Edgar, you are right!
We have both String>>asInteger and String>>asSignedInteger already. The proposed fix for String>>asInteger makes it work like the existing String>>asSignedInteger. There are no senders of #asSignedInteger in the 3.7 image.
Therefore I think that a simpler and better "fix" is to just get rid of the existing String>>asInteger, and rename #asSignedInteger to #asInteger. That will eliminate one method, make #asInteger behave as expected, and no further changes would be required.