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 file.
10 *
11 * This file contains a very simple stand-alone application that displays
12 * the text in COM (comment) markers in a JFIF file.
13 * This may be useful as an example of the minimum logic needed to parse
14 * JPEG markers.
15 */
16
17 #define JPEG_CJPEG_DJPEG /* to get the command-line config symbols */
18 #include "jinclude.h" /* get auto-config symbols, <stdio.h> */
19
20 #ifdef HAVE_LOCALE_H
21 #include <locale.h> /* Bill Allombert: use locale for isprint */
22 #endif
23 #include <ctype.h> /* to declare isupper(), tolower() */
24 #ifdef USE_SETMODE
25 #include <fcntl.h> /* to declare setmode()'s parameter macros */
26 /* If you have setmode() but not <io.h>, just delete this line: */
27 #include <io.h> /* to declare setmode() */
28 #endif
29
30 #ifdef USE_CCOMMAND /* command-line reader for Macintosh */
31 #ifdef __MWERKS__
32 #include <SIOUX.h> /* Metrowerks needs this */
33 #include <console.h> /* ... and this */
34 #endif
35 #ifdef THINK_C
36 #include <console.h> /* Think declares it here */
37 #endif
38 #endif
39
40 #ifdef DONT_USE_B_MODE /* define mode parameters for fopen() */
41 #define READ_BINARY "r"
42 #else
43 #define READ_BINARY "rb"
44 #endif
45
46 #ifndef EXIT_FAILURE /* define exit() codes if not provided */
47 #define EXIT_FAILURE 1
48 #endif
49 #ifndef EXIT_SUCCESS
50 #define EXIT_SUCCESS 0
51 #endif
52
53
54 /*
55 * These macros are used to read the input file.
56 * To reuse this code in another application, you might need to change these.
57 */
58
59 static FILE * infile; /* input JPEG file */
60
61 /* Return next input byte, or EOF if no more */
62 #define NEXTBYTE() getc(infile)
63
64
65 /* Error exit handler */
66 #define ERREXIT(msg) (fprintf(stderr, "%s\n", msg), exit(EXIT_FAILURE))
67
68
69 /* Read one byte, testing for EOF */
70 static int
read_1_byte(void)71 read_1_byte (void)
72 {
73 int c;
74
75 c = NEXTBYTE();
76 if (c == EOF)
77 ERREXIT("Premature EOF in JPEG file");
78 return c;
79 }
80
81 /* Read 2 bytes, convert to unsigned int */
82 /* All 2-byte quantities in JPEG markers are MSB first */
83 static unsigned int
read_2_bytes(void)84 read_2_bytes (void)
85 {
86 int c1, c2;
87
88 c1 = NEXTBYTE();
89 if (c1 == EOF)
90 ERREXIT("Premature EOF in JPEG file");
91 c2 = NEXTBYTE();
92 if (c2 == EOF)
93 ERREXIT("Premature EOF in JPEG file");
94 return (((unsigned int) c1) << 8) + ((unsigned int) c2);
95 }
96
97
98 /*
99 * JPEG markers consist of one or more 0xFF bytes, followed by a marker
100 * code byte (which is not an FF). Here are the marker codes of interest
101 * in this program. (See jdmarker.c for a more complete list.)
102 */
103
104 #define M_SOF0 0xC0 /* Start Of Frame N */
105 #define M_SOF1 0xC1 /* N indicates which compression process */
106 #define M_SOF2 0xC2 /* Only SOF0-SOF2 are now in common use */
107 #define M_SOF3 0xC3
108 #define M_SOF5 0xC5 /* NB: codes C4 and CC are NOT SOF markers */
109 #define M_SOF6 0xC6
110 #define M_SOF7 0xC7
111 #define M_SOF9 0xC9
112 #define M_SOF10 0xCA
113 #define M_SOF11 0xCB
114 #define M_SOF13 0xCD
115 #define M_SOF14 0xCE
116 #define M_SOF15 0xCF
117 #define M_SOI 0xD8 /* Start Of Image (beginning of datastream) */
118 #define M_EOI 0xD9 /* End Of Image (end of datastream) */
119 #define M_SOS 0xDA /* Start Of Scan (begins compressed data) */
120 #define M_APP0 0xE0 /* Application-specific marker, type N */
121 #define M_APP12 0xEC /* (we don't bother to list all 16 APPn's) */
122 #define M_COM 0xFE /* COMment */
123
124
125 /*
126 * Find the next JPEG marker and return its marker code.
