https://codereview.appspot.com/573770043/diff/553990043/lily/stencil-integral.cc File lily/stencil-integral.cc (right): https://codereview.appspot.com/573770043/diff/553990043/lily/stencil-integral... lily/stencil-integral.cc:966: copy.raise (y_pos[i] - my_y); hm, the old code passed different arguments in the two branches: copy.shift (x_pos[i] - my_x); vs. copy.shift (y_pos[i] - my_y); and copy.raise (y_pos[i] - my_y); vs. copy.raise (x_pos[i] - my_x); AFAICS this has nothing to do with commutativity of shift and raise.
https://codereview.appspot.com/573770043/diff/553990043/lily/stencil-integral.cc File lily/stencil-integral.cc (right): https://codereview.appspot.com/573770043/diff/553990043/lily/stencil-integral... lily/stencil-integral.cc:966: copy.raise (y_pos[i] - my_y); On 2020/05/01 20:59:35, hahnjo wrote: > hm, the old code passed different arguments in the two branches: > copy.shift (x_pos[i] - my_x); > vs. > copy.shift (y_pos[i] - my_y); > > and > copy.raise (y_pos[i] - my_y); > vs. > copy.raise (x_pos[i] - my_x); > > AFAICS this has nothing to do with commutativity of shift and raise. Ah, you are right. Silly me.
d'uh
against HEAD^
LGTM (provided there are no regtest differences)