/[pcre]/code/trunk/pcregrep.c

Diff of /code/trunk/pcregrep.c

Parent Directory | Revision Log | View Patch Patch

-revision 280 by ph10,
Wed Dec  5 20:56:03 2007 UTC
+revision 324 by ph10,
Fri Mar  7 19:48:32 2008 UTC
 Line 6
  its pattern matching. On a Unix or Win32 system it can recurse into
  directories.
-            Copyright (c) 1997-2007 University of Cambridge
+            Copyright (c) 1997-2008 University of Cambridge
  -----------------------------------------------------------------------------
  Redistribution and use in source and binary forms, with or without
 Line 55 
 POSSIBILITY OF SUCH DAMAGE.
  #include <unistd.h>
  #endif
+ #ifdef SUPPORT_LIBZ
+ #include <zlib.h>
+ #endif
+ #ifdef SUPPORT_LIBBZ2
+ #include <bzlib.h>
+ #endif
  #include "pcre.h"
  #define FALSE 0
-Line 76 
 all values greater than FN_DEFAULT. */
+Line 84 
 all values greater than FN_DEFAULT. */
  enum { FN_NONE, FN_DEFAULT, FN_ONLY, FN_NOMATCH_ONLY, FN_FORCE };
+ /* File reading styles */
+ enum { FR_PLAIN, FR_LIBZ, FR_LIBBZ2 };
  /* Actions for the -d and -D options */
  enum { dee_READ, dee_SKIP, dee_RECURSE };
-Line 322 
 return isatty(fileno(stdout));
+Line 334 
 return isatty(fileno(stdout));
  /* I (Philip Hazel) have no means of testing this code. It was contributed by
  Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES
- when it did not exist. */
+ when it did not exist. David Byron added a patch that moved the #include of
+ <windows.h> to before the INVALID_FILE_ATTRIBUTES definition rather than after.
+ */
  #elif HAVE_WINDOWS_H
-Line 333 
 when it did not exist. */
+Line 346 
 when it did not exist. */
  #ifndef WIN32_LEAN_AND_MEAN
  # define WIN32_LEAN_AND_MEAN
  #endif
+ #include <windows.h>
  #ifndef INVALID_FILE_ATTRIBUTES
  #define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF
  #endif
- #include <windows.h>
  typedef struct directory_type
  {
  HANDLE handle;
-Line 423 
 regular if they are not directories. */
+Line 437 
 regular if they are not directories. */
  int isregfile(char *filename)
  {
- return !isdirectory(filename)
+ return !isdirectory(filename);
  }
-Line 434 
 return !isdirectory(filename)
+Line 448 
 return !isdirectory(filename)
  static BOOL
  is_stdout_tty(void)
  {
- FALSE;
+ return FALSE;
  }
-Line 810 
 be in the middle third most of the time,
+Line 824 
 be in the middle third most of the time,
  "before" context printing.
  Arguments:
-   in           the fopened FILE stream
+   handle       the fopened FILE stream for a normal file
+                the gzFile pointer when reading is via libz
+                the BZFILE pointer when reading is via libbz2
+   frtype       FR_PLAIN, FR_LIBZ, or FR_LIBBZ2
    printname    the file name if it is to be printed for each match
                 or NULL if the file name is not to be printed
                 it cannot be NULL if filenames[_nomatch]_only is set
  Returns:       0 if there was at least one match
 otherwise (no matches)
+if there is a read error on a .bz2 file
  */
  static int
- pcregrep(FILE *in, char *printname)
+ pcregrep(void *handle, int frtype, char *printname)
  {
  int rc = 1;
  int linenumber = 1;
-Line 834 
 char *ptr = buffer;
+Line 852 
 char *ptr = buffer;
  char *endptr;
  size_t bufflength;
  BOOL endhyphenpending = FALSE;
+ FILE *in = NULL;                    /* Ensure initialized */
+ #ifdef SUPPORT_LIBZ
+ gzFile ingz = NULL;
+ #endif
- /* Do the first read into the start of the buffer and set up the pointer to
+ #ifdef SUPPORT_LIBBZ2
- end of what we have. */
+ BZFILE *inbz2 = NULL;
+ #endif
+ /* Do the first read into the start of the buffer and set up the pointer to end
+ of what we have. In the case of libz, a non-zipped .gz file will be read as a
+ plain file. However, if a .bz2 file isn't actually bzipped, the first read will
