| 1 | /* |
| 2 | * Copyright (c) 1989 The Regents of the University of California. |
| 3 | * All rights reserved. |
| 4 | * |
| 5 | * Redistribution and use in source and binary forms are permitted |
| 6 | * provided that the above copyright notice and this paragraph are |
| 7 | * duplicated in all such forms and that any documentation, |
| 8 | * advertising materials, and other materials related to such |
| 9 | * distribution and use acknowledge that the software was developed |
| 10 | * by the University of California, Berkeley. The name of the |
| 11 | * University may not be used to endorse or promote products derived |
| 12 | * from this software without specific prior written permission. |
| 13 | * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR |
| 14 | * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED |
| 15 | * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. |
| 16 | */ |
| 17 | |
| 18 | #ifndef lint |
| 19 | static char sccsid[] = "@(#)parse.c 5.1 (Berkeley) %G%"; |
| 20 | #endif /* not lint */ |
| 21 | |
| 22 | #include <sys/types.h> |
| 23 | #include <sys/file.h> |
| 24 | #include <stdio.h> |
| 25 | #include <ctype.h> |
| 26 | #include <strings.h> |
| 27 | #include "hexdump.h" |
| 28 | |
| 29 | FU *endfu; /* format at end-of-data */ |
| 30 | |
| 31 | addfile(name) |
| 32 | char *name; |
| 33 | { |
| 34 | register char *p; |
| 35 | FILE *fp; |
| 36 | int ch; |
| 37 | char buf[2048 + 1]; |
| 38 | |
| 39 | if (!(fp = fopen(name, "r"))) { |
| 40 | (void)fprintf(stderr, "hexdump: can't read %s.\n", name); |
| 41 | exit(1); |
| 42 | } |
| 43 | while (fgets(buf, sizeof(buf), fp)) { |
| 44 | if (!(p = index(buf, '\n'))) { |
| 45 | (void)fprintf(stderr, "hexdump: line too long.\n"); |
| 46 | while ((ch = getchar()) != '\n' && ch != EOF); |
| 47 | continue; |
| 48 | } |
| 49 | *p = '\0'; |
| 50 | for (p = buf; *p && isspace(*p); ++p); |
| 51 | if (!*p || *p == '#') |
| 52 | continue; |
| 53 | add(p); |
| 54 | } |
| 55 | (void)fclose(fp); |
| 56 | } |
| 57 | |
| 58 | add(fmt) |
| 59 | char *fmt; |
| 60 | { |
| 61 | register char *p; |
| 62 | static FS **nextfs; |
| 63 | FS *tfs; |
| 64 | FU *tfu, **nextfu; |
| 65 | char savech, *savep, *emalloc(), *strdup(); |
| 66 | |
| 67 | /* start new linked list of format units */ |
| 68 | /* NOSTRICT */ |
| 69 | tfs = (FS *)emalloc(sizeof(FS)); |
| 70 | if (!fshead) |
| 71 | fshead = tfs; |
| 72 | else |
| 73 | *nextfs = tfs; |
| 74 | nextfs = &tfs->nextfs; |
| 75 | nextfu = &tfs->nextfu; |
| 76 | |
| 77 | /* take the format string and break it up into format units */ |
| 78 | for (p = fmt;;) { |
| 79 | /* skip leading white space */ |
| 80 | for (; isspace(*p); ++p); |
| 81 | if (!*p) |
| 82 | break; |
| 83 | |
| 84 | /* allocate a new format unit and link it in */ |
| 85 | /* NOSTRICT */ |
| 86 | tfu = (FU *)emalloc(sizeof(FU)); |
| 87 | *nextfu = tfu; |
| 88 | nextfu = &tfu->nextfu; |
| 89 | tfu->reps = 1; |
| 90 | |
| 91 | /* if leading digit, repetition count */ |
| 92 | if (isdigit(*p)) { |
| 93 | for (savep = p; isdigit(*p); ++p); |
| 94 | if (!isspace(*p) && *p != '/') |
| 95 | badfmt(fmt); |
| 96 | /* may overwrite either white space or slash */ |
| 97 | savech = *p; |
| 98 | *p = '\0'; |
| 99 | tfu->reps = atoi(savep); |
| 100 | tfu->flags = F_SETREP; |
| 101 | *p = savech; |
| 102 | /* skip trailing white space */ |
