On Wed, Apr 22, 2009 at 03:49:37PM -0700, Matthew Dillon wrote:
>       fetch http://apollo.backplane.com/DFlyMisc/hammer01.patch

Matt, shouldn't the following chunk (in above patch) be tested/pushed
independently of the rest of the patch?

@@ -2308,9 +2336,10 @@ hammer_sync_record_callback(hammer_record_t record, void 
*data)
         * Assign the create_tid for new records.  Deletions already
         * have the record's entire key properly set up.
         */
-       if (record->type != HAMMER_MEM_RECORD_DEL)
+       if (record->type != HAMMER_MEM_RECORD_DEL) {
                record->leaf.base.create_tid = trans->tid;
                record->leaf.create_ts = trans->time32;
+       }
        for (;;) {
                error = hammer_ip_sync_record_cursor(cursor, record);
                if (error != EDEADLK)

Reply via email to