wwwstream: flesh out clone instructions for v2

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

diff --git a/lib/PublicInbox/MID.pm b/lib/PublicInbox/MID.pm
index 4ccb704da7e3d783f61a3ded49a970349ebcb86e..117d3c42f987e1fc2079cbccd2025f0176194085 100644 (file)
--- a/lib/PublicInbox/MID.pm
+++ b/lib/PublicInbox/MID.pm
@@ -10,7 +10,10 @@ our @EXPORT_OK = qw/mid_clean id_compress mid2path mid_mime mid_escape MID_ESC
        mids references/;
 use URI::Escape qw(uri_escape_utf8);
 use Digest::SHA qw/sha1_hex/;
-use constant MID_MAX => 40; # SHA-1 hex length
+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 ($mid) = @_;
@@ -47,18 +50,48 @@ sub mid2path {
        "$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);
+sub mids ($) {
+       my ($hdr) = @_;
+       my @mids;
+       my @v = $hdr->header_raw('Message-Id');
+       foreach my $v (@v) {
+               my @cur = ($v =~ /<([^>]+)>/sg);
+               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);
+       }
+
+       uniq_mids(\@mids);
+}
+
+# last References should be IRT, but some mail clients do things
+# out of order, so trust IRT over References iff IRT exists
+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));
+               }
        }
-       my @mids = (join(' ', @raw) =~ /<([^>]+)>/g);
-       my $mids = scalar(@mids) == 0 ? \@raw: \@mids;
+       uniq_mids(\@mids);
+}
+
+sub uniq_mids ($) {
+       my ($mids) = @_;
        my @ret;
+       my %seen;
        foreach (@$mids) {
                next if $seen{$_};
                push @ret, $_;
@@ -67,12 +100,6 @@ sub uniq_mids {
        \@ret;
 }
 
-sub mids { uniq_mids($_[0], 'Message-Id') }
-
-# last References should be IRT, but some mail clients do things
-# out of order, so trust IRT over References iff IRT exists
-sub references { uniq_mids($_[0], 'References', 'In-Reply-To') }
-
 # RFC3986, section 3.3:
 sub MID_ESC () { '^A-Za-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@' }
 sub mid_escape ($) { uri_escape_utf8($_[0], MID_ESC) }