execCommand('underline' / 'strikeThrough') results in text-decoration: underline / line-through, deleting all other text decorations even when more than one styles are specified.
Created attachment 33278 [details] fixes the bug
Comment on attachment 33278 [details] fixes the bug /CSSComputedStyleDeclaration.cpp change needs testing. window.getComputedStyle(foo).getPropertyCSSValue('text-decoration') RefPtr<CSSValueList> list = CSSValueList::createCommaSeparated(); 1055 RefPtr<CSSValueList> list = CSSValueList::createSpaceSeparated(); needs test too. foo.style.textDecoration 73 void CSSValueList::remove(CSSValue* val) should be called removeFirst or removeAll or should at least be documented as to its behavior. 78 if (m_values.at(index)->cssText() == val->cssText()) is really really sad. consider adding a bool return to remove and then you don't need hasValue 93 static bool applyCommandToFrame(Frame* frame, EditorCommandSource source, EditAction action, CSSMutableStyleDeclaration* style) is a great change and could even be made in a separate patch. it's OK to do in this one though. YOu can just delete: 110116 static bool executeApplyStyle(Frame* frame, EditorCommandSource source, EditAction action, int propertyID, const char* propertyValue) Spacing: 130 Node* nodeToRemove=0; This looks great! We just need more tests!
Created attachment 33304 [details] 2nd submission, no longer adds hasValue to CSSValueList
Comment on attachment 33304 [details] 2nd submission, no longer adds hasValue to CSSValueList getComputedStyle/getComputedStyle-text-decoration.html shoudl be a js test (like how editing/execCommand/resources/toggle-text-decorations.js is). Ping me if you need help converting it. missing newline at EOF: 29 var successfullyParsed = true; 030 \ No newline at end of file Would like to see the converted test. Otherwise this looks fine.
Created attachment 33308 [details] 3rd submission, test for computed style is converted to a js test
Comment on attachment 33308 [details] 3rd submission, test for computed style is converted to a js test Looks fine.
Landed in http://trac.webkit.org/changeset/46251