bzip2.c

高效率的一种通用压缩/解压程序

      case SM_I2O:
         inStr = stdin;
         outStr = stdout;
         if ( isatty ( fileno ( stdin ) ) ) {
            fprintf ( stderr,
                      "%s: I won't read compressed data from a terminal.\n",
                      progName );
            fprintf ( stderr, "%s: For help, type: `%s --help'.\n",
                              progName, progName );
            return;
         };
         break;

      case SM_F2O:
         inStr = fopen ( inName, "rb" );
         outStr = stdout;
         if ( inStr == NULL ) {
            fprintf ( stderr, "%s: Can't open input file %s, skipping.\n",
                      progName, inName );
            return;
         };
         break;

      case SM_F2F:
         inStr = fopen ( inName, "rb" );
         outStr = fopen ( outName, "wb" );
         if ( outStr == NULL) {
            fprintf ( stderr, "%s: Can't create output file %s, skipping.\n",
                      progName, outName );
            return;
         }
         if ( inStr == NULL ) {
            fprintf ( stderr, "%s: Can't open input file %s, skipping.\n",
                      progName, inName );
            return;
         };
         setInterimPermissions ( outName );
         break;

      default:
         panic ( "uncompress: bad srcMode" );
         break;
   }

   if (verbosity >= 1) {
      fprintf ( stderr, "  %s: ", inName );
      pad ( inName );
      fflush ( stderr );
   }

   /*--- Now the input and output handles are sane.  Do the Biz. ---*/
   outputHandleJustInCase = outStr;
   magicNumberOK = uncompressStream ( inStr, outStr );
   outputHandleJustInCase = NULL;

   /*--- If there was an I/O error, we won't get here. ---*/
   if ( magicNumberOK ) {
      if ( srcMode == SM_F2F ) {
         copyDatePermissionsAndOwner ( inName, outName );
         if ( !keepInputFiles ) {
            IntNative retVal = remove ( inName );
            ERROR_IF_NOT_ZERO ( retVal );
         }
      }
   } else {
      if ( srcMode == SM_F2F ) {
         IntNative retVal = remove ( outName );
         ERROR_IF_NOT_ZERO ( retVal );
      }
   }

   if ( magicNumberOK ) {
      if (verbosity >= 1)
         fprintf ( stderr, "done\n" );
   } else {
      if (verbosity >= 1)
         fprintf ( stderr, "not a bzip2 file, skipping.\n" ); else
         fprintf ( stderr,
                   "%s: %s is not a bzip2 file, skipping.\n",
                   progName, inName );
   }

}


/*---------------------------------------------*/
void testf ( Char *name )
{
   FILE *inStr;
   Bool allOK;

   if (name == NULL && srcMode != SM_I2O)
      panic ( "testf: bad modes\n" );

   copyFileName ( outName, "(none)" );
   switch (srcMode) {
      case SM_I2O: copyFileName ( inName, "(stdin)" ); break;
      case SM_F2F: copyFileName ( inName, name ); break;
      case SM_F2O: copyFileName ( inName, name ); break;
   }

   if ( srcMode != SM_I2O && containsDubiousChars ( inName ) ) {
      fprintf ( stderr, "%s: There are no files matching `%s'.\n",
                progName, inName );
      return;
   }
   if ( srcMode != SM_I2O && !fileExists ( inName ) ) {
      fprintf ( stderr, "%s: Input file %s doesn't exist, skipping.\n",
                progName, inName );
      return;
   }
   if ( srcMode != SM_I2O && !endsInBz2 ( inName )) {
      fprintf ( stderr,
                "%s: Input file name %s doesn't end in `.bz2', skipping.\n",
                progName, inName );
      return;
   }
   if ( srcMode != SM_I2O && notAStandardFile ( inName )) {
      fprintf ( stderr, "%s: Input file %s is not a normal file, skipping.\n",
                progName, inName );
      return;
   }

   switch ( srcMode ) {

      case SM_I2O:
         if ( isatty ( fileno ( stdin ) ) ) {
            fprintf ( stderr,
                      "%s: I won't read compressed data from a terminal.\n",
                      progName );
            fprintf ( stderr, "%s: For help, type: `%s --help'.\n",
                              progName, progName );
            return;
         };
         inStr = stdin;
         break;

      case SM_F2O: case SM_F2F:
         inStr = fopen ( inName, "rb" );
         if ( inStr == NULL ) {
            fprintf ( stderr, "%s: Can't open input file %s, skipping.\n",
                      progName, inName );
            return;
         };
         break;

      default:
         panic ( "testf: bad srcMode" );
         break;
   }

   if (verbosity >= 1) {
      fprintf ( stderr, "  %s: ", inName );
      pad ( inName );
      fflush ( stderr );
   }

