>Such higher level locking actions can be implemented using single-entity >locks, and hence I'd prefer not to introduce the added complexity of >multiple entity simultaneous locks. > >- Jim > If a user locks files A, B, & C with separate locks and performs a PUT to A, there is no guarantee that the locks on B & C have not been already been broken. Thus a mechanism is needed to perform actions on a group of files in such a way that each individual action is guaranteed to be atomic with respect to the lock on those files. Your own paper includes a perfect example of this - A user wishes to lock an entire web site so as to be able to guarantee global consistency even if they are only interested in changing a single file. If they have to lock every file individually, a bandwidth eating action I might add, they still have no guarantee of global consistency as one or more locks could be removed for any number of reasons. A multi-file lock will guarantee that an action will only go through if all the relevant files are still locked. > Yaron >Received on Friday, 6 September 1996 15:54:51 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:43:41 GMT