00001 /*
00002  * wrjpgcom.c
00003  *
00004  * Copyright (C) 1994-1997, Thomas G. Lane.
00005  * This file is part of the Independent JPEG Group's software.
00006  * For conditions of distribution and use, see the accompanying README file.
00007  *
00008  * This file contains a very simple stand-alone application that inserts
00009  * user-supplied text as a COM (comment) marker in a JFIF file.
00010  * This may be useful as an example of the minimum logic needed to parse
00011  * JPEG markers.
00012  */
00013 
00014 #define JPEG_CJPEG_DJPEG        /* to get the command-line config symbols */
00015 #include "jinclude.h"           /* get auto-config symbols, <stdio.h> */
00016 
00017 #ifndef HAVE_STDLIB_H           /* <stdlib.h> should declare malloc() */
00018 extern void * malloc ();
00019 #endif
00020 #include <ctype.h>              /* to declare isupper(), tolower() */
00021 #ifdef USE_SETMODE
00022 #include <fcntl.h>              /* to declare setmode()'s parameter macros */
00023 /* If you have setmode() but not <io.h>, just delete this line: */
00024 #include <io.h>                 /* to declare setmode() */
00025 #endif
00026 
00027 #ifdef USE_CCOMMAND             /* command-line reader for Macintosh */
00028 #ifdef __MWERKS__
00029 #include <SIOUX.h>              /* Metrowerks needs this */
00030 #include <console.h>            /* ... and this */
00031 #endif
00032 #ifdef THINK_C
00033 #include <console.h>            /* Think declares it here */
00034 #endif
00035 #endif
00036 
00037 #ifdef DONT_USE_B_MODE          /* define mode parameters for fopen() */
00038 #define READ_BINARY     "r"
00039 #define WRITE_BINARY    "w"
00040 #else
00041 #ifdef VMS                      /* VMS is very nonstandard */
00042 #define READ_BINARY     "rb", "ctx=stm"
00043 #define WRITE_BINARY    "wb", "ctx=stm"
00044 #else                           /* standard ANSI-compliant case */
00045 #define READ_BINARY     "rb"
00046 #define WRITE_BINARY    "wb"
00047 #endif
00048 #endif
00049 
00050 #ifndef EXIT_FAILURE            /* define exit() codes if not provided */
00051 #define EXIT_FAILURE  1
00052 #endif
00053 #ifndef EXIT_SUCCESS
00054 #ifdef VMS
00055 #define EXIT_SUCCESS  1         /* VMS is very nonstandard */
00056 #else
00057 #define EXIT_SUCCESS  0
00058 #endif
00059 #endif
00060 
00061 /* Reduce this value if your malloc() can't allocate blocks up to 64K.
00062  * On DOS, compiling in large model is usually a better solution.
00063  */
00064 
00065 #ifndef MAX_COM_LENGTH
00066 #define MAX_COM_LENGTH 65000L   /* must be <= 65533 in any case */
00067 #endif
00068 
00069 
00070 /*
00071  * These macros are used to read the input file and write the output file.
00072  * To reuse this code in another application, you might need to change these.
00073  */
00074 
00075 static FILE * infile;           /* input JPEG file */
00076 
00077 /* Return next input byte, or EOF if no more */
00078 #define NEXTBYTE()  getc(infile)
00079 
00080 static FILE * outfile;          /* output JPEG file */
00081 
00082 /* Emit an output byte */
00083 #define PUTBYTE(x)  putc((x), outfile)
00084 
00085 
00086 /* Error exit handler */
00087 #define ERREXIT(msg)  (fprintf(stderr, "%s\n", msg), exit(EXIT_FAILURE))
00088 
00089 
00090 /* Read one byte, testing for EOF */
00091 static int
00092 read_1_byte (void)
00093 {
00094   int c;
00095 
00096   c = NEXTBYTE();
00097   if (c == EOF)
00098     ERREXIT("Premature EOF in JPEG file");
00099   return c;
00100 }
00101 
00102 /* Read 2 bytes, convert to unsigned int */
00103 /* All 2-byte quantities in JPEG markers are MSB first */
00104 static unsigned int
00105 read_2_bytes (void)
00106 {
00107   int c1, c2;
00108 
00109   c1 = NEXTBYTE();
00110   if (c1 == EOF)
00111     ERREXIT("Premature EOF in JPEG file");
00112   c2 = NEXTBYTE();
00113   if (c2 == EOF)
00114     ERREXIT("Premature EOF in JPEG file");
00115   return (((unsigned int) c1) << 8) + ((unsigned int) c2);
00116 }
00117 
00118 
00119 /* Routines to write data to output file */
00120 
00121 static void
00122 write_1_byte (int c)
00123 {
00124   PUTBYTE(c);
00125 }
00126 
00127 static void
00128 write_2_bytes (unsigned int val)
00129 {
00130   PUTBYTE((val >> 8) & 0xFF);
00131   PUTBYTE(val & 0xFF);
00132 }
00133 
00134 static void
00135 write_marker (int marker)
00136 {
00137   PUTBYTE(0xFF);
00138   PUTBYTE(marker);
00139 }
00140 
00141 static void
00142 copy_rest_of_file (void)
00143 {
00144   int c;
00145 
00146   while ((c = NEXTBYTE()) != EOF)
00147     PUTBYTE(c);
00148 }
00149 
00150 
00151 /*
00152  * JPEG markers consist of one or more 0xFF bytes, followed by a marker
00153  * code byte (which is not an FF).  Here are the marker codes of interest
00154  * in this program.  (See jdmarker.c for a more complete list.)
