• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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 //  Based on original Protocol Buffers design by
33 //  Sanjay Ghemawat, Jeff Dean, and others.
34 
35 #include <google/protobuf/wire_format_lite.h>
36 
37 #include <stack>
38 #include <string>
39 #include <vector>
40 
41 #include <google/protobuf/stubs/logging.h>
42 #include <google/protobuf/stubs/common.h>
43 #include <google/protobuf/stubs/stringprintf.h>
44 #include <google/protobuf/io/coded_stream.h>
45 #include <google/protobuf/io/zero_copy_stream.h>
46 #include <google/protobuf/io/zero_copy_stream_impl_lite.h>
47 
48 
49 #include <google/protobuf/port_def.inc>
50 
51 namespace google {
52 namespace protobuf {
53 namespace internal {
54 
55 #if !defined(_MSC_VER) || _MSC_VER >= 1900
56 // Old version of MSVC doesn't like definitions of inline constants, GCC
57 // requires them.
58 const int WireFormatLite::kMessageSetItemStartTag;
59 const int WireFormatLite::kMessageSetItemEndTag;
60 const int WireFormatLite::kMessageSetTypeIdTag;
61 const int WireFormatLite::kMessageSetMessageTag;
62 
63 #endif
64 
65 // IBM xlC requires prefixing constants with WireFormatLite::
66 const size_t WireFormatLite::kMessageSetItemTagsSize =
67     io::CodedOutputStream::StaticVarintSize32<
68         WireFormatLite::kMessageSetItemStartTag>::value +
69     io::CodedOutputStream::StaticVarintSize32<
70         WireFormatLite::kMessageSetItemEndTag>::value +
71     io::CodedOutputStream::StaticVarintSize32<
72         WireFormatLite::kMessageSetTypeIdTag>::value +
73     io::CodedOutputStream::StaticVarintSize32<
74         WireFormatLite::kMessageSetMessageTag>::value;
75 
76 const WireFormatLite::CppType
77     WireFormatLite::kFieldTypeToCppTypeMap[MAX_FIELD_TYPE + 1] = {
78         static_cast<CppType>(0),  // 0 is reserved for errors
79 
80         CPPTYPE_DOUBLE,   // TYPE_DOUBLE
81         CPPTYPE_FLOAT,    // TYPE_FLOAT
82         CPPTYPE_INT64,    // TYPE_INT64
83         CPPTYPE_UINT64,   // TYPE_UINT64
84         CPPTYPE_INT32,    // TYPE_INT32
85         CPPTYPE_UINT64,   // TYPE_FIXED64
86         CPPTYPE_UINT32,   // TYPE_FIXED32
87         CPPTYPE_BOOL,     // TYPE_BOOL
88         CPPTYPE_STRING,   // TYPE_STRING
89         CPPTYPE_MESSAGE,  // TYPE_GROUP
90         CPPTYPE_MESSAGE,  // TYPE_MESSAGE
91         CPPTYPE_STRING,   // TYPE_BYTES
92         CPPTYPE_UINT32,   // TYPE_UINT32
93         CPPTYPE_ENUM,     // TYPE_ENUM
94         CPPTYPE_INT32,    // TYPE_SFIXED32
95         CPPTYPE_INT64,    // TYPE_SFIXED64
96         CPPTYPE_INT32,    // TYPE_SINT32
97         CPPTYPE_INT64,    // TYPE_SINT64
98 };
99 
100 const WireFormatLite::WireType
101     WireFormatLite::kWireTypeForFieldType[MAX_FIELD_TYPE + 1] = {
102         static_cast<WireFormatLite::WireType>(-1),  // invalid
103         WireFormatLite::WIRETYPE_FIXED64,           // TYPE_DOUBLE
104         WireFormatLite::WIRETYPE_FIXED32,           // TYPE_FLOAT
105         WireFormatLite::WIRETYPE_VARINT,            // TYPE_INT64
106         WireFormatLite::WIRETYPE_VARINT,            // TYPE_UINT64
107         WireFormatLite::WIRETYPE_VARINT,            // TYPE_INT32
108         WireFormatLite::WIRETYPE_FIXED64,           // TYPE_FIXED64
109         WireFormatLite::WIRETYPE_FIXED32,           // TYPE_FIXED32
110         WireFormatLite::WIRETYPE_VARINT,            // TYPE_BOOL
111         WireFormatLite::WIRETYPE_LENGTH_DELIMITED,  // TYPE_STRING
112         WireFormatLite::WIRETYPE_START_GROUP,       // TYPE_GROUP
113         WireFormatLite::WIRETYPE_LENGTH_DELIMITED,  // TYPE_MESSAGE
114         WireFormatLite::WIRETYPE_LENGTH_DELIMITED,  // TYPE_BYTES
115         WireFormatLite::WIRETYPE_VARINT,            // TYPE_UINT32
116         WireFormatLite::WIRETYPE_VARINT,            // TYPE_ENUM
117         WireFormatLite::WIRETYPE_FIXED32,           // TYPE_SFIXED32
118         WireFormatLite::WIRETYPE_FIXED64,           // TYPE_SFIXED64
119         WireFormatLite::WIRETYPE_VARINT,            // TYPE_SINT32
120         WireFormatLite::WIRETYPE_VARINT,            // TYPE_SINT64
121 };
122 
SkipField(io::CodedInputStream * input,uint32 tag)123 bool WireFormatLite::SkipField(io::CodedInputStream* input, uint32 tag) {
124   // Field number 0 is illegal.
