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

Diff of /code/trunk/pcregrep.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 461 by ph10, Mon Oct 5 10:59:35 2009 UTC
# Line 6  Line 6 
6  its pattern matching. On a Unix or Win32 system it can recurse into  its pattern matching. On a Unix or Win32 system it can recurse into
7  directories.  directories.
8    
9             Copyright (c) 1997-2005 University of Cambridge             Copyright (c) 1997-2009 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
40    #ifdef HAVE_CONFIG_H
41    #include "config.h"
42    #endif
43    
44  #include <ctype.h>  #include <ctype.h>
45    #include <locale.h>
46  #include <stdio.h>  #include <stdio.h>
47  #include <string.h>  #include <string.h>
48  #include <stdlib.h>  #include <stdlib.h>
# Line 45  POSSIBILITY OF SUCH DAMAGE. Line 50  POSSIBILITY OF SUCH DAMAGE.
50    
51  #include <sys/types.h>  #include <sys/types.h>
52  #include <sys/stat.h>  #include <sys/stat.h>
53    
54    #ifdef HAVE_UNISTD_H
55  #include <unistd.h>  #include <unistd.h>
56    #endif
57    
58    #ifdef SUPPORT_LIBZ
59    #include <zlib.h>
60    #endif
61    
62    #ifdef SUPPORT_LIBBZ2
63    #include <bzlib.h>
64    #endif
65    
 #include "config.h"  
66  #include "pcre.h"  #include "pcre.h"
67    
68  #define FALSE 0  #define FALSE 0
# Line 55  POSSIBILITY OF SUCH DAMAGE. Line 70  POSSIBILITY OF SUCH DAMAGE.
70    
71  typedef int BOOL;  typedef int BOOL;
72    
 #define VERSION "4.0 07-Jun-2005"  
73  #define MAX_PATTERN_COUNT 100  #define MAX_PATTERN_COUNT 100
74    #define OFFSET_SIZE 99
75    
76  #if BUFSIZ > 8192  #if BUFSIZ > 8192
77  #define MBUFTHIRD BUFSIZ  #define MBUFTHIRD BUFSIZ
# Line 64  typedef int BOOL; Line 79  typedef int BOOL;
79  #define MBUFTHIRD 8192  #define MBUFTHIRD 8192
80  #endif  #endif
81    
82    /* Values for the "filenames" variable, which specifies options for file name
83    output. The order is important; it is assumed that a file name is wanted for
84    all values greater than FN_DEFAULT. */
85    
86    enum { FN_NONE, FN_DEFAULT, FN_MATCH_ONLY, FN_NOMATCH_ONLY, FN_FORCE };
87    
88    /* File reading styles */
89    
90    enum { FR_PLAIN, FR_LIBZ, FR_LIBBZ2 };
91    
92    /* Actions for the -d and -D options */
93    
94    enum { dee_READ, dee_SKIP, dee_RECURSE };
95    enum { DEE_READ, DEE_SKIP };
96    
97    /* Actions for special processing options (flag bits) */
98    
99    #define PO_WORD_MATCH     0x0001
100    #define PO_LINE_MATCH     0x0002
101    #define PO_FIXED_STRINGS  0x0004
102    
103    /* Line ending types */
104    
105    enum { EL_LF, EL_CR, EL_CRLF, EL_ANY, EL_ANYCRLF };
106    
107    
108    
109  /*************************************************  /*************************************************
110  *               Global variables                 *  *               Global variables                 *
111  *************************************************/  *************************************************/
112    
113    /* Jeffrey Friedl has some debugging requirements that are not part of the
114    regular code. */
115    
116    #ifdef JFRIEDL_DEBUG
117    static int S_arg = -1;
118    static unsigned int jfriedl_XR = 0; /* repeat regex attempt this many times */
119    static unsigned int jfriedl_XT = 0; /* replicate text this many times */
120    static const char *jfriedl_prefix = "";
121    static const char *jfriedl_postfix = "";
122    #endif
123    
124    static int  endlinetype;
125    
126    static char *colour_string = (char *)"1;31";
127    static char *colour_option = NULL;
128    static char *dee_option = NULL;
129    static char *DEE_option = NULL;
130    static char *newline = NULL;
131  static char *pattern_filename = NULL;  static char *pattern_filename = NULL;
132  static char *stdin_name = (char *)"(standard input)";  static char *stdin_name = (char *)"(standard input)";
133    static char *locale = NULL;
134    
135    static const unsigned char *pcretables = NULL;
136    
137  static int  pattern_count = 0;  static int  pattern_count = 0;
138  static pcre **pattern_list;  static pcre **pattern_list = NULL;
139  static pcre_extra **hints_list;  static pcre_extra **hints_list = NULL;
140    
141  static char *include_pattern = NULL;  static char *include_pattern = NULL;
142  static char *exclude_pattern = NULL;  static char *exclude_pattern = NULL;
143    static char *include_dir_pattern = NULL;
144    static char *exclude_dir_pattern = NULL;
145    
146  static pcre *include_compiled = NULL;  static pcre *include_compiled = NULL;
147  static pcre *exclude_compiled = NULL;  static pcre *exclude_compiled = NULL;
148    static pcre *include_dir_compiled = NULL;
149    static pcre *exclude_dir_compiled = NULL;
150    
151  static int after_context = 0;  static int after_context = 0;
152  static int before_context = 0;  static int before_context = 0;
153  static int both_context = 0;  static int both_context = 0;
154    static int dee_action = dee_READ;
155    static int DEE_action = DEE_READ;
156    static int error_count = 0;
157    static int filenames = FN_DEFAULT;
158    static int process_options = 0;
159    
160  static BOOL count_only = FALSE;  static BOOL count_only = FALSE;
161  static BOOL filenames = TRUE;  static BOOL do_colour = FALSE;
162  static BOOL filenames_only = FALSE;  static BOOL file_offsets = FALSE;
 static BOOL filenames_nomatch_only = FALSE;  
163  static BOOL hyphenpending = FALSE;  static BOOL hyphenpending = FALSE;
164  static BOOL invert = FALSE;  static BOOL invert = FALSE;
165    static BOOL line_offsets = FALSE;
166  static BOOL multiline = FALSE;  static BOOL multiline = FALSE;
167  static BOOL number = FALSE;  static BOOL number = FALSE;
168    static BOOL omit_zero_count = FALSE;
169    static BOOL only_matching = FALSE;
170  static BOOL quiet = FALSE;  static BOOL quiet = FALSE;
 static BOOL recurse = FALSE;  
171  static BOOL silent = FALSE;  static BOOL silent = FALSE;
172  static BOOL whole_lines = FALSE;  static BOOL utf8 = FALSE;
 static BOOL word_match = FALSE;  
173    
174  /* Structure for options and list of them */  /* Structure for options and list of them */
175    
176  enum { OP_NODATA, OP_STRING, OP_NUMBER };  enum { OP_NODATA, OP_STRING, OP_OP_STRING, OP_NUMBER, OP_OP_NUMBER,
177           OP_PATLIST };
178    
179  typedef struct option_item {  typedef struct option_item {
180    int type;    int type;
# Line 112  typedef struct option_item { Line 184  typedef struct option_item {
184    const char *help_text;    const char *help_text;
185  } option_item;  } option_item;
186    
187    /* Options without a single-letter equivalent get a negative value. This can be
188    used to identify them. */
189    
190    #define N_COLOUR       (-1)
191    #define N_EXCLUDE      (-2)
192    #define N_EXCLUDE_DIR  (-3)
193    #define N_HELP         (-4)
194    #define N_INCLUDE      (-5)
195    #define N_INCLUDE_DIR  (-6)
196    #define N_LABEL        (-7)
197    #define N_LOCALE       (-8)
198    #define N_NULL         (-9)
199    #define N_LOFFSETS     (-10)
200    #define N_FOFFSETS     (-11)
201    
202  static option_item optionlist[] = {  static option_item optionlist[] = {
203    { OP_NODATA, -1,  NULL,              "",              "  terminate options" },    { OP_NODATA,    N_NULL,   NULL,              "",              "  terminate options" },
204    { OP_NODATA, -1,  NULL,              "help",          "display this help and exit" },    { OP_NODATA,    N_HELP,   NULL,              "help",          "display this help and exit" },
205    { OP_NUMBER, 'A', &after_context,    "after-context=number", "set number of following context lines" },    { OP_NUMBER,    'A',      &after_context,    "after-context=number", "set number of following context lines" },
206    { OP_NUMBER, 'B', &before_context,   "before-context=number", "set number of prior context lines" },    { OP_NUMBER,    'B',      &before_context,   "before-context=number", "set number of prior context lines" },
207    { OP_NUMBER, 'C', &both_context,     "context=number", "set number of context lines, before & after" },    { OP_OP_STRING, N_COLOUR, &colour_option,    "color=option",  "matched text color option" },
208    { OP_NODATA, 'c', NULL,              "count",         "print only a count of matching lines per FILE" },    { OP_NUMBER,    'C',      &both_context,     "context=number", "set number of context lines, before & after" },
209    { OP_STRING, 'f', &pattern_filename, "file=path",     "read patterns from file" },    { OP_NODATA,    'c',      NULL,              "count",         "print only a count of matching lines per FILE" },
210    { OP_NODATA, 'h', NULL,              "no-filename",   "suppress the prefixing filename on output" },    { OP_OP_STRING, N_COLOUR, &colour_option,    "colour=option", "matched text colour option" },
211    { OP_NODATA, 'i', NULL,              "ignore-case",   "ignore case distinctions" },    { OP_STRING,    'D',      &DEE_option,       "devices=action","how to handle devices, FIFOs, and sockets" },
212    { OP_NODATA, 'l', NULL,              "files-with-matches", "print only FILE names containing matches" },    { OP_STRING,    'd',      &dee_option,       "directories=action", "how to handle directories" },
213    { OP_NODATA, 'L', NULL,              "files-without-match","print only FILE names not containing matches" },    { OP_PATLIST,   'e',      NULL,              "regex(p)=pattern", "specify pattern (may be used more than once)" },
214    { OP_STRING, -1,  &stdin_name,       "label=name",    "set name for standard input" },    { OP_NODATA,    'F',      NULL,              "fixed-strings", "patterns are sets of newline-separated strings" },
215    { OP_NODATA, 'M', NULL,              "multiline",     "run in multiline mode" },    { OP_STRING,    'f',      &pattern_filename, "file=path",     "read patterns from file" },
216    { OP_NODATA, 'n', NULL,              "line-number",   "print line number with output lines" },    { OP_NODATA,    N_FOFFSETS, NULL,            "file-offsets",  "output file offsets, not text" },
217    { OP_NODATA, 'q', NULL,              "quiet",         "suppress output, just set return code" },    { OP_NODATA,    'H',      NULL,              "with-filename", "force the prefixing filename on output" },
218    { OP_NODATA, 'r', NULL,              "recursive",     "recursively scan sub-directories" },    { OP_NODATA,    'h',      NULL,              "no-filename",   "suppress the prefixing filename on output" },
219    { OP_STRING, -1,  &exclude_pattern,  "exclude=pattern","exclude matching files when recursing" },    { OP_NODATA,    'i',      NULL,              "ignore-case",   "ignore case distinctions" },
220    { OP_STRING, -1,  &include_pattern,  "include=pattern","include matching files when recursing" },    { OP_NODATA,    'l',      NULL,              "files-with-matches", "print only FILE names containing matches" },
221    { OP_NODATA, 's', NULL,              "no-messages",   "suppress error messages" },    { OP_NODATA,    'L',      NULL,              "files-without-match","print only FILE names not containing matches" },
222    { OP_NODATA, 'u', NULL,              "utf-8",         "use UTF-8 mode" },    { OP_STRING,    N_LABEL,  &stdin_name,       "label=name",    "set name for standard input" },
223    { OP_NODATA, 'V', NULL,              "version",       "print version information and exit" },    { OP_NODATA,    N_LOFFSETS, NULL,            "line-offsets",  "output line numbers and offsets, not text" },
224    { OP_NODATA, 'v', NULL,              "invert-match",  "select non-matching lines" },    { OP_STRING,    N_LOCALE, &locale,           "locale=locale", "use the named locale" },
225    { OP_NODATA, 'w', NULL,              "word-regex(p)", "force PATTERN to match only as a word"  },    { OP_NODATA,    'M',      NULL,              "multiline",     "run in multiline mode" },
226    { OP_NODATA, 'x', NULL,              "line-regex(p)", "force PATTERN to match only whole lines" },    { OP_STRING,    'N',      &newline,          "newline=type",  "set newline type (CR, LF, CRLF, ANYCRLF or ANY)" },
227    { OP_NODATA, 0,   NULL,               NULL,            NULL }    { OP_NODATA,    'n',      NULL,              "line-number",   "print line number with output lines" },
228      { OP_NODATA,    'o',      NULL,              "only-matching", "show only the part of the line that matched" },
229      { OP_NODATA,    'q',      NULL,              "quiet",         "suppress output, just set return code" },
230      { OP_NODATA,    'r',      NULL,              "recursive",     "recursively scan sub-directories" },
231      { OP_STRING,    N_EXCLUDE,&exclude_pattern,  "exclude=pattern","exclude matching files when recursing" },
