1 /*
2 * rdjpgcom.c
3 *
4 * Copyright (C) 1994, Thomas G. Lane.
5 * This file is part of the Independent JPEG Group's software.
6 * For conditions of distribution and use, see the accompanying README file.
7 *
8 * This file contains a very simple stand-alone application that displays
9 * the text in COM (comment) markers in a JFIF file.
10 * This may be useful as an example of the minimum logic needed to parse
11 * JPEG markers.
12 */
13
14 #define JPEG_CJPEG_DJPEG /* to get the command-line config symbols */
15 #include "jinclude.h" /* get auto-config symbols, <stdio.h> */
16
17 #include <ctype.h> /* to declare isupper(), tolower() */
18 #ifdef USE_SETMODE
19 #include <fcntl.h> /* to declare setmode()'s parameter macros */
20 /* If you have setmode() but not <io.h>, just delete this line: */
21 #include <io.h> /* to declare setmode() */
22 #endif
23
24 #ifdef USE_CCOMMAND /* command-line reader for Macintosh */
25 #ifdef __MWERKS__
26 #include <SIOUX.h> /* Metrowerks declares it here */
27 #endif
28 #ifdef THINK_C
29 #include <console.h> /* Think declares it here */
30 #endif
31 #endif
32
33 #ifdef DONT_USE_B_MODE /* define mode parameters for fopen() */
34 #define READ_BINARY "r"
35 #else
36 #define READ_BINARY "rb"
37 #endif
38
39 #ifndef EXIT_FAILURE /* define exit() codes if not provided */
40 #define EXIT_FAILURE 1
41 #endif
42 #ifndef EXIT_SUCCESS
43 #ifdef VMS
44 #define EXIT_SUCCESS 1 /* VMS is very nonstandard */
45 #else
46 #define EXIT_SUCCESS 0
47 #endif
48 #endif
49
50
51 /*
52 * These macros are used to read the input file.
53 * To reuse this code in another application, you might need to change these.
54 */
55
56 static FILE * infile; /* input JPEG file */
57
58 /* Return next input byte, or EOF if no more */
59 #define NEXTBYTE() getc(infile)
60
61
62 /* Error exit handler */
63 #define ERREXIT(msg) (fprintf(stderr, "%s\n", msg), exit(EXIT_FAILURE))
64
65
66 /* Read one byte, testing for EOF */
67 static int
read_1_byte(void)68 read_1_byte (void)
69 {
70 int c;
71
72 c = NEXTBYTE();
73 if (c == EOF)
74 ERREXIT("Premature EOF in JPEG file");
75 return c;
76 }
77
78 /* Read 2 bytes, convert to unsigned int */
79 /* All 2-byte quantities in JPEG markers are MSB first */
80 static unsigned int
read_2_bytes(void)81 read_2_bytes (void)
82 {
83 int c1, c2;
84
85 c1 = NEXTBYTE();
86 if (c1 == EOF)
87 ERREXIT("Premature EOF in JPEG file");
88 c2 = NEXTBYTE();
89 if (c2 == EOF)
90 ERREXIT("Premature EOF in JPEG file");
91 return (((unsigned int) c1) << 8) + ((unsigned int) c2);
92 }
93
94
95 /*
96 * JPEG markers consist of one or more 0xFF bytes, followed by a marker
97 * code byte (which is not an FF). Here are the marker codes of interest
98 * in this program. (See jdmarker.c for a more complete list.)
99 */
100
101 #define M_SOF0 0xC0 /* Start Of Frame N */
102 #define M_SOF1 0xC1 /* N indicates which compression process */
103 #define M_SOF2 0xC2 /* Only SOF0 and SOF1 are now in common use */
104 #define M_SOF3 0xC3
105 #define M_SOF5 0xC5
106 #define M_SOF6 0xC6
107 #define M_SOF7 0xC7
108 #define M_SOF9 0xC9
109 #define M_SOF10 0xCA
110 #define M_SOF11 0xCB
111 #define M_SOF13 0xCD
112 #define M_SOF14 0xCE
113 #define M_SOF15 0xCF
114 #define M_SOI 0xD8 /* Start Of Image (beginning of datastream) */
115 #define M_EOI 0xD9 /* End Of Image (end of datastream) */
116 #define M_SOS 0xDA /* Start Of Scan (begins compressed data) */
117 #define M_COM 0xFE /* COMment */
118
119
120 /*
121 * Find the next JPEG marker and return its marker code.
