text-utils/hexdump-parse.c

   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, with or without
   6  * modification, are permitted provided that the following conditions
   7  * are met:
   8  * 1. Redistributions of source code must retain the above copyright
   9  *    notice, this list of conditions and the following disclaimer.
  10  * 2. Redistributions in binary form must reproduce the above copyright
  11  *    notice, this list of conditions and the following disclaimer in the
  12  *    documentation and/or other materials provided with the distribution.
  13  * 3. All advertising materials mentioning features or use of this software
  14  *    must display the following acknowledgement:
  15  *      This product includes software developed by the University of
  16  *      California, Berkeley and its contributors.
  17  * 4. Neither the name of the University nor the names of its contributors
  18  *    may be used to endorse or promote products derived from this software
  19  *    without specific prior written permission.
  20  *
  21  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  22  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  23  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  24  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  25  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  26  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  27  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  28  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  29  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  30  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  31  * SUCH DAMAGE.
  32  */
  33
  34  /* 1999-02-22 Arkadiusz Miśkiewicz <misiek@pld.ORG.PL>
  35   * - added Native Language Support
  36   */
  37
  38 #include <sys/types.h>
  39 #include <sys/file.h>
  40 #include <stdio.h>
  41 #include <stdlib.h>
  42 #include <ctype.h>
  43 #include <string.h>
  44 #include "hexdump.h"
  45 #include "nls.h"
  46 #include "xalloc.h"
  47 #include "strutils.h"
  48 #include "colors.h"
  49
  50 static void escape(char *p1);
  51 static struct list_head *color_fmt(char *cfmt, int bcnt);
  52
  53 static void __attribute__ ((__noreturn__)) badcnt(const char *s)
  54 {
  55         errx(EXIT_FAILURE, _("bad byte count for conversion character %s"), s);
  56 }
  57
  58 static void __attribute__ ((__noreturn__)) badsfmt(void)
  59 {
  60         errx(EXIT_FAILURE, _("%%s requires a precision or a byte count"));
  61 }
  62
  63 static void __attribute__ ((__noreturn__)) badfmt(const char *fmt)
  64 {
  65         errx(EXIT_FAILURE, _("bad format {%s}"), fmt);
  66 }
  67
  68 static void __attribute__ ((__noreturn__)) badconv(const char *ch)
  69 {
  70         errx(EXIT_FAILURE, _("bad conversion character %%%s"), ch);
  71 }
  72
  73 #define first_letter(s,f) strchr(f, *(s))
  74
  75 struct hexdump_fu *endfu;                                       /* format at end-of-data */
  76
  77 void addfile(char *name, struct hexdump *hex)
  78 {
  79         char *fmt, *buf = NULL;
  80         FILE *fp;
  81         size_t n = 0;
  82
  83         if ((fp = fopen(name, "r")) == NULL)
  84                 err(EXIT_FAILURE, _("can't read %s"), name);
  85
  86         while (getline(&buf, &n, fp) != -1) {
  87                 fmt = buf;
  88
  89                 while (*fmt && isspace(*fmt))
  90                         ++fmt;
  91                 if (!*fmt || *fmt == '#')
  92                         continue;
  93
  94                 add_fmt(fmt, hex);
  95         }
  96
  97         free(buf);
  98         fclose(fp);
  99 }
 100
 101 static char *next_number(const char *str, int *num)
 102 {
 103         char *end = NULL;
 104
 105         errno = 0;
 106         *num = strtol(str, &end, 10);
 107
 108         if (errno || !end || end == str)
 109                 return NULL;
 110         return end;
 111 }
 112
 113 void add_fmt(const char *fmt, struct hexdump *hex)
 114 {
 115         const char *p, *savep;
 116         struct hexdump_fs *tfs;
 117         struct hexdump_fu *tfu;
 118
 119         /* Start new linked list of format units. */
 120         tfs = xcalloc(1, sizeof(struct hexdump_fs));
 121         INIT_LIST_HEAD(&tfs->fslist);
 122         INIT_LIST_HEAD(&tfs->fulist);
 123         list_add_tail(&tfs->fslist, &hex->fshead);
 124
 125         /* Take the format string and break it up into format units. */
 126         p = fmt;
 127         while (TRUE) {
 128                 /* Skip leading white space. */
 129                 if (!*(p = skip_space(p)))
 130                         break;
