# pm-jamime-save.rc -- save message's MIME attachement (one file) to a file
# $Id: pm-jamime-save.rc,v 2.9 2004/09/21 14:50:57 jaalto Exp $
#
#   File id
#
#       .Copyright (C)   1998-2004 Jari Aalto
#       .$Keywords:      procmail, subroutine, mime $
#
#       This code is free software in terms of GNU Gen. pub. Lic. v2 or later
#       Refer to http://www.gnu.org/copyleft/gpl.html
#
#   Documentation
#
#       This module saves _one_ simple file attachment (MIME) from he
#       message. The message must define following MIME headers. If
#       "filename=" does not exists, then the message will not be
#       handled.
#
#           Mime-Version: <version>
#           Content-Type: <type>
#           Content-Disposition: attachment; filename="file.txt"
#
#       The last line can also be in separate line, provided that it
#       is intended according to standard rules:
#
#           Mime-Version: <version>
#           Content-Type: <type>
#           Content-Disposition: attachment;
#               filename="file.txt"
#
#       Procmail is not very suitable for saving MIME attachments and
#       you should not think that this the right tool for you.
#       If you receive anything more than 1 attachment, this recipe
#       does nothing, because that's out of our league and you need some
#       more heavy weight mime tools. E.g. Perl CPAN has MIME libraries.
#
#       _Note_: When the attachment is in the body, it is simply written
#       to a disk and the location in message is replaced with test:
#
#           Extracted to file:/users/foo/junk/file.txt.1998-03-30
#
#       The existing mime headers that surrounded the attachment are
#       untouched, so don't try to press your Mail Agents MIME buttons
#       at that point any more. There is no such file in that spot
#       if you set `JA_MIME_SAVE_DEL' to `yes').
#
#   Required settings
#
#       PMSRC must point to source directory of procmail code. This
#       subroutine includes library:
#
#       o   pm-javar.rc
#       o   pm-jamime.rc
#       o   pm-jamime-decode.rc
#       o   pm-jadate.rc (which will call other pm-jadate*.rc files)
#
#   Call arguments (variables to set before calling)
#
#       o   `JA_MIME_SAVE_DIR', point this to directory where you
#           want to store attachments.
#       o   `JA_MIME_SAVE_DECODE', set this to "yes", if you want that
#           attachment is decoded before written to disk. This usually
#           opens quoted printable or base64 encoding.
#       o   `JA_MIME_SAVE_DEL', set this to "yes", if you want to remove
#           the attachment from the body of the message after it has
#           been filed. Be vary careful if you use this option. If you
#           keep backup cache of incoming mail, then you might try "yes".
#       o   `JA_MIME_SAVE_OVERWRITE', set this to "yes" if it's okay to
#           overwrite to an existing filename found from attachment.
#           If you get periodic attachments always with same name, then
#           you would want to set this to yes.
#
#   Core dump note
#
#       Because procmail uses LINEBUF when filtering messages, a core
#       dump may happen if the attachment being filtered is bigger than
#       the LINEBUF. The current setting accepts 524K attachments, but if
#       you expect to get bigger than that, you want to increase
#       `JA_MIME_SAVE_LINEBUF'.
#
#   Possible conflict with your awk
#
#       Awk is used because it is much more system load friendly than perl.
#       If you see an error message in the log file saying that awk failed:
#
#           procmail: Executing awk,
#           ...
#           procmail: Error while writing to "awk"
#           procmail: Rescue of unfiltered data succeeded
#
#       it means that the system's standard awk doesn't support the
#       variable passing syntax. To verify that this is the case, run
#       following test:
#
#           % awk '{print VAR; exit}' VAR="value" /etc/passwd
#
#       The proper awk should print "value". If not, then see if you have
#       `nawk' or `gawk' in your system, which should understand the
#       variable passing syntax. To change the AWK, you need to set
#       following variable somewhere at the top of your *.procmailrc*
#
#           AWK = "gawk"        # if that works better than standard "awk"
#
#   Return values (none)
#
#   Change Log (none)

# .................................................... &initialising ...

dummy = "
========================================================================
pm-jamime-save.rc: init:"

:0
* !  WSPC ?? ( )
{
    INCLUDERC = $PMSRC/pm-javar.rc
}

:0
* !  MIME_VER ?? [0-9]
{
    INCLUDERC = $PMSRC/pm-jamime.rc
}


:0                              # if has no value, set it
*$ ! LINEBUF ?? $d
{
    LINEBUF = 1024
}


MimeSaveLinebuf     = $LINEBUF

# .......................................................... &public ...
# User configurable sections

JA_MIME_SAVE_DIR       = ${JA_MIME_SAVE_DIR:-"$HOME"}
JA_MIME_SAVE_DECODE    = ${JA_MIME_SAVE_DECODE:-"no"}
JA_MIME_SAVE_DEL       = ${JA_MIME_SAVE_DEL:-"no"}
JA_MIME_SAVE_LINEBUF   = ${JA_MIME_SAVE_LINEBUF:-524280}
JA_MIME_SAVE_OVERWRITE = ${JA_MIME_SAVE_OVERWRITE:-"no"}

