I really appreciate the honesty. It is not the end of the world to delay a highly anticipated feature, especially when the delay was a consequence of prioritizing the long-term stability of the codebase over the promised delivery date.
There are costs to not shipping a feature you meant to ship. So we first get a bit of sunk cost fallacy, trying to stretch to meet the goal, then if we pull back we have to make sure the feature is either completely toggled off or reverted. And then you have to fix the docs, which is never a fast process. It always takes longer wall-clock time to do the grunt tasks than strictly necessary to perform the steps.
