1/*
2 * wrjpgcom.c
3 *
4 * Copyright (C) 1994-1997, 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 inserts
9 * user-supplied text as a COM (comment) marker 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#ifndef HAVE_STDLIB_H		/* <stdlib.h> should declare malloc() */
18extern void * malloc ();
19#endif
20#include <ctype.h>		/* to declare isupper(), tolower() */
21#ifdef USE_SETMODE
22#include <fcntl.h>		/* to declare setmode()'s parameter macros */
23/* If you have setmode() but not <io.h>, just delete this line: */
24#include <io.h>			/* to declare setmode() */
25#endif
26
27#ifdef USE_CCOMMAND		/* command-line reader for Macintosh */
28#ifdef __MWERKS__
29#include <SIOUX.h>              /* Metrowerks needs this */
30#include <console.h>		/* ... and this */
31#endif
32#ifdef THINK_C
33#include <console.h>		/* Think declares it here */
34#endif
35#endif
36
37#ifdef DONT_USE_B_MODE		/* define mode parameters for fopen() */
38#define READ_BINARY	"r"
39#define WRITE_BINARY	"w"
40#else
41#ifdef VMS			/* VMS is very nonstandard */
42#define READ_BINARY	"rb", "ctx=stm"
43#define WRITE_BINARY	"wb", "ctx=stm"
44#else				/* standard ANSI-compliant case */
45#define READ_BINARY	"rb"
46#define WRITE_BINARY	"wb"
47#endif
48#endif
49
50#ifndef EXIT_FAILURE		/* define exit() codes if not provided */
51#define EXIT_FAILURE  1
52#endif
53#ifndef EXIT_SUCCESS
54#ifdef VMS
55#define EXIT_SUCCESS  1		/* VMS is very nonstandard */
56#else
57#define EXIT_SUCCESS  0
58#endif
59#endif
60
61/* Reduce this value if your malloc() can't allocate blocks up to 64K.
62 * On DOS, compiling in large model is usually a better solution.
63 */
64
65#ifndef MAX_COM_LENGTH
66#define MAX_COM_LENGTH 65000L	/* must be <= 65533 in any case */
67#endif
68
69
70/*
71 * These macros are used to read the input file and write the output file.
72 * To reuse this code in another application, you might need to change these.
73 */
74
75static FILE * infile;		/* input JPEG file */
76
77/* Return next input byte, or EOF if no more */
78#define NEXTBYTE()  getc(infile)
79
80static FILE * outfile;		/* output JPEG file */
81
82/* Emit an output byte */
83#define PUTBYTE(x)  putc((x), outfile)
84
85
86/* Error exit handler */
87#define ERREXIT(msg)  (fprintf(stderr, "%s\n", msg), exit(EXIT_FAILURE))
88
89
90/* Read one byte, testing for EOF */
91static int
92read_1_byte (void)
93{
94  int c;
95
96  c = NEXTBYTE();
97  if (c == EOF)
98    ERREXIT("Premature EOF in JPEG file");
99  return c;
100}
101
102/* Read 2 bytes, convert to unsigned int */
103/* All 2-byte quantities in JPEG markers are MSB first */
104static unsigned int
105read_2_bytes (void)
106{
107  int c1, c2;
108
109  c1 = NEXTBYTE();
110  if (c1 == EOF)
111    ERREXIT("Premature EOF in JPEG file");
112  c2 = NEXTBYTE();
113  if (c2 == EOF)
114    ERREXIT("Premature EOF in JPEG file");
115  return (((unsigned int) c1) << 8) + ((unsigned int) c2);
116}
117
118
119/* Routines to write data to output file */
120
121static void
122write_1_byte (int c)
123{
124  PUTBYTE(c);
125}
126
127static void
128write_2_bytes (unsigned int val)
129{
130  PUTBYTE((val >> 8) & 0xFF);
131  PUTBYTE(val & 0xFF);
132}
133
134static void
135write_marker (int marker)
136{
137  PUTBYTE(0xFF);
138  PUTBYTE(marker);
139}
140
141static void
142copy_rest_of_file (void)
143{
144  int c;
145
146  while ((c = NEXTBYTE()) != EOF)
147    PUTBYTE(c);
148}
149
150
151/*
152 * JPEG markers consist of one or more 0xFF bytes, followed by a marker
153 * code byte (which is not an FF).  Here are the marker codes of interest
154 * in this program.  (See jdmarker.c for a more complete list.)
