treewide: run update-copyrights from gnulib for 2019

[public-inbox.git] / lib / PublicInbox / MID.pm

diff --git a/lib/PublicInbox/MID.pm b/lib/PublicInbox/MID.pm
index 4ccb704da7e3d783f61a3ded49a970349ebcb86e..d2bbaec71e9ef8b86b35d5f6e54e27da78b9d6fb 100644 (file)
--- a/lib/PublicInbox/MID.pm
+++ b/lib/PublicInbox/MID.pm
@@ -1,4 +1,4 @@
-# Copyright (C) 2015-2018 all contributors <meta@public-inbox.org>
+# Copyright (C) 2015-2020 all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 #
 # Various Message-ID-related functions.
@@ -7,10 +7,14 @@ 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/;
-use constant MID_MAX => 40; # SHA-1 hex length
+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')
+};
 
 sub mid_clean {
        my ($mid) = @_;
@@ -22,11 +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);
        }
@@ -47,31 +51,79 @@ sub mid2path {
        "$x2/$x38";
 }
 
-sub mid_mime ($) { $_[0]->header_obj->header_raw('Message-ID') }
+# Only for v1 code paths:
+sub mid_mime ($) { mids($_[0]->header_obj)->[0] }
 
-sub uniq_mids {
-       my ($hdr, @fields) = @_;
-       my %seen;
-       my @raw;
-       foreach my $f (@fields) {
-               push @raw, $hdr->header_raw($f);
+# only intended for Message-ID and X-Alt-Message-ID
+sub extract_mids {
+       my @mids;
+       for my $v (@_) {
+               my @cur = ($v =~ /<([^>]+)>/sg);
+               if (@cur) {
+                       push(@mids, @cur);
+               } else {
+                       push(@mids, $v);
+               }
        }
-       my @mids = (join(' ', @raw) =~ /<([^>]+)>/g);
-       my $mids = scalar(@mids) == 0 ? \@raw: \@mids;
-       my @ret;
-       foreach (@$mids) {
-               next if $seen{$_};
-               push @ret, $_;
-               $seen{$_} = 1;
-       }
-       \@ret;
+       \@mids;
+}
+
+sub mids ($) {
+       my ($hdr) = @_;
+       my @mids = $hdr->header_raw('Message-Id');
+       uniq_mids(extract_mids(@mids));
 }
 
-sub mids { uniq_mids($_[0], 'Message-Id') }
+# we allow searching on X-Alt-Message-ID since PublicInbox::NNTP uses them
+# to placate some clients, and we want to ensure NNTP-only clients can
+# import and index without relying on HTTP endpoints
+sub mids_for_index ($) {
+       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
 # out of order, so trust IRT over References iff IRT exists
-sub references { uniq_mids($_[0], 'References', 'In-Reply-To') }
+sub references ($) {
+       my ($hdr) = @_;
+       my @mids;
+       foreach my $f (qw(References In-Reply-To)) {
+               my @v = $hdr->header_raw($f);
+               foreach my $v (@v) {
+                       push(@mids, ($v =~ /<([^>]+)>/sg));
+               }
+       }
+
+       # 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, $seen) = @_;
+       my @ret;
+       $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);
+               }
+               push(@ret, $mid) unless $seen->{$mid}++;
+       }
+       \@ret;
+}
 
 # RFC3986, section 3.3:
 sub MID_ESC () { '^A-Za-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@' }