+# This module is part of debbugs, and is released
+# under the terms of the GPL version 2, or any later
+# version at your option.
+# See the file README and COPYING for more information.
+#
+# [Other people have contributed to this file; their copyrights should
+# go here too.]
+# Copyright 2006 by Don Armstrong <don@donarmstrong.com>.
+
+
package Debbugs::MIME;
+=head1 NAME
+
+Debbugs::MIME -- Mime handling routines for debbugs
+
+=head1 SYNOPSIS
+
+ use Debbugs::MIME qw(parse decode_rfc1522);
+
+=head1 DESCRIPTION
+
+
+=head1 BUGS
+
+None known.
+
+=cut
+
+use warnings;
use strict;
use base qw(Exporter);
BEGIN {
$VERSION = 1.00;
- @EXPORT_OK = qw(parse decode_rfc1522 encode_rfc1522 convert_to_utf8 create_mime_message);
+ @EXPORT_OK = qw(parse decode_rfc1522 encode_rfc1522 convert_to_utf8 create_mime_message getmailbody);
}
use File::Path;
+use File::Temp qw();
use MIME::Parser;
+use POSIX qw(strftime);
+use List::MoreUtils qw(apply);
+
# for decode_rfc1522
use MIME::WordDecoder qw();
use Encode qw(decode encode encode_utf8 decode_utf8 is_utf8);
# for encode_rfc1522
use MIME::Words qw();
-sub getmailbody ($);
-sub getmailbody ($)
+sub getmailbody
{
my $entity = shift;
my $type = $entity->effective_type;
if ($type eq 'text/plain' or
- ($type =~ m#text/# and $type ne 'text/html') or
+ ($type =~ m#text/?# and $type ne 'text/html') or
$type eq 'application/pgp') {
return $entity->bodyhandle;
} elsif ($type eq 'multipart/alternative') {
return undef;
}
-sub parse ($)
+sub parse
{
# header and decoded body respectively
my (@headerlines, @bodylines);
- my $parser = new MIME::Parser;
- mkdir "mime.tmp.$$", 0777;
- $parser->output_under("mime.tmp.$$");
+ my $parser = MIME::Parser->new();
+ my $tempdir = File::Temp::tempdir();
+ $parser->output_under($tempdir);
my $entity = eval { $parser->parse_data($_[0]) };
if ($entity and $entity->head->tags) {
@bodylines = @msg[$i .. $#msg];
}
- rmtree "mime.tmp.$$", 0, 1;
+ rmtree $tempdir, 0, 1;
# Remove blank lines.
shift @bodylines while @bodylines and $bodylines[0] !~ /\S/;
# Strip off RFC2440-style PGP clearsigning.
if (@bodylines and $bodylines[0] =~ /^-----BEGIN PGP SIGNED/) {
- shift @bodylines while @bodylines and length $bodylines[0];
+ shift @bodylines while @bodylines and
+ length $bodylines[0] and
+ # we currently don't strip \r; handle this for the
+ # time being, though eventually it should be stripped
+ # too, I think. [See #565981]
+ $bodylines[0] ne "\r";
shift @bodylines while @bodylines and $bodylines[0] !~ /\S/;
for my $findsig (0 .. $#bodylines) {
if ($bodylines[$findsig] =~ /^-----BEGIN PGP SIGNATURE/) {
=head2 create_mime_message
- create_mime_message([To=>'don@debian.org'],$body,[$attach1, $attach2]);
+ create_mime_message([To=>'don@debian.org'],$body,[$attach1, $attach2],$include_date);
Creates a MIME encoded message with headers given by the first
argument, and a message given by the second.
Optional attachments can be specified in the third arrayref argument.
+Whether to include the date in the header is the final argument; it
+defaults to true, setting the Date header if one is not already
+present.
+
Headers are passed directly to MIME::Entity::build, the message is the
first attachment.
