I assume that the intent was that all processes conceptually go through the semaphore waiting queue, even if they don't end up blocking there. So all processes requesting keys would be constrained by the FIFO order. The statement that the method returns if the specified number of keys is available would then be misleading. Either it needs the additional condition "and there are no earlier processes in the semaphore waiting queue", or perhaps the keys are not considered "available" to the new process, because the earlier requesting processes have priority claim on them. I agree that it is unclear, and could be interpreted the way that Jamie suggested, and that this is undesirable. Steven Sharp sharp@cadence.comReceived on Thu Sep 15 12:45:12 2005
This archive was generated by hypermail 2.1.8 : Thu Sep 15 2005 - 12:46:45 PDT