| 1 | // |
| 2 | // NumericString.h |
| 3 | // |
| 4 | // Library: Foundation |
| 5 | // Package: Core |
| 6 | // Module: NumericString |
| 7 | // |
| 8 | // Numeric string utility functions. |
| 9 | // |
| 10 | // Copyright (c) 2004-2006, Applied Informatics Software Engineering GmbH. |
| 11 | // and Contributors. |
| 12 | // |
| 13 | // SPDX-License-Identifier: BSL-1.0 |
| 14 | // |
| 15 | |
| 16 | |
| 17 | #ifndef Foundation_NumericString_INCLUDED |
| 18 | #define Foundation_NumericString_INCLUDED |
| 19 | |
| 20 | |
| 21 | #include "Poco/Foundation.h" |
| 22 | #include "Poco/Buffer.h" |
| 23 | #include "Poco/FPEnvironment.h" |
| 24 | #ifdef min |
| 25 | #undef min |
| 26 | #endif |
| 27 | #ifdef max |
| 28 | #undef max |
| 29 | #endif |
| 30 | #include <limits> |
| 31 | #include <cmath> |
| 32 | #include <cctype> |
| 33 | #if !defined(POCO_NO_LOCALE) |
| 34 | #include <locale> |
| 35 | #endif |
| 36 | |
| 37 | |
| 38 | // binary numbers are supported, thus 64 (bits) + 1 (string terminating zero) |
| 39 | #define POCO_MAX_INT_STRING_LEN 65 |
| 40 | // value from strtod.cc (double_conversion::kMaxSignificantDecimalDigits) |
| 41 | #define POCO_MAX_FLT_STRING_LEN 780 |
| 42 | |
| 43 | #define POCO_FLT_INF "inf" |
| 44 | #define POCO_FLT_NAN "nan" |
| 45 | #define POCO_FLT_EXP 'e' |
| 46 | |
| 47 | |
| 48 | namespace Poco { |
| 49 | |
| 50 | |
| 51 | inline char decimalSeparator() |
| 52 | /// Returns decimal separator from global locale or |
| 53 | /// default '.' for platforms where locale is unavailable. |
| 54 | { |
| 55 | #if !defined(POCO_NO_LOCALE) |
| 56 | return std::use_facet<std::numpunct<char> >(std::locale()).decimal_point(); |
| 57 | #else |
| 58 | return '.'; |
| 59 | #endif |
| 60 | } |
| 61 | |
| 62 | |
| 63 | inline char thousandSeparator() |
| 64 | /// Returns thousand separator from global locale or |
| 65 | /// default ',' for platforms where locale is unavailable. |
| 66 | { |
| 67 | #if !defined(POCO_NO_LOCALE) |
| 68 | return std::use_facet<std::numpunct<char> >(std::locale()).thousands_sep(); |
| 69 | #else |
| 70 | return ','; |
| 71 | #endif |
| 72 | } |
| 73 | |
| 74 | |
| 75 | // |
| 76 | // String to Number Conversions |
| 77 | // |
| 78 | |
| 79 | template <typename I> |
| 80 | bool strToInt(const char* pStr, I& result, short base, char thSep = ',') |
| 81 | /// Converts zero-terminated character array to integer number; |
| 82 | /// Thousand separators are recognized for base10 and current locale; |
| 83 | /// it is silently skipped but not verified for correct positioning. |
| 84 | /// Function returns true if successful. If parsing was unsuccessful, |
| 85 | /// the return value is false with the result value undetermined. |
| 86 | { |
| 87 | if (!pStr) return false; |
| 88 | while (std::isspace(*pStr)) ++pStr; |
| 89 | if (*pStr == '\0') return false; |
| 90 | short sign = 1; |
| 91 | if ((base == 10) && (*pStr == '-')) |
| 92 | { |
| 93 | // Unsigned types can't be negative so abort parsing |
| 94 | if (std::numeric_limits<I>::min() >= 0) return false; |
| 95 | sign = -1; |
| 96 | ++pStr; |
| 97 | } |
| 98 | else if (*pStr == '+') ++pStr; |
| 99 | |
| 100 | // parser states: |
| 101 | const char STATE_SIGNIFICANT_DIGITS = 1; |
| 102 | char state = 0; |
| 103 | |
| 104 | result = 0; |
| 105 | I limitCheck = std::numeric_limits<I>::max() / base; |
| 106 | for (; *pStr != '\0'; ++pStr) |
