Elliott Hughes | 7f0849f | 2016-08-26 16:17:17 -0700 | [diff] [blame] | 1 | /*- |
| 2 | * Copyright (c) 1990, 1993 |
| 3 | * The Regents of the University of California. All rights reserved. |
| 4 | * |
| 5 | * This code is derived from software contributed to Berkeley by |
| 6 | * Chris Torek. |
| 7 | * |
| 8 | * Redistribution and use in source and binary forms, with or without |
| 9 | * modification, are permitted provided that the following conditions |
| 10 | * are met: |
| 11 | * 1. Redistributions of source code must retain the above copyright |
| 12 | * notice, this list of conditions and the following disclaimer. |
| 13 | * 2. Redistributions in binary form must reproduce the above copyright |
| 14 | * notice, this list of conditions and the following disclaimer in the |
| 15 | * documentation and/or other materials provided with the distribution. |
| 16 | * 3. Neither the name of the University nor the names of its contributors |
| 17 | * may be used to endorse or promote products derived from this software |
| 18 | * without specific prior written permission. |
| 19 | * |
| 20 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
| 21 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 22 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 23 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE |
| 24 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
| 25 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
| 26 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
| 27 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
| 28 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
| 29 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
| 30 | * SUCH DAMAGE. |
| 31 | */ |
| 32 | |
| 33 | #include <ctype.h> |
| 34 | #include <stdlib.h> |
| 35 | |
| 36 | #include "local.h" |
Elliott Hughes | 7f0849f | 2016-08-26 16:17:17 -0700 | [diff] [blame] | 37 | |
| 38 | #define BUF 513 /* Maximum length of numeric string. */ |
| 39 | |
| 40 | size_t parsefloat(FILE *fp, char *buf, char *end) { |
| 41 | char *commit, *p; |
| 42 | int infnanpos = 0; |
| 43 | enum { |
| 44 | S_START, S_GOTSIGN, S_INF, S_NAN, S_MAYBEHEX, |
| 45 | S_DIGITS, S_FRAC, S_EXP, S_EXPDIGITS |
| 46 | } state = S_START; |
| 47 | unsigned char c; |
| 48 | int gotmantdig = 0, ishex = 0; |
| 49 | |
| 50 | /* |
| 51 | * We set commit = p whenever the string we have read so far |
| 52 | * constitutes a valid representation of a floating point |
| 53 | * number by itself. At some point, the parse will complete |
| 54 | * or fail, and we will ungetc() back to the last commit point. |
| 55 | * To ensure that the file offset gets updated properly, it is |
| 56 | * always necessary to read at least one character that doesn't |
| 57 | * match; thus, we can't short-circuit "infinity" or "nan(...)". |
| 58 | */ |
| 59 | commit = buf - 1; |
| 60 | for (p = buf; p < end; ) { |
| 61 | c = *fp->_p; |
| 62 | reswitch: |
| 63 | switch (state) { |
| 64 | case S_START: |
| 65 | state = S_GOTSIGN; |
| 66 | if (c == '-' || c == '+') |
| 67 | break; |
| 68 | else |
| 69 | goto reswitch; |
| 70 | case S_GOTSIGN: |
| 71 | switch (c) { |
| 72 | case '0': |
| 73 | state = S_MAYBEHEX; |
| 74 | commit = p; |
| 75 | break; |
| 76 | case 'I': |
| 77 | case 'i': |
| 78 | state = S_INF; |
| 79 | break; |
| 80 | case 'N': |
| 81 | case 'n': |
| 82 | state = S_NAN; |
