Mercurial > hgrepos > Python2 > PyMuPDF

comparison mupdf-source/thirdparty/libjpeg/rdjpgcom.c @ 2:b50eed0cc0ef upstream

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
ADD: MuPDF v1.26.7: the MuPDF source as downloaded by a default build of PyMuPDF 1.26.4. The directory name has changed: no version number in the expanded directory now.
author Franz Glasner <fzglas.hg@dom66.de>
date Mon, 15 Sep 2025 11:43:07 +0200
parents
children
comparison
equal deleted inserted replaced
1:1d09e1dec1d9 2:b50eed0cc0ef
1 /*
2 * rdjpgcom.c
3 *
4 * Copyright (C) 1994-1997, Thomas G. Lane.
5 * Modified 2009 by Bill Allombert, Guido Vollbeding.
6 * This file is part of the Independent JPEG Group's software.
7 * For conditions of distribution and use, see the accompanying README file.
8 *
9 * This file contains a very simple stand-alone application that displays
10 * the text in COM (comment) markers in a JFIF file.
11 * This may be useful as an example of the minimum logic needed to parse
12 * JPEG markers.
13 */
14
15 #define JPEG_CJPEG_DJPEG /* to get the command-line config symbols */
16 #include "jinclude.h" /* get auto-config symbols, <stdio.h> */
17
18 #ifdef HAVE_LOCALE_H
19 #include <locale.h> /* Bill Allombert: use locale for isprint */
20 #endif
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 (int raw)
226 {
227 unsigned int length;
228 int ch;
229 int lastch = 0;
230
231 /* Bill Allombert: set locale properly for isprint */
232 #ifdef HAVE_LOCALE_H
233 setlocale(LC_CTYPE, "");
234 #endif
235
236 /* Get the marker parameter length count */
237 length = read_2_bytes();
238 /* Length includes itself, so must be at least 2 */
239 if (length < 2)
240 ERREXIT("Erroneous JPEG marker length");
241 length -= 2;
242
243 while (length > 0) {
244 ch = read_1_byte();
245 if (raw) {
246 putc(ch, stdout);
247 /* Emit the character in a readable form.
248 * Nonprintables are converted to \nnn form,
249 * while \ is converted to \\.
250 * Newlines in CR, CR/LF, or LF form will be printed as one newline.
251 */
252 } else if (ch == '\r') {
253 printf("\n");
254 } else if (ch == '\n') {
255 if (lastch != '\r')
256 printf("\n");
257 } else if (ch == '\\') {
258 printf("\\\\");
259 } else if (isprint(ch)) {
260 putc(ch, stdout);
261 } else {
262 printf("\\%03o", ch);
263 }
264 lastch = ch;
265 length--;
266 }
267 printf("\n");
268
269 /* Bill Allombert: revert to C locale */
270 #ifdef HAVE_LOCALE_H
271 setlocale(LC_CTYPE, "C");
272 #endif
273 }
274
275
276 /*
277 * Process a SOFn marker.
278 * This code is only needed if you want to know the image dimensions...
279 */
280
281 static void
282 process_SOFn (int marker)
283 {
284 unsigned int length;
285 unsigned int image_height, image_width;
286 int data_precision, num_components;
287 const char * process;
288 int ci;
289
290 length = read_2_bytes(); /* usual parameter length count */
291
292 data_precision = read_1_byte();
293 image_height = read_2_bytes();
294 image_width = read_2_bytes();
295 num_components = read_1_byte();
296
297 switch (marker) {
298 case M_SOF0: process = "Baseline"; break;
299 case M_SOF1: process = "Extended sequential"; break;
300 case M_SOF2: process = "Progressive"; break;
301 case M_SOF3: process = "Lossless"; break;
302 case M_SOF5: process = "Differential sequential"; break;
303 case M_SOF6: process = "Differential progressive"; break;
304 case M_SOF7: process = "Differential lossless"; break;
305 case M_SOF9: process = "Extended sequential, arithmetic coding"; break;
306 case M_SOF10: process = "Progressive, arithmetic coding"; break;
307 case M_SOF11: process = "Lossless, arithmetic coding"; break;
308 case M_SOF13: process = "Differential sequential, arithmetic coding"; break;
309 case M_SOF14: process = "Differential progressive, arithmetic coding"; break;
310 case M_SOF15: process = "Differential lossless, arithmetic coding"; break;
311 default: process = "Unknown"; break;
312 }
313
314 printf("JPEG image is %uw * %uh, %d color components, %d bits per sample\n",
315 image_width, image_height, num_components, data_precision);
316 printf("JPEG process: %s\n", process);
317
318 if (length != (unsigned int) (8 + num_components * 3))
319 ERREXIT("Bogus SOF marker length");
320
321 for (ci = 0; ci < num_components; ci++) {
322 (void) read_1_byte(); /* Component ID code */
323 (void) read_1_byte(); /* H, V sampling factors */
324 (void) read_1_byte(); /* Quantization table number */
325 }
326 }
327
328
329 /*
330 * Parse the marker stream until SOS or EOI is seen;
331 * display any COM markers.
332 * While the companion program wrjpgcom will always insert COM markers before
333 * SOFn, other implementations might not, so we scan to SOS before stopping.
334 * If we were only interested in the image dimensions, we would stop at SOFn.
335 * (Conversely, if we only cared about COM markers, there would be no need
336 * for special code to handle SOFn; we could treat it like other markers.)
