Line data Source code
1 : /*
2 : * Copyright 2014 Google Inc. All rights reserved.
3 : *
4 : * Licensed under the Apache License, Version 2.0 (the "License");
5 : * you may not use this file except in compliance with the License.
6 : * You may obtain a copy of the License at
7 : *
8 : * http://www.apache.org/licenses/LICENSE-2.0
9 : *
10 : * Unless required by applicable law or agreed to in writing, software
11 : * distributed under the License is distributed on an "AS IS" BASIS,
12 : * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 : * See the License for the specific language governing permissions and
14 : * limitations under the License.
15 : */
16 :
17 : #ifndef FLATBUFFERS_UTIL_H_
18 : #define FLATBUFFERS_UTIL_H_
19 :
20 : #include <ctype.h>
21 : #include <errno.h>
22 :
23 : #include "flatbuffers/base.h"
24 : #include "flatbuffers/stl_emulation.h"
25 :
26 : #ifndef FLATBUFFERS_PREFER_PRINTF
27 : # include <iomanip>
28 : # include <sstream>
29 : #else // FLATBUFFERS_PREFER_PRINTF
30 : # include <float.h>
31 : # include <stdio.h>
32 : #endif // FLATBUFFERS_PREFER_PRINTF
33 :
34 : #include <string>
35 :
36 : namespace flatbuffers {
37 :
38 : // @locale-independent functions for ASCII characters set.
39 :
40 : // Fast checking that character lies in closed range: [a <= x <= b]
41 : // using one compare (conditional branch) operator.
42 0 : inline bool check_ascii_range(char x, char a, char b) {
43 0 : FLATBUFFERS_ASSERT(a <= b);
44 : // (Hacker's Delight): `a <= x <= b` <=> `(x-a) <={u} (b-a)`.
45 : // The x, a, b will be promoted to int and subtracted without overflow.
46 0 : return static_cast<unsigned int>(x - a) <= static_cast<unsigned int>(b - a);
47 : }
48 :
49 : // Case-insensitive isalpha
50 0 : inline bool is_alpha(char c) {
51 : // ASCII only: alpha to upper case => reset bit 0x20 (~0x20 = 0xDF).
52 0 : return check_ascii_range(c & 0xDF, 'a' & 0xDF, 'z' & 0xDF);
53 : }
54 :
55 : // Check for uppercase alpha
56 : inline bool is_alpha_upper(char c) { return check_ascii_range(c, 'A', 'Z'); }
57 :
58 : // Check (case-insensitive) that `c` is equal to alpha.
59 0 : inline bool is_alpha_char(char c, char alpha) {
60 0 : FLATBUFFERS_ASSERT(is_alpha(alpha));
61 : // ASCII only: alpha to upper case => reset bit 0x20 (~0x20 = 0xDF).
62 0 : return ((c & 0xDF) == (alpha & 0xDF));
63 : }
64 :
65 : // https://en.cppreference.com/w/cpp/string/byte/isxdigit
66 : // isdigit and isxdigit are the only standard narrow character classification
67 : // functions that are not affected by the currently installed C locale. although
68 : // some implementations (e.g. Microsoft in 1252 codepage) may classify
69 : // additional single-byte characters as digits.
70 0 : inline bool is_digit(char c) { return check_ascii_range(c, '0', '9'); }
71 :
72 : inline bool is_xdigit(char c) {
73 : // Replace by look-up table.