 131
 132                 /* Allocate a new format unit and link it in. */
 133                 tfu = xcalloc(1, sizeof(struct hexdump_fu));
 134                 tfu->reps = 1;
 135
 136                 INIT_LIST_HEAD(&tfu->fulist);
 137                 INIT_LIST_HEAD(&tfu->prlist);
 138                 list_add_tail(&tfu->fulist, &tfs->fulist);
 139
 140                 /* If leading digit, repetition count. */
 141                 if (isdigit(*p)) {
 142                         p = next_number(p, &tfu->reps);
 143                         if (!p || (!isspace(*p) && *p != '/'))
 144                                 badfmt(fmt);
 145
 146                         /* may overwrite either white space or slash */
 147                         tfu->flags = F_SETREP;
 148                         /* skip trailing white space */
 149                         p = skip_space(++p);
 150                 }
 151
 152                 /* Skip slash and trailing white space. */
 153                 if (*p == '/')
 154                         p = skip_space(++p);
 155
 156                 /* byte count */
 157                 if (isdigit(*p)) {
 158                         p = next_number(p, &tfu->bcnt);
 159                         if (!p || !isspace(*p))
 160                                 badfmt(fmt);
 161                         /* skip trailing white space */
 162                         p = skip_space(++p);
 163                 }
 164
 165                 /* format */
 166                 if (*p != '"')
 167                         badfmt(fmt);
 168                 savep = ++p;
 169                 while (*p != '"') {
 170                         if (!*p++)
 171                                 badfmt(fmt);
 172                 }
 173                 tfu->fmt = xmalloc(p - savep + 1);
 174                 xstrncpy(tfu->fmt, savep, p - savep + 1);
 175                 escape(tfu->fmt);
 176                 ++p;
 177         }
 178 }
 179
 180 static const char *spec = ".#-+ 0123456789";
 181
 182 int block_size(struct hexdump_fs *fs)
 183 {
 184         struct hexdump_fu *fu;
 185         int bcnt, prec, cursize = 0;
 186         char *fmt;
 187         struct list_head *p;
 188
 189         /* figure out the data block size needed for each format unit */
 190         list_for_each (p, &fs->fulist) {
 191                 fu = list_entry(p, struct hexdump_fu, fulist);
 192                 if (fu->bcnt) {
 193                         cursize += fu->bcnt * fu->reps;
 194                         continue;
 195                 }
 196                 bcnt = prec = 0;
 197                 fmt = fu->fmt;
 198                 while (*fmt) {
 199                         if (*fmt != '%') {
 200                                 ++fmt;
 201                                 continue;
 202                         }
 203                         /*
 204                          * skip any special chars -- save precision in
 205                          * case it's a %s format.
 206                          */
 207                         while (strchr(spec + 1, *++fmt) && *fmt != '\0')
 208                                 ;
 209                         if (*fmt == '.' && isdigit(*++fmt))
 210                                 fmt = next_number(fmt, &prec);
 211                         if (*fmt == '\0')
 212                                 badfmt(fu->fmt);
 213                         if (first_letter(fmt, "diouxX"))
 214                                 bcnt += 4;
 215                         else if (first_letter(fmt, "efgEG"))
 216                                 bcnt += 8;
 217                         else if (*fmt == 's')
 218                                 bcnt += prec;
 219                         else if (*fmt == 'c' || (*fmt == '_' && first_letter(++fmt, "cpu")))
 220                                 ++bcnt;
 221                         ++fmt;
 222                 }
 223                 cursize += bcnt * fu->reps;
 224         }
 225         return(cursize);
 226 }
 227
 228 void rewrite_rules(struct hexdump_fs *fs, struct hexdump *hex)
 229 {
 230         enum { NOTOKAY, USEBCNT, USEPREC } sokay;
 231         struct hexdump_pr *pr;
 232         struct hexdump_fu *fu;
 233         struct list_head *p, *q;
 234         char *p1, *p2, *fmtp;
 235         char savech, cs[4];
 236         int nconv, prec = 0;
 237
 238         list_for_each (p, &fs->fulist) {
 239                 fu = list_entry(p, struct hexdump_fu, fulist);
 240                 /*
 241                  * Break each format unit into print units; each
 242                  * conversion character gets its own.