232      { OP_STRING,    N_INCLUDE,&include_pattern,  "include=pattern","include matching files when recursing" },
233      { OP_STRING,    N_EXCLUDE_DIR,&exclude_dir_pattern, "exclude_dir=pattern","exclude matching directories when recursing" },
234      { OP_STRING,    N_INCLUDE_DIR,&include_dir_pattern, "include_dir=pattern","include matching directories when recursing" },
235    #ifdef JFRIEDL_DEBUG
236      { OP_OP_NUMBER, 'S',      &S_arg,            "jeffS",         "replace matched (sub)string with X" },
237    #endif
238      { OP_NODATA,    's',      NULL,              "no-messages",   "suppress error messages" },
239      { OP_NODATA,    'u',      NULL,              "utf-8",         "use UTF-8 mode" },
240      { OP_NODATA,    'V',      NULL,              "version",       "print version information and exit" },
241      { OP_NODATA,    'v',      NULL,              "invert-match",  "select non-matching lines" },
242      { OP_NODATA,    'w',      NULL,              "word-regex(p)", "force patterns to match only as words"  },
243      { OP_NODATA,    'x',      NULL,              "line-regex(p)", "force patterns to match only whole lines" },
244      { OP_NODATA,    0,        NULL,               NULL,            NULL }
245  };  };
246    
247    /* Tables for prefixing and suffixing patterns, according to the -w, -x, and -F
248    options. These set the 1, 2, and 4 bits in process_options, respectively. Note
249    that the combination of -w and -x has the same effect as -x on its own, so we
250    can treat them as the same. */
251    
252    static const char *prefix[] = {
253      "", "\\b", "^(?:", "^(?:", "\\Q", "\\b\\Q", "^(?:\\Q", "^(?:\\Q" };
254    
255    static const char *suffix[] = {
256      "", "\\b", ")$",   ")$",   "\\E", "\\E\\b", "\\E)$",   "\\E)$" };
257    
258    /* UTF-8 tables - used only when the newline setting is "any". */
259    
260    const int utf8_table3[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};
261    
262    const char utf8_table4[] = {
263      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
264      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
265      2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
266      3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 };
267    
268    
269    
270  /*************************************************  /*************************************************
271  *       Functions for directory scanning         *  *            OS-specific functions               *
272  *************************************************/  *************************************************/
273    
274  /* These functions are defined so that they can be made system specific,  /* These functions are defined so that they can be made system specific,
275  although at present the only ones are for Unix, Win32, and for "no directory  although at present the only ones are for Unix, Win32, and for "no support". */
 recursion support". */  
276    
277    
278  /************* Directory scanning in Unix ***********/  /************* Directory scanning in Unix ***********/
279    
280  #if IS_UNIX  #if defined HAVE_SYS_STAT_H && defined HAVE_DIRENT_H && defined HAVE_SYS_TYPES_H
281  #include <sys/types.h>  #include <sys/types.h>
282  #include <sys/stat.h>  #include <sys/stat.h>
283  #include <dirent.h>  #include <dirent.h>
# Line 184  for (;;) Line 309  for (;;)
309    if (strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0)    if (strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0)
310      return dent->d_name;      return dent->d_name;
311    }    }
312  return NULL;   /* Keep compiler happy; never executed */  /* Control never reaches here */
313  }  }
314    
315  static void  static void
# Line 194  closedir(dir); Line 319  closedir(dir);
319  }  }
320    
321    
322    /************* Test for regular file in Unix **********/
323    
324    static int
325    isregfile(char *filename)
326    {
327    struct stat statbuf;
328    if (stat(filename, &statbuf) < 0)
329      return 1;        /* In the expectation that opening as a file will fail */
330    return (statbuf.st_mode & S_IFMT) == S_IFREG;
331    }
332    
333    
334    /************* Test stdout for being a terminal in Unix **********/
335    
336    static BOOL
337    is_stdout_tty(void)
338    {
339    return isatty(fileno(stdout));
340    }
341    
342    
343  /************* Directory scanning in Win32 ***********/  /************* Directory scanning in Win32 ***********/
344    
345  /* I (Philip Hazel) have no means of testing this code. It was contributed by  /* I (Philip Hazel) have no means of testing this code. It was contributed by
346  Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES  Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES
347  when it did not exist. */  when it did not exist. David Byron added a patch that moved the #include of
348    <windows.h> to before the INVALID_FILE_ATTRIBUTES definition rather than after.
349    */
350    
351  #elif HAVE_WIN32API  #elif HAVE_WINDOWS_H
352    
353  #ifndef STRICT  #ifndef STRICT
354  # define STRICT  # define STRICT
# Line 209  when it did not exist. */ Line 356  when it did not exist. */
356  #ifndef WIN32_LEAN_AND_MEAN  #ifndef WIN32_LEAN_AND_MEAN
357  # define WIN32_LEAN_AND_MEAN  # define WIN32_LEAN_AND_MEAN
358  #endif  #endif
359    
360    #include <windows.h>
361    
362  #ifndef INVALID_FILE_ATTRIBUTES  #ifndef INVALID_FILE_ATTRIBUTES
363  #define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF  #define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF
364  #endif  #endif
365    
 #include <windows.h>  
   
366  typedef struct directory_type  typedef struct directory_type
367  {  {
368  HANDLE handle;  HANDLE handle;
# Line 292  free(dir); Line 440  free(dir);
440  }  }
441    
442    
443    /************* Test for regular file in Win32 **********/
444    
445    /* I don't know how to do this, or if it can be done; assume all paths are
446    regular if they are not directories. */
447    
448    int isregfile(char *filename)
449    {
450    return !isdirectory(filename);
451    }
452    
453    
454    /************* Test stdout for being a terminal in Win32 **********/
455    
456    /* I don't know how to do this; assume never */
457    
458    static BOOL
459    is_stdout_tty(void)
460    {
461    return FALSE;
462    }
463    
464    
465  /************* Directory scanning when we can't do it ***********/  /************* Directory scanning when we can't do it ***********/
466    
467  /* The type is void, and apart from isdirectory(), the functions do nothing. */  /* The type is void, and apart from isdirectory(), the functions do nothing. */
# Line 300  free(dir); Line 470  free(dir);
470    
471  typedef void directory_type;  typedef void directory_type;
472    
473  int isdirectory(char *filename) { return FALSE; }  int isdirectory(char *filename) { return 0; }
474  directory_type * opendirectory(char *filename) {}  directory_type * opendirectory(char *filename) { return (directory_type*)0;}
475  char *readdirectory(directory_type *dir) {}  char *readdirectory(directory_type *dir) { return (char*)0;}
476  void closedirectory(directory_type *dir) {}  void closedirectory(directory_type *dir) {}
477    
478    
479    /************* Test for regular when we can't do it **********/
480    
481    /* Assume all files are regular. */
482    
483    int isregfile(char *filename) { return 1; }
484    
485    
486    /************* Test stdout for being a terminal when we can't do it **********/
487    
488    static BOOL
489    is_stdout_tty(void)
490    {
491    return FALSE;
492    }
493    
494    
495  #endif  #endif
496    
497    
498    
499  #if ! HAVE_STRERROR  #ifndef HAVE_STRERROR
500  /*************************************************  /*************************************************
501  *     Provide strerror() for non-ANSI libraries  *  *     Provide strerror() for non-ANSI libraries  *
502  *************************************************/  *************************************************/
# Line 332  return sys_errlist[n]; Line 519  return sys_errlist[n];
519    
520    
521  /*************************************************  /*************************************************
522    *             Find end of line                   *
523    *************************************************/
524    
525    /* The length of the endline sequence that is found is set via lenptr. This may
526    be zero at the very end of the file if there is no line-ending sequence there.
527    
528    Arguments:
529      p         current position in line
530      endptr    end of available data
531      lenptr    where to put the length of the eol sequence
532    
533    Returns:    pointer to the last byte of the line
534    */
535    
536    static char *
537    end_of_line(char *p, char *endptr, int *lenptr)
538    {
539    switch(endlinetype)
540      {
541      default:      /* Just in case */
542      case EL_LF:
543      while (p < endptr && *p != '\n') p++;
544      if (p < endptr)
545        {
546        *lenptr = 1;
547        return p + 1;
548        }
549      *lenptr = 0;
550      return endptr;
551    
552      case EL_CR:
553      while (p < endptr && *p != '\r') p++;
554      if (p < endptr)
555        {
556        *lenptr = 1;
557        return p + 1;
558        }
559      *lenptr = 0;
560      return endptr;
561    
562      case EL_CRLF:
563      for (;;)
564        {
565        while (p < endptr && *p != '\r') p++;
566        if (++p >= endptr)
567          {
568          *lenptr = 0;
569          return endptr;
570          }
571        if (*p == '\n')
572          {
573          *lenptr = 2;
574          return p + 1;
575          }
576        }
577      break;
578    
579      case EL_ANYCRLF:
580      while (p < endptr)
581        {
582        int extra = 0;
583        register int c = *((unsigned char *)p);
584    
585        if (utf8 && c >= 0xc0)
586          {
587          int gcii, gcss;
588          extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
589          gcss = 6*extra;
590          c = (c & utf8_table3[extra]) << gcss;
591          for (gcii = 1; gcii <= extra; gcii++)
592            {
593            gcss -= 6;
594            c |= (p[gcii] & 0x3f) << gcss;
595            }
596          }
597    
598        p += 1 + extra;
599    
600        switch (c)
601          {
602          case 0x0a:    /* LF */
603          *lenptr = 1;
604          return p;
605    
606          case 0x0d:    /* CR */
607          if (p < endptr && *p == 0x0a)
608            {
609            *lenptr = 2;
610            p++;
611            }
612          else *lenptr = 1;
613          return p;
614    
615          default:
616          break;
617          }
618        }   /* End of loop for ANYCRLF case */
619    
620      *lenptr = 0;  /* Must have hit the end */
621      return endptr;
622    
623      case EL_ANY:
624      while (p < endptr)
625        {
626        int extra = 0;
627        register int c = *((unsigned char *)p);
628    
629        if (utf8 && c >= 0xc0)
630          {
631          int gcii, gcss;
632          extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
633          gcss = 6*extra;
634          c = (c & utf8_table3[extra]) << gcss;
635          for (gcii = 1; gcii <= extra; gcii++)
636            {
637            gcss -= 6;
638            c |= (p[gcii] & 0x3f) << gcss;
639            }
640          }
641    
642        p += 1 + extra;
643    
644        switch (c)
645          {
646          case 0x0a:    /* LF */
647          case 0x0b:    /* VT */
648          case 0x0c:    /* FF */
649          *lenptr = 1;
650          return p;
651    
652          case 0x0d:    /* CR */
653          if (p < endptr && *p == 0x0a)
654            {
655            *lenptr = 2;
656            p++;
657            }
658          else *lenptr = 1;
659          return p;
660    
661          case 0x85:    /* NEL */
662          *lenptr = utf8? 2 : 1;
663          return p;
664    
665          case 0x2028:  /* LS */
666          case 0x2029:  /* PS */
667          *lenptr = 3;
668          return p;
669    
670          default:
671          break;
672          }
673        }   /* End of loop for ANY case */
674    
675      *lenptr = 0;  /* Must have hit the end */
676      return endptr;
677      }     /* End of overall switch */
678    }
679    
680    
681    
682    /*************************************************
683    *         Find start of previous line            *
684    *************************************************/
685    
686    /* This is called when looking back for before lines to print.