00155  */
00156 
00157 #define M_SOF0  0xC0            /* Start Of Frame N */
00158 #define M_SOF1  0xC1            /* N indicates which compression process */
00159 #define M_SOF2  0xC2            /* Only SOF0-SOF2 are now in common use */
00160 #define M_SOF3  0xC3
00161 #define M_SOF5  0xC5            /* NB: codes C4 and CC are NOT SOF markers */
00162 #define M_SOF6  0xC6
00163 #define M_SOF7  0xC7
00164 #define M_SOF9  0xC9
00165 #define M_SOF10 0xCA
00166 #define M_SOF11 0xCB
00167 #define M_SOF13 0xCD
00168 #define M_SOF14 0xCE
00169 #define M_SOF15 0xCF
00170 #define M_SOI   0xD8            /* Start Of Image (beginning of datastream) */
00171 #define M_EOI   0xD9            /* End Of Image (end of datastream) */
00172 #define M_SOS   0xDA            /* Start Of Scan (begins compressed data) */
00173 #define M_COM   0xFE            /* COMment */
00174 
00175 
00176 /*
00177  * Find the next JPEG marker and return its marker code.
00178  * We expect at least one FF byte, possibly more if the compressor used FFs
00179  * to pad the file.  (Padding FFs will NOT be replicated in the output file.)
00180  * There could also be non-FF garbage between markers.  The treatment of such
00181  * garbage is unspecified; we choose to skip over it but emit a warning msg.
00182  * NB: this routine must not be used after seeing SOS marker, since it will
00183  * not deal correctly with FF/00 sequences in the compressed image data...
00184  */
00185 
00186 static int
00187 next_marker (void)
00188 {
00189   int c;
00190   int discarded_bytes = 0;
00191 
00192   /* Find 0xFF byte; count and skip any non-FFs. */
00193   c = read_1_byte();
00194   while (c != 0xFF) {
00195     discarded_bytes++;
00196     c = read_1_byte();
00197   }
00198   /* Get marker code byte, swallowing any duplicate FF bytes.  Extra FFs
00199    * are legal as pad bytes, so don't count them in discarded_bytes.
00200    */
00201   do {
00202     c = read_1_byte();
00203   } while (c == 0xFF);
00204 
00205   if (discarded_bytes != 0) {
00206     fprintf(stderr, "Warning: garbage data found in JPEG file\n");
00207   }
00208 
00209   return c;
00210 }
00211 
00212 
00213 /*
00214  * Read the initial marker, which should be SOI.
00215  * For a JFIF file, the first two bytes of the file should be literally
00216  * 0xFF M_SOI.  To be more general, we could use next_marker, but if the
00217  * input file weren't actually JPEG at all, next_marker might read the whole
00218  * file and then return a misleading error message...
00219  */
00220 
00221 static int
00222 first_marker (void)
00223 {
00224   int c1, c2;
00225 
00226   c1 = NEXTBYTE();
00227   c2 = NEXTBYTE();
00228   if (c1 != 0xFF || c2 != M_SOI)
00229     ERREXIT("Not a JPEG file");
00230   return c2;
00231 }
00232 
00233 
00234 /*
00235  * Most types of marker are followed by a variable-length parameter segment.
00236  * This routine skips over the parameters for any marker we don't otherwise
00237  * want to process.
00238  * Note that we MUST skip the parameter segment explicitly in order not to
00239  * be fooled by 0xFF bytes that might appear within the parameter segment;
00240  * such bytes do NOT introduce new markers.