# ........................................................... &do-it ...

dummy = "pm-jamime-save.rc: HEADER; $MIME_H_ATTACHMENT"

:0
*  MIME_H_ATTACHMENT ?? [a-z]
{
    file = $JA_MIME_SAVE_DIR/$MIME_H_ATTACHMENT

    :0
    * JA_MIME_SAVE_DECODE ?? yes
    {
        #   decode regardless of body content
        JA_MIME_DECODE_REGEXP = ".*"

        INCLUDERC = $RC_MIME_DECODE
    }

    # ..................................................... filename ...

    :0
    *    JA_MIME_SAVE_OVERWRITE ?? no
    *$ ? $IS_EXIST $file
    {
        :0
        *$ ! YYYY ?? $d
        {
            INCLUDERC = $RC_DATE
        }

        file = "$file.$YYYY-$MM-$DD"

        #   Still not unique?

        :0
        *$ ? $IS_EXIST $file
        {
            :0 fhw
            | $FORMAIL -I "X-MimeSave-Error: (file exists) $file"

            # kill variable to prevent next recipe from running
            file
        }
    }

    # .................................................... write out ...

    :0
    * file ?? [a-z]
    {
        :0 bwc:                                 # do not modify body
        * JA_MIME_SAVE_DEL ?? no
        | $CAT > $file

        #       Write out the attachment and replace body
        #       with reference to the file.

        :0 E
        {
            LINEBUF = $JA_MIME_SAVE_LINEBUF

            :0 fbw:
            | ( $CAT > $file; echo "Saved to $file" )

            LINEBUF = $MimeSaveLinebuf
        }
    }
}

dummy = "pm-jamime-save.rc: BODY check"

:0 E
*  MIME_B_ATTACHMENT               ??  [a-z]
*  MIME_B_ATTACHMENT_FILE_COUNT    ??  ^^1^^
{
    dummy = "pm-jamime-save.rc: BODY ENTERED: can do only some limited operations."
    file  = $JA_MIME_SAVE_DIR/$MIME_B_ATTACHMENT

    #   Try to locate the header where the content type is defined for
    #   file attachment. Note: there is no typo in the regexp, the
    #   caret(^) matches newline. See http://pm-doc.sourceforge.net/
    #
    #   Content-Type: application/octet-stream
    #   Content-Disposition: attachment; filename="file.txt"
    #   Content-Transfer-Encoding: 7bit

    dummy = "pm-jamime-save.rc: Check contentType BODY 1"

    :0
    *$ B ?? ^Content-Type:$s+\/.*\
            ^Content-Disposition:.*($MIME_B_ATTACHMENT\
            |^$s+.*$MIME_B_ATTACHMENT)
    *  MATCH ?? \/.*
    {
        contentType = $MATCH
    }

    dummy = "pm-jamime-save.rc: Check contentType BODY 2"

    :0 E
    *$ B ?? ^Content-Type:$s+\/.*\
            ^Content-Transfer-Encoding:.*\
            ^Content-Disposition:.*($MIME_B_ATTACHMENT\
            |^$s+.*$MIME_B_ATTACHMENT)
    *  MATCH ?? \/.*
    {
        contentType = $MATCH

        #   The last "Content-Disposition" match allows two cases
        #
        #   1.  Content-Disposition: attachment; filename="file.txt"
        #
        #   2.  Content-Disposition: attachment;
        #               filename="file.txt"

        :0
        *$ B ?? ^Content-Type:.*\
                ^Content-Transfer-Encoding:$s\/.*\
                ^Content-Disposition:.*($MIME_B_ATTACHMENT\
                |^$s+.*$MIME_B_ATTACHMENT)
        *  MATCH ?? \/.*
        {
            contentEncoding = $MATCH
        }
    }

    #   We expect that the headers come in this order.
    #   If they don't, then we can't know the encoding.
    #   We don't even try anything else: Procmail is not the right tool
    #   for complete MIME handling.

    dummy = "pm-jamime-save.rc: Check contentEncoding BODY"

    :0
    *  contentEncoding ?? ^^^^
    *$ B ?? ^Content-Type:.*\
            ^Content-Disposition:.*\
            ^Content-Transfer-Encoding:$s+\/.*
    {
        contentEncoding = $MATCH
    }

    :0 E
    *  contentEncoding ?? ^^^^
    *$ B ?? ^Content-Type:.*\
            ^Content-Transfer-Encoding:$s+\/.*
    {
        contentEncoding = $MATCH
    }

    :0
    *  contentEncoding ?? ^^^^
    *$ B ?? ^Content-Type:.*\
            ^Content-Transfer-Encoding:$s+\/.*
    {
        contentEncoding = $MATCH
    }

