user32: 0 vs. INFINITE timeout

[Francois Gouget]

In Windows 9x SendMessageTimeout() and others used INFINITE (i.e. -1) to 
mean an infinite timeout but since NT it is 0 that denotes an infinite 
timeout. Wine's behavior changed to match NT but a couple of timeouts 
got left to -1 in the server calls.

Changelog:
  * dlls/user/message.c

    Francois Gouget <fgouget at codeweavers.com>
    It is 0 that denotes an infinite timeout for the wine server now, 
not -1 (INFINITE).
    Document put_message_in_queue()'s check for INFINITE (done for 
backward compatibility with Windows 9x).

-- 
Francois Gouget
fgouget at codeweavers.com

-------------- next part --------------
Index: dlls/user/message.c
===================================================================
RCS file: /home/wine/wine/dlls/user/message.c,v
retrieving revision 1.100
diff -u -p -r1.100 message.c
--- dlls/user/message.c	30 Jan 2006 18:46:17 -0000	1.100
+++ dlls/user/message.c	24 Feb 2006 10:57:09 -0000
@@ -1372,7 +1372,7 @@ static BOOL post_dde_message( DWORD dest
         req->wparam  = info->wparam;
         req->lparam  = lp;
         req->time    = GetCurrentTime();
-        req->timeout = -1;
+        req->timeout = 0;
         for (i = 0; i < data->count; i++)
             wine_server_add_data( req, data->data[i], data->size[i] );
         if ((res = wine_server_call( req )))
@@ -2151,8 +2151,11 @@ static BOOL put_message_in_queue( DWORD 
 {
     struct packed_message data;
     unsigned int res;
-    int i, timeout = -1;
+    int i, timeout = 0;
 
+    /* Check for INFINITE timeout for compatibility with Win9x,
+     * although Windows >= NT does not do so
+     */
     if (info->type != MSG_NOTIFY &&
         info->type != MSG_CALLBACK &&
         info->type != MSG_POSTED &&