1 // Protocol Buffers - Google's data interchange format
2 // Copyright 2008 Google Inc. All rights reserved.
3 // https://developers.google.com/protocol-buffers/
4 //
5 // Redistribution and use in source and binary forms, with or without
6 // modification, are permitted provided that the following conditions are
7 // met:
8 //
9 // * Redistributions of source code must retain the above copyright
10 // notice, this list of conditions and the following disclaimer.
11 // * Redistributions in binary form must reproduce the above
12 // copyright notice, this list of conditions and the following disclaimer
13 // in the documentation and/or other materials provided with the
14 // distribution.
15 // * Neither the name of Google Inc. nor the names of its
16 // contributors may be used to endorse or promote products derived from
17 // this software without specific prior written permission.
18 //
19 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30
31 // Author: kenton@google.com (Kenton Varda)
32 // atenasio@google.com (Chris Atenasio) (ZigZag transform)
33 // Based on original Protocol Buffers design by
34 // Sanjay Ghemawat, Jeff Dean, and others.
35 //
36 // This header is logically internal, but is made public because it is used
37 // from protocol-compiler-generated code, which may reside in other components.
38
39 #ifndef GOOGLE_PROTOBUF_WIRE_FORMAT_H__
40 #define GOOGLE_PROTOBUF_WIRE_FORMAT_H__
41
42 #include <string>
43 #include <google/protobuf/stubs/common.h>
44 #include <google/protobuf/descriptor.pb.h>
45 #include <google/protobuf/descriptor.h>
46 #include <google/protobuf/message.h>
47 #include <google/protobuf/wire_format_lite.h>
48
49 // Do UTF-8 validation on string type in Debug build only
50 #ifndef NDEBUG
51 #define GOOGLE_PROTOBUF_UTF8_VALIDATION_ENABLED
52 #endif
53
54 namespace google {
55 namespace protobuf {
56 namespace io {
57 class CodedInputStream; // coded_stream.h
58 class CodedOutputStream; // coded_stream.h
59 }
60 class UnknownFieldSet; // unknown_field_set.h
61 }
62
63 namespace protobuf {
64 namespace internal {
65
66 // This class is for internal use by the protocol buffer library and by
67 // protocol-complier-generated message classes. It must not be called
68 // directly by clients.
69 //
70 // This class contains code for implementing the binary protocol buffer
71 // wire format via reflection. The WireFormatLite class implements the
72 // non-reflection based routines.
73 //
74 // This class is really a namespace that contains only static methods
75 class LIBPROTOBUF_EXPORT WireFormat {
76 public:
77
78 // Given a field return its WireType
79 static inline WireFormatLite::WireType WireTypeForField(
80 const FieldDescriptor* field);
81
82 // Given a FieldDescriptor::Type return its WireType
83 static inline WireFormatLite::WireType WireTypeForFieldType(
84 FieldDescriptor::Type type);
85
86 // Compute the byte size of a tag. For groups, this includes both the start
87 // and end tags.
88 static inline int TagSize(int field_number, FieldDescriptor::Type type);
89
90 // These procedures can be used to implement the methods of Message which
91 // handle parsing and serialization of the protocol buffer wire format
92 // using only the Reflection interface. When you ask the protocol
93 // compiler to optimize for code size rather than speed, it will implement
94 // those methods in terms of these procedures. Of course, these are much
95 // slower than the specialized implementations which the protocol compiler
96 // generates when told to optimize for speed.
97
98 // Read a message in protocol buffer wire format.
99 //
100 // This procedure reads either to the end of the input stream or through
101 // a WIRETYPE_END_GROUP tag ending the message, whichever comes first.
102 // It returns false if the input is invalid.
103 //
104 // Required fields are NOT checked by this method. You must call
105 // IsInitialized() on the resulting message yourself.
106 static bool ParseAndMergePartial(io::CodedInputStream* input,
107 Message* message);
108
109 // Serialize a message in protocol buffer wire format.
110 //
111 // Any embedded messages within the message must have their correct sizes
112 // cached. However, the top-level message need not; its size is passed as
113 // a parameter to this procedure.
114 //
115 // These return false iff the underlying stream returns a write error.
116 static void SerializeWithCachedSizes(
117 const Message& message,
118 int size, io::CodedOutputStream* output);
119
120 // Implements Message::ByteSize() via reflection. WARNING: The result
121 // of this method is *not* cached anywhere. However, all embedded messages
122 // will have their ByteSize() methods called, so their sizes will be cached.
123 // Therefore, calling this method is sufficient to allow you to call
124 // WireFormat::SerializeWithCachedSizes() on the same object.