125   if (WireFormatLite::GetTagFieldNumber(tag) == 0) return false;
126   switch (WireFormatLite::GetTagWireType(tag)) {
127     case WireFormatLite::WIRETYPE_VARINT: {
128       uint64 value;
129       if (!input->ReadVarint64(&value)) return false;
130       return true;
131     }
132     case WireFormatLite::WIRETYPE_FIXED64: {
133       uint64 value;
134       if (!input->ReadLittleEndian64(&value)) return false;
135       return true;
136     }
137     case WireFormatLite::WIRETYPE_LENGTH_DELIMITED: {
138       uint32 length;
139       if (!input->ReadVarint32(&length)) return false;
140       if (!input->Skip(length)) return false;
141       return true;
142     }
143     case WireFormatLite::WIRETYPE_START_GROUP: {
144       if (!input->IncrementRecursionDepth()) return false;
145       if (!SkipMessage(input)) return false;
146       input->DecrementRecursionDepth();
147       // Check that the ending tag matched the starting tag.
148       if (!input->LastTagWas(
149               WireFormatLite::MakeTag(WireFormatLite::GetTagFieldNumber(tag),
150                                       WireFormatLite::WIRETYPE_END_GROUP))) {
151         return false;
152       }
153       return true;
154     }
155     case WireFormatLite::WIRETYPE_END_GROUP: {
156       return false;
157     }
158     case WireFormatLite::WIRETYPE_FIXED32: {
159       uint32 value;
160       if (!input->ReadLittleEndian32(&value)) return false;
161       return true;
162     }
163     default: {
164       return false;
165     }
166   }
167 }
168 
SkipField(io::CodedInputStream * input,uint32 tag,io::CodedOutputStream * output)169 bool WireFormatLite::SkipField(io::CodedInputStream* input, uint32 tag,
170                                io::CodedOutputStream* output) {
171   // Field number 0 is illegal.
172   if (WireFormatLite::GetTagFieldNumber(tag) == 0) return false;
173   switch (WireFormatLite::GetTagWireType(tag)) {
174     case WireFormatLite::WIRETYPE_VARINT: {
175       uint64 value;
176       if (!input->ReadVarint64(&value)) return false;
177       output->WriteVarint32(tag);
178       output->WriteVarint64(value);
179       return true;
180     }
181     case WireFormatLite::WIRETYPE_FIXED64: {
182       uint64 value;
183       if (!input->ReadLittleEndian64(&value)) return false;
184       output->WriteVarint32(tag);
185       output->WriteLittleEndian64(value);
186       return true;
187     }
188     case WireFormatLite::WIRETYPE_LENGTH_DELIMITED: {
189       uint32 length;
190       if (!input->ReadVarint32(&length)) return false;
191       output->WriteVarint32(tag);
192       output->WriteVarint32(length);
193       // TODO(mkilavuz): Provide API to prevent extra string copying.
194       std::string temp;
195       if (!input->ReadString(&temp, length)) return false;
196       output->WriteString(temp);
197       return true;
198     }
199     case WireFormatLite::WIRETYPE_START_GROUP: {
200       output->WriteVarint32(tag);
201       if (!input->IncrementRecursionDepth()) return false;
202       if (!SkipMessage(input, output)) return false;
203       input->DecrementRecursionDepth();
204       // Check that the ending tag matched the starting tag.
205       if (!input->LastTagWas(
206               WireFormatLite::MakeTag(WireFormatLite::GetTagFieldNumber(tag),
207                                       WireFormatLite::WIRETYPE_END_GROUP))) {
208         return false;
209       }
210       return true;
211     }
212     case WireFormatLite::WIRETYPE_END_GROUP: {
213       return false;
214     }
215     case WireFormatLite::WIRETYPE_FIXED32: {
216       uint32 value;
217       if (!input->ReadLittleEndian32(&value)) return false;
218       output->WriteVarint32(tag);
219       output->WriteLittleEndian32(value);
220       return true;
221     }
222     default: {
223       return false;
224     }
225   }
226 }
227 
SkipMessage(io::CodedInputStream * input)228 bool WireFormatLite::SkipMessage(io::CodedInputStream* input) {
229   while (true) {
230     uint32 tag = input->ReadTag();
231     if (tag == 0) {
232       // End of input.  This is a valid place to end, so return true.
