Index: third_party/WebKit/Source/core/loader/TextResourceDecoderBuilder.cpp |
diff --git a/third_party/WebKit/Source/core/loader/TextResourceDecoderBuilder.cpp b/third_party/WebKit/Source/core/loader/TextResourceDecoderBuilder.cpp |
index 077ad6d0f5013afd3cd332a3a30ec748e3eb6825..63f29db7f71adec747f84e721a36989574117e13 100644 |
--- a/third_party/WebKit/Source/core/loader/TextResourceDecoderBuilder.cpp |
+++ b/third_party/WebKit/Source/core/loader/TextResourceDecoderBuilder.cpp |
@@ -34,6 +34,7 @@ |
#include "core/frame/LocalFrame.h" |
#include "core/frame/Settings.h" |
#include "platform/weborigin/SecurityOrigin.h" |
+#include "wtf/HashMap.h" |
namespace blink { |
@@ -42,6 +43,99 @@ static inline bool canReferToParentFrameEncoding(const LocalFrame* frame, const |
return parentFrame && parentFrame->document()->securityOrigin()->canAccess(frame->document()->securityOrigin()); |
} |
+static WTF::HashMap<String, WTF::TextEncoding>* getEncodingMap() |
esprehn
2016/02/26 09:43:13
if you use a Map you should return by const ref he
Jinsuk Kim
2016/03/02 03:46:08
I chose to use a simple array. Please see if it lo
|
+{ |
+ static WTF::HashMap<String, WTF::TextEncoding>* map {}; |
+ if (!map) { |
+ map = new WTF::HashMap<String, WTF::TextEncoding>(); |
aelias_OOO_until_Jul13
2016/02/26 09:07:24
Hmm, in Chromium we would use something like LAZY_
esprehn
2016/02/26 09:43:13
This should use DEFINE_STATIC_LOCAL and map.isEmpt
Jinsuk Kim
2016/03/02 03:46:08
Please see the reply above.
Jinsuk Kim
2016/03/02 03:46:08
I agree. It got complicated that I had wished. Rew
|
+ const WTF::TextEncoding windows1250("windows-1250"); |
+ const WTF::TextEncoding windows1251("windows-1251"); |
+ const WTF::TextEncoding windows1252("windows-1252"); |
+ const WTF::TextEncoding windows1255("windows-1255"); |
+ const WTF::TextEncoding windows1256("windows-1256"); |
+ const WTF::TextEncoding windows1257("windows-1257"); |
+ const WTF::TextEncoding windows1258("windows-1258"); |
+ const WTF::TextEncoding windows874("windows-874"); |
+ const WTF::TextEncoding windows949("windows-949"); |
+ const WTF::TextEncoding iso88592("ISO-8859-2"); |
+ const WTF::TextEncoding iso88597("ISO-8859-7"); |
+ const WTF::TextEncoding iso88599("ISO-8859-9"); |
+ const WTF::TextEncoding big5("Big5"); |
+ const WTF::TextEncoding gbk("GBK"); |
+ const WTF::TextEncoding shiftJis("Shift_JIS"); |
+ |
+ struct { |
+ const char* domain; |
+ const WTF::TextEncoding& encoding; |
+ } encodingMap[] = { |
+ { "au", windows1252 }, |
+ { "az", iso88599 }, |
+ { "bd", windows1252 }, |
+ { "bg", windows1251 }, |
+ { "br", windows1252 }, |
+ { "ca", windows1252 }, |
+ { "ch", windows1252 }, |
+ { "cn", gbk }, |
+ { "cz", windows1250 }, |
+ { "de", windows1252 }, |
+ { "dk", windows1252 }, |
+ { "ee", windows1257 }, |
+ { "eg", windows1256 }, |
+ { "et", windows1252 }, |
+ { "fi", windows1252 }, |
+ { "fr", windows1252 }, |
+ { "gb", windows1252 }, |
+ { "gr", iso88597 }, |
