1 // Copyright 2007, Google Inc.
2 // All rights reserved.
3 //
4 // Redistribution and use in source and binary forms, with or without
5 // modification, are permitted provided that the following conditions are
6 // met:
7 //
8 // * Redistributions of source code must retain the above copyright
9 // notice, this list of conditions and the following disclaimer.
10 // * Redistributions in binary form must reproduce the above
11 // copyright notice, this list of conditions and the following disclaimer
12 // in the documentation and/or other materials provided with the
13 // distribution.
14 // * Neither the name of Google Inc. nor the names of its
15 // contributors may be used to endorse or promote products derived from
16 // this software without specific prior written permission.
17 //
18 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
21 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
22 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
23 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
24 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
25 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
26 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
28 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
29
30
31 // Google Test - The Google C++ Testing and Mocking Framework
32 //
33 // This file implements a universal value printer that can print a
34 // value of any type T:
35 //
36 // void ::testing::internal::UniversalPrinter<T>::Print(value, ostream_ptr);
37 //
38 // It uses the << operator when possible, and prints the bytes in the
39 // object otherwise. A user can override its behavior for a class
40 // type Foo by defining either operator<<(::std::ostream&, const Foo&)
41 // or void PrintTo(const Foo&, ::std::ostream*) in the namespace that
42 // defines Foo.
43
44 #include "gtest/gtest-printers.h"
45 #include <stdio.h>
46 #include <cctype>
47 #include <cwchar>
48 #include <ostream> // NOLINT
49 #include <string>
50 #include "gtest/internal/gtest-port.h"
51 #include "src/gtest-internal-inl.h"
52
53 namespace testing {
54
55 namespace {
56
57 using ::std::ostream;
58
59 // Prints a segment of bytes in the given object.
60 GTEST_ATTRIBUTE_NO_SANITIZE_MEMORY_
61 GTEST_ATTRIBUTE_NO_SANITIZE_ADDRESS_
62 GTEST_ATTRIBUTE_NO_SANITIZE_HWADDRESS_
63 GTEST_ATTRIBUTE_NO_SANITIZE_THREAD_
PrintByteSegmentInObjectTo(const unsigned char * obj_bytes,size_t start,size_t count,ostream * os)64 void PrintByteSegmentInObjectTo(const unsigned char* obj_bytes, size_t start,
65 size_t count, ostream* os) {
66 char text[5] = "";
67 for (size_t i = 0; i != count; i++) {
68 const size_t j = start + i;
69 if (i != 0) {
70 // Organizes the bytes into groups of 2 for easy parsing by
71 // human.
72 if ((j % 2) == 0)
73 *os << ' ';
74 else
75 *os << '-';
76 }
77 GTEST_SNPRINTF_(text, sizeof(text), "%02X", obj_bytes[j]);
78 *os << text;
79 }
80 }
81
82 // Prints the bytes in the given value to the given ostream.
PrintBytesInObjectToImpl(const unsigned char * obj_bytes,size_t count,ostream * os)83 void PrintBytesInObjectToImpl(const unsigned char* obj_bytes, size_t count,
84 ostream* os) {
85 // Tells the user how big the object is.
86 *os << count << "-byte object <";
87
88 const size_t kThreshold = 132;
89 const size_t kChunkSize = 64;
90 // If the object size is bigger than kThreshold, we'll have to omit
91 // some details by printing only the first and the last kChunkSize
92 // bytes.
93 if (count < kThreshold) {
94 PrintByteSegmentInObjectTo(obj_bytes, 0, count, os);
95 } else {
96 PrintByteSegmentInObjectTo(obj_bytes, 0, kChunkSize, os);
97 *os << " ... ";
98 // Rounds up to 2-byte boundary.
99 const size_t resume_pos = (count - kChunkSize + 1)/2*2;
100 PrintByteSegmentInObjectTo(obj_bytes, resume_pos, count - resume_pos, os);
101 }
102 *os << ">";
103 }
104
105 } // namespace
106
107 namespace internal {
108
109 // Delegates to PrintBytesInObjectToImpl() to print the bytes in the
110 // given object. The delegation simplifies the implementation, which
111 // uses the << operator and thus is easier done outside of the
112 // ::testing::internal namespace, which contains a << operator that
113 // sometimes conflicts with the one in STL.
PrintBytesInObjectTo(const unsigned char * obj_bytes,size_t count,ostream * os)114 void PrintBytesInObjectTo(const unsigned char* obj_bytes, size_t count,
115 ostream* os) {
116 PrintBytesInObjectToImpl(obj_bytes, count, os);
117 }
118
119 // Depending on the value of a char (or wchar_t), we print it in one
120 // of three formats:
121 // - as is if it's a printable ASCII (e.g. 'a', '2', ' '),
122 // - as a hexadecimal escape sequence (e.g. '\x7F'), or
123 // - as a special escape sequence (e.g. '\r', '\n').