233       return true;
234     }
235 
236     WireFormatLite::WireType wire_type = WireFormatLite::GetTagWireType(tag);
237 
238     if (wire_type == WireFormatLite::WIRETYPE_END_GROUP) {
239       // Must be the end of the message.
240       return true;
241     }
242 
243     if (!SkipField(input, tag)) return false;
244   }
245 }
246 
SkipMessage(io::CodedInputStream * input,io::CodedOutputStream * output)247 bool WireFormatLite::SkipMessage(io::CodedInputStream* input,
248                                  io::CodedOutputStream* output) {
249   while (true) {
250     uint32 tag = input->ReadTag();
251     if (tag == 0) {
252       // End of input.  This is a valid place to end, so return true.
253       return true;
254     }
255 
256     WireFormatLite::WireType wire_type = WireFormatLite::GetTagWireType(tag);
257 
258     if (wire_type == WireFormatLite::WIRETYPE_END_GROUP) {
259       output->WriteVarint32(tag);
260       // Must be the end of the message.
261       return true;
262     }
263 
264     if (!SkipField(input, tag, output)) return false;
265   }
266 }
267 
SkipField(io::CodedInputStream * input,uint32 tag)268 bool FieldSkipper::SkipField(io::CodedInputStream* input, uint32 tag) {
269   return WireFormatLite::SkipField(input, tag);
270 }
271 
SkipMessage(io::CodedInputStream * input)272 bool FieldSkipper::SkipMessage(io::CodedInputStream* input) {
273   return WireFormatLite::SkipMessage(input);
274 }
275 
SkipUnknownEnum(int,int)276 void FieldSkipper::SkipUnknownEnum(int /* field_number */, int /* value */) {
277   // Nothing.
278 }
279 
SkipField(io::CodedInputStream * input,uint32 tag)280 bool CodedOutputStreamFieldSkipper::SkipField(io::CodedInputStream* input,
281                                               uint32 tag) {
282   return WireFormatLite::SkipField(input, tag, unknown_fields_);
283 }
284 
SkipMessage(io::CodedInputStream * input)285 bool CodedOutputStreamFieldSkipper::SkipMessage(io::CodedInputStream* input) {
286   return WireFormatLite::SkipMessage(input, unknown_fields_);
287 }
288 
SkipUnknownEnum(int field_number,int value)289 void CodedOutputStreamFieldSkipper::SkipUnknownEnum(int field_number,
290                                                     int value) {
291   unknown_fields_->WriteVarint32(field_number);
292   unknown_fields_->WriteVarint64(value);
293 }
294 
ReadPackedEnumPreserveUnknowns(io::CodedInputStream * input,int field_number,bool (* is_valid)(int),io::CodedOutputStream * unknown_fields_stream,RepeatedField<int> * values)295 bool WireFormatLite::ReadPackedEnumPreserveUnknowns(
296     io::CodedInputStream* input, int field_number, bool (*is_valid)(int),
297     io::CodedOutputStream* unknown_fields_stream, RepeatedField<int>* values) {
298   uint32 length;
299   if (!input->ReadVarint32(&length)) return false;
300   io::CodedInputStream::Limit limit = input->PushLimit(length);
301   while (input->BytesUntilLimit() > 0) {
302     int value;
303     if (!ReadPrimitive<int, WireFormatLite::TYPE_ENUM>(input, &value)) {
304       return false;
305     }
306     if (is_valid == NULL || is_valid(value)) {
307       values->Add(value);
308     } else {
309       uint32 tag = WireFormatLite::MakeTag(field_number,
310                                            WireFormatLite::WIRETYPE_VARINT);
311       unknown_fields_stream->WriteVarint32(tag);
312       unknown_fields_stream->WriteVarint32(value);
313     }
314   }
315   input->PopLimit(limit);
316   return true;
317 }
318 
319 #if !defined(PROTOBUF_LITTLE_ENDIAN)
320 
321 namespace {
EncodeFixedSizeValue(float v,uint8 * dest)322 void EncodeFixedSizeValue(float v, uint8* dest) {
323   WireFormatLite::WriteFloatNoTagToArray(v, dest);
324 }
325 
EncodeFixedSizeValue(double v,uint8 * dest)326 void EncodeFixedSizeValue(double v, uint8* dest) {
327   WireFormatLite::WriteDoubleNoTagToArray(v, dest);
328 }
329 
EncodeFixedSizeValue(uint32 v,uint8 * dest)330 void EncodeFixedSizeValue(uint32 v, uint8* dest) {
331   WireFormatLite::WriteFixed32NoTagToArray(v, dest);
332 }
333 
EncodeFixedSizeValue(uint64 v,uint8 * dest)334 void EncodeFixedSizeValue(uint64 v, uint8* dest) {
335   WireFormatLite::WriteFixed64NoTagToArray(v, dest);
336 }
337 
EncodeFixedSizeValue(int32 v,uint8 * dest)338 void EncodeFixedSizeValue(int32 v, uint8* dest) {
339   WireFormatLite::WriteSFixed32NoTagToArray(v, dest);
340 }
341 
EncodeFixedSizeValue(int64 v,uint8 * dest)342 void EncodeFixedSizeValue(int64 v, uint8* dest) {
343   WireFormatLite::WriteSFixed64NoTagToArray(v, dest);
344 }
345 