| 83 | break; |
| 84 | default: |
| 85 | state = S_DIGITS; |
| 86 | goto reswitch; |
| 87 | } |
| 88 | break; |
| 89 | case S_INF: |
| 90 | if (infnanpos > 6 || |
| 91 | (c != "nfinity"[infnanpos] && |
| 92 | c != "NFINITY"[infnanpos])) |
| 93 | goto parsedone; |
| 94 | if (infnanpos == 1 || infnanpos == 6) |
| 95 | commit = p; /* inf or infinity */ |
| 96 | infnanpos++; |
| 97 | break; |
| 98 | case S_NAN: |
| 99 | switch (infnanpos) { |
| 100 | case -1: /* XXX kludge to deal with nan(...) */ |
| 101 | goto parsedone; |
| 102 | case 0: |
| 103 | if (c != 'A' && c != 'a') |
| 104 | goto parsedone; |
| 105 | break; |
| 106 | case 1: |
| 107 | if (c != 'N' && c != 'n') |
| 108 | goto parsedone; |
| 109 | else |
| 110 | commit = p; |
| 111 | break; |
| 112 | case 2: |
| 113 | if (c != '(') |
| 114 | goto parsedone; |
| 115 | break; |
| 116 | default: |
| 117 | if (c == ')') { |
| 118 | commit = p; |
| 119 | infnanpos = -2; |
| 120 | } else if (!isalnum(c) && c != '_') |
| 121 | goto parsedone; |
| 122 | break; |
| 123 | } |
| 124 | infnanpos++; |
| 125 | break; |
| 126 | case S_MAYBEHEX: |
| 127 | state = S_DIGITS; |
| 128 | if (c == 'X' || c == 'x') { |
| 129 | ishex = 1; |
| 130 | break; |
| 131 | } else { /* we saw a '0', but no 'x' */ |
| 132 | gotmantdig = 1; |
| 133 | goto reswitch; |
| 134 | } |
| 135 | case S_DIGITS: |
| 136 | if ((ishex && isxdigit(c)) || isdigit(c)) |
| 137 | gotmantdig = 1; |
| 138 | else { |
| 139 | state = S_FRAC; |
| 140 | if (c != '.') |
| 141 | goto reswitch; |
| 142 | } |
| 143 | if (gotmantdig) |
| 144 | commit = p; |
| 145 | break; |
| 146 | case S_FRAC: |
| 147 | if (((c == 'E' || c == 'e') && !ishex) || |
| 148 | ((c == 'P' || c == 'p') && ishex)) { |
| 149 | if (!gotmantdig) |
| 150 | goto parsedone; |
| 151 | else |
| 152 | state = S_EXP; |
| 153 | } else if ((ishex && isxdigit(c)) || isdigit(c)) { |
| 154 | commit = p; |
| 155 | gotmantdig = 1; |
| 156 | } else |
| 157 | goto parsedone; |
| 158 | break; |
| 159 | case S_EXP: |
| 160 | state = S_EXPDIGITS; |
| 161 | if (c == '-' || c == '+') |
| 162 | break; |
| 163 | else |
| 164 | goto reswitch; |
| 165 | case S_EXPDIGITS: |
| 166 | if (isdigit(c)) |
| 167 | commit = p; |
| 168 | else |
| 169 | goto parsedone; |
| 170 | break; |
| 171 | default: |
| 172 | abort(); |
| 173 | } |
| 174 | *p++ = c; |
| 175 | if (--fp->_r > 0) |
| 176 | fp->_p++; |
| 177 | else if (__srefill(fp)) |
| 178 | break; /* EOF */ |
| 179 | } |
| 180 | |
| 181 | parsedone: |
| 182 | while (commit < --p) |
| 183 | (void)ungetc(*(unsigned char *)p, fp); |
| 184 | *++commit = '\0'; |
| 185 | return commit - buf; |
| 186 | } |
| 187 | |
| 188 | size_t wparsefloat(FILE *fp, wchar_t *buf, wchar_t *end) { |
| 189 | wchar_t *commit, *p; |
| 190 | int infnanpos = 0; |
| 191 | enum { |
| 192 | S_START, S_GOTSIGN, S_INF, S_NAN, S_MAYBEHEX, |
| 193 | S_DIGITS, S_FRAC, S_EXP, S_EXPDIGITS |
| 194 | } state = S_START; |
| 195 | wint_t c; |
| 196 | int gotmantdig = 0, ishex = 0; |
| 197 | |
| 198 | /* |
| 199 | * We set commit = p whenever the string we have read so far |
| 200 | * constitutes a valid representation of a floating point |
| 201 | * number by itself. At some point, the parse will complete |
| 202 | * or fail, and we will ungetc() back to the last commit point. |
| 203 | * To ensure that the file offset gets updated properly, it is |
