Lines Matching full:be
75 // stop() should be safe to call, because either we don't have a worker in ~this()
76 // thread running and it is a no-op anyway, or it is guaranteed to be in ~this()
78 // be garbage collected yet. in ~this()
91 // Keep track that there is a new work item to be processed.
96 // We should be able to send the control message as a whole – we currently
97 // assume to be able to receive it at once as well. If this proves to be
100 // be changed to a more sophisticated scheme.
229 // Technically, only half barriers would be required here, but adding the in ensureWorkerThreadRunning()
259 // should be available, as we got notified by libevent). in receiveControlMsg()
269 "never be executed: %s", socketErrnoString(errno)); in receiveControlMsg()
313 "instead of %s), some work item will possibly never be executed.", in receiveControlMsg()
333 // This should never happen, just to be sure the worker thread in executeWork()
341 // re-lookup this on every iteration. This could be solved, but I'd in executeWork()
415 /// ignored and can be used to wake up the worker thread.
420 /// expected that work items can always be read at once from it, i.e. that
421 /// there will never be short reads.
428 // TODO: This should really be of some queue type, not an array slice, but