 243                  */
 244                 nconv = 0;
 245                 fmtp = fu->fmt;
 246                 while (*fmtp) {
 247                         pr = xcalloc(1, sizeof(struct hexdump_pr));
 248                         INIT_LIST_HEAD(&pr->prlist);
 249                         list_add_tail(&pr->prlist, &fu->prlist);
 250
 251                         /* Skip preceding text and up to the next % sign. */
 252                         p1 = fmtp;
 253                         while (*p1 && *p1 != '%')
 254                                 ++p1;
 255
 256                         /* Only text in the string. */
 257                         if (!*p1) {
 258                                 pr->fmt = xstrdup(fmtp);
 259                                 pr->flags = F_TEXT;
 260                                 break;
 261                         }
 262
 263                         /*
 264                          * Get precision for %s -- if have a byte count, don't
 265                          * need it.
 266                          */
 267                         if (fu->bcnt) {
 268                                 sokay = USEBCNT;
 269                                 /* skip to conversion character */
 270                                 for (p1++; strchr(spec, *p1); p1++)
 271                                         ;
 272                         } else {
 273                                 /* skip any special chars, field width */
 274                                 while (strchr(spec + 1, *++p1))
 275                                         ;
 276                                 if (*p1 == '.' && isdigit(*++p1)) {
 277                                         sokay = USEPREC;
 278                                         p1 = next_number(p1, &prec);
 279                                 } else
 280                                         sokay = NOTOKAY;
 281                         }
 282
 283                         p2 = p1 + 1;            /* Set end pointer. */
 284                         cs[0] = *p1;            /* Set conversion string. */
 285                         cs[1] = 0;
 286
 287                         /*
 288                          * Figure out the byte count for each conversion;
 289                          * rewrite the format as necessary, set up blank-
 290                          * padding for end of data.
 291                          */
 292                         if (*cs == 'c') {
 293                                 pr->flags = F_CHAR;
 294                                 switch(fu->bcnt) {
 295                                         case 0:
 296                                         case 1:
 297                                                 pr->bcnt = 1;
 298                                                 break;
 299                                         default:
 300                                                 p1[1] = '\0';
 301                                                 badcnt(p1);
 302                                 }
 303                         } else if (first_letter(cs, "di")) {
 304                                 pr->flags = F_INT;
 305                                 goto isint;
 306                         } else if (first_letter(cs, "ouxX")) {
 307                                 pr->flags = F_UINT;
 308 isint:                          cs[3] = '\0';
 309                                 cs[2] = cs[0];
 310                                 cs[1] = 'l';
 311                                 cs[0] = 'l';
 312                                 switch(fu->bcnt) {
 313                                         case 0:
 314                                                 pr->bcnt = 4;
 315                                                 break;
 316                                         case 1:
 317                                         case 2:
 318                                         case 4:
 319                                         case 8:
 320                                                 pr->bcnt = fu->bcnt;
 321                                                 break;
 322                                         default:
 323                                                 p1[1] = '\0';
 324                                                 badcnt(p1);
 325                                 }
 326                         } else if (first_letter(cs, "efgEG")) {
 327                                 pr->flags = F_DBL;
 328                                 switch(fu->bcnt) {
 329                                         case 0:
 330                                                 pr->bcnt = 8;
 331                                                 break;
 332                                         case 4:
 333                                         case 8:
 334                                                 pr->bcnt = fu->bcnt;
 335                                                 break;
 336                                         default:
 337                                                 p1[1] = '\0';
 338                                                 badcnt(p1);
 339                                 }
 340                         } else if(*cs == 's') {
 341                                 pr->flags = F_STR;
 342                                 switch(sokay) {
 343                                         case NOTOKAY:
 344                                                 badsfmt();
 345                                         case USEBCNT:
 346                                                 pr->bcnt = fu->bcnt;
 347                                                 break;
 348                                         case USEPREC:
 349                                                 pr->bcnt = prec;
