Nouveau script pour d=c3=a9coder les ent=c3=aates MIME

Bonjour,

J'ai fait une nouvelle version de mon script. Au lieu de prendre en
paramètre une ou plusieurs chaînes à décoder, il prend en paramètre
un ou plusieurs noms de fichiers contenant les entêtes complets.
On peut aussi lui faire lire les données sur stdin si on ne met pas
de paramètre, ou qu'on met « - » à la place d'un nom de fichier.

Par exemple, si on saisit :

decode_headers <<POTIRON
Message-ID: <c5a9-1@part.org>
From: Arvo =?UTF-8?Q?P=C3=A4rt?= <arvo@part.org>
To: =?Latin1?Q?Fr=E9d=E9ric_Chopin?= <fred@chopin.org>,
=?Latin2?Q?Anton=EDn_Dvo=F8=E1k?= <anton@dvorak.org>
References: <A65R-4d@chopin.org> <c5a7-3@part.org>
<A72Q-5a@chopin.org>
In-Reply-To: <A72Q-5a@chopin.org>
Subject: Re: Going to =?Shift-JIS?B?k4yLngo=?= (Tokyo) =?UTF-8?B?zpEK?=
=?UTF-8?B?zrjOrs69zrEK?= (Athens) and =?ISO-8859-5?Q?=BC=DE=E1=DA?=
=?ISO-8859-5?Q?=D2=D0?= (Moscow)
POTIRON

On obtiendra :

Message-ID: <c5a9-1@part.org>
From: Arvo Pärt <arvo@part.org>
To: Frédéric Chopin <fred@chopin.org>, Antonín Dvořák <anton@dvorak.org>
References: <A65R-4d@chopin.org> <c5a7-3@part.org> <A72Q-5a@chopin.org>
In-Reply-To: <A72Q-5a@chopin.org>
Subject: Re: Going to 東京 (Tokyo) Αθήνα (Athens) and Москва (Moscow)

(Note : le mot POTIRON n'a rien de spécial, c'est juste pour aider
OuiOui à améliorer son karma)


Voici le script. Attention, pour l'adapter au Mac il faudra modifier les
variables CUT, DECODE_QP, DECODE_B64, TR et ICONV définies vers le début
du fichier.

========================================================================
#!/bin/bash
####################################################################
# Name:
# decode_headers
#
# Description:
# Script used for decoding RFC 2047 MIME encoded headers.
#
# Example:
# decode_headers <<POTIRON
# Message-ID: <c5a9-1@part.org>
# From: Arvo =?UTF-8?Q?P=C3=A4rt?= <arvo@part.org>
# To: =?Latin1?Q?Fr=E9d=E9ric_Chopin?= <fred@chopin.org>,
# =?Latin2?Q?Anton=EDn_Dvo=F8=E1k?= <anton@dvorak.org>
# References: <A65R-4d@chopin.org> <c5a7-3@part.org>
# <A72Q-5a@chopin.org>
# In-Reply-To: <A72Q-5a@chopin.org>
# Subject: Re: Going to =?Shift-JIS?B?k4yLngo=?= (Tokyo) =?UTF-8?B?zpEK?=
# =?UTF-8?B?zrjOrs69zrEK?= (Athens) and =?ISO-8859-5?Q?=BC=DE=E1=DA?=
# =?ISO-8859-5?Q?=D2=D0?= (Moscow)
# POTIRON
# ->
# Message-ID: <c5a9-1@part.org>
# From: Arvo Pärt <arvo@part.org>
# To: Frédéric Chopin <fred@chopin.org>, Antonín Dvořák <anton@dvorak.org>
# References: <A65R-4d@chopin.org> <c5a7-3@part.org> <A72Q-5a@chopin.org>
# In-Reply-To: <A72Q-5a@chopin.org>
# Subject: Re: Going to 東京 (Tokyo) Αθήνα (Athens) and Москва (Moscow)
#
# Known bugs:
# Can't handle encoded-words which are not clearly separated
# from non-blank characters, e.g. in comments: (=?...?.?...?=)
####################################################################

#
# The following constants should be adapted for your system.
# For example, the executable base64 could be at /usr/local/bin instead
# of /usr/bin, and it could require parameter -D instead of -d.
#
CUT="/usr/bin/cut"
DECODE_QP="/usr/bin/qprint -d"
DECODE_B64="/usr/bin/base64 -d"
TR="/usr/bin/tr"
ICONV="/usr/bin/iconv"

#
# This script uses the following global variables
# VARY
# used when a function has to return a string and not only a number
# DECODED_LINE
# The current state of a header being currently decoded
# STATUS
# Has three possible values:
# - "none" at the beginning of a new header
# - "decoded-word" after a correctly decoded MIME part
# - "normal" after any other string
#
VARY=""
DECODED_LINE=""
STATUS="none"

