1/*
2 * rdjpgcom.c
3 *
4 * This file was part of the Independent JPEG Group's software:
5 * Copyright (C) 1994-1997, Thomas G. Lane.
6 * Modified 2009 by Bill Allombert, Guido Vollbeding.
7 * It was modified by The libjpeg-turbo Project to include only code relevant
8 * to libjpeg-turbo.
9 * For conditions of distribution and use, see the accompanying README.ijg
10 * 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#ifdef HAVE_LOCALE_H
22#include <locale.h>             /* Bill Allombert: use locale for isprint */
23#endif
24#include <ctype.h>              /* to declare isupper(), tolower() */
25#ifdef USE_SETMODE
26#include <fcntl.h>              /* to declare setmode()'s parameter macros */
27/* If you have setmode() but not <io.h>, just delete this line: */
28#include <io.h>                 /* to declare setmode() */
29#endif
30
31#ifdef USE_CCOMMAND             /* command-line reader for Macintosh */
32#ifdef __MWERKS__
33#include <SIOUX.h>              /* Metrowerks needs this */
34#include <console.h>            /* ... and this */
35#endif
36#ifdef THINK_C
37#include <console.h>            /* Think declares it here */
38#endif
39#endif
40
41#ifdef DONT_USE_B_MODE          /* define mode parameters for fopen() */
42#define READ_BINARY     "r"
43#else
44#define READ_BINARY     "rb"
45#endif
46
47#ifndef EXIT_FAILURE            /* define exit() codes if not provided */
48#define EXIT_FAILURE  1
49#endif
50#ifndef EXIT_SUCCESS
51#define EXIT_SUCCESS  0
52#endif
53
54
55/*
56 * These macros are used to read the input file.
57 * To reuse this code in another application, you might need to change these.
58 */
59
60static FILE *infile;            /* input JPEG file */
61
62/* Return next input byte, or EOF if no more */
63#define NEXTBYTE()  getc(infile)
64
65
66/* Error exit handler */
67#define ERREXIT(msg)  (fprintf(stderr, "%s\n", msg), exit(EXIT_FAILURE))
68
69
70/* Read one byte, testing for EOF */
71static int
72read_1_byte (void)
73{
74  int c;
75
76  c = NEXTBYTE();
77  if (c == EOF)
78    ERREXIT("Premature EOF in JPEG file");
79  return c;
80}
81
82/* Read 2 bytes, convert to unsigned int */
83/* All 2-byte quantities in JPEG markers are MSB first */
84static unsigned int
85read_2_bytes (void)
86{
87  int c1, c2;
88
89  c1 = NEXTBYTE();
90  if (c1 == EOF)
91    ERREXIT("Premature EOF in JPEG file");
92  c2 = NEXTBYTE();
93  if (c2 == EOF)
94    ERREXIT("Premature EOF in JPEG file");
95  return (((unsigned int) c1) << 8) + ((unsigned int) c2);
96}
97
98
99/*
100 * JPEG markers consist of one or more 0xFF bytes, followed by a marker
101 * code byte (which is not an FF).  Here are the marker codes of interest
102 * in this program.  (See jdmarker.c for a more complete list.)
103 */
104
105#define M_SOF0  0xC0            /* Start Of Frame N */
106#define M_SOF1  0xC1            /* N indicates which compression process */
107#define M_SOF2  0xC2            /* Only SOF0-SOF2 are now in common use */
108#define M_SOF3  0xC3
109#define M_SOF5  0xC5            /* NB: codes C4 and CC are NOT SOF markers */
110#define M_SOF6  0xC6
111#define M_SOF7  0xC7
112#define M_SOF9  0xC9
113#define M_SOF10 0xCA
114#define M_SOF11 0xCB
115#define M_SOF13 0xCD
116#define M_SOF14 0xCE
117#define M_SOF15 0xCF
118#define M_SOI   0xD8            /* Start Of Image (beginning of datastream) */
119#define M_EOI   0xD9            /* End Of Image (end of datastream) */
120#define M_SOS   0xDA            /* Start Of Scan (begins compressed data) */
121#define M_APP0  0xE0            /* Application-specific marker, type N */
122#define M_APP12 0xEC            /* (we don't bother to list all 16 APPn's) */
123#define M_COM   0xFE            /* COMment */
124
125
126/*
127 * Find the next JPEG marker and return its marker code.
