1 files changed, 38 insertions, 29 deletions

diff --git a/lib/PublicInbox/SearchThread.pm b/lib/PublicInbox/SearchThread.pm
index 38d1aa6e..60f692b2 100644
--- a/lib/PublicInbox/SearchThread.pm
+++ b/lib/PublicInbox/SearchThread.pm
@@ -24,7 +24,16 @@ use PublicInbox::MID qw($MID_EXTRACT);
 
 sub thread {
         my ($msgs, $ordersub, $ctx) = @_;
-        my $id_table = {};
+
+        # A. put all current $msgs (non-ghosts) into %id_table
+        my %id_table = map {;
+                # this delete saves around 4K across 1K messages
+                # TODO: move this to a more appropriate place, breaks tests
+                # if we do it during psgi_cull
+                delete $_->{num};
+
+                $_->{mid} => PublicInbox::SearchThread::Msg::cast($_);
+        } @$msgs;
 
         # Sadly, we sort here anyways since the fill-in-the-blanks References:
         # can be shakier if somebody used In-Reply-To with multiple, disparate
@@ -32,36 +41,21 @@ sub thread {
         # always determine ordering when somebody uses multiple In-Reply-To.
         # We'll trust the client Date: header here instead of the Received:
         # time since this is for display (and not retrieval)
-        _add_message($id_table, $_) for sort { $a->{ds} <=> $b->{ds} } @$msgs;
+        _set_parent(\%id_table, $_) for sort { $a->{ds} <=> $b->{ds} } @$msgs;
         my $ibx = $ctx->{-inbox};
         my $rootset = [ grep {
                         !delete($_->{parent}) && $_->visible($ibx)
-                } values %$id_table ];
-        $id_table = undef;
+                } values %id_table ];
         $rootset = $ordersub->($rootset);
         $_->order_children($ordersub, $ctx) for @$rootset;
         $rootset;
 }
 
-sub _get_cont_for_id ($$) {
-        my ($id_table, $mid) = @_;
-        $id_table->{$mid} ||= PublicInbox::SearchThread::Msg->new($mid);
-}
-
-sub _add_message ($$) {
-        my ($id_table, $smsg) = @_;
-
-        # A. if id_table...
-        my $this = _get_cont_for_id($id_table, $smsg->{mid});
-        $this->{smsg} = $smsg;
-
-        # saves around 4K across 1K messages
-        # TODO: move this to a more appropriate place, breaks tests
-        # if we do it during psgi_cull
-        delete $smsg->{num};
+sub _set_parent ($$) {
+        my ($id_table, $this) = @_;
 
         # B. For each element in the message's References field:
-        defined(my $refs = $smsg->{references}) or return;
+        defined(my $refs = $this->{references}) or return;
 
         # This loop exists to help fill in gaps left from missing
         # messages.  It is not needed in a perfect world where
@@ -70,7 +64,8 @@ sub _add_message ($$) {
         my $prev;
         foreach my $ref ($refs =~ m/$MID_EXTRACT/go) {
                 # Find a Container object for the given Message-ID
-                my $cont = _get_cont_for_id($id_table, $ref);
+                my $cont = $id_table->{$ref} //=
+                        PublicInbox::SearchThread::Msg::ghost($ref);
 
                 # Link the References field's Containers together in
                 # the order implied by the References header
@@ -96,22 +91,31 @@ sub _add_message ($$) {
 }
 
 package PublicInbox::SearchThread::Msg;
+use base qw(PublicInbox::Smsg);
 use strict;
 use warnings;
 use Carp qw(croak);
 
-sub new {
+# declare a ghost smsg (determined by absence of {blob})
+sub ghost {
         bless {
-                id => $_[1],
+                mid => $_[0],
                 children => {}, # becomes an array when sorted by ->order(...)
-        }, $_[0];
+        }, __PACKAGE__;
+}
+
+# give a existing smsg the methods of this class
+sub cast {
+        my ($smsg) = @_;
+        $smsg->{children} = {};
+        bless $smsg, __PACKAGE__;
 }
 
 sub topmost {
         my ($self) = @_;
         my @q = ($self);
         while (my $cont = shift @q) {
-                return $cont if $cont->{smsg};
+                return $cont if $cont->{blob};
                 push @q, values %{$cont->{children}};
         }
         undef;
@@ -122,7 +126,7 @@ sub add_child {
         croak "Cowardly refusing to become my own parent: $self"
           if $self == $child;
 
-        my $cid = $child->{id};
+        my $cid = $child->{mid};
 
         # reparenting:
         if (defined(my $parent = $child->{parent})) {
@@ -148,8 +152,13 @@ sub has_descendent {
 # being folded/mangled by a MUA, and not a missing message.
 sub visible ($$) {
         my ($self, $ibx) = @_;
-        ($self->{smsg} ||= eval { $ibx->smsg_by_mid($self->{id}) }) ||
-         (scalar values %{$self->{children}});
+        return 1 if $self->{blob};
+        if (my $by_mid = $ibx->smsg_by_mid($self->{mid})) {
+                %$self = (%$self, %$by_mid);
+                1;
+        } else {
+                (scalar values %{$self->{children}});
+        }
 }
 
 sub order_children {