00241  */
00242 
00243 static void
00244 copy_variable (void)
00245 /* Copy an unknown or uninteresting variable-length marker */
00246 {
00247   unsigned int length;
00248 
00249   /* Get the marker parameter length count */
00250   length = read_2_bytes();
00251   write_2_bytes(length);
00252   /* Length includes itself, so must be at least 2 */
00253   if (length < 2)
00254     ERREXIT("Erroneous JPEG marker length");
00255   length -= 2;
00256   /* Skip over the remaining bytes */
00257   while (length > 0) {
00258     write_1_byte(read_1_byte());
00259     length--;
00260   }
00261 }
00262 
00263 static void
00264 skip_variable (void)
00265 /* Skip over an unknown or uninteresting variable-length marker */
00266 {
00267   unsigned int length;
00268 
00269   /* Get the marker parameter length count */
00270   length = read_2_bytes();
00271   /* Length includes itself, so must be at least 2 */
00272   if (length < 2)
00273     ERREXIT("Erroneous JPEG marker length");
00274   length -= 2;
00275   /* Skip over the remaining bytes */
00276   while (length > 0) {
00277     (void) read_1_byte();
00278     length--;
00279   }
00280 }
00281 
00282 
00283 /*
00284  * Parse the marker stream until SOFn or EOI is seen;
00285  * copy data to output, but discard COM markers unless keep_COM is true.
00286  */
00287 
00288 static int
00289 scan_JPEG_header (int keep_COM)
00290 {
00291   int marker;
00292 
00293   /* Expect SOI at start of file */
00294   if (first_marker() != M_SOI)
00295     ERREXIT("Expected SOI marker first");
00296   write_marker(M_SOI);
00297 
00298   /* Scan miscellaneous markers until we reach SOFn. */
00299   for (;;) {
00300     marker = next_marker();
00301     switch (marker) {
00302       /* Note that marker codes 0xC4, 0xC8, 0xCC are not, and must not be,
00303        * treated as SOFn.  C4 in particular is actually DHT.
00304        */
00305     case M_SOF0:                /* Baseline */
00306     case M_SOF1:                /* Extended sequential, Huffman */
00307     case M_SOF2:                /* Progressive, Huffman */
00308     case M_SOF3:                /* Lossless, Huffman */
00309     case M_SOF5:                /* Differential sequential, Huffman */
00310     case M_SOF6:                /* Differential progressive, Huffman */
00311     case M_SOF7:                /* Differential lossless, Huffman */
00312     case M_SOF9:                /* Extended sequential, arithmetic */
00313     case M_SOF10:               /* Progressive, arithmetic */
00314     case M_SOF11:               /* Lossless, arithmetic */
00315     case M_SOF13:               /* Differential sequential, arithmetic */
00316     case M_SOF14:               /* Differential progressive, arithmetic */
00317     case M_SOF15:               /* Differential lossless, arithmetic */
00318       return marker;
00319 
00320     case M_SOS:                 /* should not see compressed data before SOF */
00321       ERREXIT("SOS without prior SOFn");
00322       break;
00323 
00324     case M_EOI:                 /* in case it's a tables-only JPEG stream */
00325       return marker;
00326 
00327     case M_COM:                 /* Existing COM: conditionally discard */
00328       if (keep_COM) {
00329         write_marker(marker);
00330         copy_variable();
00331       } else {
00332         skip_variable();
00333       }
00334       break;
00335 
00336     default:                    /* Anything else just gets copied */
00337       write_marker(marker);
00338       copy_variable();          /* we assume it has a parameter count... */
00339       break;
00340     }
00341   } /* end loop */
00342 }
00343 
00344 
00345 /* Command line parsing code */
00346 
00347 static const char * progname;   /* program name for error messages */
00348 
00349 
00350 static void
00351 usage (void)
00352 /* complain about bad command line */
00353 {
00354   fprintf(stderr, "wrjpgcom inserts a textual comment in a JPEG file.\n");
00355   fprintf(stderr, "You can add to or replace any existing comment(s).\n");
00356 
00357   fprintf(stderr, "Usage: %s [switches] ", progname);
00358 #ifdef TWO_FILE_COMMANDLINE
00359   fprintf(stderr, "inputfile outputfile\n");
00360 #else
00361   fprintf(stderr, "[inputfile]\n");
00362 #endif
00363 
00364   fprintf(stderr, "Switches (names may be abbreviated):\n");
00365   fprintf(stderr, "  -replace         Delete any existing comments\n");
00366   fprintf(stderr, "  -comment \"text\"  Insert comment with given text\n");