74 : return is_digit(c) || check_ascii_range(c & 0xDF, 'a' & 0xDF, 'f' & 0xDF);
75 : }
76 :
77 : // Case-insensitive isalnum
78 0 : inline bool is_alnum(char c) { return is_alpha(c) || is_digit(c); }
79 :
80 : inline char CharToUpper(char c) {
81 : return static_cast<char>(::toupper(static_cast<unsigned char>(c)));
82 : }
83 :
84 : inline char CharToLower(char c) {
85 : return static_cast<char>(::tolower(static_cast<unsigned char>(c)));
86 : }
87 :
88 : // @end-locale-independent functions for ASCII character set
89 :
90 : #ifdef FLATBUFFERS_PREFER_PRINTF
91 : template<typename T> size_t IntToDigitCount(T t) {
92 : size_t digit_count = 0;
93 : // Count the sign for negative numbers
94 : if (t < 0) digit_count++;
95 : // Count a single 0 left of the dot for fractional numbers
96 : if (-1 < t && t < 1) digit_count++;
97 : // Count digits until fractional part
98 : T eps = std::numeric_limits<T>::epsilon();
99 : while (t <= (-1 + eps) || (1 - eps) <= t) {
100 : t /= 10;
101 : digit_count++;
102 : }
103 : return digit_count;
104 : }
105 :
106 : template<typename T> size_t NumToStringWidth(T t, int precision = 0) {
107 : size_t string_width = IntToDigitCount(t);
108 : // Count the dot for floating point numbers
109 : if (precision) string_width += (precision + 1);
110 : return string_width;
111 : }
112 :
113 : template<typename T>
114 : std::string NumToStringImplWrapper(T t, const char *fmt, int precision = 0) {
115 : size_t string_width = NumToStringWidth(t, precision);
116 : std::string s(string_width, 0x00);
117 : // Allow snprintf to use std::string trailing null to detect buffer overflow
118 : snprintf(const_cast<char *>(s.data()), (s.size() + 1), fmt, string_width, t);
119 : return s;
120 : }
121 : #endif // FLATBUFFERS_PREFER_PRINTF
122 :
123 : // Convert an integer or floating point value to a string.
124 : // In contrast to std::stringstream, "char" values are
125 : // converted to a string of digits, and we don't use scientific notation.
126 8162 : template<typename T> std::string NumToString(T t) {
127 : // clang-format off
128 :
129 : #ifndef FLATBUFFERS_PREFER_PRINTF
130 16324 : std::stringstream ss;
131 8162 : ss << t;
132 16324 : return ss.str();
133 : #else // FLATBUFFERS_PREFER_PRINTF
134 : auto v = static_cast<long long>(t);
135 : return NumToStringImplWrapper(v, "%.*lld");
136 : #endif // FLATBUFFERS_PREFER_PRINTF
137 : // clang-format on
138 : }
139 : // Avoid char types used as character data.
140 : template<> inline std::string NumToString<signed char>(signed char t) {
141 : return NumToString(static_cast<int>(t));
142 : }
143 : template<> inline std::string NumToString<unsigned char>(unsigned char t) {
144 : return NumToString(static_cast<int>(t));
145 : }
146 : template<> inline std::string NumToString<char>(char t) {
147 : return NumToString(static_cast<int>(t));
148 : }
149 :
150 : // Special versions for floats/doubles.
151 0 : template<typename T> std::string FloatToString(T t, int precision) {
152 : // clang-format off
153 :
154 : #ifndef FLATBUFFERS_PREFER_PRINTF
155 : // to_string() prints different numbers of digits for floats depending on
156 : // platform and isn't available on Android, so we use stringstream
157 0 : std::stringstream ss;
158 : // Use std::fixed to suppress scientific notation.
159 0 : ss << std::fixed;
160 : // Default precision is 6, we want that to be higher for doubles.
161 0 : ss << std::setprecision(precision);
162 0 : ss << t;
163 0 : auto s = ss.str();
164 : #else // FLATBUFFERS_PREFER_PRINTF
165 : auto v = static_cast<double>(t);
166 : auto s = NumToStringImplWrapper(v, "%0.*f", precision);
167 : #endif // FLATBUFFERS_PREFER_PRINTF
168 : // clang-format on
169 : // Sadly, std::fixed turns "1" into "1.00000", so here we undo that.
170 0 : auto p = s.find_last_not_of('0');
171 0 : if (p != std::string::npos) {
172 : // Strip trailing zeroes. If it is a whole number, keep one zero.
173 0 : s.resize(p + (s[p] == '.' ? 2 : 1));
174 : }
175 0 : return s;
176 : }
177 :
178 0 : template<> inline std::string NumToString<double>(double t) {
179 0 : return FloatToString(t, 12);
180 : }
181 : template<> inline std::string NumToString<float>(float t) {
182 : return FloatToString(t, 6);
183 : }
184 :
185 : // Convert an integer value to a hexadecimal string.
