To go back a step, SDC's example is a good illustration of why I'm skeptical that when a chimney penetrates a rake, the chimney's flashings should direct water over the rake rather than back toward the roof side of the chimney.
We know that real-world, even at new construction, wall cladding details will usually be imperfect if not blatantly incorrect, and that as a result at least some water will be entering behind the water resistant barriers and flashings details.
And we know that at older buildings, there will often have been a series of modifications (siding and window installs, additions, wall penetrations for dryer and bath exhaust fans and the like, etc.) which compromise the original wall water-control provisions, often in ways we can't visually assess without removing portions of the cladding and/or trim.
We also know that the investigation and and correction of many of these defects will usually be outside of the scope of the work being performed during a re-roof or even the installation of a roof at new construction. For example, in SDC's examples few roofers are going to be pulling off siding to investigate its installation, except perhaps for a small area at the roof, chimney junction itself - even if they are aware of the potential problems, or even strongly suspect they are present, convincing homeowners to pay for the investigation and repair of such problems will be somewhere between difficult and impossible).
So the the questions confronting someone installing a roof (if they even bother to ask it) is not "Is water getting into the walls?", it's "How much water is getting in, and how fast is it deteriorating the materials it reaches?".
And real-world, given the quality of the actual work by others we see every day, the answer usually is "If it's not too much water, and it's not creating too much damage" (i.e. there is only minor deterioration, and it's going to take decades to cause problems), then "given the nature of the beast, that's about the best we can expect, but let's at least try not to make it any worse...".
And from that perspective, why would a roofer want to direct substantial quantities of water down a wall of unknown but likely imperfect water-resistance, when there is the alternative of sending it down the roof to be eventually conducted away from the rest of the structure by an appropriate roof drainage system?