/[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 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 345 by ph10, Mon Apr 28 15:10:02 2008 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) 1997-2008 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 "4.1 05-Sep-2005"  
73  #define MAX_PATTERN_COUNT 100  #define MAX_PATTERN_COUNT 100
74    
75  #if BUFSIZ > 8192  #if BUFSIZ > 8192
# Line 64  typedef int BOOL; Line 78  typedef int BOOL;
78  #define MBUFTHIRD 8192  #define MBUFTHIRD 8192
79  #endif  #endif
80    
81    /* Values for the "filenames" variable, which specifies options for file name
82    output. The order is important; it is assumed that a file name is wanted for
83    all values greater than FN_DEFAULT. */
84    
85    enum { FN_NONE, FN_DEFAULT, FN_ONLY, FN_NOMATCH_ONLY, FN_FORCE };
86    
87    /* File reading styles */
88    
89    enum { FR_PLAIN, FR_LIBZ, FR_LIBBZ2 };
90    
91    /* Actions for the -d and -D options */
92    
93    enum { dee_READ, dee_SKIP, dee_RECURSE };
94    enum { DEE_READ, DEE_SKIP };
95    
96    /* Actions for special processing options (flag bits) */
97    
98    #define PO_WORD_MATCH     0x0001
99    #define PO_LINE_MATCH     0x0002
100    #define PO_FIXED_STRINGS  0x0004
101    
102    /* Line ending types */
103    
104    enum { EL_LF, EL_CR, EL_CRLF, EL_ANY, EL_ANYCRLF };
105    
106    
107    
108  /*************************************************  /*************************************************
109  *               Global variables                 *  *               Global variables                 *
110  *************************************************/  *************************************************/
111    
112    /* Jeffrey Friedl has some debugging requirements that are not part of the
113    regular code. */
114    
115    #ifdef JFRIEDL_DEBUG
116    static int S_arg = -1;
117    static unsigned int jfriedl_XR = 0; /* repeat regex attempt this many times */
118    static unsigned int jfriedl_XT = 0; /* replicate text this many times */
119    static const char *jfriedl_prefix = "";
120    static const char *jfriedl_postfix = "";
121    #endif
122    
123    static int  endlinetype;
124    
125    static char *colour_string = (char *)"1;31";
126    static char *colour_option = NULL;
127    static char *dee_option = NULL;
128    static char *DEE_option = NULL;
129    static char *newline = NULL;
130  static char *pattern_filename = NULL;  static char *pattern_filename = NULL;
131  static char *stdin_name = (char *)"(standard input)";  static char *stdin_name = (char *)"(standard input)";
132    static char *locale = NULL;
133    
134    static const unsigned char *pcretables = NULL;
135    
136  static int  pattern_count = 0;  static int  pattern_count = 0;
137  static pcre **pattern_list;  static pcre **pattern_list = NULL;
138  static pcre_extra **hints_list;  static pcre_extra **hints_list = NULL;
139    
140  static char *include_pattern = NULL;  static char *include_pattern = NULL;
141  static char *exclude_pattern = NULL;  static char *exclude_pattern = NULL;
142    static char *include_dir_pattern = NULL;
143    static char *exclude_dir_pattern = NULL;
144    
145  static pcre *include_compiled = NULL;  static pcre *include_compiled = NULL;
146  static pcre *exclude_compiled = NULL;  static pcre *exclude_compiled = NULL;
147    static pcre *include_dir_compiled = NULL;
148    static pcre *exclude_dir_compiled = NULL;
149    
150  static int after_context = 0;  static int after_context = 0;
151  static int before_context = 0;  static int before_context = 0;
152  static int both_context = 0;  static int both_context = 0;
153    static int dee_action = dee_READ;
154    static int DEE_action = DEE_READ;
155    static int error_count = 0;
156    static int filenames = FN_DEFAULT;
157    static int process_options = 0;
158    
159  static BOOL count_only = FALSE;  static BOOL count_only = FALSE;
160  static BOOL filenames = TRUE;  static BOOL do_colour = FALSE;
161  static BOOL filenames_only = FALSE;  static BOOL file_offsets = FALSE;
 static BOOL filenames_nomatch_only = FALSE;  
162  static BOOL hyphenpending = FALSE;  static BOOL hyphenpending = FALSE;
163  static BOOL invert = FALSE;  static BOOL invert = FALSE;
164    static BOOL line_offsets = FALSE;
165  static BOOL multiline = FALSE;  static BOOL multiline = FALSE;
166  static BOOL number = FALSE;  static BOOL number = FALSE;
167    static BOOL only_matching = FALSE;
168  static BOOL quiet = FALSE;  static BOOL quiet = FALSE;
 static BOOL recurse = FALSE;  
169  static BOOL silent = FALSE;  static BOOL silent = FALSE;
170  static BOOL whole_lines = FALSE;  static BOOL utf8 = FALSE;
 static BOOL word_match = FALSE;  
171    
172  /* Structure for options and list of them */  /* Structure for options and list of them */
173    
174  enum { OP_NODATA, OP_STRING, OP_NUMBER };  enum { OP_NODATA, OP_STRING, OP_OP_STRING, OP_NUMBER, OP_OP_NUMBER,
175           OP_PATLIST };
176    
177  typedef struct option_item {  typedef struct option_item {
178    int type;    int type;
# Line 112  typedef struct option_item { Line 182  typedef struct option_item {
182    const char *help_text;    const char *help_text;
183  } option_item;  } option_item;
184    
185    /* Options without a single-letter equivalent get a negative value. This can be
186    used to identify them. */
187    
188    #define N_COLOUR       (-1)
189    #define N_EXCLUDE      (-2)
190    #define N_EXCLUDE_DIR  (-3)
191    #define N_HELP         (-4)
192    #define N_INCLUDE      (-5)
193    #define N_INCLUDE_DIR  (-6)
194    #define N_LABEL        (-7)
195    #define N_LOCALE       (-8)
196    #define N_NULL         (-9)
197    #define N_LOFFSETS     (-10)
198    #define N_FOFFSETS     (-11)
199    
200  static option_item optionlist[] = {  static option_item optionlist[] = {
201    { OP_NODATA, -1,  NULL,              "",              "  terminate options" },    { OP_NODATA,    N_NULL,   NULL,              "",              "  terminate options" },
202    { OP_NODATA, -1,  NULL,              "help",          "display this help and exit" },    { OP_NODATA,    N_HELP,   NULL,              "help",          "display this help and exit" },
203    { 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" },
204    { 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" },
205    { 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" },
206    { 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" },
207    { 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" },
208    { 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" },
209    { OP_NODATA, 'i', NULL,              "ignore-case",   "ignore case distinctions" },    { OP_STRING,    'D',      &DEE_option,       "devices=action","how to handle devices, FIFOs, and sockets" },
210    { OP_NODATA, 'l', NULL,              "files-with-matches", "print only FILE names containing matches" },    { OP_STRING,    'd',      &dee_option,       "directories=action", "how to handle directories" },
211    { OP_NODATA, 'L', NULL,              "files-without-match","print only FILE names not containing matches" },    { OP_PATLIST,   'e',      NULL,              "regex(p)",      "specify pattern (may be used more than once)" },
212    { 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" },
213    { OP_NODATA, 'M', NULL,              "multiline",     "run in multiline mode" },    { OP_STRING,    'f',      &pattern_filename, "file=path",     "read patterns from file" },
214    { OP_NODATA, 'n', NULL,              "line-number",   "print line number with output lines" },    { OP_NODATA,    N_FOFFSETS, NULL,            "file-offsets",  "output file offsets, not text" },
215    { OP_NODATA, 'q', NULL,              "quiet",         "suppress output, just set return code" },    { OP_NODATA,    'H',      NULL,              "with-filename", "force the prefixing filename on output" },
216    { OP_NODATA, 'r', NULL,              "recursive",     "recursively scan sub-directories" },    { OP_NODATA,    'h',      NULL,              "no-filename",   "suppress the prefixing filename on output" },
217    { OP_STRING, -1,  &exclude_pattern,  "exclude=pattern","exclude matching files when recursing" },    { OP_NODATA,    'i',      NULL,              "ignore-case",   "ignore case distinctions" },
218    { 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" },
219    { OP_NODATA, 's', NULL,              "no-messages",   "suppress error messages" },    { OP_NODATA,    'L',      NULL,              "files-without-match","print only FILE names not containing matches" },
220    { OP_NODATA, 'u', NULL,              "utf-8",         "use UTF-8 mode" },    { OP_STRING,    N_LABEL,  &stdin_name,       "label=name",    "set name for standard input" },
221    { OP_NODATA, 'V', NULL,              "version",       "print version information and exit" },    { OP_NODATA,    N_LOFFSETS, NULL,            "line-offsets",  "output line numbers and offsets, not text" },
222    { OP_NODATA, 'v', NULL,              "invert-match",  "select non-matching lines" },    { OP_STRING,    N_LOCALE, &locale,           "locale=locale", "use the named locale" },
223    { OP_NODATA, 'w', NULL,              "word-regex(p)", "force PATTERN to match only as a word"  },    { OP_NODATA,    'M',      NULL,              "multiline",     "run in multiline mode" },
224    { 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)" },
225    { OP_NODATA, 0,   NULL,               NULL,            NULL }    { OP_NODATA,    'n',      NULL,              "line-number",   "print line number with output lines" },
226      { OP_NODATA,    'o',      NULL,              "only-matching", "show only the part of the line that matched" },
227      { OP_NODATA,    'q',      NULL,              "quiet",         "suppress output, just set return code" },
228      { OP_NODATA,    'r',      NULL,              "recursive",     "recursively scan sub-directories" },
229      { OP_STRING,    N_EXCLUDE,&exclude_pattern,  "exclude=pattern","exclude matching files when recursing" },
230      { OP_STRING,    N_INCLUDE,&include_pattern,  "include=pattern","include matching files when recursing" },
231      { OP_STRING,    N_EXCLUDE_DIR,&exclude_dir_pattern, "exclude_dir=pattern","exclude matching directories when recursing" },
232      { OP_STRING,    N_INCLUDE_DIR,&include_dir_pattern, "include_dir=pattern","include matching directories when recursing" },
233    #ifdef JFRIEDL_DEBUG
234      { OP_OP_NUMBER, 'S',      &S_arg,            "jeffS",         "replace matched (sub)string with X" },
235    #endif
236      { OP_NODATA,    's',      NULL,              "no-messages",   "suppress error messages" },
237      { OP_NODATA,    'u',      NULL,              "utf-8",         "use UTF-8 mode" },
238      { OP_NODATA,    'V',      NULL,              "version",       "print version information and exit" },
239      { OP_NODATA,    'v',      NULL,              "invert-match",  "select non-matching lines" },
240      { OP_NODATA,    'w',      NULL,              "word-regex(p)", "force patterns to match only as words"  },
241      { OP_NODATA,    'x',      NULL,              "line-regex(p)", "force patterns to match only whole lines" },
