Flesh things out more

Author: foolip

source

@@ -13894,6 +13894,7 @@ https://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20HTML%3E%
    <li><code data-x="attr-autocorrect">autocorrect</code></li>
    <li><code data-x="attr-fe-autofocus">autofocus</code></li>
    <li><code data-x="attr-contenteditable">contenteditable</code></li>
+   <li><code data-x="attr-contentname">contentname</code></li>
    <li><code data-x="attr-dir">dir</code></li>
    <li><code data-x="attr-draggable">draggable</code></li>
    <li><code data-x="attr-enterkeyhint">enterkeyhint</code></li>
@@ -67930,6 +67931,17 @@ interface <dfn interface>HTMLTemplateElement</dfn> : <span>HTMLElement</span> {
   default">missing value default</i> are both the <dfn
   data-x="attr-contentmethod-none-state">None</dfn> state.</p>
 
+  <p>The <dfn element-attr for="htmlsvg-global"><code data-x="attr-contentname">contentname</code></dfn>
+  content points to the target for template contentmethod, like an ID attribute.</p>
+  <!-- TODO better words, and should this be in another section since it's not for <template>? -->
+
+  <p class="XXX">Should contentname work on only HTML elements, HTML+SVG+MathML, or any element?</p>
+
+  <p>If specified, the <code data-x="attr-contentname">contentname</code> attribute value must equal
+  the <code data-x="attr-contentname">contentname</code> of another element in the same
+  <span>tree</span>.</p>
+  <!-- TODO: make this non-symmetrical and matching the processing model -->
+
   <p>The <dfn element-attr for="template"><code
   data-x="attr-template-shadowrootmode">shadowrootmode</code></dfn> content attribute is an
   <span>enumerated attribute</span> with the following keywords and states:</p>
@@ -139294,6 +139306,13 @@ dictionary <dfn dictionary>StorageEventInit</dfn> : <span>EventInit</span> {
       parser.</p>
      </dd>
 
+     <dt>If <var>target</var> is a <code>template</code> element whose "contentmethod target" is non-null:</dt>
+
+     <dd>
+      <p>Let <var>adjusted insertion location</var> be inside <var>target</var>'s "contentmethod target",
+      after its last child (if any).</p>
+     </dd>
+
      <dt>Otherwise</dt>
 
      <dd>
@@ -139305,8 +139324,31 @@ dictionary <dfn dictionary>StorageEventInit</dfn> : <span>EventInit</span> {
 
    <li>
     <p>If the <var>adjusted insertion location</var> is inside a <code>template</code>
-    element, let it instead be inside the <code>template</code> element's <span>template
-    contents</span>, after its last child (if any).</p>
+    element:</p>
+
+    <ol>
+     <li><p>If <var>target</var> is a <code>template</code> element whose "contentmethod target" is
+     non-null, let <var>adjusted insertion location</var> be inside <var>target</var>'s
+     "contentmethod target", after its last child (if any).</p></li>
+
+     <li><p>Otherwise, let <var>adjusted insertion location</var> be inside the
+     <code>template</code> element's <span>template contents</span>, after its last child (if
+     any).</p></li>
+    </ol>
+
+    <p class="XXX">We shouldn't handle contentmethod in both places above. But which is correct?
+    If we use the first place, should the above "Otherwise" be able to further update the
+    adjusted insertion location? The test case for this would be something like:</p>
+
+    <pre><code class="html">
+    &lt;template contentname="regular-template">
+    &lt;template contentmethod="append">
+     &lt;template contentname="regular-template">
+      &lt;div class="where-does-this-end-up">&lt;/div>
+     &lt;/template>
+    &lt;/template></code></pre>
+
+    <p class="XXX">Also, does "inside" above mean any descendent, or only children?</p>
    </li>
 
    <li>
@@ -139446,40 +139488,54 @@ dictionary <dfn dictionary>StorageEventInit</dfn> : <span>EventInit</span> {
   <var>element</var>:</p>
 
