Adding <keygen> Content Setting (Blink)

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

Index: third_party/WebKit/Source/core/html/HTMLKeygenElement.cpp
diff --git a/third_party/WebKit/Source/core/html/HTMLKeygenElement.cpp b/third_party/WebKit/Source/core/html/HTMLKeygenElement.cpp
index a7eaec25442f22b924b0ac68a650a238583c2303..2c26131278e49c55a681f13ca29240bc58a8e6c2 100644
--- a/third_party/WebKit/Source/core/html/HTMLKeygenElement.cpp
+++ b/third_party/WebKit/Source/core/html/HTMLKeygenElement.cpp
@@ -35,6 +35,7 @@
 #include "core/html/HTMLOptionElement.h"
 #include "core/html/HTMLSelectElement.h"
 #include "core/layout/LayoutBlockFlow.h"
+#include "core/loader/FrameLoaderClient.h"
 #include "platform/text/PlatformLocale.h"
 #include "public/platform/Platform.h"
 #include "public/platform/WebLocalizedString.h"
@@ -61,6 +62,9 @@ PassRefPtrWillBeRawPtr<HTMLKeygenElement> HTMLKeygenElement::create(Document& do
 
 LayoutObject* HTMLKeygenElement::createLayoutObject(const ComputedStyle& style)
 {
+    if (!document().frame()->loader().client()->allowKeygen())
+        document().frame()->loader().client()->didNotAllowKeygen();

[jochen (gone - plz use gerrit), 2015/11/11 22:14:51]

ok, so this should not be a content setting like thing at all:

it's not a check whether we allow keygen, and the didNotAllowKeygen callback is
also lying about what it does.

instead, the HTMLKeygenElement ctor should just ping the embedder about the fact
that it's there.

[svaldez, 2015/11/23 20:13:50]

Done.

+
     // TODO(mstensho): While it's harmful and meaningless to allow most display types on replaced
     // content (e.g. table, table-row or flex), it would be useful to honor at least some of
     // them. Table-cell (and maybe table-caption too), for instance. See crbug.com/335040