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;
  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 void add_fmt(const char *fmt, struct hexdump *hex)
 102 {
 103         const char *p, *savep;
 104         struct hexdump_fs *tfs;
 105         struct hexdump_fu *tfu;
 106
 107         /* Start new linked list of format units. */
 108         tfs = xcalloc(1, sizeof(struct hexdump_fs));
 109         INIT_LIST_HEAD(&tfs->fslist);
 110         INIT_LIST_HEAD(&tfs->fulist);
 111         list_add_tail(&tfs->fslist, &hex->fshead);
 112
 113         /* Take the format string and break it up into format units. */
 114         p = fmt;
 115         while (TRUE) {
 116                 /* Skip leading white space. */
 117                 if (!*(p = skip_space(p)))
 118                         break;
 119
 120                 /* Allocate a new format unit and link it in. */
 121                 tfu = xcalloc(1, sizeof(struct hexdump_fu));
 122                 tfu->reps = 1;
 123
 124                 INIT_LIST_HEAD(&tfu->fulist);
 125                 INIT_LIST_HEAD(&tfu->prlist);
 126                 list_add_tail(&tfu->fulist, &tfs->fulist);
 127
 128                 /* If leading digit, repetition count. */
 129                 if (isdigit(*p)) {
 130                         savep = p;
 131                         while (isdigit(*p) && ++p)
 132                                 ;
 133                         if (!isspace(*p) && *p != '/')
 134                                 badfmt(fmt);
 135                         /* may overwrite either white space or slash */
 136                         tfu->reps = atoi(savep);
 137                         tfu->flags = F_SETREP;
 138                         /* skip trailing white space */
 139                         p = skip_space(++p);
 140                 }
 141
 142                 /* Skip slash and trailing white space. */
 143                 if (*p == '/')
 144                         p = skip_space(p);
 145
 146                 /* byte count */
 147                 if (isdigit(*p)) {
 148                         savep = p;
 149                         while (isdigit(*p) && ++p)
 150                                 ;
 151                         if (!isspace(*p))
 152                                 badfmt(fmt);
 153                         tfu->bcnt = atoi(savep);
 154                         /* skip trailing white space */
 155                         p = skip_space(++p);
 156                 }
 157
 158                 /* format */
 159                 if (*p != '"')
 160                         badfmt(fmt);
 161                 savep = ++p;
 162                 while (*p != '"') {
 163                         if (!*p++)
 164                                 badfmt(fmt);
 165                 }
 166                 tfu->fmt = xmalloc(p - savep + 1);
 167                 xstrncpy(tfu->fmt, savep, p - savep + 1);
 168                 escape(tfu->fmt);
 169                 ++p;
 170         }
 171 }
 172
 173 static const char *spec = ".#-+ 0123456789";
 174
 175 int block_size(struct hexdump_fs *fs)
 176 {
 177         struct hexdump_fu *fu;
 178         int bcnt, prec, cursize = 0;
 179         char *fmt;
 180         struct list_head *p;
 181
 182         /* figure out the data block size needed for each format unit */
 183         list_for_each (p, &fs->fulist) {
 184                 fu = list_entry(p, struct hexdump_fu, fulist);
 185                 if (fu->bcnt) {
 186                         cursize += fu->bcnt * fu->reps;
 187                         continue;
 188                 }
 189                 bcnt = prec = 0;
 190                 fmt = fu->fmt;
 191                 while (*fmt) {
 192                         if (*fmt != '%') {
 193                                 ++fmt;
 194                                 continue;
 195                         }
 196                         /*
 197                          * skip any special chars -- save precision in
 198                          * case it's a %s format.