EncodeFixedSizeValue(bool v,uint8 * dest)346 void EncodeFixedSizeValue(bool v, uint8* dest) {
347   WireFormatLite::WriteBoolNoTagToArray(v, dest);
348 }
349 }  // anonymous namespace
350 
351 #endif  // !defined(PROTOBUF_LITTLE_ENDIAN)
352 
353 template <typename CType>
WriteArray(const CType * a,int n,io::CodedOutputStream * output)354 static void WriteArray(const CType* a, int n, io::CodedOutputStream* output) {
355 #if defined(PROTOBUF_LITTLE_ENDIAN)
356   output->WriteRaw(reinterpret_cast<const char*>(a), n * sizeof(a[0]));
357 #else
358   const int kAtATime = 128;
359   uint8 buf[sizeof(CType) * kAtATime];
360   for (int i = 0; i < n; i += kAtATime) {
361     int to_do = std::min(kAtATime, n - i);
362     uint8* ptr = buf;
363     for (int j = 0; j < to_do; j++) {
364       EncodeFixedSizeValue(a[i + j], ptr);
365       ptr += sizeof(a[0]);
366     }
367     output->WriteRaw(buf, to_do * sizeof(a[0]));
368   }
369 #endif
370 }
371 
WriteFloatArray(const float * a,int n,io::CodedOutputStream * output)372 void WireFormatLite::WriteFloatArray(const float* a, int n,
373                                      io::CodedOutputStream* output) {
374   WriteArray<float>(a, n, output);
375 }
376 
WriteDoubleArray(const double * a,int n,io::CodedOutputStream * output)377 void WireFormatLite::WriteDoubleArray(const double* a, int n,
378                                       io::CodedOutputStream* output) {
379   WriteArray<double>(a, n, output);
380 }
381 
WriteFixed32Array(const uint32 * a,int n,io::CodedOutputStream * output)382 void WireFormatLite::WriteFixed32Array(const uint32* a, int n,
383                                        io::CodedOutputStream* output) {
384   WriteArray<uint32>(a, n, output);
385 }
386 
WriteFixed64Array(const uint64 * a,int n,io::CodedOutputStream * output)387 void WireFormatLite::WriteFixed64Array(const uint64* a, int n,
388                                        io::CodedOutputStream* output) {
389   WriteArray<uint64>(a, n, output);
390 }
391 
WriteSFixed32Array(const int32 * a,int n,io::CodedOutputStream * output)392 void WireFormatLite::WriteSFixed32Array(const int32* a, int n,
393                                         io::CodedOutputStream* output) {
394   WriteArray<int32>(a, n, output);
395 }
396 
WriteSFixed64Array(const int64 * a,int n,io::CodedOutputStream * output)397 void WireFormatLite::WriteSFixed64Array(const int64* a, int n,
398                                         io::CodedOutputStream* output) {
399   WriteArray<int64>(a, n, output);
400 }
401 
WriteBoolArray(const bool * a,int n,io::CodedOutputStream * output)402 void WireFormatLite::WriteBoolArray(const bool* a, int n,
403                                     io::CodedOutputStream* output) {
404   WriteArray<bool>(a, n, output);
405 }
406 
WriteInt32(int field_number,int32 value,io::CodedOutputStream * output)407 void WireFormatLite::WriteInt32(int field_number, int32 value,
408                                 io::CodedOutputStream* output) {
409   WriteTag(field_number, WIRETYPE_VARINT, output);
410   WriteInt32NoTag(value, output);
411 }
WriteInt64(int field_number,int64 value,io::CodedOutputStream * output)412 void WireFormatLite::WriteInt64(int field_number, int64 value,
413                                 io::CodedOutputStream* output) {
414   WriteTag(field_number, WIRETYPE_VARINT, output);
415   WriteInt64NoTag(value, output);
416 }
WriteUInt32(int field_number,uint32 value,io::CodedOutputStream * output)417 void WireFormatLite::WriteUInt32(int field_number, uint32 value,
418                                  io::CodedOutputStream* output) {
419   WriteTag(field_number, WIRETYPE_VARINT, output);
420   WriteUInt32NoTag(value, output);
421 }
WriteUInt64(int field_number,uint64 value,io::CodedOutputStream * output)422 void WireFormatLite::WriteUInt64(int field_number, uint64 value,
423                                  io::CodedOutputStream* output) {
424   WriteTag(field_number, WIRETYPE_VARINT, output);
425   WriteUInt64NoTag(value, output);
426 }
WriteSInt32(int field_number,int32 value,io::CodedOutputStream * output)427 void WireFormatLite::WriteSInt32(int field_number, int32 value,
428                                  io::CodedOutputStream* output) {
429   WriteTag(field_number, WIRETYPE_VARINT, output);
430   WriteSInt32NoTag(value, output);
431 }
WriteSInt64(int field_number,int64 value,io::CodedOutputStream * output)432 void WireFormatLite::WriteSInt64(int field_number, int64 value,
433                                  io::CodedOutputStream* output) {
