[sup-compat] Don't index mime parts with content-disposition of attachment

[notmuch] / notmuch-index-message.cc

diff --git a/notmuch-index-message.cc b/notmuch-index-message.cc
index dfcb3ba50c67b56f3757c4c8311b4aa96146fec5..6795430b8c72aa8d8e8e34c6027c561851366a92 100644 (file)
--- a/notmuch-index-message.cc
+++ b/notmuch-index-message.cc
@@ -269,12 +269,23 @@ static void
 insert_thread_id (GHashTable *thread_ids, Xapian::Document doc)
 {
     string value_string;
-    const char *value;
+    const char *value, *id, *comma;
 
     value_string = doc.get_value (NOTMUCH_VALUE_THREAD);
     value = value_string.c_str();
-    if (strlen (value))
-       g_hash_table_insert (thread_ids, strdup (value), NULL);
+    if (strlen (value)) {
+       id = value;
+       while (*id) {
+           comma = strchr (id, ',');
+           if (comma == NULL)
+               comma = id + strlen (id);
+           g_hash_table_insert (thread_ids,
+                                strndup (id, comma - id), NULL);
+           id = comma;
+           if (*id)
+               id++;
+       }
+    }
 }
 
 /* Return one or more thread_ids, (as a GPtrArray of strings), for the
@@ -377,9 +388,6 @@ gen_terms_body_str (Xapian::TermGenerator term_gen,
            continue;
 
        /* Also skip lines introducing a quote on the next line. */
-       while (line_end > line && isspace (*line_end))
-           line_end--;
-
        if (*line_end == ':' && *next_line == '>')
            continue;
 
@@ -398,22 +406,47 @@ gen_terms_body_str (Xapian::TermGenerator term_gen,
 
 /* Callback to generate terms for each mime part of a message. */
 static void
-gen_terms_part (GMimeObject *parent,
-               GMimeObject *part,
-               gpointer user_data)
+gen_terms_part (Xapian::TermGenerator term_gen,
+               GMimeObject *part)
 {
-    Xapian::TermGenerator *term_gen = (Xapian::TermGenerator *) user_data;
     GMimeStream *stream;
     GMimeDataWrapper *wrapper;
     GByteArray *byte_array;
+    GMimeContentDisposition *disposition;
     char *body;
 
+    if (GMIME_IS_MULTIPART (part)) {
+       GMimeMultipart *multipart = GMIME_MULTIPART (part);
+       int i;
+
+       for (i = 0; i < g_mime_multipart_get_count (multipart); i++) {
+           if (GMIME_IS_MULTIPART_SIGNED (multipart)) {
+               /* Don't index the signature. */
+               if (i == 1)
+                   continue;
+               if (i > 1)
+                   fprintf (stderr, "Warning: Unexpected extra parts of mutlipart/signed. Indexing anyway.\n");
+           }
+           gen_terms_part (term_gen,
+                           g_mime_multipart_get_part (multipart, i));
+       }
+       return;
+    }
+
     if (! GMIME_IS_PART (part)) {
        fprintf (stderr, "Warning: Not indexing unknown mime part: %s.\n",
                 g_type_name (G_OBJECT_TYPE (part)));
        return;
     }
 
+    disposition = g_mime_object_get_content_disposition (GMIME_OBJECT (part));
+    if (disposition &&
+       strcmp (disposition->disposition, GMIME_DISPOSITION_ATTACHMENT) == 0)
+    {
+       add_term (term_gen.get_document (), "label", "attachment");
+       return;
+    }
+
     byte_array = g_byte_array_new ();
 
     stream = g_mime_stream_mem_new_with_byte_array (byte_array);
@@ -423,9 +456,10 @@ gen_terms_part (GMimeObject *parent,
 
     g_object_unref (stream);
 
+    g_byte_array_append (byte_array, (guint8 *) "\0", 1);
     body = (char *) g_byte_array_free (byte_array, FALSE);
 
-    gen_terms_body_str (*term_gen, body);
+    gen_terms_body_str (term_gen, body);
 
     free (body);
 }
@@ -486,7 +520,7 @@ index_file (Xapian::WritableDatabase db,
     gen_terms (term_gen, "subject", subject);
     gen_terms (term_gen, "body", subject);
 
-    g_mime_message_foreach (message, gen_terms_part, &term_gen);
+    gen_terms_part (term_gen, g_mime_message_get_mime_part (message));
 
     parents = g_ptr_array_new ();