Typo in chapter "11.3.1 The bootstrap"?

[Epi17]

In the function for the bootstrapping procedure, the propensity score model (seven_dwarfs_9) is currently fitted using the original dataset(seven_dwarfs_9). However, based on my understanding, to properly account for the uncertainty in estimating the propensity scores, the model should instead be fitted on the bootstrapped dataset (.df), i.e., the resampled data in the current bootstrap iteration.

Is this an oversight, or is there a reason for fitting the model to the original data rather than the resampled one?