#
# Function: usage
#
usage()
{
printf "Usage: %s [OPTION...] [FILE...]\n" $0
printf "Decode headers from given files for RFC 2047 MIME encodings.\n"
printf "\n"
printf "With no FILE, or when FILE is -, read standard input.\n"
printf "\n"
printf " -h, --help Give this help list\n"
exit 1
}

#
# Function: decode_word
#
# Description:
# Check that the parameter is an encoded word, then decode it and
# convert it to UTF-8.
#
# Parameter:
# A single (possibly MIME-encoded) word.
#
# Return value:
# If decoding is ok, set the result into VARY and return 0
# Otherwise return 1
#
decode_word()
{
word="$*"
###################################################################
# An encoded word contains only ASCII characters in range from
# '!' (Ascii value 0x21) to '~' (Ascii value 0x7e). This excludes
# in particular the SPACE (Ascii 0x20) and the TAB (Ascii 0x09).
#
# More specifically, it consists of five parts separated by
# question marks '?'
# 1. A character "="
# 2. The charset, e.g. "UTF-8" or "ISO-8859-1"
# 3. The encoding, B or Q in upper or lower case
# 4. The encoded text
# 5. A character "="
###################################################################

# Check that:
# - there is no character outside range from '!' to '~'
# - the 1st part is a "="
# - the 5th part is a "=" and it is the end of the string
if [ $(LANG=C expr "_$word" : '_=?[!-~]*=$') = 0 ]; then return 1; fi
end=$(printf "$word" | $CUT -f 5- -d '?')
if [ "$end" != "=" ]; then return 1; fi

# Extract charset, encoding, and encoded text
charset=$(printf "$word" | $CUT -f 2 -d '?')
encoding=$(printf "$word" | $CUT -f 3 -d '?')
encoded=$(printf "$word" | $CUT -f 4 -d '?')

case $encoding in
B | b)
decoded=$(printf "$encoded" | $DECODE_B64 2>/dev/null)
if [ $? != 0 ]; then return 1; fi
;;
Q | q)
decoded=$(printf "$encoded" | $TR "_" " " | $DECODE_QP 2>/dev/null)
if [ $? != 0 ]; then return 1; fi
;;
*)
return 1
;;
esac

VARY=$(printf "$decoded" | $ICONV -f $charset -t UTF-8 2>/dev/null)
return $?
}

#
# Function: add_word
#
# Description:
# Try to decode a new word, and update DECODED_LINE and STATUS
# depending on the result and the previous STATUS.
#
# Parameter:
# A single (possibly MIME-encoded) word.
#
# Return value:
# None
#
# Side effects:
# Change DECODED_LINE and STATUS
#
add_word()
{
word="$*"
if decode_word "$word"; then
if [ "$STATUS" = "normal" ]; then
DECODED_LINE="${DECODED_LINE} "
fi
DECODED_LINE="${DECODED_LINE}${VARY}"
STATUS="decoded-word"
else
if [ "$STATUS" != "none" ]; then
DECODED_LINE="${DECODED_LINE} "
fi
DECODED_LINE="${DECODED_LINE}${word}"
STATUS="normal"
fi
}

#
# Function: flush_line
#
# Description:
# Before beginning to manage a new header, or just before ending
# the script, print the pending DECODED_LINE if any.
#
# Parameter:
# None
#
# Return value:
# None
#
# Side effects:
# Print things to stdout
# Change DECODED_LINE and STATUS
#
flush_line()
{
if [ -n "${DECODED_LINE}" ]; then
printf "%s\n" "${DECODED_LINE}"
DECODED_LINE=""
STATUS="none"
fi
}

#
# Function: manage_line
#
# Description:
# Manage a new line, which can be either the beginning or the
# continuation of a mail/news header.
# This function prints the previous line if this is a new one,
# then it adds successive parts to the new DECODED_LINE, while
# updating STATUS as needed.
#
# Parameter:
# An input
line.========================================================================

#
# Return value:
# None
#
# Side effects:
# Print things to stdout
# Change DECODED_LINE and STATUS
#
manage_line()
{
line="$*"

# Is it a continuation line?
if [ $(LANG=C expr "$line" : "[ \t]") = 0 ]; then
# No: new header
flush_line
fi

for word in $line; do
add_word "$word"
done
}

#
# Function: manage_file
#
# Description:
# Call manage_line for each line in a given file
#
# Parameter:
# A file name, or "-" for stdin
#
# Return value:
# None
#
# Side effects:
# Same as manage_line
#
manage_file()
{
file=${1--} # POSIX-compliant; ${1:--} can be used either.
while IFS= read -r line; do
manage_line "$line"
done < <(cat -- "$file")
}

#
# Parse arguments for -h or --help
#
for i in "$@"; do
case $i in
-h | --help)
usage
;;
-)
;;
-*)
printf "Unknown argument '%s'\n" "$i"
usage
;;
esac
done

#
# Main loop.
# Call manage_file for each filename in parameters,
# then print the last pending DECODED_LINE if any.
#
for file in "$@"; do
manage_file "$file"
done
flush_line

exit 0
========================================================================

--
Olivier Miakinen