pm-zh-decode.rc

# set syntax=procmail
#
#   Notes
#
#       You should install the following binary tools in your system:
#           procmail    (Obviously!)
#           formail     (usually comes with procmail)
#           mmencode    (may be ``mimencode'' in some system)
#           w3m         (with M17N support)
#           piconv      (iconv written in Perl, you can use ``iconv'' instead)
#
#   Documentation
#
#       If your system do not have ``w3m'', but you still want to use this 
#       module digressively, you can set the PM_ZH_USE_W3M to disable 
#       ``w3m'', but words in mail body that encoded in url-encode or escaped 
#       hexadecimal will not be processed.
#
#       You can disable w3m processing by putting the following line before 
#       the INCLUDERC setting:
#
#           PM_ZH_USE_W3M = "no"
#
#      All the return values described below are encoded in UTF-8, so I 
#      recommend you also save your .procmailrc in UTF-8 format.
#
#   Required settings
#
#       In your .procmailrc config file, make sure the binary tool mentioned 
#       above can be found in the ${PATH} environment.
#
#   Call arguments (variables to set before calling)
#
#       (none)
#
#   Return values
#
#       o   PM_ZH_SUBJECT contains the Subject string from the header in UTF-8
#       o   PM_ZH_STRIP_SUBJECT contains the Subject in UTF-8 without fake 
#           characters.
#       o   PM_ZH_FROM contains the From string from the header in UTF-8, 
#           usually contains sender's name and e-mail address.
#       o   PM_ZH_BODY contains the mail body content in UTF-8
#       o   E_CHARSET contains the charset name used to decoded the mail body
#       o   E_ENCODING contains the encoded name used to decoded the mail body
#       o   HAS_HTML_CONTENT is "yes" if there is text/html section in mail 
#           body.
#       o   E_HTML_CHARSET contains the charset name used to decoded the 
#           text/html section in mail body.
#
#   Examples
#
#       PATH = /bin:/usr/bin:/usr/local/bin
#       # if you don't have ``w3m'', uncomment the following line
#       # PM_ZH_USE_W3M = "no"
#       INCLUDERC = /absolute/path/to/pm-zh-decode.rc
#
#       # filter the mail to 'spam' folder if mail body has any one of the 
#       # string ['賺錢', '貸款', '免利息', '直銷']
#       :0:
#       * PM_ZH_BODY ?? (賺錢|\
#       貸款|\
#       免利息|\
#       直銷)
#       spam
#
#   Known issues
#
#       There is some problem when procmail do long string processing in 
#       back-quotes(`), so we can just do the extracting command in one time.
#
#       Environment variable assignment may sometime not work normally caused 
#       by procmail either.
#
#   Change Log
#
#       2012-05-04
#           Create module.
#
#       2012-05-07
#           Rewrite parsing procedure, use w3m.
#
#

#
# Program Settings
# ===========================================

FORMAIL		= ${FORMAIL:-'formail'}

MME_QP		= ${MME_QP:-'mmencode -u -q'}	# decode Quoted printable
MME_B64		= ${MME_B64:-'mmencode -u -b'}	# decode base64

CONV_B2U	= ${CONV_B2U:-'piconv -f big5 -t utf8'}
CONV_G2U	= ${CONV_G2U:-'piconv -f gb2312 -t utf8'}

# bumped the LINEBUF up to be large enough to hold string
LINEBUF = 4096

#
# Extract Fields
# ===========================================

# decoding subject, sender

PM_ZH_SUBJECT = `$FORMAIL -cxSubject: | perl -MEncode -ne 'print encode("utf8", decode("MIME-Header", $_));'`
PM_ZH_FROM = `$FORMAIL -cxFrom: | perl -MEncode -ne 'print encode("utf8", decode("MIME-Header", $_));'`

# strip special characters 
# long string reuse may failed, so we do it again. :~
PM_ZH_STRIP_SUBJECT = `$FORMAIL -cxSubject: | perl -MEncode -ne 'print encode("utf8", decode("MIME-Header", $_));' | sed 's/'"[-+~^|#\@!& ',;:?|.=_(){}<>\/*%0-9]"/'''/g' | sed 's/\[//g' | sed 's/]//g' | sed 's/"//g'`

PM_ZH_BODY = ""

#
# Charset, Encoding pasing
# ===========================================

# Now, we want to test if mail BODY has any Charset, Encoding settings...

# default charset, encoding
E_CHARSET = "iso-8859-1"
E_ENCODING = "7-bit"

HAS_HTML_CONTENT = "no"

:0
* HB ?? ^Content-Type:
{
    # CHARSET
    :0
    *$ B ?? ^(Content-Type:.*|	)charset *= *\"?\/[^\";]+
    { E_CHARSET = $MATCH }

    :0 E
    *$ H ?? ^(Content-Type:.*|	)charset *= *\"?\/[^\";]+
    { E_CHARSET = $MATCH }

    # ENCODING
    :0
    *$ B ?? ^Content-Transfer-Encoding: +\/.+
    { E_ENCODING = $MATCH }

    :0 E
    *$ H ?? ^Content-Transfer-Encoding: +\/.+
    { E_ENCODING = $MATCH }

    # html content processing
    :0
    * B ?? ^Content-Type: *text/html;^?( |	)*charset *= *\"?\/[^\";]+
    {
	HAS_HTML_CONTENT = "yes"
	E_HTML_CHARSET = $MATCH
    }
}

#
# Decoding method assignment
# ===========================================

:0
{
    # default filter
    F_CHARSET = 'cat'
    F_ENCODING = 'cat'
    F_HTML = 'cat'

    # CHARSET
    :0
    * E_CHARSET ?? big-?5
    { F_CHARSET = $CONV_B2U }

    :0 E
    * E_CHARSET ?? gb2312-?.*
    { F_CHARSET = $CONV_G2U }

    # ENCODING
    :0
    * E_ENCODING ?? quoted-printable
    { F_ENCODING = $MME_QP }

    :0 E
    * E_ENCODING ?? base64
    { F_ENCODING = $MME_B64 }

    # html content processing
    :0
    * ! PM_ZH_USE_W3M ?? ^^no^^
    * HAS_HTML_CONTENT ?? ^^yes^^
    {
	:0
	* E_HTML_CHARSET ?? big-?5
	{ F_HTML = 'w3m -I big5 -O big5 -T text/html' }

	:0 E
	* E_HTML_CHARSET ?? gb2312-?.*
	{ F_HTML = 'w3m -I gb2312 -O gb2312 -T text/html' }

	:0 E
	* E_HTML_CHARSET ?? utf-?8
	{ F_HTML = 'w3m -I utf8 -O utf8 -T text/html' }
    }
}

#
# After parsing the mail BODY, we can remove any MIME header from it
# We extract the mail BODY here
# NOTE: any space is NOT allowed between ``variable'' and ``='' here, it's a special usage excluding environment variable assignmnet

:0 fW
PM_ZH_BODY= | $FORMAIL -I "" | awk ' ! /^--[a-zA-Z0-9\-="._:]+$/ && ! /^Content-[DT]/ && ! /^\tcharset=/ && ! /^\tboundary=/  {print}' | $F_ENCODING | $F_HTML | $F_CHARSET | tr -d \\n