| 103 | for (++p; isspace(*p); ++p); |
| 104 | } |
| 105 | |
| 106 | /* skip slash and trailing white space */ |
| 107 | if (*p == '/') |
| 108 | while (isspace(*++p)); |
| 109 | |
| 110 | /* byte count */ |
| 111 | if (isdigit(*p)) { |
| 112 | for (savep = p; isdigit(*p); ++p); |
| 113 | if (!isspace(*p)) |
| 114 | badfmt(fmt); |
| 115 | savech = *p; |
| 116 | *p = '\0'; |
| 117 | tfu->bcnt = atoi(savep); |
| 118 | *p = savech; |
| 119 | /* skip trailing white space */ |
| 120 | for (++p; isspace(*p); ++p); |
| 121 | } |
| 122 | |
| 123 | /* format */ |
| 124 | if (*p != '"') |
| 125 | badfmt(fmt); |
| 126 | for (savep = ++p; *p != '"'; ++p); |
| 127 | if (*p != '"') |
| 128 | badfmt(fmt); |
| 129 | savech = *p; |
| 130 | *p = '\0'; |
| 131 | if (!(tfu->fmt = strdup(savep))) |
| 132 | nomem(); |
| 133 | escape(tfu->fmt); |
| 134 | *p++ = savech; |
| 135 | } |
| 136 | } |
| 137 | |
| 138 | static char *spec = ".#-+ 0123456789"; |
| 139 | size(fs) |
| 140 | FS *fs; |
| 141 | { |
| 142 | register FU *fu; |
| 143 | register int bcnt, cursize; |
| 144 | register char *fmt; |
| 145 | int prec; |
| 146 | |
| 147 | /* figure out the data block size needed for each format unit */ |
| 148 | for (cursize = 0, fu = fs->nextfu; fu; fu = fu->nextfu) { |
| 149 | if (fu->bcnt) { |
| 150 | cursize += fu->bcnt * fu->reps; |
| 151 | continue; |
| 152 | } |
| 153 | for (bcnt = prec = 0, fmt = fu->fmt; *fmt; ++fmt) { |
| 154 | if (*fmt != '%') |
| 155 | continue; |
| 156 | /* |
| 157 | * skip any special chars -- save precision in |
| 158 | * case it's a %s format. |
| 159 | */ |
| 160 | while (index(spec + 1, *++fmt)); |
| 161 | if (*fmt == '.' && isdigit(*++fmt)) { |
| 162 | prec = atoi(fmt); |
| 163 | while (isdigit(*++fmt)); |
| 164 | } |
| 165 | switch(*fmt) { |
| 166 | case 'c': |
| 167 | bcnt += 1; |
| 168 | break; |
| 169 | case 'd': case 'i': case 'o': case 'u': |
| 170 | case 'x': case 'X': |
| 171 | bcnt += 4; |
| 172 | break; |
| 173 | case 'e': case 'E': case 'f': case 'g': case 'G': |
| 174 | bcnt += 8; |
| 175 | break; |
| 176 | case 's': |
| 177 | bcnt += prec; |
| 178 | break; |
| 179 | case '_': |
| 180 | switch(*++fmt) { |
| 181 | case 'c': case 'p': case 'u': |
| 182 | bcnt += 1; |
| 183 | break; |
| 184 | } |
| 185 | } |
| 186 | } |
| 187 | cursize += bcnt * fu->reps; |
| 188 | } |
| 189 | return(cursize); |
| 190 | } |
| 191 | |
| 192 | rewrite(fs) |
| 193 | FS *fs; |
| 194 | { |
| 195 | enum { NOTOKAY, USEBCNT, USEPREC } sokay; |
| 196 | register PR *pr, **nextpr; |
| 197 | register FU *fu; |
| 198 | register char *p1, *p2; |
| 199 | char savech, *fmtp; |
| 200 | int nconv, prec; |
| 201 | |
| 202 | for (fu = fs->nextfu; fu; fu = fu->nextfu) { |
| 203 | /* |
| 204 | * break each format unit into print units; each |
| 205 | * conversion character gets its own. |
| 206 | */ |
| 207 | for (nconv = 0, fmtp = fu->fmt; *fmtp; nextpr = &pr->nextpr) { |
| 208 | /* NOSTRICT */ |
| 209 | pr = (PR *)emalloc(sizeof(PR)); |
| 210 | if (!fu->nextpr) |
| 211 | fu->nextpr = pr; |
| 212 | else |
| 213 | *nextpr = pr; |
| 214 | |
| 215 | /* skip preceding text and up to the next % sign */ |
| 216 | for (p1 = fmtp; *p1 && *p1 != '%'; ++p1); |
| 217 | |
| 218 | /* only text in the string */ |
| 219 | if (!*p1) { |
| 220 | pr->fmt = fmtp; |
| 221 | pr->flags = F_TEXT; |
| 222 | break; |
| 223 | } |
| 224 | |