   /*--- Now the input handle is sane.  Do the Biz. ---*/
   allOK = testStream ( inStr );

   if (allOK && verbosity >= 1) fprintf ( stderr, "ok\n" );
   if (!allOK) testFailsExist = True;
}


/*---------------------------------------------*/
void license ( void )
{
   fprintf ( stderr,

    "bzip2, a block-sorting file compressor.  "
    "Version 0.9.0c, 18-Oct-98.\n"
    "   \n"
    "   Copyright (C) 1996, 1997, 1998 by Julian Seward.\n"
    "   \n"
    "   This program is free software; you can redistribute it and/or modify\n"
    "   it under the terms set out in the LICENSE file, which is included\n"
    "   in the bzip2-0.9.0c source distribution.\n"
    "   \n"
    "   This program is distributed in the hope that it will be useful,\n"
    "   but WITHOUT ANY WARRANTY; without even the implied warranty of\n"
    "   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n"
    "   LICENSE file for more details.\n"
    "   \n"
   );
}


/*---------------------------------------------*/
void usage ( Char *fullProgName )
{
   fprintf (
      stderr,
      "bzip2, a block-sorting file compressor.  "
      "Version 0.9.0c, 18-Oct-98.\n"
      "\n   usage: %s [flags and input files in any order]\n"
      "\n"
      "   -h --help           print this message\n"
      "   -d --decompress     force decompression\n"
      "   -z --compress       force compression\n"
      "   -k --keep           keep (don't delete) input files\n"
      "   -f --force          overwrite existing output filess\n"
      "   -t --test           test compressed file integrity\n"
      "   -c --stdout         output to standard out\n"
      "   -v --verbose        be verbose (a 2nd -v gives more)\n"
      "   -L --license        display software version & license\n"
      "   -V --version        display software version & license\n"
      "   -s --small          use less memory (at most 2500k)\n"
      "   -1 .. -9            set block size to 100k .. 900k\n"
      "   --repetitive-fast   compress repetitive blocks faster\n"
      "   --repetitive-best   compress repetitive blocks better\n"
      "\n"
      "   If invoked as `bzip2', default action is to compress.\n"
      "              as `bunzip2',  default action is to decompress.\n"
      "              as `bz2cat', default action is to decompress to stdout.\n"
      "\n"
      "   If no file names are given, bzip2 compresses or decompresses\n"
      "   from standard input to standard output.  You can combine\n"
      "   short flags, so `-v -4' means the same as -v4 or -4v, &c.\n"
#if BZ_UNIX
      "\n"
#endif
      ,

      fullProgName
   );
}


/*---------------------------------------------*/
/*--
  All the garbage from here to main() is purely to
  implement a linked list of command-line arguments,
  into which main() copies argv[1 .. argc-1].

  The purpose of this ridiculous exercise is to
  facilitate the expansion of wildcard characters
  * and ? in filenames for halfwitted OSs like
  MSDOS, Windows 95 and NT.

  The actual Dirty Work is done by the platform-specific
  macro APPEND_FILESPEC.
--*/

typedef
   struct zzzz {
      Char        *name;
      struct zzzz *link;
   }
   Cell;


/*---------------------------------------------*/
void *myMalloc ( Int32 n )
{
   void* p;

   p = malloc ( (size_t)n );
   if (p == NULL) outOfMemory ();
   return p;
}


/*---------------------------------------------*/
Cell *mkCell ( void )
{
   Cell *c;

   c = (Cell*) myMalloc ( sizeof ( Cell ) );
   c->name = NULL;
   c->link = NULL;
   return c;
}


