On Thu, 2016-11-17 at 21:47 +0000, Adel Boutros wrote: > Hello Alan, > > > I have probably asked this question but suffering from amnesia I have > to ask again if that is the case: > > > For clients using the C++ bindings of proton-c and the > messaging_handler, what should they change to profit from the > proactor or is it an implicit benefit?
Implicit. We will use the proactor to provide a better IO layer underneath but the C++ API will not change. The proactor design is influenced by what we did with the C++ binding, so it should be an easy fit. The cross-thread wakeup feature you reported as slow in the existing C++ binding is part implemented directly on the underlying IO in the proactor, whereas the existing C++ version is cobbled on top of the old reactor - hence I'm hoping it will solve your performance issue. > > > Regards > > Adel > > ________________________________ > From: Alan Conway <[email protected]> > Sent: Thursday, November 17, 2016 9:24:43 PM > To: users > Subject: Proton C "proactor" API for multi-threaded proton > applications > > The proactor API and an example implementation using libuv are now on > master. There is an overview at ~/proton/proton-c/docs/api/index.md > and > API doc in the header files. index.md and proactor.h attached for > your > convenience. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