124 enum CharFormat {
125 kAsIs,
126 kHexEscape,
127 kSpecialEscape
128 };
129
130 // Returns true if c is a printable ASCII character. We test the
131 // value of c directly instead of calling isprint(), which is buggy on
132 // Windows Mobile.
IsPrintableAscii(wchar_t c)133 inline bool IsPrintableAscii(wchar_t c) {
134 return 0x20 <= c && c <= 0x7E;
135 }
136
137 // Prints a wide or narrow char c as a character literal without the
138 // quotes, escaping it when necessary; returns how c was formatted.
139 // The template argument UnsignedChar is the unsigned version of Char,
140 // which is the type of c.
141 template <typename UnsignedChar, typename Char>
PrintAsCharLiteralTo(Char c,ostream * os)142 static CharFormat PrintAsCharLiteralTo(Char c, ostream* os) {
143 wchar_t w_c = static_cast<wchar_t>(c);
144 switch (w_c) {
145 case L'\0':
146 *os << "\\0";
147 break;
148 case L'\'':
149 *os << "\\'";
150 break;
151 case L'\\':
152 *os << "\\\\";
153 break;
154 case L'\a':
155 *os << "\\a";
156 break;
157 case L'\b':
158 *os << "\\b";
159 break;
160 case L'\f':
161 *os << "\\f";
162 break;
163 case L'\n':
164 *os << "\\n";
165 break;
166 case L'\r':
167 *os << "\\r";
168 break;
169 case L'\t':
170 *os << "\\t";
171 break;
172 case L'\v':
173 *os << "\\v";
174 break;
175 default:
176 if (IsPrintableAscii(w_c)) {
177 *os << static_cast<char>(c);
178 return kAsIs;
179 } else {
180 ostream::fmtflags flags = os->flags();
181 *os << "\\x" << std::hex << std::uppercase
182 << static_cast<int>(static_cast<UnsignedChar>(c));
183 os->flags(flags);
184 return kHexEscape;
185 }
186 }
187 return kSpecialEscape;
188 }
189
190 // Prints a wchar_t c as if it's part of a string literal, escaping it when
191 // necessary; returns how c was formatted.
PrintAsStringLiteralTo(wchar_t c,ostream * os)192 static CharFormat PrintAsStringLiteralTo(wchar_t c, ostream* os) {
193 switch (c) {
194 case L'\'':
195 *os << "'";
196 return kAsIs;
197 case L'"':
198 *os << "\\\"";
199 return kSpecialEscape;
200 default:
201 return PrintAsCharLiteralTo<wchar_t>(c, os);
202 }
203 }
204
205 // Prints a char c as if it's part of a string literal, escaping it when
206 // necessary; returns how c was formatted.
PrintAsStringLiteralTo(char c,ostream * os)207 static CharFormat PrintAsStringLiteralTo(char c, ostream* os) {
208 return PrintAsStringLiteralTo(
209 static_cast<wchar_t>(static_cast<unsigned char>(c)), os);
210 }
211
212 // Prints a wide or narrow character c and its code. '\0' is printed
213 // as "'\\0'", other unprintable characters are also properly escaped
214 // using the standard C++ escape sequence. The template argument
215 // UnsignedChar is the unsigned version of Char, which is the type of c.
216 template <typename UnsignedChar, typename Char>
PrintCharAndCodeTo(Char c,ostream * os)217 void PrintCharAndCodeTo(Char c, ostream* os) {
218 // First, print c as a literal in the most readable form we can find.
219 *os << ((sizeof(c) > 1) ? "L'" : "'");
220 const CharFormat format = PrintAsCharLiteralTo<UnsignedChar>(c, os);
221 *os << "'";
222
223 // To aid user debugging, we also print c's code in decimal, unless
224 // it's 0 (in which case c was printed as '\\0', making the code
225 // obvious).
226 if (c == 0)
227 return;
228 *os << " (" << static_cast<int>(c);
229
230 // For more convenience, we print c's code again in hexadecimal,
231 // unless c was already printed in the form '\x##' or the code is in
232 // [1, 9].
233 if (format == kHexEscape || (1 <= c && c <= 9)) {
234 // Do nothing.
