Set active document url to <template> content URL.

third_party/WebKit/Source/core/html/HTMLTemplateElement.cpp

 /*
  * Copyright (C) 2012 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 28 matching lines @@
 using namespace HTMLNames;
 
 inline HTMLTemplateElement::HTMLTemplateElement(Document& document)
     : HTMLElement(templateTag, document), content_(this, nullptr) {}
 
 DEFINE_NODE_FACTORY(HTMLTemplateElement)
 
 HTMLTemplateElement::~HTMLTemplateElement() {}
 
 DocumentFragment* HTMLTemplateElement::content() const {
-  if (!content_)
+  if (!content_) {
     content_ = TemplateContentDocumentFragment::Create(
         GetDocument().EnsureTemplateDocument(),
         const_cast<HTMLTemplateElement*>(this));
+    content_->GetDocument().SetURL(GetDocument().Url());

[fs, 2017/04/25 09:13:45]

Shouldn't this rather be handled when creating the template document?

[foolip, 2017/04/25 09:23:44]

The relevant bit of spec:
https://html.spec.whatwg.org/multipage/scripting.html#template-contents
https://html.spec.whatwg.org/multipage/scripting.html#appropriate-template-co...

AFAICT, there's nothing that sets that document's URL.

[fs, 2017/04/25 10:20:47]

Right, https://dom.spec.whatwg.org/#interface-document only calls for 'setting'
the origin. (Maybe that should be asserted by the test as well then.)

[foolip, 2017/04/25 14:23:28]

Do you mean "The Document() constructor, when invoked, must return a new
document whose origin is the origin of current global object’s associated
Document"? It's a bit odd, but I'm pretty sure that is for the script-exposed
constructor only, and that when HTML says "Let /new doc/ be a new Document",
it's using a magical internal constructor.

[fs, 2017/04/25 14:39:26]

That may very well be the case. But shouldn't the magic internal constructor be
well-defined? (Based on the discrepancies observed here, one might argue that it
isn't.)

[fs, 2017/04/25 14:51:08]

Found what I was looking for:

'Unless stated otherwise, a document’s encoding is the utf-8 encoding, content
type is "application/xml", URL is "about:blank", origin is an opaque origin,
type is "xml", and its mode is "no-quirks".'

So, our current behavior seems right.

+  }
 
   return content_.Get();
 }
 
 Node* HTMLTemplateElement::cloneNode(bool deep, ExceptionState&) {
   if (!deep)
     return CloneElementWithoutChildren();
 
   Node* clone = CloneElementWithChildren();
   if (content_)
@@ skipping 12 matching lines @@
   visitor->Trace(content_);
   HTMLElement::Trace(visitor);
 }
 
 DEFINE_TRACE_WRAPPERS(HTMLTemplateElement) {
   visitor->TraceWrappers(content_);
   HTMLElement::TraceWrappers(visitor);
 }
 
 }  // namespace blink