| 204 | * always necessary to read at least one character that doesn't |
| 205 | * match; thus, we can't short-circuit "infinity" or "nan(...)". |
| 206 | */ |
| 207 | commit = buf - 1; |
| 208 | c = WEOF; |
| 209 | for (p = buf; p < end; ) { |
| 210 | if ((c = __fgetwc_unlock(fp)) == WEOF) |
| 211 | break; |
| 212 | reswitch: |
| 213 | switch (state) { |
| 214 | case S_START: |
| 215 | state = S_GOTSIGN; |
| 216 | if (c == '-' || c == '+') |
| 217 | break; |
| 218 | else |
| 219 | goto reswitch; |
| 220 | case S_GOTSIGN: |
| 221 | switch (c) { |
| 222 | case '0': |
| 223 | state = S_MAYBEHEX; |
| 224 | commit = p; |
| 225 | break; |
| 226 | case 'I': |
| 227 | case 'i': |
| 228 | state = S_INF; |
| 229 | break; |
| 230 | case 'N': |
| 231 | case 'n': |
| 232 | state = S_NAN; |
| 233 | break; |
| 234 | default: |
| 235 | state = S_DIGITS; |
| 236 | goto reswitch; |
| 237 | } |
| 238 | break; |
| 239 | case S_INF: |
| 240 | if (infnanpos > 6 || |
| 241 | (c != (wint_t)"nfinity"[infnanpos] && |
| 242 | c != (wint_t)"NFINITY"[infnanpos])) |
| 243 | goto parsedone; |
| 244 | if (infnanpos == 1 || infnanpos == 6) |
| 245 | commit = p; /* inf or infinity */ |
| 246 | infnanpos++; |
| 247 | break; |
| 248 | case S_NAN: |
| 249 | switch (infnanpos) { |
| 250 | case -1: /* XXX kludge to deal with nan(...) */ |
| 251 | goto parsedone; |
| 252 | case 0: |
| 253 | if (c != 'A' && c != 'a') |
| 254 | goto parsedone; |
| 255 | break; |
| 256 | case 1: |
| 257 | if (c != 'N' && c != 'n') |
| 258 | goto parsedone; |
| 259 | else |
| 260 | commit = p; |
| 261 | break; |
| 262 | case 2: |
| 263 | if (c != '(') |
| 264 | goto parsedone; |
| 265 | break; |
| 266 | default: |
| 267 | if (c == ')') { |
| 268 | commit = p; |
| 269 | infnanpos = -2; |
| 270 | } else if (!iswalnum(c) && c != '_') |
| 271 | goto parsedone; |
| 272 | break; |
| 273 | } |
| 274 | infnanpos++; |
| 275 | break; |
| 276 | case S_MAYBEHEX: |
| 277 | state = S_DIGITS; |
| 278 | if (c == 'X' || c == 'x') { |
| 279 | ishex = 1; |
| 280 | break; |
| 281 | } else { /* we saw a '0', but no 'x' */ |
| 282 | gotmantdig = 1; |
| 283 | goto reswitch; |
| 284 | } |
| 285 | case S_DIGITS: |
| 286 | if ((ishex && iswxdigit(c)) || iswdigit(c)) |
| 287 | gotmantdig = 1; |
| 288 | else { |
| 289 | state = S_FRAC; |
| 290 | if (c != L'.') |
| 291 | goto reswitch; |
| 292 | } |
| 293 | if (gotmantdig) |
| 294 | commit = p; |
| 295 | break; |
| 296 | case S_FRAC: |
| 297 | if (((c == 'E' || c == 'e') && !ishex) || |
| 298 | ((c == 'P' || c == 'p') && ishex)) { |
| 299 | if (!gotmantdig) |
| 300 | goto parsedone; |
| 301 | else |
| 302 | state = S_EXP; |
| 303 | } else if ((ishex && iswxdigit(c)) || iswdigit(c)) { |
| 304 | commit = p; |
| 305 | gotmantdig = 1; |
| 306 | } else |
| 307 | goto parsedone; |
| 308 | break; |
| 309 | case S_EXP: |
| 310 | state = S_EXPDIGITS; |
| 311 | if (c == '-' || c == '+') |
| 312 | break; |
| 313 | else |
| 314 | goto reswitch; |
| 315 | case S_EXPDIGITS: |
| 316 | if (iswdigit(c)) |
| 317 | commit = p; |
| 318 | else |
| 319 | goto parsedone; |
| 320 | break; |
| 321 | default: |
| 322 | abort(); |
| 323 | } |
| 324 | *p++ = c; |
| 325 | c = WEOF; |
| 326 | } |
| 327 | |
| 328 | parsedone: |
| 329 | if (c != WEOF) |
| 330 | ungetwc(c, fp); |
| 331 | while (commit < --p) |
| 332 | ungetwc(*p, fp); |
| 333 | *++commit = '\0'; |
| 334 | return (int)(commit - buf); |
| 335 | } |