Implement MediaMetadata artwork in content

content/renderer/manifest/manifest_parser.h

Index: content/renderer/manifest/manifest_parser.h
diff --git a/content/renderer/manifest/manifest_parser.h b/content/renderer/manifest/manifest_parser.h
index 0f12c5a474566306c2947aefb33f3aa18b7dc947..ad9e65634a349103e33149627f02843389ab2072 100644
--- a/content/renderer/manifest/manifest_parser.h
+++ b/content/renderer/manifest/manifest_parser.h
@@ -24,6 +24,8 @@ class DictionaryValue;
 
 namespace content {
 
+std::vector<gfx::Size> ParseIconSizesHTML(const base::string16& sizes_str16);

[mlamouri (slow - plz ping), 2016/06/10 10:54:58]

Not a big fan of re-using the parsing algorithm of the Manifest but let see how
the other CL goes. Worse case, we can land with a TODO and switch to the Blink
API when available.

[Zhiqiang Zhang (Slow), 2016/06/14 13:16:17]

Rebased onto the common parser in Blink.

+
 // ManifestParser handles the logic of parsing the Web Manifest from a string.
 // It implements:
 // http://w3c.github.io/manifest/#dfn-steps-for-processing-a-manifest