/*---------------------------------------------*/
Cell *snocString ( Cell *root, Char *name )
{
   if (root == NULL) {
      Cell *tmp = mkCell();
      tmp->name = (Char*) myMalloc ( 5 + strlen(name) );
      strcpy ( tmp->name, name );
      return tmp;
   } else {
      Cell *tmp = root;
      while (tmp->link != NULL) tmp = tmp->link;
      tmp->link = snocString ( tmp->link, name );
      return root;
   }
}


/*---------------------------------------------*/
#define ISFLAG(s) (strcmp(aa->name, (s))==0)


IntNative main ( IntNative argc, Char *argv[] )
{
   Int32  i, j;
   Char   *tmp;
   Cell   *argList;
   Cell   *aa;

   /*-- Be really really really paranoid :-) --*/
   if (sizeof(Int32) != 4 || sizeof(UInt32) != 4  ||
       sizeof(Int16) != 2 || sizeof(UInt16) != 2  ||
       sizeof(Char)  != 1 || sizeof(UChar)  != 1) {
      fprintf ( stderr,
                "bzip2: I'm not configured correctly for this platform!\n"
                "\tI require Int32, Int16 and Char to have sizes\n"
                "\tof 4, 2 and 1 bytes to run properly, and they don't.\n"
                "\tProbably you can fix this by defining them correctly,\n"
                "\tand recompiling.  Bye!\n" );
      exit(3);
   }


   /*-- Set up signal handlers --*/
   signal (SIGINT,  mySignalCatcher);
   signal (SIGTERM, mySignalCatcher);
   signal (SIGSEGV, mySIGSEGVorSIGBUScatcher);
#if BZ_UNIX
   signal (SIGHUP,  mySignalCatcher);
   signal (SIGBUS,  mySIGSEGVorSIGBUScatcher);
#endif


   /*-- Initialise --*/
   outputHandleJustInCase  = NULL;
   smallMode               = False;
   keepInputFiles          = False;
   forceOverwrite          = False;
   verbosity               = 0;
   blockSize100k           = 9;
   testFailsExist          = False;
   numFileNames            = 0;
   numFilesProcessed       = 0;
   workFactor              = 30;

   copyFileName ( inName,  "(none)" );
   copyFileName ( outName, "(none)" );

   copyFileName ( progNameReally, argv[0] );
   progName = &progNameReally[0];
   for (tmp = &progNameReally[0]; *tmp != '\0'; tmp++)
      if (*tmp == PATH_SEP) progName = tmp + 1;


   /*-- Expand filename wildcards in arg list --*/
   argList = NULL;
   for (i = 1; i <= argc-1; i++)
      APPEND_FILESPEC(argList, argv[i]);


   /*-- Find the length of the longest filename --*/
   longestFileName = 7;
   numFileNames    = 0;
   for (aa = argList; aa != NULL; aa = aa->link)
      if (aa->name[0] != '-') {
         numFileNames++;
         if (longestFileName < (Int32)strlen(aa->name) )
            longestFileName = (Int32)strlen(aa->name);
      }


   /*-- Determine source modes; flag handling may change this too. --*/
   if (numFileNames == 0)
      srcMode = SM_I2O; else srcMode = SM_F2F;


   /*-- Determine what to do (compress/uncompress/test/cat). --*/
   /*-- Note that subsequent flag handling may change this. --*/
   opMode = OM_Z;

   if ( (strstr ( progName, "unzip" ) != 0) ||
        (strstr ( progName, "UNZIP" ) != 0) )
      opMode = OM_UNZ;

   if ( (strstr ( progName, "z2cat" ) != 0) ||
        (strstr ( progName, "Z2CAT" ) != 0) ||
        (strstr ( progName, "zcat" ) != 0)  ||
        (strstr ( progName, "ZCAT" ) != 0) )  {
      opMode = OM_UNZ;
      srcMode = (numFileNames == 0) ? SM_I2O : SM_F2O;
   }


