On 28-Feb-08, at 10:29 AM, Anastasia Cheetham wrote: > Some observations (so far):
Further observations: Using jsUnit, the test HTML markup is not actually rendered on-screen, but is essentially 'virtually' rendered in an off-screen model. This actually worked in our favor, because the Lightbox grid flows the images according to the window width, and the off-screen model indirectly resulted in a consistent window size, allowing the tests to make assumptions about where the thumbnails are in relation to each other. With the new testing framework, the test HTML markup is actually rendered, which means that the spatial arrangement of the thumbnails is dependent on the window size, and our assumptions no longer hold. We have ways of forcing the grid into a certain size, and I'm working with that now, so we should be able to address this. Also: The tests are not terribly successful in IE on Windows. I haven't fully identified what's going on, but the first test runs, but seems to simply fail outright in its attempt to move an image by simulating a control+arrow. Subsequent tests encounter an undefined error attempting to focus the Lightbox. This seems to be related to setup and teardown issues, because this happens regardless of the order the tests are run (first test runs, next test has error). Still investigating... -- Anastasia Cheetham [EMAIL PROTECTED] Software Designer, Fluid Project fluidproject.org Adaptive Technology Resource Centre / University of Toronto _______________________________________________ fluid-work mailing list [email protected] http://fluidproject.org/mailman/listinfo/fluid-work