337 */
338
339 static int
340 scan_JPEG_header (int verbose, int raw)
341 {
342 int marker;
343
344 /* Expect SOI at start of file */
345 if (first_marker() != M_SOI)
346 ERREXIT("Expected SOI marker first");
347
348 /* Scan miscellaneous markers until we reach SOS. */
349 for (;;) {
350 marker = next_marker();
351 switch (marker) {
352 /* Note that marker codes 0xC4, 0xC8, 0xCC are not, and must not be,
353 * treated as SOFn. C4 in particular is actually DHT.
354 */
355 case M_SOF0: /* Baseline */
356 case M_SOF1: /* Extended sequential, Huffman */
357 case M_SOF2: /* Progressive, Huffman */
358 case M_SOF3: /* Lossless, Huffman */
359 case M_SOF5: /* Differential sequential, Huffman */
360 case M_SOF6: /* Differential progressive, Huffman */
361 case M_SOF7: /* Differential lossless, Huffman */
362 case M_SOF9: /* Extended sequential, arithmetic */
363 case M_SOF10: /* Progressive, arithmetic */
364 case M_SOF11: /* Lossless, arithmetic */
365 case M_SOF13: /* Differential sequential, arithmetic */
366 case M_SOF14: /* Differential progressive, arithmetic */
367 case M_SOF15: /* Differential lossless, arithmetic */
368 if (verbose)
369 process_SOFn(marker);
370 else
371 skip_variable();
372 break;
373
374 case M_SOS: /* stop before hitting compressed data */
375 return marker;
376
377 case M_EOI: /* in case it's a tables-only JPEG stream */
378 return marker;
379
380 case M_COM:
381 process_COM(raw);
382 break;
383
384 case M_APP12:
385 /* Some digital camera makers put useful textual information into
386 * APP12 markers, so we print those out too when in -verbose mode.
387 */
388 if (verbose) {
389 printf("APP12 contains:\n");
390 process_COM(raw);
391 } else
392 skip_variable();
393 break;
394
395 default: /* Anything else just gets skipped */
396 skip_variable(); /* we assume it has a parameter count... */
397 break;
398 }
399 } /* end loop */
400 }
401
402
403 /* Command line parsing code */
404
405 static const char * progname; /* program name for error messages */
406
407
408 static void
409 usage (void)
410 /* complain about bad command line */
411 {
412 fprintf(stderr, "rdjpgcom displays any textual comments in a JPEG file.\n");
413
414 fprintf(stderr, "Usage: %s [switches] [inputfile]\n", progname);
415
416 fprintf(stderr, "Switches (names may be abbreviated):\n");
417 fprintf(stderr, " -raw Display non-printable characters in comments (unsafe)\n");
418 fprintf(stderr, " -verbose Also display dimensions of JPEG image\n");
419
420 exit(EXIT_FAILURE);
421 }
422
423
424 static int
425 keymatch (char * arg, const char * keyword, int minchars)
426 /* Case-insensitive matching of (possibly abbreviated) keyword switches. */
427 /* keyword is the constant keyword (must be lower case already), */
428 /* minchars is length of minimum legal abbreviation. */
429 {
430 register int ca, ck;
431 register int nmatched = 0;
432
433 while ((ca = *arg++) != '\0') {
434 if ((ck = *keyword++) == '\0')
435 return 0; /* arg longer than keyword, no good */
436 if (isupper(ca)) /* force arg to lcase (assume ck is already) */
437 ca = tolower(ca);
438 if (ca != ck)
439 return 0; /* no good */
440 nmatched++; /* count matched characters */
441 }
442 /* reached end of argument; fail if it's too short for unique abbrev */
443 if (nmatched < minchars)
444 return 0;
445 return 1; /* A-OK */
446 }
447
448
449 /*
450 * The main program.
451 */
452
453 int
454 main (int argc, char **argv)
455 {
456 int argn;
457 char * arg;
458 int verbose = 0, raw = 0;
459
460 /* On Mac, fetch a command line. */
461 #ifdef USE_CCOMMAND
462 argc = ccommand(&argv);
463 #endif
464
465 progname = argv[0];
466 if (progname == NULL || progname[0] == 0)
467 progname = "rdjpgcom"; /* in case C library doesn't provide it */
468
469 /* Parse switches, if any */
470 for (argn = 1; argn < argc; argn++) {
471 arg = argv[argn];
472 if (arg[0] != '-')
473 break; /* not switch, must be file name */
474 arg++; /* advance over '-' */
475 if (keymatch(arg, "verbose", 1)) {
476 verbose++;
477 } else if (keymatch(arg, "raw", 1)) {
478 raw = 1;
479 } else
480 usage();
481 }
482
483 /* Open the input file. */
484 /* Unix style: expect zero or one file name */
485 if (argn < argc-1) {
486 fprintf(stderr, "%s: only one input file\n", progname);
487 usage();
488 }
489 if (argn < argc) {
490 if ((infile = fopen(argv[argn], READ_BINARY)) == NULL) {
491 fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
492 exit(EXIT_FAILURE);
493 }
494 } else {
495 /* default input file is stdin */
496 #ifdef USE_SETMODE /* need to hack file mode? */
497 setmode(fileno(stdin), O_BINARY);
498 #endif
499 #ifdef USE_FDOPEN /* need to re-open in binary mode? */
500 if ((infile = fdopen(fileno(stdin), READ_BINARY)) == NULL) {
501 fprintf(stderr, "%s: can't open stdin\n", progname);
502 exit(EXIT_FAILURE);
503 }
504 #else
505 infile = stdin;
506 #endif
507 }
508
509 /* Scan the JPEG headers. */
510 (void) scan_JPEG_header(verbose, raw);
511
512 /* All done. */
513 exit(EXIT_SUCCESS);
514 return 0; /* suppress no-return-value warnings */
515 }