Thanks Lex. Responses inline. I fixed the comment and am stumped on the test, would love to hear ideas. I'll take your LGTM as an invitation to submit w/the test as is.
http://gwt-code-reviews.appspot.com/64812/diff/1/5 File user/src/com/google/gwt/uibinder/rebind/model/ImplicitCssResource.java (right): http://gwt-code-reviews.appspot.com/64812/diff/1/5#newcode67 Line 67: * file rather than look load a resource. On 2009/09/10 18:12:43, Lex wrote: > Sounds just great. > "rather than look load" --> "rather than load" Done. http://gwt-code-reviews.appspot.com/64812/diff/1/5#newcode138 Line 138: * this package It looks like it starts with ResourceOracle and then tries the thread class loader if that fails. Dunno why. I think you just implied that I should change the UiBinder code in general to start using ResourceOracle? If so, I'll put that on the todo list. http://gwt-code-reviews.appspot.com/64812/diff/1/9 File user/test/com/google/gwt/uibinder/sample/client/UiBinderTest.java (right): http://gwt-code-reviews.appspot.com/64812/diff/1/9#newcode281 Line 281: widgetUi.totallyPrivateStyleSpan.getClassName().length() > 0); Ideally I'd look at the computed style rather than the class name. We don't offer an api for that, and I suspect it's not possible or non-trivial cross browser. (Love to hear that I'm wrong.) I've been racking my brain for an alternative and I don't see a way to do it. I'm testing styles that are only accessible via methods on a generated interface, that's the whole point. To see them for real I'd have to declare these methods in a non-generated interface implemented by ui:style, but that defeats the purpose of the test. http://gwt-code-reviews.appspot.com/64812 --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---