From: Stefan B. <bo...@ap...> - 2007-03-29 04:14:52
|
[I haven't seen David's post to this list so far. I guess it is sittting in the moderation queue for some reason. Unfortunately I am not a moderator of this list.] On Wed, 28 Mar 2007, David Carver <dc...@st...> wrote: >> The later means that in cases where the control document contains >> nodes not present in the test document the type of difference would >> change. This would be a backwards incompatible change (right now >> you'd get a "element tag name is different") but only affect people >> using DetailedDiff. >> >> What do you think? >> >> Stefan >> > Any possibility of deprecating the old methods, and creating some > new methods to handle the new functionality. The way I envision it there is no new method. Consider Control: <a><b/><c/></a> Test: <a><b/></a> Today you get two differences from DetailedDiff: (1) "number of childern on <a> is different" and (2) "expected <c/> but was <b/>". I'd like to turn the second into "element <c/> not found". The other way around Control: <a><b/></a> Test: <a><b/><c/></a> only leads to a single difference from DetailedDiff today: (1) "number of childern on <a> is different". I'd like to add "element <c/> not found". I wouldn't want to touch Control: <a><b/><c/></a> Test: <a><b/><d/></a> where the only difference would be "expected <c> but was <d>". Stefan |