242      { OP_NODATA,    0,        NULL,               NULL,            NULL }
243  };  };
244    
245    /* Tables for prefixing and suffixing patterns, according to the -w, -x, and -F
246    options. These set the 1, 2, and 4 bits in process_options, respectively. Note
247    that the combination of -w and -x has the same effect as -x on its own, so we
248    can treat them as the same. */
249    
250    static const char *prefix[] = {
251      "", "\\b", "^(?:", "^(?:", "\\Q", "\\b\\Q", "^(?:\\Q", "^(?:\\Q" };
252    
253    static const char *suffix[] = {
254      "", "\\b", ")$",   ")$",   "\\E", "\\E\\b", "\\E)$",   "\\E)$" };
255    
256    /* UTF-8 tables - used only when the newline setting is "any". */
257    
258    const int utf8_table3[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};
259    
260    const char utf8_table4[] = {
261      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
262      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
263      2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
264      3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 };
265    
266    
267    
268  /*************************************************  /*************************************************
269  *       Functions for directory scanning         *  *            OS-specific functions               *
270  *************************************************/  *************************************************/
271    
272  /* These functions are defined so that they can be made system specific,  /* These functions are defined so that they can be made system specific,
273  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". */  
274    
275    
276  /************* Directory scanning in Unix ***********/  /************* Directory scanning in Unix ***********/
277    
278  #if IS_UNIX  #if defined HAVE_SYS_STAT_H && defined HAVE_DIRENT_H && defined HAVE_SYS_TYPES_H
279  #include <sys/types.h>  #include <sys/types.h>
280  #include <sys/stat.h>  #include <sys/stat.h>
281  #include <dirent.h>  #include <dirent.h>
# Line 184  for (;;) Line 307  for (;;)
307    if (strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0)    if (strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0)
308      return dent->d_name;      return dent->d_name;
309    }    }
310  return NULL;   /* Keep compiler happy; never executed */  /* Control never reaches here */
311  }  }
312    
313  static void  static void
# Line 194  closedir(dir); Line 317  closedir(dir);
317  }  }
318    
319    
320    /************* Test for regular file in Unix **********/
321    
322    static int
323    isregfile(char *filename)
324    {
325    struct stat statbuf;
326    if (stat(filename, &statbuf) < 0)
327      return 1;        /* In the expectation that opening as a file will fail */
328    return (statbuf.st_mode & S_IFMT) == S_IFREG;
329    }
330    
331    
332    /************* Test stdout for being a terminal in Unix **********/
333    
334    static BOOL
335    is_stdout_tty(void)
336    {
337    return isatty(fileno(stdout));
338    }
339    
340    
341  /************* Directory scanning in Win32 ***********/  /************* Directory scanning in Win32 ***********/
342    
343  /* 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
344  Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES  Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES
345  when it did not exist. */  when it did not exist. David Byron added a patch that moved the #include of
346    <windows.h> to before the INVALID_FILE_ATTRIBUTES definition rather than after.
347    */
348    
349  #elif HAVE_WIN32API  #elif HAVE_WINDOWS_H
350    
351  #ifndef STRICT  #ifndef STRICT
352  # define STRICT  # define STRICT
# Line 209  when it did not exist. */ Line 354  when it did not exist. */
354  #ifndef WIN32_LEAN_AND_MEAN  #ifndef WIN32_LEAN_AND_MEAN
355  # define WIN32_LEAN_AND_MEAN  # define WIN32_LEAN_AND_MEAN
356  #endif  #endif
357    
358    #include <windows.h>
359    
360  #ifndef INVALID_FILE_ATTRIBUTES  #ifndef INVALID_FILE_ATTRIBUTES
361  #define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF  #define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF
362  #endif  #endif
363    
 #include <windows.h>  
   
364  typedef struct directory_type  typedef struct directory_type
365  {  {
366  HANDLE handle;  HANDLE handle;
# Line 292  free(dir); Line 438  free(dir);
438  }  }
439    
440    
441    /************* Test for regular file in Win32 **********/
442    
443    /* I don't know how to do this, or if it can be done; assume all paths are
444    regular if they are not directories. */
445    
446    int isregfile(char *filename)
447    {
448    return !isdirectory(filename);
449    }
450    
451    
452    /************* Test stdout for being a terminal in Win32 **********/
453    
454    /* I don't know how to do this; assume never */
455    
456    static BOOL
457    is_stdout_tty(void)
458    {
459    return FALSE;
460    }
461    
462    
463  /************* Directory scanning when we can't do it ***********/  /************* Directory scanning when we can't do it ***********/
464    
465  /* 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 468  free(dir);
468    
469  typedef void directory_type;  typedef void directory_type;
470    
471  int isdirectory(char *filename) { return FALSE; }  int isdirectory(char *filename) { return 0; }
472  directory_type * opendirectory(char *filename) {}  directory_type * opendirectory(char *filename) { return (directory_type*)0;}
473  char *readdirectory(directory_type *dir) {}  char *readdirectory(directory_type *dir) { return (char*)0;}
474  void closedirectory(directory_type *dir) {}  void closedirectory(directory_type *dir) {}
475    
476    
477    /************* Test for regular when we can't do it **********/
478    
479    /* Assume all files are regular. */
480    
481    int isregfile(char *filename) { return 1; }
482    
483    
484    /************* Test stdout for being a terminal when we can't do it **********/
485    
486    static BOOL
487    is_stdout_tty(void)
488    {
489    return FALSE;
490    }
491    
492    
493  #endif  #endif
494    
495    
496    
497  #if ! HAVE_STRERROR  #ifndef HAVE_STRERROR
498  /*************************************************  /*************************************************
499  *     Provide strerror() for non-ANSI libraries  *  *     Provide strerror() for non-ANSI libraries  *
500  *************************************************/  *************************************************/
# Line 332  return sys_errlist[n]; Line 517  return sys_errlist[n];
517    
518    
519  /*************************************************  /*************************************************
520    *             Find end of line                   *
521    *************************************************/
522    
523    /* The length of the endline sequence that is found is set via lenptr. This may
524    be zero at the very end of the file if there is no line-ending sequence there.
525    
526    Arguments:
527      p         current position in line
528      endptr    end of available data
529      lenptr    where to put the length of the eol sequence
530    
531    Returns:    pointer to the last byte of the line
532    */
533    
534    static char *
535    end_of_line(char *p, char *endptr, int *lenptr)
536    {
537    switch(endlinetype)
538      {
539      default:      /* Just in case */
540      case EL_LF:
541      while (p < endptr && *p != '\n') p++;
542      if (p < endptr)
543        {
544        *lenptr = 1;
545        return p + 1;
546        }
547      *lenptr = 0;
548      return endptr;
549    
550      case EL_CR:
551      while (p < endptr && *p != '\r') p++;
552      if (p < endptr)
553        {
554        *lenptr = 1;
555        return p + 1;
556        }
557      *lenptr = 0;
558      return endptr;
559    
560      case EL_CRLF:
561      for (;;)
562        {
563        while (p < endptr && *p != '\r') p++;
564        if (++p >= endptr)
565          {
566          *lenptr = 0;
567          return endptr;
568          }
569        if (*p == '\n')
570          {
571          *lenptr = 2;
572          return p + 1;
573          }
574        }
575      break;
576    
577      case EL_ANYCRLF:
578      while (p < endptr)
579        {
580        int extra = 0;
581        register int c = *((unsigned char *)p);
582    
583        if (utf8 && c >= 0xc0)
584          {
585          int gcii, gcss;
586          extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
587          gcss = 6*extra;
588          c = (c & utf8_table3[extra]) << gcss;
589          for (gcii = 1; gcii <= extra; gcii++)
590            {
591            gcss -= 6;
592            c |= (p[gcii] & 0x3f) << gcss;
593            }
594          }
595    
596        p += 1 + extra;
597    
598        switch (c)
599          {
600          case 0x0a:    /* LF */
601          *lenptr = 1;
602          return p;
603    
604          case 0x0d:    /* CR */
605          if (p < endptr && *p == 0x0a)
606            {
607            *lenptr = 2;
608            p++;
609            }
610          else *lenptr = 1;
611          return p;
612    
613          default:
614          break;
615          }
616        }   /* End of loop for ANYCRLF case */
617    
618      *lenptr = 0;  /* Must have hit the end */
619      return endptr;
620    
621      case EL_ANY:
622      while (p < endptr)
623        {
624        int extra = 0;
625        register int c = *((unsigned char *)p);
626    
627        if (utf8 && c >= 0xc0)
628          {
629          int gcii, gcss;
630          extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
631          gcss = 6*extra;
632          c = (c & utf8_table3[extra]) << gcss;
633          for (gcii = 1; gcii <= extra; gcii++)
634            {
635            gcss -= 6;
636            c |= (p[gcii] & 0x3f) << gcss;
637            }
638          }
639    
640        p += 1 + extra;
641    
642        switch (c)
643          {
644          case 0x0a:    /* LF */
645          case 0x0b:    /* VT */
646          case 0x0c:    /* FF */
647          *lenptr = 1;
648          return p;
649    
650          case 0x0d:    /* CR */
651          if (p < endptr && *p == 0x0a)
652            {
653            *lenptr = 2;
654            p++;
655            }
656          else *lenptr = 1;
657          return p;
658    
659          case 0x85:    /* NEL */
660          *lenptr = utf8? 2 : 1;
661          return p;
662    
663          case 0x2028:  /* LS */
664          case 0x2029:  /* PS */
665          *lenptr = 3;
666          return p;
667    
668          default:
669          break;
670          }
671        }   /* End of loop for ANY case */
672    
673      *lenptr = 0;  /* Must have hit the end */
674      return endptr;
675      }     /* End of overall switch */
676    }
677    
678    
679    
680    /*************************************************
681    *         Find start of previous line            *
682    *************************************************/
683    
684    /* This is called when looking back for before lines to print.