 199                          */
 200                         while (strchr(spec + 1, *++fmt))
 201                                 ;
 202                         if (*fmt == '.' && isdigit(*++fmt)) {
 203                                 prec = atoi(fmt);
 204                                 while (isdigit(*++fmt))
 205                                         ;
 206                         }
 207                         if (first_letter(fmt, "diouxX"))
 208                                 bcnt += 4;
 209                         else if (first_letter(fmt, "efgEG"))
 210                                 bcnt += 8;
 211                         else if (*fmt == 's')
 212                                 bcnt += prec;
 213                         else if (*fmt == 'c' || (*fmt == '_' && first_letter(++fmt, "cpu")))
 214                                 ++bcnt;
 215                         ++fmt;
 216                 }
 217                 cursize += bcnt * fu->reps;
 218         }
 219         return(cursize);
 220 }
 221
 222 void rewrite_rules(struct hexdump_fs *fs, struct hexdump *hex)
 223 {
 224         enum { NOTOKAY, USEBCNT, USEPREC } sokay;
 225         struct hexdump_pr *pr;
 226         struct hexdump_fu *fu;
 227         struct list_head *p, *q;
 228         char *p1, *p2, *fmtp;
 229         char savech, cs[3];
 230         int nconv, prec = 0;
 231
 232         list_for_each (p, &fs->fulist) {
 233                 fu = list_entry(p, struct hexdump_fu, fulist);
 234                 /*
 235                  * Break each format unit into print units; each
 236                  * conversion character gets its own.
 237                  */
 238                 nconv = 0;
 239                 fmtp = fu->fmt;
 240                 while (*fmtp) {
 241                         pr = xcalloc(1, sizeof(struct hexdump_pr));
 242                         INIT_LIST_HEAD(&pr->prlist);
 243                         list_add_tail(&pr->prlist, &fu->prlist);
 244
 245                         /* Skip preceding text and up to the next % sign. */
 246                         p1 = fmtp;
 247                         while (*p1 && *p1 != '%')
 248                                 ++p1;
 249
 250                         /* Only text in the string. */
 251                         if (!*p1) {
 252                                 pr->fmt = xstrdup(fmtp);
 253                                 pr->flags = F_TEXT;
 254                                 break;
 255                         }
 256
 257                         /*
 258                          * Get precision for %s -- if have a byte count, don't
 259                          * need it.
 260                          */
 261                         if (fu->bcnt) {
 262                                 sokay = USEBCNT;
 263                                 /* skip to conversion character */
 264                                 while (++p1 && strchr(spec, *p1))
 265                                         ;
 266                         } else {
 267                                 /* skip any special chars, field width */
 268                                 while (strchr(spec + 1, *++p1))
 269                                         ;
 270                                 if (*p1 == '.' && isdigit(*++p1)) {
 271                                         sokay = USEPREC;
 272                                         prec = atoi(p1);
 273                                         while (isdigit(*++p1))
 274                                                 ;
 275                                 } else
 276                                         sokay = NOTOKAY;
 277                         }
 278
 279                         p2 = p1 + 1;            /* Set end pointer. */
 280                         cs[0] = *p1;            /* Set conversion string. */
 281                         cs[1] = 0;
 282
 283                         /*
 284                          * Figure out the byte count for each conversion;
 285                          * rewrite the format as necessary, set up blank-
 286                          * padding for end of data.