| 107 | { |
| 108 | switch (*pStr) |
| 109 | { |
| 110 | case '0': |
| 111 | if (state < STATE_SIGNIFICANT_DIGITS) break; |
| 112 | |
| 113 | case '1': case '2': case '3': case '4': |
| 114 | case '5': case '6': case '7': |
| 115 | if (state < STATE_SIGNIFICANT_DIGITS) state = STATE_SIGNIFICANT_DIGITS; |
| 116 | if (result > limitCheck) return false; |
| 117 | result = result * base + (*pStr - '0'); |
| 118 | |
| 119 | break; |
| 120 | |
| 121 | case '8': case '9': |
| 122 | if ((base == 10) || (base == 0x10)) |
| 123 | { |
| 124 | if (state < STATE_SIGNIFICANT_DIGITS) state = STATE_SIGNIFICANT_DIGITS; |
| 125 | if (result > limitCheck) return false; |
| 126 | result = result * base + (*pStr - '0'); |
| 127 | } |
| 128 | else return false; |
| 129 | |
| 130 | break; |
| 131 | |
| 132 | case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': |
| 133 | if (base != 0x10) return false; |
| 134 | if (state < STATE_SIGNIFICANT_DIGITS) state = STATE_SIGNIFICANT_DIGITS; |
| 135 | if (result > limitCheck) return false; |
| 136 | result = result * base + (10 + *pStr - 'a'); |
| 137 | |
| 138 | break; |
| 139 | |
| 140 | case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': |
| 141 | if (base != 0x10) return false; |
| 142 | if (state < STATE_SIGNIFICANT_DIGITS) state = STATE_SIGNIFICANT_DIGITS; |
| 143 | if (result > limitCheck) return false; |
| 144 | result = result * base + (10 + *pStr - 'A'); |
| 145 | |
| 146 | break; |
| 147 | |
| 148 | case '.': |
| 149 | if ((base == 10) && (thSep == '.')) break; |
| 150 | else return false; |
| 151 | |
| 152 | case ',': |
| 153 | if ((base == 10) && (thSep == ',')) break; |
| 154 | else return false; |
| 155 | |
| 156 | case ' ': |
| 157 | if ((base == 10) && (thSep == ' ')) break; |
| 158 | // fallthrough |
| 159 | |
| 160 | default: |
| 161 | return false; |
| 162 | } |
| 163 | } |
| 164 | |
| 165 | if ((sign < 0) && (base == 10)) result *= sign; |
| 166 | |
| 167 | return true; |
| 168 | } |
| 169 | |
| 170 | |
| 171 | template <typename I> |
| 172 | bool strToInt(const std::string& str, I& result, short base, char thSep = ',') |
| 173 | /// Converts string to integer number; |
| 174 | /// This is a wrapper function, for details see see the |
| 175 | /// bool strToInt(const char*, I&, short, char) implementation. |
| 176 | { |
| 177 | return strToInt(str.c_str(), result, base, thSep); |
| 178 | } |
| 179 | |
| 180 | |
| 181 | // |
| 182 | // Number to String Conversions |
| 183 | // |
| 184 | |
| 185 | namespace Impl { |
| 186 | |
| 187 | class Ptr |
| 188 | /// Utility char pointer wrapper class. |
| 189 | /// Class ensures increment/decrement remain within boundaries. |
| 190 | { |
| 191 | public: |
| 192 | Ptr(char* ptr, std::size_t offset): _beg(ptr), _cur(ptr), _end(ptr + offset) |
| 193 | { |
| 194 | } |
| 195 | |
| 196 | char*& operator ++ () // prefix |
| 197 | { |
| 198 | check(_cur + 1); |
| 199 | return ++_cur; |
| 200 | } |
| 201 | |
| 202 | char* operator ++ (int) // postfix |
| 203 | { |
| 204 | check(_cur + 1); |
| 205 | char* tmp = _cur++; |
| 206 | return tmp; |
| 207 | } |
| 208 | |
| 209 | char*& operator -- () // prefix |
| 210 | { |
| 211 | check(_cur - 1); |
| 212 | return --_cur; |
| 213 | } |
| 214 | |
| 215 | char* operator -- (int) // postfix |
| 216 | { |
| 217 | check(_cur - 1); |
| 218 | char* tmp = _cur--; |
| 219 | return tmp; |
