#!/bin/sh

# (27.08.07) Fixed problem detecting html from munpack info
# (01.10.07) Didn't remove tmp files on error condition.
# (09.10.07) Replaced 'dc' subshell with shell arithmetic

# Given an html email (either as body or attached as multi-part) it will unpack
# the html and all the attachments. It rewrites the Content-ID resource
# locators with the unpacked filenames so you can see attached images and such.
#
# Also accepts 'safe' argument. When received it kills all external URLs (for
# images, javascript, etc). I use this when I want to see the html but don't
# trust the source.
#
# All temp files are cleaned up.
#
# To use from mutt include this keybinding in the approparite config file. 
# Replace <F> keys with whatever you like.
#
## view an html email with firefox
#macro  index  <F12>  "<copy-message>/tmp/tmpbox\n<enter><shell-escape>~/bin/mutt-view-html.sh\n" "View HTML in browser"
#macro  pager  <F12>  "<copy-message>/tmp/tmpbox\n<enter><shell-escape>~/bin/mutt-view-html.sh\n" "View HTML in browser"
#
## safe version (no external images loaded)
#macro  index  <F11>  "<copy-message>/tmp/tmpbox\n<enter><shell-escape>~/bin/mutt-view-html.sh safe\n" "View (safe) HTML in browser"
#macro  pager  <F11>  "<copy-message>/tmp/tmpbox\n<enter><shell-escape>~/bin/mutt-view-html.sh safe\n" "View (safe) HTML in browser"

tmpdir=/tmp/mutthtmltmpdir
mbox=/tmp/tmpbox

mkdir $tmpdir
cd $tmpdir
file_info=`munpack -q -t -C $tmpdir $mbox`

if [ "$file_info" = "Did not find anything to unpack from $mbox" ]; then
    cat $mbox | formail -x content-type: | grep -q text/html
    if [ $? -eq 0 ]; then
        file_info="part1 (text/html)"
        # sed here strips out mail headers
        cat $mbox | sed -e '1,/^$/ d' > part1
    fi
fi

echo $file_info | grep -iq text/html
if [ $? -ne 0 ]; then
    echo "No html found!"
    rmdir $tmpdir
    rm $mbox
    sleep 2
    exit 1
fi

file_html=`echo $file_info | sed 's/.*\(part[0-9]\+\) (text\/html).*/\1/g'`
files=`echo $file_info | sed 's/\([^()]\+\) ([^()]\+)/\1/g'`
# need list of datafiles in case attachments are unnamed
data_files=`echo "$file_info" | grep -v "\(text/.*\)" | \
        sed 's/\([^()]\+\) ([^()]\+)/\1/g'`

html_file=$file_html.html
mv $file_html $html_file
files=`echo $files | sed "s/$file_html/$html_file/"`

#/usr/bin/tidy -mq -wrap 0 $html_file 2> /dev/null

if [ -n "$html_file" ]; then
    grep -qi src=\"cid: $html_file
    if [ $? -eq 0 ]; then
        filenames=`grep -i content-disposition $mbox | \
            sed -e 's/.*filename="\([^"]\+\)"/\1/'`
        if [ -z $filesnames ]; then
            # unnamed file attachments :P
            filenames="$data_files"
        fi
        cids=`grep -i content-id $mbox | \
            sed -e 's/Content-ID: <\([^<>]\+\)>/\1/'`

        count=1
        for fn in $filenames; do
            cid=`echo $cids | awk -v count=$count '{ print $count }'`
            #count=`dc -e "1 $count + p"`
            count=$(( 1 + $count ))
            sed -i "s/cid:$cid/$fn/g" $html_file
        done
    fi
    if [ "$1" = "safe" ]; then
        grep -qi src=\"http $html_file
        if [ $? -eq 0 ]; then
            # misses javascript places images, but noscript should 
            # take care of those
            sed -i "s/src\s*=\s*\"*https*:\/\/[^\">]*\"*/src=\"#\"/g" $html_file
            sed -i "s/background\s*=\s*\"*https*:\/\/[^\">]*\"*/background=\"#\"/g" $html_file
        fi
    fi
else
    rm $files $mbox
    rmdir $tmpdir
    exit 1
fi

(x-www-browser $tmpdir/$html_file; sleep 3; rm $files $mbox; rmdir $tmpdir) &
#(x-www-browser $tmpdir/$html_file; sleep 3; rm $files; rmdir $tmpdir) &

## vim: ft=sh