Add support for CSS unicod encoding to the text codec.

third_party/WebKit/Source/wtf/text/TextCodec.cpp

Index: third_party/WebKit/Source/wtf/text/TextCodec.cpp
diff --git a/third_party/WebKit/Source/wtf/text/TextCodec.cpp b/third_party/WebKit/Source/wtf/text/TextCodec.cpp
index 191ebb5689055add7eb204a4518e9cc21fad1ce9..50b334f2c2cea88a4c345e06d0d6dcaf06fc5a06 100644
--- a/third_party/WebKit/Source/wtf/text/TextCodec.cpp
+++ b/third_party/WebKit/Source/wtf/text/TextCodec.cpp
@@ -47,6 +47,10 @@ int TextCodec::getUnencodableReplacement(unsigned codePoint, UnencodableHandling
     case URLEncodedEntitiesForUnencodables:
         snprintf(replacement, sizeof(UnencodableReplacementArray), "%%26%%23%u%%3B", codePoint);
         return static_cast<int>(strlen(replacement));
+
+    case CSSEncodedEntitiesForUnencodables:
+        snprintf(replacement, sizeof(UnencodableReplacementArray), "\\%x", codePoint);

[Łukasz Anforowicz, 2016/04/27 04:28:12]

I don't understand how we handle "A 'real' space after the escape sequence must
be doubled." mentioned by the spec you linked to (thanks for the link) at
https://www.w3.org/TR/css-syntax-3/#escaping

Would it be reasonable to always include a space here (i.e. use "\\%x " format
string)?

Also - it might be also worthwhile to tweak 1.css to test this space-handling
scenario (i.e. content: "\e003 \e003  \e003x"; ?) and verify via
SavePageOriginalVsSavedComparisonTest (as I suggest in another comment).

[bburns, 2016/04/27 20:24:21]

Added the space as suggested and augmented the test as well.

+        return static_cast<int>(strlen(replacement));
     }
     ASSERT_NOT_REACHED();
     replacement[0] = 0;