| 220 | } |
| 221 | |
| 222 | char*& operator += (int incr) |
| 223 | { |
| 224 | check(_cur + incr); |
| 225 | return _cur += incr; |
| 226 | } |
| 227 | |
| 228 | char*& operator -= (int decr) |
| 229 | { |
| 230 | check(_cur - decr); |
| 231 | return _cur -= decr; |
| 232 | } |
| 233 | |
| 234 | operator char* () const |
| 235 | { |
| 236 | return _cur; |
| 237 | } |
| 238 | |
| 239 | std::size_t span() const |
| 240 | { |
| 241 | return _end - _beg; |
| 242 | } |
| 243 | |
| 244 | private: |
| 245 | void check(char* ptr) |
| 246 | { |
| 247 | if (ptr > _end) throw RangeException(); |
| 248 | } |
| 249 | |
| 250 | const char* _beg; |
| 251 | char* _cur; |
| 252 | const char* _end; |
| 253 | }; |
| 254 | |
| 255 | } // namespace Impl |
| 256 | |
| 257 | |
| 258 | template <typename T> |
| 259 | bool intToStr(T value, |
| 260 | unsigned short base, |
| 261 | char* result, |
| 262 | std::size_t& size, |
| 263 | bool prefix = false, |
| 264 | int width = -1, |
| 265 | char fill = ' ', |
| 266 | char thSep = 0) |
| 267 | /// Converts integer to string. Numeric bases from binary to hexadecimal are supported. |
| 268 | /// If width is non-zero, it pads the return value with fill character to the specified width. |
| 269 | /// When padding is zero character ('0'), it is prepended to the number itself; all other |
| 270 | /// paddings are prepended to the formatted result with minus sign or base prefix included |
| 271 | /// If prefix is true and base is octal or hexadecimal, respective prefix ('0' for octal, |
| 272 | /// "0x" for hexadecimal) is prepended. For all other bases, prefix argument is ignored. |
| 273 | /// Formatted string has at least [width] total length. |
| 274 | { |
| 275 | if (base < 2 || base > 0x10) |
| 276 | { |
| 277 | *result = '\0'; |
| 278 | return false; |
| 279 | } |
| 280 | |
| 281 | Impl::Ptr ptr(result, size); |
| 282 | int thCount = 0; |
| 283 | T tmpVal; |
| 284 | do |
| 285 | { |
| 286 | tmpVal = value; |
| 287 | value /= base; |
| 288 | *ptr++ = "FEDCBA9876543210123456789ABCDEF" [15 + (tmpVal - value * base)]; |
| 289 | if (thSep && (base == 10) && (++thCount == 3)) |
| 290 | { |
| 291 | *ptr++ = thSep; |
| 292 | thCount = 0; |
| 293 | } |
| 294 | } while (value); |
| 295 | |
| 296 | if ('0' == fill) |
| 297 | { |
| 298 | if (tmpVal < 0) --width; |
| 299 | if (prefix && base == 010) --width; |
| 300 | if (prefix && base == 0x10) width -= 2; |
| 301 | while ((ptr - result) < width) *ptr++ = fill; |
| 302 | } |
| 303 | |
| 304 | if (prefix && base == 010) *ptr++ = '0'; |
| 305 | else if (prefix && base == 0x10) |
| 306 | { |
| 307 | *ptr++ = 'x'; |
| 308 | *ptr++ = '0'; |
| 309 | } |
| 310 | |
| 311 | if (tmpVal < 0) *ptr++ = '-'; |
| 312 | |
| 313 | if ('0' != fill) |
| 314 | { |
| 315 | while ((ptr - result) < width) *ptr++ = fill; |
| 316 | } |
| 317 | |
| 318 | size = ptr - result; |
| 319 | poco_assert_dbg (size <= ptr.span()); |
| 320 | poco_assert_dbg ((-1 == width) || (size >= size_t(width))); |
| 321 | *ptr-- = '\0'; |
| 322 | |
| 323 | char* ptrr = result; |
| 324 | char tmp; |
| 325 | while(ptrr < ptr) |
| 326 | { |
| 327 | tmp = *ptr; |
| 328 | *ptr-- = *ptrr; |
| 329 | *ptrr++ = tmp; |
| 330 | } |
| 331 | |
| 332 | return true; |
| 333 | } |
| 334 | |
| 335 | |
| 336 | template <typename T> |
| 337 | bool uIntToStr(T value, |
| 338 | unsigned short base, |
| 339 | char* result, |