186 : // The returned string length is always xdigits long, prefixed by 0 digits.
187 : // For example, IntToStringHex(0x23, 8) returns the string "00000023".
188 0 : inline std::string IntToStringHex(int i, int xdigits) {
189 0 : FLATBUFFERS_ASSERT(i >= 0);
190 : // clang-format off
191 :
192 : #ifndef FLATBUFFERS_PREFER_PRINTF
193 0 : std::stringstream ss;
194 0 : ss << std::setw(xdigits) << std::setfill('0') << std::hex << std::uppercase
195 0 : << i;
196 0 : return ss.str();
197 : #else // FLATBUFFERS_PREFER_PRINTF
198 : return NumToStringImplWrapper(i, "%.*X", xdigits);
199 : #endif // FLATBUFFERS_PREFER_PRINTF
200 : // clang-format on
201 : }
202 :
203 : // clang-format off
204 : // Use locale independent functions {strtod_l, strtof_l, strtoll_l, strtoull_l}.
205 : #if defined(FLATBUFFERS_LOCALE_INDEPENDENT) && (FLATBUFFERS_LOCALE_INDEPENDENT > 0)
206 : class ClassicLocale {
207 : #ifdef _MSC_VER
208 : typedef _locale_t locale_type;
209 : #else
210 : typedef locale_t locale_type; // POSIX.1-2008 locale_t type
211 : #endif
212 : ClassicLocale();
213 : ~ClassicLocale();
214 : locale_type locale_;
215 : static ClassicLocale instance_;
216 : public:
217 : static locale_type Get() { return instance_.locale_; }
218 : };
219 :
220 : #ifdef _MSC_VER
221 : #define __strtoull_impl(s, pe, b) _strtoui64_l(s, pe, b, ClassicLocale::Get())
222 : #define __strtoll_impl(s, pe, b) _strtoi64_l(s, pe, b, ClassicLocale::Get())
223 : #define __strtod_impl(s, pe) _strtod_l(s, pe, ClassicLocale::Get())
224 : #define __strtof_impl(s, pe) _strtof_l(s, pe, ClassicLocale::Get())
225 : #else
226 : #define __strtoull_impl(s, pe, b) strtoull_l(s, pe, b, ClassicLocale::Get())
227 : #define __strtoll_impl(s, pe, b) strtoll_l(s, pe, b, ClassicLocale::Get())
228 : #define __strtod_impl(s, pe) strtod_l(s, pe, ClassicLocale::Get())
229 : #define __strtof_impl(s, pe) strtof_l(s, pe, ClassicLocale::Get())
230 : #endif
231 : #else
232 : #define __strtod_impl(s, pe) strtod(s, pe)
233 : #define __strtof_impl(s, pe) static_cast<float>(strtod(s, pe))
234 : #ifdef _MSC_VER
235 : #define __strtoull_impl(s, pe, b) _strtoui64(s, pe, b)
236 : #define __strtoll_impl(s, pe, b) _strtoi64(s, pe, b)
237 : #else
238 : #define __strtoull_impl(s, pe, b) strtoull(s, pe, b)
239 : #define __strtoll_impl(s, pe, b) strtoll(s, pe, b)
240 : #endif
241 : #endif
242 :
243 0 : inline void strtoval_impl(int64_t *val, const char *str, char **endptr,
244 : int base) {
245 0 : *val = __strtoll_impl(str, endptr, base);
246 0 : }
247 :
248 0 : inline void strtoval_impl(uint64_t *val, const char *str, char **endptr,
249 : int base) {
250 0 : *val = __strtoull_impl(str, endptr, base);
251 0 : }
252 :
253 0 : inline void strtoval_impl(double *val, const char *str, char **endptr) {
254 0 : *val = __strtod_impl(str, endptr);
255 0 : }
256 :
257 : // UBSAN: double to float is safe if numeric_limits<float>::is_iec559 is true.