00367   fprintf(stderr, "  -cfile name      Read comment from named file\n");
00368   fprintf(stderr, "Notice that you must put quotes around the comment text\n");
00369   fprintf(stderr, "when you use -comment.\n");
00370   fprintf(stderr, "If you do not give either -comment or -cfile on the command line,\n");
00371   fprintf(stderr, "then the comment text is read from standard input.\n");
00372   fprintf(stderr, "It can be multiple lines, up to %u characters total.\n",
00373           (unsigned int) MAX_COM_LENGTH);
00374 #ifndef TWO_FILE_COMMANDLINE
00375   fprintf(stderr, "You must specify an input JPEG file name when supplying\n");
00376   fprintf(stderr, "comment text from standard input.\n");
00377 #endif
00378 
00379   exit(EXIT_FAILURE);
00380 }
00381 
00382 
00383 static int
00384 keymatch (char * arg, const char * keyword, int minchars)
00385 /* Case-insensitive matching of (possibly abbreviated) keyword switches. */
00386 /* keyword is the constant keyword (must be lower case already), */
00387 /* minchars is length of minimum legal abbreviation. */
00388 {
00389   register int ca, ck;
00390   register int nmatched = 0;
00391 
00392   while ((ca = *arg++) != '\0') {
00393     if ((ck = *keyword++) == '\0')
00394       return 0;                 /* arg longer than keyword, no good */
00395     if (isupper(ca))            /* force arg to lcase (assume ck is already) */
00396       ca = tolower(ca);
00397     if (ca != ck)
00398       return 0;                 /* no good */
00399     nmatched++;                 /* count matched characters */
00400   }
00401   /* reached end of argument; fail if it's too short for unique abbrev */
00402   if (nmatched < minchars)
00403     return 0;
00404   return 1;                     /* A-OK */
00405 }
00406 
00407 
00408 /*
00409  * The main program.
00410  */
00411 
00412 int
00413 main (int argc, char **argv)
00414 {
00415   int argn;
00416   char * arg;
00417   int keep_COM = 1;
00418   char * comment_arg = NULL;
00419   FILE * comment_file = NULL;
00420   unsigned int comment_length = 0;
00421   int marker;
00422 
00423   /* On Mac, fetch a command line. */
00424 #ifdef USE_CCOMMAND
00425   argc = ccommand(&argv);
00426 #endif
00427 
00428   progname = argv[0];
00429   if (progname == NULL || progname[0] == 0)
00430     progname = "wrjpgcom";      /* in case C library doesn't provide it */
00431 
00432   /* Parse switches, if any */
00433   for (argn = 1; argn < argc; argn++) {
00434     arg = argv[argn];
00435     if (arg[0] != '-')
00436       break;                    /* not switch, must be file name */
00437     arg++;                      /* advance over '-' */
00438     if (keymatch(arg, "replace", 1)) {
00439       keep_COM = 0;
00440     } else if (keymatch(arg, "cfile", 2)) {
00441       if (++argn >= argc) usage();
00442       if ((comment_file = fopen(argv[argn], "r")) == NULL) {
00443         fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
00444         exit(EXIT_FAILURE);
00445       }
00446     } else if (keymatch(arg, "comment", 1)) {
00447       if (++argn >= argc) usage();
00448       comment_arg = argv[argn];
00449       /* If the comment text starts with '"', then we are probably running
00450        * under MS-DOG and must parse out the quoted string ourselves.  Sigh.
00451        */
00452       if (comment_arg[0] == '"') {
00453         comment_arg = (char *) malloc((size_t) MAX_COM_LENGTH);
00454         if (comment_arg == NULL)
00455           ERREXIT("Insufficient memory");
00456         strcpy(comment_arg, argv[argn]+1);
00457         for (;;) {
00458           comment_length = (unsigned int) strlen(comment_arg);
00459           if (comment_length > 0 && comment_arg[comment_length-1] == '"') {
00460             comment_arg[comment_length-1] = '\0'; /* zap terminating quote */
00461             break;
00462           }
00463           if (++argn >= argc)
00464             ERREXIT("Missing ending quote mark");
00465           strcat(comment_arg, " ");
00466           strcat(comment_arg, argv[argn]);
00467         }
00468       }
00469       comment_length = (unsigned int) strlen(comment_arg);
00470     } else
00471       usage();
00472   }
00473 
00474   /* Cannot use both -comment and -cfile. */
00475   if (comment_arg != NULL && comment_file != NULL)
00476     usage();
00477   /* If there is neither -comment nor -cfile, we will read the comment text
00478    * from stdin; in this case there MUST be an input JPEG file name.
