One thought, plus could I get a quick test for this?
http://codereview.appspot.com/1722047/diff/1/2 File java/gadgets/src/main/java/org/apache/shindig/gadgets/rewrite/image/BasicImageRewriter.java (right): http://codereview.appspot.com/1722047/diff/1/2#newcode149 java/gadgets/src/main/java/org/apache/shindig/gadgets/rewrite/image/BasicImageRewriter.java:149: (requestedWidth == null || imageInfo.getWidth() <= requestedWidth)) { This is clearly better than status quo, and I'll submit it. Even so, it would be nice to support no_expand=1 with one or the other of resize_w/resize_h (providing a scaled bounding-box). Doing so mandates moving the logic @175 et al higher in the class, to compute whichever is missing from whichever is present. Thoughts? http://codereview.appspot.com/1722047/show