| 225 | /* |
| 226 | * get precision for %s -- if have a byte count, don't |
| 227 | * need it. |
| 228 | */ |
| 229 | if (fu->bcnt) { |
| 230 | sokay = USEBCNT; |
| 231 | /* skip to conversion character */ |
| 232 | for (++p1; index(spec, *p1); ++p1); |
| 233 | } else { |
| 234 | /* skip any special chars, field width */ |
| 235 | while (index(spec + 1, *++p1)); |
| 236 | if (*p1 == '.' && isdigit(*++p1)) { |
| 237 | sokay = USEPREC; |
| 238 | prec = atoi(p1); |
| 239 | while (isdigit(*++p1)); |
| 240 | } |
| 241 | else |
| 242 | sokay = NOTOKAY; |
| 243 | } |
| 244 | |
| 245 | p2 = p1 + 1; /* set end pointer */ |
| 246 | |
| 247 | /* |
| 248 | * figure out the byte count for each conversion; |
| 249 | * rewrite the format as necessary, set up blank- |
| 250 | * padding for end of data. |
| 251 | */ |
| 252 | switch(*p1) { |
| 253 | case 'c': |
| 254 | pr->flags = F_CHAR; |
| 255 | switch(fu->bcnt) { |
| 256 | case 0: case 1: |
| 257 | pr->bcnt = 1; |
| 258 | break; |
| 259 | default: |
| 260 | p1[1] = '\0'; |
| 261 | badcnt(p1); |
| 262 | } |
| 263 | break; |
| 264 | case 'd': case 'i': |
| 265 | pr->flags = F_INT; |
| 266 | goto sw1; |
| 267 | case 'o': case 'u': case 'x': case 'X': |
| 268 | pr->flags = F_UINT; |
| 269 | sw1: switch(fu->bcnt) { |
| 270 | case 0: case 4: |
| 271 | pr->bcnt = 4; |
| 272 | break; |
| 273 | case 1: |
| 274 | pr->bcnt = 1; |
| 275 | break; |
| 276 | case 2: |
| 277 | pr->bcnt = 2; |
| 278 | break; |
| 279 | default: |
| 280 | p1[1] = '\0'; |
| 281 | badcnt(p1); |
| 282 | } |
| 283 | break; |
| 284 | case 'e': case 'E': case 'f': case 'g': case 'G': |
| 285 | pr->flags = F_DBL; |
| 286 | switch(fu->bcnt) { |
| 287 | case 0: case 8: |
| 288 | pr->bcnt = 8; |
| 289 | break; |
| 290 | case 4: |
| 291 | pr->bcnt = 4; |
| 292 | break; |
| 293 | default: |
| 294 | p1[1] = '\0'; |
| 295 | badcnt(p1); |
| 296 | } |
| 297 | break; |
| 298 | case 's': |
| 299 | pr->flags = F_STR; |
| 300 | switch(sokay) { |
| 301 | case NOTOKAY: |
| 302 | badsfmt(); |
| 303 | case USEBCNT: |
| 304 | pr->bcnt = fu->bcnt; |
| 305 | break; |
| 306 | case USEPREC: |
| 307 | pr->bcnt = prec; |
| 308 | break; |
| 309 | } |
| 310 | break; |
| 311 | case '_': |
| 312 | ++p2; |
| 313 | switch(p1[1]) { |
| 314 | case 'A': |
| 315 | endfu = fu; |
| 316 | fu->flags |= F_IGNORE; |
| 317 | /* FALLTHROUGH */ |
| 318 | case 'a': |
| 319 | pr->flags = F_ADDRESS; |
| 320 | ++p2; |
| 321 | switch(p1[2]) { |
| 322 | case 'd': case 'o': case'x': |
| 323 | *p1 = p1[2]; |
| 324 | break; |
| 325 | default: |
| 326 | p1[3] = '\0'; |
| 327 | badconv(p1); |
| 328 | } |
| 329 | break; |
| 330 | case 'c': |
| 331 | pr->flags = F_C; |
| 332 | /* *p1 = 'c'; set in conv_c */ |
| 333 | goto sw2; |
| 334 | case 'p': |
| 335 | pr->flags = F_P; |
| 336 | *p1 = 'c'; |
| 337 | goto sw2; |
| 338 | case 'u': |
| 339 | pr->flags = F_U; |
| 340 | /* *p1 = 'c'; set in conv_u */ |
| 341 | sw2: switch(fu->bcnt) { |
| 342 | case 0: case 1: |
| 343 | pr->bcnt = 1; |
| 344 | break; |
| 345 | default: |
| 346 | p1[2] = '\0'; |
| 347 | badcnt(p1); |
| 348 | } |
| 349 | break; |
| 350 | default: |
| 351 | p1[2] = '\0'; |
| 352 | badconv(p1); |
| 353 | } |
| 354 | break; |
| 355 | default: |
| 356 | p1[1] = '\0'; |
| 357 | badconv(p1); |
| 358 | } |
| 359 | |
| 360 | /* |
| 361 | * copy to PR format string, set conversion character |