127 * We expect at least one FF byte, possibly more if the compressor used FFs
128 * to pad the file.
129 * There could also be non-FF garbage between markers. The treatment of such
130 * garbage is unspecified; we choose to skip over it but emit a warning msg.
131 * NB: this routine must not be used after seeing SOS marker, since it will
132 * not deal correctly with FF/00 sequences in the compressed image data...
133 */
134
135 static int
next_marker(void)136 next_marker (void)
137 {
138 int c;
139 int discarded_bytes = 0;
140
141 /* Find 0xFF byte; count and skip any non-FFs. */
142 c = read_1_byte();
143 while (c != 0xFF) {
144 discarded_bytes++;
145 c = read_1_byte();
146 }
147 /* Get marker code byte, swallowing any duplicate FF bytes. Extra FFs
148 * are legal as pad bytes, so don't count them in discarded_bytes.
149 */
150 do {
151 c = read_1_byte();
152 } while (c == 0xFF);
153
154 if (discarded_bytes != 0) {
155 fprintf(stderr, "Warning: garbage data found in JPEG file\n");
156 }
157
158 return c;
159 }
160
161
162 /*
163 * Read the initial marker, which should be SOI.
164 * For a JFIF file, the first two bytes of the file should be literally
165 * 0xFF M_SOI. To be more general, we could use next_marker, but if the
166 * input file weren't actually JPEG at all, next_marker might read the whole
167 * file and then return a misleading error message...
168 */
169
170 static int
first_marker(void)171 first_marker (void)
172 {
173 int c1, c2;
174
175 c1 = NEXTBYTE();
176 c2 = NEXTBYTE();
177 if (c1 != 0xFF || c2 != M_SOI)
178 ERREXIT("Not a JPEG file");
179 return c2;
180 }
181
182
183 /*
184 * Most types of marker are followed by a variable-length parameter segment.
185 * This routine skips over the parameters for any marker we don't otherwise
186 * want to process.
187 * Note that we MUST skip the parameter segment explicitly in order not to
188 * be fooled by 0xFF bytes that might appear within the parameter segment;
189 * such bytes do NOT introduce new markers.
190 */
191
192 static void
skip_variable(void)193 skip_variable (void)
194 /* Skip over an unknown or uninteresting variable-length marker */
195 {
196 unsigned int length;
197
198 /* Get the marker parameter length count */
199 length = read_2_bytes();
200 /* Length includes itself, so must be at least 2 */
201 if (length < 2)
202 ERREXIT("Erroneous JPEG marker length");
203 length -= 2;
204 /* Skip over the remaining bytes */
205 while (length > 0) {
206 (void) read_1_byte();
207 length--;
208 }
209 }
210
211
212 /*
213 * Process a COM marker.
214 * We want to print out the marker contents as legible text;
215 * we must guard against non-text junk and varying newline representations.
216 */
217
218 static void
process_COM(int raw)219 process_COM (int raw)
220 {
221 unsigned int length;
222 int ch;
223 int lastch = 0;
224
225 /* Bill Allombert: set locale properly for isprint */
226 #ifdef HAVE_LOCALE_H
227 setlocale(LC_CTYPE, "");
228 #endif
229
230 /* Get the marker parameter length count */
231 length = read_2_bytes();
232 /* Length includes itself, so must be at least 2 */
233 if (length < 2)
234 ERREXIT("Erroneous JPEG marker length");
235 length -= 2;
236
237 while (length > 0) {
238 ch = read_1_byte();
239 if (raw) {
240 putc(ch, stdout);
241 /* Emit the character in a readable form.
242 * Nonprintables are converted to \nnn form,
243 * while \ is converted to \\.
244 * Newlines in CR, CR/LF, or LF form will be printed as one newline.
245 */
246 } else if (ch == '\r') {
247 printf("\n");
248 } else if (ch == '\n') {
249 if (lastch != '\r')
250 printf("\n");
251 } else if (ch == '\\') {
252 printf("\\\\");
253 } else if (isprint(ch)) {
254 putc(ch, stdout);
255 } else {
256 printf("\\%03o", ch);
257 }
258 lastch = ch;
259 length--;
260 }
261 printf("\n");
262
263 /* Bill Allombert: revert to C locale */
264 #ifdef HAVE_LOCALE_H
265 setlocale(LC_CTYPE, "C");
266 #endif
267 }
268
269
270 /*
271 * Process a SOFn marker.