| 340 | std::size_t& size, |
| 341 | bool prefix = false, |
| 342 | int width = -1, |
| 343 | char fill = ' ', |
| 344 | char thSep = 0) |
| 345 | /// Converts unsigned integer to string. Numeric bases from binary to hexadecimal are supported. |
| 346 | /// If width is non-zero, it pads the return value with fill character to the specified width. |
| 347 | /// When padding is zero character ('0'), it is prepended to the number itself; all other |
| 348 | /// paddings are prepended to the formatted result with minus sign or base prefix included |
| 349 | /// If prefix is true and base is octal or hexadecimal, respective prefix ('0' for octal, |
| 350 | /// "0x" for hexadecimal) is prepended. For all other bases, prefix argument is ignored. |
| 351 | /// Formatted string has at least [width] total length. |
| 352 | { |
| 353 | if (base < 2 || base > 0x10) |
| 354 | { |
| 355 | *result = '\0'; |
| 356 | return false; |
| 357 | } |
| 358 | |
| 359 | Impl::Ptr ptr(result, size); |
| 360 | int thCount = 0; |
| 361 | T tmpVal; |
| 362 | do |
| 363 | { |
| 364 | tmpVal = value; |
| 365 | value /= base; |
| 366 | *ptr++ = "FEDCBA9876543210123456789ABCDEF" [15 + (tmpVal - value * base)]; |
| 367 | if (thSep && (base == 10) && (++thCount == 3)) |
| 368 | { |
| 369 | *ptr++ = thSep; |
| 370 | thCount = 0; |
| 371 | } |
| 372 | } while (value); |
| 373 | |
| 374 | if ('0' == fill) |
| 375 | { |
| 376 | if (prefix && base == 010) --width; |
| 377 | if (prefix && base == 0x10) width -= 2; |
| 378 | while ((ptr - result) < width) *ptr++ = fill; |
| 379 | } |
| 380 | |
| 381 | if (prefix && base == 010) *ptr++ = '0'; |
| 382 | else if (prefix && base == 0x10) |
| 383 | { |
| 384 | *ptr++ = 'x'; |
| 385 | *ptr++ = '0'; |
| 386 | } |
| 387 | |
| 388 | if ('0' != fill) |
| 389 | { |
| 390 | while ((ptr - result) < width) *ptr++ = fill; |
| 391 | } |
| 392 | |
| 393 | size = ptr - result; |
| 394 | poco_assert_dbg (size <= ptr.span()); |
| 395 | poco_assert_dbg ((-1 == width) || (size >= size_t(width))); |
| 396 | *ptr-- = '\0'; |
| 397 | |
| 398 | char* ptrr = result; |
| 399 | char tmp; |
| 400 | while(ptrr < ptr) |
| 401 | { |
| 402 | tmp = *ptr; |
| 403 | *ptr-- = *ptrr; |
| 404 | *ptrr++ = tmp; |
| 405 | } |
| 406 | |
| 407 | return true; |
| 408 | } |
| 409 | |
| 410 | |
| 411 | template <typename T> |
| 412 | bool intToStr (T number, unsigned short base, std::string& result, bool prefix = false, int width = -1, char fill = ' ', char thSep = 0) |
| 413 | /// Converts integer to string; This is a wrapper function, for details see see the |
| 414 | /// bool intToStr(T, unsigned short, char*, int, int, char, char) implementation. |
| 415 | { |
| 416 | char res[POCO_MAX_INT_STRING_LEN] = {0}; |
| 417 | std::size_t size = POCO_MAX_INT_STRING_LEN; |
| 418 | bool ret = intToStr(number, base, res, size, prefix, width, fill, thSep); |
| 419 | result.assign(res, size); |
| 420 | return ret; |
| 421 | } |
| 422 | |
| 423 | |
| 424 | template <typename T> |
| 425 | bool uIntToStr (T number, unsigned short base, std::string& result, bool prefix = false, int width = -1, char fill = ' ', char thSep = 0) |
| 426 | /// Converts unsigned integer to string; This is a wrapper function, for details see see the |
| 427 | /// bool uIntToStr(T, unsigned short, char*, int, int, char, char) implementation. |
| 428 | { |
| 429 | char res[POCO_MAX_INT_STRING_LEN] = {0}; |