155 */
156
157#define M_SOF0  0xC0		/* Start Of Frame N */
158#define M_SOF1  0xC1		/* N indicates which compression process */
159#define M_SOF2  0xC2		/* Only SOF0-SOF2 are now in common use */
160#define M_SOF3  0xC3
161#define M_SOF5  0xC5		/* NB: codes C4 and CC are NOT SOF markers */
162#define M_SOF6  0xC6
163#define M_SOF7  0xC7
164#define M_SOF9  0xC9
165#define M_SOF10 0xCA
166#define M_SOF11 0xCB
167#define M_SOF13 0xCD
168#define M_SOF14 0xCE
169#define M_SOF15 0xCF
170#define M_SOI   0xD8		/* Start Of Image (beginning of datastream) */
171#define M_EOI   0xD9		/* End Of Image (end of datastream) */
172#define M_SOS   0xDA		/* Start Of Scan (begins compressed data) */
173#define M_COM   0xFE		/* COMment */
174
175
176/*
177 * Find the next JPEG marker and return its marker code.
178 * We expect at least one FF byte, possibly more if the compressor used FFs
179 * to pad the file.  (Padding FFs will NOT be replicated in the output file.)
180 * There could also be non-FF garbage between markers.  The treatment of such
181 * garbage is unspecified; we choose to skip over it but emit a warning msg.
182 * NB: this routine must not be used after seeing SOS marker, since it will
183 * not deal correctly with FF/00 sequences in the compressed image data...
184 */
185
186static int
187next_marker (void)
188{
189  int c;
190  int discarded_bytes = 0;
191
192  /* Find 0xFF byte; count and skip any non-FFs. */
193  c = read_1_byte();
194  while (c != 0xFF) {
195    discarded_bytes++;
196    c = read_1_byte();
197  }
198  /* Get marker code byte, swallowing any duplicate FF bytes.  Extra FFs
199   * are legal as pad bytes, so don't count them in discarded_bytes.
200   */
201  do {
202    c = read_1_byte();
203  } while (c == 0xFF);
204
205  if (discarded_bytes != 0) {
206    fprintf(stderr, "Warning: garbage data found in JPEG file\n");
207  }
208
209  return c;
210}
211
212
213/*
214 * Read the initial marker, which should be SOI.
215 * For a JFIF file, the first two bytes of the file should be literally
216 * 0xFF M_SOI.  To be more general, we could use next_marker, but if the
217 * input file weren't actually JPEG at all, next_marker might read the whole
218 * file and then return a misleading error message...
219 */
220
221static int
222first_marker (void)
223{
224  int c1, c2;
225
226  c1 = NEXTBYTE();
227  c2 = NEXTBYTE();
228  if (c1 != 0xFF || c2 != M_SOI)
229    ERREXIT("Not a JPEG file");
230  return c2;
231}
232
233
234/*
235 * Most types of marker are followed by a variable-length parameter segment.
236 * This routine skips over the parameters for any marker we don't otherwise
237 * want to process.
238 * Note that we MUST skip the parameter segment explicitly in order not to
239 * be fooled by 0xFF bytes that might appear within the parameter segment;
240 * such bytes do NOT introduce new markers.
241 */
242
243static void
244copy_variable (void)
245/* Copy an unknown or uninteresting variable-length marker */
246{
247  unsigned int length;
248
249  /* Get the marker parameter length count */
250  length = read_2_bytes();
251  write_2_bytes(length);
252  /* Length includes itself, so must be at least 2 */
253  if (length < 2)
254    ERREXIT("Erroneous JPEG marker length");
255  length -= 2;
256  /* Skip over the remaining bytes */
257  while (length > 0) {
258    write_1_byte(read_1_byte());
259    length--;
260  }
261}
262
263static void
264skip_variable (void)
265/* Skip over an unknown or uninteresting variable-length marker */
266{
267  unsigned int length;
268
269  /* Get the marker parameter length count */
270  length = read_2_bytes();
271  /* Length includes itself, so must be at least 2 */
272  if (length < 2)
273    ERREXIT("Erroneous JPEG marker length");
274  length -= 2;
275  /* Skip over the remaining bytes */
276  while (length > 0) {
277    (void) read_1_byte();
278    length--;
279  }
280}
281
282
283/*
284 * Parse the marker stream until SOFn or EOI is seen;
285 * copy data to output, but discard COM markers unless keep_COM is true.
286 */
287
288static int
289scan_JPEG_header (int keep_COM)
290{
291  int marker;
292
293  /* Expect SOI at start of file */
294  if (first_marker() != M_SOI)
295    ERREXIT("Expected SOI marker first");
296  write_marker(M_SOI);
297
298  /* Scan miscellaneous markers until we reach SOFn. */
299  for (;;) {
300    marker = next_marker();
301    switch (marker) {
302      /* Note that marker codes 0xC4, 0xC8, 0xCC are not, and must not be,
303       * treated as SOFn.  C4 in particular is actually DHT.