272 * This code is only needed if you want to know the image dimensions...
273 */
274
275 static void
process_SOFn(int marker)276 process_SOFn (int marker)
277 {
278 unsigned int length;
279 unsigned int image_height, image_width;
280 int data_precision, num_components;
281 const char * process;
282 int ci;
283
284 length = read_2_bytes(); /* usual parameter length count */
285
286 data_precision = read_1_byte();
287 image_height = read_2_bytes();
288 image_width = read_2_bytes();
289 num_components = read_1_byte();
290
291 switch (marker) {
292 case M_SOF0: process = "Baseline"; break;
293 case M_SOF1: process = "Extended sequential"; break;
294 case M_SOF2: process = "Progressive"; break;
295 case M_SOF3: process = "Lossless"; break;
296 case M_SOF5: process = "Differential sequential"; break;
297 case M_SOF6: process = "Differential progressive"; break;
298 case M_SOF7: process = "Differential lossless"; break;
299 case M_SOF9: process = "Extended sequential, arithmetic coding"; break;
300 case M_SOF10: process = "Progressive, arithmetic coding"; break;
301 case M_SOF11: process = "Lossless, arithmetic coding"; break;
302 case M_SOF13: process = "Differential sequential, arithmetic coding"; break;
303 case M_SOF14: process = "Differential progressive, arithmetic coding"; break;
304 case M_SOF15: process = "Differential lossless, arithmetic coding"; break;
305 default: process = "Unknown"; break;
306 }
307
308 printf("JPEG image is %uw * %uh, %d color components, %d bits per sample\n",
309 image_width, image_height, num_components, data_precision);
310 printf("JPEG process: %s\n", process);
311
312 if (length != (unsigned int) (8 + num_components * 3))
313 ERREXIT("Bogus SOF marker length");
314
315 for (ci = 0; ci < num_components; ci++) {
316 (void) read_1_byte(); /* Component ID code */
317 (void) read_1_byte(); /* H, V sampling factors */
318 (void) read_1_byte(); /* Quantization table number */
319 }
320 }
321
322
323 /*
324 * Parse the marker stream until SOS or EOI is seen;
325 * display any COM markers.
326 * While the companion program wrjpgcom will always insert COM markers before
327 * SOFn, other implementations might not, so we scan to SOS before stopping.
328 * If we were only interested in the image dimensions, we would stop at SOFn.
329 * (Conversely, if we only cared about COM markers, there would be no need
330 * for special code to handle SOFn; we could treat it like other markers.)
331 */
332
333 static int
scan_JPEG_header(int verbose,int raw)334 scan_JPEG_header (int verbose, int raw)
335 {
336 int marker;
337
338 /* Expect SOI at start of file */
339 if (first_marker() != M_SOI)
340 ERREXIT("Expected SOI marker first");
341
342 /* Scan miscellaneous markers until we reach SOS. */
343 for (;;) {
344 marker = next_marker();
345 switch (marker) {
346 /* Note that marker codes 0xC4, 0xC8, 0xCC are not, and must not be,
347 * treated as SOFn. C4 in particular is actually DHT.
348 */
349 case M_SOF0: /* Baseline */
350 case M_SOF1: /* Extended sequential, Huffman */
351 case M_SOF2: /* Progressive, Huffman */
352 case M_SOF3: /* Lossless, Huffman */
353 case M_SOF5: /* Differential sequential, Huffman */
354 case M_SOF6: /* Differential progressive, Huffman */
355 case M_SOF7: /* Differential lossless, Huffman */
356 case M_SOF9: /* Extended sequential, arithmetic */
357 case M_SOF10: /* Progressive, arithmetic */
358 case M_SOF11: /* Lossless, arithmetic */
359 case M_SOF13: /* Differential sequential, arithmetic */
360 case M_SOF14: /* Differential progressive, arithmetic */
361 case M_SOF15: /* Differential lossless, arithmetic */
362 if (verbose)
363 process_SOFn(marker);
364 else
365 skip_variable();
366 break;
367
368 case M_SOS: /* stop before hitting compressed data */
369 return marker;
370
371 case M_EOI: /* in case it's a tables-only JPEG stream */
372 return marker;
373
374 case M_COM:
375 process_COM(raw);
376 break;
377
378 case M_APP12:
379 /* Some digital camera makers put useful textual information into
380 * APP12 markers, so we print those out too when in -verbose mode.
