Both, as both edits improve the post.

Why do you believe that it is ok to rejected an edit which improves the post?

@Ramhound I doubt that this is the case, if you look at the data you will find that those users are part of the lower range of the acceptance rate, while there are also reviewers with thousands reviews on the upper end of the acceptance rate. But let's focus on the real question: Why are substantial edits rejected when they don't also address all, potential minor, issues?

Superuser was probably my first SX account ever. I doubt that the 80% are representative regarding the current situation as the time range is over mutliple years.

I am not aware that such a rule exists anywhere within the stackexchange network. I believe that this is policy is followed by a small but active superuser reviwers group for the false reasons. Eventually, I fear, causing harm to the community by driving away contributors.

I always try to state questions in a neutral emotionless manner. If I didn't achieve this in your opinion, then I'm happy about your suggestions how to improve it. Feel free to suggest an edit.

"Just because you don't like seeing your edits rejected isn't a good reason." Granted, I was affected by this. But this merely caused me to realize that it appears that many edits on superuser get rejected for no good reason. I am for example unlikely going to edit again until edits that improve a post are generally accepted.

"Not every user has the free time to click the 'improve' or 'reject' and 'edit' buttons either." That is a argument to use the 'accept' button in my eyes. If the edit does improve the text, why not? And if in doubt there is always the 'Skip' button.

"…and not cause unnecessary churn on the front page." If this is really considered a problem by the majority, then it shouldn't be used as argument to reject good edits. Instead edits simply should not cause posts to get bumped on the front page.