On Jun 28, 2006, at 9:35 AM, Jesus Cea wrote:

> David Binger wrote:
>> Maybe we can address this by making sure that new ghosts don't
>> have their _p_touched set to  the current sync_count.
>
> +1

Now that I have looked at the code, it appears that new ghosts
have their _p_touched set to 0, which is only the same as the
connection.sync_count when connection.sync_count also happens
to be 0.

Currently, the sync_count's initial value is 0.  I intend
to change it so that the initial value is 1.  I think this will
remove this corner case conflict possibility.