128 * We expect at least one FF byte, possibly more if the compressor used FFs
129 * to pad the file.
130 * There could also be non-FF garbage between markers.  The treatment of such
131 * garbage is unspecified; we choose to skip over it but emit a warning msg.
132 * NB: this routine must not be used after seeing SOS marker, since it will
133 * not deal correctly with FF/00 sequences in the compressed image data...
134 */
135
136static int
137next_marker (void)
138{
139  int c;
140  int discarded_bytes = 0;
141
142  /* Find 0xFF byte; count and skip any non-FFs. */
143  c = read_1_byte();
144  while (c != 0xFF) {
145    discarded_bytes++;
146    c = read_1_byte();
147  }
148  /* Get marker code byte, swallowing any duplicate FF bytes.  Extra FFs
149   * are legal as pad bytes, so don't count them in discarded_bytes.
150   */
151  do {
152    c = read_1_byte();
153  } while (c == 0xFF);
154
155  if (discarded_bytes != 0) {
156    fprintf(stderr, "Warning: garbage data found in JPEG file\n");
157  }
158
159  return c;
160}
161
162
163/*
164 * Read the initial marker, which should be SOI.
165 * For a JFIF file, the first two bytes of the file should be literally
166 * 0xFF M_SOI.  To be more general, we could use next_marker, but if the
167 * input file weren't actually JPEG at all, next_marker might read the whole
168 * file and then return a misleading error message...
169 */
170
171static int
172first_marker (void)
173{
174  int c1, c2;
175
176  c1 = NEXTBYTE();
177  c2 = NEXTBYTE();
178  if (c1 != 0xFF || c2 != M_SOI)
179    ERREXIT("Not a JPEG file");
180  return c2;
181}
182
183
184/*
185 * Most types of marker are followed by a variable-length parameter segment.
186 * This routine skips over the parameters for any marker we don't otherwise
187 * want to process.
188 * Note that we MUST skip the parameter segment explicitly in order not to
189 * be fooled by 0xFF bytes that might appear within the parameter segment;
190 * such bytes do NOT introduce new markers.
191 */
192
193static void
194skip_variable (void)
195/* Skip over an unknown or uninteresting variable-length marker */
196{
197  unsigned int length;
198
199  /* Get the marker parameter length count */
200  length = read_2_bytes();
201  /* Length includes itself, so must be at least 2 */
202  if (length < 2)
203    ERREXIT("Erroneous JPEG marker length");
204  length -= 2;
205  /* Skip over the remaining bytes */
206  while (length > 0) {
207    (void) read_1_byte();
208    length--;
209  }
210}
211
212
213/*
214 * Process a COM marker.
215 * We want to print out the marker contents as legible text;
216 * we must guard against non-text junk and varying newline representations.
217 */
218
219static void
220process_COM (int raw)
221{
222  unsigned int length;
223  int ch;
224  int lastch = 0;
225
226  /* Bill Allombert: set locale properly for isprint */
227#ifdef HAVE_LOCALE_H
228  setlocale(LC_CTYPE, "");
229#endif
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    if (raw) {
241      putc(ch, stdout);
242    /* Emit the character in a readable form.
243     * Nonprintables are converted to \nnn form,
244     * while \ is converted to \\.
245     * Newlines in CR, CR/LF, or LF form will be printed as one newline.
246     */
247    } else if (ch == '\r') {
248      printf("\n");
249    } else if (ch == '\n') {
250      if (lastch != '\r')
251        printf("\n");
252    } else if (ch == '\\') {
253      printf("\\\\");
254    } else if (isprint(ch)) {
255      putc(ch, stdout);
256    } else {
257      printf("\\%03o", ch);
258    }
259    lastch = ch;
260    length--;
261  }
262  printf("\n");
263
264  /* Bill Allombert: revert to C locale */
265#ifdef HAVE_LOCALE_H
266  setlocale(LC_CTYPE, "C");
267#endif
268}
269
270
271/*
272 * Process a SOFn marker.
