Alban Gruin [off-list ref] writes:
This introduces a visible change: as we do not switch on the tip of the
branch to rebase, no reflog entry is created at the beginning of the
rebase for it.
Fortunately, this does not mean "git log @{-1}.." during a rebase
starts to behave differently. If it were the case, that would have
been a bad regression, but that is not the case.
If you omit one reflog entry, you are creating TWO changes. The
detaching of the HEAD to the "onto" commit would record, just like
any other reflog entry, would record from which commit we detached
HEAD from. That reflog entry will also be different, as you'd be
switching from a different commit.
I am not sure what the implication of this difference will be in
practice, though, but it must be smaller than the effect of the
missing reflog entry discussed earlier.