258 : __suppress_ubsan__("float-cast-overflow")
259 : inline void strtoval_impl(float *val, const char *str, char **endptr) {
260 : *val = __strtof_impl(str, endptr);
261 : }
262 : #undef __strtoull_impl
263 : #undef __strtoll_impl
264 : #undef __strtod_impl
265 : #undef __strtof_impl
266 : // clang-format on
267 :
268 : // Adaptor for strtoull()/strtoll().
269 : // Flatbuffers accepts numbers with any count of leading zeros (-009 is -9),
270 : // while strtoll with base=0 interprets first leading zero as octal prefix.
271 : // In future, it is possible to add prefixed 0b0101.
272 : // 1) Checks errno code for overflow condition (out of range).
273 : // 2) If base <= 0, function try to detect base of number by prefix.
274 : //
275 : // Return value (like strtoull and strtoll, but reject partial result):
276 : // - If successful, an integer value corresponding to the str is returned.
277 : // - If full string conversion can't be performed, 0 is returned.
278 : // - If the converted value falls out of range of corresponding return type, a
279 : // range error occurs. In this case value MAX(T)/MIN(T) is returned.
280 : template<typename T>
281 0 : inline bool StringToIntegerImpl(T *val, const char *const str,
282 : const int base = 0,
283 : const bool check_errno = true) {
284 : // T is int64_t or uint64_T
285 0 : FLATBUFFERS_ASSERT(str);
286 0 : if (base <= 0) {
287 0 : auto s = str;
288 0 : while (*s && !is_digit(*s)) s++;
289 0 : if (s[0] == '0' && is_alpha_char(s[1], 'X'))
290 0 : return StringToIntegerImpl(val, str, 16, check_errno);
291 : // if a prefix not match, try base=10
292 0 : return StringToIntegerImpl(val, str, 10, check_errno);
293 : } else {
294 0 : if (check_errno) errno = 0; // clear thread-local errno
295 0 : auto endptr = str;
296 0 : strtoval_impl(val, str, const_cast<char **>(&endptr), base);
297 0 : if ((*endptr != '\0') || (endptr == str)) {
298 0 : *val = 0; // erase partial result
299 0 : return false; // invalid string
300 : }
301 : // errno is out-of-range, return MAX/MIN
302 0 : if (check_errno && errno) return false;
303 0 : return true;
304 : }
305 : }
306 :
307 : template<typename T>
308 0 : inline bool StringToFloatImpl(T *val, const char *const str) {
309 : // Type T must be either float or double.
310 0 : FLATBUFFERS_ASSERT(str && val);
311 0 : auto end = str;
312 0 : strtoval_impl(val, str, const_cast<char **>(&end));
313 0 : auto done = (end != str) && (*end == '\0');
314 0 : if (!done) *val = 0; // erase partial result
315 0 : return done;
316 : }
317 :
318 : // Convert a string to an instance of T.
319 : // Return value (matched with StringToInteger64Impl and strtod):
320 : // - If successful, a numeric value corresponding to the str is returned.
321 : // - If full string conversion can't be performed, 0 is returned.
322 : // - If the converted value falls out of range of corresponding return type, a
323 : // range error occurs. In this case value MAX(T)/MIN(T) is returned.
324 : template<typename T> inline bool StringToNumber(const char *s, T *val) {
325 : // Assert on `unsigned long` and `signed long` on LP64.
326 : // If it is necessary, it could be solved with flatbuffers::enable_if<B,T>.
327 : static_assert(sizeof(T) < sizeof(int64_t), "unexpected type T");
328 : FLATBUFFERS_ASSERT(s && val);
329 : int64_t i64;
330 : // The errno check isn't needed, will return MAX/MIN on overflow.
331 : if (StringToIntegerImpl(&i64, s, 0, false)) {
332 : const int64_t max = (flatbuffers::numeric_limits<T>::max)();
333 : const int64_t min = flatbuffers::numeric_limits<T>::lowest();
334 : if (i64 > max) {
335 : *val = static_cast<T>(max);
336 : return false;
337 : }
338 : if (i64 < min) {
339 : // For unsigned types return max to distinguish from
340 : // "no conversion can be performed" when 0 is returned.
