Hello,

The "hack" I planned actually turned out to be no hack. After over 2
hours of code study, I found that the required change was in just one
line :) And moreover, that it's in another file and does not interfere
with my first patch.

So, here goes a second patch. I've done a test on whether it works, but
no massive performance testing yet. Whoever wants a max-speedy FETCH
should allpy both of these patches to 2.0 and test.

Yours, Mikhail Ramendik


--- dbmail-orig/db.c	2004-09-27 16:32:07.000000000 +0400
+++ dbmail-new/db.c	2004-10-26 19:07:47.000000000 +0400
@@ -2,6 +2,8 @@
 /*
   Copyright (C) 1999-2004 IC & S  [EMAIL PROTECTED]
 
+  Modified by Mikhail Ramendik [EMAIL PROTECTED] (MR)
+
   This program is free software; you can redistribute it and/or 
   modify it under the terms of the GNU General Public License 
   as published by the Free Software Foundation; either 
@@ -3576,7 +3578,8 @@
 		 "FROM dbmail_messageblks blk, dbmail_messages msg "
 		 "WHERE blk.physmessage_id = msg.physmessage_id "
 		 "AND msg.message_idnr = '%llu' "
-		 "ORDER BY blk.messageblk_idnr ASC", msg_idnr);
+		 "ORDER BY blk.messageblk_idnr ASC LIMIT 1", msg_idnr);
+/*MR added LIMIT 1 - we are only getting the header, and therefore only use one row anyway */
 
 	if (db_query(query) == -1) {
 		trace(TRACE_ERROR, "%s,%s: could not get message header",

Reply via email to