No ext_urls

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

diff --git a/lib/PublicInbox/MID.pm b/lib/PublicInbox/MID.pm
index e9a3b0c094e52b3fce1d330374d03b6bd7520235..4819cc256f575f7463ddf030569ef07c8ae6c6a4 100644 (file)
--- a/lib/PublicInbox/MID.pm
+++ b/lib/PublicInbox/MID.pm
@@ -1,18 +1,18 @@
-# Copyright (C) 2015-2020 all contributors <meta@public-inbox.org>
+# Copyright (C) all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 #
 # Various Message-ID-related functions.
 package PublicInbox::MID;
 use strict;
-use warnings;
-use base qw/Exporter/;
+use v5.10.1; # TODO: check unicode_strings compat for v5.12
+use parent qw(Exporter);
 our @EXPORT_OK = qw(mid_clean id_compress mid2path mid_escape MID_ESC
-       mids references mids_for_index $MID_EXTRACT);
+       mids references mids_for_index mids_in $MID_EXTRACT);
 use URI::Escape qw(uri_escape_utf8);
-use Digest::SHA qw/sha1_hex/;
+use PublicInbox::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')
 };
 
@@ -32,7 +32,7 @@ sub mid_clean {
 sub id_compress {
        my ($id, $force) = @_;
 
-       if ($force || $id =~ /[^a-zA-Z0-9_\-]/ || length($id) > MID_MAX) {
+       if ($force || $id =~ /[^a-zA-Z0-9_\-]/ || length($id) > ID_MAX) {
                utf8::encode($id);
                return sha1_hex($id);
        }
@@ -73,14 +73,17 @@ sub mids ($) {
        uniq_mids(extract_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 ($) {
-       my ($hdr) = @_;
-       my @mids = $hdr->header_raw('Message-ID');
-       my @alts = $hdr->header_raw('X-Alt-Message-ID');
-       uniq_mids(extract_mids(@mids, @alts));
+       mids_in($_[0], qw(Message-ID X-Alt-Message-ID));
 }
 
 # last References should be IRT, but some mail clients do things
@@ -119,13 +122,13 @@ sub uniq_mids ($;$) {
                        warn "Message-ID: <$mid> too long, truncating\n";
                        $mid = substr($mid, 0, MAX_MID_SIZE);
                }
-               push(@ret, $mid) unless $seen->{$mid}++;
+               $seen->{$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;