+ fail. */
+ #ifdef SUPPORT_LIBZ
+ if (frtype == FR_LIBZ)
+   {
+   ingz = (gzFile)handle;
+   bufflength = gzread (ingz, buffer, 3*MBUFTHIRD);
+   }
+ else
+ #endif
+ #ifdef SUPPORT_LIBBZ2
+ if (frtype == FR_LIBBZ2)
+   {
+   inbz2 = (BZFILE *)handle;
+   bufflength = BZ2_bzread(inbz2, buffer, 3*MBUFTHIRD);
+   if ((int)bufflength < 0) return 2;   /* Gotcha: bufflength is size_t; */
+   }                                    /* without the cast it is unsigned. */
+ else
+ #endif
+   {
+   in = (FILE *)handle;
+   bufflength = fread(buffer, 1, 3*MBUFTHIRD, in);
+   }
- bufflength = fread(buffer, 1, 3*MBUFTHIRD, in);
  endptr = buffer + bufflength;
  /* Loop while the current pointer is not at the end of the file. For large
-Line 851 
 while (ptr < endptr)
+Line 904 
 while (ptr < endptr)
    int i, endlinelength;
    int mrc = 0;
    BOOL match = FALSE;
    char *matchptr = ptr;
    char *t = ptr;
    size_t length, linelength;
-Line 914 
 while (ptr < endptr)
+Line 967 
 while (ptr < endptr)
    }
  #endif
    /* We come back here after a match when the -o option (only_matching) is set,
    in order to find any further matches in the same line. */
    ONLY_MATCHING_RESTART:
    /* Run through all the patterns until one matches. Note that we don't include
    the final newline in the subject string. */
-Line 979 
 while (ptr < endptr)
+Line 1032 
 while (ptr < endptr)
      else if (quiet) return 0;
      /* The --only-matching option prints just the substring that matched, and
      the --file-offsets and --line-offsets options output offsets for the
      matching substring (they both force --only-matching). None of these options
      prints any context. Afterwards, adjust the start and length, and then jump
      back to look for further matches in the same line. If we are in invert
-Line 989 
 while (ptr < endptr)
+Line 1042 
 while (ptr < endptr)
      else if (only_matching)
        {
        if (!invert)
          {
          if (printname != NULL) fprintf(stdout, "%s:", printname);
          if (number) fprintf(stdout, "%d:", linenumber);
          if (line_offsets)
            fprintf(stdout, "%d,%d", matchptr + offsets[0] - ptr,
              offsets[1] - offsets[0]);
          else if (file_offsets)
            fprintf(stdout, "%d,%d", filepos + matchptr + offsets[0] - ptr,
              offsets[1] - offsets[0]);
          else
            fwrite(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
          fprintf(stdout, "\n");
          matchptr += offsets[1];
          length -= offsets[1];
          match = FALSE;
          goto ONLY_MATCHING_RESTART;
          }
        }
      /* This is the default case when none of the above options is set. We print
-Line 1179 
 while (ptr < endptr)
+Line 1232 
 while (ptr < endptr)
      linelength = endmatch - ptr - ellength;
      }
    /* Advance to after the newline and increment the line number. The file
    offset to the current line is maintained in filepos. */
    ptr += linelength + endlinelength;
-Line 1205 
 while (ptr < endptr)
+Line 1258 
 while (ptr < endptr)
      memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD);
      ptr -= MBUFTHIRD;
+ #ifdef SUPPORT_LIBZ
+     if (frtype == FR_LIBZ)
+       bufflength = 2*MBUFTHIRD +
+         gzread (ingz, buffer + 2*MBUFTHIRD, MBUFTHIRD);
+     else
+ #endif
+ #ifdef SUPPORT_LIBBZ2
+     if (frtype == FR_LIBBZ2)
+       bufflength = 2*MBUFTHIRD +
+         BZ2_bzread(inbz2, buffer + 2*MBUFTHIRD, MBUFTHIRD);
+     else
+ #endif
      bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in);
      endptr = buffer + bufflength;
      /* Adjust any last match point */
-Line 1269 
 grep_or_recurse(char *pathname, BOOL dir
+Line 1338 
 grep_or_recurse(char *pathname, BOOL dir
  {