685    
686    Arguments:
687      p         start of the subsequent line
688      startptr  start of available data
689    
690    Returns:    pointer to the start of the previous line
691    */
692    
693    static char *
694    previous_line(char *p, char *startptr)
695    {
696    switch(endlinetype)
697      {
698      default:      /* Just in case */
699      case EL_LF:
700      p--;
701      while (p > startptr && p[-1] != '\n') p--;
702      return p;
703    
704      case EL_CR:
705      p--;
706      while (p > startptr && p[-1] != '\n') p--;
707      return p;
708    
709      case EL_CRLF:
710      for (;;)
711        {
712        p -= 2;
713        while (p > startptr && p[-1] != '\n') p--;
714        if (p <= startptr + 1 || p[-2] == '\r') return p;
715        }
716      return p;   /* But control should never get here */
717    
718      case EL_ANY:
719      case EL_ANYCRLF:
720      if (*(--p) == '\n' && p > startptr && p[-1] == '\r') p--;
721      if (utf8) while ((*p & 0xc0) == 0x80) p--;
722    
723      while (p > startptr)
724        {
725        register int c;
726        char *pp = p - 1;
727    
728        if (utf8)
729          {
730          int extra = 0;
731          while ((*pp & 0xc0) == 0x80) pp--;
732          c = *((unsigned char *)pp);
733          if (c >= 0xc0)
734            {
735            int gcii, gcss;
736            extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
737            gcss = 6*extra;
738            c = (c & utf8_table3[extra]) << gcss;
739            for (gcii = 1; gcii <= extra; gcii++)
740              {
741              gcss -= 6;
742              c |= (pp[gcii] & 0x3f) << gcss;
743              }
744            }
745          }
746        else c = *((unsigned char *)pp);
747    
748        if (endlinetype == EL_ANYCRLF) switch (c)
749          {
750          case 0x0a:    /* LF */
751          case 0x0d:    /* CR */
752          return p;
753    
754          default:
755          break;
756          }
757    
758        else switch (c)
759          {
760          case 0x0a:    /* LF */
761          case 0x0b:    /* VT */
762          case 0x0c:    /* FF */
763          case 0x0d:    /* CR */
764          case 0x85:    /* NEL */
765          case 0x2028:  /* LS */
766          case 0x2029:  /* PS */
767          return p;
768    
769          default:
770          break;
771          }
772    
773        p = pp;  /* Back one character */
774        }        /* End of loop for ANY case */
775    
776      return startptr;  /* Hit start of data */
777      }     /* End of overall switch */
778    }
779    
780    
781    
782    
783    
784    /*************************************************
785  *       Print the previous "after" lines         *  *       Print the previous "after" lines         *
786  *************************************************/  *************************************************/
787    
788  /* 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,
789  and at the end of the file.  and at the end of the file. The data in the line is written using fwrite() so
790    that a binary zero does not terminate it.
791    
792  Arguments:  Arguments:
793    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 806  if (after_context > 0 && lastmatchnumber
806    int count = 0;    int count = 0;
807    while (lastmatchrestart < endptr && count++ < after_context)    while (lastmatchrestart < endptr && count++ < after_context)
808      {      {
809        int ellength;
810      char *pp = lastmatchrestart;      char *pp = lastmatchrestart;
811      if (printname != NULL) fprintf(stdout, "%s-", printname);      if (printname != NULL) fprintf(stdout, "%s-", printname);
812      if (number) fprintf(stdout, "%d-", lastmatchnumber++);      if (number) fprintf(stdout, "%d-", lastmatchnumber++);
813      while (*pp != '\n') pp++;      pp = end_of_line(pp, endptr, &ellength);
814      fprintf(stdout, "%.*s", pp - lastmatchrestart + 1, lastmatchrestart);      fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout);
815      lastmatchrestart = pp + 1;      lastmatchrestart = pp;
816      }      }
817    hyphenpending = TRUE;    hyphenpending = TRUE;
818    }    }
# Line 380  be in the middle third most of the time, Line 832  be in the middle third most of the time,
832  "before" context printing.  "before" context printing.
833    
834  Arguments:  Arguments:
835    in           the fopened FILE stream    handle       the fopened FILE stream for a normal file
836                   the gzFile pointer when reading is via libz
837                   the BZFILE pointer when reading is via libbz2
838      frtype       FR_PLAIN, FR_LIBZ, or FR_LIBBZ2
839    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
840                 or NULL if the file name is not to be printed                 or NULL if the file name is not to be printed
841                 it cannot be NULL if filenames[_nomatch]_only is set                 it cannot be NULL if filenames[_nomatch]_only is set
842    
843  Returns:       0 if there was at least one match  Returns:       0 if there was at least one match
844                 1 otherwise (no matches)                 1 otherwise (no matches)
845                   2 if there is a read error on a .bz2 file
846  */  */
847    
848  static int  static int
849  pcregrep(FILE *in, char *printname)  pcregrep(void *handle, int frtype, char *printname)
850  {  {
851  int rc = 1;  int rc = 1;
852  int linenumber = 1;  int linenumber = 1;
853  int lastmatchnumber = 0;  int lastmatchnumber = 0;
854  int count = 0;  int count = 0;
855    int filepos = 0;
856  int offsets[99];  int offsets[99];
857  char *lastmatchrestart = NULL;  char *lastmatchrestart = NULL;
858  char buffer[3*MBUFTHIRD];  char buffer[3*MBUFTHIRD];
# Line 403  char *ptr = buffer; Line 860  char *ptr = buffer;
860  char *endptr;  char *endptr;
861  size_t bufflength;  size_t bufflength;
862  BOOL endhyphenpending = FALSE;  BOOL endhyphenpending = FALSE;
863    FILE *in = NULL;                    /* Ensure initialized */
864    
865  /* Do the first read into the start of the buffer and set up the pointer to  #ifdef SUPPORT_LIBZ
866  end of what we have. */  gzFile ingz = NULL;
867    #endif
868    
869    #ifdef SUPPORT_LIBBZ2
870    BZFILE *inbz2 = NULL;
871    #endif
872    
873    
874    /* Do the first read into the start of the buffer and set up the pointer to end
875    of what we have. In the case of libz, a non-zipped .gz file will be read as a
876    plain file. However, if a .bz2 file isn't actually bzipped, the first read will
877    fail. */
878    
879    #ifdef SUPPORT_LIBZ
880    if (frtype == FR_LIBZ)
881      {
882      ingz = (gzFile)handle;
883      bufflength = gzread (ingz, buffer, 3*MBUFTHIRD);
884      }
885    else
886    #endif
887    
888    #ifdef SUPPORT_LIBBZ2
889    if (frtype == FR_LIBBZ2)
890      {
891      inbz2 = (BZFILE *)handle;
892      bufflength = BZ2_bzread(inbz2, buffer, 3*MBUFTHIRD);
893      if ((int)bufflength < 0) return 2;   /* Gotcha: bufflength is size_t; */
894      }                                    /* without the cast it is unsigned. */
895    else
896    #endif
897    
898      {
899      in = (FILE *)handle;
900      bufflength = fread(buffer, 1, 3*MBUFTHIRD, in);
901      }
902    
 bufflength = fread(buffer, 1, 3*MBUFTHIRD, in);  
903  endptr = buffer + bufflength;  endptr = buffer + bufflength;
904    
905  /* 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 909  way, the buffer is shifted left and re-f
909    
910  while (ptr < endptr)  while (ptr < endptr)
911    {    {
912    int i;    int i, endlinelength;
913      int mrc = 0;
914    BOOL match = FALSE;    BOOL match = FALSE;
915      char *matchptr = ptr;
916    char *t = ptr;    char *t = ptr;
917    size_t length, linelength;    size_t length, linelength;
918    
# Line 429  while (ptr < endptr) Line 923  while (ptr < endptr)
923    line. In multiline mode the PCRE_FIRSTLINE option is used for compiling, so    line. In multiline mode the PCRE_FIRSTLINE option is used for compiling, so
924    that any match is constrained to be in the first line. */    that any match is constrained to be in the first line. */
925    
926    linelength = 0;    t = end_of_line(t, endptr, &endlinelength);
927    while (t < endptr && *t++ != '\n') linelength++;    linelength = t - ptr - endlinelength;
928    length = multiline? endptr - ptr : linelength;    length = multiline? (size_t)(endptr - ptr) : linelength;
929    
930      /* Extra processing for Jeffrey Friedl's debugging. */
931    
932    #ifdef JFRIEDL_DEBUG
933      if (jfriedl_XT || jfriedl_XR)
934      {
935          #include <sys/time.h>
936          #include <time.h>
937          struct timeval start_time, end_time;
938          struct timezone dummy;
939    
940          if (jfriedl_XT)
941          {
942              unsigned long newlen = length * jfriedl_XT + strlen(jfriedl_prefix) + strlen(jfriedl_postfix);
943              const char *orig = ptr;
