Fault injection and unexpected requirement injection

For what its worth, Sripathi and I have discussed Ingo's position and we are in no way "turned off" or even discouraged by it. As maintainer Ingo has veto power and we respect his opinion enough to accept it without much argument. I don't agree with the "butt-ugly" statement, but his other arguments are sufficient in my opinion. Depending on the scope of the new work, this may or may not be something we will be able to commit to working on. This is the risk we take for quality, and it isn't necessarily a bad one. If we were getting a lot of requests for this feature we might push back harder, but as it is, this is a useful feature for our development, and we can use it as is for now. If others really need it, they can apply this patch to their builds or help in the effort to integrate fault injection with perf.