122 * We expect at least one FF byte, possibly more if the compressor used FFs
123 * to pad the file.
124 * There could also be non-FF garbage between markers. The treatment of such
125 * garbage is unspecified; we choose to skip over it but emit a warning msg.
126 * NB: this routine must not be used after seeing SOS marker, since it will
127 * not deal correctly with FF/00 sequences in the compressed image data...
128 */
129
130 static int
next_marker(void)131 next_marker (void)
132 {
133 int c;
134 int discarded_bytes = 0;
135
136 /* Find 0xFF byte; count and skip any non-FFs. */
137 c = read_1_byte();
138 while (c != 0xFF) {
139 discarded_bytes++;
140 c = read_1_byte();
141 }
142 /* Get marker code byte, swallowing any duplicate FF bytes. Extra FFs
143 * are legal as pad bytes, so don't count them in discarded_bytes.
144 */
145 do {
146 c = read_1_byte();
147 } while (c == 0xFF);
148
149 if (discarded_bytes != 0) {
150 fprintf(stderr, "Warning: garbage data found in JPEG file\n");
151 }
152
153 return c;
154 }
155
156
157 /*
158 * Read the initial marker, which should be SOI.
159 * For a JFIF file, the first two bytes of the file should be literally
160 * 0xFF M_SOI. To be more general, we could use next_marker, but if the
161 * input file weren't actually JPEG at all, next_marker might read the whole
162 * file and then return a misleading error message...
163 */
164
165 static int
first_marker(void)166 first_marker (void)
167 {
168 int c1, c2;
169
170 c1 = NEXTBYTE();
171 c2 = NEXTBYTE();
172 if (c1 != 0xFF || c2 != M_SOI)
173 ERREXIT("Not a JPEG file");
174 return c2;
175 }
176
177
178 /*
179 * Most types of marker are followed by a variable-length parameter segment.
180 * This routine skips over the parameters for any marker we don't otherwise
181 * want to process.
182 * Note that we MUST skip the parameter segment explicitly in order not to
183 * be fooled by 0xFF bytes that might appear within the parameter segment;
184 * such bytes do NOT introduce new markers.
185 */
186
187 static void
skip_variable(void)188 skip_variable (void)
189 /* Skip over an unknown or uninteresting variable-length marker */
190 {
191 unsigned int length;
192
193 /* Get the marker parameter length count */
194 length = read_2_bytes();
195 /* Length includes itself, so must be at least 2 */
196 if (length < 2)
197 ERREXIT("Erroneous JPEG marker length");
198 length -= 2;
199 /* Skip over the remaining bytes */
200 while (length > 0) {
201 (void) read_1_byte();
202 length--;
203 }
204 }
205
206
207 /*
208 * Process a COM marker.
209 * We want to print out the marker contents as legible text;
210 * we must guard against random junk and varying newline representations.
211 */
212
213 static void
process_COM(void)214 process_COM (void)
215 {
216 unsigned int length;
217 int ch;
218 int lastch = 0;
219
220 /* Get the marker parameter length count */
221 length = read_2_bytes();
222 /* Length includes itself, so must be at least 2 */
223 if (length < 2)
224 ERREXIT("Erroneous JPEG marker length");
225 length -= 2;
226
227 while (length > 0) {
228 ch = read_1_byte();
229 /* Emit the character in a readable form.
230 * Nonprintables are converted to \nnn form,
231 * while \ is converted to \\.
232 * Newlines in CR, CR/LF, or LF form will be printed as one newline.
