The way I read this is that even given the source code you still wouldn't trust that the binaries it produced are the same as the existing compiled binaries* so I am forced to conclude that nothing will satisfy you.
* - because apparently the golang compiler does not create bit-identical output when run on different machines despite bit-identical sources.
That would be about the size of it, though I already wrote as much in my initial reply to you on this topic. It should be a fairly uncontroversial concept (not sure why b9ron felt the need to push back), though not perhaps an obvious one, so I mentioned it in order to help readers better understand exactly where they are putting their trust.
It's not a controversial concept, I just wonder what is the point of arguing about all the ways the source code could be compromised when, by your own admission, you don't trust any code you didn't write yourself? It's a bit of an ontological conundrum, really.