 350                                                 break;
 351                                 }
 352                         } else if (*cs == '_') {
 353                                 ++p2;
 354                                 switch(p1[1]) {
 355                                         case 'A':
 356                                                 endfu = fu;
 357                                                 fu->flags |= F_IGNORE;
 358                                                 /* fallthrough */
 359                                         case 'a':
 360                                                 pr->flags = F_ADDRESS;
 361                                                 ++p2;
 362                                                 if (first_letter(p1 + 2, "dox")) {
 363                                                         cs[0] = 'l';
 364                                                         cs[1] = 'l';
 365                                                         cs[2] = p1[2];
 366                                                         cs[3] = '\0';
 367                                                 } else {
 368                                                         p1[3] = '\0';
 369                                                         badconv(p1);
 370                                                 }
 371                                                 break;
 372                                         case 'c':
 373                                                 pr->flags = F_C;
 374                                                 /* cs[0] = 'c'; set in conv_c */
 375                                                 goto isint2;
 376                                         case 'p':
 377                                                 pr->flags = F_P;
 378                                                 cs[0] = 'c';
 379                                                 goto isint2;
 380                                         case 'u':
 381                                                 pr->flags = F_U;
 382                                                 /* cs[0] = 'c'; set in conv_u */
 383         isint2:                                 switch(fu->bcnt) {
 384                                                         case 0:
 385                                                         case 1:
 386                                                                 pr->bcnt = 1;
 387                                                                 break;
 388                                                         default:
 389                                                                 p1[2] = '\0';
 390                                                                 badcnt(p1);
 391                                                 }
 392                                                 break;
 393                                         default:
 394                                                 p1[2] = '\0';
 395                                                 badconv(p1);
 396                                 }
 397                         } else {
 398                                 p1[1] = '\0';
 399                                 badconv(p1);
 400                         }
 401
 402                         /* Color unit(s) specified */
 403                         if (*p2 == '_' && p2[1] == 'L') {
 404                                 if (colors_wanted()) {
 405                                         char *a;
 406
 407                                         /* "cut out" the color_unit(s) */
 408                                         a = strchr(p2, '[');
 409                                         p2 = strrchr(p2, ']');
 410                                         if (a++ && p2)
 411                                                 pr->colorlist = color_fmt(xstrndup(a, p2++ - a), pr->bcnt);
 412                                         else
 413                                                 badconv(p2);
 414                                 }
 415                                 /* we don't want colors, quietly skip over them */
 416                                 else {
 417                                         p2 = strrchr(p2, ']');
 418                                         /* be a bit louder if we don't know how to skip over them */
 419                                         if (!p2)
 420                                                 badconv("_L");
 421                                         ++p2;
 422                                 }
 423                         }
 424                         /*
 425                          * Copy to hexdump_pr format string, set conversion character
 426                          * pointer, update original.
 427                          */
 428                         savech = *p2;
 429                         p1[0] = '\0';
 430                         pr->fmt = xmalloc(strlen(fmtp) + strlen(cs) + 1);
 431                         strcpy(pr->fmt, fmtp);
 432                         strcat(pr->fmt, cs);
 433                         *p2 = savech;
 434                         pr->cchar = pr->fmt + (p1 - fmtp);
 435                         fmtp = p2;
 436
 437                         /* Only one conversion character if byte count */
 438                         if (!(pr->flags&F_ADDRESS) && fu->bcnt && nconv++)
 439                                 errx(EXIT_FAILURE,
 440                                     _("byte count with multiple conversion characters"));
 441                 }
 442                 /*
 443                  * If format unit byte count not specified, figure it out
 444                  * so can adjust rep count later.