944              ptr = malloc(newlen + 1);
945              if (!ptr) {
946                      printf("out of memory");
947                      exit(2);
948              }
949              endptr = ptr;
950              strcpy(endptr, jfriedl_prefix); endptr += strlen(jfriedl_prefix);
951              for (i = 0; i < jfriedl_XT; i++) {
952                      strncpy(endptr, orig,  length);
953                      endptr += length;
954              }
955              strcpy(endptr, jfriedl_postfix); endptr += strlen(jfriedl_postfix);
956              length = newlen;
957          }
958    
959          if (gettimeofday(&start_time, &dummy) != 0)
960                  perror("bad gettimeofday");
961    
962    
963          for (i = 0; i < jfriedl_XR; i++)
964              match = (pcre_exec(pattern_list[0], hints_list[0], ptr, length, 0, 0, offsets, 99) >= 0);
965    
966          if (gettimeofday(&end_time, &dummy) != 0)
967                  perror("bad gettimeofday");
968    
969          double delta = ((end_time.tv_sec + (end_time.tv_usec / 1000000.0))
970                          -
971                          (start_time.tv_sec + (start_time.tv_usec / 1000000.0)));
972    
973          printf("%s TIMER[%.4f]\n", match ? "MATCH" : "FAIL", delta);
974          return 0;
975      }
976    #endif
977    
978      /* We come back here after a match when the -o option (only_matching) is set,
979      in order to find any further matches in the same line. */
980    
981      ONLY_MATCHING_RESTART:
982    
983    /* Run through all the patterns until one matches. Note that we don't include    /* Run through all the patterns until one matches. Note that we don't include
984    the final newline in the subject string. */    the final newline in the subject string. */
985    
986    for (i = 0; !match && i < pattern_count; i++)    for (i = 0; i < pattern_count; i++)
987      {      {
988      match = pcre_exec(pattern_list[i], hints_list[i], ptr, length, 0, 0,      mrc = pcre_exec(pattern_list[i], hints_list[i], matchptr, length, 0, 0,
989        offsets, 99) >= 0;        offsets, 99);
990        if (mrc >= 0) { match = TRUE; break; }
991        if (mrc != PCRE_ERROR_NOMATCH)
992          {
993          fprintf(stderr, "pcregrep: pcre_exec() error %d while matching ", mrc);
994          if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1);
995          fprintf(stderr, "this line:\n");
996          fwrite(matchptr, 1, linelength, stderr);  /* In case binary zero included */
997          fprintf(stderr, "\n");
998          if (error_count == 0 &&
999              (mrc == PCRE_ERROR_MATCHLIMIT || mrc == PCRE_ERROR_RECURSIONLIMIT))
1000            {
1001            fprintf(stderr, "pcregrep: error %d means that a resource limit "
1002              "was exceeded\n", mrc);
1003            fprintf(stderr, "pcregrep: check your regex for nested unlimited loops\n");
1004            }
1005          if (error_count++ > 20)
1006            {
1007            fprintf(stderr, "pcregrep: too many errors - abandoned\n");
1008            exit(2);
1009            }
1010          match = invert;    /* No more matching; don't show the line again */
1011          break;
1012          }
1013      }      }
1014    
1015    /* If it's a match or a not-match (as required), print what's wanted. */    /* If it's a match or a not-match (as required), do what's wanted. */
1016    
1017    if (match != invert)    if (match != invert)
1018      {      {
1019      BOOL hyphenprinted = FALSE;      BOOL hyphenprinted = FALSE;
1020    
1021      if (filenames_nomatch_only) return 1;      /* We've failed if we want a file that doesn't have any matches. */
1022    
1023        if (filenames == FN_NOMATCH_ONLY) return 1;
1024    
1025        /* Just count if just counting is wanted. */
1026    
1027      if (count_only) count++;      if (count_only) count++;
1028    
1029      else if (filenames_only)      /* If all we want is a file name, there is no need to scan any more lines
1030        in the file. */
1031    
1032        else if (filenames == FN_ONLY)
1033        {        {
1034        fprintf(stdout, "%s\n", printname);        fprintf(stdout, "%s\n", printname);
1035        return 0;        return 0;
1036        }        }
1037    
1038        /* Likewise, if all we want is a yes/no answer. */
1039    
1040      else if (quiet) return 0;      else if (quiet) return 0;
1041    
1042        /* The --only-matching option prints just the substring that matched, and
1043        the --file-offsets and --line-offsets options output offsets for the
1044        matching substring (they both force --only-matching). None of these options
1045        prints any context. Afterwards, adjust the start and length, and then jump
1046        back to look for further matches in the same line. If we are in invert
1047        mode, however, nothing is printed - this could be still useful because the
1048        return code is set. */
1049    
1050        else if (only_matching)
1051          {
1052          if (!invert)
1053            {
1054            if (printname != NULL) fprintf(stdout, "%s:", printname);
1055            if (number) fprintf(stdout, "%d:", linenumber);
1056            if (line_offsets)
1057              fprintf(stdout, "%d,%d", matchptr + offsets[0] - ptr,
1058                offsets[1] - offsets[0]);
1059            else if (file_offsets)
1060              fprintf(stdout, "%d,%d", filepos + matchptr + offsets[0] - ptr,
1061                offsets[1] - offsets[0]);
1062            else
1063              fwrite(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
1064            fprintf(stdout, "\n");
1065            matchptr += offsets[1];
1066            length -= offsets[1];
1067            match = FALSE;
1068            goto ONLY_MATCHING_RESTART;
1069            }
1070          }
1071    
1072        /* This is the default case when none of the above options is set. We print
1073        the matching lines(s), possibly preceded and/or followed by other lines of
1074        context. */
1075    
1076      else      else
1077        {        {
1078        /* 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 1080  while (ptr < endptr)
1080    
1081        if (after_context > 0 && lastmatchnumber > 0)        if (after_context > 0 && lastmatchnumber > 0)
1082          {          {
1083            int ellength;
1084          int linecount = 0;          int linecount = 0;
1085          char *p = lastmatchrestart;          char *p = lastmatchrestart;
1086    
1087          while (p < ptr && linecount < after_context)          while (p < ptr && linecount < after_context)
1088            {            {
1089            while (*p != '\n') p++;            p = end_of_line(p, ptr, &ellength);
           p++;  
1090            linecount++;            linecount++;
1091            }            }
1092    
1093          /* It is important to advance lastmatchrestart during this printing so          /* It is important to advance lastmatchrestart during this printing so
1094          that it interacts correctly with any "before" printing below. */          that it interacts correctly with any "before" printing below. Print
1095            each line's data using fwrite() in case there are binary zeroes. */
1096    
1097          while (lastmatchrestart < p)          while (lastmatchrestart < p)
1098            {            {
1099            char *pp = lastmatchrestart;            char *pp = lastmatchrestart;
1100            if (printname != NULL) fprintf(stdout, "%s-", printname);            if (printname != NULL) fprintf(stdout, "%s-", printname);
1101            if (number) fprintf(stdout, "%d-", lastmatchnumber++);            if (number) fprintf(stdout, "%d-", lastmatchnumber++);
1102            while (*pp != '\n') pp++;            pp = end_of_line(pp, endptr, &ellength);
1103            fprintf(stdout, "%.*s", pp - lastmatchrestart + 1, lastmatchrestart);            fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout);
1104            lastmatchrestart = pp + 1;            lastmatchrestart = pp;
1105            }            }
1106          if (lastmatchrestart != ptr) hyphenpending = TRUE;          if (lastmatchrestart != ptr) hyphenpending = TRUE;
1107          }          }
# Line 510  while (ptr < endptr) Line 1124  while (ptr < endptr)
1124          char *p = ptr;          char *p = ptr;
1125    
1126          while (p > buffer && (lastmatchnumber == 0 || p > lastmatchrestart) &&          while (p > buffer && (lastmatchnumber == 0 || p > lastmatchrestart) &&
1127                 linecount++ < before_context)                 linecount < before_context)
1128            {            {
1129            p--;            linecount++;
1130            while (p > buffer && p[-1] != '\n') p--;            p = previous_line(p, buffer);
1131            }            }
1132    
1133          if (lastmatchnumber > 0 && p > lastmatchrestart && !hyphenprinted)          if (lastmatchnumber > 0 && p > lastmatchrestart && !hyphenprinted)