341 : *val = static_cast<T>(flatbuffers::is_unsigned<T>::value ? max : min);
342 : return false;
343 : }
344 : *val = static_cast<T>(i64);
345 : return true;
346 : }
347 : *val = 0;
348 : return false;
349 : }
350 :
351 : template<> inline bool StringToNumber<int64_t>(const char *str, int64_t *val) {
352 : return StringToIntegerImpl(val, str);
353 : }
354 :
355 : template<>
356 : inline bool StringToNumber<uint64_t>(const char *str, uint64_t *val) {
357 : if (!StringToIntegerImpl(val, str)) return false;
358 : // The strtoull accepts negative numbers:
359 : // If the minus sign was part of the input sequence, the numeric value
360 : // calculated from the sequence of digits is negated as if by unary minus
361 : // in the result type, which applies unsigned integer wraparound rules.
362 : // Fix this behaviour (except -0).
363 : if (*val) {
364 : auto s = str;
365 : while (*s && !is_digit(*s)) s++;
366 : s = (s > str) ? (s - 1) : s; // step back to one symbol
367 : if (*s == '-') {
368 : // For unsigned types return the max to distinguish from
369 : // "no conversion can be performed".
370 : *val = (flatbuffers::numeric_limits<uint64_t>::max)();
371 : return false;
372 : }
373 : }
374 : return true;
375 : }
376 :
377 : template<> inline bool StringToNumber(const char *s, float *val) {
378 : return StringToFloatImpl(val, s);
379 : }
380 :
381 0 : template<> inline bool StringToNumber(const char *s, double *val) {
382 0 : return StringToFloatImpl(val, s);
383 : }
384 :
385 0 : inline int64_t StringToInt(const char *s, int base = 10) {
386 : int64_t val;
387 0 : return StringToIntegerImpl(&val, s, base) ? val : 0;
388 : }
389 :
390 0 : inline uint64_t StringToUInt(const char *s, int base = 10) {
391 : uint64_t val;
392 0 : return StringToIntegerImpl(&val, s, base) ? val : 0;
393 : }
394 :
395 : typedef bool (*LoadFileFunction)(const char *filename, bool binary,
396 : std::string *dest);
397 : typedef bool (*FileExistsFunction)(const char *filename);
398 :
399 : LoadFileFunction SetLoadFileFunction(LoadFileFunction load_file_function);
400 :
401 : FileExistsFunction SetFileExistsFunction(
402 : FileExistsFunction file_exists_function);
403 :
404 : // Check if file "name" exists.
405 : bool FileExists(const char *name);
406 :
407 : // Check if "name" exists and it is also a directory.
408 : bool DirExists(const char *name);
409 :
410 : // Load file "name" into "buf" returning true if successful
411 : // false otherwise. If "binary" is false data is read
412 : // using ifstream's text mode, otherwise data is read with
413 : // no transcoding.
414 : bool LoadFile(const char *name, bool binary, std::string *buf);
415 :
416 : // Save data "buf" of length "len" bytes into a file
417 : // "name" returning true if successful, false otherwise.
418 : // If "binary" is false data is written using ifstream's
419 : // text mode, otherwise data is written with no
420 : // transcoding.
421 : bool SaveFile(const char *name, const char *buf, size_t len, bool binary);
422 :
423 : // Save data "buf" into file "name" returning true if
424 : // successful, false otherwise. If "binary" is false
425 : // data is written using ifstream's text mode, otherwise
426 : // data is written with no transcoding.
427 : inline bool SaveFile(const char *name, const std::string &buf, bool binary) {
428 : return SaveFile(name, buf.c_str(), buf.size(), binary);
429 : }
430 :
431 : // Functionality for minimalistic portable path handling.
432 :
433 : // The functions below behave correctly regardless of whether posix ('/') or
434 : // Windows ('/' or '\\') separators are used.
435 :
436 : // Any new separators inserted are always posix.
437 : FLATBUFFERS_CONSTEXPR char kPathSeparator = '/';
438 :
439 : // Returns the path with the extension, if any, removed.
440 : std::string StripExtension(const std::string &filepath);
441 :
442 : // Returns the extension, if any.