 445                  */
 446                 if (!fu->bcnt)
 447                         list_for_each(q, &fu->prlist)
 448                                 fu->bcnt
 449                                   += (list_entry(q, struct hexdump_pr, prlist))->bcnt;
 450         }
 451         /*
 452          * If the format string interprets any data at all, and it's
 453          * not the same as the blocksize, and its last format unit
 454          * interprets any data at all, and has no iteration count,
 455          * repeat it as necessary.
 456          *
 457          * If rep count is greater than 1, no trailing whitespace
 458          * gets output from the last iteration of the format unit.
 459          */
 460         list_for_each (p, &fs->fulist) {
 461                 fu = list_entry(p, struct hexdump_fu, fulist);
 462
 463                 if (list_entry_is_last(&fu->fulist, &fs->fulist) &&
 464                         fs->bcnt < hex->blocksize &&
 465                         !(fu->flags&F_SETREP) && fu->bcnt)
 466                                 fu->reps += (hex->blocksize - fs->bcnt) / fu->bcnt;
 467                 if (fu->reps > 1 && !list_empty(&fu->prlist)) {
 468                         pr = list_last_entry(&fu->prlist, struct hexdump_pr, prlist);
 469                         if (!pr)
 470                                 continue;
 471                         for (p1 = pr->fmt, p2 = NULL; *p1; ++p1)
 472                                 p2 = isspace(*p1) ? p1 : NULL;
 473                         if (p2)
 474                                 pr->nospace = p2;
 475                 }
 476         }
 477 }
 478
 479 /* [!]color[:string|:hex_number|:oct_number][@offt|@offt_start-offt_end],... */
 480 static struct list_head *color_fmt(char *cfmt, int bcnt)
 481 {
 482         struct hexdump_clr *hc, *hcnext;
 483         struct list_head *ret_head;
 484         char *clr, *fmt;
 485
 486         ret_head = xmalloc(sizeof(struct list_head));
 487         hcnext = hc = xcalloc(1, sizeof(struct hexdump_clr));
 488
 489         INIT_LIST_HEAD(&hc->colorlist);
 490         INIT_LIST_HEAD(ret_head);
 491         list_add_tail(&hc->colorlist, ret_head);
 492
 493         fmt = cfmt;
 494         while (cfmt && *cfmt) {
 495                 char *end;
 496                 /* invert this condition */
 497                 if (*cfmt == '!') {
 498                         hcnext->invert = 1;
 499                         ++cfmt;
 500                 }
 501
 502                 clr = xstrndup(cfmt, strcspn(cfmt, ":@,"));
 503                 cfmt += strlen(clr);
 504                 hcnext->fmt = color_sequence_from_colorname(clr);
 505                 free(clr);
 506
 507                 if (!hcnext->fmt)
 508                         return NULL;
 509
 510                 /* only colorize this specific value */
 511                 if (*cfmt == ':') {
 512                         ++cfmt;
 513                         /* a hex or oct value */
 514                         if (*cfmt == '0') {
 515                                 /* hex */
 516                                 errno = 0;
 517                                 end = NULL;
 518                                 if (cfmt[1] == 'x' || cfmt[1] == 'X')
 519                                         hcnext->val = strtoul(cfmt + 2, &end, 16);
 520                                 else
 521                                         hcnext->val = strtoul(cfmt, &end, 8);
 522                                 if (errno || end == cfmt)
 523                                         badfmt(fmt);
 524                                 cfmt = end;
 525                         /* a string */
 526                         } else {
 527                                 off_t fmt_end;
 528                                 char endchar;
 529                                 char *endstr;
 530
 531                                 hcnext->val = -1;
 532                                 /* temporarily null-delimit the format, so we can reverse-search
 533                                  * for the start of an offset specifier */
 534                                 fmt_end = strcspn(cfmt, ",");
 535                                 endchar = cfmt[fmt_end];
 536                                 cfmt[fmt_end] = '\0';
 537                                 endstr = strrchr(cfmt, '@');
 538
 539                                 if (endstr) {
 540                                         if (endstr[1] != '\0')
 541                                                 --endstr;
 542                                         hcnext->str = xstrndup(cfmt, endstr - cfmt + 1);
 543                                 } else
