OLD | NEW |
1 /* | 1 /* |
2 Copyright (C) 1999 Lars Knoll (knoll@mpi-hd.mpg.de) | 2 Copyright (C) 1999 Lars Knoll (knoll@mpi-hd.mpg.de) |
3 Copyright (C) 2006 Alexey Proskuryakov (ap@nypop.com) | 3 Copyright (C) 2006 Alexey Proskuryakov (ap@nypop.com) |
4 Copyright (C) 2006, 2008 Apple Inc. All rights reserved. | 4 Copyright (C) 2006, 2008 Apple Inc. All rights reserved. |
5 | 5 |
6 This library is free software; you can redistribute it and/or | 6 This library is free software; you can redistribute it and/or |
7 modify it under the terms of the GNU Library General Public | 7 modify it under the terms of the GNU Library General Public |
8 License as published by the Free Software Foundation; either | 8 License as published by the Free Software Foundation; either |
9 version 2 of the License, or (at your option) any later version. | 9 version 2 of the License, or (at your option) any later version. |
10 | 10 |
(...skipping 61 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
72 | 72 |
73 void setHintEncoding(const WTF::TextEncoding& encoding) | 73 void setHintEncoding(const WTF::TextEncoding& encoding) |
74 { | 74 { |
75 m_hintEncoding = encoding.name(); | 75 m_hintEncoding = encoding.name(); |
76 } | 76 } |
77 | 77 |
78 void useLenientXMLDecoding() { m_useLenientXMLDecoding = true; } | 78 void useLenientXMLDecoding() { m_useLenientXMLDecoding = true; } |
79 bool sawError() const { return m_sawError; } | 79 bool sawError() const { return m_sawError; } |
80 size_t checkForBOM(const char*, size_t); | 80 size_t checkForBOM(const char*, size_t); |
81 | 81 |
82 private: | 82 protected: |
83 | |
84 // TextResourceDecoder does three kind of encoding detection: | 83 // TextResourceDecoder does three kind of encoding detection: |
85 // 1. By BOM, | 84 // 1. By BOM, |
86 // 2. By Content if |m_contentType| is not |PlainTextContext| | 85 // 2. By Content if |m_contentType| is not |PlainTextContext| |
87 // (e.g. <meta> tag for HTML), and | 86 // (e.g. <meta> tag for HTML), and |
88 // 3. By detectTextEncoding(). | 87 // 3. By detectTextEncoding(). |
89 enum EncodingDetectionOption { | 88 enum EncodingDetectionOption { |
90 // Use 1. + 2. + 3. | 89 // Use 1. + 2. + 3. |
91 UseAllAutoDetection, | 90 UseAllAutoDetection, |
92 | 91 |
93 // Use 1. + 2. | 92 // Use 1. + 2. |
94 UseContentAndBOMBasedDetection, | 93 UseContentAndBOMBasedDetection, |
95 | 94 |
96 // Use None of them. | 95 // Use None of them. |
97 // |m_contentType| must be |PlainTextContent| and | 96 // |m_contentType| must be |PlainTextContent| and |
98 // |m_encoding| must be UTF8Encoding. | 97 // |m_encoding| must be UTF8Encoding. |
99 // This doesn't change encoding based on BOMs, but still processes | 98 // This doesn't change encoding based on BOMs, but still processes |
100 // utf-8 BOMs so that utf-8 BOMs don't appear in the decoded result. | 99 // utf-8 BOMs so that utf-8 BOMs don't appear in the decoded result. |
101 AlwaysUseUTF8ForText | 100 AlwaysUseUTF8ForText |
102 }; | 101 }; |
103 | 102 |
104 TextResourceDecoder(const String& mimeType, const WTF::TextEncoding& default
Encoding, EncodingDetectionOption); | 103 TextResourceDecoder(const String& mimeType, const WTF::TextEncoding& default
Encoding, EncodingDetectionOption); |
105 | 104 |
| 105 private: |
106 enum ContentType { PlainTextContent, HTMLContent, XMLContent, CSSContent };
// PlainText only checks for BOM. | 106 enum ContentType { PlainTextContent, HTMLContent, XMLContent, CSSContent };
// PlainText only checks for BOM. |
107 static ContentType determineContentType(const String& mimeType); | 107 static ContentType determineContentType(const String& mimeType); |
108 static const WTF::TextEncoding& defaultEncoding(ContentType, const WTF::Text
Encoding& defaultEncoding); | 108 static const WTF::TextEncoding& defaultEncoding(ContentType, const WTF::Text
Encoding& defaultEncoding); |
109 | 109 |
110 bool checkForCSSCharset(const char*, size_t, bool& movedDataToBuffer); | 110 bool checkForCSSCharset(const char*, size_t, bool& movedDataToBuffer); |
111 bool checkForXMLCharset(const char*, size_t, bool& movedDataToBuffer); | 111 bool checkForXMLCharset(const char*, size_t, bool& movedDataToBuffer); |
112 void checkForMetaCharset(const char*, size_t); | 112 void checkForMetaCharset(const char*, size_t); |
113 bool shouldAutoDetect() const; | 113 bool shouldAutoDetect() const; |
114 | 114 |
115 ContentType m_contentType; | 115 ContentType m_contentType; |
116 WTF::TextEncoding m_encoding; | 116 WTF::TextEncoding m_encoding; |
117 std::unique_ptr<TextCodec> m_codec; | 117 std::unique_ptr<TextCodec> m_codec; |
118 EncodingSource m_source; | 118 EncodingSource m_source; |
119 const char* m_hintEncoding; | 119 const char* m_hintEncoding; |
120 Vector<char> m_buffer; | 120 Vector<char> m_buffer; |
121 bool m_checkedForBOM; | 121 bool m_checkedForBOM; |
122 bool m_checkedForCSSCharset; | 122 bool m_checkedForCSSCharset; |
123 bool m_checkedForXMLCharset; | 123 bool m_checkedForXMLCharset; |
124 bool m_checkedForMetaCharset; | 124 bool m_checkedForMetaCharset; |
125 bool m_useLenientXMLDecoding; // Don't stop on XML decoding errors. | 125 bool m_useLenientXMLDecoding; // Don't stop on XML decoding errors. |
126 bool m_sawError; | 126 bool m_sawError; |
127 EncodingDetectionOption m_encodingDetectionOption; | 127 EncodingDetectionOption m_encodingDetectionOption; |
128 | 128 |
129 std::unique_ptr<HTMLMetaCharsetParser> m_charsetParser; | 129 std::unique_ptr<HTMLMetaCharsetParser> m_charsetParser; |
130 }; | 130 }; |
131 | 131 |
132 } // namespace blink | 132 } // namespace blink |
133 | 133 |
134 #endif | 134 #endif |
OLD | NEW |