Maybe, though as a hard and fast rule I strongly disagree.
Delaying merge increases risk. It can also make the diff impossible to review. Overhead can be increased by way of merge conflict and re-doubling of efforts that would not be necessary had the update been merged.
Other times, sometimes you want the one step to be launched so you can ensure there is no regression. Sometimes the next steps are going to take a long time and themselves are multi-pronged and would be merged in multiple parallel phases.
Sometimes you just want to get rid of the inappropriate abstraction so that someone can complete task X where inappopriate abstraction was hindering that. In other times, "worse" is in the eye of the beholder, WET'ing the code first can sometimes be the right thing to do, even if it is never re-DRY'ied.
Delaying merge increases risk. It can also make the diff impossible to review. Overhead can be increased by way of merge conflict and re-doubling of efforts that would not be necessary had the update been merged.
Other times, sometimes you want the one step to be launched so you can ensure there is no regression. Sometimes the next steps are going to take a long time and themselves are multi-pronged and would be merged in multiple parallel phases.
Sometimes you just want to get rid of the inappropriate abstraction so that someone can complete task X where inappopriate abstraction was hindering that. In other times, "worse" is in the eye of the beholder, WET'ing the code first can sometimes be the right thing to do, even if it is never re-DRY'ied.