I've checked the rear group. It's firmly seated. I've gone ahead and firmly screwed the front group back in. Then, shutter in B, cable release and f/3.5, I've taped a yashicamat 124G focus screen on the back of the film chamber. Using an 8x loupe I've placed a far away (>300 feet ) pole in the middle of the frame, where the focusing screen horizontal split it is. I have tried to the best of my capability to achieve perfect focus and then noted down the position on the focusing knob.
I can confirm that perfect (based on my setup above) infinity was reached well before the infinity mark on the focusing knob. After a couple of tests I could reliably place the best infinity focus somewhere between 30 ft and 60 ft on the knob (close to 30).
To check the agreement with the viewing lens, I've then dropped my loupe onto the camera's own focusing screen (the one above the mirror). Again, I've tried to focus on the same far away object and noted down the mark on the focusing knob. This was in line with the 30ft noted above for the taking lens (a tiny bit closer to 30ft than before), suggesting perhaps the two lenses might be in sync, and that the whole assembly is focusing beyond infinite.
One point worth noting - I found it * much * more difficult to achieve perfect infinity focus using the camera's focusing screen into the viewing lens. The screen is very dim and does not have a split image in the centre. I am therefore wondering if the reason for the focusing inaccuracy I'm experiencing is due to my incapability of focusing with sufficient precision using the camera's focusing screen.