-# Copyright (C) 2015-2018 all contributors <meta@public-inbox.org>
+# Copyright (C) 2015-2021 all contributors <meta@public-inbox.org>
# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
#
# Various Message-ID-related functions.
use strict;
use warnings;
use base qw/Exporter/;
-our @EXPORT_OK = qw/mid_clean id_compress mid2path mid_mime mid_escape MID_ESC
- mids references/;
+our @EXPORT_OK = qw(mid_clean id_compress mid2path mid_escape MID_ESC
+ mids references mids_for_index mids_in $MID_EXTRACT);
use URI::Escape qw(uri_escape_utf8);
use Digest::SHA qw/sha1_hex/;
+require PublicInbox::Address;
use constant {
- MID_MAX => 40, # SHA-1 hex length # TODO: get rid of this
+ ID_MAX => 40, # SHA-1 hex length for HTML id anchors
MAX_MID_SIZE => 244, # max term size (Xapian limitation) - length('Q')
};
+our $MID_EXTRACT = qr/<([^>]+)>/s;
+
sub mid_clean {
my ($mid) = @_;
defined($mid) or die "no Message-ID";
# MDA->precheck did more checking for us
- if ($mid =~ /<([^>]+)>/) {
+ if ($mid =~ $MID_EXTRACT) {
$mid = $1;
}
$mid;
}
-# this is idempotent
+# this is idempotent, used for HTML anchor/ids and such
sub id_compress {
my ($id, $force) = @_;
- if ($force || $id =~ /[^\w\-]/ || length($id) > MID_MAX) {
+ if ($force || $id =~ /[^a-zA-Z0-9_\-]/ || length($id) > ID_MAX) {
utf8::encode($id);
return sha1_hex($id);
}
"$x2/$x38";
}
-sub mid_mime ($) { $_[0]->header_obj->header_raw('Message-ID') }
-
-sub mids ($) {
- my ($hdr) = @_;
+# only intended for Message-ID and X-Alt-Message-ID
+sub extract_mids {
my @mids;
- my @v = $hdr->header_raw('Message-Id');
- foreach my $v (@v) {
- my @cur = ($v =~ /<([^>]+)>/sg);
+ for my $v (@_) {
+ my @cur = ($v =~ /$MID_EXTRACT/g);
if (@cur) {
push(@mids, @cur);
} else {
push(@mids, $v);
}
}
- foreach my $i (0..$#mids) {
- next if length($mids[$i]) <= MAX_MID_SIZE;
- warn "Message-ID: <$mids[$i]> too long, truncating\n";
- $mids[$i] = substr($mids[$i], 0, MAX_MID_SIZE);
- }
+ \@mids;
+}
+
+sub mids ($) {
+ my ($hdr) = @_;
+ my @mids = $hdr->header_raw('Message-ID');
+ uniq_mids(extract_mids(@mids));
+}
- uniq_mids(\@mids);
+# for Resent-Message-ID and maybe others
+sub mids_in ($@) {
+ my ($eml, @headers) = @_;
+ uniq_mids(extract_mids(map { ($eml->header_raw($_)) } @headers));
+}
+
+# we allow searching on X-Alt-Message-ID since PublicInbox::NNTP uses them
+# to placate some clients, and we want to ensure NNTP-only clients can
+# import and index without relying on HTTP endpoints
+sub mids_for_index ($) {
+ mids_in($_[0], qw(Message-ID X-Alt-Message-ID));
}
# last References should be IRT, but some mail clients do things
foreach my $f (qw(References In-Reply-To)) {
my @v = $hdr->header_raw($f);
foreach my $v (@v) {
- push(@mids, ($v =~ /<([^>]+)>/sg));
+ push(@mids, ($v =~ /$MID_EXTRACT/g));
}
}
- uniq_mids(\@mids);
+
+ # old versions of git-send-email would prompt users for
+ # In-Reply-To and users' muscle memory would use 'y' or 'n'
+ # as responses:
+ my %addr = ( y => 1, n => 1 );
+
+ foreach my $f (qw(To From Cc)) {
+ my @v = $hdr->header_raw($f);
+ foreach my $v (@v) {
+ $addr{$_} = 1 for (PublicInbox::Address::emails($v));
+ }
+ }
+ uniq_mids(\@mids, \%addr);
}
-sub uniq_mids ($) {
- my ($mids) = @_;
+sub uniq_mids ($;$) {
+ my ($mids, $seen) = @_;
my @ret;
- my %seen;
- foreach (@$mids) {
- next if $seen{$_};
- push @ret, $_;
- $seen{$_} = 1;
+ $seen ||= {};
+ foreach my $mid (@$mids) {
+ $mid =~ tr/\n\t\r//d;
+ if (length($mid) > MAX_MID_SIZE) {
+ warn "Message-ID: <$mid> too long, truncating\n";
+ $mid = substr($mid, 0, MAX_MID_SIZE);
+ }
+ $seen->{$mid} //= push(@ret, $mid);
}
\@ret;
}
# RFC3986, section 3.3:
-sub MID_ESC () { '^A-Za-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@' }
+sub MID_ESC () { '^A-Za-z0-9\-\._~!\$\&\'\(\)\*\+,;=:@' }
sub mid_escape ($) { uri_escape_utf8($_[0], MID_ESC) }
1;