From: Jonathan Kamens <jik@kamens.brookline.ma.us>
To: nelson@crynwr.com
Subject: Re: "movemail.pl" which works with maildirs
Date: Mon, 20 Dec 1999 14:23:28 -0500

Here's yet another update to my maildir movemail.pl script.

#!/usr/bin/perl

# mdmovemail.pl - a replacement for the Emacs "movemail" program which
# understands how to read mail from a qmail-style maildir.
#
# Install it somewhere, make it executable, and then set
# rmail-movemail-program to point to it.  If you don't want the script
# to automatically strip out duplicate messages (defined as messages
# with the same message ID and the same body checksum), set the
# $strip_duplicates variable below to undef (it's enabled by default).
#
# Then create an empty file inside your maildir.  The name is
# irrelevant, although you might want to start it with a period to
# make it hidden when the directory is listed.  Personally, I
# recommend ".rmail-is-too-smart" (because that's why this flag file
# is needed).
# 
# You'll then need to modify rmail-primary-inbox-list to point at that
# file.  For example:
# 
# (setq rmail-primary-inbox-list 
#       (list (expand-file-name "~/Maildir/.rmail-is-too-smart")))
# 
# Written by Jonathan Kamens <jik@kamens.brookline.ma.us>.  Please
# send comments, corrections, etc. to me.

$strip_duplicates = 1;

($whoami = $0) =~ s,.*/,,;

$usage = "Usage: $whoami [-p] maildir destfile\n";

if ($ARGV[0] eq '-p') {
    shift;
    $preserve++;
}

die $usage if (! ($maildir = shift));
die $usage if (! ($destfile = shift));

# Perhaps you are wondering why this is necessary.  It's necessary
# because if rmail-primary-inbox-list contains a path name which is
# actually a directory, i.e., it contains your maildir, Rmail mode
# isn't content to just leave it alone -- it'll append a slash and
# your username to the end of it.  Then, it'll discover that the
# resulting path doesn't exist, so it will decide there's no new mail
# to retrieve.  This hack avoids the problem by using an empty flag
# file in the maildir, as described in the comment at the top of this
# file.

if (-f $maildir) {
    ($parent = $maildir) =~ s,/[^/]+$,,;
    $maildir = $parent if (-d "$parent/new");
}

die "$whoami: $maildir does not exist\n$usage" if (! -e $maildir);
die "$whoami: $maildir is not a directory\n$usage" if (! -d $maildir);
die "$whoami: $destfile already exists\n$usage" if (-e $destfile);

$msg_dir = "$maildir/new";
$message_start = "\f\n0, unseen,,\n";
$message_end = "\037";

die "$whoami: opening $msg_dir: $!\n" if (! opendir(DIR, $msg_dir));
exit 0 if (! (@dirfiles = sort readdir(DIR)));
closedir(DIR);

foreach $file (@dirfiles) {
    next if ($file eq '.');
    next if ($file eq '..');
    next if (! -f "$msg_dir/$file");
    &output_file($file);
}

die "$whoami: closing $destfile: $!\n" if ($dest_created && !close(OUTPUT));
unlink(@output_files) if (! $preserve);

sub output_file {
    my($basename) = @_;
    my($inpath) = "$msg_dir/$basename";
    local($/) = undef;
    my($contents);

    if (! $dest_created++) {
        open(OUTPUT, ">$destfile") || die "$whoami: opening $destfile: $!\n";
    }

    open(INPUT, $inpath) || die "$whoami: opening $inpath: $!\n";
    die "$whoami: reading $inpath: $!\n" if (! defined($contents = <INPUT>));
    close(INPUT);

    $contents =~ s/([^\n])$/$1\n/;

    if ($strip_duplicates && &is_duplicate($contents)) {
        push(@output_files, $inpath);
        return;
    }

    print(OUTPUT $message_start, $contents, $message_end) ||
        die "$whoami: writing to $destfile: $!\n";

    push(@output_files, $inpath);
}

sub is_duplicate {
    my($contents) = @_;
    my($head, $body, $checksum, $id);

    ($head, $body) = split(/\n\n/, $contents, 2);
    $head =~ s/\n+$//;
    $body =~ s/^\n+//;
    $body =~ s/\n+$//;

    $checksum = unpack("%16C*", $body);

    return undef if ($head !~ /(^|\n)\bmessage-id:\s*(\S+)\s*(\n|$)/i);
    $id = $2;

    return 1 if (defined($checksums{$id}) && $checksums{$id} == $checksum);

    $checksums{$id} = $checksum;

    return undef;
}