304       */
305    case M_SOF0:		/* Baseline */
306    case M_SOF1:		/* Extended sequential, Huffman */
307    case M_SOF2:		/* Progressive, Huffman */
308    case M_SOF3:		/* Lossless, Huffman */
309    case M_SOF5:		/* Differential sequential, Huffman */
310    case M_SOF6:		/* Differential progressive, Huffman */
311    case M_SOF7:		/* Differential lossless, Huffman */
312    case M_SOF9:		/* Extended sequential, arithmetic */
313    case M_SOF10:		/* Progressive, arithmetic */
314    case M_SOF11:		/* Lossless, arithmetic */
315    case M_SOF13:		/* Differential sequential, arithmetic */
316    case M_SOF14:		/* Differential progressive, arithmetic */
317    case M_SOF15:		/* Differential lossless, arithmetic */
318      return marker;
319
320    case M_SOS:			/* should not see compressed data before SOF */
321      ERREXIT("SOS without prior SOFn");
322      break;
323
324    case M_EOI:			/* in case it's a tables-only JPEG stream */
325      return marker;
326
327    case M_COM:			/* Existing COM: conditionally discard */
328      if (keep_COM) {
329	write_marker(marker);
330	copy_variable();
331      } else {
332	skip_variable();
333      }
334      break;
335
336    default:			/* Anything else just gets copied */
337      write_marker(marker);
338      copy_variable();		/* we assume it has a parameter count... */
339      break;
340    }
341  } /* end loop */
342}
343
344
345/* Command line parsing code */
346
347static const char * progname;	/* program name for error messages */
348
349
350static void
351usage (void)
352/* complain about bad command line */
353{
354  fprintf(stderr, "wrjpgcom inserts a textual comment in a JPEG file.\n");
355  fprintf(stderr, "You can add to or replace any existing comment(s).\n");
356
357  fprintf(stderr, "Usage: %s [switches] ", progname);
358#ifdef TWO_FILE_COMMANDLINE
359  fprintf(stderr, "inputfile outputfile\n");
360#else
361  fprintf(stderr, "[inputfile]\n");
362#endif
363
364  fprintf(stderr, "Switches (names may be abbreviated):\n");
365  fprintf(stderr, "  -replace         Delete any existing comments\n");
366  fprintf(stderr, "  -comment \"text\"  Insert comment with given text\n");
367  fprintf(stderr, "  -cfile name      Read comment from named file\n");
368  fprintf(stderr, "Notice that you must put quotes around the comment text\n");
369  fprintf(stderr, "when you use -comment.\n");
370  fprintf(stderr, "If you do not give either -comment or -cfile on the command line,\n");
371  fprintf(stderr, "then the comment text is read from standard input.\n");
372  fprintf(stderr, "It can be multiple lines, up to %u characters total.\n",
373	  (unsigned int) MAX_COM_LENGTH);
374#ifndef TWO_FILE_COMMANDLINE
375  fprintf(stderr, "You must specify an input JPEG file name when supplying\n");
376  fprintf(stderr, "comment text from standard input.\n");
377#endif
378
379  exit(EXIT_FAILURE);
380}
381
382
383static int
384keymatch (char * arg, const char * keyword, int minchars)
385/* Case-insensitive matching of (possibly abbreviated) keyword switches. */
386/* keyword is the constant keyword (must be lower case already), */
387/* minchars is length of minimum legal abbreviation. */
388{
389  register int ca, ck;
390  register int nmatched = 0;
391
392  while ((ca = *arg++) != '\0') {
393    if ((ck = *keyword++) == '\0')
394      return 0;			/* arg longer than keyword, no good */
395    if (isupper(ca))		/* force arg to lcase (assume ck is already) */
396      ca = tolower(ca);
397    if (ca != ck)
398      return 0;			/* no good */
399    nmatched++;			/* count matched characters */
400  }
401  /* reached end of argument; fail if it's too short for unique abbrev */
402  if (nmatched < minchars)
403    return 0;
404  return 1;			/* A-OK */
405}
406
407
408/*
409 * The main program.
410 */
411
412int
413main (int argc, char **argv)
414{
415  int argn;
416  char * arg;
417  int keep_COM = 1;
418  char * comment_arg = NULL;
419  FILE * comment_file = NULL;
420  unsigned int comment_length = 0;
421  int marker;
422
423  /* On Mac, fetch a command line. */
424#ifdef USE_CCOMMAND
425  argc = ccommand(&argv);
426#endif
427
428  progname = argv[0];
429  if (progname == NULL || progname[0] == 0)
430    progname = "wrjpgcom";	/* in case C library doesn't provide it */
431
432  /* Parse switches, if any */
433  for (argn = 1; argn < argc; argn++) {
434    arg = argv[argn];
435    if (arg[0] != '-')
436      break;			/* not switch, must be file name */
437    arg++;			/* advance over '-' */
438    if (keymatch(arg, "replace", 1)) {
439      keep_COM = 0;
440    } else if (keymatch(arg, "cfile", 2)) {
441      if (++argn >= argc) usage();
442      if ((comment_file = fopen(argv[argn], "r")) == NULL) {
443	fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
444	exit(EXIT_FAILURE);
445      }
446    } else if (keymatch(arg, "comment", 1)) {
447      if (++argn >= argc) usage();
448      comment_arg = argv[argn];
449      /* If the comment text starts with '"', then we are probably running
450       * under MS-DOG and must parse out the quoted string ourselves.  Sigh.
