compiler/libjpeg/test/rdjpgcom.c

   1 /*
   2     $Id$
   3 */
   4
   5 /*
   6  * rdjpgcom.c
   7  *
   8  * Copyright (C) 1994-1997, Thomas G. Lane.
   9  * This file is part of the Independent JPEG Group's software.
  10  * For conditions of distribution and use, see the accompanying README file.
  11  *
  12  * This file contains a very simple stand-alone application that displays
  13  * the text in COM (comment) markers in a JFIF file.
  14  * This may be useful as an example of the minimum logic needed to parse
  15  * JPEG markers.
  16  */
  17
  18 #define JPEG_CJPEG_DJPEG        /* to get the command-line config symbols */
  19 #include "jinclude.h"           /* get auto-config symbols, <stdio.h> */
  20
  21 #include <ctype.h>              /* to declare isupper(), tolower() */
  22 #ifdef USE_SETMODE
  23 #include <fcntl.h>              /* to declare setmode()'s parameter macros */
  24 /* If you have setmode() but not <io.h>, just delete this line: */
  25 #include <io.h>                 /* to declare setmode() */
  26 #endif
  27
  28 #ifdef USE_CCOMMAND             /* command-line reader for Macintosh */
  29 #ifdef __MWERKS__
  30 #include <SIOUX.h>              /* Metrowerks needs this */
  31 #include <console.h>            /* ... and this */
  32 #endif
  33 #ifdef THINK_C
  34 #include <console.h>            /* Think declares it here */
  35 #endif
  36 #endif
  37
  38 #ifdef DONT_USE_B_MODE          /* define mode parameters for fopen() */
  39 #define READ_BINARY     "r"
  40 #else
  41 #ifdef VMS                      /* VMS is very nonstandard */
  42 #define READ_BINARY     "rb", "ctx=stm"
  43 #else                           /* standard ANSI-compliant case */
  44 #define READ_BINARY     "rb"
  45 #endif
  46 #endif
  47
  48 #ifndef EXIT_FAILURE            /* define exit() codes if not provided */
  49 #define EXIT_FAILURE  1
  50 #endif
  51 #ifndef EXIT_SUCCESS
  52 #ifdef VMS
  53 #define EXIT_SUCCESS  1         /* VMS is very nonstandard */
  54 #else
  55 #define EXIT_SUCCESS  0
  56 #endif
  57 #endif
  58
  59
  60 /*
  61  * These macros are used to read the input file.
  62  * To reuse this code in another application, you might need to change these.
  63  */
  64
  65 static FILE * infile;           /* input JPEG file */
  66
  67 /* Return next input byte, or EOF if no more */
  68 #define NEXTBYTE()  getc(infile)
  69
  70
  71 /* Error exit handler */
  72 #define ERREXIT(msg)  (fprintf(stderr, "%s\n", msg), exit(EXIT_FAILURE))
  73
  74
  75 /* Read one byte, testing for EOF */
  76 static int
  77 read_1_byte (void)
  78 {
  79   int c;
  80
  81   c = NEXTBYTE();
  82   if (c == EOF)
  83     ERREXIT("Premature EOF in JPEG file");
  84   return c;
  85 }
  86
  87 /* Read 2 bytes, convert to unsigned int */
  88 /* All 2-byte quantities in JPEG markers are MSB first */
  89 static unsigned int
  90 read_2_bytes (void)
  91 {
  92   int c1, c2;
  93
  94   c1 = NEXTBYTE();
  95   if (c1 == EOF)
  96     ERREXIT("Premature EOF in JPEG file");
  97   c2 = NEXTBYTE();
  98   if (c2 == EOF)
  99     ERREXIT("Premature EOF in JPEG file");
 100   return (((unsigned int) c1) << 8) + ((unsigned int) c2);
 101 }
 102
 103
 104 /*
 105  * JPEG markers consist of one or more 0xFF bytes, followed by a marker
 106  * code byte (which is not an FF).  Here are the marker codes of interest
 107  * in this program.  (See jdmarker.c for a more complete list.)