=cut
sub create_mime_message{
- my ($headers,$body,$attachments) = @_;
+ my ($headers,$body,$attachments,$include_date) = @_;
$attachments = [] if not defined $attachments;
+ $include_date = 1 if not defined $include_date;
die "The first argument to create_mime_message must be an arrayref" unless ref($headers) eq 'ARRAY';
die "The third argument to create_mime_message must be an arrayref" unless ref($attachments) eq 'ARRAY';
+ if ($include_date) {
+ my %headers = apply {lc($_)} @{$headers};
+ if (not exists $headers{date}) {
+ push @{$headers},
+ ('Date',
+ strftime("%a, %d %b %Y %H:%M:%S +0000",gmtime)
+ );
+ }
+ }
+
# Build the message
- my $msg = MIME::Entity->build(@{$headers},
+ # MIME::Entity is stupid, and doesn't rfc1522 encode its headers, so we do it for it.
+ my $msg = MIME::Entity->build('Content-Type' => 'text/plain; charset=utf-8',
+ 'Encoding' => 'quoted-printable',
+ (map{encode_rfc1522($_)} @{$headers}),
Data => $body
);
else {
# This is *craptacular*, but because various MTAs
# (sendmail and exim4, at least) appear to eat From
- # lines in message/rfc822 attachments, we need to make
- # sure the From line is the first thing in the
- # attachement, not the second, so if it gets eaten, the
- # headers don't collide into the body.
+ # lines in message/rfc822 attachments, we need eat
+ # the entire From line ourselves so the MTA doesn't
+ # leave \n detrius around.
if (ref($attachment) eq 'ARRAY' and $attachment->[1] =~ /^From /) {
# make a copy so that we don't screw up anything
# that is expecting this arrayref to stay constant
return $data if $charset eq 'raw' or is_utf8($data,1);
my $result;
eval {
- # this encode/decode madness is to make sure that the data
- # really is valid utf8 and that the is_utf8 flag is off.
- $result = encode("utf8",decode($charset,$data))
+ $result = decode($charset,$data);
};
if ($@) {
warn "Unable to decode charset; '$charset' and '$data': $@";
]));
}
-sub decode_rfc1522 ($)
-{
+sub decode_rfc1522 {
my ($string) = @_;
+ # this is craptacular, but leading space is hacked off by unmime.
+ # Save it.
+ my $leading_space = '';
+ $leading_space = $1 if $string =~ s/^(\s+)//;
# unmime calls the default MIME::WordDecoder handler set up at
# initialization time.
- return MIME::WordDecoder::unmime($string);
+ return $leading_space . MIME::WordDecoder::unmime($string);
}
=head2 encode_rfc1522
# We cannot use MIME::Words::encode_mimewords because that function
# does not handle spaces properly at all.
-sub encode_rfc1522 ($) {
+sub encode_rfc1522 {
my ($rawstr) = @_;
+ # handle being passed undef properly
+ return undef if not defined $rawstr;
+ if (is_utf8($rawstr)) {
+ $rawstr= encode_utf8($rawstr);
+ }
# We process words in reverse so we can preserve spacing between
# encoded words. This regex splits on word|nonword boundaries and
- # nonword|nonword boundaries.
- my @words = reverse split /(?:(?<=[\s\n])|(?=[\s\n]))/m, $rawstr;
+ # nonword|nonword boundaries. We also consider parenthesis and "
+ # to be nonwords to avoid escaping them in comments in violation
+ # of RFC1522
+ my @words = reverse split /(?:(?<=[\s\n\)\(\"])|(?=[\s\n\)\(\"]))/m, $rawstr;
my $previous_word_encoded = 0;
my $string = '';
if (length $encoded > 75) {
# Turn utf8 into the internal perl representation
# so . is a character, not a byte.
- my $tempstr = decode_utf8($word,Encode::FB_DEFAULT);
+ my $tempstr = is_utf8($word)?$word:decode_utf8($word,Encode::FB_DEFAULT);
my @encoded;
# Strip it into 10 character long segments, and encode
# the segments