LayoutTests/fast/dom/Range/range-clone-contents.html only fails on Windows bots. It would be interesting to know if disabling optimization on WebCore/dom/Range.cpp affects the results of the test. The Range::processContents() method reuses a lot of variables, also shadows many variables, and uses multiple variables within for() loop conditions.
Added the test to the Windows skip list: Committed r75889. <http://trac.webkit.org/changeset/75889>
<rdar://problem/8870569>
Brian/Adam: Are there any known issues where DOMAllInOne.cpp isn't always rebuilt properly for Windows if one of its constituent files is changed?
Does anyone on Windows see this failure in their local build?
Created attachment 79114 [details] Patch to try Patch to try if there is a codegen issue with MSVC compiler.
(In reply to comment #3) > Brian/Adam: Are there any known issues where DOMAllInOne.cpp isn't always rebuilt properly for Windows if one of its constituent files is changed? I don't know of any such issues.
A link to the failure diff would be handy.
(In reply to comment #7) > A link to the failure diff would be handy. Here are links to the failure diffs from Bug 52512 Comment #7: Windows 7 Release bot: <http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r75882%20(8257)/fast/dom/Range/range-clone-contents-pretty-diff.html> Windows XP Debug bot: <http://build.webkit.org/results/Windows%20XP%20Debug%20(Tests)/r75882%20(24216)/fast/dom/Range/range-clone-contents-pretty-diff.html>
(In reply to comment #8) > (In reply to comment #7) > > A link to the failure diff would be handy. > > Here are links to the failure diffs from Bug 52512 Comment #7: Thanks! FYI, our standard procedure in cases like this (where the test is just failing, as opposed to crashing or timing out) is to check in platform-specific results that reflect the current behavior, rather than skipping the test.
(In reply to comment #8) > (In reply to comment #7) > > A link to the failure diff would be handy. > > Here are links to the failure diffs from Bug 52512 Comment #7: > > Windows 7 Release bot: <http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r75882%20(8257)/fast/dom/Range/range-clone-contents-pretty-diff.html> Here's the Windows 7 Release build of r75882: <http://build.webkit.org/builders/Windows%20Release%20%28Build%29/builds/10555> > Windows XP Debug bot: <http://build.webkit.org/results/Windows%20XP%20Debug%20(Tests)/r75882%20(24216)/fast/dom/Range/range-clone-contents-pretty-diff.html> And Windows XP Debug build of r75882: <http://build.webkit.org/builders/Windows%20Debug%20%28Build%29/builds/29578>
A local Release build of WebKit r75949 on Windows XP shows that fast/dom/Range/range-clone-contents.html passes. This appears to be an issue with needing a clean build on the Windows buildbots. However, I'll wait for independent verification from another source before removing the test from the Skipped list.
Created attachment 79192 [details] Unskip the test
Comment on attachment 79192 [details] Unskip the test Clearing flags on attachment: 79192 Committed r75970: <http://trac.webkit.org/changeset/75970>
All reviewed patches have been landed. Closing bug.
(In reply to comment #13) > (From update of attachment 79192 [details]) > Clearing flags on attachment: 79192 > > Committed r75970: <http://trac.webkit.org/changeset/75970> Windows 7 Release tests for r75970: <http://build.webkit.org/builders/Windows%207%20Release%20%28Tests%29/builds/8298> Windows XP Debug tests for r75970: <http://build.webkit.org/builders/Windows%20XP%20Debug%20%28Tests%29/builds/24245> The new test now passes.