> This is not likely an issue of dodgy drivers as it is a non standard DAW ASIO practice SONAR uses. Well RME is seemingly the only interface company having issues right now, and like I say the EXACT problem happened with Focusrite which they fixed. RME pointing to "allow arm change during playback/record" is a total smoking gun. Did RME test their drivers with Cakewalk? Obviously not until now. Go figure! :)
BTW ASIO is protocol, there really isn't a "non standard" way of consuming it, it either works or it doesn't And Cakewalk supplies an off button as a workaround for those drivers that do not support Cakewalk (or are buggy) so they've covered that scenario as well.
Expect a fix from RME and not from Cakewalk, and when RME releases the fix and it works without any change in Cake's code, then expect everybody to somehow believe it's still Cake's fault (as is the norm) even though what they do is actually to help the third parties to fix their own issues when consuming Cake's hook in points :)
Cheers...!