| 362 | * pointer, update original. |
| 363 | */ |
| 364 | savech = *p2; |
| 365 | p1[1] = '\0'; |
| 366 | if (!(pr->fmt = strdup(fmtp))) |
| 367 | nomem(); |
| 368 | *p2 = savech; |
| 369 | pr->cchar = pr->fmt + (p1 - fmtp); |
| 370 | fmtp = p2; |
| 371 | |
| 372 | /* only one conversion character if byte count */ |
| 373 | if (!(pr->flags&F_ADDRESS) && fu->bcnt && nconv++) { |
| 374 | (void)fprintf(stderr, |
| 375 | "hexdump: byte count with multiple conversion characters.\n"); |
| 376 | exit(1); |
| 377 | } |
| 378 | } |
| 379 | /* |
| 380 | * if format unit byte count not specified, figure it out |
| 381 | * so can adjust rep count later. |
| 382 | */ |
| 383 | if (!fu->bcnt) |
| 384 | for (pr = fu->nextpr; pr; pr = pr->nextpr) |
| 385 | fu->bcnt += pr->bcnt; |
| 386 | } |
| 387 | /* |
| 388 | * if the format string interprets any data at all, and it's |
| 389 | * not the same as the blocksize, and its last format unit |
| 390 | * interprets any data at all, and has no iteration count, |
| 391 | * repeat it as necessary. |
| 392 | * |
| 393 | * if, rep count is greater than 1, no trailing whitespace |
| 394 | * gets output from the last iteration of the format unit. |
| 395 | */ |
| 396 | for (fu = fs->nextfu;; fu = fu->nextfu) { |
| 397 | if (!fu->nextfu && fs->bcnt < blocksize && |
| 398 | !(fu->flags&F_SETREP) && fu->bcnt) |
| 399 | fu->reps += (blocksize - fs->bcnt) / fu->bcnt; |
| 400 | if (fu->reps > 1) { |
| 401 | for (pr = fu->nextpr;; pr = pr->nextpr) |
| 402 | if (!pr->nextpr) |
| 403 | break; |
| 404 | for (p1 = pr->fmt, p2 = NULL; *p1; ++p1) |
| 405 | p2 = isspace(*p1) ? p1 : NULL; |
| 406 | if (p2) |
| 407 | pr->nospace = p2; |
| 408 | } |
| 409 | if (!fu->nextfu) |
| 410 | break; |
| 411 | } |
| 412 | } |
| 413 | |
| 414 | |
| 415 | escape(p1) |
| 416 | register char *p1; |
| 417 | { |
| 418 | register char *p2; |
| 419 | |
| 420 | /* alphabetic escape sequences have to be done in place */ |
| 421 | for (p2 = p1;; ++p1, ++p2) { |
| 422 | if (!*p1) { |
| 423 | *p2 = *p1; |
| 424 | break; |
| 425 | } |
| 426 | if (*p1 == '\\') |
| 427 | switch(*++p1) { |
| 428 | case 'a': |
| 429 | /* *p2 = '\a'; */ |
| 430 | *p2 = '\007'; |
| 431 | break; |
| 432 | case 'b': |
| 433 | *p2 = '\b'; |
| 434 | break; |
| 435 | case 'f': |
| 436 | *p2 = '\f'; |
| 437 | break; |
| 438 | case 'n': |
| 439 | *p2 = '\n'; |
| 440 | break; |
| 441 | case 'r': |
| 442 | *p2 = '\r'; |
| 443 | break; |
| 444 | case 't': |
| 445 | *p2 = '\t'; |
| 446 | break; |
| 447 | case 'v': |
| 448 | *p2 = '\v'; |
| 449 | break; |
| 450 | default: |
| 451 | *p2 = *p1; |
| 452 | break; |
| 453 | } |
| 454 | } |
| 455 | } |
| 456 | |
| 457 | badcnt(s) |
| 458 | char *s; |
| 459 | { |
| 460 | (void)fprintf(stderr, |
| 461 | "hexdump: bad byte count for conversion character %s.\n", s); |
| 462 | exit(1); |
| 463 | } |
| 464 | |
| 465 | badsfmt() |
| 466 | { |
| 467 | (void)fprintf(stderr, |
| 468 | "hexdump: %%s requires a precision or a byte count.\n"); |
| 469 | exit(1); |
| 470 | } |
| 471 | |
| 472 | badfmt(fmt) |
| 473 | char *fmt; |
| 474 | { |
| 475 | (void)fprintf(stderr, "hexdump: bad format {%s}\n", fmt); |
| 476 | exit(1); |
| 477 | } |
| 478 | |
| 479 | badconv(ch) |
| 480 | char *ch; |
| 481 | { |
| 482 | (void)fprintf(stderr, "hexdump: bad conversion character %s.\n", ch); |
| 483 | exit(1); |
| 484 | } |