| 430 | std::size_t size = POCO_MAX_INT_STRING_LEN; |
| 431 | bool ret = uIntToStr(number, base, res, size, prefix, width, fill, thSep); |
| 432 | result.assign(res, size); |
| 433 | return ret; |
| 434 | } |
| 435 | |
| 436 | |
| 437 | // |
| 438 | // Wrappers for double-conversion library (http://code.google.com/p/double-conversion/). |
| 439 | // |
| 440 | // Library is the implementation of the algorithm described in Florian Loitsch's paper: |
| 441 | // http://florian.loitsch.com/publications/dtoa-pldi2010.pdf |
| 442 | // |
| 443 | |
| 444 | |
| 445 | Foundation_API void floatToStr(char* buffer, |
| 446 | int bufferSize, |
| 447 | float value, |
| 448 | int lowDec = -std::numeric_limits<float>::digits10, |
| 449 | int highDec = std::numeric_limits<float>::digits10); |
| 450 | /// Converts a float value to string. Converted string must be shorter than bufferSize. |
| 451 | /// Conversion is done by computing the shortest string of digits that correctly represents |
| 452 | /// the input number. Depending on lowDec and highDec values, the function returns |
| 453 | /// decimal or exponential representation. |
| 454 | |
| 455 | Foundation_API void floatToFixedStr(char* buffer, |
| 456 | int bufferSize, |
| 457 | float value, |
| 458 | int precision); |
| 459 | /// Converts a float value to string. Converted string must be shorter than bufferSize. |
| 460 | /// Computes a decimal representation with a fixed number of digits after the |
| 461 | /// decimal point. |
| 462 | |
| 463 | |
| 464 | Foundation_API std::string& floatToStr(std::string& str, |
| 465 | float value, |
| 466 | int precision = -1, |
| 467 | int width = 0, |
| 468 | char thSep = 0, |
| 469 | char decSep = 0); |
| 470 | /// Converts a float value, assigns it to the supplied string and returns the reference. |
| 471 | /// This function calls floatToStr(char*, int, float, int, int) and formats the result according to |
| 472 | /// precision (total number of digits after the decimal point, -1 means ignore precision argument) |
| 473 | /// and width (total length of formatted string). |
| 474 | |
| 475 | |
| 476 | Foundation_API std::string& floatToFixedStr(std::string& str, |
| 477 | float value, |
| 478 | int precision, |
| 479 | int width = 0, |
| 480 | char thSep = 0, |
| 481 | char decSep = 0); |
| 482 | /// Converts a float value, assigns it to the supplied string and returns the reference. |
| 483 | /// This function calls floatToFixedStr(char*, int, float, int) and formats the result according to |
| 484 | /// precision (total number of digits after the decimal point) and width (total length of formatted string). |
| 485 | |
| 486 | |
| 487 | Foundation_API void doubleToStr(char* buffer, |
| 488 | int bufferSize, |
| 489 | double value, |
| 490 | int lowDec = -std::numeric_limits<double>::digits10, |
| 491 | int highDec = std::numeric_limits<double>::digits10); |
| 492 | /// Converts a double value to string. Converted string must be shorter than bufferSize. |
| 493 | /// Conversion is done by computing the shortest string of digits that correctly represents |
| 494 | /// the input number. Depending on lowDec and highDec values, the function returns |
| 495 | /// decimal or exponential representation. |
| 496 | |
| 497 | |
| 498 | Foundation_API void doubleToFixedStr(char* buffer, |
| 499 | int bufferSize, |
