> <env> lock_detect [default|oldest|youngest|random]
This command runs the deadlock detector. It directly translates to the lock_detect DB call. It returns either a 0 (for success), a DB error message or it throws a Tcl error with a system message. The first argument sets the policy for deadlock as follows:
This command returns a list of name/value pairs where the names correspond to the C-structure field names of DB_LOCK_STAT and the values are the data returned. This command is a direct translation of the lock_stat DB call.
This command returns a unique locker ID value. It directly translates
to the lock_id DB call.
This command frees the locker allockated by the lock_id call. It directly translates to the lock_id_free DB call.
This is a diagnostic command to set the locker id that will get
allocated next and the maximum id that
will trigger the id reclaim algorithm.
This command gets a lock. It will invoke the lock_get function. After it successfully gets a handle to a lock, we bind it to a new Tcl command of the form $env.lockX, where X is an integer starting at 0 (e.g. $env.lock0, $env.lock1, etc). We use the Tcl_CreateObjCommand() to create the top level locking command function. It is through this handle that the user can release the lock. Internally, the handle we get back from DB will be stored as the ClientData portion of the new command set so that future locking calls will have that handle readily available.
The arguments are:
This command releases the lock referenced by the command. It is
a direct translation of the lock_put
function. It returns either a 0 (for success), a DB error message
or it throws a Tcl error with a system message. Additionally, since
the handle is no longer valid, we will call
Tcl_DeleteCommand()
so
that further uses of the handle will be dealt with properly by Tcl itself.
This command performs a series of lock calls. It is a direct translation of the lock_vec function. This command will return a list of the return values from each operation specified in the argument list. For the 'put' operations the entry in the return value list is either a 0 (for success) or an error. For the 'get' operation, the entry is the lock widget handle, $env.lockN (as described above in <env> lock_get) or an error. If an error occurs, the return list will contain the return values for all the successful operations up the erroneous one and the error code for that operation. Subsequent operations will be ignored.
As for the other operations, if we are doing a 'get' we will create the commands and if we are doing a 'put' we will have to delete the commands. Additionally, we will have to do this after the call to the DB lock_vec and iterate over the results, creating and/or deleting Tcl commands. It is possible that we may return a lock widget from a get operation that is considered invalid, if, for instance, there was a put_all operation performed later in the vector of operations. The arguments are:
This command sets the lock timeout for all future locks in this environment.
The timeout is in micorseconds.