687    
688    Arguments:
689      p         start of the subsequent line
690      startptr  start of available data
691    
692    Returns:    pointer to the start of the previous line
693    */
694    
695    static char *
696    previous_line(char *p, char *startptr)
697    {
698    switch(endlinetype)
699      {
700      default:      /* Just in case */
701      case EL_LF:
702      p--;
703      while (p > startptr && p[-1] != '\n') p--;
704      return p;
705    
706      case EL_CR:
707      p--;
708      while (p > startptr && p[-1] != '\n') p--;
709      return p;
710    
711      case EL_CRLF:
712      for (;;)
713        {
714        p -= 2;
715        while (p > startptr && p[-1] != '\n') p--;
716        if (p <= startptr + 1 || p[-2] == '\r') return p;
717        }
718      return p;   /* But control should never get here */
719    
720      case EL_ANY:
721      case EL_ANYCRLF:
722      if (*(--p) == '\n' && p > startptr && p[-1] == '\r') p--;
723      if (utf8) while ((*p & 0xc0) == 0x80) p--;
724    
725      while (p > startptr)
726        {
727        register int c;
728        char *pp = p - 1;
729    
730        if (utf8)
731          {
732          int extra = 0;
733          while ((*pp & 0xc0) == 0x80) pp--;
734          c = *((unsigned char *)pp);
735          if (c >= 0xc0)
736            {
737            int gcii, gcss;
738            extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
739            gcss = 6*extra;
740            c = (c & utf8_table3[extra]) << gcss;
741            for (gcii = 1; gcii <= extra; gcii++)
742              {
743              gcss -= 6;
744              c |= (pp[gcii] & 0x3f) << gcss;
745              }
746            }
747          }
748        else c = *((unsigned char *)pp);
749    
750        if (endlinetype == EL_ANYCRLF) switch (c)
751          {
752          case 0x0a:    /* LF */
753          case 0x0d:    /* CR */
754          return p;
755    
756          default:
757          break;
758          }
759    
760        else switch (c)
761          {
762          case 0x0a:    /* LF */
763          case 0x0b:    /* VT */
764          case 0x0c:    /* FF */
765          case 0x0d:    /* CR */
766          case 0x85:    /* NEL */
767          case 0x2028:  /* LS */
768          case 0x2029:  /* PS */
769          return p;
770    
771          default:
772          break;
773          }
774    
775        p = pp;  /* Back one character */
776        }        /* End of loop for ANY case */
777    
778      return startptr;  /* Hit start of data */
779      }     /* End of overall switch */
780    }
781    
782    
783    
784    
785    
786    /*************************************************
787  *       Print the previous "after" lines         *  *       Print the previous "after" lines         *
788  *************************************************/  *************************************************/
789    
790  /* This is called if we are about to lose said lines because of buffer filling,  /* This is called if we are about to lose said lines because of buffer filling,
791  and at the end of the file.  and at the end of the file. The data in the line is written using fwrite() so
792    that a binary zero does not terminate it.
793    
794  Arguments:  Arguments:
795    lastmatchnumber   the number of the last matching line, plus one    lastmatchnumber   the number of the last matching line, plus one
# Line 355  if (after_context > 0 && lastmatchnumber Line 808  if (after_context > 0 && lastmatchnumber
808    int count = 0;    int count = 0;
809    while (lastmatchrestart < endptr && count++ < after_context)    while (lastmatchrestart < endptr && count++ < after_context)
810      {      {
811        int ellength;
812      char *pp = lastmatchrestart;      char *pp = lastmatchrestart;
813      if (printname != NULL) fprintf(stdout, "%s-", printname);      if (printname != NULL) fprintf(stdout, "%s-", printname);
814      if (number) fprintf(stdout, "%d-", lastmatchnumber++);      if (number) fprintf(stdout, "%d-", lastmatchnumber++);
815      while (*pp != '\n') pp++;      pp = end_of_line(pp, endptr, &ellength);
816      fprintf(stdout, "%.*s", pp - lastmatchrestart + 1, lastmatchrestart);      fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout);
817      lastmatchrestart = pp + 1;      lastmatchrestart = pp;
818      }      }
819    hyphenpending = TRUE;    hyphenpending = TRUE;
820    }    }
# Line 369  if (after_context > 0 && lastmatchnumber Line 823  if (after_context > 0 && lastmatchnumber
823    
824    
825  /*************************************************  /*************************************************
826    *   Apply patterns to subject till one matches   *
827    *************************************************/
828    
829    /* This function is called to run through all patterns, looking for a match. It
830    is used multiple times for the same subject when colouring is enabled, in order
831    to find all possible matches.
832    
833    Arguments:
834      matchptr    the start of the subject
835      length      the length of the subject to match
836      offsets     the offets vector to fill in
837      mrc         address of where to put the result of pcre_exec()
838    
839    Returns:      TRUE if there was a match
840                  FALSE if there was no match
841                  invert if there was a non-fatal error
842    */
843    
844    static BOOL
845    match_patterns(char *matchptr, size_t length, int *offsets, int *mrc)
846    {
847    int i;
848    for (i = 0; i < pattern_count; i++)
849      {
850      *mrc = pcre_exec(pattern_list[i], hints_list[i], matchptr, length, 0,
851        PCRE_NOTEMPTY, offsets, OFFSET_SIZE);
852      if (*mrc >= 0) return TRUE;
853      if (*mrc == PCRE_ERROR_NOMATCH) continue;
854      fprintf(stderr, "pcregrep: pcre_exec() error %d while matching ", *mrc);
855      if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1);
856      fprintf(stderr, "this text:\n");
857      fwrite(matchptr, 1, length, stderr);  /* In case binary zero included */
858      fprintf(stderr, "\n");
859      if (error_count == 0 &&
860          (*mrc == PCRE_ERROR_MATCHLIMIT || *mrc == PCRE_ERROR_RECURSIONLIMIT))
861        {
862        fprintf(stderr, "pcregrep: error %d means that a resource limit "
863          "was exceeded\n", *mrc);
864        fprintf(stderr, "pcregrep: check your regex for nested unlimited loops\n");
865        }
866      if (error_count++ > 20)
867        {
868        fprintf(stderr, "pcregrep: too many errors - abandoned\n");
869        exit(2);
870        }
871      return invert;    /* No more matching; don't show the line again */
872      }
873    
874    return FALSE;  /* No match, no errors */
875    }
876    
877    
878    
879    /*************************************************
880  *            Grep an individual file             *  *            Grep an individual file             *
881  *************************************************/  *************************************************/
882    
# Line 380  be in the middle third most of the time, Line 888  be in the middle third most of the time,
888  "before" context printing.  "before" context printing.
889    
890  Arguments:  Arguments:
891    in           the fopened FILE stream    handle       the fopened FILE stream for a normal file
892                   the gzFile pointer when reading is via libz
893                   the BZFILE pointer when reading is via libbz2
894      frtype       FR_PLAIN, FR_LIBZ, or FR_LIBBZ2
895    printname    the file name if it is to be printed for each match    printname    the file name if it is to be printed for each match
896                 or NULL if the file name is not to be printed                 or NULL if the file name is not to be printed
897                 it cannot be NULL if filenames[_nomatch]_only is set                 it cannot be NULL if filenames[_nomatch]_only is set
898    
899  Returns:       0 if there was at least one match  Returns:       0 if there was at least one match
900                 1 otherwise (no matches)                 1 otherwise (no matches)
901                   2 if there is a read error on a .bz2 file
902  */  */
903    
904  static int  static int
905  pcregrep(FILE *in, char *printname)  pcregrep(void *handle, int frtype, char *printname)
906  {  {
907  int rc = 1;  int rc = 1;
908  int linenumber = 1;  int linenumber = 1;
909  int lastmatchnumber = 0;  int lastmatchnumber = 0;
910  int count = 0;  int count = 0;
911  int offsets[99];  int filepos = 0;
912    int offsets[OFFSET_SIZE];
913  char *lastmatchrestart = NULL;  char *lastmatchrestart = NULL;
914  char buffer[3*MBUFTHIRD];  char buffer[3*MBUFTHIRD];
915  char *ptr = buffer;  char *ptr = buffer;
916  char *endptr;  char *endptr;
917  size_t bufflength;  size_t bufflength;
918  BOOL endhyphenpending = FALSE;  BOOL endhyphenpending = FALSE;
919    FILE *in = NULL;                    /* Ensure initialized */
920    
921    #ifdef SUPPORT_LIBZ
922    gzFile ingz = NULL;
923    #endif
924    
925    #ifdef SUPPORT_LIBBZ2
926    BZFILE *inbz2 = NULL;
927    #endif
928    
929    
930    /* Do the first read into the start of the buffer and set up the pointer to end
931    of what we have. In the case of libz, a non-zipped .gz file will be read as a
932    plain file. However, if a .bz2 file isn't actually bzipped, the first read will
933    fail. */
934    
935    #ifdef SUPPORT_LIBZ
936    if (frtype == FR_LIBZ)
937      {
938      ingz = (gzFile)handle;
939      bufflength = gzread (ingz, buffer, 3*MBUFTHIRD);
940      }
941    else
942    #endif
943    
944    #ifdef SUPPORT_LIBBZ2
945    if (frtype == FR_LIBBZ2)
946      {
947      inbz2 = (BZFILE *)handle;
948      bufflength = BZ2_bzread(inbz2, buffer, 3*MBUFTHIRD);
949      if ((int)bufflength < 0) return 2;   /* Gotcha: bufflength is size_t; */
950      }                                    /* without the cast it is unsigned. */
951    else
952    #endif
953    
954  /* Do the first read into the start of the buffer and set up the pointer to    {
955  end of what we have. */    in = (FILE *)handle;
956      bufflength = fread(buffer, 1, 3*MBUFTHIRD, in);
957      }
958    
 bufflength = fread(buffer, 1, 3*MBUFTHIRD, in);  
959  endptr = buffer + bufflength;  endptr = buffer + bufflength;
960    
961  /* Loop while the current pointer is not at the end of the file. For large  /* Loop while the current pointer is not at the end of the file. For large
# Line 417  way, the buffer is shifted left and re-f Line 965  way, the buffer is shifted left and re-f
965    
966  while (ptr < endptr)  while (ptr < endptr)
967    {    {
968    int i;    int endlinelength;
969    BOOL match = FALSE;    int mrc = 0;
970      BOOL match;
971      char *matchptr = ptr;
972    char *t = ptr;    char *t = ptr;
973    size_t length, linelength;    size_t length, linelength;
974    
975    /* At this point, ptr is at the start of a line. We need to find the length    /* At this point, ptr is at the start of a line. We need to find the length
976    of the subject string to pass to pcre_exec(). In multiline mode, it is the    of the subject string to pass to pcre_exec(). In multiline mode, it is the
977    length remainder of the data in the buffer. Otherwise, it is the length of    length remainder of the data in the buffer. Otherwise, it is the length of
978    the next line. After matching, we always advance by the length of the next    the next line, excluding the terminating newline. After matching, we always
979    line. In multiline mode the PCRE_FIRSTLINE option is used for compiling, so    advance by the length of the next line. In multiline mode the PCRE_FIRSTLINE
980    that any match is constrained to be in the first line. */    option is used for compiling, so that any match is constrained to be in the
981      first line. */
982    
983      t = end_of_line(t, endptr, &endlinelength);
984      linelength = t - ptr - endlinelength;
985      length = multiline? (size_t)(endptr - ptr) : linelength;
986    
987      /* Extra processing for Jeffrey Friedl's debugging. */
988    
989    #ifdef JFRIEDL_DEBUG
990      if (jfriedl_XT || jfriedl_XR)
991      {
992          #include <sys/time.h>
993          #include <time.h>
994          struct timeval start_time, end_time;
995          struct timezone dummy;
996          int i;
997    
998    linelength = 0;        if (jfriedl_XT)
999    while (t < endptr && *t++ != '\n') linelength++;        {
1000    length = multiline? endptr - ptr : linelength;            unsigned long newlen = length * jfriedl_XT + strlen(jfriedl_prefix) + strlen(jfriedl_postfix);
1001              const char *orig = ptr;
1002              ptr = malloc(newlen + 1);
1003              if (!ptr) {
1004                      printf("out of memory");
1005                      exit(2);
1006              }
1007              endptr = ptr;
1008              strcpy(endptr, jfriedl_prefix); endptr += strlen(jfriedl_prefix);
1009              for (i = 0; i < jfriedl_XT; i++) {
1010                      strncpy(endptr, orig,  length);
1011                      endptr += length;
1012              }
1013              strcpy(endptr, jfriedl_postfix); endptr += strlen(jfriedl_postfix);
1014              length = newlen;
1015          }
1016    
1017    /* Run through all the patterns until one matches. Note that we don't include        if (gettimeofday(&start_time, &dummy) != 0)
1018    the final newline in the subject string. */                perror("bad gettimeofday");
1019    
   for (i = 0; !match && i < pattern_count; i++)  
     {  
     match = pcre_exec(pattern_list[i], hints_list[i], ptr, length, 0, 0,  
       offsets, 99) >= 0;  
     }  
1020    
1021    /* If it's a match or a not-match (as required), print what's wanted. */        for (i = 0; i < jfriedl_XR; i++)
1022              match = (pcre_exec(pattern_list[0], hints_list[0], ptr, length, 0,
1023                  PCRE_NOTEMPTY, offsets, OFFSET_SIZE) >= 0);
1024    
1025          if (gettimeofday(&end_time, &dummy) != 0)
1026                  perror("bad gettimeofday");
1027    
1028          double delta = ((end_time.tv_sec + (end_time.tv_usec / 1000000.0))
1029                          -
1030                          (start_time.tv_sec + (start_time.tv_usec / 1000000.0)));
1031    
1032          printf("%s TIMER[%.4f]\n", match ? "MATCH" : "FAIL", delta);
1033          return 0;
1034      }
1035    #endif
1036    
1037      /* We come back here after a match when the -o option (only_matching) is set,
1038      in order to find any further matches in the same line. */
1039    
1040      ONLY_MATCHING_RESTART:
1041    
1042      /* Run through all the patterns until one matches or there is an error other
1043      than NOMATCH. This code is in a subroutine so that it can be re-used for
1044      finding subsequent matches when colouring matched lines. */
1045    
1046      match = match_patterns(matchptr, length, offsets, &mrc);
1047    
1048      /* If it's a match or a not-match (as required), do what's wanted. */
1049    
1050    if (match != invert)    if (match != invert)
1051      {      {
1052      BOOL hyphenprinted = FALSE;      BOOL hyphenprinted = FALSE;
1053    
1054      if (filenames_nomatch_only) return 1;      /* We've failed if we want a file that doesn't have any matches. */
1055    