| 500 | double value, |
| 501 | int precision); |
| 502 | /// Converts a double value to string. Converted string must be shorter than bufferSize. |
| 503 | /// Computes a decimal representation with a fixed number of digits after the |
| 504 | /// decimal point. |
| 505 | |
| 506 | |
| 507 | Foundation_API std::string& doubleToStr(std::string& str, |
| 508 | double value, |
| 509 | int precision = -1, |
| 510 | int width = 0, |
| 511 | char thSep = 0, |
| 512 | char decSep = 0); |
| 513 | /// Converts a double value, assigns it to the supplied string and returns the reference. |
| 514 | /// This function calls doubleToStr(char*, int, double, int, int) and formats the result according to |
| 515 | /// precision (total number of digits after the decimal point, -1 means ignore precision argument) |
| 516 | /// and width (total length of formatted string). |
| 517 | |
| 518 | |
| 519 | Foundation_API std::string& doubleToFixedStr(std::string& str, |
| 520 | double value, |
| 521 | int precision = -1, |
| 522 | int width = 0, |
| 523 | char thSep = 0, |
| 524 | char decSep = 0); |
| 525 | /// Converts a double value, assigns it to the supplied string and returns the reference. |
| 526 | /// This function calls doubleToFixedStr(char*, int, double, int) and formats the result according to |
| 527 | /// precision (total number of digits after the decimal point) and width (total length of formatted string). |
| 528 | |
| 529 | |
| 530 | Foundation_API float strToFloat(const char* str, |
| 531 | const char* inf = POCO_FLT_INF, const char* nan = POCO_FLT_NAN); |
| 532 | /// Converts the string of characters into single-precision floating point number. |
| 533 | /// Function uses double_convesrion::DoubleToStringConverter to do the conversion. |
| 534 | |
| 535 | |
| 536 | Foundation_API bool strToFloat(const std::string&, float& result, |
| 537 | char decSep = '.', char thSep = ',', |
| 538 | const char* inf = POCO_FLT_INF, const char* nan = POCO_FLT_NAN); |
| 539 | /// Converts the string of characters into single-precision floating point number. |
| 540 | /// The conversion result is assigned to the result parameter. |
| 541 | /// If decimal separator and/or thousand separator are different from defaults, they should be |
| 542 | /// supplied to ensure proper conversion. |
| 543 | /// |
| 544 | /// Returns true if successful, false otherwise. |
| 545 | |
| 546 | |
| 547 | Foundation_API double strToDouble(const char* str, |
| 548 | const char* inf = POCO_FLT_INF, const char* nan = POCO_FLT_NAN); |
| 549 | /// Converts the string of characters into double-precision floating point number. |
| 550 | |
| 551 | |
| 552 | Foundation_API bool strToDouble(const std::string& str, double& result, |
| 553 | char decSep = '.', char thSep = ',', |
| 554 | const char* inf = POCO_FLT_INF, const char* nan = POCO_FLT_NAN); |
| 555 | /// Converts the string of characters into double-precision floating point number. |
| 556 | /// The conversion result is assigned to the result parameter. |
| 557 | /// If decimal separator and/or thousand separator are different from defaults, they should be |
| 558 | /// supplied to ensure proper conversion. |
| 559 | /// |
| 560 | /// Returns true if successful, false otherwise. |
| 561 | |
| 562 | |
| 563 | } // namespace Poco |
| 564 | |
| 565 | |
| 566 | #endif // Foundation_NumericString_INCLUDED |
| 567 | |