434   WriteTag(field_number, WIRETYPE_VARINT, output);
435   WriteSInt64NoTag(value, output);
436 }
WriteFixed32(int field_number,uint32 value,io::CodedOutputStream * output)437 void WireFormatLite::WriteFixed32(int field_number, uint32 value,
438                                   io::CodedOutputStream* output) {
439   WriteTag(field_number, WIRETYPE_FIXED32, output);
440   WriteFixed32NoTag(value, output);
441 }
WriteFixed64(int field_number,uint64 value,io::CodedOutputStream * output)442 void WireFormatLite::WriteFixed64(int field_number, uint64 value,
443                                   io::CodedOutputStream* output) {
444   WriteTag(field_number, WIRETYPE_FIXED64, output);
445   WriteFixed64NoTag(value, output);
446 }
WriteSFixed32(int field_number,int32 value,io::CodedOutputStream * output)447 void WireFormatLite::WriteSFixed32(int field_number, int32 value,
448                                    io::CodedOutputStream* output) {
449   WriteTag(field_number, WIRETYPE_FIXED32, output);
450   WriteSFixed32NoTag(value, output);
451 }
WriteSFixed64(int field_number,int64 value,io::CodedOutputStream * output)452 void WireFormatLite::WriteSFixed64(int field_number, int64 value,
453                                    io::CodedOutputStream* output) {
454   WriteTag(field_number, WIRETYPE_FIXED64, output);
455   WriteSFixed64NoTag(value, output);
456 }
WriteFloat(int field_number,float value,io::CodedOutputStream * output)457 void WireFormatLite::WriteFloat(int field_number, float value,
458                                 io::CodedOutputStream* output) {
459   WriteTag(field_number, WIRETYPE_FIXED32, output);
460   WriteFloatNoTag(value, output);
461 }
WriteDouble(int field_number,double value,io::CodedOutputStream * output)462 void WireFormatLite::WriteDouble(int field_number, double value,
463                                  io::CodedOutputStream* output) {
464   WriteTag(field_number, WIRETYPE_FIXED64, output);
465   WriteDoubleNoTag(value, output);
466 }
WriteBool(int field_number,bool value,io::CodedOutputStream * output)467 void WireFormatLite::WriteBool(int field_number, bool value,
468                                io::CodedOutputStream* output) {
469   WriteTag(field_number, WIRETYPE_VARINT, output);
470   WriteBoolNoTag(value, output);
471 }
WriteEnum(int field_number,int value,io::CodedOutputStream * output)472 void WireFormatLite::WriteEnum(int field_number, int value,
473                                io::CodedOutputStream* output) {
474   WriteTag(field_number, WIRETYPE_VARINT, output);
475   WriteEnumNoTag(value, output);
476 }
477 
WriteString(int field_number,const std::string & value,io::CodedOutputStream * output)478 void WireFormatLite::WriteString(int field_number, const std::string& value,
479                                  io::CodedOutputStream* output) {
480   // String is for UTF-8 text only
481   WriteTag(field_number, WIRETYPE_LENGTH_DELIMITED, output);
482   GOOGLE_CHECK_LE(value.size(), kint32max);
483   output->WriteVarint32(value.size());
484   output->WriteString(value);
485 }
WriteStringMaybeAliased(int field_number,const std::string & value,io::CodedOutputStream * output)486 void WireFormatLite::WriteStringMaybeAliased(int field_number,
487                                              const std::string& value,
488                                              io::CodedOutputStream* output) {
489   // String is for UTF-8 text only
490   WriteTag(field_number, WIRETYPE_LENGTH_DELIMITED, output);
491   GOOGLE_CHECK_LE(value.size(), kint32max);
492   output->WriteVarint32(value.size());
493   output->WriteRawMaybeAliased(value.data(), value.size());
494 }
WriteBytes(int field_number,const std::string & value,io::CodedOutputStream * output)495 void WireFormatLite::WriteBytes(int field_number, const std::string& value,
496                                 io::CodedOutputStream* output) {
497   WriteTag(field_number, WIRETYPE_LENGTH_DELIMITED, output);
498   GOOGLE_CHECK_LE(value.size(), kint32max);
499   output->WriteVarint32(value.size());
500   output->WriteString(value);
501 }
WriteBytesMaybeAliased(int field_number,const std::string & value,io::CodedOutputStream * output)502 void WireFormatLite::WriteBytesMaybeAliased(int field_number,
503                                             const std::string& value,
504                                             io::CodedOutputStream* output) {
505   WriteTag(field_number, WIRETYPE_LENGTH_DELIMITED, output);
506   GOOGLE_CHECK_LE(value.size(), kint32max);
507   output->WriteVarint32(value.size());