233 */
234 if (ch == '\r') {
235 printf("\n");
236 } else if (ch == '\n') {
237 if (lastch != '\r')
238 printf("\n");
239 } else if (ch == '\\') {
240 printf("\\\\");
241 } else if (isprint(ch)) {
242 putc(ch, stdout);
243 } else {
244 printf("\\%03o", ch);
245 }
246 lastch = ch;
247 length--;
248 }
249 printf("\n");
250 }
251
252
253 /*
254 * Process a SOFn marker.
255 * This code is only needed if you want to know the image dimensions...
256 */
257
258 static void
process_SOFn(int marker)259 process_SOFn (int marker)
260 {
261 unsigned int length;
262 unsigned int image_height, image_width;
263 int data_precision, num_components;
264 const char * process;
265 int ci, c1, c2, c3;
266
267 length = read_2_bytes(); /* usual parameter length count */
268
269 data_precision = read_1_byte();
270 image_height = read_2_bytes();
271 image_width = read_2_bytes();
272 num_components = read_1_byte();
273
274 switch (marker) {
275 case M_SOF0: process = "Baseline"; break;
276 case M_SOF1: process = "Extended sequential"; break;
277 case M_SOF2: process = "Progressive"; break;
278 case M_SOF3: process = "Lossless"; break;
279 case M_SOF5: process = "Differential sequential"; break;
280 case M_SOF6: process = "Differential progressive"; break;
281 case M_SOF7: process = "Differential lossless"; break;
282 case M_SOF9: process = "Extended sequential, arithmetic coding"; break;
283 case M_SOF10: process = "Progressive, arithmetic coding"; break;
284 case M_SOF11: process = "Lossless, arithmetic coding"; break;
285 case M_SOF13: process = "Differential sequential, arithmetic coding"; break;
286 case M_SOF14: process = "Differential progressive, arithmetic coding"; break;
287 case M_SOF15: process = "Differential lossless, arithmetic coding"; break;
288 default: process = "Unknown"; break;
289 }
290
291 printf("JPEG image is %uw * %uh, %d color components, %d bits per sample\n",
292 image_width, image_height, num_components, data_precision);
293 printf("JPEG process: %s\n", process);
294
295 if (length != (unsigned int) (8 + num_components * 3))
296 ERREXIT("Bogus SOF marker length");
297
298 for (ci = 0; ci < num_components; ci++) {
299 c1 = read_1_byte(); /* Component ID code */
300 c2 = read_1_byte(); /* H, V sampling factors */
301 c3 = read_1_byte(); /* Quant table number */
302 }
303 }
304
305
306 /*
307 * Parse the marker stream until SOS or EOI is seen;
308 * display any COM markers.
309 * While the companion program wrjpgcom will always insert COM markers before
310 * SOFn, other implementations might not, so we scan to SOS before stopping.
311 * If we were only interested in the image dimensions, we would stop at SOFn.
312 * (Conversely, if we only cared about COM markers, there would be no need
313 * for special code to handle SOFn; we could treat it like other markers.)
314 */
315
316 static int
scan_JPEG_header(int verbose)317 scan_JPEG_header (int verbose)
318 {
319 int marker;
320
321 /* Expect SOI at start of file */
322 if (first_marker() != M_SOI)
323 ERREXIT("Expected SOI marker first");
324
325 /* Scan miscellaneous markers until we reach SOS. */
326 for (;;) {
327 marker = next_marker();
328 switch (marker) {
329 case M_SOF0: /* Baseline */
330 case M_SOF1: /* Extended sequential, Huffman */
331 case M_SOF2: /* Progressive, Huffman */
332 case M_SOF3: /* Lossless, Huffman */
333 case M_SOF5: /* Differential sequential, Huffman */
334 case M_SOF6: /* Differential progressive, Huffman */
335 case M_SOF7: /* Differential lossless, Huffman */
336 case M_SOF9: /* Extended sequential, arithmetic */
337 case M_SOF10: /* Progressive, arithmetic */
338 case M_SOF11: /* Lossless, arithmetic */
339 case M_SOF13: /* Differential sequential, arithmetic */
