How do you know for sure there is oil in the water?

A compression test ,while quick ,is less useful than a leak down test. A compression test tells you there is a compression leak, a leakdown test tells you there is a leak, how much and where the leak is.

For engines in general, the only way for oil to get in the coolant _THROUGH A HEAD GASKET_ is when there is a breach between the oil pressure feed and a coolant passage. This type of failure isn't very common but is possible depending on engine design. However, a broken head bolt can cause this type of failure.

A more common though still less likely point of failure is the head cracking between a oil passage and coolant.

More rare is a block crack that extends from a oil pressure passage and coolant. This was a common failure point on the 80's GM 2.8 V6 where they would crack in the lifter valley.

Does the car have a engine oil cooler that connects to the cooling system? This is a common source of failure. I'm not sure of the configuration on the 350 but it may be a laminated stack on the oil filter housing.

Another possible source is the turbo center housing if it is water cooled, though I don't have any data to say how likely it is the problem.

The 350SD is reported to be a rod bender due to much too thin rods. There was a factory fix though I do not know how to ID what motors have been upgraded.