You have to look at each loading condition and design for the worst case scenario (each wind direction and case will provide a different combination of load on the wall, load on the roof, and MWFRS loading). That said, it's typical to provide a strap for worst case uplift, nails for worst case lateral loads perpendicular to the wall, and blocking for worst case lateral loads from the roof diaphragm going down into the wall below. That will cover it in most houses. If you have an atypical connection situation or the number of fasteners exceeds what you can reasonably put into the connection, you have to look closer at each load case and design accordingly.

Thank you for your valuable input

Pham don't we need blocking from roof diaphram to wall below regardless? maybe not the entire length and every second or third spacing to allow for vented ceilings. If we don't provide this are we relying on the joists nailing alone to transfer that load? if so what's preventing all the joists rolling over all at once

Said the Sky - absolutely that's good detailing practice. But, at least in my little corner of the world, it's far from standard construction practice and I have to fight it every time (or the contractor just doesn't put it in). As for what prevents it...I'd say it's usually the fascia board combined with what little rotational stiffness is present in the typical rafter to top plate connection and sheathing to rafter connection. We may ignore them in design, but they are there in varying degrees and influence the observed behavior.

For larger heel heights, we have transferred the lateral loading down from the roof deck to the wall below using 2x blocking at the top chord and placing sheathing rated for the shear wall loading req from that blocking to the stud wall (or CMU wall) below.

- Molly Richardson, EIT (awaiting PE licensure)

“The Enemy of the best is the good. If you’re always settling with what’s good, you’ll never be the best.“ --Jerry Rice