    # ............................................... fix mixed-part ...
    #   In most typical message, sender "says" something in text/plain and
    #   then adds an atatchement
    #
    #   But, due to order of the MIME headers we may have picked the
    #   text/plain. Change it to application/octet-stream if found.
    #
    #   Mime-Version: 1.0
    #   Content-Type: multipart/mixed; boundary="----------118D218634724256"
    #
    #   ------------118D218634724256
    #   Content-Type: text/plain; charset=us-ascii
    #   Content-Transfer-Encoding: 7bit
    #
    #
    #   ------------118D218634724256
    #   Content-Type: application/octet-stream; name="Ass_rake.dbf"
    #   Content-Transfer-Encoding: base64
    #   Content-Disposition: attachment; filename="Ass_rake.dbf"

    dummy = "pm-jamime-save.rc: Check contentType ($contentType) text/plain => application"

    :0
    *  contentType ?? text/
    *$ B           ?? ^Content-Type:$s+\/application.*
    {
        contentType = $MATCH

        dummy = "pm-jamime-save.rc: finding Transfer-Encoding for application/*"

        :0
        *$ B     ?? ^Content-Type:$s+application.*\
                    ^Content-Transfer-Encoding:$s\/.*
        {
            contentEncoding = $MATCH
        }
    }

    # ............................................... check filename ...
    #   Change filename if there is already that file

    dummy = "pm-jamime-save.rc: check filename"

    :0
    *    JA_MIME_SAVE_OVERWRITE ?? no
    *$ ? $IS_EXIST $file
    {
        :0
        *$ ! YYYY ?? $d
        {
            INCLUDERC = $RC_DATE
        }

        file = "$file.$YYYY-$MM-$DD"

        #   Still not unique?

        :0
        *$ ? $IS_EXIST $file
        {
            :0 fhw
            | $FORMAIL -I "X-MimeSave-Error: (file exists) $file"

            # Kill variable to prevent next recipe from running
            file
        }
    }

    # ........................................................ &save ...
    # That nice backslash lines in the AWK are automatically handled
    # by Emacs procmail mode (tinyprocmail.el)
    # See http://tiny-tools.sourceforge.net/

    # 1) Locate the positions where the attachment starts
    # 2) When found, start looking for empty line which endS the mime
    #    headers. GO is set to 1, when attachment starts
    # 3) Attachment ends when mime boundary is hit. Actually
    #    it ended one line before that, which was a empty line.

    MimeSaveShellmetas  = $SHELLMETAS
    LINEBUF             = $JA_MIME_SAVE_LINEBUF
    SHELLMETAS

    :0 fbw
    * COUNT ?? [1-9]
    * file  ?? [a-z]
    | $AWK                                                                  \
      ' { BEGIN found = 0 }                                                 \
        {                                                                   \
            if ( match($0,MATCH) > 0 )                                      \
            {                                                               \
                 found++;                                                   \
            }                                                               \
            if ( (found = $COUNT - 1) &&  match($0, "^[ \t]*$") )           \
            {                                                               \
                go = 1;                                                     \
                next;                                                       \
            }                                                               \
            if ( match($0, MIME) )                                          \
            {                                                               \
                go = 0;                                                     \
            }                                                               \
            if ( go )                                                       \
            {                                                               \
                if ( DELETE == "yes")                                       \
                {                                                           \
                    if ( urlFlag == 0 )                                     \
                    {                                                       \
                        urlFlag = 1;                                        \
                        printf "\extracted to file:%s\n\n", FILE;           \
                    }                                                       \
                }                                                           \
                else                                                        \
                {                                                           \
                    print;                                                  \
                }                                                           \
                print $0 >> FILE;                                           \
                next;                                                       \
            }                                                               \
            print;                                                          \
        }                                                                   \
      ' DELETE="$JA_MIME_SAVE_DEL"                                          \
        MIME="$MIME_BOUNDARY"                                               \
        COUNT="$MIME_BOUNDARY_COUNT"                                        \
        MATCH="Disposition:.*$MIME_B_ATTACHMENT"                            \
        FILE="$file"


    LINEBUF     = $MimeSaveLinebuf
    SHELLMETAS  = $MimeSaveShellmetas

    # ....................................................... decode ...
    dummy = "pm-jamime-save.rc: Encoding($contentEncoding) of $file"

    :0
    *$ file             ?? $a
    *$ contentEncoding  ?? $a
    {
        file2 = "$file.raw";

        :0 wc
        * contentEncoding ?? base64
        | $CAT $file | $MIME_BIN_64 > $file2; $MV $file2 $file

        :0 Ewc
        * contentEncoding ?? quoted-printable
        | $CAT $file | $MIME_BIN_QP > $file2; $MV $file2 $file

    }
}

dummy = "pm-jamime-save.rc: end:"

# End of file pm-jamime-save.rc