125 static int ByteSize(const Message& message);
126
127 // -----------------------------------------------------------------
128 // Helpers for dealing with unknown fields
129
130 // Skips a field value of the given WireType. The input should start
131 // positioned immediately after the tag. If unknown_fields is non-NULL,
132 // the contents of the field will be added to it.
133 static bool SkipField(io::CodedInputStream* input, uint32 tag,
134 UnknownFieldSet* unknown_fields);
135
136 // Reads and ignores a message from the input. If unknown_fields is non-NULL,
137 // the contents will be added to it.
138 static bool SkipMessage(io::CodedInputStream* input,
139 UnknownFieldSet* unknown_fields);
140
141 // Read a packed enum field. If the is_valid function is not NULL, values for
142 // which is_valid(value) returns false are appended to unknown_fields_stream.
143 static bool ReadPackedEnumPreserveUnknowns(io::CodedInputStream* input,
144 uint32 field_number,
145 bool (*is_valid)(int),
146 UnknownFieldSet* unknown_fields,
147 RepeatedField<int>* values);
148
149 // Write the contents of an UnknownFieldSet to the output.
150 static void SerializeUnknownFields(const UnknownFieldSet& unknown_fields,
151 io::CodedOutputStream* output);
152 // Same as above, except writing directly to the provided buffer.
153 // Requires that the buffer have sufficient capacity for
154 // ComputeUnknownFieldsSize(unknown_fields).
155 //
156 // Returns a pointer past the last written byte.
157 static uint8* SerializeUnknownFieldsToArray(
158 const UnknownFieldSet& unknown_fields,
159 uint8* target);
160
161 // Same thing except for messages that have the message_set_wire_format
162 // option.
163 static void SerializeUnknownMessageSetItems(
164 const UnknownFieldSet& unknown_fields,
165 io::CodedOutputStream* output);
166 // Same as above, except writing directly to the provided buffer.
167 // Requires that the buffer have sufficient capacity for
168 // ComputeUnknownMessageSetItemsSize(unknown_fields).
169 //
170 // Returns a pointer past the last written byte.
171 static uint8* SerializeUnknownMessageSetItemsToArray(
172 const UnknownFieldSet& unknown_fields,
173 uint8* target);
174
175 // Compute the size of the UnknownFieldSet on the wire.
176 static int ComputeUnknownFieldsSize(const UnknownFieldSet& unknown_fields);
177
178 // Same thing except for messages that have the message_set_wire_format
179 // option.
180 static int ComputeUnknownMessageSetItemsSize(
181 const UnknownFieldSet& unknown_fields);
182
183
184 // Helper functions for encoding and decoding tags. (Inlined below and in
185 // _inl.h)
186 //
187 // This is different from MakeTag(field->number(), field->type()) in the case
188 // of packed repeated fields.
189 static uint32 MakeTag(const FieldDescriptor* field);
190
191 // Parse a single field. The input should start out positioned immediately
192 // after the tag.
193 static bool ParseAndMergeField(
194 uint32 tag,
195 const FieldDescriptor* field, // May be NULL for unknown
196 Message* message,
197 io::CodedInputStream* input);
198
199 // Serialize a single field.
200 static void SerializeFieldWithCachedSizes(
201 const FieldDescriptor* field, // Cannot be NULL
202 const Message& message,
203 io::CodedOutputStream* output);
204
205 // Compute size of a single field. If the field is a message type, this
206 // will call ByteSize() for the embedded message, insuring that it caches
207 // its size.
208 static int FieldByteSize(
209 const FieldDescriptor* field, // Cannot be NULL
210 const Message& message);
211
212 // Parse/serialize a MessageSet::Item group. Used with messages that use
213 // opion message_set_wire_format = true.
214 static bool ParseAndMergeMessageSetItem(
215 io::CodedInputStream* input,
216 Message* message);
217 static void SerializeMessageSetItemWithCachedSizes(
218 const FieldDescriptor* field,
219 const Message& message,
220 io::CodedOutputStream* output);
221 static int MessageSetItemByteSize(
222 const FieldDescriptor* field,
223 const Message& message);
224
225 // Computes the byte size of a field, excluding tags. For packed fields, it
226 // only includes the size of the raw data, and not the size of the total
227 // length, but for other length-delimited types, the size of the length is
228 // included.
229 static int FieldDataOnlyByteSize(
230 const FieldDescriptor* field, // Cannot be NULL
231 const Message& message);
232
233 enum Operation {
234 PARSE = 0,
235 SERIALIZE = 1,
236 };
237
238 // Verifies that a string field is valid UTF8, logging an error if not.