   <ol>
-   <!-- Either here and before every other call to "adjusted insertion location", or maybe inside
-    "adjusted insertion location", adjust the insertion location based on contentmethod and the
-    first child element of the template. -->
+   <li><p>Let the <var>adjusted insertion location</var> be the <span>appropriate place for
+   inserting a node</span>.</p></li>
+
+   <li><p>If it is not possible to insert <var>element</var> at the <var>adjusted insertion
+   location</var>, abort these steps.</p></li>
+
+   <li><p>If the parser was not created as part of the <span>HTML fragment parsing
+   algorithm</span>, then push a new <span>element queue</span> onto <var>element</var>'s
+   <span>relevant agent</span>'s <span>custom element reactions stack</span>.</p></li>
+
    <li>
-    <p>contentmethod handling:</p>
+    <p>If <var>element</var>'s has a <code data-x="attr-contentname">contentname</code> attribute
+    specified and <var>adjusted insertion location</var> is inside a <code>template</code> element whose
+    <code data-x="attr-template-contentmethod">contentmethod</code> attribute is not in the <span
+    data-x="attr-contentmethod-none-state">None</span> state:</p>
 
     <ol>
-     <li><p>If <var>element</var> has a contentname attribute and is being inserted into a template
-     element with a valid contentmethod attribute, find the target element with the same namespace,
-     localName, and a contentname attribute with the same value.</p></li>
+     <li><p>Let <var>scope</var> be the <code>template</code> element's parent.</p></li>
 
-     <li><p>TODO: is there any whitespace stripping in contentname attribute values?</p></li>
+     <li><p>If <var>scope</var> is a <code>body</code> element, set <var>scope</var> to its
+     parent.</p></li>
 
-     <li><p>If the template element is in <code>body</code>, instead search the whole document so
-      that things in <code>head</code> can be patched.</p></li>
+     <li><p>Let <var>target</var> be the first element in tree-order within <var>scope</var> with
+     a <code data-x="attr-contentname">contentname</code> attribute exactly matching that of
+     <var>element</var>.</p></li>
 
-      <li><p>Return if the target element wasn't found</p></li>
+     <li><p>If <var>target</var> is null, TODO and return.</p></li>
 
-      <li><p>Depending on the contentmethod state, remove existing nodes and update the insertion point.</p></li>
+     <li><p>If <var>target</var> doesn't have the same namespace and localName as <var>element</var>,
+     TODO and return.</p></li>
 
-      <li><p>Question: do we adjust the insertion point simply by setting <var>adjusted insertion location</var>,
-      or is there other bookkeeping that's needed here?</p></li>
-    </ol>
-   </li>
+     <li>
+      <p>Set the <code>template</code> element's "contentmethod target" to <var>target</var>.</p>
 
-   <li><p>Let the <var>adjusted insertion location</var> be the <span>appropriate place for
-   inserting a node</span>.</p></li>
+      <p class="note">The "contentmethod target" will be used later when nodes are to be inserted
+      into the child of the template element that has a contentname attribute.</p>
+     </li>
 
-   <li><p>If it is not possible to insert <var>element</var> at the <var>adjusted insertion
-   location</var>, abort these steps.</p></li>
+     <li>
+      <p>Depending on the <code data-x="attr-template-contentmethod">contentmethod</code> attribute
+      state, potentially remove <var>target</var> or its children and adjust the
+      <code>template</code> element's "contentmethod target" and additionally keep track of the
+      element we should insert before for the "prepend" and "replace" modes.</p>
 
-   <li><p>If the parser was not created as part of the <span>HTML fragment parsing
-   algorithm</span>, then push a new <span>element queue</span> onto <var>element</var>'s
-   <span>relevant agent</span>'s <span>custom element reactions stack</span>.</p></li>
+      <p class="XXX">Should we still insert the element in the following steps? Does it matter?</p>
+     </li>
+    </ol>
+   </li>
 
    <li><p>Insert <var>element</var> at the <var>adjusted insertion location</var>.</p></li>
 
@@ -140290,8 +140346,7 @@ document.body.appendChild(text);
 
       <ol>
        <li><p>TODO: what to we need to do so that elements are inserted into the template element
-       and redirected by our pre-insertion steps? We need to undo whatever it is that causes nodes
-       to be inserted into the DocumentFragment normally.</p></li>
+       and redirected by our pre-insertion steps?</p></li>
       </ol>
      </li>
 
@@ -152811,7 +152866,7 @@ interface <dfn interface>External</dfn> {
           "<code data-x="attr-contentmethod-prepend">prepend</code>"
     <tr>
      <th> <code data-x="">contentname</code>
-     <td> <span data-x="attr-id">HTML elements</span>
+     <td> <span data-x="attr-contentname">HTML elements</span>
      <td> Unique string to target what contentmethod affects <!-- TODO better words -->
      <td> <a href="#attribute-text">Text</a>*
     <tr>