 108  */
 109
 110 #define M_SOF0  0xC0            /* Start Of Frame N */
 111 #define M_SOF1  0xC1            /* N indicates which compression process */
 112 #define M_SOF2  0xC2            /* Only SOF0-SOF2 are now in common use */
 113 #define M_SOF3  0xC3
 114 #define M_SOF5  0xC5            /* NB: codes C4 and CC are NOT SOF markers */
 115 #define M_SOF6  0xC6
 116 #define M_SOF7  0xC7
 117 #define M_SOF9  0xC9
 118 #define M_SOF10 0xCA
 119 #define M_SOF11 0xCB
 120 #define M_SOF13 0xCD
 121 #define M_SOF14 0xCE
 122 #define M_SOF15 0xCF
 123 #define M_SOI   0xD8            /* Start Of Image (beginning of datastream) */
 124 #define M_EOI   0xD9            /* End Of Image (end of datastream) */
 125 #define M_SOS   0xDA            /* Start Of Scan (begins compressed data) */
 126 #define M_APP0  0xE0            /* Application-specific marker, type N */
 127 #define M_APP12 0xEC            /* (we don't bother to list all 16 APPn's) */
 128 #define M_COM   0xFE            /* COMment */
 129
 130
 131 /*
 132  * Find the next JPEG marker and return its marker code.
 133  * We expect at least one FF byte, possibly more if the compressor used FFs
 134  * to pad the file.
 135  * There could also be non-FF garbage between markers.  The treatment of such
 136  * garbage is unspecified; we choose to skip over it but emit a warning msg.
 137  * NB: this routine must not be used after seeing SOS marker, since it will
 138  * not deal correctly with FF/00 sequences in the compressed image data...
 139  */
 140
 141 static int
 142 next_marker (void)
 143 {
 144   int c;
 145   int discarded_bytes = 0;
 146
 147   /* Find 0xFF byte; count and skip any non-FFs. */
 148   c = read_1_byte();
 149   while (c != 0xFF) {
 150     discarded_bytes++;
 151     c = read_1_byte();
 152   }
 153   /* Get marker code byte, swallowing any duplicate FF bytes.  Extra FFs
 154    * are legal as pad bytes, so don't count them in discarded_bytes.
 155    */
 156   do {
 157     c = read_1_byte();
 158   } while (c == 0xFF);
 159
 160   if (discarded_bytes != 0) {
 161     fprintf(stderr, "Warning: garbage data found in JPEG file\n");
 162   }
 163
 164   return c;
 165 }
 166
 167
 168 /*
 169  * Read the initial marker, which should be SOI.
 170  * For a JFIF file, the first two bytes of the file should be literally
 171  * 0xFF M_SOI.  To be more general, we could use next_marker, but if the
 172  * input file weren't actually JPEG at all, next_marker might read the whole
 173  * file and then return a misleading error message...
 174  */
 175
 176 static int
 177 first_marker (void)
 178 {
 179   int c1, c2;
 180
 181   c1 = NEXTBYTE();
 182   c2 = NEXTBYTE();
 183   if (c1 != 0xFF || c2 != M_SOI)
 184     ERREXIT("Not a JPEG file");
 185   return c2;
 186 }
 187
 188
 189 /*
 190  * Most types of marker are followed by a variable-length parameter segment.
 191  * This routine skips over the parameters for any marker we don't otherwise
 192  * want to process.
 193  * Note that we MUST skip the parameter segment explicitly in order not to
 194  * be fooled by 0xFF bytes that might appear within the parameter segment;
 195  * such bytes do NOT introduce new markers.
 196  */
 197
 198 static void
 199 skip_variable (void)
 200 /* Skip over an unknown or uninteresting variable-length marker */
 201 {
 202   unsigned int length;
 203
 204   /* Get the marker parameter length count */
 205   length = read_2_bytes();
 206   /* Length includes itself, so must be at least 2 */
 207   if (length < 2)
 208     ERREXIT("Erroneous JPEG marker length");
 209   length -= 2;
 210   /* Skip over the remaining bytes */
 211   while (length > 0) {
 212     (void) read_1_byte();
 213     length--;
 214   }
 215 }
 216
 217
 218 /*
 219  * Process a COM marker.
 220  * We want to print out the marker contents as legible text;
 221  * we must guard against non-text junk and varying newline representations.
 222  */
 223
 224 static void
 225 process_COM (void)
 226 {
 227   unsigned int length;
 228   int ch;
 229   int lastch = 0;
 230
 231   /* Get the marker parameter length count */
 232   length = read_2_bytes();
 233   /* Length includes itself, so must be at least 2 */
 234   if (length < 2)
 235     ERREXIT("Erroneous JPEG marker length");
 236   length -= 2;
 237
 238   while (length > 0) {
 239     ch = read_1_byte();
 240     /* Emit the character in a readable form.
 241      * Nonprintables are converted to \nnn form,
 242      * while \ is converted to \\.
 243      * Newlines in CR, CR/LF, or LF form will be printed as one newline.