443 : std::string GetExtension(const std::string &filepath);
444 :
445 : // Return the last component of the path, after the last separator.
446 : std::string StripPath(const std::string &filepath);
447 :
448 : // Strip the last component of the path + separator.
449 : std::string StripFileName(const std::string &filepath);
450 :
451 : std::string StripPrefix(const std::string &filepath,
452 : const std::string &prefix_to_remove);
453 :
454 : // Concatenates a path with a filename, regardless of whether the path
455 : // ends in a separator or not.
456 : std::string ConCatPathFileName(const std::string &path,
457 : const std::string &filename);
458 :
459 : // Replaces any '\\' separators with '/'
460 : std::string PosixPath(const char *path);
461 : std::string PosixPath(const std::string &path);
462 :
463 : // This function ensure a directory exists, by recursively
464 : // creating dirs for any parts of the path that don't exist yet.
465 : void EnsureDirExists(const std::string &filepath);
466 :
467 : // Obtains the absolute path from any other path.
468 : // Returns the input path if the absolute path couldn't be resolved.
469 : std::string AbsolutePath(const std::string &filepath);
470 :
471 : // Returns files relative to the --project_root path, prefixed with `//`.
472 : std::string RelativeToRootPath(const std::string &project,
473 : const std::string &filepath);
474 :
475 : // To and from UTF-8 unicode conversion functions
476 :
477 : // Convert a unicode code point into a UTF-8 representation by appending it
478 : // to a string. Returns the number of bytes generated.
479 : inline int ToUTF8(uint32_t ucc, std::string *out) {
480 : FLATBUFFERS_ASSERT(!(ucc & 0x80000000)); // Top bit can't be set.
481 : // 6 possible encodings: http://en.wikipedia.org/wiki/UTF-8
482 : for (int i = 0; i < 6; i++) {
483 : // Max bits this encoding can represent.
484 : uint32_t max_bits = 6 + i * 5 + static_cast<int>(!i);
485 : if (ucc < (1u << max_bits)) { // does it fit?
486 : // Remaining bits not encoded in the first byte, store 6 bits each
487 : uint32_t remain_bits = i * 6;
488 : // Store first byte:
489 : (*out) += static_cast<char>((0xFE << (max_bits - remain_bits)) |
490 : (ucc >> remain_bits));
491 : // Store remaining bytes:
492 : for (int j = i - 1; j >= 0; j--) {
493 : (*out) += static_cast<char>(((ucc >> (j * 6)) & 0x3F) | 0x80);
494 : }
495 : return i + 1; // Return the number of bytes added.
496 : }
497 : }
498 : FLATBUFFERS_ASSERT(0); // Impossible to arrive here.
499 : return -1;
500 : }
501 :
502 : // Converts whatever prefix of the incoming string corresponds to a valid
503 : // UTF-8 sequence into a unicode code. The incoming pointer will have been
504 : // advanced past all bytes parsed.
505 : // returns -1 upon corrupt UTF-8 encoding (ignore the incoming pointer in
506 : // this case).
507 0 : inline int FromUTF8(const char **in) {
508 0 : int len = 0;
509 : // Count leading 1 bits.
510 0 : for (int mask = 0x80; mask >= 0x04; mask >>= 1) {
511 0 : if (**in & mask) {
512 0 : len++;
513 : } else {
514 0 : break;
515 : }
516 : }
517 0 : if ((static_cast<unsigned char>(**in) << len) & 0x80)
518 0 : return -1; // Bit after leading 1's must be 0.
519 0 : if (!len) return *(*in)++;
520 : // UTF-8 encoded values with a length are between 2 and 4 bytes.
521 0 : if (len < 2 || len > 4) { return -1; }
522 : // Grab initial bits of the code.
523 0 : int ucc = *(*in)++ & ((1 << (7 - len)) - 1);
524 0 : for (int i = 0; i < len - 1; i++) {
525 0 : if ((**in & 0xC0) != 0x80) return -1; // Upper bits must 1 0.
526 0 : ucc <<= 6;
527 0 : ucc |= *(*in)++ & 0x3F; // Grab 6 more bits of the code.