1056        if (filenames == FN_NOMATCH_ONLY) return 1;
1057    
1058        /* Just count if just counting is wanted. */
1059    
1060      if (count_only) count++;      if (count_only) count++;
1061    
1062      else if (filenames_only)      /* If all we want is a file name, there is no need to scan any more lines
1063        in the file. */
1064    
1065        else if (filenames == FN_MATCH_ONLY)
1066        {        {
1067        fprintf(stdout, "%s\n", printname);        fprintf(stdout, "%s\n", printname);
1068        return 0;        return 0;
1069        }        }
1070    
1071        /* Likewise, if all we want is a yes/no answer. */
1072    
1073      else if (quiet) return 0;      else if (quiet) return 0;
1074    
1075        /* The --only-matching option prints just the substring that matched, and
1076        the --file-offsets and --line-offsets options output offsets for the
1077        matching substring (they both force --only-matching). None of these options
1078        prints any context. Afterwards, adjust the start and length, and then jump
1079        back to look for further matches in the same line. If we are in invert
1080        mode, however, nothing is printed - this could be still useful because the
1081        return code is set. */
1082    
1083        else if (only_matching)
1084          {
1085          if (!invert)
1086            {
1087            if (printname != NULL) fprintf(stdout, "%s:", printname);
1088            if (number) fprintf(stdout, "%d:", linenumber);
1089            if (line_offsets)
1090              fprintf(stdout, "%d,%d", (int)(matchptr + offsets[0] - ptr),
1091                offsets[1] - offsets[0]);
1092            else if (file_offsets)
1093              fprintf(stdout, "%d,%d", (int)(filepos + matchptr + offsets[0] - ptr),
1094                offsets[1] - offsets[0]);
1095            else
1096              {
1097              if (do_colour) fprintf(stdout, "%c[%sm", 0x1b, colour_string);
1098              fwrite(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
1099              if (do_colour) fprintf(stdout, "%c[00m", 0x1b);
1100              }
1101            fprintf(stdout, "\n");
1102            matchptr += offsets[1];
1103            length -= offsets[1];
1104            match = FALSE;
1105            goto ONLY_MATCHING_RESTART;
1106            }
1107          }
1108    
1109        /* This is the default case when none of the above options is set. We print
1110        the matching lines(s), possibly preceded and/or followed by other lines of
1111        context. */
1112    
1113      else      else
1114        {        {
1115        /* See if there is a requirement to print some "after" lines from a        /* See if there is a requirement to print some "after" lines from a
# Line 467  while (ptr < endptr) Line 1117  while (ptr < endptr)
1117    
1118        if (after_context > 0 && lastmatchnumber > 0)        if (after_context > 0 && lastmatchnumber > 0)
1119          {          {
1120            int ellength;
1121          int linecount = 0;          int linecount = 0;
1122          char *p = lastmatchrestart;          char *p = lastmatchrestart;
1123    
1124          while (p < ptr && linecount < after_context)          while (p < ptr && linecount < after_context)
1125            {            {
1126            while (*p != '\n') p++;            p = end_of_line(p, ptr, &ellength);
           p++;  
1127            linecount++;            linecount++;
1128            }            }
1129    
1130          /* It is important to advance lastmatchrestart during this printing so          /* It is important to advance lastmatchrestart during this printing so
1131          that it interacts correctly with any "before" printing below. */          that it interacts correctly with any "before" printing below. Print
1132            each line's data using fwrite() in case there are binary zeroes. */
1133    
1134          while (lastmatchrestart < p)          while (lastmatchrestart < p)
1135            {            {
1136            char *pp = lastmatchrestart;            char *pp = lastmatchrestart;
1137            if (printname != NULL) fprintf(stdout, "%s-", printname);            if (printname != NULL) fprintf(stdout, "%s-", printname);
1138            if (number) fprintf(stdout, "%d-", lastmatchnumber++);            if (number) fprintf(stdout, "%d-", lastmatchnumber++);
1139            while (*pp != '\n') pp++;            pp = end_of_line(pp, endptr, &ellength);
1140            fprintf(stdout, "%.*s", pp - lastmatchrestart + 1, lastmatchrestart);            fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout);
1141            lastmatchrestart = pp + 1;            lastmatchrestart = pp;
1142            }            }
1143          if (lastmatchrestart != ptr) hyphenpending = TRUE;          if (lastmatchrestart != ptr) hyphenpending = TRUE;
1144          }          }
# Line 510  while (ptr < endptr) Line 1161  while (ptr < endptr)
1161          char *p = ptr;          char *p = ptr;
1162    
1163          while (p > buffer && (lastmatchnumber == 0 || p > lastmatchrestart) &&          while (p > buffer && (lastmatchnumber == 0 || p > lastmatchrestart) &&
1164                 linecount++ < before_context)                 linecount < before_context)
1165            {            {
1166            p--;            linecount++;
1167            while (p > buffer && p[-1] != '\n') p--;            p = previous_line(p, buffer);
1168            }            }
1169    
1170          if (lastmatchnumber > 0 && p > lastmatchrestart && !hyphenprinted)          if (lastmatchnumber > 0 && p > lastmatchrestart && !hyphenprinted)
# Line 521  while (ptr < endptr) Line 1172  while (ptr < endptr)
1172    
1173          while (p < ptr)          while (p < ptr)
1174            {            {
1175              int ellength;
1176            char *pp = p;            char *pp = p;
1177            if (printname != NULL) fprintf(stdout, "%s-", printname);            if (printname != NULL) fprintf(stdout, "%s-", printname);
1178            if (number) fprintf(stdout, "%d-", linenumber - linecount--);            if (number) fprintf(stdout, "%d-", linenumber - linecount--);
1179            while (*pp != '\n') pp++;            pp = end_of_line(pp, endptr, &ellength);
1180            fprintf(stdout, "%.*s", pp - p + 1, p);            fwrite(p, 1, pp - p, stdout);
1181            p = pp + 1;            p = pp;
1182            }            }
1183          }          }
1184    
1185        /* Now print the matching line(s); ensure we set hyphenpending at the end        /* Now print the matching line(s); ensure we set hyphenpending at the end
1186        of the file. */        of the file if any context lines are being output. */
1187    
1188          if (after_context > 0 || before_context > 0)
1189            endhyphenpending = TRUE;
1190    
       endhyphenpending = TRUE;  
1191        if (printname != NULL) fprintf(stdout, "%s:", printname);        if (printname != NULL) fprintf(stdout, "%s:", printname);
1192        if (number) fprintf(stdout, "%d:", linenumber);        if (number) fprintf(stdout, "%d:", linenumber);
1193    
1194        /* In multiline mode, we want to print to the end of the line in which        /* In multiline mode, we want to print to the end of the line in which
1195        the end of the matched string is found, so we adjust linelength and the        the end of the matched string is found, so we adjust linelength and the
1196        line number appropriately. Because the PCRE_FIRSTLINE option is set, the        line number appropriately, but only when there actually was a match
1197        start of the match will always be before the first \n character. */        (invert not set). Because the PCRE_FIRSTLINE option is set, the start of
1198          the match will always be before the first newline sequence. */
1199    
1200        if (multiline)        if (multiline)
1201          {          {
1202          char *endmatch = ptr + offsets[1];          int ellength;
1203          t = ptr;          char *endmatch = ptr;
1204          while (t < endmatch) { if (*t++ == '\n') linenumber++; }          if (!invert)
1205          while (endmatch < endptr && *endmatch != '\n') endmatch++;            {
1206          linelength = endmatch - ptr;            endmatch += offsets[1];
1207              t = ptr;
1208              while (t < endmatch)
1209                {
1210                t = end_of_line(t, endptr, &ellength);
1211                if (t <= endmatch) linenumber++; else break;
1212                }
1213              }
1214            endmatch = end_of_line(endmatch, endptr, &ellength);
1215            linelength = endmatch - ptr - ellength;
1216          }          }
1217    
1218        fprintf(stdout, "%.*s\n", linelength, ptr);        /*** NOTE: Use only fwrite() to output the data line, so that binary
1219        }        zeroes are treated as just another data character. */
1220    
1221      rc = 0;    /* Had some success */        /* This extra option, for Jeffrey Friedl's debugging requirements,
1222          replaces the matched string, or a specific captured string if it exists,
1223          with X. When this happens, colouring is ignored. */
1224    
1225      /* Remember where the last match happened for after_context. We remember  #ifdef JFRIEDL_DEBUG
1226      where we are about to restart, and that line's number. */        if (S_arg >= 0 && S_arg < mrc)
1227            {
1228            int first = S_arg * 2;
1229            int last  = first + 1;
1230            fwrite(ptr, 1, offsets[first], stdout);
1231            fprintf(stdout, "X");
1232            fwrite(ptr + offsets[last], 1, linelength - offsets[last], stdout);
1233            }
1234          else
1235    #endif
1236    
1237      lastmatchrestart = ptr + linelength + 1;        /* We have to split the line(s) up if colouring, and search for further
1238      lastmatchnumber = linenumber + 1;        matches. */
     }  
1239    
1240    /* Advance to after the newline and increment the line number. */        if (do_colour)
1241            {
1242            int last_offset = 0;
1243            fwrite(ptr, 1, offsets[0], stdout);
1244            fprintf(stdout, "%c[%sm", 0x1b, colour_string);
1245            fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
1246            fprintf(stdout, "%c[00m", 0x1b);
1247            for (;;)
1248              {
1249              last_offset += offsets[1];
1250              matchptr += offsets[1];
1251              length -= offsets[1];
1252              if (!match_patterns(matchptr, length, offsets, &mrc)) break;
1253              fwrite(matchptr, 1, offsets[0], stdout);
1254              fprintf(stdout, "%c[%sm", 0x1b, colour_string);
1255              fwrite(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
1256              fprintf(stdout, "%c[00m", 0x1b);
1257              }
1258            fwrite(ptr + last_offset, 1, (linelength + endlinelength) - last_offset,
1259              stdout);
1260            }
1261    
1262          /* Not colouring; no need to search for further matches */
1263    
1264          else fwrite(ptr, 1, linelength + endlinelength, stdout);
1265          }
1266    
1267        /* End of doing what has to be done for a match */
1268    
1269        rc = 0;    /* Had some success */
1270    
1271        /* Remember where the last match happened for after_context. We remember
1272        where we are about to restart, and that line's number. */
1273    
1274        lastmatchrestart = ptr + linelength + endlinelength;
1275        lastmatchnumber = linenumber + 1;
1276        }
1277    
1278      /* For a match in multiline inverted mode (which of course did not cause
1279      anything to be printed), we have to move on to the end of the match before
1280      proceeding. */
1281    
1282      if (multiline && invert && match)
1283        {
1284        int ellength;
1285        char *endmatch = ptr + offsets[1];
1286        t = ptr;
1287        while (t < endmatch)
1288          {
1289          t = end_of_line(t, endptr, &ellength);
1290          if (t <= endmatch) linenumber++; else break;
1291          }
1292        endmatch = end_of_line(endmatch, endptr, &ellength);
1293        linelength = endmatch - ptr - ellength;
1294        }
1295    
1296      /* Advance to after the newline and increment the line number. The file
1297      offset to the current line is maintained in filepos. */
1298    
1299    ptr += linelength + 1;    ptr += linelength + endlinelength;
1300      filepos += linelength + endlinelength;
1301    linenumber++;    linenumber++;
1302    
1303    /* If we haven't yet reached the end of the file (the buffer is full), and    /* If we haven't yet reached the end of the file (the buffer is full), and
# Line 587  while (ptr < endptr) Line 1319  while (ptr < endptr)
1319    
1320      memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD);      memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD);
1321      ptr -= MBUFTHIRD;      ptr -= MBUFTHIRD;
1322    
1323    #ifdef SUPPORT_LIBZ
1324        if (frtype == FR_LIBZ)
1325          bufflength = 2*MBUFTHIRD +
1326            gzread (ingz, buffer + 2*MBUFTHIRD, MBUFTHIRD);
1327        else
1328    #endif
1329    
1330    #ifdef SUPPORT_LIBBZ2
1331        if (frtype == FR_LIBBZ2)
1332          bufflength = 2*MBUFTHIRD +
1333            BZ2_bzread(inbz2, buffer + 2*MBUFTHIRD, MBUFTHIRD);
1334        else
1335    #endif
1336    
1337      bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in);      bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in);
1338    
1339      endptr = buffer + bufflength;      endptr = buffer + bufflength;
1340    
1341      /* Adjust any last match point */      /* Adjust any last match point */
# Line 599  while (ptr < endptr) Line 1347  while (ptr < endptr)
1347  /* End of file; print final "after" lines if wanted; do_after_lines sets  /* End of file; print final "after" lines if wanted; do_after_lines sets
1348  hyphenpending if it prints something. */  hyphenpending if it prints something. */
1349    
1350  do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname);  if (!only_matching && !count_only)
1351  hyphenpending |= endhyphenpending;    {
1352      do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname);
1353      hyphenpending |= endhyphenpending;
1354      }
1355    
1356  /* Print the file name if we are looking for those without matches and there  /* Print the file name if we are looking for those without matches and there
1357  were none. If we found a match, we won't have got this far. */  were none. If we found a match, we won't have got this far. */
1358    
1359  if (filenames_nomatch_only)  if (filenames == FN_NOMATCH_ONLY)
1360    {    {
1361    fprintf(stdout, "%s\n", printname);    fprintf(stdout, "%s\n", printname);
1362    return 0;    return 0;
# Line 615  if (filenames_nomatch_only) Line 1366  if (filenames_nomatch_only)
1366    
1367  if (count_only)  if (count_only)
1368    {    {
1369    if (printname != NULL) fprintf(stdout, "%s:", printname);    if (count > 0 || !omit_zero_count)
1370    fprintf(stdout, "%d\n", count);      {
1371        if (printname != NULL && filenames != FN_NONE)
1372          fprintf(stdout, "%s:", printname);
1373        fprintf(stdout, "%d\n", count);
1374        }
1375    }    }
1376    
1377  return rc;  return rc;