 287                          */
 288                         if (*cs == 'c') {
 289                                 pr->flags = F_CHAR;
 290                                 switch(fu->bcnt) {
 291                                         case 0:
 292                                         case 1:
 293                                                 pr->bcnt = 1;
 294                                                 break;
 295                                         default:
 296                                                 p1[1] = '\0';
 297                                                 badcnt(p1);
 298                                 }
 299                         } else if (first_letter(cs, "di")) {
 300                                 pr->flags = F_INT;
 301                                 goto isint;
 302                         } else if (first_letter(cs, "ouxX")) {
 303                                 pr->flags = F_UINT;
 304 isint:                          cs[2] = '\0';
 305                                 cs[1] = cs[0];
 306                                 cs[0] = 'q';
 307                                 switch(fu->bcnt) {
 308                                         case 0:
 309                                                 pr->bcnt = 4;
 310                                                 break;
 311                                         case 1:
 312                                         case 2:
 313                                         case 4:
 314                                         case 8:
 315                                                 pr->bcnt = fu->bcnt;
 316                                                 break;
 317                                         default:
 318                                                 p1[1] = '\0';
 319                                                 badcnt(p1);
 320                                 }
 321                         } else if (first_letter(cs, "efgEG")) {
 322                                 pr->flags = F_DBL;
 323                                 switch(fu->bcnt) {
 324                                         case 0:
 325                                                 pr->bcnt = 8;
 326                                                 break;
 327                                         case 4:
 328                                         case 8:
 329                                                 pr->bcnt = fu->bcnt;
 330                                                 break;
 331                                         default:
 332                                                 p1[1] = '\0';
 333                                                 badcnt(p1);
 334                                 }
 335                         } else if(*cs == 's') {
 336                                 pr->flags = F_STR;
 337                                 switch(sokay) {
 338                                         case NOTOKAY:
 339                                                 badsfmt();
 340                                         case USEBCNT:
 341                                                 pr->bcnt = fu->bcnt;
 342                                                 break;
 343                                         case USEPREC:
 344                                                 pr->bcnt = prec;
 345                                                 break;
 346                                 }
 347                         } else if (*cs == '_') {
 348                                 ++p2;
 349                                 switch(p1[1]) {
 350                                         case 'A':
 351                                                 endfu = fu;
 352                                                 fu->flags |= F_IGNORE;
 353                                                 /* FALLTHROUGH */
 354                                         case 'a':
 355                                                 pr->flags = F_ADDRESS;
 356                                                 ++p2;
 357                                                 if (first_letter(p1 + 2, "dox")) {
 358                                                         cs[0] = 'q';
 359                                                         cs[1] = p1[2];
 360                                                         cs[2] = '\0';
 361                                                 } else {
 362                                                         p1[3] = '\0';
 363                                                         badconv(p1);
 364                                                 }
 365                                                 break;
 366                                         case 'c':
 367                                                 pr->flags = F_C;
 368                                                 /* cs[0] = 'c'; set in conv_c */
 369                                                 goto isint2;
 370                                         case 'p':
 371                                                 pr->flags = F_P;
 372                                                 cs[0] = 'c';
 373                                                 goto isint2;
 374                                         case 'u':
 375                                                 pr->flags = F_U;
 376                                                 /* cs[0] = 'c'; set in conv_u */
 377         isint2:                                 switch(fu->bcnt) {
 378                                                         case 0:
 379                                                         case 1:
 380                                                                 pr->bcnt = 1;
 381                                                                 break;
 382                                                         default:
 383                                                                 p1[2] = '\0';
 384                                                                 badcnt(p1);
 385                                                 }
 386                                                 break;
 387                                         default:
 388                                                 p1[2] = '\0';
 389                                                 badconv(p1);
 390                                 }
 391                         } else {
 392                                 p1[1] = '\0';
 393                                 badconv(p1);
 394                         }
 395
 396                         /* Color unit(s) specified */
 397                         if (*p2 == '_' && p2[1] == 'L') {
 398                                 if (colors_wanted()) {
 399                                         char *a;
 400
 401                                         /* "cut out" the color_unit(s) */
 402                                         a = strchr(p2, '[');
 403                                         p2 = strrchr(p2, ']');
 404                                         if (a++ && p2)
 405                                                 pr->colorlist = color_fmt(xstrndup(a, p2++ - a), pr->bcnt);
 406                                         else
 407                                                 badconv(p2);
 408                                 }
 409                                 /* we don't want colors, quietly skip over them */
 410                                 else {
 411                                         p2 = strrchr(p2, ']');
 412                                         /* be a bit louder if we don't know how to skip over them */
 413                                         if (!p2)
 414                                                 badconv("_L");
 415                                         ++p2;
 416                                 }
 417                         }
 418                         /*
 419                          * Copy to hexdump_pr format string, set conversion character
 420                          * pointer, update original.