  int rc = 1;
  int sep;
- FILE *in;
+ int frtype;
+ int pathlen;
+ void *handle;
+ FILE *in = NULL;           /* Ensure initialized */
+ #ifdef SUPPORT_LIBZ
+ gzFile ingz = NULL;
+ #endif
+ #ifdef SUPPORT_LIBBZ2
+ BZFILE *inbz2 = NULL;
+ #endif
  /* If the file name is "-" we scan stdin */
  if (strcmp(pathname, "-") == 0)
    {
-   return pcregrep(stdin,
+   return pcregrep(stdin, FR_PLAIN,
      (filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))?
        stdin_name : NULL);
    }
  /* If the file is a directory, skip if skipping or if we are recursing, scan
  each file within it, subject to any include or exclude patterns that were set.
  The scanning code is localized so it can be made system-specific. */
-Line 1304 
 if ((sep = isdirectory(pathname)) != 0)
+Line 1383 
 if ((sep = isdirectory(pathname)) != 0)
      while ((nextfile = readdirectory(dir)) != NULL)
        {
-       int frc, blen;
+       int frc, nflen;
        sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile);
-       blen = strlen(buffer);
+       nflen = strlen(nextfile);
-       if (exclude_compiled != NULL &&
+       if (!isdirectory(buffer))
-           pcre_exec(exclude_compiled, NULL, buffer, blen, 0, 0, NULL, 0) >= 0)
+         {
-         continue;
+         if (exclude_compiled != NULL &&
+             pcre_exec(exclude_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0)
-       if (include_compiled != NULL &&
+           continue;
-           pcre_exec(include_compiled, NULL, buffer, blen, 0, 0, NULL, 0) < 0)
-         continue;
+         if (include_compiled != NULL &&
+             pcre_exec(include_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0)
+           continue;
+         }
        frc = grep_or_recurse(buffer, dir_recurse, FALSE);
        if (frc > 1) rc = frc;
-Line 1337 
 skipping was not requested. The scan pro
+Line 1419 
 skipping was not requested. The scan pro
  argument at top level, we don't show the file name, unless we are only showing
  the file name, or the filename was forced (-H). */
- in = fopen(pathname, "r");
+ pathlen = strlen(pathname);
- if (in == NULL)
+ /* Open using zlib if it is supported and the file name ends with .gz. */
+ #ifdef SUPPORT_LIBZ
+ if (pathlen > 3 && strcmp(pathname + pathlen - 3, ".gz") == 0)
+   {
+   ingz = gzopen(pathname, "rb");
+   if (ingz == NULL)
+     {
+     if (!silent)
+       fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
+         strerror(errno));
+     return 2;
+     }
+   handle = (void *)ingz;
+   frtype = FR_LIBZ;
+   }
+ else
+ #endif
+ /* Otherwise open with bz2lib if it is supported and the name ends with .bz2. */
+ #ifdef SUPPORT_LIBBZ2
+ if (pathlen > 4 && strcmp(pathname + pathlen - 4, ".bz2") == 0)
+   {
+   inbz2 = BZ2_bzopen(pathname, "rb");
+   handle = (void *)inbz2;
+   frtype = FR_LIBBZ2;
+   }
+ else
+ #endif
+ /* Otherwise use plain fopen(). The label is so that we can come back here if
+ an attempt to read a .bz2 file indicates that it really is a plain file. */
+ #ifdef SUPPORT_LIBBZ2
+ PLAIN_FILE:
+ #endif
+   {
+   in = fopen(pathname, "r");
+   handle = (void *)in;
+   frtype = FR_PLAIN;
+   }
+ /* All the opening methods return errno when they fail. */
+ if (handle == NULL)
    {
    if (!silent)
      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
-Line 1346 
 if (in == NULL)
+Line 1474 
 if (in == NULL)
    return 2;
    }
- rc = pcregrep(in, (filenames > FN_DEFAULT ||
+ /* Now grep the file */
+ rc = pcregrep(handle, frtype, (filenames > FN_DEFAULT ||
    (filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL);
+ /* Close in an appropriate manner. */
+ #ifdef SUPPORT_LIBZ
+ if (frtype == FR_LIBZ)
+   gzclose(ingz);
+ else
+ #endif
+ /* If it is a .bz2 file and the result is 2, it means that the first attempt to
