Dear Maintainer,

what gimp 2.x does and what 3.x should probably do

1) select a region of an image with mouse
2) edit->copy
3) edit->paste-as-new-image

new image is created, with same size as copied subset of original image

what 3.x actually does

new image is created, with same size as original image, with the only
part of that image which is
defined as something other than transparent is what was copied

this makes making a new image from a subset of an image a non-trivial
task, as before the user *does not need to concern themselves* with
the size of the subset of the image they want to copy, whereas now you
have to measure out how big of an image you need to create a new image
for and then copy/paste into that new image.