239 // This function will not be called by newly generated protobuf code
240 // but remains present to support existing code.
241 static void VerifyUTF8String(const char* data, int size, Operation op);
242 // The NamedField variant takes a field name in order to produce an
243 // informative error message if verification fails.
244 static void VerifyUTF8StringNamedField(const char* data,
245 int size,
246 Operation op,
247 const char* field_name);
248
249 private:
250 // Skip a MessageSet field.
251 static bool SkipMessageSetField(io::CodedInputStream* input,
252 uint32 field_number,
253 UnknownFieldSet* unknown_fields);
254
255 // Parse a MessageSet field.
256 static bool ParseAndMergeMessageSetField(uint32 field_number,
257 const FieldDescriptor* field,
258 Message* message,
259 io::CodedInputStream* input);
260
261 GOOGLE_DISALLOW_EVIL_CONSTRUCTORS(WireFormat);
262 };
263
264 // Subclass of FieldSkipper which saves skipped fields to an UnknownFieldSet.
265 class LIBPROTOBUF_EXPORT UnknownFieldSetFieldSkipper : public FieldSkipper {
266 public:
UnknownFieldSetFieldSkipper(UnknownFieldSet * unknown_fields)267 UnknownFieldSetFieldSkipper(UnknownFieldSet* unknown_fields)
268 : unknown_fields_(unknown_fields) {}
~UnknownFieldSetFieldSkipper()269 virtual ~UnknownFieldSetFieldSkipper() {}
270
271 // implements FieldSkipper -----------------------------------------
272 virtual bool SkipField(io::CodedInputStream* input, uint32 tag);
273 virtual bool SkipMessage(io::CodedInputStream* input);
274 virtual void SkipUnknownEnum(int field_number, int value);
275
276 protected:
277 UnknownFieldSet* unknown_fields_;
278 };
279
280 // inline methods ====================================================
281
WireTypeForField(const FieldDescriptor * field)282 inline WireFormatLite::WireType WireFormat::WireTypeForField(
283 const FieldDescriptor* field) {
284 if (field->is_packed()) {
285 return WireFormatLite::WIRETYPE_LENGTH_DELIMITED;
286 } else {
287 return WireTypeForFieldType(field->type());
288 }
289 }
290
WireTypeForFieldType(FieldDescriptor::Type type)291 inline WireFormatLite::WireType WireFormat::WireTypeForFieldType(
292 FieldDescriptor::Type type) {
293 // Some compilers don't like enum -> enum casts, so we implicit_cast to
294 // int first.
295 return WireFormatLite::WireTypeForFieldType(
296 static_cast<WireFormatLite::FieldType>(
297 implicit_cast<int>(type)));
298 }
299
MakeTag(const FieldDescriptor * field)300 inline uint32 WireFormat::MakeTag(const FieldDescriptor* field) {
301 return WireFormatLite::MakeTag(field->number(), WireTypeForField(field));
302 }
303
TagSize(int field_number,FieldDescriptor::Type type)304 inline int WireFormat::TagSize(int field_number, FieldDescriptor::Type type) {
305 // Some compilers don't like enum -> enum casts, so we implicit_cast to
306 // int first.
307 return WireFormatLite::TagSize(field_number,
308 static_cast<WireFormatLite::FieldType>(
309 implicit_cast<int>(type)));
310 }
311
VerifyUTF8String(const char * data,int size,WireFormat::Operation op)312 inline void WireFormat::VerifyUTF8String(const char* data, int size,
313 WireFormat::Operation op) {
314 #ifdef GOOGLE_PROTOBUF_UTF8_VALIDATION_ENABLED
315 WireFormatLite::VerifyUtf8String(
316 data, size, static_cast<WireFormatLite::Operation>(op), NULL);
317 #else
318 // Avoid the compiler warning about unused variables.
319 (void)data; (void)size; (void)op;
320 #endif
321 }
322
VerifyUTF8StringNamedField(const char * data,int size,WireFormat::Operation op,const char * field_name)323 inline void WireFormat::VerifyUTF8StringNamedField(
324 const char* data, int size, WireFormat::Operation op,
325 const char* field_name) {
326 #ifdef GOOGLE_PROTOBUF_UTF8_VALIDATION_ENABLED
327 WireFormatLite::VerifyUtf8String(
328 data, size, static_cast<WireFormatLite::Operation>(op), field_name);
329 #endif
330 }
331
332
333 } // namespace internal
334 } // namespace protobuf
335
336 } // namespace google
337 #endif // GOOGLE_PROTOBUF_WIRE_FORMAT_H__
338