# Line 633  recursing; if it's a file, grep it. Line 1388  recursing; if it's a file, grep it.
1388    
1389  Arguments:  Arguments:
1390    pathname          the path to investigate    pathname          the path to investigate
1391    dir_recurse       TRUE if recursing is wanted (-r)    dir_recurse       TRUE if recursing is wanted (-r or -drecurse)
   show_filenames    TRUE if file names are wanted for multiple files, except  
                       for the only file at top level when not filenames_only  
1392    only_one_at_top   TRUE if the path is the only one at toplevel    only_one_at_top   TRUE if the path is the only one at toplevel
1393    
1394  Returns:   0 if there was at least one match  Returns:   0 if there was at least one match
# Line 646  However, file opening failures are suppr Line 1399  However, file opening failures are suppr
1399  */  */
1400    
1401  static int  static int
1402  grep_or_recurse(char *pathname, BOOL dir_recurse, BOOL show_filenames,  grep_or_recurse(char *pathname, BOOL dir_recurse, BOOL only_one_at_top)
   BOOL only_one_at_top)  
1403  {  {
1404  int rc = 1;  int rc = 1;
1405  int sep;  int sep;
1406  FILE *in;  int frtype;
1407  char *printname;  int pathlen;
1408    void *handle;
1409    FILE *in = NULL;           /* Ensure initialized */
1410    
1411    #ifdef SUPPORT_LIBZ
1412    gzFile ingz = NULL;
1413    #endif
1414    
1415    #ifdef SUPPORT_LIBBZ2
1416    BZFILE *inbz2 = NULL;
1417    #endif
1418    
1419  /* If the file name is "-" we scan stdin */  /* If the file name is "-" we scan stdin */
1420    
1421  if (strcmp(pathname, "-") == 0)  if (strcmp(pathname, "-") == 0)
1422    {    {
1423    return pcregrep(stdin,    return pcregrep(stdin, FR_PLAIN,
1424      (filenames_only || filenames_nomatch_only ||      (filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))?
     (show_filenames && !only_one_at_top))?  
1425        stdin_name : NULL);        stdin_name : NULL);
1426    }    }
1427    
1428  /* If the file is a directory and we are recursing, scan each file within it,  /* If the file is a directory, skip if skipping or if we are recursing, scan
1429  subject to any include or exclude patterns that were set. The scanning code is  each file and directory within it, subject to any include or exclude patterns
1430  localized so it can be made system-specific. */  that were set. The scanning code is localized so it can be made
1431    system-specific. */
1432    
1433  if ((sep = isdirectory(pathname)) != 0 && dir_recurse)  if ((sep = isdirectory(pathname)) != 0)
1434    {    {
1435    char buffer[1024];    if (dee_action == dee_SKIP) return 1;
1436    char *nextfile;    if (dee_action == dee_RECURSE)
1437    directory_type *dir = opendirectory(pathname);      {
1438        char buffer[1024];
1439        char *nextfile;
1440        directory_type *dir = opendirectory(pathname);
1441    
1442        if (dir == NULL)
1443          {
1444          if (!silent)
1445            fprintf(stderr, "pcregrep: Failed to open directory %s: %s\n", pathname,
1446              strerror(errno));
1447          return 2;
1448          }
1449    
1450        while ((nextfile = readdirectory(dir)) != NULL)
1451          {
1452          int frc, nflen;
1453          sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile);
1454          nflen = strlen(nextfile);
1455    
1456    if (dir == NULL)        if (isdirectory(buffer))
1457            {
1458            if (exclude_dir_compiled != NULL &&
1459                pcre_exec(exclude_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0)
1460              continue;
1461    
1462            if (include_dir_compiled != NULL &&
1463                pcre_exec(include_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0)
1464              continue;
1465            }
1466          else
1467            {
1468            if (exclude_compiled != NULL &&
1469                pcre_exec(exclude_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0)
1470              continue;
1471    
1472            if (include_compiled != NULL &&
1473                pcre_exec(include_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0)
1474              continue;
1475            }
1476    
1477          frc = grep_or_recurse(buffer, dir_recurse, FALSE);
1478          if (frc > 1) rc = frc;
1479           else if (frc == 0 && rc == 1) rc = 0;
1480          }
1481    
1482        closedirectory(dir);
1483        return rc;
1484        }
1485      }
1486    
1487    /* If the file is not a directory and not a regular file, skip it if that's
1488    been requested. */
1489    
1490    else if (!isregfile(pathname) && DEE_action == DEE_SKIP) return 1;
1491    
1492    /* Control reaches here if we have a regular file, or if we have a directory
1493    and recursion or skipping was not requested, or if we have anything else and
1494    skipping was not requested. The scan proceeds. If this is the first and only
1495    argument at top level, we don't show the file name, unless we are only showing
1496    the file name, or the filename was forced (-H). */
1497    
1498    pathlen = strlen(pathname);
1499    
1500    /* Open using zlib if it is supported and the file name ends with .gz. */
1501    
1502    #ifdef SUPPORT_LIBZ
1503    if (pathlen > 3 && strcmp(pathname + pathlen - 3, ".gz") == 0)
1504      {
1505      ingz = gzopen(pathname, "rb");
1506      if (ingz == NULL)
1507      {      {
1508      if (!silent)      if (!silent)
1509        fprintf(stderr, "pcregrep: Failed to open directory %s: %s\n", pathname,        fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
1510          strerror(errno));          strerror(errno));
1511      return 2;      return 2;
1512      }      }
1513      handle = (void *)ingz;
1514      frtype = FR_LIBZ;
1515      }
1516    else
1517    #endif
1518    
1519    while ((nextfile = readdirectory(dir)) != NULL)  /* Otherwise open with bz2lib if it is supported and the name ends with .bz2. */
     {  
     int frc, blen;  
     sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile);  
     blen = strlen(buffer);  
   
     if (exclude_compiled != NULL &&  
         pcre_exec(exclude_compiled, NULL, buffer, blen, 0, 0, NULL, 0) >= 0)  
       continue;  
1520    
1521      if (include_compiled != NULL &&  #ifdef SUPPORT_LIBBZ2
1522          pcre_exec(include_compiled, NULL, buffer, blen, 0, 0, NULL, 0) < 0)  if (pathlen > 4 && strcmp(pathname + pathlen - 4, ".bz2") == 0)
1523        continue;    {
1524      inbz2 = BZ2_bzopen(pathname, "rb");
1525      handle = (void *)inbz2;
1526      frtype = FR_LIBBZ2;
1527      }
1528    else
1529    #endif
1530    
1531      frc = grep_or_recurse(buffer, dir_recurse, TRUE, FALSE);  /* Otherwise use plain fopen(). The label is so that we can come back here if
1532      if (frc > 1) rc = frc;  an attempt to read a .bz2 file indicates that it really is a plain file. */
      else if (frc == 0 && rc == 1) rc = 0;  
     }  
1533    
1534    closedirectory(dir);  #ifdef SUPPORT_LIBBZ2
1535    return rc;  PLAIN_FILE:
1536    #endif
1537      {
1538      in = fopen(pathname, "rb");
1539      handle = (void *)in;
1540      frtype = FR_PLAIN;
1541    }    }
1542    
1543  /* If the file is not a directory, or we are not recursing, scan it. If this is  /* All the opening methods return errno when they fail. */
 the first and only argument at top level, we don't show the file name (unless  
 we are only showing the file name). Otherwise, control is via the  
 show_filenames variable. */  
1544    
1545  in = fopen(pathname, "r");  if (handle == NULL)
 if (in == NULL)  
1546    {    {
1547    if (!silent)    if (!silent)
1548      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
# Line 719  if (in == NULL) Line 1550  if (in == NULL)
1550    return 2;    return 2;
1551    }    }
1552    
1553  printname =  (filenames_only || filenames_nomatch_only ||  /* Now grep the file */
1554    (show_filenames && !only_one_at_top))? pathname : NULL;  
1555    rc = pcregrep(handle, frtype, (filenames > FN_DEFAULT ||
1556      (filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL);
1557    
1558    /* Close in an appropriate manner. */
1559    
1560    #ifdef SUPPORT_LIBZ
1561    if (frtype == FR_LIBZ)
1562      gzclose(ingz);
1563    else
1564    #endif
1565    
1566    /* If it is a .bz2 file and the result is 2, it means that the first attempt to
1567    read failed. If the error indicates that the file isn't in fact bzipped, try
1568    again as a normal file. */
1569    
1570    #ifdef SUPPORT_LIBBZ2
1571    if (frtype == FR_LIBBZ2)
1572      {
1573      if (rc == 2)
1574        {
1575        int errnum;
1576        const char *err = BZ2_bzerror(inbz2, &errnum);
1577        if (errnum == BZ_DATA_ERROR_MAGIC)
1578          {
1579          BZ2_bzclose(inbz2);
1580          goto PLAIN_FILE;
1581          }
1582        else if (!silent)
1583          fprintf(stderr, "pcregrep: Failed to read %s using bzlib: %s\n",
1584            pathname, err);
1585        }
1586      BZ2_bzclose(inbz2);
1587      }
1588    else
1589    #endif
1590    
1591  rc = pcregrep(in, printname);  /* Normal file close */
1592    
1593  fclose(in);  fclose(in);
1594    
1595    /* Pass back the yield from pcregrep(). */
1596    
1597  return rc;  return rc;
1598  }  }
1599    
# Line 738  return rc; Line 1607  return rc;
1607  static int  static int
1608  usage(int rc)  usage(int rc)
1609  {  {
1610  fprintf(stderr, "Usage: pcregrep [-LMVcfhilnqrsvwx] [long-options] [pattern] [file1 file2 ...]\n");  option_item *op;
1611  fprintf(stderr, "Type `pcregrep --help' for more information.\n");  fprintf(stderr, "Usage: pcregrep [-");
1612    for (op = optionlist; op->one_char != 0; op++)
1613      {
1614      if (op->one_char > 0) fprintf(stderr, "%c", op->one_char);
1615      }
1616    fprintf(stderr, "] [long options] [pattern] [files]\n");
1617    fprintf(stderr, "Type `pcregrep --help' for more information and the long "
1618      "options.\n");
1619  return rc;  return rc;
1620  }  }
1621    
# Line 757  option_item *op; Line 1633  option_item *op;
1633    
1634  printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n");  printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n");
1635  printf("Search for PATTERN in each FILE or standard input.\n");  printf("Search for PATTERN in each FILE or standard input.\n");
1636  printf("PATTERN must be present if -f is not used.\n");  printf("PATTERN must be present if neither -e nor -f is used.\n");
1637  printf("\"-\" can be used as a file name to mean STDIN.\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");  
1638    
1639    #ifdef SUPPORT_LIBZ
1640    printf("Files whose names end in .gz are read using zlib.\n");
1641    #endif
1642    
1643    #ifdef SUPPORT_LIBBZ2