528 : }
529 : // UTF-8 cannot encode values between 0xD800 and 0xDFFF (reserved for
530 : // UTF-16 surrogate pairs).
531 0 : if (ucc >= 0xD800 && ucc <= 0xDFFF) { return -1; }
532 : // UTF-8 must represent code points in their shortest possible encoding.
533 0 : switch (len) {
534 0 : case 2:
535 : // Two bytes of UTF-8 can represent code points from U+0080 to U+07FF.
536 0 : if (ucc < 0x0080 || ucc > 0x07FF) { return -1; }
537 0 : break;
538 0 : case 3:
539 : // Three bytes of UTF-8 can represent code points from U+0800 to U+FFFF.
540 0 : if (ucc < 0x0800 || ucc > 0xFFFF) { return -1; }
541 0 : break;
542 0 : case 4:
543 : // Four bytes of UTF-8 can represent code points from U+10000 to U+10FFFF.
544 0 : if (ucc < 0x10000 || ucc > 0x10FFFF) { return -1; }
545 0 : break;
546 : }
547 0 : return ucc;
548 : }
549 :
550 : #ifndef FLATBUFFERS_PREFER_PRINTF
551 : // Wraps a string to a maximum length, inserting new lines where necessary. Any
552 : // existing whitespace will be collapsed down to a single space. A prefix or
553 : // suffix can be provided, which will be inserted before or after a wrapped
554 : // line, respectively.
555 : inline std::string WordWrap(const std::string in, size_t max_length,
556 : const std::string wrapped_line_prefix,
557 : const std::string wrapped_line_suffix) {
558 : std::istringstream in_stream(in);
559 : std::string wrapped, line, word;
560 :
561 : in_stream >> word;
562 : line = word;
563 :
564 : while (in_stream >> word) {
565 : if ((line.length() + 1 + word.length() + wrapped_line_suffix.length()) <
566 : max_length) {
567 : line += " " + word;
568 : } else {
569 : wrapped += line + wrapped_line_suffix + "\n";
570 : line = wrapped_line_prefix + word;
571 : }
572 : }
573 : wrapped += line;
574 :
575 : return wrapped;
576 : }
577 : #endif // !FLATBUFFERS_PREFER_PRINTF
578 :
579 47221 : inline bool EscapeString(const char *s, size_t length, std::string *_text,
580 : bool allow_non_utf8, bool natural_utf8) {
581 47221 : std::string &text = *_text;
582 47221 : text += "\"";
583 803235 : for (uoffset_t i = 0; i < length; i++) {
584 756014 : char c = s[i];
585 756014 : switch (c) {
586 0 : case '\n': text += "\\n"; break;
587 0 : case '\t': text += "\\t"; break;
588 0 : case '\r': text += "\\r"; break;
589 0 : case '\b': text += "\\b"; break;
590 0 : case '\f': text += "\\f"; break;
591 0 : case '\"': text += "\\\""; break;
592 0 : case '\\': text += "\\\\"; break;
593 756014 : default:
594 756014 : if (c >= ' ' && c <= '~') {
595 756014 : text += c;
596 : } else {
597 : // Not printable ASCII data. Let's see if it's valid UTF-8 first:
598 0 : const char *utf8 = s + i;
599 0 : int ucc = FromUTF8(&utf8);
600 0 : if (ucc < 0) {
601 0 : if (allow_non_utf8) {
602 0 : text += "\\x";
603 0 : text += IntToStringHex(static_cast<uint8_t>(c), 2);
604 : } else {
605 : // There are two cases here:
606 : //
607 : // 1) We reached here by parsing an IDL file. In that case,
608 : // we previously checked for non-UTF-8, so we shouldn't reach
609 : // here.
610 : //
611 : // 2) We reached here by someone calling GenerateText()
612 : // on a previously-serialized flatbuffer. The data might have
613 : // non-UTF-8 Strings, or might be corrupt.
614 : //
615 : // In both cases, we have to give up and inform the caller
616 : // they have no JSON.
