           rawChar = substr(rawURL,nonURLPos,1)
           replaceChar = ord(rawChar)
           cleanURL = cleanURL substr(rawURL,1,nonURLPos-1) "%" replaceChar
           rawURL = substr(rawURL,nonURLPos+1)
        }
    } while (nonURLPos > 0)

    cleanURL = cleanURL rawURL
    return cleanURL
}

echo "$teststr" | perl -nwe '
        foreach my $char (split //) {
                if ($char =~ /([[:alnum:]\/.:\n])/) {
                        print $char;
                } else {
                        printf "%%%02x",ord($char);
                }
        }
'

echo "$teststr" | perl -nwe '
        print join "", map { $_ =~ /([[:alnum:]\/.:\n])/ ? $_ : sprintf("%%%02x",ord($_)) } split //;
'

:
##########################################################################
# Title      :    urlencode - encode URL data
# Author     :    Heiner Steven (heiner.steven@odn.de)
# Date       :    2000-03-15
# Requires   :    awk
# Categories :    File Conversion, WWW, CGI
# SCCS-Id.   :    @(#) urlencode    1.4 06/10/29
##########################################################################
# Description
#    Encode data according to
#        RFC 1738: "Uniform Resource Locators (URL)" and
#        RFC 1866: "Hypertext Markup Language - 2.0" (HTML)
#
#    This encoding is used i.e. for the MIME type
#    "application/x-www-form-urlencoded"
#
# Notes
#    o    The default behaviour is not to encode the line endings. This
#    may not be what was intended, because the result will be
#    multiple lines of output (which cannot be used in an URL or a
#    HTTP "POST" request). If the desired output should be one
#    line, use the "-l" option.
#
#    o    The "-l" option assumes, that the end-of-line is denoted by
#    the character LF (ASCII 10). This is not true for Windows or
#    Mac systems, where the end of a line is denoted by the two
#    characters CR LF (ASCII 13 10).
#    We use this for symmetry; data processed in the following way:
#        cat | urlencode -l | urldecode -l
#    should (and will) result in the original data
#
#    o    Large lines (or binary files) will break many AWK
#        implementations. If you get the message
#        awk: record `...' too long
#         record number xxx
#    consider using GNU AWK (gawk).
#
#    o    urlencode will always terminate it's output with an EOL
#        character
#
# Thanks to Stefan Brozinski for pointing out a bug related to non-standard
# locales.
#
# See also
#    urldecode
##########################################################################

PN=`basename "$0"`            # Program name
VER='1.4'

: ${AWK=awk}

Usage () {
    echo >&2 "$PN - encode URL data, $VER
usage: $PN [-l] [file ...]
    -l:  encode line endings (result will be one line of output)

The default is to encode each input line on its own."
    exit 1
}

Msg () {
    for MsgLine
    do echo "$PN: $MsgLine" >&2
    done
}

Fatal () { Msg "$@"; exit 1; }

set -- `getopt hl "$@" 2>/dev/null` || Usage
[ $# -lt 1 ] && Usage            # "getopt" detected an error

EncodeEOL=no
while [ $# -gt 0 ]
do
    case "$1" in
        -l)    EncodeEOL=yes;;
    --)    shift; break;;
    -h)    Usage;;
    -*)    Usage;;
    *)    break;;            # First file name
    esac
    shift
done

LANG=C    export LANG
$AWK '
    BEGIN {
    # We assume an awk implementation that is just plain dumb.
    # We will convert an character to its ASCII value with the
    # table ord[], and produce two-digit hexadecimal output
    # without the printf("%02X") feature.

    EOL = "%0A"        # "end of line" string (encoded)
    split ("1 2 3 4 5 6 7 8 9 A B C D E F", hextab, " ")
    hextab [0] = 0
    for ( i=1; i<=255; ++i ) ord [ sprintf ("%c", i) "" ] = i + 0
    if ("'"$EncodeEOL"'" == "yes") EncodeEOL = 1; else EncodeEOL = 0
    }
    {
    encoded = ""
    for ( i=1; i<=length ($0); ++i ) {
        c = substr ($0, i, 1)
        if ( c ~ /[a-zA-Z0-9.-]/ ) {
        encoded = encoded c        # safe character
        } else if ( c == " " ) {
        encoded = encoded "+"    # special handling
        } else {
        # unsafe character, encode it as a two-digit hex-number
        lo = ord [c] % 16
        hi = int (ord [c] / 16);
        encoded = encoded "%" hextab [hi] hextab [lo]
        }
    }
    if ( EncodeEOL ) {
        printf ("%s", encoded EOL)
    } else {
        print encoded
    }
    }
    END {
        #if ( EncodeEOL ) print ""
    }
' "$@"