451       */
452      if (comment_arg[0] == '"') {
453	comment_arg = (char *) malloc((size_t) MAX_COM_LENGTH);
454	if (comment_arg == NULL)
455	  ERREXIT("Insufficient memory");
456	strcpy(comment_arg, argv[argn]+1);
457	for (;;) {
458	  comment_length = (unsigned int) strlen(comment_arg);
459	  if (comment_length > 0 && comment_arg[comment_length-1] == '"') {
460	    comment_arg[comment_length-1] = '\0'; /* zap terminating quote */
461	    break;
462	  }
463	  if (++argn >= argc)
464	    ERREXIT("Missing ending quote mark");
465	  strcat(comment_arg, " ");
466	  strcat(comment_arg, argv[argn]);
467	}
468      }
469      comment_length = (unsigned int) strlen(comment_arg);
470    } else
471      usage();
472  }
473
474  /* Cannot use both -comment and -cfile. */
475  if (comment_arg != NULL && comment_file != NULL)
476    usage();
477  /* If there is neither -comment nor -cfile, we will read the comment text
478   * from stdin; in this case there MUST be an input JPEG file name.
479   */
480  if (comment_arg == NULL && comment_file == NULL && argn >= argc)
481    usage();
482
483  /* Open the input file. */
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  /* Open the output file. */
505#ifdef TWO_FILE_COMMANDLINE
506  /* Must have explicit output file name */
507  if (argn != argc-2) {
508    fprintf(stderr, "%s: must name one input and one output file\n",
509	    progname);
510    usage();
511  }
512  if ((outfile = fopen(argv[argn+1], WRITE_BINARY)) == NULL) {
513    fprintf(stderr, "%s: can't open %s\n", progname, argv[argn+1]);
514    exit(EXIT_FAILURE);
515  }
516#else
517  /* Unix style: expect zero or one file name */
518  if (argn < argc-1) {
519    fprintf(stderr, "%s: only one input file\n", progname);
520    usage();
521  }
522  /* default output file is stdout */
523#ifdef USE_SETMODE		/* need to hack file mode? */
524  setmode(fileno(stdout), O_BINARY);
525#endif
526#ifdef USE_FDOPEN		/* need to re-open in binary mode? */
527  if ((outfile = fdopen(fileno(stdout), WRITE_BINARY)) == NULL) {
528    fprintf(stderr, "%s: can't open stdout\n", progname);
529    exit(EXIT_FAILURE);
530  }
531#else
532  outfile = stdout;
533#endif
534#endif /* TWO_FILE_COMMANDLINE */
535
536  /* Collect comment text from comment_file or stdin, if necessary */
537  if (comment_arg == NULL) {
538    FILE * src_file;
539    int c;
540
541    comment_arg = (char *) malloc((size_t) MAX_COM_LENGTH);
542    if (comment_arg == NULL)
543      ERREXIT("Insufficient memory");
544    comment_length = 0;
545    src_file = (comment_file != NULL ? comment_file : stdin);
546    while ((c = getc(src_file)) != EOF) {
547      if (comment_length >= (unsigned int) MAX_COM_LENGTH) {
548	fprintf(stderr, "Comment text may not exceed %u bytes\n",
549		(unsigned int) MAX_COM_LENGTH);
550	exit(EXIT_FAILURE);
551      }
552      comment_arg[comment_length++] = (char) c;
553    }
554    if (comment_file != NULL)
555      fclose(comment_file);
556  }
557
558  /* Copy JPEG headers until SOFn marker;
559   * we will insert the new comment marker just before SOFn.
560   * This (a) causes the new comment to appear after, rather than before,
561   * existing comments; and (b) ensures that comments come after any JFIF
562   * or JFXX markers, as required by the JFIF specification.
563   */
564  marker = scan_JPEG_header(keep_COM);
565  /* Insert the new COM marker, but only if nonempty text has been supplied */
566  if (comment_length > 0) {
567    write_marker(M_COM);
568    write_2_bytes(comment_length + 2);
569    while (comment_length > 0) {
570      write_1_byte(*comment_arg++);
571      comment_length--;
572    }
573  }
574  /* Duplicate the remainder of the source file.
575   * Note that any COM markers occuring after SOF will not be touched.
576   */
577  write_marker(marker);
578  copy_rest_of_file();
579
580  /* All done. */
581  exit(EXIT_SUCCESS);
582  return 0;			/* suppress no-return-value warnings */
583}
584