00479    */
00480   if (comment_arg == NULL && comment_file == NULL && argn >= argc)
00481     usage();
00482 
00483   /* Open the input file. */
00484   if (argn < argc) {
00485     if ((infile = fopen(argv[argn], READ_BINARY)) == NULL) {
00486       fprintf(stderr, "%s: can't open %s\n", progname, argv[argn]);
00487       exit(EXIT_FAILURE);
00488     }
00489   } else {
00490     /* default input file is stdin */
00491 #ifdef USE_SETMODE              /* need to hack file mode? */
00492     setmode(fileno(stdin), O_BINARY);
00493 #endif
00494 #ifdef USE_FDOPEN               /* need to re-open in binary mode? */
00495     if ((infile = fdopen(fileno(stdin), READ_BINARY)) == NULL) {
00496       fprintf(stderr, "%s: can't open stdin\n", progname);
00497       exit(EXIT_FAILURE);
00498     }
00499 #else
00500     infile = stdin;
00501 #endif
00502   }
00503 
00504   /* Open the output file. */
00505 #ifdef TWO_FILE_COMMANDLINE
00506   /* Must have explicit output file name */
00507   if (argn != argc-2) {
00508     fprintf(stderr, "%s: must name one input and one output file\n",
00509             progname);
00510     usage();
00511   }
00512   if ((outfile = fopen(argv[argn+1], WRITE_BINARY)) == NULL) {
00513     fprintf(stderr, "%s: can't open %s\n", progname, argv[argn+1]);
00514     exit(EXIT_FAILURE);
00515   }
00516 #else
00517   /* Unix style: expect zero or one file name */
00518   if (argn < argc-1) {
00519     fprintf(stderr, "%s: only one input file\n", progname);
00520     usage();
00521   }
00522   /* default output file is stdout */
00523 #ifdef USE_SETMODE              /* need to hack file mode? */
00524   setmode(fileno(stdout), O_BINARY);
00525 #endif
00526 #ifdef USE_FDOPEN               /* need to re-open in binary mode? */
00527   if ((outfile = fdopen(fileno(stdout), WRITE_BINARY)) == NULL) {
00528     fprintf(stderr, "%s: can't open stdout\n", progname);
00529     exit(EXIT_FAILURE);
00530   }
00531 #else
00532   outfile = stdout;
00533 #endif
00534 #endif /* TWO_FILE_COMMANDLINE */
00535 
00536   /* Collect comment text from comment_file or stdin, if necessary */
00537   if (comment_arg == NULL) {
00538     FILE * src_file;
00539     int c;
00540 
00541     comment_arg = (char *) malloc((size_t) MAX_COM_LENGTH);
00542     if (comment_arg == NULL)
00543       ERREXIT("Insufficient memory");
00544     comment_length = 0;
00545     src_file = (comment_file != NULL ? comment_file : stdin);
00546     while ((c = getc(src_file)) != EOF) {
00547       if (comment_length >= (unsigned int) MAX_COM_LENGTH) {
00548         fprintf(stderr, "Comment text may not exceed %u bytes\n",
00549                 (unsigned int) MAX_COM_LENGTH);
00550         exit(EXIT_FAILURE);
00551       }
00552       comment_arg[comment_length++] = (char) c;
00553     }
00554     if (comment_file != NULL)
00555       fclose(comment_file);
00556   }
00557 
00558   /* Copy JPEG headers until SOFn marker;
00559    * we will insert the new comment marker just before SOFn.
00560    * This (a) causes the new comment to appear after, rather than before,
00561    * existing comments; and (b) ensures that comments come after any JFIF
00562    * or JFXX markers, as required by the JFIF specification.
00563    */
00564   marker = scan_JPEG_header(keep_COM);
00565   /* Insert the new COM marker, but only if nonempty text has been supplied */
00566   if (comment_length > 0) {
00567     write_marker(M_COM);
00568     write_2_bytes(comment_length + 2);
00569     while (comment_length > 0) {
00570       write_1_byte(*comment_arg++);
00571       comment_length--;
00572     }
00573   }
00574   /* Duplicate the remainder of the source file.
00575    * Note that any COM markers occuring after SOF will not be touched.
00576    */
00577   write_marker(marker);
00578   copy_rest_of_file();
00579 
00580   /* All done. */
00581   exit(EXIT_SUCCESS);
00582   return 0;                     /* suppress no-return-value warnings */
00583 }