340 case M_SOF14: /* Differential progressive, arithmetic */
341 case M_SOF15: /* Differential lossless, arithmetic */
342 if (verbose)
343 process_SOFn(marker);
344 else
345 skip_variable();
346 break;
347
348 case M_SOS: /* stop before hitting compressed data */
349 return marker;
350
351 case M_EOI: /* in case it's a tables-only JPEG stream */
352 return marker;
353
354 case M_COM:
355 process_COM();
356 break;
357
358 default: /* Anything else just gets skipped */
359 skip_variable(); /* we assume it has a parameter count... */
360 break;
361 }
362 } /* end loop */
363 }
364
365
366 /* Command line parsing code */
367
368 static const char * progname; /* program name for error messages */
369
370
371 static void
usage(void)372 usage (void)
373 /* complain about bad command line */
374 {
375 fprintf(stderr, "rdjpgcom displays any textual comments in a JPEG file.\n");
376
377 fprintf(stderr, "Usage: %s [switches] [inputfile]\n", progname);
378
379 fprintf(stderr, "Switches (names may be abbreviated):\n");
380 fprintf(stderr, " -verbose Also display dimensions of JPEG image\n");
381
382 exit(EXIT_FAILURE);
383 }
384
385
386 static int
keymatch(char * arg,const char * keyword,int minchars)387 keymatch (char * arg, const char * keyword, int minchars)
388 /* Case-insensitive matching of (possibly abbreviated) keyword switches. */
389 /* keyword is the constant keyword (must be lower case already), */
390 /* minchars is length of minimum legal abbreviation. */
391 {
392 register int ca, ck;
393 register int nmatched = 0;
394
395 while ((ca = *arg++) != '\0') {
396 if ((ck = *keyword++) == '\0')
397 return 0; /* arg longer than keyword, no good */
398 if (isupper(ca)) /* force arg to lcase (assume ck is already) */
399 ca = tolower(ca);
400 if (ca != ck)
401 return 0; /* no good */
402 nmatched++; /* count matched characters */
403 }
404 /* reached end of argument; fail if it's too short for unique abbrev */
405 if (nmatched < minchars)
406 return 0;
407 return 1; /* A-OK */
408 }
409
410
411 /*
412 * The main program.
413 */
414
415 int
main(int argc,char ** argv)416 main (int argc, char **argv)
417 {
418 int argn;
419 char * arg;
420 int verbose = 0;
421
422 /* On Mac, fetch a command line. */
423 #ifdef USE_CCOMMAND
424 argc = ccommand(&argv);
425 #endif
426
427 progname = argv[0];
428 if (progname == NULL || progname[0] == 0)
429 progname = "rdjpgcom"; /* in case C library doesn't provide it */
430
431 /* Parse switches, if any */
432 for (argn = 1; argn < argc; argn++) {
433 arg = argv[argn];
434 if (arg[0] != '-')
435 break; /* not switch, must be file name */
436 arg++; /* advance over '-' */
437 if (keymatch(arg, "verbose", 1)) {
438 verbose++;
439 } else
440 usage();
441 }
442
443 /* Open the input file. */
444 /* Unix style: expect zero or one file name */
445 if (argn < argc-1) {
446 fprintf(stderr, "%s: only one input file\n", progname);
447 usage();
448 }
449 if (argn < argc) {
450 if ((infile = fopen(argv[argn], READ_BINARY)) == NULL) {
451 fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
452 exit(EXIT_FAILURE);
453 }
454 } else {
455 /* default input file is stdin */
456 #ifdef USE_SETMODE /* need to hack file mode? */
457 setmode(fileno(stdin), O_BINARY);
458 #endif
459 #ifdef USE_FDOPEN /* need to re-open in binary mode? */
460 if ((infile = fdopen(fileno(stdin), READ_BINARY)) == NULL) {
461 fprintf(stderr, "%s: can't open stdin\n", progname);
462 exit(EXIT_FAILURE);
463 }
464 #else
465 infile = stdin;
466 #endif
467 }
468
469 /* Scan the JPEG headers. */
470 (void) scan_JPEG_header(verbose);
471
472 /* All done. */
473 exit(EXIT_SUCCESS);
474 return 0; /* suppress no-return-value warnings */
475 }
476