headless: Require the user to pass in an initial URL

headless/lib/browser/headless_browser_impl.cc

Index: headless/lib/browser/headless_browser_impl.cc
diff --git a/headless/lib/browser/headless_browser_impl.cc b/headless/lib/browser/headless_browser_impl.cc
index 77f463d8caa04484e52b22f22375fbad7943bfed..57564d7ddc8828994e61fb7eaa84003b5e88462a 100644
--- a/headless/lib/browser/headless_browser_impl.cc
+++ b/headless/lib/browser/headless_browser_impl.cc
@@ -31,10 +31,18 @@ HeadlessBrowserImpl::HeadlessBrowserImpl(
 HeadlessBrowserImpl::~HeadlessBrowserImpl() {}
 
 std::unique_ptr<HeadlessWebContents> HeadlessBrowserImpl::CreateWebContents(
+    const GURL& initial_url,
     const gfx::Size& size) {
   DCHECK(BrowserMainThread()->BelongsToCurrentThread());
-  return base::WrapUnique(new HeadlessWebContentsImpl(
-      browser_context(), window_tree_host_->window(), size));
+  std::unique_ptr<HeadlessWebContentsImpl> web_contents =
+      base::WrapUnique(new HeadlessWebContentsImpl(
+          browser_context(), window_tree_host_->window(), size));
+  // We require the user to pass in an initial URL to ensure that the renderer
+  // gets initialized and eventually becomes ready to be inspected. See
+  // HeadlessWebContents::Observer::WebContentsReady.
+  if (!web_contents->OpenURL(initial_url))

[alex clarke (OOO till 29th), 2016/04/06 15:34:53]

Suppose this failed for some reason, how does the user find out why? (i.e. a 404
or maybe a dns lookup fail etc...)

[Sami, 2016/04/06 15:40:16]

The only reason this would fail immediately is if the URL is badly formed. 404s
and everything else will be communicated via the client API.

+    return nullptr;
+  return std::move(web_contents);
 }
 
 scoped_refptr<base::SingleThreadTaskRunner>