Subj : Re: Lockable objects To : comp.programming.threads From : Torsten Robitzki Date : Thu May 19 2005 11:57 pm Uenal Mutlu wrote: > > As I told you: A-U-T-O-M-A-T-I-O-N-!-!-! Did you? If a tool can make any difference out of std::pair > and mutlu::vector than due to the fact that the second container class is into the mutlu namespace. Is this what you have in mind? > If it is documented then you, and any other object, knows the object has > these synchronization members. I'm not an object of yours kind. I'm not a robot! With the same argumentation you can stuff in everything but the kitchen sink into an object. Would that convince you? > You, and any other object, can see > whether it is safe to change the object or not. Huh? Silly robot! It might be save to change the state of an object now. But what will tell you this about the state of you application. Do yourself a favour and read again everything David Schwartz told you. He did a great afford to explain you what you should know about the basics. > In your approach nobody knows of your standalone mutex, whether > if it exists or not, and for what purpose. This is not the case with an > integrated mutex. Can you elaborate on the protocol you have in mind that will make the difference? As for now you just introduced just a mutex with every "not trivial" object. .