Having
moved to Mutt, I needed a good way to search my mails.
I heard good things about
grepmail, so I decided use it. And if it had just worked, that would have been the end of the matter.
However, as it turns out, grepmail uses Perl's
Mail::Mbox::MessageParser. This was not installed on my machine so I fired up
cpan and went hunting. Unfortunately,
cpan complained that
Mail::Mbox::MessageParser was not passing it's test cases properly and so I was stuck.
Then I figured - how hard could it be to search mails?
Turns out - it's not so hard after all. Presenting -
grepmail.sh! A simple and powerful mail searching script.
FEATURES
-
Any regular expression supported by sed can be used for the search (in header and body).
- Able to search in
compressed (bzip-ed) mailboxes.
- Ignores false matches in Base64 encoded regions.
- Checks the environment variable
GREPMAILPROG for the MUA. If not set, defaults to
mutt -f.
- Implements
smartcase which means it will ignore case in the match
except if you use uppercase alphabets, in which case it will do an exact match.
apple matches "APPLE", "apple", or "aPpLe"
Apple only matches "Apple"
- Displays the progress of the search in a nicely colored and formatted terminal.
- Allows the user to cancel the search in the middle by trapping
Ctrl+C.
It works pretty well - give it a try!
UPDATE: 2009-08-25
Rithish recommends
Xobni for Outlook users. I haven't tried it out but it looks good so if you are looking for an Outlook solution try it out.
grepmail.sh
#!/bin/bash
# This program greps for mails matching a given regexp
# and runs mutt on the resulting MBox.
#
# Author: http://the-brown-dragon.com
#
# History:
# 2009-08-28
# - Added ability to search in compressed (bzip-ed) mailboxes.
# - Now taking list of mboxes to search from user (clean solution).
# - BUGFIX: Search expression with spaces
# 2009-08-25
# - Instead of ignoring /archive/ subdirectories, relegated
# to the end of the search.
# 2009-08-21
# - Added ability for users to cancel search (trapping Ctrl+C)
# - Somewhat improved UI
# - Ignoring "/archive/" subdirectories
# - Accepts an environment variable (GREPMAILPROG) for MUA
# If not provided, defaults to mutt
# - BUGFIX: Name display incorrect when only email present
# 2009-08-20
# - Searches all MBox mailboxes under ~/Mail directory
# - Does a smartsearch
# - Ignores Base64-Blocks
# - Runs Mutt on the search results
# - Displays progress of search in color
##############################################################################
# Check args
if [ -z "$2" ];
then
echo "Usage: `basename $0` regexp-to-search list-of-mboxes"
exit 1
fi
# Set up variables
REGEXP=$1
SEDFILE1=grepmail1.$$.sed
SEDFILE2=grepmail2.$$.sed
RESULTMB=grepmail-` echo "${REGEXP}" | sed "s/ .*//g"`.$$.mbox
# Remaining are file names
shift
# Create sed script to find messages
# with the given regexp
# If no caps are given, ignore case (good default behaviour)
if [ -n "`echo "${REGEXP}" | sed "/[A-Z]/p;d;"`" ];
then
IGNORECASE="" # Don't ignore case
else
IGNORECASE="y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/"
fi
cat > ${SEDFILE1} <<SEDFILE1END
1{/^From /!q} # Not a mailbox - just quit
/^From /{ # New mail
x # Get previously stored mail
/^MATCH/{ # Has it been marked as a match?
s/^MATCH// # Remove marker
s/^\nFrom /From / # Clean up any trailing newlines
p # Save it!
}
s/.*// # Empty previous message
x # And start afresh
}
H # Save this message line also
\${ # End of file?
x # Get previously stored mail
/^MATCH/{ # Has it been marked as a match?
s/^MATCH// # Remove marker
s/^\nFrom /From / # Clean up any trailing newlines
p # Save it!
}
d # Empty last message
}
# Ignore Base64 encoded
/^[A-Za-z0-9+/=]\+$/d
# Match regular expression
${IGNORECASE}
/${REGEXP}/{
g # Match found! Get the stored message
/^MATCH/!s/^/MATCH/ # Mark it as matched (if not already marked)
h # Put it back
}
d # Just continue
SEDFILE1END
# Create nice output sed script
cat > ${SEDFILE2} <<SEDFILE2END
/^From /{s/.*//;h;} # New mail - empty hold area
x # Get hold area
/DONE/{x;d;} # Already processed message
x # Not yet, get back current line
/^From: /{ # Found a "From:" line
x # Save it and get stored text
/^From: /{x;d;} # "From:" already found, revert!
/^Subject:/H # (else)Stored text has "Subject:"? Add it
}
/^Subject:/{ # Found a "Subject:" line
x # Save it and get stored text
/\n*Subject:/{x;d;} # "Subject:" already found, revert!
/^From:/{G;h;} # (else)Stored text has "From:"? Add it
}
g # Get stored text
/^From:.*\nSubject:.*$/{ # Got both "From:" and "Subject"?
s/^From: *\(.*\) *\nSubject: *\(.*\) *$/\1\n\2/ # Break 'em up
h;s/.*\n// # Isolate "Subject:" value
x;s/\n.*// # Isolate "From:" value
s/"//g # Remove quotes
s/\(..*\)[[<(].*/\1/ # Remove email if name is present
s/[ \n]*$// # Clean trailing spaces
s/^<//;s/>$//; # Remove separators if email left
G # Append "Subject" value
s/\(.*\)\n\(.*\)/ [32m(\1) \2[m/ # Format...
p # ...and SHOW!
s/.*/DONE/ # Mark as done
h # Save DONE marker
}
d # Just continue...
SEDFILE2END
# Set up system to break out of search early (trap Ctrl+C)
USERINTERRUPT=0
stoploop ()
{
USERINTERRUPT=1
}
trap stoploop SIGINT
# Prepare result mailbox
> ${RESULTMB}
# Disable internal file separator to handle filenames with spaces
IFS=$'\n'
# Search in specified mailboxes
for FNAME in $*
do
if [ ${USERINTERRUPT} == 1 ]; then
echo User Interrupted Search...
break
fi
echo Searching ${FNAME}
case ${FNAME} in
*.bz2 )
bzcat ${FNAME} | sed -f ${SEDFILE1} | tee -a ${RESULTMB} | sed -f ${SEDFILE2}
;;
* )
sed -f ${SEDFILE1} ${FNAME} | tee -a ${RESULTMB} | sed -f ${SEDFILE2}
;;
esac
done
# Done
# Clean up
rm ${SEDFILE1} ${SEDFILE2}
# Run mutt on results
if [ -s ${RESULTMB} ]
then
echo "Search results saved in: [32;1m${RESULTMB}[m"
/usr/bin/bash -c "${GREPMAILPROG:-"mutt -f"} ${RESULTMB}"
else
echo No results found!
rm -f ${RESULTMB}
fi
# TODO
# - Improve ability to specify mailboxes
# - Allow user to turn off color and/or progress
# - Ignore HTML tags and other meta information
# - Allow search by date range