1644    printf("Files whose names end in .bz2 are read using bzlib2.\n");
1645    #endif
1646    
1647    #if defined SUPPORT_LIBZ || defined SUPPORT_LIBBZ2
1648    printf("Other files and the standard input are read as plain files.\n\n");
1649    #else
1650    printf("All files are read as plain files, without any interpretation.\n\n");
1651    #endif
1652    
1653    printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n");
1654  printf("Options:\n");  printf("Options:\n");
1655    
1656  for (op = optionlist; op->one_char != 0; op++)  for (op = optionlist; op->one_char != 0; op++)
# Line 768  for (op = optionlist; op->one_char != 0; Line 1658  for (op = optionlist; op->one_char != 0;
1658    int n;    int n;
1659    char s[4];    char s[4];
1660    if (op->one_char > 0) sprintf(s, "-%c,", op->one_char); else strcpy(s, "   ");    if (op->one_char > 0) sprintf(s, "-%c,", op->one_char); else strcpy(s, "   ");
1661    printf("  %s --%s%n", s, op->long_name, &n);    n = 30 - printf("  %s --%s", s, op->long_name);
   n = 30 - n;  
1662    if (n < 1) n = 1;    if (n < 1) n = 1;
1663    printf("%.*s%s\n", n, "                    ", op->help_text);    printf("%.*s%s\n", n, "                    ", op->help_text);
1664    }    }
# Line 794  handle_option(int letter, int options) Line 1683  handle_option(int letter, int options)
1683  {  {
1684  switch(letter)  switch(letter)
1685    {    {
1686    case -1:  help(); exit(0);    case N_FOFFSETS: file_offsets = TRUE; break;
1687      case N_HELP: help(); exit(0);
1688      case N_LOFFSETS: line_offsets = number = TRUE; break;
1689    case 'c': count_only = TRUE; break;    case 'c': count_only = TRUE; break;
1690    case 'h': filenames = FALSE; break;    case 'F': process_options |= PO_FIXED_STRINGS; break;
1691      case 'H': filenames = FN_FORCE; break;
1692      case 'h': filenames = FN_NONE; break;
1693    case 'i': options |= PCRE_CASELESS; break;    case 'i': options |= PCRE_CASELESS; break;
1694    case 'l': filenames_only = TRUE; break;    case 'l': omit_zero_count = TRUE; filenames = FN_MATCH_ONLY; break;
1695    case 'L': filenames_nomatch_only = TRUE; break;    case 'L': filenames = FN_NOMATCH_ONLY; break;
1696    case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break;    case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break;
1697    case 'n': number = TRUE; break;    case 'n': number = TRUE; break;
1698      case 'o': only_matching = TRUE; break;
1699    case 'q': quiet = TRUE; break;    case 'q': quiet = TRUE; break;
1700    case 'r': recurse = TRUE; break;    case 'r': dee_action = dee_RECURSE; break;
1701    case 's': silent = TRUE; break;    case 's': silent = TRUE; break;
1702    case 'u': options |= PCRE_UTF8; break;    case 'u': options |= PCRE_UTF8; utf8 = TRUE; break;
1703    case 'v': invert = TRUE; break;    case 'v': invert = TRUE; break;
1704    case 'w': word_match = TRUE; break;    case 'w': process_options |= PO_WORD_MATCH; break;
1705    case 'x': whole_lines = TRUE; break;    case 'x': process_options |= PO_LINE_MATCH; break;
1706    
1707    case 'V':    case 'V':
1708    fprintf(stderr, "pcregrep version %s using ", VERSION);    fprintf(stderr, "pcregrep version %s\n", pcre_version());
   fprintf(stderr, "PCRE version %s\n", pcre_version());  
1709    exit(0);    exit(0);
1710    break;    break;
1711    
# Line 828  return options; Line 1721  return options;
1721    
1722    
1723  /*************************************************  /*************************************************
1724    *          Construct printed ordinal             *
1725    *************************************************/
1726    
1727    /* This turns a number into "1st", "3rd", etc. */
1728    
1729    static char *
1730    ordin(int n)
1731    {
1732    static char buffer[8];
1733    char *p = buffer;
1734    sprintf(p, "%d", n);
1735    while (*p != 0) p++;
1736    switch (n%10)
1737      {
1738      case 1: strcpy(p, "st"); break;
1739      case 2: strcpy(p, "nd"); break;
1740      case 3: strcpy(p, "rd"); break;
1741      default: strcpy(p, "th"); break;
1742      }
1743    return buffer;
1744    }
1745    
1746    
1747    
1748    /*************************************************
1749    *          Compile a single pattern              *
1750    *************************************************/
1751    
1752    /* When the -F option has been used, this is called for each substring.
1753    Otherwise it's called for each supplied pattern.
1754    
1755    Arguments:
1756      pattern        the pattern string
1757      options        the PCRE options
1758      filename       the file name, or NULL for a command-line pattern
1759      count          0 if this is the only command line pattern, or
1760                     number of the command line pattern, or
1761                     linenumber for a pattern from a file
1762    
1763    Returns:         TRUE on success, FALSE after an error
1764    */
1765    
1766    static BOOL
1767    compile_single_pattern(char *pattern, int options, char *filename, int count)
1768    {
1769    char buffer[MBUFTHIRD + 16];
1770    const char *error;
1771    int errptr;
1772    
1773    if (pattern_count >= MAX_PATTERN_COUNT)
1774      {
1775      fprintf(stderr, "pcregrep: Too many %spatterns (max %d)\n",
1776        (filename == NULL)? "command-line " : "", MAX_PATTERN_COUNT);
1777      return FALSE;
1778      }
1779    
1780    sprintf(buffer, "%s%.*s%s", prefix[process_options], MBUFTHIRD, pattern,
1781      suffix[process_options]);
1782    pattern_list[pattern_count] =
1783      pcre_compile(buffer, options, &error, &errptr, pcretables);
1784    if (pattern_list[pattern_count] != NULL)
1785      {
1786      pattern_count++;
1787      return TRUE;
1788      }
1789    
1790    /* Handle compile errors */
1791    
1792    errptr -= (int)strlen(prefix[process_options]);
1793    if (errptr > (int)strlen(pattern)) errptr = (int)strlen(pattern);
1794    
1795    if (filename == NULL)
1796      {
1797      if (count == 0)
1798        fprintf(stderr, "pcregrep: Error in command-line regex "
1799          "at offset %d: %s\n", errptr, error);
1800      else
1801        fprintf(stderr, "pcregrep: Error in %s command-line regex "
1802          "at offset %d: %s\n", ordin(count), errptr, error);
1803      }
1804    else
1805      {
1806      fprintf(stderr, "pcregrep: Error in regex in line %d of %s "
1807        "at offset %d: %s\n", count, filename, errptr, error);
1808      }
1809    
1810    return FALSE;
1811    }
1812    
1813    
1814    
1815    /*************************************************
1816    *           Compile one supplied pattern         *
1817    *************************************************/
1818    
1819    /* When the -F option has been used, each string may be a list of strings,
1820    separated by line breaks. They will be matched literally.
1821    
1822    Arguments:
1823      pattern        the pattern string
1824      options        the PCRE options
1825      filename       the file name, or NULL for a command-line pattern
1826      count          0 if this is the only command line pattern, or
1827                     number of the command line pattern, or
1828                     linenumber for a pattern from a file
1829    
1830    Returns:         TRUE on success, FALSE after an error
1831    */
1832    
1833    static BOOL
1834    compile_pattern(char *pattern, int options, char *filename, int count)
1835    {
1836    if ((process_options & PO_FIXED_STRINGS) != 0)
1837      {
1838      char *eop = pattern + strlen(pattern);
1839      char buffer[MBUFTHIRD];
1840      for(;;)
1841        {
1842        int ellength;
1843        char *p = end_of_line(pattern, eop, &ellength);
1844        if (ellength == 0)
1845          return compile_single_pattern(pattern, options, filename, count);
1846        sprintf(buffer, "%.*s", (int)(p - pattern - ellength), pattern);
1847        pattern = p;
1848        if (!compile_single_pattern(buffer, options, filename, count))
1849          return FALSE;
1850        }
1851      }
1852    else return compile_single_pattern(pattern, options, filename, count);
1853    }
1854    
1855    
1856    
1857    /*************************************************
1858  *                Main program                    *  *                Main program                    *
1859  *************************************************/  *************************************************/
1860    
# Line 838  main(int argc, char **argv) Line 1865  main(int argc, char **argv)
1865  {  {
1866  int i, j;  int i, j;
1867  int rc = 1;  int rc = 1;
1868  int options = 0;  int pcre_options = 0;
1869    int cmd_pattern_count = 0;
1870    int hint_count = 0;
1871  int errptr;  int errptr;
 const char *error;  
1872  BOOL only_one_at_top;  BOOL only_one_at_top;
1873    char *patterns[MAX_PATTERN_COUNT];
1874    const char *locale_from = "--locale";
1875    const char *error;
1876    
1877    /* Set the default line ending value from the default in the PCRE library;
1878    "lf", "cr", "crlf", and "any" are supported. Anything else is treated as "lf".
1879    Note that the return values from pcre_config(), though derived from the ASCII
1880    codes, are the same in EBCDIC environments, so we must use the actual values
1881    rather than escapes such as as '\r'. */
1882    
1883    (void)pcre_config(PCRE_CONFIG_NEWLINE, &i);
1884    switch(i)
1885      {
1886      default:               newline = (char *)"lf"; break;
1887      case 13:               newline = (char *)"cr"; break;
1888      case (13 << 8) | 10:   newline = (char *)"crlf"; break;
1889      case -1:               newline = (char *)"any"; break;
1890      case -2:               newline = (char *)"anycrlf"; break;
1891      }
1892    
1893  /* Process the options */  /* Process the options */
1894    
# Line 855  for (i = 1; i < argc; i++) Line 1902  for (i = 1; i < argc; i++)
1902    if (argv[i][0] != '-') break;    if (argv[i][0] != '-') break;
1903    
1904    /* If we hit an argument that is just "-", it may be a reference to STDIN,    /* If we hit an argument that is just "-", it may be a reference to STDIN,
1905    but only if we have previously had -f to define the patterns. */    but only if we have previously had -e or -f to define the patterns. */
1906    
1907    if (argv[i][1] == 0)    if (argv[i][1] == 0)
1908      {      {
1909      if (pattern_filename != NULL) break;      if (pattern_filename != NULL || pattern_count > 0) break;
1910        else exit(usage(2));        else exit(usage(2));
1911      }      }
1912    
# Line 881  for (i = 1; i < argc; i++) Line 1928  for (i = 1; i < argc; i++)
1928      /* Some long options have data that follows after =, for example file=name.      /* Some long options have data that follows after =, for example file=name.
