* SUMMARY When LayoutTests/fast/frames/viewsource-empty-attribute-value.html is run stand-alone, it fails due to a different frame name. * STEPS TO REPRODUCE 0. ./WebKitTools/Scripts/build-webkit 1. ./WebKitTools/Scripts/run-webkit-tests fast/frames/viewsource-empty-attribute-value.html * RESULTS Tests fails. * REGRESSION Unknown. Only tested with WebKit r32774. * NOTES I thought a similar issue had already been fixed in DumpRenderTree (where the frame number wasn't getting reset between tests). I can't seem to find it, though.
The failure looks like this: --- /tmp/layout-test-results/fast/frames/viewsource-empty-attribute-value-expected.txt 2008-05-07 03:51:40.000000000 -0700 +++ /tmp/layout-test-results/fast/frames/viewsource-empty-attribute-value-actual.txt 2008-05-07 03:51:40.000000000 -0700 @@ -5,7 +5,7 @@ -------- -Frame: '<!--framePath /1/<!--frame0-->-->' +Frame: '<!--framePath //<!--frame0-->-->' -------- <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd"> <hr noshade width=75%>
Unique frame IDs are assigned in sequence through the lifetime of a WebView, so starting with a fresh WebView you get different identifiers.