[PATCH 6/6] lib: choose oldest message when breaking reference loops

[David Bremner]

This preserves a sensible thread order
---
 lib/thread.cc    | 33 ++++++++++++++++++++++++---------
 test/T050-new.sh |  1 -
 2 files changed, 24 insertions(+), 10 deletions(-)

diff --git a/lib/thread.cc b/lib/thread.cc
index dbac002f..a6dc4e5a 100644
--- a/lib/thread.cc
+++ b/lib/thread.cc
@@ -390,20 +390,18 @@ _thread_add_matched_message (notmuch_thread_t *thread,
 static void
 _resolve_thread_relationships (notmuch_thread_t *thread)
 {
-    notmuch_message_node_t *node;
+    notmuch_message_node_t *node, *first_node;
     notmuch_message_t *message, *parent;
     const char *in_reply_to;
 
-    for (node = thread->message_list->head; node; node = node->next) {
+    first_node = thread->message_list->head;
+    if (!first_node)
+	return;
+
+    for (node = first_node->next; node; node = node->next) {
 	message = node->message;
 	in_reply_to = _notmuch_message_get_in_reply_to (message);
-	/*
-	 * if we reach the end of the list without finding a top-level
-	 * message, that means the thread is a cycle (or set of
-	 * cycles) and any message can be considered top-level
-	 */
-	if ((thread->toplevel_list->head || node->next) &&
-	     in_reply_to && strlen (in_reply_to) &&
+	if (in_reply_to && strlen (in_reply_to) &&
 	    g_hash_table_lookup_extended (thread->message_hash,
 					  in_reply_to, NULL,
 					  (void **) &parent))
@@ -412,6 +410,23 @@ _resolve_thread_relationships (notmuch_thread_t *thread)
 	    _notmuch_message_list_add_message (thread->toplevel_list, message);
     }
 
+    /*
+     * if we reach the end of the list without finding a top-level
+     * message, that means the thread is a cycle (or set of cycles)
+     * and any message can be considered top-level.  Choose the oldest
+     * message, which happens to be first in our list.
+     */
+    message=first_node->message;
+    in_reply_to = _notmuch_message_get_in_reply_to (message);
+    if (thread->toplevel_list->head &&
+	in_reply_to && strlen (in_reply_to) &&
+	g_hash_table_lookup_extended (thread->message_hash,
+				      in_reply_to, NULL,
+				      (void **) &parent))
+	_notmuch_message_add_reply (parent, message);
+    else
+	_notmuch_message_list_add_message (thread->toplevel_list, message);
+
     /* XXX: After scanning through the entire list looking for parents
      * via "In-Reply-To", we should do a second pass that looks at the
      * list of messages IDs in the "References" header instead. (And
diff --git a/test/T050-new.sh b/test/T050-new.sh
index 320a7646..9025fa7a 100755
--- a/test/T050-new.sh
+++ b/test/T050-new.sh
@@ -359,7 +359,6 @@ test_begin_subtest "reference loop does not crash"
 test_expect_code 0 "notmuch show --format=json id:mid-loop-12 at example.org id:mid-loop-21 at example.org > OUTPUT"
 
 test_begin_subtest "reference loop ordered by date"
-test_subtest_known_broken
 threadid=$(notmuch search --output=threads  id:mid-loop-12 at example.org)
 notmuch show --format=mbox $threadid | grep '^Date'  > OUTPUT
 cat <<EOF > EXPECTED
-- 
2.17.0