381 */
382 if (verbose) {
383 printf("APP12 contains:\n");
384 process_COM(raw);
385 } else
386 skip_variable();
387 break;
388
389 default: /* Anything else just gets skipped */
390 skip_variable(); /* we assume it has a parameter count... */
391 break;
392 }
393 } /* end loop */
394 }
395
396
397 /* Command line parsing code */
398
399 static const char * progname; /* program name for error messages */
400
401
402 static void
usage(void)403 usage (void)
404 /* complain about bad command line */
405 {
406 fprintf(stderr, "rdjpgcom displays any textual comments in a JPEG file.\n");
407
408 fprintf(stderr, "Usage: %s [switches] [inputfile]\n", progname);
409
410 fprintf(stderr, "Switches (names may be abbreviated):\n");
411 fprintf(stderr, " -raw Display non-printable characters in comments (unsafe)\n");
412 fprintf(stderr, " -verbose Also display dimensions of JPEG image\n");
413
414 exit(EXIT_FAILURE);
415 }
416
417
418 static int
keymatch(char * arg,const char * keyword,int minchars)419 keymatch (char * arg, const char * keyword, int minchars)
420 /* Case-insensitive matching of (possibly abbreviated) keyword switches. */
421 /* keyword is the constant keyword (must be lower case already), */
422 /* minchars is length of minimum legal abbreviation. */
423 {
424 register int ca, ck;
425 register int nmatched = 0;
426
427 while ((ca = *arg++) != '\0') {
428 if ((ck = *keyword++) == '\0')
429 return 0; /* arg longer than keyword, no good */
430 if (isupper(ca)) /* force arg to lcase (assume ck is already) */
431 ca = tolower(ca);
432 if (ca != ck)
433 return 0; /* no good */
434 nmatched++; /* count matched characters */
435 }
436 /* reached end of argument; fail if it's too short for unique abbrev */
437 if (nmatched < minchars)
438 return 0;
439 return 1; /* A-OK */
440 }
441
442
443 /*
444 * The main program.
445 */
446
447 int
main(int argc,char ** argv)448 main (int argc, char **argv)
449 {
450 int argn;
451 char * arg;
452 int verbose = 0, raw = 0;
453
454 /* On Mac, fetch a command line. */
455 #ifdef USE_CCOMMAND
456 argc = ccommand(&argv);
457 #endif
458
459 progname = argv[0];
460 if (progname == NULL || progname[0] == 0)
461 progname = "rdjpgcom"; /* in case C library doesn't provide it */
462
463 /* Parse switches, if any */
464 for (argn = 1; argn < argc; argn++) {
465 arg = argv[argn];
466 if (arg[0] != '-')
467 break; /* not switch, must be file name */
468 arg++; /* advance over '-' */
469 if (keymatch(arg, "verbose", 1)) {
470 verbose++;
471 } else if (keymatch(arg, "raw", 1)) {
472 raw = 1;
473 } else
474 usage();
475 }
476
477 /* Open the input file. */
478 /* Unix style: expect zero or one file name */
479 if (argn < argc-1) {
480 fprintf(stderr, "%s: only one input file\n", progname);
481 usage();
482 }
483 if (argn < argc) {
484 if ((infile = fopen(argv[argn], READ_BINARY)) == NULL) {
485 fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
486 exit(EXIT_FAILURE);
487 }
488 } else {
489 /* default input file is stdin */
490 #ifdef USE_SETMODE /* need to hack file mode? */
491 setmode(fileno(stdin), O_BINARY);
492 #endif
493 #ifdef USE_FDOPEN /* need to re-open in binary mode? */
494 if ((infile = fdopen(fileno(stdin), READ_BINARY)) == NULL) {
495 fprintf(stderr, "%s: can't open stdin\n", progname);
496 exit(EXIT_FAILURE);
497 }
498 #else
499 infile = stdin;
500 #endif
501 }
502
503 /* Scan the JPEG headers. */
504 (void) scan_JPEG_header(verbose, raw);
505
506 /* All done. */
507 exit(EXIT_SUCCESS);
508 return 0; /* suppress no-return-value warnings */
509 }
510