617 0 : return false;
618 : }
619 : } else {
620 0 : if (natural_utf8) {
621 : // utf8 points to past all utf-8 bytes parsed
622 0 : text.append(s + i, static_cast<size_t>(utf8 - s - i));
623 0 : } else if (ucc <= 0xFFFF) {
624 : // Parses as Unicode within JSON's \uXXXX range, so use that.
625 0 : text += "\\u";
626 0 : text += IntToStringHex(ucc, 4);
627 0 : } else if (ucc <= 0x10FFFF) {
628 : // Encode Unicode SMP values to a surrogate pair using two \u
629 : // escapes.
630 0 : uint32_t base = ucc - 0x10000;
631 0 : auto high_surrogate = (base >> 10) + 0xD800;
632 0 : auto low_surrogate = (base & 0x03FF) + 0xDC00;
633 0 : text += "\\u";
634 0 : text += IntToStringHex(high_surrogate, 4);
635 0 : text += "\\u";
636 0 : text += IntToStringHex(low_surrogate, 4);
637 : }
638 : // Skip past characters recognized.
639 0 : i = static_cast<uoffset_t>(utf8 - s - 1);
640 : }
641 : }
642 756014 : break;
643 : }
644 : }
645 47221 : text += "\"";
646 47221 : return true;
647 : }
648 :
649 : inline std::string BufferToHexText(const void *buffer, size_t buffer_size,
650 : size_t max_length,
651 : const std::string &wrapped_line_prefix,
652 : const std::string &wrapped_line_suffix) {
653 : std::string text = wrapped_line_prefix;
654 : size_t start_offset = 0;
655 : const char *s = reinterpret_cast<const char *>(buffer);
656 : for (size_t i = 0; s && i < buffer_size; i++) {
657 : // Last iteration or do we have more?
658 : bool have_more = i + 1 < buffer_size;
659 : text += "0x";
660 : text += IntToStringHex(static_cast<uint8_t>(s[i]), 2);
661 : if (have_more) { text += ','; }
662 : // If we have more to process and we reached max_length
663 : if (have_more &&
664 : text.size() + wrapped_line_suffix.size() >= start_offset + max_length) {
665 : text += wrapped_line_suffix;
666 : text += '\n';
667 : start_offset = text.size();
668 : text += wrapped_line_prefix;
669 : }
670 : }
671 : text += wrapped_line_suffix;
672 : return text;
673 : }
674 :
675 : // Remove paired quotes in a string: "text"|'text' -> text.
676 : std::string RemoveStringQuotes(const std::string &s);
677 :
678 : // Change th global C-locale to locale with name <locale_name>.
679 : // Returns an actual locale name in <_value>, useful if locale_name is "" or
680 : // null.
681 : bool SetGlobalTestLocale(const char *locale_name,
682 : std::string *_value = nullptr);
683 :
684 : // Read (or test) a value of environment variable.
685 : bool ReadEnvironmentVariable(const char *var_name,
686 : std::string *_value = nullptr);
687 :
688 : // MSVC specific: Send all assert reports to STDOUT to prevent CI hangs.
689 : void SetupDefaultCRTReportMode();
690 :
691 : enum class Case {
692 : kUnknown = 0,
693 : // TheQuickBrownFox
694 : kUpperCamel = 1,
695 : // theQuickBrownFox
696 : kLowerCamel = 2,
697 : // the_quick_brown_fox
698 : kSnake = 3,
699 : // THE_QUICK_BROWN_FOX
700 : kScreamingSnake = 4,
701 : // THEQUICKBROWNFOX
702 : kAllUpper = 5,
703 : // thequickbrownfox
704 : kAllLower = 6,
705 : // the-quick-brown-fox
706 : kDasher = 7,
707 : // THEQuiCKBr_ownFox (or whatever you want, we won't change it)
708 : kKeep = 8,
709 : // the_quick_brown_fox123 (as opposed to the_quick_brown_fox_123)
710 : kSnake2 = 9,
711 : };
712 :
713 : // Convert the `input` string of case `input_case` to the specified `output_case`.
714 : std::string ConvertCase(const std::string &input, Case output_case,
715 : Case input_case = Case::kSnake);
716 :
717 : } // namespace flatbuffers
718 :
719 : #endif // FLATBUFFERS_UTIL_H_
|
