X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FMID.pm;h=d7a42c38573b17111458ddccd32e821355f79b91;hb=e67dc3e9fc73029332a632d022002d3ce80c1306;hp=c82e84013ec0d68730423ab4b691074c5b690dc5;hpb=1cb230c4d47c8ef5d03f0b8de2f8ad93c303d816;p=public-inbox.git
diff --git a/lib/PublicInbox/MID.pm b/lib/PublicInbox/MID.pm
index c82e8401..d7a42c38 100644
--- a/lib/PublicInbox/MID.pm
+++ b/lib/PublicInbox/MID.pm
@@ -1,4 +1,4 @@
-# Copyright (C) 2015-2018 all contributors
+# Copyright (C) 2015-2019 all contributors
# License: AGPL-3.0+
#
# Various Message-ID-related functions.
@@ -7,9 +7,10 @@ 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/;
+ mids references mids_for_index/;
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
MAX_MID_SIZE => 244, # max term size (Xapian limitation) - length('Q')
@@ -25,11 +26,11 @@ sub mid_clean {
$mid;
}
-# this is idempotent
+# this is idempotent, used for HTML anchor/ids and such
sub id_compress {
my ($id, $force) = @_;
- if ($force || $id =~ /[^\w\-]/ || length($id) > MID_MAX) {
+ if ($force || $id =~ /[^a-zA-Z0-9_\-]/ || length($id) > MID_MAX) {
utf8::encode($id);
return sha1_hex($id);
}
@@ -53,11 +54,10 @@ sub mid2path {
# Only for v1 code paths:
sub mid_mime ($) { mids($_[0]->header_obj)->[0] }
-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) {
+ for my $v (@_) {
my @cur = ($v =~ /<([^>]+)>/sg);
if (@cur) {
push(@mids, @cur);
@@ -65,7 +65,23 @@ sub mids ($) {
push(@mids, $v);
}
}
- uniq_mids(\@mids);
+ \@mids;
+}
+
+sub mids ($) {
+ my ($hdr) = @_;
+ my @mids = $hdr->header_raw('Message-Id');
+ uniq_mids(extract_mids(@mids));
+}
+
+# we allow searching on X-Alt-Message-ID since PublicInbox::NNTP uses them
+# to placate some clients, and we want to ensure NNTP-only clients can
+# import and index without relying on HTTP endpoints
+sub mids_for_index ($) {
+ my ($hdr) = @_;
+ my @mids = $hdr->header_raw('Message-Id');
+ my @alts = $hdr->header_raw('X-Alt-Message-ID');
+ uniq_mids(extract_mids(@mids, @alts));
}
# last References should be IRT, but some mail clients do things
@@ -79,21 +95,34 @@ sub references ($) {
push(@mids, ($v =~ /<([^>]+)>/sg));
}
}
- 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;
+ $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);
}
- next if $seen{$mid};
+ next if $seen->{$mid};
push @ret, $mid;
- $seen{$mid} = 1;
+ $seen->{$mid} = 1;
}
\@ret;
}