UTF-8 detector for pages missing encoding info

third_party/WebKit/Source/core/html/parser/TextResourceDecoder.cpp

Index: third_party/WebKit/Source/core/html/parser/TextResourceDecoder.cpp
diff --git a/third_party/WebKit/Source/core/html/parser/TextResourceDecoder.cpp b/third_party/WebKit/Source/core/html/parser/TextResourceDecoder.cpp
index 49897d7c3ac409bd860dbaf44bbc4786a39ad3b4..5e27ba9ba25127e309e0083ef6604eaf0ac442d4 100644
--- a/third_party/WebKit/Source/core/html/parser/TextResourceDecoder.cpp
+++ b/third_party/WebKit/Source/core/html/parser/TextResourceDecoder.cpp
@@ -28,6 +28,7 @@
 #include "wtf/StringExtras.h"
 #include "wtf/text/TextCodec.h"
 #include "wtf/text/TextEncodingRegistry.h"
+#include "wtf/text/UTF8.h"
 
 using namespace WTF;
 
@@ -402,11 +403,7 @@ String TextResourceDecoder::decode(const char* data, size_t len)
     if (m_contentType == HTMLContent && !m_checkedForMetaCharset)
         checkForMetaCharset(dataForDecode, lengthForDecode);
 
-    if (shouldAutoDetect()) {
-        WTF::TextEncoding detectedEncoding;
-        if (detectTextEncoding(data, len, m_hintEncoding, &detectedEncoding))
-            setEncoding(detectedEncoding, EncodingFromContentSniffing);
-    }
+    detectTextEncoding(data, len);
 
     ASSERT(m_encoding.isValid());
 
@@ -419,16 +416,28 @@ String TextResourceDecoder::decode(const char* data, size_t len)
     return result;
 }
 
+void TextResourceDecoder::detectTextEncoding(const char* data, size_t len)
+{
+    if (shouldAutoDetect()) {
+        WTF::TextEncoding detectedEncoding;
+        if (detectTextEncodingUniversal(data, len, m_hintEncoding, &detectedEncoding))
+            setEncoding(detectedEncoding, EncodingFromContentSniffing);
+        return;
+    }
+    if ((m_source == DefaultEncoding || (m_source == EncodingFromParentFrame && m_hintEncoding))) {

[jungshik at Google, 2016/03/24 06:15:08]

nit:   The above condition is shared by shouldAutoDetect() except that
shouldAutoDetect has an additional check to see if 'autodetector' is ON.  How
about refactoring the above condition to a helper function
|shouldDetectEncoding|?  Then, this function can be something like:

if (shouldDetectEncoding()) {
    if (  .... isUTF8Encoded( ...) )  
       setEncoding ....; 
       return;
    if (isUniveralDetectorOn()) 
         .... call Universal encoding detector 

.....

[Jinsuk Kim, 2016/03/25 02:15:42]

Done.

+        if (WTF::Unicode::isUTF8Encoded(data, len))
+            setEncoding(UTF8Encoding(), EncodingFromContentSniffing);

[jungshik at Google, 2016/03/24 06:15:08]

Given that isUTF8Encoded excludes 'ASCII' (by checking the output of U8_NEXT is
larger than 0x7F), you may as well consider being more aggressive and putting
this *before* the Universal detector. 

[Jinsuk Kim, 2016/03/25 02:15:42]

Makes sense. Done.

+    }
+}
+
 String TextResourceDecoder::flush()
 {
     // If we can not identify the encoding even after a document is completely
     // loaded, we need to detect the encoding if other conditions for
     // autodetection is satisfied.
-    if (m_buffer.size() && shouldAutoDetect()
+    if (m_buffer.size()
         && ((!m_checkedForXMLCharset && (m_contentType == HTMLContent || m_contentType == XMLContent)) || (!m_checkedForCSSCharset && (m_contentType == CSSContent)))) {
-        WTF::TextEncoding detectedEncoding;
-        if (detectTextEncoding(m_buffer.data(), m_buffer.size(), m_hintEncoding, &detectedEncoding))
-            setEncoding(detectedEncoding, EncodingFromContentSniffing);
+        detectTextEncoding(m_buffer.data(), m_buffer.size());
     }
 
     if (!m_codec)