508   output->WriteRawMaybeAliased(value.data(), value.size());
509 }
510 
511 
WriteGroup(int field_number,const MessageLite & value,io::CodedOutputStream * output)512 void WireFormatLite::WriteGroup(int field_number, const MessageLite& value,
513                                 io::CodedOutputStream* output) {
514   WriteTag(field_number, WIRETYPE_START_GROUP, output);
515   value.SerializeWithCachedSizes(output);
516   WriteTag(field_number, WIRETYPE_END_GROUP, output);
517 }
518 
WriteMessage(int field_number,const MessageLite & value,io::CodedOutputStream * output)519 void WireFormatLite::WriteMessage(int field_number, const MessageLite& value,
520                                   io::CodedOutputStream* output) {
521   WriteTag(field_number, WIRETYPE_LENGTH_DELIMITED, output);
522   const int size = value.GetCachedSize();
523   output->WriteVarint32(size);
524   value.SerializeWithCachedSizes(output);
525 }
526 
WriteSubMessageMaybeToArray(int size,const MessageLite & value,io::CodedOutputStream * output)527 void WireFormatLite::WriteSubMessageMaybeToArray(
528     int size, const MessageLite& value, io::CodedOutputStream* output) {
529   output->SetCur(value._InternalSerialize(output->Cur(), output->EpsCopy()));
530 }
531 
WriteGroupMaybeToArray(int field_number,const MessageLite & value,io::CodedOutputStream * output)532 void WireFormatLite::WriteGroupMaybeToArray(int field_number,
533                                             const MessageLite& value,
534                                             io::CodedOutputStream* output) {
535   WriteTag(field_number, WIRETYPE_START_GROUP, output);
536   const int size = value.GetCachedSize();
537   WriteSubMessageMaybeToArray(size, value, output);
538   WriteTag(field_number, WIRETYPE_END_GROUP, output);
539 }
540 
WriteMessageMaybeToArray(int field_number,const MessageLite & value,io::CodedOutputStream * output)541 void WireFormatLite::WriteMessageMaybeToArray(int field_number,
542                                               const MessageLite& value,
543                                               io::CodedOutputStream* output) {
544   WriteTag(field_number, WIRETYPE_LENGTH_DELIMITED, output);
545   const int size = value.GetCachedSize();
546   output->WriteVarint32(size);
547   WriteSubMessageMaybeToArray(size, value, output);
548 }
549 
550 PROTOBUF_ALWAYS_INLINE static bool ReadBytesToString(
551     io::CodedInputStream* input, std::string* value);
ReadBytesToString(io::CodedInputStream * input,std::string * value)552 inline static bool ReadBytesToString(io::CodedInputStream* input,
553                                      std::string* value) {
554   uint32 length;
555   return input->ReadVarint32(&length) && input->ReadString(value, length);
556 }
557 
ReadBytes(io::CodedInputStream * input,std::string * value)558 bool WireFormatLite::ReadBytes(io::CodedInputStream* input,
559                                std::string* value) {
560   return ReadBytesToString(input, value);
561 }
562 
ReadBytes(io::CodedInputStream * input,std::string ** p)563 bool WireFormatLite::ReadBytes(io::CodedInputStream* input, std::string** p) {
564   if (*p == &GetEmptyStringAlreadyInited()) {
565     *p = new std::string();
566   }
567   return ReadBytesToString(input, *p);
568 }
569 
PrintUTF8ErrorLog(const char * field_name,const char * operation_str,bool emit_stacktrace)570 void PrintUTF8ErrorLog(const char* field_name, const char* operation_str,
571                        bool emit_stacktrace) {
572   std::string stacktrace;
573   std::string quoted_field_name = "";
574   if (field_name != nullptr) {
575     quoted_field_name = StringPrintf(" '%s'", field_name);
576   }
577   GOOGLE_LOG(ERROR) << "String field" << quoted_field_name << " contains invalid "
578              << "UTF-8 data when " << operation_str << " a protocol "
579              << "buffer. Use the 'bytes' type if you intend to send raw "
580              << "bytes. " << stacktrace;
581 }
582 
VerifyUtf8String(const char * data,int size,Operation op,const char * field_name)583 bool WireFormatLite::VerifyUtf8String(const char* data, int size, Operation op,
584                                       const char* field_name) {
585   if (!IsStructurallyValidUTF8(data, size)) {
586     const char* operation_str = NULL;
587     switch (op) {
588       case PARSE:
589         operation_str = "parsing";
590         break;
591       case SERIALIZE:
592         operation_str = "serializing";
593         break;
594         // no default case: have the compiler warn if a case is not covered.
