http://code.google.com/p/chromium/issues/detail?id=131575
will check this issue.
Created attachment 146569 [details] patch1 Root node of radionode list needs to updated to form if detached from dom tree
Comment on attachment 146569 [details] patch1 View in context: https://bugs.webkit.org/attachment.cgi?id=146569&action=review Please fix nits in the test before you land. > Source/WebCore/dom/Node.cpp:2941 > + ASSERT(hasTagName(formTag)); > + if (!hasRareData() || !rareData()->nodeLists()) > + return; > + > + NodeListsNodeData::RadioNodeListCache cache = rareData()->nodeLists()->m_radioNodeListCache; > + for (NodeListsNodeData::RadioNodeListCache::iterator it = cache.begin(); it != cache.end(); ++it) > + it->second->setRootElement(toElement(this)); It seems like we have the same problem in LabelsNodeList and others that can be rooted at the document node. We might want to fix them as well in a follow up patch. > LayoutTests/fast/forms/radionodelist-whose-form-element-detached-from-domtree.html:10 > + <input type="radio" name="type1" value="rbarrval1"> Why is the value "rbarrval1"? Can we use some other human-understandable value instead? > LayoutTests/fast/forms/radionodelist-whose-form-element-detached-from-domtree.html:18 > + <input type="checkbox" name="cb" value="cbval1"> Ditto.
Created attachment 146634 [details] patch_for_landing Fixed nits suggested.
(In reply to comment #3) Thanks for reviewing this patch. > > It seems like we have the same problem in LabelsNodeList and others that can be rooted at the document node. > We might want to fix them as well in a follow up patch. > Will check those. > Why is the value "rbarrval1"? Can we use some other human-understandable value instead? > Fixed.
Comment on attachment 146634 [details] patch_for_landing View in context: https://bugs.webkit.org/attachment.cgi?id=146634&action=review > LayoutTests/fast/forms/radionodelist-whose-form-element-detached-from-domtree-expected.txt:10 > +PASS radioNodeList1[0].value is 'val1' Please spell out value.
Created attachment 146647 [details] patch Better element values
Comment on attachment 146647 [details] patch Rejecting attachment 146647 [details] from commit-queue. Failed to run "['/mnt/git/webkit-commit-queue/Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '-..." exit_code: 1 ERROR: /mnt/git/webkit-commit-queue/LayoutTests/ChangeLog neither lists a valid reviewer nor contains the string "Unreviewed" or "Rubber stamp" (case insensitive). Full output: http://queues.webkit.org/results/12918649
Comment on attachment 146647 [details] patch Clearing flags on attachment: 146647 Committed r119891: <http://trac.webkit.org/changeset/119891>
All reviewed patches have been landed. Closing bug.
Thanks rniwa, I did not see that commit failed, I was not available, it was 2AM already.
On my second thought, this patch doesn't fix the bug completely in some edge cases.
I'm rolling out the patch because the fix is incorrect.
Apparently sheriffbot can't roll out this patch. I'll file a new bug and manually revert this change and make the right fix.