   /*-- Look at the flags. --*/
   for (aa = argList; aa != NULL; aa = aa->link)
      if (aa->name[0] == '-' && aa->name[1] != '-')
         for (j = 1; aa->name[j] != '\0'; j++)
            switch (aa->name[j]) {
               case 'c': srcMode          = SM_F2O; break;
               case 'd': opMode           = OM_UNZ; break;
               case 'z': opMode           = OM_Z; break;
               case 'f': forceOverwrite   = True; break;
               case 't': opMode           = OM_TEST; break;
               case 'k': keepInputFiles   = True; break;
               case 's': smallMode        = True; break;
               case '1': blockSize100k    = 1; break;
               case '2': blockSize100k    = 2; break;
               case '3': blockSize100k    = 3; break;
               case '4': blockSize100k    = 4; break;
               case '5': blockSize100k    = 5; break;
               case '6': blockSize100k    = 6; break;
               case '7': blockSize100k    = 7; break;
               case '8': blockSize100k    = 8; break;
               case '9': blockSize100k    = 9; break;
               case 'V':
               case 'L': license();            break;
               case 'v': verbosity++; break;
               case 'h': usage ( progName );
                         exit ( 1 );
                         break;
               default:  fprintf ( stderr, "%s: Bad flag `%s'\n",
                                   progName, aa->name );
                         usage ( progName );
                         exit ( 1 );
                         break;
         }

   /*-- And again ... --*/
   for (aa = argList; aa != NULL; aa = aa->link) {
      if (ISFLAG("--stdout"))            srcMode          = SM_F2O;  else
      if (ISFLAG("--decompress"))        opMode           = OM_UNZ;  else
      if (ISFLAG("--compress"))          opMode           = OM_Z;    else
      if (ISFLAG("--force"))             forceOverwrite   = True;    else
      if (ISFLAG("--test"))              opMode           = OM_TEST; else
      if (ISFLAG("--keep"))              keepInputFiles   = True;    else
      if (ISFLAG("--small"))             smallMode        = True;    else
      if (ISFLAG("--version"))           license();                  else
      if (ISFLAG("--license"))           license();                  else
      if (ISFLAG("--repetitive-fast"))   workFactor = 5;             else
      if (ISFLAG("--repetitive-best"))   workFactor = 150;           else
      if (ISFLAG("--verbose"))           verbosity++;                else
      if (ISFLAG("--help"))              { usage ( progName ); exit ( 1 ); }
         else
         if (strncmp ( aa->name, "--", 2) == 0) {
            fprintf ( stderr, "%s: Bad flag `%s'\n", progName, aa->name );
            usage ( progName );
            exit ( 1 );
         }
   }

   if (verbosity > 4) verbosity = 4;
   if (opMode == OM_Z && smallMode) blockSize100k = 2;

   if (srcMode == SM_F2O && numFileNames == 0) {
      fprintf ( stderr, "%s: -c expects at least one filename.\n",
                progName );
      exit ( 1 );
   }

   if (opMode == OM_TEST && srcMode == SM_F2O) {
      fprintf ( stderr, "%s: -c and -t cannot be used together.\n",
                progName );
      exit ( 1 );
   }

   if (opMode != OM_Z) blockSize100k = 0;

   if (opMode == OM_Z) {
      if (srcMode == SM_I2O)
         compress ( NULL );
         else
         for (aa = argList; aa != NULL; aa = aa->link)
            if (aa->name[0] != '-') {
               numFilesProcessed++;
               compress ( aa->name );
            }
   } else
   if (opMode == OM_UNZ) {
      if (srcMode == SM_I2O)
         uncompress ( NULL );
         else
         for (aa = argList; aa != NULL; aa = aa->link)
            if (aa->name[0] != '-') {
               numFilesProcessed++;
               uncompress ( aa->name );
            }
   } else {
      testFailsExist = False;
      if (srcMode == SM_I2O)
         testf ( NULL );
         else
         for (aa = argList; aa != NULL; aa = aa->link)
            if (aa->name[0] != '-') {
               numFilesProcessed++;
               testf ( aa->name );
            }
      if (testFailsExist) {
         fprintf ( stderr,
           "\n"
           "You can use the `bzip2recover' program to *attempt* to recover\n"
           "data from undamaged sections of corrupted files.\n\n"
         );
         exit(2);
      }
   }
   return 0;
}


/*-----------------------------------------------------------*/
/*--- end                                         bzip2.c ---*/
/*-----------------------------------------------------------*/