Should the row heights switcher (state set by the user):

1. Override the rowHeights overrides (state set by the consumer)?

2. Or, should the control be automatically disabled if rowHeights exists (we assume that the consumer does not want their overrides to be overridden)?

See #5372 (review)

[...] when the user selects an option, change it for all the rows even the very specifically set by index ones [or] If the setting seems to be completely controlled/set by the consumer, or a selection can't be determined, just hide it.
This is probably the easier option and probably aligns more with the consumer's intention. If they're explicitly setting row heights for individual rows, then they probably don't expect the content to fit any other way.

If a consumer has deliberately chose specific heights for specific rows, my guess is they don't want the consumer to adjust the row heights at all.
As a user, if I chose to specifically set the row heights to single, and saw that row 2, 4, and 5 were the only ones not adjusting, I'd think there was a bug in the UI. So I'm more trying to find a solution that aligns the consumer's intent with the user's expectation. We can't truly know both, but we can predict by defaulting to specific outcomes and providing an "out".