# Line 521  while (ptr < endptr) Line 1135  while (ptr < endptr)
1135    
1136          while (p < ptr)          while (p < ptr)
1137            {            {
1138              int ellength;
1139            char *pp = p;            char *pp = p;
1140            if (printname != NULL) fprintf(stdout, "%s-", printname);            if (printname != NULL) fprintf(stdout, "%s-", printname);
1141            if (number) fprintf(stdout, "%d-", linenumber - linecount--);            if (number) fprintf(stdout, "%d-", linenumber - linecount--);
1142            while (*pp != '\n') pp++;            pp = end_of_line(pp, endptr, &ellength);
1143            fprintf(stdout, "%.*s", pp - p + 1, p);            fwrite(p, 1, pp - p, stdout);
1144            p = pp + 1;            p = pp;
1145            }            }
1146          }          }
1147    
# Line 541  while (ptr < endptr) Line 1156  while (ptr < endptr)
1156    
1157        /* 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
1158        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
1159        line number appropriately. Because the PCRE_FIRSTLINE option is set, the        line number appropriately, but only when there actually was a match
1160        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
1161          the match will always be before the first newline sequence. */
1162    
1163        if (multiline)        if (multiline)
1164          {          {
1165          char *endmatch = ptr + offsets[1];          int ellength;
1166          t = ptr;          char *endmatch = ptr;
1167          while (t < endmatch) { if (*t++ == '\n') linenumber++; }          if (!invert)
1168          while (endmatch < endptr && *endmatch != '\n') endmatch++;            {
1169          linelength = endmatch - ptr;            endmatch += offsets[1];
1170              t = ptr;
1171              while (t < endmatch)
1172                {
1173                t = end_of_line(t, endptr, &ellength);
1174                if (t <= endmatch) linenumber++; else break;
1175                }
1176              }
1177            endmatch = end_of_line(endmatch, endptr, &ellength);
1178            linelength = endmatch - ptr - ellength;
1179            }
1180    
1181          /*** NOTE: Use only fwrite() to output the data line, so that binary
1182          zeroes are treated as just another data character. */
1183    
1184          /* This extra option, for Jeffrey Friedl's debugging requirements,
1185          replaces the matched string, or a specific captured string if it exists,
1186          with X. When this happens, colouring is ignored. */
1187    
1188    #ifdef JFRIEDL_DEBUG
1189          if (S_arg >= 0 && S_arg < mrc)
1190            {
1191            int first = S_arg * 2;
1192            int last  = first + 1;
1193            fwrite(ptr, 1, offsets[first], stdout);
1194            fprintf(stdout, "X");
1195            fwrite(ptr + offsets[last], 1, linelength - offsets[last], stdout);
1196          }          }
1197          else
1198    #endif
1199    
1200          /* We have to split the line(s) up if colouring. */
1201    
1202        fprintf(stdout, "%.*s\n", linelength, ptr);        if (do_colour)
1203            {
1204            fwrite(ptr, 1, offsets[0], stdout);
1205            fprintf(stdout, "%c[%sm", 0x1b, colour_string);
1206            fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
1207            fprintf(stdout, "%c[00m", 0x1b);
1208            fwrite(ptr + offsets[1], 1, (linelength + endlinelength) - offsets[1],
1209              stdout);
1210            }
1211          else fwrite(ptr, 1, linelength + endlinelength, stdout);
1212        }        }
1213    
1214        /* End of doing what has to be done for a match */
1215    
1216      rc = 0;    /* Had some success */      rc = 0;    /* Had some success */
1217    
1218      /* Remember where the last match happened for after_context. We remember      /* Remember where the last match happened for after_context. We remember
1219      where we are about to restart, and that line's number. */      where we are about to restart, and that line's number. */
1220    
1221      lastmatchrestart = ptr + linelength + 1;      lastmatchrestart = ptr + linelength + endlinelength;
1222      lastmatchnumber = linenumber + 1;      lastmatchnumber = linenumber + 1;
1223      }      }
1224    
1225    /* Advance to after the newline and increment the line number. */    /* For a match in multiline inverted mode (which of course did not cause
1226      anything to be printed), we have to move on to the end of the match before
1227      proceeding. */
1228    
1229      if (multiline && invert && match)
1230        {
1231        int ellength;
1232        char *endmatch = ptr + offsets[1];
1233        t = ptr;
1234        while (t < endmatch)
1235          {
1236          t = end_of_line(t, endptr, &ellength);
1237          if (t <= endmatch) linenumber++; else break;
1238          }
1239        endmatch = end_of_line(endmatch, endptr, &ellength);
1240        linelength = endmatch - ptr - ellength;
1241        }
1242    
1243      /* Advance to after the newline and increment the line number. The file
1244      offset to the current line is maintained in filepos. */
1245    
1246    ptr += linelength + 1;    ptr += linelength + endlinelength;
1247      filepos += linelength + endlinelength;
1248    linenumber++;    linenumber++;
1249    
1250    /* 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 589  while (ptr < endptr) Line 1266  while (ptr < endptr)
1266    
1267      memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD);      memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD);
1268      ptr -= MBUFTHIRD;      ptr -= MBUFTHIRD;
     bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in);  
     endptr = buffer + bufflength;  
1269    
1270      /* Adjust any last match point */  #ifdef SUPPORT_LIBZ
1271        if (frtype == FR_LIBZ)
1272          bufflength = 2*MBUFTHIRD +
1273            gzread (ingz, buffer + 2*MBUFTHIRD, MBUFTHIRD);
1274        else
1275    #endif
1276    
1277      if (lastmatchnumber > 0) lastmatchrestart -= MBUFTHIRD;  #ifdef SUPPORT_LIBBZ2
1278      }      if (frtype == FR_LIBBZ2)
1279          bufflength = 2*MBUFTHIRD +
1280            BZ2_bzread(inbz2, buffer + 2*MBUFTHIRD, MBUFTHIRD);
1281        else
1282    #endif
1283    
1284        bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in);
1285    
1286        endptr = buffer + bufflength;
1287    
1288        /* Adjust any last match point */
1289    
1290        if (lastmatchnumber > 0) lastmatchrestart -= MBUFTHIRD;
1291        }
1292    }     /* Loop through the whole file */    }     /* Loop through the whole file */
1293    
1294  /* 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
1295  hyphenpending if it prints something. */  hyphenpending if it prints something. */
1296    
1297  do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname);  if (!only_matching && !count_only)
1298  hyphenpending |= endhyphenpending;    {
1299      do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname);
1300      hyphenpending |= endhyphenpending;
1301      }
1302    
1303  /* 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
1304  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. */
1305    
1306  if (filenames_nomatch_only)  if (filenames == FN_NOMATCH_ONLY)
1307    {    {
1308    fprintf(stdout, "%s\n", printname);    fprintf(stdout, "%s\n", printname);
1309    return 0;    return 0;
# Line 635  recursing; if it's a file, grep it. Line 1331  recursing; if it's a file, grep it.
1331    
1332  Arguments:  Arguments:
1333    pathname          the path to investigate    pathname          the path to investigate
1334    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  
1335    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
1336    
1337  Returns:   0 if there was at least one match  Returns:   0 if there was at least one match
# Line 648  However, file opening failures are suppr Line 1342  However, file opening failures are suppr
1342  */  */
1343    
1344  static int  static int
1345  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)  
1346  {  {
1347  int rc = 1;  int rc = 1;
1348  int sep;  int sep;
1349  FILE *in;  int frtype;
1350  char *printname;  int pathlen;
1351    void *handle;
1352    FILE *in = NULL;           /* Ensure initialized */
1353    
1354    #ifdef SUPPORT_LIBZ
1355    gzFile ingz = NULL;
1356    #endif
1357    
1358    #ifdef SUPPORT_LIBBZ2
1359    BZFILE *inbz2 = NULL;
1360    #endif
1361    
1362  /* If the file name is "-" we scan stdin */  /* If the file name is "-" we scan stdin */
1363    
1364  if (strcmp(pathname, "-") == 0)  if (strcmp(pathname, "-") == 0)
1365    {    {
1366    return pcregrep(stdin,    return pcregrep(stdin, FR_PLAIN,
1367      (filenames_only || filenames_nomatch_only ||      (filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))?
     (show_filenames && !only_one_at_top))?  