 544                                         hcnext->str = xstrndup(cfmt, fmt_end);
 545
 546                                 /* restore the character */
 547                                 cfmt[fmt_end] = endchar;
 548                                 cfmt += strlen(hcnext->str);
 549                         }
 550
 551                 /* no specific value */
 552                 } else
 553                         hcnext->val = -1;
 554
 555                 /* only colorize at this offset */
 556                 hcnext->range = bcnt;
 557                 if (cfmt && *cfmt == '@') {
 558                         errno = 0;
 559                         hcnext->offt = strtoul(++cfmt, &cfmt, 10);
 560                         if (errno)
 561                                 badfmt(fmt);
 562
 563                         /* offset range */
 564                         if (*cfmt == '-') {
 565                                 ++cfmt;
 566                                 errno = 0;
 567
 568                                 hcnext->range =
 569                                   strtoul(cfmt, &cfmt, 10) - hcnext->offt + 1;
 570                                 if (errno)
 571                                         badfmt(fmt);
 572                                 /* offset range must be between 0 and format byte count */
 573                                 if (hcnext->range < 0)
 574                                         badcnt("_L");
 575                                 /* the offset extends over several print units, clone
 576                                  * the condition, link it in and adjust the address/offset */
 577                                 while (hcnext->range > bcnt) {
 578                                         hc = xcalloc(1, sizeof(struct hexdump_clr));
 579                                         memcpy(hc, hcnext, sizeof(struct hexdump_clr));
 580
 581                                         hc->range = bcnt;
 582
 583                                         INIT_LIST_HEAD(&hc->colorlist);
 584                                         list_add_tail(&hc->colorlist, ret_head);
 585
 586                                         hcnext->offt += bcnt;
 587                                         hcnext->range -= bcnt;
 588                                 }
 589                         }
 590                 /* no specific offset */
 591                 } else
 592                         hcnext->offt = (off_t)-1;
 593
 594                 /* check if the string we're looking for is the same length as the range */
 595                 if (hcnext->str && (int)strlen(hcnext->str) != hcnext->range)
 596                         badcnt("_L");
 597
 598                 /* link in another condition */
 599                 if (cfmt && *cfmt == ',') {
 600                         ++cfmt;
 601
 602                         hcnext = xcalloc(1, sizeof(struct hexdump_clr));
 603                         INIT_LIST_HEAD(&hcnext->colorlist);
 604                         list_add_tail(&hcnext->colorlist, ret_head);
 605                 }
 606         }
 607         return ret_head;
 608 }
 609
 610 static void escape(char *p1)
 611 {
 612         char *p2;
 613
 614         /* alphabetic escape sequences have to be done in place */
 615         p2 = p1;
 616         while (TRUE) {
 617                 if (!*p1) {
 618                         *p2 = *p1;
 619                         break;
 620                 }
 621                 if (*p1 == '\\')
 622                         switch(*++p1) {
 623                         case 'a':
 624                              /* *p2 = '\a'; */
 625                                 *p2 = '\007';
 626                                 break;
 627                         case 'b':
 628                                 *p2 = '\b';
 629                                 break;
 630                         case 'f':
 631                                 *p2 = '\f';
 632                                 break;
 633                         case 'n':
 634                                 *p2 = '\n';
 635                                 break;
 636                         case 'r':
 637                                 *p2 = '\r';
 638                                 break;
 639                         case 't':
 640                                 *p2 = '\t';
 641                                 break;
 642                         case 'v':
 643                                 *p2 = '\v';
 644                                 break;
 645                         default:
 646                                 *p2 = *p1;
 647                                 break;
 648                         }
 649                 ++p1; ++p2;
 650         }
 651 }