encodedstream.h

Public Defines

RAPIDJSON_ENCODINGS_FUNC(x) <Ch>::x, <Ch>::x, <Ch>::x, <Ch>::x, <Ch>::x

Public Defines

RAPIDJSON_ENCODINGS_FUNC(x) <Ch>::x, <Ch>::x, <Ch>::x, <Ch>::x, <Ch>::x

  1
  2// Tencent is pleased to support the open source community by making RapidJSON available.
  3// 
  4// Copyright (C) 2015 THL A29 Limited, a Tencent company, and Milo Yip. All rights reserved.
  5//
  6// Licensed under the MIT License (the "License"); you may not use this file except
  7// in compliance with the License. You may obtain a copy of the License at
  8//
  9// http://opensource.org/licenses/MIT
 10//
 11// Unless required by applicable law or agreed to in writing, software distributed 
 12// under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR 
 13// CONDITIONS OF ANY KIND, either express or implied. See the License for the 
 14// specific language governing permissions and limitations under the License.
 15
 16#ifndef RAPIDJSON_ENCODEDSTREAM_H_
 17#define RAPIDJSON_ENCODEDSTREAM_H_
 18
 19#include "stream.h"
 20#include "memorystream.h"
 21
 22#ifdef __GNUC__
 23RAPIDJSON_DIAG_PUSH
 24RAPIDJSON_DIAG_OFF(effc++)
 25#endif
 26
 27#ifdef __clang__
 28RAPIDJSON_DIAG_PUSH
 29RAPIDJSON_DIAG_OFF(padded)
 30#endif
 31
 32RAPIDJSON_NAMESPACE_BEGIN
 33
 34//! Input byte stream wrapper with a statically bound encoding.
 35/*!
 36    \tparam Encoding The interpretation of encoding of the stream. Either UTF8, UTF16LE, UTF16BE, UTF32LE, UTF32BE.
 37    \tparam InputByteStream Type of input byte stream. For example, FileReadStream.
 38*/
 39template <typename Encoding, typename InputByteStream>
 40class EncodedInputStream {
 41    RAPIDJSON_STATIC_ASSERT(sizeof(typename InputByteStream::Ch) == 1);
 42public:
 43    typedef typename Encoding::Ch Ch;
 44
 45    EncodedInputStream(InputByteStream& is) : is_(is) { 
 46        current_ = Encoding::TakeBOM(is_);
 47    }
 48
 49    Ch Peek() const { return current_; }
 50    Ch Take() { Ch c = current_; current_ = Encoding::Take(is_); return c; }
 51    size_t Tell() const { return is_.Tell(); }
 52
 53    // Not implemented
 54    void Put(Ch) { RAPIDJSON_ASSERT(false); }
 55    void Flush() { RAPIDJSON_ASSERT(false); } 
 56    Ch* PutBegin() { RAPIDJSON_ASSERT(false); return 0; }
 57    size_t PutEnd(Ch*) { RAPIDJSON_ASSERT(false); return 0; }
 58
 59private:
 60    EncodedInputStream(const EncodedInputStream&);
 61    EncodedInputStream& operator=(const EncodedInputStream&);
 62
 63    InputByteStream& is_;
 64    Ch current_;
 65};
 66
 67//! Specialized for UTF8 MemoryStream.
 68template <>
 69class EncodedInputStream<UTF8<>, MemoryStream> {
 70public:
 71    typedef UTF8<>::Ch Ch;
 72
 73    EncodedInputStream(MemoryStream& is) : is_(is) {
 74        if (static_cast<unsigned char>(is_.Peek()) == 0xEFu) is_.Take();
 75        if (static_cast<unsigned char>(is_.Peek()) == 0xBBu) is_.Take();
 76        if (static_cast<unsigned char>(is_.Peek()) == 0xBFu) is_.Take();
 77    }
 78    Ch Peek() const { return is_.Peek(); }
 79    Ch Take() { return is_.Take(); }
 80    size_t Tell() const { return is_.Tell(); }
 81
 82    // Not implemented
 83    void Put(Ch) {}
 84    void Flush() {} 
 85    Ch* PutBegin() { return 0; }
 86    size_t PutEnd(Ch*) { return 0; }
 87
 88    MemoryStream& is_;
 89
 90private:
 91    EncodedInputStream(const EncodedInputStream&);
 92    EncodedInputStream& operator=(const EncodedInputStream&);
 93};
 94
 95//! Output byte stream wrapper with statically bound encoding.
 96/*!
 97    \tparam Encoding The interpretation of encoding of the stream. Either UTF8, UTF16LE, UTF16BE, UTF32LE, UTF32BE.
 98    \tparam OutputByteStream Type of input byte stream. For example, FileWriteStream.
 99*/
100template <typename Encoding, typename OutputByteStream>
101class EncodedOutputStream {
102    RAPIDJSON_STATIC_ASSERT(sizeof(typename OutputByteStream::Ch) == 1);
103public:
104    typedef typename Encoding::Ch Ch;
105
106    EncodedOutputStream(OutputByteStream& os, bool putBOM = true) : os_(os) { 
107        if (putBOM)
108            Encoding::PutBOM(os_);
109    }
110
111    void Put(Ch c) { Encoding::Put(os_, c);  }
112    void Flush() { os_.Flush(); }
113
114    // Not implemented
115    Ch Peek() const { RAPIDJSON_ASSERT(false); return 0;}
116    Ch Take() { RAPIDJSON_ASSERT(false); return 0;}
117    size_t Tell() const { RAPIDJSON_ASSERT(false);  return 0; }
118    Ch* PutBegin() { RAPIDJSON_ASSERT(false); return 0; }
119    size_t PutEnd(Ch*) { RAPIDJSON_ASSERT(false); return 0; }
120
121private:
122    EncodedOutputStream(const EncodedOutputStream&);
123    EncodedOutputStream& operator=(const EncodedOutputStream&);
124
125    OutputByteStream& os_;
126};
127
128#define RAPIDJSON_ENCODINGS_FUNC(x) UTF8<Ch>::x, UTF16LE<Ch>::x, UTF16BE<Ch>::x, UTF32LE<Ch>::x, UTF32BE<Ch>::x
129
130//! Input stream wrapper with dynamically bound encoding and automatic encoding detection.
131/*!
132    \tparam CharType Type of character for reading.
133    \tparam InputByteStream type of input byte stream to be wrapped.
134*/
135template <typename CharType, typename InputByteStream>
136class AutoUTFInputStream {
137    RAPIDJSON_STATIC_ASSERT(sizeof(typename InputByteStream::Ch) == 1);
138public:
139    typedef CharType Ch;
140
141    //! Constructor.
142    /*!
143        \param is input stream to be wrapped.
144        \param type UTF encoding type if it is not detected from the stream.
145    */
146    AutoUTFInputStream(InputByteStream& is, UTFType type = kUTF8) : is_(&is), type_(type), hasBOM_(false) {
147        RAPIDJSON_ASSERT(type >= kUTF8 && type <= kUTF32BE);        
148        DetectType();
149        static const TakeFunc f[] = { RAPIDJSON_ENCODINGS_FUNC(Take) };
150        takeFunc_ = f[type_];
151        current_ = takeFunc_(*is_);
152    }
153
154    UTFType GetType() const { return type_; }
155    bool HasBOM() const { return hasBOM_; }
156
157    Ch Peek() const { return current_; }
158    Ch Take() { Ch c = current_; current_ = takeFunc_(*is_); return c; }
159    size_t Tell() const { return is_->Tell(); }
160
161    // Not implemented
162    void Put(Ch) { RAPIDJSON_ASSERT(false); }
163    void Flush() { RAPIDJSON_ASSERT(false); } 
164    Ch* PutBegin() { RAPIDJSON_ASSERT(false); return 0; }
165    size_t PutEnd(Ch*) { RAPIDJSON_ASSERT(false); return 0; }
166
167private:
168    AutoUTFInputStream(const AutoUTFInputStream&);
169    AutoUTFInputStream& operator=(const AutoUTFInputStream&);
170
171    // Detect encoding type with BOM or RFC 4627
172    void DetectType() {
173        // BOM (Byte Order Mark):
174        // 00 00 FE FF  UTF-32BE
175        // FF FE 00 00  UTF-32LE
176        // FE FF        UTF-16BE
177        // FF FE        UTF-16LE
178        // EF BB BF     UTF-8
179
180        const unsigned char* c = reinterpret_cast<const unsigned char *>(is_->Peek4());
181        if (!c)
182            return;
183
184        unsigned bom = static_cast<unsigned>(c[0] | (c[1] << 8) | (c[2] << 16) | (c[3] << 24));
185        hasBOM_ = false;
186        if (bom == 0xFFFE0000)                  { type_ = kUTF32BE; hasBOM_ = true; is_->Take(); is_->Take(); is_->Take(); is_->Take(); }
187        else if (bom == 0x0000FEFF)             { type_ = kUTF32LE; hasBOM_ = true; is_->Take(); is_->Take(); is_->Take(); is_->Take(); }
188        else if ((bom & 0xFFFF) == 0xFFFE)      { type_ = kUTF16BE; hasBOM_ = true; is_->Take(); is_->Take();                           }
189        else if ((bom & 0xFFFF) == 0xFEFF)      { type_ = kUTF16LE; hasBOM_ = true; is_->Take(); is_->Take();                           }
190        else if ((bom & 0xFFFFFF) == 0xBFBBEF)  { type_ = kUTF8;    hasBOM_ = true; is_->Take(); is_->Take(); is_->Take();              }
191
192        // RFC 4627: Section 3
193        // "Since the first two characters of a JSON text will always be ASCII