 244      */
 245     if (ch == '\r') {
 246       printf("\n");
 247     } else if (ch == '\n') {
 248       if (lastch != '\r')
 249         printf("\n");
 250     } else if (ch == '\\') {
 251       printf("\\\\");
 252     } else if (isprint(ch)) {
 253       putc(ch, stdout);
 254     } else {
 255       printf("\\%03o", ch);
 256     }
 257     lastch = ch;
 258     length--;
 259   }
 260   printf("\n");
 261 }
 262
 263
 264 /*
 265  * Process a SOFn marker.
 266  * This code is only needed if you want to know the image dimensions...
 267  */
 268
 269 static void
 270 process_SOFn (int marker)
 271 {
 272   unsigned int length;
 273   unsigned int image_height, image_width;
 274   int data_precision, num_components;
 275   const char * process;
 276   int ci;
 277
 278   length = read_2_bytes();      /* usual parameter length count */
 279
 280   data_precision = read_1_byte();
 281   image_height = read_2_bytes();
 282   image_width = read_2_bytes();
 283   num_components = read_1_byte();
 284
 285   switch (marker) {
 286   case M_SOF0:  process = "Baseline";  break;
 287   case M_SOF1:  process = "Extended sequential";  break;
 288   case M_SOF2:  process = "Progressive";  break;
 289   case M_SOF3:  process = "Lossless";  break;
 290   case M_SOF5:  process = "Differential sequential";  break;
 291   case M_SOF6:  process = "Differential progressive";  break;
 292   case M_SOF7:  process = "Differential lossless";  break;
 293   case M_SOF9:  process = "Extended sequential, arithmetic coding";  break;
 294   case M_SOF10: process = "Progressive, arithmetic coding";  break;
 295   case M_SOF11: process = "Lossless, arithmetic coding";  break;
 296   case M_SOF13: process = "Differential sequential, arithmetic coding";  break;
 297   case M_SOF14: process = "Differential progressive, arithmetic coding"; break;
 298   case M_SOF15: process = "Differential lossless, arithmetic coding";  break;
 299   default:      process = "Unknown";  break;
 300   }
 301
 302   printf("JPEG image is %uw * %uh, %d color components, %d bits per sample\n",
 303          image_width, image_height, num_components, data_precision);
 304   printf("JPEG process: %s\n", process);
 305
 306   if (length != (unsigned int) (8 + num_components * 3))
 307     ERREXIT("Bogus SOF marker length");
 308
 309   for (ci = 0; ci < num_components; ci++) {
 310     (void) read_1_byte();       /* Component ID code */
 311     (void) read_1_byte();       /* H, V sampling factors */
 312     (void) read_1_byte();       /* Quantization table number */
 313   }
 314 }
 315
 316
 317 /*
 318  * Parse the marker stream until SOS or EOI is seen;
 319  * display any COM markers.
 320  * While the companion program wrjpgcom will always insert COM markers before
 321  * SOFn, other implementations might not, so we scan to SOS before stopping.
 322  * If we were only interested in the image dimensions, we would stop at SOFn.
 323  * (Conversely, if we only cared about COM markers, there would be no need
 324  * for special code to handle SOFn; we could treat it like other markers.)
 325  */
 326
 327 static int
 328 scan_JPEG_header (int verbose)
 329 {
 330   int marker;
 331
 332   /* Expect SOI at start of file */
 333   if (first_marker() != M_SOI)
 334     ERREXIT("Expected SOI marker first");
 335
 336   /* Scan miscellaneous markers until we reach SOS. */
 337   for (;;) {
 338     marker = next_marker();
 339     switch (marker) {
 340       /* Note that marker codes 0xC4, 0xC8, 0xCC are not, and must not be,
 341        * treated as SOFn.  C4 in particular is actually DHT.
 342        */
 343     case M_SOF0:                /* Baseline */
 344     case M_SOF1:                /* Extended sequential, Huffman */
 345     case M_SOF2:                /* Progressive, Huffman */
 346     case M_SOF3:                /* Lossless, Huffman */
 347     case M_SOF5:                /* Differential sequential, Huffman */
 348     case M_SOF6:                /* Differential progressive, Huffman */
 349     case M_SOF7:                /* Differential lossless, Huffman */
 350     case M_SOF9:                /* Extended sequential, arithmetic */
 351     case M_SOF10:               /* Progressive, arithmetic */
 352     case M_SOF11:               /* Lossless, arithmetic */
 353     case M_SOF13:               /* Differential sequential, arithmetic */
 354     case M_SOF14:               /* Differential progressive, arithmetic */
 355     case M_SOF15:               /* Differential lossless, arithmetic */
 356       if (verbose)
 357         process_SOFn(marker);
 358       else
 359         skip_variable();
 360       break;
 361
 362     case M_SOS:                 /* stop before hitting compressed data */
 363       return marker;
 364
 365     case M_EOI:                 /* in case it's a tables-only JPEG stream */
 366       return marker;
 367
 368     case M_COM:
 369       process_COM();
 370       break;
 371
 372     case M_APP12:
 373       /* Some digital camera makers put useful textual information into
 374        * APP12 markers, so we print those out too when in -verbose mode.