595     }
596     PrintUTF8ErrorLog(field_name, operation_str, false);
597     return false;
598   }
599   return true;
600 }
601 
602 // this code is deliberately written such that clang makes it into really
603 // efficient SSE code.
604 template <bool ZigZag, bool SignExtended, typename T>
VarintSize(const T * data,const int n)605 static size_t VarintSize(const T* data, const int n) {
606 #if __cplusplus >= 201103L
607   static_assert(sizeof(T) == 4, "This routine only works for 32 bit integers");
608   // is_unsigned<T> => !ZigZag
609   static_assert(
610       (std::is_unsigned<T>::value ^ ZigZag) || std::is_signed<T>::value,
611       "Cannot ZigZag encode unsigned types");
612   // is_unsigned<T> => !SignExtended
613   static_assert(
614       (std::is_unsigned<T>::value ^ SignExtended) || std::is_signed<T>::value,
615       "Cannot SignExtended unsigned types");
616   static_assert(!(SignExtended && ZigZag),
617                 "Cannot SignExtended and ZigZag on the same type");
618 #endif
619   uint32 sum = n;
620   uint32 msb_sum = 0;
621   for (int i = 0; i < n; i++) {
622     uint32 x = data[i];
623     if (ZigZag) {
624       x = WireFormatLite::ZigZagEncode32(x);
625     } else if (SignExtended) {
626       msb_sum += x >> 31;
627     }
628     // clang is so smart that it produces optimal SSE sequence unrolling
629     // the loop 8 ints at a time. With a sequence of 4
630     // cmpres = cmpgt x, sizeclass  ( -1 or 0)
631     // sum = sum - cmpres
632     if (x > 0x7F) sum++;
633     if (x > 0x3FFF) sum++;
634     if (x > 0x1FFFFF) sum++;
635     if (x > 0xFFFFFFF) sum++;
636   }
637   if (SignExtended) sum += msb_sum * 5;
638   return sum;
639 }
640 
641 template <bool ZigZag, typename T>
VarintSize64(const T * data,const int n)642 static size_t VarintSize64(const T* data, const int n) {
643 #if __cplusplus >= 201103L
644   static_assert(sizeof(T) == 8, "This routine only works for 64 bit integers");
645   // is_unsigned<T> => !ZigZag
646   static_assert(!ZigZag || !std::is_unsigned<T>::value,
647                 "Cannot ZigZag encode unsigned types");
648 #endif
649   uint64 sum = n;
650   for (int i = 0; i < n; i++) {
651     uint64 x = data[i];
652     if (ZigZag) {
653       x = WireFormatLite::ZigZagEncode64(x);
654     }
655     // First step is a binary search, we can't branch in sse so we use the
656     // result of the compare to adjust sum and appropriately. This code is
657     // written to make clang recognize the vectorization.
658     uint64 tmp = x >= (static_cast<uint64>(1) << 35) ? -1 : 0;
659     sum += 5 & tmp;
660     x >>= 35 & tmp;
661     if (x > 0x7F) sum++;
662     if (x > 0x3FFF) sum++;
663     if (x > 0x1FFFFF) sum++;
664     if (x > 0xFFFFFFF) sum++;
665   }
666   return sum;
667 }
668 
669 // GCC does not recognize the vectorization opportunity
670 // and other platforms are untested, in those cases using the optimized
671 // varint size routine for each element is faster.