1368        stdin_name : NULL);        stdin_name : NULL);
1369    }    }
1370    
1371  /* 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
1372  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
1373  localized so it can be made system-specific. */  that were set. The scanning code is localized so it can be made
1374    system-specific. */
1375    
1376  if ((sep = isdirectory(pathname)) != 0 && dir_recurse)  if ((sep = isdirectory(pathname)) != 0)
1377    {    {
1378    char buffer[1024];    if (dee_action == dee_SKIP) return 1;
1379    char *nextfile;    if (dee_action == dee_RECURSE)
1380    directory_type *dir = opendirectory(pathname);      {
1381        char buffer[1024];
1382        char *nextfile;
1383        directory_type *dir = opendirectory(pathname);
1384    
1385        if (dir == NULL)
1386          {
1387          if (!silent)
1388            fprintf(stderr, "pcregrep: Failed to open directory %s: %s\n", pathname,
1389              strerror(errno));
1390          return 2;
1391          }
1392    
1393        while ((nextfile = readdirectory(dir)) != NULL)
1394          {
1395          int frc, nflen;
1396          sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile);
1397          nflen = strlen(nextfile);
1398    
1399          if (isdirectory(buffer))
1400            {
1401            if (exclude_dir_compiled != NULL &&
1402                pcre_exec(exclude_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0)
1403              continue;
1404    
1405            if (include_dir_compiled != NULL &&
1406                pcre_exec(include_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0)
1407              continue;
1408            }
1409          else
1410            {
1411            if (exclude_compiled != NULL &&
1412                pcre_exec(exclude_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0)
1413              continue;
1414    
1415            if (include_compiled != NULL &&
1416                pcre_exec(include_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0)
1417              continue;
1418            }
1419    
1420          frc = grep_or_recurse(buffer, dir_recurse, FALSE);
1421          if (frc > 1) rc = frc;
1422           else if (frc == 0 && rc == 1) rc = 0;
1423          }
1424    
1425        closedirectory(dir);
1426        return rc;
1427        }
1428      }
1429    
1430    /* If the file is not a directory and not a regular file, skip it if that's
1431    been requested. */
1432    
1433    else if (!isregfile(pathname) && DEE_action == DEE_SKIP) return 1;
1434    
1435    /* Control reaches here if we have a regular file, or if we have a directory
1436    and recursion or skipping was not requested, or if we have anything else and
1437    skipping was not requested. The scan proceeds. If this is the first and only
1438    argument at top level, we don't show the file name, unless we are only showing
1439    the file name, or the filename was forced (-H). */
1440    
1441    if (dir == NULL)  pathlen = strlen(pathname);
1442    
1443    /* Open using zlib if it is supported and the file name ends with .gz. */
1444    
1445    #ifdef SUPPORT_LIBZ
1446    if (pathlen > 3 && strcmp(pathname + pathlen - 3, ".gz") == 0)
1447      {
1448      ingz = gzopen(pathname, "rb");
1449      if (ingz == NULL)
1450      {      {
1451      if (!silent)      if (!silent)
1452        fprintf(stderr, "pcregrep: Failed to open directory %s: %s\n", pathname,        fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
1453          strerror(errno));          strerror(errno));
1454      return 2;      return 2;
1455      }      }
1456      handle = (void *)ingz;
1457      frtype = FR_LIBZ;
1458      }
1459    else
1460    #endif
1461    
1462    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;  
1463    
1464      if (include_compiled != NULL &&  #ifdef SUPPORT_LIBBZ2
1465          pcre_exec(include_compiled, NULL, buffer, blen, 0, 0, NULL, 0) < 0)  if (pathlen > 4 && strcmp(pathname + pathlen - 4, ".bz2") == 0)
1466        continue;    {
1467      inbz2 = BZ2_bzopen(pathname, "rb");
1468      handle = (void *)inbz2;
1469      frtype = FR_LIBBZ2;
1470      }
1471    else
1472    #endif
1473    
1474      frc = grep_or_recurse(buffer, dir_recurse, TRUE, FALSE);  /* Otherwise use plain fopen(). The label is so that we can come back here if
1475      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;  
     }  
1476    
1477    closedirectory(dir);  #ifdef SUPPORT_LIBBZ2
1478    return rc;  PLAIN_FILE:
1479    #endif
1480      {
1481      in = fopen(pathname, "r");
1482      handle = (void *)in;
1483      frtype = FR_PLAIN;
1484    }    }
1485    
1486  /* 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. */  
1487    
1488  in = fopen(pathname, "r");  if (handle == NULL)
 if (in == NULL)  
1489    {    {
1490    if (!silent)    if (!silent)
1491      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
# Line 721  if (in == NULL) Line 1493  if (in == NULL)
1493    return 2;    return 2;
1494    }    }
1495    
1496  printname =  (filenames_only || filenames_nomatch_only ||  /* Now grep the file */
   (show_filenames && !only_one_at_top))? pathname : NULL;  
1497    
1498  rc = pcregrep(in, printname);  rc = pcregrep(handle, frtype, (filenames > FN_DEFAULT ||
1499      (filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL);
1500    
1501    /* Close in an appropriate manner. */
1502    
1503    #ifdef SUPPORT_LIBZ
1504    if (frtype == FR_LIBZ)
1505      gzclose(ingz);
1506    else
1507    #endif
1508    
1509    /* If it is a .bz2 file and the result is 2, it means that the first attempt to
1510    read failed. If the error indicates that the file isn't in fact bzipped, try
1511    again as a normal file. */
1512    
1513    #ifdef SUPPORT_LIBBZ2
1514    if (frtype == FR_LIBBZ2)
1515      {
1516      if (rc == 2)
1517        {
1518        int errnum;
1519        const char *err = BZ2_bzerror(inbz2, &errnum);
1520        if (errnum == BZ_DATA_ERROR_MAGIC)
1521          {
1522          BZ2_bzclose(inbz2);
1523          goto PLAIN_FILE;
1524          }
1525        else if (!silent)
1526          fprintf(stderr, "pcregrep: Failed to read %s using bzlib: %s\n",
1527            pathname, err);
1528        }
1529      BZ2_bzclose(inbz2);
1530      }
1531    else
1532    #endif
1533    
1534    /* Normal file close */
1535    
1536  fclose(in);  fclose(in);
1537    
1538    /* Pass back the yield from pcregrep(). */
1539    
1540  return rc;  return rc;
1541  }  }
1542    
# Line 740  return rc; Line 1550  return rc;
1550  static int  static int
1551  usage(int rc)  usage(int rc)
1552  {  {
1553  fprintf(stderr, "Usage: pcregrep [-LMVcfhilnqrsvwx] [long-options] [pattern] [file1 file2 ...]\n");  option_item *op;
1554  fprintf(stderr, "Type `pcregrep --help' for more information.\n");  fprintf(stderr, "Usage: pcregrep [-");
1555    for (op = optionlist; op->one_char != 0; op++)
1556      {
1557      if (op->one_char > 0) fprintf(stderr, "%c", op->one_char);
1558      }
1559    fprintf(stderr, "] [long options] [pattern] [files]\n");
1560    fprintf(stderr, "Type `pcregrep --help' for more information and the long "
1561      "options.\n");
1562  return rc;  return rc;
1563  }  }
1564    
# Line 759  option_item *op; Line 1576  option_item *op;
1576    
1577  printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n");  printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n");
1578  printf("Search for PATTERN in each FILE or standard input.\n");  printf("Search for PATTERN in each FILE or standard input.\n");
1579  printf("PATTERN must be present if -f is not used.\n");  printf("PATTERN must be present if neither -e nor -f is used.\n");
1580  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");  
1581    
1582    #ifdef SUPPORT_LIBZ
1583    printf("Files whose names end in .gz are read using zlib.\n");
1584    #endif
1585    
1586    #ifdef SUPPORT_LIBBZ2
1587    printf("Files whose names end in .bz2 are read using bzlib2.\n");
1588    #endif
1589    
1590    #if defined SUPPORT_LIBZ || defined SUPPORT_LIBBZ2
1591    printf("Other files and the standard input are read as plain files.\n\n");
1592    #else
1593    printf("All files are read as plain files, without any interpretation.\n\n");
1594    #endif
1595    
1596    printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n");
1597  printf("Options:\n");  printf("Options:\n");
1598    
1599  for (op = optionlist; op->one_char != 0; op++)  for (op = optionlist; op->one_char != 0; op++)
# Line 770  for (op = optionlist; op->one_char != 0; Line 1601  for (op = optionlist; op->one_char != 0;
1601    int n;    int n;
1602    char s[4];    char s[4];
1603    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, "   ");
1604    printf("  %s --%s%n", s, op->long_name, &n);    n = 30 - printf("  %s --%s", s, op->long_name);
   n = 30 - n;  
1605    if (n < 1) n = 1;    if (n < 1) n = 1;
1606    printf("%.*s%s\n", n, "                    ", op->help_text);    printf("%.*s%s\n", n, "                    ", op->help_text);
1607    }    }
# Line 796  handle_option(int letter, int options) Line 1626  handle_option(int letter, int options)
1626  {  {
1627  switch(letter)  switch(letter)
1628    {    {
1629    case -1:  help(); exit(0);    case N_FOFFSETS: file_offsets = TRUE; break;
1630      case N_HELP: help(); exit(0);
1631      case N_LOFFSETS: line_offsets = number = TRUE; break;
1632    case 'c': count_only = TRUE; break;    case 'c': count_only = TRUE; break;
1633    case 'h': filenames = FALSE; break;    case 'F': process_options |= PO_FIXED_STRINGS; break;
1634      case 'H': filenames = FN_FORCE; break;
1635      case 'h': filenames = FN_NONE; break;
1636    case 'i': options |= PCRE_CASELESS; break;    case 'i': options |= PCRE_CASELESS; break;
1637    case 'l': filenames_only = TRUE; break;    case 'l': filenames = FN_ONLY; break;
1638    case 'L': filenames_nomatch_only = TRUE; break;    case 'L': filenames = FN_NOMATCH_ONLY; break;
1639    case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break;    case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break;
1640    case 'n': number = TRUE; break;    case 'n': number = TRUE; break;
1641      case 'o': only_matching = TRUE; break;
1642    case 'q': quiet = TRUE; break;    case 'q': quiet = TRUE; break;
1643    case 'r': recurse = TRUE; break;    case 'r': dee_action = dee_RECURSE; break;
1644    case 's': silent = TRUE; break;    case 's': silent = TRUE; break;
1645    case 'u': options |= PCRE_UTF8; break;    case 'u': options |= PCRE_UTF8; utf8 = TRUE; break;
1646    case 'v': invert = TRUE; break;    case 'v': invert = TRUE; break;
1647    case 'w': word_match = TRUE; break;    case 'w': process_options |= PO_WORD_MATCH; break;
1648    case 'x': whole_lines = TRUE; break;    case 'x': process_options |= PO_LINE_MATCH; break;
1649    
1650    case 'V':    case 'V':
1651    fprintf(stderr, "pcregrep version %s using ", VERSION);    fprintf(stderr, "pcregrep version %s\n", pcre_version());
   fprintf(stderr, "PCRE version %s\n", pcre_version());  
1652    exit(0);    exit(0);
1653    break;    break;
1654    
# Line 830  return options; Line 1664  return options;
1664    
1665    
1666  /*************************************************  /*************************************************
1667    *          Construct printed ordinal             *
1668    *************************************************/
1669    
1670    /* This turns a number into "1st", "3rd", etc. */
1671    
1672    static char *
1673    ordin(int n)
1674    {
1675    static char buffer[8];
1676    char *p = buffer;
1677    sprintf(p, "%d", n);
1678    while (*p != 0) p++;
1679    switch (n%10)
1680      {
1681      case 1: strcpy(p, "st"); break;
1682      case 2: strcpy(p, "nd"); break;
1683      case 3: strcpy(p, "rd"); break;
1684      default: strcpy(p, "th"); break;
1685      }
1686    return buffer;
1687    }
1688    
1689    
1690    
1691    /*************************************************
1692    *          Compile a single pattern              *
1693    *************************************************/
1694    
1695    /* When the -F option has been used, this is called for each substring.