1929      Some options have variations in the long name spelling: specifically, we      Some options have variations in the long name spelling: specifically, we
1930      allow "regexp" because GNU grep allows it, though I personally go along      allow "regexp" because GNU grep allows it, though I personally go along
1931      with Jeff Friedl in preferring "regex" without the "p". These options are      with Jeffrey Friedl and Larry Wall in preferring "regex" without the "p".
1932      entered in the table as "regex(p)". No option is in both these categories,      These options are entered in the table as "regex(p)". Options can be in
1933      fortunately. */      both these categories. */
1934    
1935      for (op = optionlist; op->one_char != 0; op++)      for (op = optionlist; op->one_char != 0; op++)
1936        {        {
1937        char *opbra = strchr(op->long_name, '(');        char *opbra = strchr(op->long_name, '(');
1938        char *equals = strchr(op->long_name, '=');        char *equals = strchr(op->long_name, '=');
1939        if (opbra == NULL)     /* Not a (p) case */  
1940          /* Handle options with only one spelling of the name */
1941    
1942          if (opbra == NULL)     /* Does not contain '(' */
1943          {          {
1944          if (equals == NULL)  /* Not thing=data case */          if (equals == NULL)  /* Not thing=data case */
1945            {            {
# Line 898  for (i = 1; i < argc; i++) Line 1948  for (i = 1; i < argc; i++)
1948          else                 /* Special case xxx=data */          else                 /* Special case xxx=data */
1949            {            {
1950            int oplen = equals - op->long_name;            int oplen = equals - op->long_name;
1951            int arglen = (argequals == NULL)? strlen(arg) : argequals - arg;            int arglen = (argequals == NULL)? (int)strlen(arg) : argequals - arg;
1952            if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0)            if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0)
1953              {              {
1954              option_data = arg + arglen;              option_data = arg + arglen;
# Line 911  for (i = 1; i < argc; i++) Line 1961  for (i = 1; i < argc; i++)
1961              }              }
1962            }            }
1963          }          }
1964        else                   /* Special case xxxx(p) */  
1965          /* Handle options with an alternate spelling of the name */
1966    
1967          else
1968          {          {
1969          char buff1[24];          char buff1[24];
1970          char buff2[24];          char buff2[24];
1971    
1972          int baselen = opbra - op->long_name;          int baselen = opbra - op->long_name;
1973            int fulllen = strchr(op->long_name, ')') - op->long_name + 1;
1974            int arglen = (argequals == NULL || equals == NULL)?
1975              (int)strlen(arg) : argequals - arg;
1976    
1977          sprintf(buff1, "%.*s", baselen, op->long_name);          sprintf(buff1, "%.*s", baselen, op->long_name);
1978          sprintf(buff2, "%s%.*s", buff1, strlen(op->long_name) - baselen - 2,          sprintf(buff2, "%s%.*s", buff1, fulllen - baselen - 2, opbra + 1);
1979            opbra + 1);  
1980          if (strcmp(arg, buff1) == 0 || strcmp(arg, buff2) == 0)          if (strncmp(arg, buff1, arglen) == 0 ||
1981               strncmp(arg, buff2, arglen) == 0)
1982              {
1983              if (equals != NULL && argequals != NULL)
1984                {
1985                option_data = argequals;
1986                if (*option_data == '=')
1987                  {
1988                  option_data++;
1989                  longopwasequals = TRUE;
1990                  }
1991                }
1992            break;            break;
1993              }
1994          }          }
1995        }        }
1996    
# Line 931  for (i = 1; i < argc; i++) Line 2001  for (i = 1; i < argc; i++)
2001        }        }
2002      }      }
2003    
2004      /* Jeffrey Friedl's debugging harness uses these additional options which
2005      are not in the right form for putting in the option table because they use
2006      only one hyphen, yet are more than one character long. By putting them
2007      separately here, they will not get displayed as part of the help() output,
2008      but I don't think Jeffrey will care about that. */
2009    
2010    #ifdef JFRIEDL_DEBUG
2011      else if (strcmp(argv[i], "-pre") == 0) {
2012              jfriedl_prefix = argv[++i];
2013              continue;
2014      } else if (strcmp(argv[i], "-post") == 0) {
2015              jfriedl_postfix = argv[++i];
2016              continue;
2017      } else if (strcmp(argv[i], "-XT") == 0) {
2018              sscanf(argv[++i], "%d", &jfriedl_XT);
2019              continue;
2020      } else if (strcmp(argv[i], "-XR") == 0) {
2021              sscanf(argv[++i], "%d", &jfriedl_XR);
2022              continue;
2023      }
2024    #endif
2025    
2026    
2027    /* One-char options; many that have no data may be in a single argument; we    /* One-char options; many that have no data may be in a single argument; we
2028    continue till we hit the last one or one that needs data. */    continue till we hit the last one or one that needs data. */
2029    
# Line 953  for (i = 1; i < argc; i++) Line 2046  for (i = 1; i < argc; i++)
2046          option_data = s+1;          option_data = s+1;
2047          break;          break;
2048          }          }
2049        options = handle_option(*s++, options);        pcre_options = handle_option(*s++, pcre_options);
2050        }        }
2051      }      }
2052    
2053    /* At this point we should have op pointing to a matched option */    /* At this point we should have op pointing to a matched option. If the type
2054      is NO_DATA, it means that there is no data, and the option might set
2055      something in the PCRE options. */
2056    
2057    if (op->type == OP_NODATA)    if (op->type == OP_NODATA)
     options = handle_option(op->one_char, options);  
   else  
2058      {      {
2059      if (*option_data == 0)      pcre_options = handle_option(op->one_char, pcre_options);
2060        continue;
2061        }
2062    
2063      /* If the option type is OP_OP_STRING or OP_OP_NUMBER, it's an option that
2064      either has a value or defaults to something. It cannot have data in a
2065      separate item. At the moment, the only such options are "colo(u)r" and
2066      Jeffrey Friedl's special -S debugging option. */
2067    
2068      if (*option_data == 0 &&
2069          (op->type == OP_OP_STRING || op->type == OP_OP_NUMBER))
2070        {
2071        switch (op->one_char)
2072        {        {
2073        if (i >= argc - 1 || longopwasequals)        case N_COLOUR:
2074          {        colour_option = (char *)"auto";
2075          fprintf(stderr, "pcregrep: Data missing after %s\n", argv[i]);        break;
2076          exit(usage(2));  #ifdef JFRIEDL_DEBUG
2077          }        case 'S':
2078        option_data = argv[++i];        S_arg = 0;
2079          break;
2080    #endif
2081          }
2082        continue;
2083        }
2084    
2085      /* Otherwise, find the data string for the option. */
2086    
2087      if (*option_data == 0)
2088        {
2089        if (i >= argc - 1 || longopwasequals)
2090          {
2091          fprintf(stderr, "pcregrep: Data missing after %s\n", argv[i]);
2092          exit(usage(2));
2093          }
2094        option_data = argv[++i];
2095        }
2096    
2097      /* If the option type is OP_PATLIST, it's the -e option, which can be called
2098      multiple times to create a list of patterns. */
2099    
2100      if (op->type == OP_PATLIST)
2101        {
2102        if (cmd_pattern_count >= MAX_PATTERN_COUNT)
2103          {
2104          fprintf(stderr, "pcregrep: Too many command-line patterns (max %d)\n",
2105            MAX_PATTERN_COUNT);
2106          return 2;
2107        }        }
2108        patterns[cmd_pattern_count++] = option_data;
2109        }
2110    
2111      /* Otherwise, deal with single string or numeric data values. */
2112    
2113      if (op->type == OP_STRING) *((char **)op->dataptr) = option_data; else    else if (op->type != OP_NUMBER && op->type != OP_OP_NUMBER)
2114        {
2115        *((char **)op->dataptr) = option_data;
2116        }
2117      else
2118        {
2119        char *endptr;
2120        int n = strtoul(option_data, &endptr, 10);
2121        if (*endptr != 0)
2122        {        {
2123        char *endptr;        if (longop)
       int n = strtoul(option_data, &endptr, 10);  
       if (*endptr != 0)  
2124          {          {
2125          if (longop)          char *equals = strchr(op->long_name, '=');
2126            fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%s\n",          int nlen = (equals == NULL)? (int)strlen(op->long_name) :
2127              option_data, op->long_name);            equals - op->long_name;
2128          else          fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%.*s\n",
2129            fprintf(stderr, "pcregrep: Malformed number \"%s\" after -%c\n",            option_data, nlen, op->long_name);
             option_data, op->one_char);  
         exit(usage(2));  
2130          }          }
2131        *((int *)op->dataptr) = n;        else
2132            fprintf(stderr, "pcregrep: Malformed number \"%s\" after -%c\n",
2133              option_data, op->one_char);
2134          exit(usage(2));
2135        }        }
2136        *((int *)op->dataptr) = n;
2137      }      }
2138    }    }
2139    
# Line 1001  if (both_context > 0) Line 2146  if (both_context > 0)
2146    if (before_context == 0) before_context = both_context;    if (before_context == 0) before_context = both_context;
2147    }    }
2148    
2149  pattern_list = (pcre **)malloc(MAX_PATTERN_COUNT * sizeof(pcre *));  /* Only one of --only-matching, --file-offsets, or --line-offsets is permitted.