194        // characters [RFC0020], it is possible to determine whether an octet
195        // stream is UTF-8, UTF-16 (BE or LE), or UTF-32 (BE or LE) by looking
196        // at the pattern of nulls in the first four octets."
197        // 00 00 00 xx  UTF-32BE
198        // 00 xx 00 xx  UTF-16BE
199        // xx 00 00 00  UTF-32LE
200        // xx 00 xx 00  UTF-16LE
201        // xx xx xx xx  UTF-8
202
203        if (!hasBOM_) {
204            int pattern = (c[0] ? 1 : 0) | (c[1] ? 2 : 0) | (c[2] ? 4 : 0) | (c[3] ? 8 : 0);
205            switch (pattern) {
206            case 0x08: type_ = kUTF32BE; break;
207            case 0x0A: type_ = kUTF16BE; break;
208            case 0x01: type_ = kUTF32LE; break;
209            case 0x05: type_ = kUTF16LE; break;
210            case 0x0F: type_ = kUTF8;    break;
211            default: break; // Use type defined by user.
212            }
213        }
214
215        // Runtime check whether the size of character type is sufficient. It only perform checks with assertion.
216        if (type_ == kUTF16LE || type_ == kUTF16BE) RAPIDJSON_ASSERT(sizeof(Ch) >= 2);
217        if (type_ == kUTF32LE || type_ == kUTF32BE) RAPIDJSON_ASSERT(sizeof(Ch) >= 4);
218    }
219
220    typedef Ch (*TakeFunc)(InputByteStream& is);
221    InputByteStream* is_;
222    UTFType type_;
223    Ch current_;
224    TakeFunc takeFunc_;
225    bool hasBOM_;
226};
227
228//! Output stream wrapper with dynamically bound encoding and automatic encoding detection.
229/*!
230    \tparam CharType Type of character for writing.
231    \tparam OutputByteStream type of output byte stream to be wrapped.
232*/
233template <typename CharType, typename OutputByteStream>
234class AutoUTFOutputStream {
235    RAPIDJSON_STATIC_ASSERT(sizeof(typename OutputByteStream::Ch) == 1);
236public:
237    typedef CharType Ch;
238
239    //! Constructor.
240    /*!
241        \param os output stream to be wrapped.
242        \param type UTF encoding type.
243        \param putBOM Whether to write BOM at the beginning of the stream.
244    */
245    AutoUTFOutputStream(OutputByteStream& os, UTFType type, bool putBOM) : os_(&os), type_(type) {
246        RAPIDJSON_ASSERT(type >= kUTF8 && type <= kUTF32BE);
247
248        // Runtime check whether the size of character type is sufficient. It only perform checks with assertion.
249        if (type_ == kUTF16LE || type_ == kUTF16BE) RAPIDJSON_ASSERT(sizeof(Ch) >= 2);
250        if (type_ == kUTF32LE || type_ == kUTF32BE) RAPIDJSON_ASSERT(sizeof(Ch) >= 4);
251
252        static const PutFunc f[] = { RAPIDJSON_ENCODINGS_FUNC(Put) };
253        putFunc_ = f[type_];
254
255        if (putBOM)
256            PutBOM();
257    }
258
259    UTFType GetType() const { return type_; }
260
261    void Put(Ch c) { putFunc_(*os_, c); }
262    void Flush() { os_->Flush(); } 
263
264    // Not implemented
265    Ch Peek() const { RAPIDJSON_ASSERT(false); return 0;}
266    Ch Take() { RAPIDJSON_ASSERT(false); return 0;}
267    size_t Tell() const { RAPIDJSON_ASSERT(false); return 0; }
268    Ch* PutBegin() { RAPIDJSON_ASSERT(false); return 0; }
269    size_t PutEnd(Ch*) { RAPIDJSON_ASSERT(false); return 0; }
270
271private:
272    AutoUTFOutputStream(const AutoUTFOutputStream&);
273    AutoUTFOutputStream& operator=(const AutoUTFOutputStream&);
274
275    void PutBOM() { 
276        typedef void (*PutBOMFunc)(OutputByteStream&);
277        static const PutBOMFunc f[] = { RAPIDJSON_ENCODINGS_FUNC(PutBOM) };
278        f[type_](*os_);
279    }
280
281    typedef void (*PutFunc)(OutputByteStream&, Ch);
282
283    OutputByteStream* os_;
284    UTFType type_;
285    PutFunc putFunc_;
286};
287
288#undef RAPIDJSON_ENCODINGS_FUNC
289
290RAPIDJSON_NAMESPACE_END
291
292#ifdef __clang__
293RAPIDJSON_DIAG_POP
294#endif
295
296#ifdef __GNUC__
297RAPIDJSON_DIAG_POP
298#endif
299
300#endif // RAPIDJSON_FILESTREAM_H_
301