1696    Otherwise it's called for each supplied pattern.
1697    
1698    Arguments:
1699      pattern        the pattern string
1700      options        the PCRE options
1701      filename       the file name, or NULL for a command-line pattern
1702      count          0 if this is the only command line pattern, or
1703                     number of the command line pattern, or
1704                     linenumber for a pattern from a file
1705    
1706    Returns:         TRUE on success, FALSE after an error
1707    */
1708    
1709    static BOOL
1710    compile_single_pattern(char *pattern, int options, char *filename, int count)
1711    {
1712    char buffer[MBUFTHIRD + 16];
1713    const char *error;
1714    int errptr;
1715    
1716    if (pattern_count >= MAX_PATTERN_COUNT)
1717      {
1718      fprintf(stderr, "pcregrep: Too many %spatterns (max %d)\n",
1719        (filename == NULL)? "command-line " : "", MAX_PATTERN_COUNT);
1720      return FALSE;
1721      }
1722    
1723    sprintf(buffer, "%s%.*s%s", prefix[process_options], MBUFTHIRD, pattern,
1724      suffix[process_options]);
1725    pattern_list[pattern_count] =
1726      pcre_compile(buffer, options, &error, &errptr, pcretables);
1727    if (pattern_list[pattern_count] != NULL)
1728      {
1729      pattern_count++;
1730      return TRUE;
1731      }
1732    
1733    /* Handle compile errors */
1734    
1735    errptr -= (int)strlen(prefix[process_options]);
1736    if (errptr > (int)strlen(pattern)) errptr = (int)strlen(pattern);
1737    
1738    if (filename == NULL)
1739      {
1740      if (count == 0)
1741        fprintf(stderr, "pcregrep: Error in command-line regex "
1742          "at offset %d: %s\n", errptr, error);
1743      else
1744        fprintf(stderr, "pcregrep: Error in %s command-line regex "
1745          "at offset %d: %s\n", ordin(count), errptr, error);
1746      }
1747    else
1748      {
1749      fprintf(stderr, "pcregrep: Error in regex in line %d of %s "
1750        "at offset %d: %s\n", count, filename, errptr, error);
1751      }
1752    
1753    return FALSE;
1754    }
1755    
1756    
1757    
1758    /*************************************************
1759    *           Compile one supplied pattern         *
1760    *************************************************/
1761    
1762    /* When the -F option has been used, each string may be a list of strings,
1763    separated by line breaks. They will be matched literally.
1764    
1765    Arguments:
1766      pattern        the pattern string
1767      options        the PCRE options
1768      filename       the file name, or NULL for a command-line pattern
1769      count          0 if this is the only command line pattern, or
1770                     number of the command line pattern, or
1771                     linenumber for a pattern from a file
1772    
1773    Returns:         TRUE on success, FALSE after an error
1774    */
1775    
1776    static BOOL
1777    compile_pattern(char *pattern, int options, char *filename, int count)
1778    {
1779    if ((process_options & PO_FIXED_STRINGS) != 0)
1780      {
1781      char *eop = pattern + strlen(pattern);
1782      char buffer[MBUFTHIRD];
1783      for(;;)
1784        {
1785        int ellength;
1786        char *p = end_of_line(pattern, eop, &ellength);
1787        if (ellength == 0)
1788          return compile_single_pattern(pattern, options, filename, count);
1789        sprintf(buffer, "%.*s", (int)(p - pattern - ellength), pattern);
1790        pattern = p;
1791        if (!compile_single_pattern(buffer, options, filename, count))
1792          return FALSE;
1793        }
1794      }
1795    else return compile_single_pattern(pattern, options, filename, count);
1796    }
1797    
1798    
1799    
1800    /*************************************************
1801  *                Main program                    *  *                Main program                    *
1802  *************************************************/  *************************************************/
1803    
# Line 840  main(int argc, char **argv) Line 1808  main(int argc, char **argv)
1808  {  {
1809  int i, j;  int i, j;
1810  int rc = 1;  int rc = 1;
1811  int options = 0;  int pcre_options = 0;
1812    int cmd_pattern_count = 0;
1813    int hint_count = 0;
1814  int errptr;  int errptr;
 const char *error;  
1815  BOOL only_one_at_top;  BOOL only_one_at_top;
1816    char *patterns[MAX_PATTERN_COUNT];
1817    const char *locale_from = "--locale";
1818    const char *error;
1819    
1820    /* Set the default line ending value from the default in the PCRE library;
1821    "lf", "cr", "crlf", and "any" are supported. Anything else is treated as "lf".
1822    */
1823    
1824    (void)pcre_config(PCRE_CONFIG_NEWLINE, &i);
1825    switch(i)
1826      {
1827      default:                 newline = (char *)"lf"; break;
1828      case '\r':               newline = (char *)"cr"; break;
1829      case ('\r' << 8) | '\n': newline = (char *)"crlf"; break;
1830      case -1:                 newline = (char *)"any"; break;
1831      case -2:                 newline = (char *)"anycrlf"; break;
1832      }
1833    
1834  /* Process the options */  /* Process the options */
1835    
# Line 857  for (i = 1; i < argc; i++) Line 1843  for (i = 1; i < argc; i++)
1843    if (argv[i][0] != '-') break;    if (argv[i][0] != '-') break;
1844    
1845    /* 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,
1846    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. */
1847    
1848    if (argv[i][1] == 0)    if (argv[i][1] == 0)
1849      {      {
1850      if (pattern_filename != NULL) break;      if (pattern_filename != NULL || pattern_count > 0) break;
1851        else exit(usage(2));        else exit(usage(2));
1852      }      }
1853    
# Line 883  for (i = 1; i < argc; i++) Line 1869  for (i = 1; i < argc; i++)
1869      /* Some long options have data that follows after =, for example file=name.      /* Some long options have data that follows after =, for example file=name.
1870      Some options have variations in the long name spelling: specifically, we      Some options have variations in the long name spelling: specifically, we
1871      allow "regexp" because GNU grep allows it, though I personally go along      allow "regexp" because GNU grep allows it, though I personally go along
1872      with Jeff Friedl in preferring "regex" without the "p". These options are      with Jeffrey Friedl and Larry Wall in preferring "regex" without the "p".