273 * This code is only needed if you want to know the image dimensions...
274 */
275
276static void
277process_SOFn (int marker)
278{
279  unsigned int length;
280  unsigned int image_height, image_width;
281  int data_precision, num_components;
282  const char *process;
283  int ci;
284
285  length = read_2_bytes();      /* usual parameter length count */
286
287  data_precision = read_1_byte();
288  image_height = read_2_bytes();
289  image_width = read_2_bytes();
290  num_components = read_1_byte();
291
292  switch (marker) {
293  case M_SOF0:  process = "Baseline";  break;
294  case M_SOF1:  process = "Extended sequential";  break;
295  case M_SOF2:  process = "Progressive";  break;
296  case M_SOF3:  process = "Lossless";  break;
297  case M_SOF5:  process = "Differential sequential";  break;
298  case M_SOF6:  process = "Differential progressive";  break;
299  case M_SOF7:  process = "Differential lossless";  break;
300  case M_SOF9:  process = "Extended sequential, arithmetic coding";  break;
301  case M_SOF10: process = "Progressive, arithmetic coding";  break;
302  case M_SOF11: process = "Lossless, arithmetic coding";  break;
303  case M_SOF13: process = "Differential sequential, arithmetic coding";  break;
304  case M_SOF14: process = "Differential progressive, arithmetic coding"; break;
305  case M_SOF15: process = "Differential lossless, arithmetic coding";  break;
306  default:      process = "Unknown";  break;
307  }
308
309  printf("JPEG image is %uw * %uh, %d color components, %d bits per sample\n",
310         image_width, image_height, num_components, data_precision);
311  printf("JPEG process: %s\n", process);
312
313  if (length != (unsigned int) (8 + num_components * 3))
314    ERREXIT("Bogus SOF marker length");
315
316  for (ci = 0; ci < num_components; ci++) {
317    (void) read_1_byte();       /* Component ID code */
318    (void) read_1_byte();       /* H, V sampling factors */
319    (void) read_1_byte();       /* Quantization table number */
320  }
321}
322
323
324/*
325 * Parse the marker stream until SOS or EOI is seen;
326 * display any COM markers.
327 * While the companion program wrjpgcom will always insert COM markers before
328 * SOFn, other implementations might not, so we scan to SOS before stopping.
329 * If we were only interested in the image dimensions, we would stop at SOFn.
330 * (Conversely, if we only cared about COM markers, there would be no need
331 * for special code to handle SOFn; we could treat it like other markers.)
332 */
333
334static int
335scan_JPEG_header (int verbose, int raw)
336{
337  int marker;
338
339  /* Expect SOI at start of file */
340  if (first_marker() != M_SOI)
341    ERREXIT("Expected SOI marker first");
342
343  /* Scan miscellaneous markers until we reach SOS. */
344  for (;;) {
345    marker = next_marker();
346    switch (marker) {
347      /* Note that marker codes 0xC4, 0xC8, 0xCC are not, and must not be,
348       * treated as SOFn.  C4 in particular is actually DHT.
349       */
350    case M_SOF0:                /* Baseline */
351    case M_SOF1:                /* Extended sequential, Huffman */
352    case M_SOF2:                /* Progressive, Huffman */
353    case M_SOF3:                /* Lossless, Huffman */
354    case M_SOF5:                /* Differential sequential, Huffman */
355    case M_SOF6:                /* Differential progressive, Huffman */
356    case M_SOF7:                /* Differential lossless, Huffman */
357    case M_SOF9:                /* Extended sequential, arithmetic */
358    case M_SOF10:               /* Progressive, arithmetic */
359    case M_SOF11:               /* Lossless, arithmetic */
360    case M_SOF13:               /* Differential sequential, arithmetic */
361    case M_SOF14:               /* Differential progressive, arithmetic */
362    case M_SOF15:               /* Differential lossless, arithmetic */
363      if (verbose)
364        process_SOFn(marker);
365      else
366        skip_variable();
367      break;
368
369    case M_SOS:                 /* stop before hitting compressed data */
370      return marker;
371
372    case M_EOI:                 /* in case it's a tables-only JPEG stream */
373      return marker;
374
375    case M_COM:
376      process_COM(raw);
377      break;
378
379    case M_APP12:
380      /* Some digital camera makers put useful textual information into
381       * APP12 markers, so we print those out too when in -verbose mode.