 421                          */
 422                         savech = *p2;
 423                         p1[0] = '\0';
 424                         pr->fmt = xmalloc(strlen(fmtp) + strlen(cs) + 1);
 425                         strcpy(pr->fmt, fmtp);
 426                         strcat(pr->fmt, cs);
 427                         *p2 = savech;
 428                         pr->cchar = pr->fmt + (p1 - fmtp);
 429                         fmtp = p2;
 430
 431                         /* Only one conversion character if byte count */
 432                         if (!(pr->flags&F_ADDRESS) && fu->bcnt && nconv++)
 433                                 errx(EXIT_FAILURE,
 434                                     _("byte count with multiple conversion characters"));
 435                 }
 436                 /*
 437                  * If format unit byte count not specified, figure it out
 438                  * so can adjust rep count later.
 439                  */
 440                 if (!fu->bcnt)
 441                         list_for_each(q, &fu->prlist)
 442                                 fu->bcnt
 443                                   += (list_entry(q, struct hexdump_pr, prlist))->bcnt;
 444         }
 445         /*
 446          * If the format string interprets any data at all, and it's
 447          * not the same as the blocksize, and its last format unit
 448          * interprets any data at all, and has no iteration count,
 449          * repeat it as necessary.
 450          *
 451          * If rep count is greater than 1, no trailing whitespace
 452          * gets output from the last iteration of the format unit.
 453          */
 454         list_for_each (p, &fs->fulist) {
 455                 fu = list_entry(p, struct hexdump_fu, fulist);
 456
 457                 if (list_entry_is_last(&fu->fulist, &fs->fulist) &&
 458                         fs->bcnt < hex->blocksize &&
 459                         !(fu->flags&F_SETREP) && fu->bcnt)
 460                                 fu->reps += (hex->blocksize - fs->bcnt) / fu->bcnt;
 461                 if (fu->reps > 1) {
 462                         if (!list_empty(&fu->prlist)) {
 463                                 pr = list_last_entry(&fu->prlist,
 464                                   struct hexdump_pr, prlist);
 465                                 for (p1 = pr->fmt, p2 = NULL; *p1; ++p1)
 466                                         p2 = isspace(*p1) ? p1 : NULL;
 467                                 if (p2)
 468                                         pr->nospace = p2;
 469                         }
 470                 }
 471         }
 472 }
 473
 474 /* [!]color[:string|:hex_number|:oct_number][@offt|@offt_start-offt_end],... */
 475 static struct list_head *color_fmt(char *cfmt, int bcnt)
 476 {
 477         struct hexdump_clr *hc, *hcnext;
 478         struct list_head *ret_head;
 479         char *clr, *fmt;
 480
 481         ret_head = xmalloc(sizeof(struct list_head));
 482         hcnext = hc = xcalloc(1, sizeof(struct hexdump_clr));
 483
 484         INIT_LIST_HEAD(&hc->colorlist);
 485         INIT_LIST_HEAD(ret_head);
 486         list_add_tail(&hc->colorlist, ret_head);
 487
 488         fmt = cfmt;
 489         while (cfmt && *cfmt) {
 490                 char *end;
 491                 /* invert this condition */
 492                 if (*cfmt == '!') {
 493                         hcnext->invert = 1;
 494                         ++cfmt;
 495                 }
 496
 497                 clr = xstrndup(cfmt, strcspn(cfmt, ":@,"));
 498                 cfmt += strlen(clr);
 499                 hcnext->fmt = colorscheme_from_string(clr);
 500                 free(clr);
 501
 502                 if (!hcnext->fmt)
 503                         return NULL;
 504
 505                 /* only colorize this specific value */
 506                 if (*cfmt == ':') {
 507                         ++cfmt;
 508                         /* a hex or oct value */
 509                         if (*cfmt == '0') {
 510                                 /* hex */
 511                                 errno = 0;
 512                                 end = NULL;
 513                                 if (cfmt[1] == 'x' || cfmt[1] == 'X')
 514                                         hcnext->val = strtoul(cfmt + 2, &end, 16);
 515                                 else
 516                                         hcnext->val = strtoul(cfmt, &end, 8);
 517                                 if (errno || end == cfmt)
 518                                         badfmt(fmt);
 519                                 cfmt = end;
 520                         /* a string */
 521                         } else {
 522                                 off_t fmt_end;
 523                                 char endchar;