+ read failed. If the error indicates that the file isn't in fact bzipped, try
+ again as a normal file. */
+ #ifdef SUPPORT_LIBBZ2
+ if (frtype == FR_LIBBZ2)
+   {
+   if (rc == 2)
+     {
+     int errnum;
+     const char *err = BZ2_bzerror(inbz2, &errnum);
+     if (errnum == BZ_DATA_ERROR_MAGIC)
+       {
+       BZ2_bzclose(inbz2);
+       goto PLAIN_FILE;
+       }
+     else if (!silent)
+       fprintf(stderr, "pcregrep: Failed to read %s using bzlib: %s\n",
+         pathname, err);
+     }
+   BZ2_bzclose(inbz2);
+   }
+ else
+ #endif
+ /* Normal file close */
  fclose(in);
+ /* Pass back the yield from pcregrep(). */
  return rc;
  }
-Line 1390 
 option_item *op;
+Line 1558 
 option_item *op;
  printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n");
  printf("Search for PATTERN in each FILE or standard input.\n");
  printf("PATTERN must be present if neither -e nor -f is used.\n");
- printf("\"-\" can be used as a file name to mean STDIN.\n\n");
+ printf("\"-\" can be used as a file name to mean STDIN.\n");
- printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n");
+ #ifdef SUPPORT_LIBZ
+ printf("Files whose names end in .gz are read using zlib.\n");
+ #endif
+ #ifdef SUPPORT_LIBBZ2
+ printf("Files whose names end in .bz2 are read using bzlib2.\n");
+ #endif
+ #if defined SUPPORT_LIBZ || defined SUPPORT_LIBBZ2
+ printf("Other files and the standard input are read as plain files.\n\n");
+ #else
+ printf("All files are read as plain files, without any interpretation.\n\n");
+ #endif
+ printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n");
  printf("Options:\n");
  for (op = optionlist; op->one_char != 0; op++)
-Line 1400 
 for (op = optionlist; op->one_char != 0;
+Line 1582 
 for (op = optionlist; op->one_char != 0;
    int n;
    char s[4];
    if (op->one_char > 0) sprintf(s, "-%c,", op->one_char); else strcpy(s, "   ");
-   printf("  %s --%s%n", s, op->long_name, &n);
+   n = 30 - printf("  %s --%s", s, op->long_name);
-   n = 30 - n;
    if (n < 1) n = 1;
    printf("%.*s%s\n", n, "                    ", op->help_text);
    }
-Line 1426 
 handle_option(int letter, int options)
+Line 1607 
 handle_option(int letter, int options)
  {
  switch(letter)
    {
    case N_FOFFSETS: file_offsets = TRUE; break;
    case N_HELP: help(); exit(0);
    case N_LOFFSETS: line_offsets = number = TRUE; break;
    case 'c': count_only = TRUE; break;
    case 'F': process_options |= PO_FIXED_STRINGS; break;
    case 'H': filenames = FN_FORCE; break;
-Line 1864 
 if (both_context > 0)
+Line 2045 
 if (both_context > 0)
    if (after_context == 0) after_context = both_context;
    if (before_context == 0) before_context = both_context;
    }
  /* Only one of --only-matching, --file-offsets, or --line-offsets is permitted.
  However, the latter two set the only_matching flag. */
  if ((only_matching && (file_offsets || line_offsets)) ||
      (file_offsets && line_offsets))
    {
    fprintf(stderr, "pcregrep: Cannot mix --only-matching, --file-offsets "
      "and/or --line-offsets\n");
    exit(usage(2));
    }
  if (file_offsets || line_offsets) only_matching = TRUE;
  /* If a locale has not been provided as an option, see if the LC_CTYPE or
  LC_ALL environment variable is set, and if so, use it. */
-Line 2115 
 if (include_pattern != NULL)
+Line 2296 
 if (include_pattern != NULL)
  if (i >= argc)
    {
-   rc = pcregrep(stdin, (filenames > FN_DEFAULT)? stdin_name : NULL);
+   rc = pcregrep(stdin, FR_PLAIN, (filenames > FN_DEFAULT)? stdin_name : NULL);
    goto EXIT;
    }

 Legend:



Removed from v.280
 


changed lines


 
Added in v.324
 Legend:



Removed from v.280
 


changed lines


 
Added in v.324
-Removed from v.280
+Added in v.324

webmaster@exim.org	ViewVC Help
Powered by ViewVC 1.1.12