26248 2014-07-01 15:01:23 +0000 Spec mismatch between createHTMLDocument and textContent 2014-07-02 09:19:15 +0000 1 1 1 Unclassified WebAppsWG DOM unspecified All All RESOLVED WONTFIX P2 normal --- 1 kangil.han annevk bugs bzbarsky mike Ms2ger www-dom public-webapps-bugzilla oldest_to_newest 108612 0 kangil.han 2014-07-01 15:01:23 +0000 DOMImplementation.createHTMLDocument(title)[1] describes like, 'Create a Text node, set its data to title (which could be the empty string), and append it to the title element created in the previous step.' While textContent attribute[2] says, 'If new value is not the empty string, set node to a new Text node whose data is new value.' So it is confusing how to handle empty title DOMString on implementation. [1] http://dom.spec.whatwg.org/#dom-domimplementation-createhtmldocument [2] http://dom.spec.whatwg.org/#dom-node-textcontent test case: http://w3c-test.org/dom/nodes/DOMImplementation-createHTMLDocument.html test name: DOMImplementation.createHTMLDocument 108613 1 kangil.han 2014-07-01 15:03:53 +0000 For the information,(alphabetical order) chrome: fail ff: pass ie: fail 108614 2 bugs 2014-07-01 15:17:41 +0000 What is exactly confusing? textnode.data = ""; sets text node's content to "". element.textContent = ""; removes all the child nodes. element.textContent = "foo"; removes all the child nodes and creates a text node with data "foo" and appends it to element. 108618 3 bzbarsky 2014-07-01 15:50:14 +0000 Basically, createHTMLDocument("foo") does NOT set .textContent = "foo" on the <title> element. It does something slightly different that guarantees there is a text node inside that <title>. Looking at Chrome's code, it implements createHTMLDocument("foo") by doing: document.title = "foo"; which is defined in the spec in terms of setting textContent on the <title> element, which in fact does not match what the DOM spec says to do for createHTMLDocument. 108624 4 Ms2ger 2014-07-01 19:25:50 +0000 (In reply to Kangil Han from comment #1) > For the information,(alphabetical order) > chrome: fail > ff: pass > ie: fail Servo: pass 108627 5 annevk 2014-07-01 19:45:32 +0000 The specification does not seem confusing, but I guess comment 0 is actually asking if we want to align the two cases. I have no opinion really. 108637 6 kangil.han 2014-07-02 09:19:15 +0000 Thanks for comments! I am fixing chromium behavior as per comment 3 suggested by Boris. https://codereview.chromium.org/366883003/