235 } else {
236 *os << ", 0x" << String::FormatHexInt(static_cast<int>(c));
237 }
238 *os << ")";
239 }
240
PrintTo(unsigned char c,::std::ostream * os)241 void PrintTo(unsigned char c, ::std::ostream* os) {
242 PrintCharAndCodeTo<unsigned char>(c, os);
243 }
PrintTo(signed char c,::std::ostream * os)244 void PrintTo(signed char c, ::std::ostream* os) {
245 PrintCharAndCodeTo<unsigned char>(c, os);
246 }
247
248 // Prints a wchar_t as a symbol if it is printable or as its internal
249 // code otherwise and also as its code. L'\0' is printed as "L'\\0'".
PrintTo(wchar_t wc,ostream * os)250 void PrintTo(wchar_t wc, ostream* os) {
251 PrintCharAndCodeTo<wchar_t>(wc, os);
252 }
253
PrintTo(char32_t c,::std::ostream * os)254 void PrintTo(char32_t c, ::std::ostream* os) {
255 *os << std::hex << "U+" << std::uppercase << std::setfill('0') << std::setw(4)
256 << c;
257 }
258
259 // Prints the given array of characters to the ostream. CharType must be either
260 // char or wchar_t.
261 // The array starts at begin, the length is len, it may include '\0' characters
262 // and may not be NUL-terminated.
263 template <typename CharType>
264 GTEST_ATTRIBUTE_NO_SANITIZE_MEMORY_
265 GTEST_ATTRIBUTE_NO_SANITIZE_ADDRESS_
266 GTEST_ATTRIBUTE_NO_SANITIZE_HWADDRESS_
267 GTEST_ATTRIBUTE_NO_SANITIZE_THREAD_
PrintCharsAsStringTo(const CharType * begin,size_t len,ostream * os)268 static CharFormat PrintCharsAsStringTo(
269 const CharType* begin, size_t len, ostream* os) {
270 const char* const kQuoteBegin = sizeof(CharType) == 1 ? "\"" : "L\"";
271 *os << kQuoteBegin;
272 bool is_previous_hex = false;
273 CharFormat print_format = kAsIs;
274 for (size_t index = 0; index < len; ++index) {
275 const CharType cur = begin[index];
276 if (is_previous_hex && IsXDigit(cur)) {
277 // Previous character is of '\x..' form and this character can be
278 // interpreted as another hexadecimal digit in its number. Break string to
279 // disambiguate.
280 *os << "\" " << kQuoteBegin;
281 }
282 is_previous_hex = PrintAsStringLiteralTo(cur, os) == kHexEscape;
283 // Remember if any characters required hex escaping.
284 if (is_previous_hex) {
285 print_format = kHexEscape;
286 }
287 }
288 *os << "\"";
289 return print_format;
290 }
291
292 // Prints a (const) char/wchar_t array of 'len' elements, starting at address
293 // 'begin'. CharType must be either char or wchar_t.
294 template <typename CharType>
295 GTEST_ATTRIBUTE_NO_SANITIZE_MEMORY_
296 GTEST_ATTRIBUTE_NO_SANITIZE_ADDRESS_
297 GTEST_ATTRIBUTE_NO_SANITIZE_HWADDRESS_
298 GTEST_ATTRIBUTE_NO_SANITIZE_THREAD_
UniversalPrintCharArray(const CharType * begin,size_t len,ostream * os)299 static void UniversalPrintCharArray(
300 const CharType* begin, size_t len, ostream* os) {
301 // The code
302 // const char kFoo[] = "foo";
303 // generates an array of 4, not 3, elements, with the last one being '\0'.
304 //
305 // Therefore when printing a char array, we don't print the last element if
306 // it's '\0', such that the output matches the string literal as it's
307 // written in the source code.
308 if (len > 0 && begin[len - 1] == '\0') {
309 PrintCharsAsStringTo(begin, len - 1, os);
310 return;
311 }
312
313 // If, however, the last element in the array is not '\0', e.g.
314 // const char kFoo[] = { 'f', 'o', 'o' };
315 // we must print the entire array. We also print a message to indicate
316 // that the array is not NUL-terminated.
317 PrintCharsAsStringTo(begin, len, os);
318 *os << " (no terminating NUL)";
319 }
320
321 // Prints a (const) char array of 'len' elements, starting at address 'begin'.
UniversalPrintArray(const char * begin,size_t len,ostream * os)322 void UniversalPrintArray(const char* begin, size_t len, ostream* os) {
323 UniversalPrintCharArray(begin, len, os);
324 }
325
326 // Prints a (const) wchar_t array of 'len' elements, starting at address
327 // 'begin'.
UniversalPrintArray(const wchar_t * begin,size_t len,ostream * os)328 void UniversalPrintArray(const wchar_t* begin, size_t len, ostream* os) {
329 UniversalPrintCharArray(begin, len, os);
330 }
331
332 // Prints the given C string to the ostream.