+ { "hk", big5 }, |
+ { "hr", windows1250 }, |
+ { "hu", iso88592 }, |
+ { "il", windows1255 }, |
+ { "ir", windows1256 }, |
+ { "is", windows1252 }, |
+ { "it", windows1252 }, |
+ { "jp", shiftJis }, |
+ { "kr", windows949 }, |
+ { "lt", windows1257 }, |
+ { "lv", windows1257 }, |
+ { "mk", windows1251 }, |
+ { "nl", windows1252 }, |
+ { "no", windows1252 }, |
+ { "pl", iso88592 }, |
+ { "pt", windows1252 }, |
+ { "ro", iso88592 }, |
+ { "rs", windows1251 }, |
+ { "ru", windows1251 }, |
+ { "se", windows1252 }, |
+ { "si", iso88592 }, |
+ { "sk", windows1250 }, |
+ { "th", windows874 }, |
+ { "tr", iso88599 }, |
+ { "tw", big5 }, |
+ { "tz", windows1252 }, |
+ { "ua", windows1251 }, |
+ { "us", windows1252 }, |
+ { "vn", windows1258 }, |
+ { "xa", windows1252 }, |
+ { "xb", windows1256 } |
+ }; |
+ for (size_t i = 0; i < WTF_ARRAY_LENGTH(encodingMap); ++i) |
+ map->add(String(encodingMap[i].domain), encodingMap[i].encoding); |
esprehn
2016/02/26 09:43:13
lets not do this, make this method into something
Jinsuk Kim
2016/03/02 03:46:08
Rewrote it with an array.
|
+ } |
+ return map; |
+} |
+ |
+static const WTF::TextEncoding getEncodingFromDomain(const KURL& url) |
+{ |
+ Vector<String> tokens; |
+ url.host().split(String(".", 1), tokens); |
esprehn
2016/02/26 09:43:13
you don't need to do String(".", 1), just pass "."
Jinsuk Kim
2016/03/02 03:46:08
Better. Thanks. Done.
|
+ if (tokens.size() > 0) { |
+ String tld = tokens[tokens.size()-1]; |
+ auto map = getEncodingMap(); |
+ if (map->find(tld) != map->end()) |
+ return map->get(tld); |
esprehn
2016/02/26 09:43:13
lets use std::lower_bound on a static list instead
Jinsuk Kim
2016/03/02 03:46:08
Rewrote it with an array.
|
+ } |
+ return WTF::TextEncoding(); |
+} |
TextResourceDecoderBuilder::TextResourceDecoderBuilder(const AtomicString& mimeType, const AtomicString& encoding) |
: m_mimeType(mimeType) |
@@ -56,12 +150,13 @@ TextResourceDecoderBuilder::~TextResourceDecoderBuilder() |
inline PassOwnPtr<TextResourceDecoder> TextResourceDecoderBuilder::createDecoderInstance(Document* document) |
{ |
+ const WTF::TextEncoding encodingFromDomain = getEncodingFromDomain(document->url()); |
if (LocalFrame* frame = document->frame()) { |
if (Settings* settings = frame->settings()) |
- return TextResourceDecoder::create(m_mimeType, settings->defaultTextEncodingName(), settings->usesEncodingDetector()); |
+ return TextResourceDecoder::create(m_mimeType, encodingFromDomain.isValid() ? encodingFromDomain : settings->defaultTextEncodingName(), settings->usesEncodingDetector()); |
aelias_OOO_until_Jul13
2016/02/26 09:07:24
Looks OK, can you add a test verifying (for one or
Jinsuk Kim
2016/03/02 03:46:08
Added tests.
|
} |
- return TextResourceDecoder::create(m_mimeType, String()); |
+ return TextResourceDecoder::create(m_mimeType, encodingFromDomain); |
} |
inline void TextResourceDecoderBuilder::setupEncoding(TextResourceDecoder* decoder, Document* document) |