2150  hints_list = (pcre_extra **)malloc(MAX_PATTERN_COUNT * sizeof(pcre_extra *));  However, the latter two set the only_matching flag. */
2151    
2152  if (pattern_list == NULL || hints_list == NULL)  if ((only_matching && (file_offsets || line_offsets)) ||
2153        (file_offsets && line_offsets))
2154    {    {
2155    fprintf(stderr, "pcregrep: malloc failed\n");    fprintf(stderr, "pcregrep: Cannot mix --only-matching, --file-offsets "
2156    return 2;      "and/or --line-offsets\n");
2157      exit(usage(2));
2158    }    }
2159    
2160  /* Compile the regular expression(s). */  if (file_offsets || line_offsets) only_matching = TRUE;
2161    
2162  if (pattern_filename != NULL)  /* If a locale has not been provided as an option, see if the LC_CTYPE or
2163    LC_ALL environment variable is set, and if so, use it. */
2164    
2165    if (locale == NULL)
2166    {    {
2167    FILE *f = fopen(pattern_filename, "r");    locale = getenv("LC_ALL");
2168    char buffer[MBUFTHIRD + 16];    locale_from = "LCC_ALL";
2169    char *rdstart;    }
   int adjust = 0;  
2170    
2171    if (f == NULL)  if (locale == NULL)
2172      {
2173      locale = getenv("LC_CTYPE");
2174      locale_from = "LC_CTYPE";
2175      }
2176    
2177    /* If a locale has been provided, set it, and generate the tables the PCRE
2178    needs. Otherwise, pcretables==NULL, which causes the use of default tables. */
2179    
2180    if (locale != NULL)
2181      {
2182      if (setlocale(LC_CTYPE, locale) == NULL)
2183      {      {
2184      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pattern_filename,      fprintf(stderr, "pcregrep: Failed to set locale %s (obtained from %s)\n",
2185        strerror(errno));        locale, locale_from);
2186      return 2;      return 2;
2187      }      }
2188      pcretables = pcre_maketables();
2189      }
2190    
2191    if (whole_lines)  /* Sort out colouring */
2192    
2193    if (colour_option != NULL && strcmp(colour_option, "never") != 0)
2194      {
2195      if (strcmp(colour_option, "always") == 0) do_colour = TRUE;
2196      else if (strcmp(colour_option, "auto") == 0) do_colour = is_stdout_tty();
2197      else
2198      {      {
2199      strcpy(buffer, "^(?:");      fprintf(stderr, "pcregrep: Unknown colour setting \"%s\"\n",
2200      adjust = 4;        colour_option);
2201        return 2;
2202      }      }
2203    else if (word_match)    if (do_colour)
2204      {      {
2205      strcpy(buffer, "\\b");      char *cs = getenv("PCREGREP_COLOUR");
2206      adjust = 2;      if (cs == NULL) cs = getenv("PCREGREP_COLOR");
2207        if (cs != NULL) colour_string = cs;
2208      }      }
2209      }
2210    
2211    rdstart = buffer + adjust;  /* Interpret the newline type; the default settings are Unix-like. */
2212    while (fgets(rdstart, MBUFTHIRD, f) != NULL)  
2213    if (strcmp(newline, "cr") == 0 || strcmp(newline, "CR") == 0)
2214      {
2215      pcre_options |= PCRE_NEWLINE_CR;
2216      endlinetype = EL_CR;
2217      }
2218    else if (strcmp(newline, "lf") == 0 || strcmp(newline, "LF") == 0)
2219      {
2220      pcre_options |= PCRE_NEWLINE_LF;
2221      endlinetype = EL_LF;
2222      }
2223    else if (strcmp(newline, "crlf") == 0 || strcmp(newline, "CRLF") == 0)
2224      {
2225      pcre_options |= PCRE_NEWLINE_CRLF;
2226      endlinetype = EL_CRLF;
2227      }
2228    else if (strcmp(newline, "any") == 0 || strcmp(newline, "ANY") == 0)
2229      {
2230      pcre_options |= PCRE_NEWLINE_ANY;
2231      endlinetype = EL_ANY;
2232      }
2233    else if (strcmp(newline, "anycrlf") == 0 || strcmp(newline, "ANYCRLF") == 0)
2234      {
2235      pcre_options |= PCRE_NEWLINE_ANYCRLF;
2236      endlinetype = EL_ANYCRLF;
2237      }
2238    else
2239      {
2240      fprintf(stderr, "pcregrep: Invalid newline specifier \"%s\"\n", newline);
2241      return 2;
2242      }
2243    
2244    /* Interpret the text values for -d and -D */
2245    
2246    if (dee_option != NULL)
2247      {
2248      if (strcmp(dee_option, "read") == 0) dee_action = dee_READ;
2249      else if (strcmp(dee_option, "recurse") == 0) dee_action = dee_RECURSE;
2250      else if (strcmp(dee_option, "skip") == 0) dee_action = dee_SKIP;
2251      else
2252      {      {
2253      char *s = rdstart + (int)strlen(rdstart);      fprintf(stderr, "pcregrep: Invalid value \"%s\" for -d\n", dee_option);
2254      if (pattern_count >= MAX_PATTERN_COUNT)      return 2;
       {  
       fprintf(stderr, "pcregrep: Too many patterns in file (max %d)\n",  
         MAX_PATTERN_COUNT);  
       return 2;  
       }  
     while (s > rdstart && isspace((unsigned char)(s[-1]))) s--;  
     if (s == rdstart) continue;  
     if (whole_lines) strcpy(s, ")$");  
       else if (word_match)strcpy(s, "\\b");  
         else *s = 0;  
     pattern_list[pattern_count] = pcre_compile(buffer, options, &error,  
       &errptr, NULL);  
     if (pattern_list[pattern_count++] == NULL)  
       {  
       fprintf(stderr, "pcregrep: Error in regex number %d at offset %d: %s\n",  
         pattern_count, errptr - adjust, error);  
       return 2;  
       }  
2255      }      }
   fclose(f);  
2256    }    }
2257    
2258  /* If no file name, a single regex must be given inline. */  if (DEE_option != NULL)
2259      {
2260      if (strcmp(DEE_option, "read") == 0) DEE_action = DEE_READ;
2261      else if (strcmp(DEE_option, "skip") == 0) DEE_action = DEE_SKIP;
2262      else
2263        {
2264        fprintf(stderr, "pcregrep: Invalid value \"%s\" for -D\n", DEE_option);
2265        return 2;
2266        }
2267      }
2268    
2269  else  /* Check the values for Jeffrey Friedl's debugging options. */
2270    
2271    #ifdef JFRIEDL_DEBUG
2272    if (S_arg > 9)
2273      {
2274      fprintf(stderr, "pcregrep: bad value for -S option\n");
2275      return 2;
2276      }
2277    if (jfriedl_XT != 0 || jfriedl_XR != 0)
2278      {
2279      if (jfriedl_XT == 0) jfriedl_XT = 1;
2280      if (jfriedl_XR == 0) jfriedl_XR = 1;
2281      }
2282    #endif
2283    
2284    /* Get memory to store the pattern and hints lists. */
2285    
2286    pattern_list = (pcre **)malloc(MAX_PATTERN_COUNT * sizeof(pcre *));
2287    hints_list = (pcre_extra **)malloc(MAX_PATTERN_COUNT * sizeof(pcre_extra *));
2288    
2289    if (pattern_list == NULL || hints_list == NULL)
2290    {    {
2291    char buffer[MBUFTHIRD + 16];    fprintf(stderr, "pcregrep: malloc failed\n");
2292    char *pat;    goto EXIT2;
2293    int adjust = 0;    }
2294    
2295    /* If no patterns were provided by -e, and there is no file provided by -f,
2296    the first argument is the one and only pattern, and it must exist. */
2297    
2298    if (cmd_pattern_count == 0 && pattern_filename == NULL)
2299      {
2300    if (i >= argc) return usage(2);    if (i >= argc) return usage(2);
2301      patterns[cmd_pattern_count++] = argv[i++];
2302      }
2303    
2304    /* Compile the patterns that were provided on the command line, either by
2305    multiple uses of -e or as a single unkeyed pattern. */
2306    
2307    for (j = 0; j < cmd_pattern_count; j++)
2308      {
2309      if (!compile_pattern(patterns[j], pcre_options, NULL,
2310           (j == 0 && cmd_pattern_count == 1)? 0 : j + 1))
2311        goto EXIT2;
2312      }
2313    
2314    if (whole_lines)  /* Compile the regular expressions that are provided in a file. */
2315    
2316    if (pattern_filename != NULL)
2317      {
2318      int linenumber = 0;
2319      FILE *f;
2320      char *filename;
2321      char buffer[MBUFTHIRD];
2322    
2323      if (strcmp(pattern_filename, "-") == 0)
2324      {      {
2325      sprintf(buffer, "^(?:%.*s)$", MBUFTHIRD, argv[i++]);      f = stdin;
2326      pat = buffer;      filename = stdin_name;
     adjust = 4;  
2327      }      }
2328    else if (word_match)    else
2329      {      {
2330      sprintf(buffer, "\\b%.*s\\b", MBUFTHIRD, argv[i++]);      f = fopen(pattern_filename, "r");
2331      pat = buffer;      if (f == NULL)
2332      adjust = 2;        {
2333          fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pattern_filename,
2334            strerror(errno));
2335          goto EXIT2;
2336          }
2337        filename = pattern_filename;
2338      }      }
   else pat = argv[i++];  
2339    
2340    pattern_list[0] = pcre_compile(pat, options, &error, &errptr, NULL);    while (fgets(buffer, MBUFTHIRD, f) != NULL)
   
   if (pattern_list[0] == NULL)  
2341      {      {
2342      fprintf(stderr, "pcregrep: Error in regex at offset %d: %s\n",      char *s = buffer + (int)strlen(buffer);
2343        errptr - adjust, error);      while (s > buffer && isspace((unsigned char)(s[-1]))) s--;
2344      return 2;      *s = 0;
2345        linenumber++;
2346        if (buffer[0] == 0) continue;   /* Skip blank lines */
2347        if (!compile_pattern(buffer, pcre_options, filename, linenumber))
2348          goto EXIT2;
2349      }      }
2350    pattern_count++;  
2351      if (f != stdin) fclose(f);
2352    }    }
2353    
2354  /* Study the regular expressions, as we will be running them many times */  /* Study the regular expressions, as we will be running them many times */
# Line 1109  for (j = 0; j < pattern_count; j++) Line 2361  for (j = 0; j < pattern_count; j++)
2361      char s[16];      char s[16];
2362      if (pattern_count == 1) s[0] = 0; else sprintf(s, " number %d", j);      if (pattern_count == 1) s[0] = 0; else sprintf(s, " number %d", j);
2363      fprintf(stderr, "pcregrep: Error while studying regex%s: %s\n", s, error);      fprintf(stderr, "pcregrep: Error while studying regex%s: %s\n", s, error);
2364      return 2;      goto EXIT2;
2365      }      }
2366      hint_count++;
2367    }    }
2368    
2369  /* If there are include or exclude patterns, compile them. */  /* If there are include or exclude patterns, compile them. */
2370    
2371  if (exclude_pattern != NULL)  if (exclude_pattern != NULL)
2372    {    {
2373    exclude_compiled = pcre_compile(exclude_pattern, 0, &error, &errptr, NULL);    exclude_compiled = pcre_compile(exclude_pattern, 0, &error, &errptr,
2374        pcretables);
2375    if (exclude_compiled == NULL)    if (exclude_compiled == NULL)
2376      {      {
2377      fprintf(stderr, "pcregrep: Error in 'exclude' regex at offset %d: %s\n",      fprintf(stderr, "pcregrep: Error in 'exclude' regex at offset %d: %s\n",
2378        errptr, error);        errptr, error);
2379      return 2;      goto EXIT2;
2380      }      }
2381    }    }
2382    
2383  if (include_pattern != NULL)  if (include_pattern != NULL)
2384    {    {
2385    include_compiled = pcre_compile(include_pattern, 0, &error, &errptr, NULL);    include_compiled = pcre_compile(include_pattern, 0, &error, &errptr,
2386        pcretables);
2387    if (include_compiled == NULL)    if (include_compiled == NULL)
2388      {      {
2389      fprintf(stderr, "pcregrep: Error in 'include' regex at offset %d: %s\n",      fprintf(stderr, "pcregrep: Error in 'include' regex at offset %d: %s\n",
2390        errptr, error);        errptr, error);
2391      return 2;      goto EXIT2;
2392      }      }
2393    }    }
2394    
2395  /* If there are no further arguments, do the business on stdin and exit */  if (exclude_dir_pattern != NULL)
2396      {
2397      exclude_dir_compiled = pcre_compile(exclude_dir_pattern, 0, &error, &errptr,
2398        pcretables);
2399      if (exclude_dir_compiled == NULL)
2400        {
2401        fprintf(stderr, "pcregrep: Error in 'exclude_dir' regex at offset %d: %s\n",
2402          errptr, error);
2403        goto EXIT2;
2404        }
2405      }
2406    
2407    if (include_dir_pattern != NULL)
2408      {
2409      include_dir_compiled = pcre_compile(include_dir_pattern, 0, &error, &errptr,
2410        pcretables);
2411      if (include_dir_compiled == NULL)
2412        {
2413        fprintf(stderr, "pcregrep: Error in 'include_dir' regex at offset %d: %s\n",
2414          errptr, error);
2415        goto EXIT2;
2416        }
2417      }
2418    
2419  if (i >= argc) return pcregrep(stdin,  /* If there are no further arguments, do the business on stdin and exit. */
2420    (filenames_only || filenames_nomatch_only)? stdin_name : NULL);  
2421    if (i >= argc)
2422      {
2423      rc = pcregrep(stdin, FR_PLAIN, (filenames > FN_DEFAULT)? stdin_name : NULL);
2424      goto EXIT;
2425      }
2426    
2427  /* Otherwise, work through the remaining arguments as files or directories.  /* Otherwise, work through the remaining arguments as files or directories.
2428  Pass in the fact that there is only one argument at top level - this suppresses  Pass in the fact that there is only one argument at top level - this suppresses
2429  the file name if the argument is not a directory and filenames_only is not set.  the file name if the argument is not a directory and filenames are not
2430  */  otherwise forced. */
2431    
2432  only_one_at_top = (i == argc - 1);  only_one_at_top = i == argc - 1;   /* Catch initial value of i */
2433    
2434  for (; i < argc; i++)  for (; i < argc; i++)
2435    {    {
2436    int frc = grep_or_recurse(argv[i], recurse, filenames, only_one_at_top);    int frc = grep_or_recurse(argv[i], dee_action == dee_RECURSE,
2437        only_one_at_top);
2438    if (frc > 1) rc = frc;    if (frc > 1) rc = frc;
2439      else if (frc == 0 && rc == 1) rc = 0;      else if (frc == 0 && rc == 1) rc = 0;
2440    }    }
2441    
2442    EXIT:
2443    if (pattern_list != NULL)
2444      {
2445      for (i = 0; i < pattern_count; i++) free(pattern_list[i]);
2446      free(pattern_list);
2447      }
2448    if (hints_list != NULL)
2449      {
2450      for (i = 0; i < hint_count; i++) free(hints_list[i]);
2451      free(hints_list);
2452      }
2453  return rc;  return rc;
2454    
2455    EXIT2:
2456    rc = 2;
2457    goto EXIT;
2458  }  }
2459    
2460  /* End of pcregrep */  /* End of pcregrep */

Legend:
Removed from v.77  
changed lines
  Added in v.461

  ViewVC Help
Powered by ViewVC 1.1.5