In loader/reload-subresource-when-type-changes.html: <link rel="prefetch" href="resources/image1.png"> <img src="resources/image1.png" onerror="log('LOAD ERROR - FAIL 2 of 3)" onload="checkLoadedImage(event, 2)"> <script src="resources/image2.png"></script> <img src="resources/image2.png" onerror="log('LOAD ERROR - FAIL 2 of 3)" onload="checkLoadedImage(event, 3)"> Because the last two images might be loaded asynchronously, onload of the third image might be triggered before that of the second image, thus the actual result might be: PASS 3 of 3 PASS 2 of 3 The situation might not happen in normal test environment because the images are all accessed as local files. However, in an environment with slow access (such as NFS, or tests accessed through HTTP), this might happen.
Created attachment 85768 [details] patch
Good catch, and it looks good to me. I like that you save expectations. I'm not a reviewer though. Can I recommend ap or abarth peek at this to r+ it for you?
(In reply to comment #2) > Good catch, and it looks good to me. I like that you save expectations. > > I'm not a reviewer though. Can I recommend ap or abarth peek at this to r+ it for you? Sure. Thanks!
ping...
Comment on attachment 85768 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=85768&action=review > LayoutTests/ChangeLog:6 > + https://bugs.webkit.org/show_bug.cgi?id=56360 You have good information in the first comment of the bug about why and how this fails. Usually it is best practice to put that information in the ChangeLog so that when people go back to understand why things are the way they are they can see it here instead of having to wade through the bug.
Created attachment 89151 [details] patch with updated ChangeLog
Comment on attachment 89151 [details] patch with updated ChangeLog Clearing flags on attachment: 89151 Committed r83554: <http://trac.webkit.org/changeset/83554>
All reviewed patches have been landed. Closing bug.