PrintTo(const char * s,ostream * os)333 void PrintTo(const char* s, ostream* os) {
334 if (s == nullptr) {
335 *os << "NULL";
336 } else {
337 *os << ImplicitCast_<const void*>(s) << " pointing to ";
338 PrintCharsAsStringTo(s, strlen(s), os);
339 }
340 }
341
342 // MSVC compiler can be configured to define whar_t as a typedef
343 // of unsigned short. Defining an overload for const wchar_t* in that case
344 // would cause pointers to unsigned shorts be printed as wide strings,
345 // possibly accessing more memory than intended and causing invalid
346 // memory accesses. MSVC defines _NATIVE_WCHAR_T_DEFINED symbol when
347 // wchar_t is implemented as a native type.
348 #if !defined(_MSC_VER) || defined(_NATIVE_WCHAR_T_DEFINED)
349 // Prints the given wide C string to the ostream.
PrintTo(const wchar_t * s,ostream * os)350 void PrintTo(const wchar_t* s, ostream* os) {
351 if (s == nullptr) {
352 *os << "NULL";
353 } else {
354 *os << ImplicitCast_<const void*>(s) << " pointing to ";
355 PrintCharsAsStringTo(s, wcslen(s), os);
356 }
357 }
358 #endif // wchar_t is native
359
360 namespace {
361
ContainsUnprintableControlCodes(const char * str,size_t length)362 bool ContainsUnprintableControlCodes(const char* str, size_t length) {
363 const unsigned char *s = reinterpret_cast<const unsigned char *>(str);
364
365 for (size_t i = 0; i < length; i++) {
366 unsigned char ch = *s++;
367 if (std::iscntrl(ch)) {
368 switch (ch) {
369 case '\t':
370 case '\n':
371 case '\r':
372 break;
373 default:
374 return true;
375 }
376 }
377 }
378 return false;
379 }
380
IsUTF8TrailByte(unsigned char t)381 bool IsUTF8TrailByte(unsigned char t) { return 0x80 <= t && t<= 0xbf; }
382
IsValidUTF8(const char * str,size_t length)383 bool IsValidUTF8(const char* str, size_t length) {
384 const unsigned char *s = reinterpret_cast<const unsigned char *>(str);
385
386 for (size_t i = 0; i < length;) {
387 unsigned char lead = s[i++];
388
389 if (lead <= 0x7f) {
390 continue; // single-byte character (ASCII) 0..7F
391 }
392 if (lead < 0xc2) {
393 return false; // trail byte or non-shortest form
394 } else if (lead <= 0xdf && (i + 1) <= length && IsUTF8TrailByte(s[i])) {
395 ++i; // 2-byte character
396 } else if (0xe0 <= lead && lead <= 0xef && (i + 2) <= length &&
397 IsUTF8TrailByte(s[i]) &&
398 IsUTF8TrailByte(s[i + 1]) &&
399 // check for non-shortest form and surrogate
400 (lead != 0xe0 || s[i] >= 0xa0) &&
401 (lead != 0xed || s[i] < 0xa0)) {
402 i += 2; // 3-byte character
403 } else if (0xf0 <= lead && lead <= 0xf4 && (i + 3) <= length &&
404 IsUTF8TrailByte(s[i]) &&
405 IsUTF8TrailByte(s[i + 1]) &&
406 IsUTF8TrailByte(s[i + 2]) &&
407 // check for non-shortest form
408 (lead != 0xf0 || s[i] >= 0x90) &&
409 (lead != 0xf4 || s[i] < 0x90)) {
410 i += 3; // 4-byte character
411 } else {
412 return false;
413 }
414 }
415 return true;
416 }
417
ConditionalPrintAsText(const char * str,size_t length,ostream * os)418 void ConditionalPrintAsText(const char* str, size_t length, ostream* os) {
419 if (!ContainsUnprintableControlCodes(str, length) &&
420 IsValidUTF8(str, length)) {
421 *os << "\n As Text: \"" << str << "\"";
422 }
423 }
424
425 } // anonymous namespace
426
PrintStringTo(const::std::string & s,ostream * os)427 void PrintStringTo(const ::std::string& s, ostream* os) {
428 if (PrintCharsAsStringTo(s.data(), s.size(), os) == kHexEscape) {
429 if (GTEST_FLAG(print_utf8)) {
430 ConditionalPrintAsText(s.data(), s.size(), os);
431 }
432 }
433 }
434
435 #if GTEST_HAS_STD_WSTRING
PrintWideStringTo(const::std::wstring & s,ostream * os)436 void PrintWideStringTo(const ::std::wstring& s, ostream* os) {
437 PrintCharsAsStringTo(s.data(), s.size(), os);
438 }
439 #endif // GTEST_HAS_STD_WSTRING
440
441 } // namespace internal
442
443 } // namespace testing
444