 524                                 char *endstr;
 525
 526                                 hcnext->val = -1;
 527                                 /* temporarily null-delimit the format, so we can reverse-search
 528                                  * for the start of an offset specifier */
 529                                 fmt_end = strcspn(cfmt, ",");
 530                                 endchar = cfmt[fmt_end];
 531                                 cfmt[fmt_end] = '\0';
 532                                 endstr = strrchr(cfmt, '@');
 533
 534                                 if (endstr) {
 535                                         if (endstr[1] != '\0')
 536                                                 --endstr;
 537                                         hcnext->str = xstrndup(cfmt, endstr - cfmt + 1);
 538                                 } else
 539                                         hcnext->str = xstrndup(cfmt, fmt_end);
 540
 541                                 /* restore the character */
 542                                 cfmt[fmt_end] = endchar;
 543                                 cfmt += strlen(hcnext->str);
 544                         }
 545
 546                 /* no specific value */
 547                 } else
 548                         hcnext->val = -1;
 549
 550                 /* only colorize at this offset */
 551                 hcnext->range = bcnt;
 552                 if (cfmt && *cfmt == '@') {
 553                         errno = 0;
 554                         hcnext->offt = strtoul(++cfmt, &cfmt, 10);
 555                         if (errno)
 556                                 badfmt(fmt);
 557
 558                         /* offset range */
 559                         if (*cfmt == '-') {
 560                                 ++cfmt;
 561                                 errno = 0;
 562
 563                                 hcnext->range =
 564                                   strtoul(cfmt, &cfmt, 10) - hcnext->offt + 1;
 565                                 if (errno)
 566                                         badfmt(fmt);
 567                                 /* offset range must be between 0 and format byte count */
 568                                 if (!(hcnext->range >= 0 && hcnext->range <= bcnt))
 569                                         badcnt("_L");
 570                         }
 571                 /* no specific offset */
 572                 } else
 573                         hcnext->offt = (off_t)-1;
 574
 575                 /* check if the string we're looking for is the same length as the range */
 576                 if (hcnext->str && (int)strlen(hcnext->str) != hcnext->range)
 577                         badcnt("_L");
 578
 579                 /* link in another condition */
 580                 if (cfmt && *cfmt == ',') {
 581                         ++cfmt;
 582
 583                         hcnext = xcalloc(1, sizeof(struct hexdump_clr));
 584                         INIT_LIST_HEAD(&hcnext->colorlist);
 585                         list_add_tail(&hcnext->colorlist, ret_head);
 586                 }
 587         }
 588         return ret_head;
 589 }
 590
 591 static void escape(char *p1)
 592 {
 593         char *p2;
 594
 595         /* alphabetic escape sequences have to be done in place */
 596         p2 = p1;
 597         while (TRUE) {
 598                 if (!*p1) {
 599                         *p2 = *p1;
 600                         break;
 601                 }
 602                 if (*p1 == '\\')
 603                         switch(*++p1) {
 604                         case 'a':
 605                              /* *p2 = '\a'; */
 606                                 *p2 = '\007';
 607                                 break;
 608                         case 'b':
 609                                 *p2 = '\b';
 610                                 break;
 611                         case 'f':
 612                                 *p2 = '\f';
 613                                 break;
 614                         case 'n':
 615                                 *p2 = '\n';
 616                                 break;
 617                         case 'r':
 618                                 *p2 = '\r';
 619                                 break;
 620                         case 't':
 621                                 *p2 = '\t';
 622                                 break;
 623                         case 'v':
 624                                 *p2 = '\v';
 625                                 break;
 626                         default:
 627                                 *p2 = *p1;
 628                                 break;
 629                         }
 630                 ++p1; ++p2;
 631         }
 632 }