382       */
383      if (verbose) {
384        printf("APP12 contains:\n");
385        process_COM(raw);
386      } else
387        skip_variable();
388      break;
389
390    default:                    /* Anything else just gets skipped */
391      skip_variable();          /* we assume it has a parameter count... */
392      break;
393    }
394  } /* end loop */
395}
396
397
398/* Command line parsing code */
399
400static const char *progname;    /* program name for error messages */
401
402
403static void
404usage (void)
405/* complain about bad command line */
406{
407  fprintf(stderr, "rdjpgcom displays any textual comments in a JPEG file.\n");
408
409  fprintf(stderr, "Usage: %s [switches] [inputfile]\n", progname);
410
411  fprintf(stderr, "Switches (names may be abbreviated):\n");
412  fprintf(stderr, "  -raw        Display non-printable characters in comments (unsafe)\n");
413  fprintf(stderr, "  -verbose    Also display dimensions of JPEG image\n");
414
415  exit(EXIT_FAILURE);
416}
417
418
419static int
420keymatch (char *arg, const char *keyword, int minchars)
421/* Case-insensitive matching of (possibly abbreviated) keyword switches. */
422/* keyword is the constant keyword (must be lower case already), */
423/* minchars is length of minimum legal abbreviation. */
424{
425  register int ca, ck;
426  register int nmatched = 0;
427
428  while ((ca = *arg++) != '\0') {
429    if ((ck = *keyword++) == '\0')
430      return 0;                 /* arg longer than keyword, no good */
431    if (isupper(ca))            /* force arg to lcase (assume ck is already) */
432      ca = tolower(ca);
433    if (ca != ck)
434      return 0;                 /* no good */
435    nmatched++;                 /* count matched characters */
436  }
437  /* reached end of argument; fail if it's too short for unique abbrev */
438  if (nmatched < minchars)
439    return 0;
440  return 1;                     /* A-OK */
441}
442
443
444/*
445 * The main program.
446 */
447
448int
449main (int argc, char **argv)
450{
451  int argn;
452  char *arg;
453  int verbose = 0, raw = 0;
454
455  /* On Mac, fetch a command line. */
456#ifdef USE_CCOMMAND
457  argc = ccommand(&argv);
458#endif
459
460  progname = argv[0];
461  if (progname == NULL || progname[0] == 0)
462    progname = "rdjpgcom";      /* in case C library doesn't provide it */
463
464  /* Parse switches, if any */
465  for (argn = 1; argn < argc; argn++) {
466    arg = argv[argn];
467    if (arg[0] != '-')
468      break;                    /* not switch, must be file name */
469    arg++;                      /* advance over '-' */
470    if (keymatch(arg, "verbose", 1)) {
471      verbose++;
472    } else if (keymatch(arg, "raw", 1)) {
473      raw = 1;
474    } else
475      usage();
476  }
477
478  /* Open the input file. */
479  /* Unix style: expect zero or one file name */
480  if (argn < argc-1) {
481    fprintf(stderr, "%s: only one input file\n", progname);
482    usage();
483  }
484  if (argn < argc) {
485    if ((infile = fopen(argv[argn], READ_BINARY)) == NULL) {
486      fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
487      exit(EXIT_FAILURE);
488    }
489  } else {
490    /* default input file is stdin */
491#ifdef USE_SETMODE              /* need to hack file mode? */
492    setmode(fileno(stdin), O_BINARY);
493#endif
494#ifdef USE_FDOPEN               /* need to re-open in binary mode? */
495    if ((infile = fdopen(fileno(stdin), READ_BINARY)) == NULL) {
496      fprintf(stderr, "%s: can't open stdin\n", progname);
497      exit(EXIT_FAILURE);
498    }
499#else
500    infile = stdin;
501#endif
502  }
503
504  /* Scan the JPEG headers. */
505  (void) scan_JPEG_header(verbose, raw);
506
507  /* All done. */
508  exit(EXIT_SUCCESS);
509  return 0;                     /* suppress no-return-value warnings */
510}
511