Re: [w3c/permissions] Introduce "Automation" section (#151)

foolip commented on this pull request.

Did a bit of surface-level reviewing before realizing I'm not the best person to review the details of this. Can you try to find people working on WebDriver implementations for each vendor and ask for their review? At least one would be good.

> @@ -1074,6 +1085,221 @@ spec: webidl
     </p>
   </section>
 </section>
+<section>
+  <h2 id="automation">
+    Automation
+  </h2>
+  <p>
+    For the purposes of user-agent automation and application testing, this
+    document defines the following <a>errors</a>, algorithms, and <a>extension
+    commands</a> for the [[WebDriver]] specification.
+  </p>
+
+  The <dfn>no such request</dfn> <a>error</a> has a [=response/status=] of 404

This is adding a line to the table in https://w3c.github.io/webdriver/webdriver-spec.html#handling-errors, but the link to https://fetch.spec.whatwg.org/#concept-response-status and no link to and no use of "JSON Error Code" (can't be linked?) made that less than clear. A table in the same style with a single row would work, but I'll leave that to your editorial discretion.

> @@ -1074,6 +1085,221 @@ spec: webidl
     </p>
   </section>
 </section>
+<section>
+  <h2 id="automation">
+    Automation
+  </h2>
+  <p>
+    For the purposes of user-agent automation and application testing, this
+    document defines the following <a>errors</a>, algorithms, and <a>extension
+    commands</a> for the [[WebDriver]] specification.

Rendered as "[WebDriver]", I think missing a "!" to make it a normative reference and presumably fix the problem?

> +          </li>
+        </ol>
+      </li>
+      <li>Return false.
+    </ol>
+  </div>
+
+  <section>
+    <h3 id="grant-command">
+      Grant
+    </h3>
+    <table>
+      <tbody>
+        <tr>
+          <th>HTTP Method</th>
+          <th><a lt="extension command prefix">Prefix</a></th>

What will the resulting URLs look like? https://w3c.github.io/webdriver/webdriver-spec.html#protocol-extensions talks about "vendor-specific needs" and "Commands that are specific to a user agent are called extension commands", which isn't the case here.

Should we perhaps just be defining new commands in the style of other commands in the WebDriver spec with a "Path Template"? If the resulting URLs are the same it doesn't matter, but I suspect not, and it'd be a bit silly if you could tell from the URL where something was specified.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/w3c/permissions/pull/151#pullrequestreview-45267575