nit: this is already calculated.

Do we need an instance variable? In _update we always need to recompute the value so maybe we can use a local variable instead?

Can you take another look at this? Sorry, I was in the middle of changing the PR, I should have used a WIP label. As it is now, I need the previousFocalPoint to calculate delta (since now it's just the delta since the last call, not since the start). So I think I either need _delta or _previousFocalPoint as an instance variable.

Ah sorry I thought ScaleUpdateDetails was constructed in _update nvm.

if the receiving render object moved then _lastTransform is also changed right (I'm assuming _lastTransform is the transform from the global coordinate space to the local coordinate space). Should we use a local variable to track localFocalPoint?

Done, good idea.

You don't have to compute localPreviousFocalPoint I think? It's just _localFocalPoint before you updated it in line 425.

Ah, I was working around the fact that _localFocalPoint is late and won't be initialized on the first run here, so I can't grab the value out of it here.

I've updated it to an if/else that avoids the recalculation, let me know if you like the old way better.

Maybe we should use a more self-explanatory name instead of delta?

I'm trying to mirror the delta parameter in DragUpdateDetails, but it's true it's not exactly the same thing. Maybe focalPointDelta?

Yeah that sounds good! I was a bit confused about what delta was since there are at least 2 pointers involved in a scale gesture.

nit: instead of using an extra boolean to store _update, is it possible to make _currentFocalPoint an Offset? ?

Yes if I add a few exclamation points where it's used, good call.

You don't have to compute localPreviousFocalPoint I think? It's just _localFocalPoint before you updated it in line 425.