Please review: Bug 571677 - Busy replica on consumers when directly deleting a replication conflict

[Rich Megginson]

https://bugzilla.redhat.com/show_bug.cgi?id=571677

The problem is that a bepreop can attempt to lock an entry in the cache 
that has already been locked by the caller.  I noticed that the modrdn 
code does not lock the target entries in the cache until after the 
bepreops have been called.  So I did the same for mod and del, and it 
worked, but I just don't know what else this may effect.  Comments?

https://bugzilla.redhat.com/attachment.cgi?id=402171&action=diff

https://bugzilla.redhat.com/attachment.cgi?id=402171&action=edit