ash: HiDPI user avatar for SessionController

ui/gfx/image/mojo/image.typemap

Index: ui/gfx/image/mojo/image.typemap
diff --git a/ui/gfx/image/mojo/image.typemap b/ui/gfx/image/mojo/image.typemap
new file mode 100644
index 0000000000000000000000000000000000000000..281f7b802db5e125d80d21de466a71175d638182
--- /dev/null
+++ b/ui/gfx/image/mojo/image.typemap
@@ -0,0 +1,19 @@
+# Copyright 2017 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+mojom = "//ui/gfx/image/mojo/image.mojom"
+public_headers = [
+  "//third_party/skia/include/core/SkBitmap.h",
+  "//ui/gfx/image/image_skia_rep.h",
+  "//ui/gfx/image/image_skia.h",
+]
+traits_headers = [ "//ui/gfx/image/mojo/image_skia_struct_traits.h" ]
+public_deps = [
+  "//ui/gfx/image/mojo:struct_traits",
+]
+type_mappings = [
+  "gfx.mojom.ImageSkia=gfx::ImageSkia[nullable_is_same_type]",
+  "gfx.mojom.ImageSkiaRep=gfx::ImageSkiaRep[nullable_is_same_type]",
+  "gfx.mojom.SharedBufferSkBitmap=SkBitmap[nullable_is_same_type]",

[msw, 2017/03/24 17:59:32]

Is there any worry in aliasing two mojom types to one c++ type?

[xiyuan, 2017/03/24 18:36:53]

It seems to me that the mojo side must be unique but C++ side does not have to
be. The mojo template is smart enough to figure that out because it knows which
mojo type will be used.

And yes, we should move this into //skia, not only it is cleaner but also saves
two memory copies. I will have a follow up CL, hopefully next week.

+]