672 // Hence we enable it only for clang
673 #if defined(__SSE__) && defined(__clang__)
Int32Size(const RepeatedField<int32> & value)674 size_t WireFormatLite::Int32Size(const RepeatedField<int32>& value) {
675   return VarintSize<false, true>(value.data(), value.size());
676 }
677 
UInt32Size(const RepeatedField<uint32> & value)678 size_t WireFormatLite::UInt32Size(const RepeatedField<uint32>& value) {
679   return VarintSize<false, false>(value.data(), value.size());
680 }
681 
SInt32Size(const RepeatedField<int32> & value)682 size_t WireFormatLite::SInt32Size(const RepeatedField<int32>& value) {
683   return VarintSize<true, false>(value.data(), value.size());
684 }
685 
EnumSize(const RepeatedField<int> & value)686 size_t WireFormatLite::EnumSize(const RepeatedField<int>& value) {
687   // On ILP64, sizeof(int) == 8, which would require a different template.
688   return VarintSize<false, true>(value.data(), value.size());
689 }
690 
691 #else  // !(defined(__SSE4_1__) && defined(__clang__))
692 
Int32Size(const RepeatedField<int32> & value)693 size_t WireFormatLite::Int32Size(const RepeatedField<int32>& value) {
694   size_t out = 0;
695   const int n = value.size();
696   for (int i = 0; i < n; i++) {
697     out += Int32Size(value.Get(i));
698   }
699   return out;
700 }
701 
UInt32Size(const RepeatedField<uint32> & value)702 size_t WireFormatLite::UInt32Size(const RepeatedField<uint32>& value) {
703   size_t out = 0;
704   const int n = value.size();
705   for (int i = 0; i < n; i++) {
706     out += UInt32Size(value.Get(i));
707   }
708   return out;
709 }
710 
SInt32Size(const RepeatedField<int32> & value)711 size_t WireFormatLite::SInt32Size(const RepeatedField<int32>& value) {
712   size_t out = 0;
713   const int n = value.size();
714   for (int i = 0; i < n; i++) {
715     out += SInt32Size(value.Get(i));
716   }
717   return out;
718 }
719 
EnumSize(const RepeatedField<int> & value)720 size_t WireFormatLite::EnumSize(const RepeatedField<int>& value) {
721   size_t out = 0;
722   const int n = value.size();
723   for (int i = 0; i < n; i++) {
724     out += EnumSize(value.Get(i));
725   }
726   return out;
727 }
728 
729 #endif
730 
731 // Micro benchmarks show that the SSE improved loop only starts beating
732 // the normal loop on Haswell platforms and then only for >32 ints. We
733 // disable this for now. Some specialized users might find it worthwhile to
734 // enable this.
735 #define USE_SSE_FOR_64_BIT_INTEGER_ARRAYS 0
736 #if USE_SSE_FOR_64_BIT_INTEGER_ARRAYS
Int64Size(const RepeatedField<int64> & value)737 size_t WireFormatLite::Int64Size(const RepeatedField<int64>& value) {
738   return VarintSize64<false>(value.data(), value.size());
739 }
740 
UInt64Size(const RepeatedField<uint64> & value)741 size_t WireFormatLite::UInt64Size(const RepeatedField<uint64>& value) {
742   return VarintSize64<false>(value.data(), value.size());
743 }
744 
SInt64Size(const RepeatedField<int64> & value)745 size_t WireFormatLite::SInt64Size(const RepeatedField<int64>& value) {
746   return VarintSize64<true>(value.data(), value.size());
747 }
748 
749 #else
750 
Int64Size(const RepeatedField<int64> & value)751 size_t WireFormatLite::Int64Size(const RepeatedField<int64>& value) {
752   size_t out = 0;
753   const int n = value.size();
754   for (int i = 0; i < n; i++) {
755     out += Int64Size(value.Get(i));
756   }
757   return out;
758 }
759 
UInt64Size(const RepeatedField<uint64> & value)760 size_t WireFormatLite::UInt64Size(const RepeatedField<uint64>& value) {
761   size_t out = 0;
762   const int n = value.size();
763   for (int i = 0; i < n; i++) {
764     out += UInt64Size(value.Get(i));
765   }
766   return out;
767 }
768 
SInt64Size(const RepeatedField<int64> & value)769 size_t WireFormatLite::SInt64Size(const RepeatedField<int64>& value) {
770   size_t out = 0;
771   const int n = value.size();
772   for (int i = 0; i < n; i++) {
773     out += SInt64Size(value.Get(i));
774   }
775   return out;
776 }
777 
778 #endif
779 
780 }  // namespace internal
781 }  // namespace protobuf
782 }  // namespace google
783