 375        */
 376       if (verbose) {
 377         printf("APP12 contains:\n");
 378         process_COM();
 379       } else
 380         skip_variable();
 381       break;
 382
 383     default:                    /* Anything else just gets skipped */
 384       skip_variable();          /* we assume it has a parameter count... */
 385       break;
 386     }
 387   } /* end loop */
 388 }
 389
 390
 391 /* Command line parsing code */
 392
 393 static const char * progname;   /* program name for error messages */
 394
 395
 396 static void
 397 usage (void)
 398 /* complain about bad command line */
 399 {
 400   fprintf(stderr, "rdjpgcom displays any textual comments in a JPEG file.\n");
 401
 402   fprintf(stderr, "Usage: %s [switches] [inputfile]\n", progname);
 403
 404   fprintf(stderr, "Switches (names may be abbreviated):\n");
 405   fprintf(stderr, "  -verbose    Also display dimensions of JPEG image\n");
 406
 407   exit(EXIT_FAILURE);
 408 }
 409
 410
 411 static int
 412 keymatch (char * arg, const char * keyword, int minchars)
 413 /* Case-insensitive matching of (possibly abbreviated) keyword switches. */
 414 /* keyword is the constant keyword (must be lower case already), */
 415 /* minchars is length of minimum legal abbreviation. */
 416 {
 417   register int ca, ck;
 418   register int nmatched = 0;
 419
 420   while ((ca = *arg++) != '\0') {
 421     if ((ck = *keyword++) == '\0')
 422       return 0;                 /* arg longer than keyword, no good */
 423     if (isupper(ca))            /* force arg to lcase (assume ck is already) */
 424       ca = tolower(ca);
 425     if (ca != ck)
 426       return 0;                 /* no good */
 427     nmatched++;                 /* count matched characters */
 428   }
 429   /* reached end of argument; fail if it's too short for unique abbrev */
 430   if (nmatched < minchars)
 431     return 0;
 432   return 1;                     /* A-OK */
 433 }
 434
 435
 436 /*
 437  * The main program.
 438  */
 439
 440 int
 441 main (int argc, char **argv)
 442 {
 443   int argn;
 444   char * arg;
 445   int verbose = 0;
 446
 447   /* On Mac, fetch a command line. */
 448 #ifdef USE_CCOMMAND
 449   argc = ccommand(&argv);
 450 #endif
 451
 452   progname = argv[0];
 453   if (progname == NULL || progname[0] == 0)
 454     progname = "rdjpgcom";      /* in case C library doesn't provide it */
 455
 456   /* Parse switches, if any */
 457   for (argn = 1; argn < argc; argn++) {
 458     arg = argv[argn];
 459     if (arg[0] != '-')
 460       break;                    /* not switch, must be file name */
 461     arg++;                      /* advance over '-' */
 462     if (keymatch(arg, "verbose", 1)) {
 463       verbose++;
 464     } else
 465       usage();
 466   }
 467
 468   /* Open the input file. */
 469   /* Unix style: expect zero or one file name */
 470   if (argn < argc-1) {
 471     fprintf(stderr, "%s: only one input file\n", progname);
 472     usage();
 473   }
 474   if (argn < argc) {
 475     if ((infile = fopen(argv[argn], READ_BINARY)) == NULL) {
 476       fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
 477       exit(EXIT_FAILURE);
 478     }
 479   } else {
 480     /* default input file is stdin */
 481 #ifdef USE_SETMODE              /* need to hack file mode? */
 482     setmode(fileno(stdin), O_BINARY);
 483 #endif
 484 #ifdef USE_FDOPEN               /* need to re-open in binary mode? */
 485     if ((infile = fdopen(fileno(stdin), READ_BINARY)) == NULL) {
 486       fprintf(stderr, "%s: can't open stdin\n", progname);
 487       exit(EXIT_FAILURE);
 488     }
 489 #else
 490     infile = stdin;
 491 #endif
 492   }
 493
 494   /* Scan the JPEG headers. */
 495   (void) scan_JPEG_header(verbose);
 496
 497   /* All done. */
 498   exit(EXIT_SUCCESS);
 499   return 0;                     /* suppress no-return-value warnings */
 500 }