1873      entered in the table as "regex(p)". No option is in both these categories,      These options are entered in the table as "regex(p)". No option is in both
1874      fortunately. */      these categories, fortunately. */
1875    
1876      for (op = optionlist; op->one_char != 0; op++)      for (op = optionlist; op->one_char != 0; op++)
1877        {        {
# Line 900  for (i = 1; i < argc; i++) Line 1886  for (i = 1; i < argc; i++)
1886          else                 /* Special case xxx=data */          else                 /* Special case xxx=data */
1887            {            {
1888            int oplen = equals - op->long_name;            int oplen = equals - op->long_name;
1889            int arglen = (argequals == NULL)? strlen(arg) : argequals - arg;            int arglen = (argequals == NULL)? (int)strlen(arg) : argequals - arg;
1890            if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0)            if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0)
1891              {              {
1892              option_data = arg + arglen;              option_data = arg + arglen;
# Line 919  for (i = 1; i < argc; i++) Line 1905  for (i = 1; i < argc; i++)
1905          char buff2[24];          char buff2[24];
1906          int baselen = opbra - op->long_name;          int baselen = opbra - op->long_name;
1907          sprintf(buff1, "%.*s", baselen, op->long_name);          sprintf(buff1, "%.*s", baselen, op->long_name);
1908          sprintf(buff2, "%s%.*s", buff1, strlen(op->long_name) - baselen - 2,          sprintf(buff2, "%s%.*s", buff1,
1909            opbra + 1);            (int)strlen(op->long_name) - baselen - 2, opbra + 1);
1910          if (strcmp(arg, buff1) == 0 || strcmp(arg, buff2) == 0)          if (strcmp(arg, buff1) == 0 || strcmp(arg, buff2) == 0)
1911            break;            break;
1912          }          }
# Line 933  for (i = 1; i < argc; i++) Line 1919  for (i = 1; i < argc; i++)
1919        }        }
1920      }      }
1921    
1922    
1923      /* Jeffrey Friedl's debugging harness uses these additional options which
1924      are not in the right form for putting in the option table because they use
1925      only one hyphen, yet are more than one character long. By putting them
1926      separately here, they will not get displayed as part of the help() output,
1927      but I don't think Jeffrey will care about that. */
1928    
1929    #ifdef JFRIEDL_DEBUG
1930      else if (strcmp(argv[i], "-pre") == 0) {
1931              jfriedl_prefix = argv[++i];
1932              continue;
1933      } else if (strcmp(argv[i], "-post") == 0) {
1934              jfriedl_postfix = argv[++i];
1935              continue;
1936      } else if (strcmp(argv[i], "-XT") == 0) {
1937              sscanf(argv[++i], "%d", &jfriedl_XT);
1938              continue;
1939      } else if (strcmp(argv[i], "-XR") == 0) {
1940              sscanf(argv[++i], "%d", &jfriedl_XR);
1941              continue;
1942      }
1943    #endif
1944    
1945    
1946    /* 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
1947    continue till we hit the last one or one that needs data. */    continue till we hit the last one or one that needs data. */
1948    
# Line 955  for (i = 1; i < argc; i++) Line 1965  for (i = 1; i < argc; i++)
1965          option_data = s+1;          option_data = s+1;
1966          break;          break;
1967          }          }
1968        options = handle_option(*s++, options);        pcre_options = handle_option(*s++, pcre_options);
1969        }        }
1970      }      }
1971    
1972    /* 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
1973      is NO_DATA, it means that there is no data, and the option might set
1974      something in the PCRE options. */
1975    
1976    if (op->type == OP_NODATA)    if (op->type == OP_NODATA)
     options = handle_option(op->one_char, options);  
   else  
1977      {      {
1978      if (*option_data == 0)      pcre_options = handle_option(op->one_char, pcre_options);
1979        continue;
1980        }
1981    
1982      /* If the option type is OP_OP_STRING or OP_OP_NUMBER, it's an option that
1983      either has a value or defaults to something. It cannot have data in a
1984      separate item. At the moment, the only such options are "colo(u)r" and
1985      Jeffrey Friedl's special -S debugging option. */
1986    
1987      if (*option_data == 0 &&
1988          (op->type == OP_OP_STRING || op->type == OP_OP_NUMBER))
1989        {
1990        switch (op->one_char)
1991        {        {
1992        if (i >= argc - 1 || longopwasequals)        case N_COLOUR:
1993          {        colour_option = (char *)"auto";
1994          fprintf(stderr, "pcregrep: Data missing after %s\n", argv[i]);        break;
1995          exit(usage(2));  #ifdef JFRIEDL_DEBUG
1996          }        case 'S':
1997        option_data = argv[++i];        S_arg = 0;
1998          break;
1999    #endif
2000        }        }
2001        continue;
2002        }
2003    
2004      /* Otherwise, find the data string for the option. */
2005    
2006      if (op->type == OP_STRING) *((char **)op->dataptr) = option_data; else    if (*option_data == 0)
2007        {
2008        if (i >= argc - 1 || longopwasequals)
2009        {        {
2010        char *endptr;        fprintf(stderr, "pcregrep: Data missing after %s\n", argv[i]);
2011        int n = strtoul(option_data, &endptr, 10);        exit(usage(2));
2012        if (*endptr != 0)        }
2013        option_data = argv[++i];
2014        }
2015    
2016      /* If the option type is OP_PATLIST, it's the -e option, which can be called
2017      multiple times to create a list of patterns. */
2018    
2019      if (op->type == OP_PATLIST)
2020        {
2021        if (cmd_pattern_count >= MAX_PATTERN_COUNT)
2022          {
2023          fprintf(stderr, "pcregrep: Too many command-line patterns (max %d)\n",
2024            MAX_PATTERN_COUNT);
2025          return 2;
2026          }
2027        patterns[cmd_pattern_count++] = option_data;
2028        }
2029    
2030      /* Otherwise, deal with single string or numeric data values. */
2031    
2032      else if (op->type != OP_NUMBER && op->type != OP_OP_NUMBER)
2033        {
2034        *((char **)op->dataptr) = option_data;
2035        }
2036      else
2037        {
2038        char *endptr;
2039        int n = strtoul(option_data, &endptr, 10);
2040        if (*endptr != 0)
2041          {
2042          if (longop)
2043          {          {
2044          if (longop)          char *equals = strchr(op->long_name, '=');
2045            fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%s\n",          int nlen = (equals == NULL)? (int)strlen(op->long_name) :
2046              option_data, op->long_name);            equals - op->long_name;
2047          else          fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%.*s\n",
2048            fprintf(stderr, "pcregrep: Malformed number \"%s\" after -%c\n",            option_data, nlen, op->long_name);
             option_data, op->one_char);  
         exit(usage(2));  
2049          }          }
2050        *((int *)op->dataptr) = n;        else
2051            fprintf(stderr, "pcregrep: Malformed number \"%s\" after -%c\n",
2052              option_data, op->one_char);
2053          exit(usage(2));
2054        }        }
2055        *((int *)op->dataptr) = n;
2056      }      }
2057    }    }
2058    
# Line 1003  if (both_context > 0) Line 2065  if (both_context > 0)
2065    if (before_context == 0) before_context = both_context;    if (before_context == 0) before_context = both_context;
2066    }    }
2067    
2068  pattern_list = (pcre **)malloc(MAX_PATTERN_COUNT * sizeof(pcre *));  /* Only one of --only-matching, --file-offsets, or --line-offsets is permitted.
2069  hints_list = (pcre_extra **)malloc(MAX_PATTERN_COUNT * sizeof(pcre_extra *));  However, the latter two set the only_matching flag. */
2070    
2071  if (pattern_list == NULL || hints_list == NULL)  if ((only_matching && (file_offsets || line_offsets)) ||
2072        (file_offsets && line_offsets))
2073    {    {
2074    fprintf(stderr, "pcregrep: malloc failed\n");    fprintf(stderr, "pcregrep: Cannot mix --only-matching, --file-offsets "
2075    return 2;      "and/or --line-offsets\n");
2076      exit(usage(2));
2077    }    }
2078    
2079  /* Compile the regular expression(s). */  if (file_offsets || line_offsets) only_matching = TRUE;
2080    
2081  if (pattern_filename != NULL)  /* If a locale has not been provided as an option, see if the LC_CTYPE or
2082    LC_ALL environment variable is set, and if so, use it. */
2083    
2084    if (locale == NULL)
2085    {    {
2086    FILE *f = fopen(pattern_filename, "r");    locale = getenv("LC_ALL");
2087    char buffer[MBUFTHIRD + 16];    locale_from = "LCC_ALL";
2088    char *rdstart;    }
2089    int adjust = 0;  
2090    if (locale == NULL)
2091      {
2092      locale = getenv("LC_CTYPE");
2093      locale_from = "LC_CTYPE";
2094      }
2095    
2096    if (f == NULL)  /* If a locale has been provided, set it, and generate the tables the PCRE
2097    needs. Otherwise, pcretables==NULL, which causes the use of default tables. */
2098    
2099    if (locale != NULL)
2100      {
2101      if (setlocale(LC_CTYPE, locale) == NULL)
2102      {      {
2103      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pattern_filename,      fprintf(stderr, "pcregrep: Failed to set locale %s (obtained from %s)\n",
2104        strerror(errno));        locale, locale_from);
2105      return 2;      return 2;
2106      }      }
2107      pcretables = pcre_maketables();
2108      }
2109    
2110    /* Sort out colouring */
2111    
2112    if (whole_lines)  if (colour_option != NULL && strcmp(colour_option, "never") != 0)
2113      {
2114      if (strcmp(colour_option, "always") == 0) do_colour = TRUE;
2115      else if (strcmp(colour_option, "auto") == 0) do_colour = is_stdout_tty();
2116      else
2117      {      {
2118      strcpy(buffer, "^(?:");      fprintf(stderr, "pcregrep: Unknown colour setting \"%s\"\n",
2119      adjust = 4;        colour_option);
2120        return 2;
2121      }      }
2122    else if (word_match)    if (do_colour)
2123      {      {
2124      strcpy(buffer, "\\b");      char *cs = getenv("PCREGREP_COLOUR");
2125      adjust = 2;      if (cs == NULL) cs = getenv("PCREGREP_COLOR");
2126        if (cs != NULL) colour_string = cs;
2127      }      }
2128      }
2129    
2130    rdstart = buffer + adjust;  /* Interpret the newline type; the default settings are Unix-like. */
2131    while (fgets(rdstart, MBUFTHIRD, f) != NULL)  
2132    if (strcmp(newline, "cr") == 0 || strcmp(newline, "CR") == 0)
2133      {
2134      pcre_options |= PCRE_NEWLINE_CR;
2135      endlinetype = EL_CR;
2136      }
2137    else if (strcmp(newline, "lf") == 0 || strcmp(newline, "LF") == 0)
2138      {
2139      pcre_options |= PCRE_NEWLINE_LF;
2140      endlinetype = EL_LF;
2141      }
2142    else if (strcmp(newline, "crlf") == 0 || strcmp(newline, "CRLF") == 0)
2143      {
2144      pcre_options |= PCRE_NEWLINE_CRLF;
2145      endlinetype = EL_CRLF;
2146      }
2147    else if (strcmp(newline, "any") == 0 || strcmp(newline, "ANY") == 0)
2148      {
2149      pcre_options |= PCRE_NEWLINE_ANY;
2150      endlinetype = EL_ANY;
2151      }
2152    else if (strcmp(newline, "anycrlf") == 0 || strcmp(newline, "ANYCRLF") == 0)
2153      {
2154      pcre_options |= PCRE_NEWLINE_ANYCRLF;
2155      endlinetype = EL_ANYCRLF;
2156      }
2157    else
2158      {
2159      fprintf(stderr, "pcregrep: Invalid newline specifier \"%s\"\n", newline);
2160      return 2;
2161      }
2162    
2163    /* Interpret the text values for -d and -D */
2164    
2165    if (dee_option != NULL)
2166      {
2167      if (strcmp(dee_option, "read") == 0) dee_action = dee_READ;
2168      else if (strcmp(dee_option, "recurse") == 0) dee_action = dee_RECURSE;
2169      else if (strcmp(dee_option, "skip") == 0) dee_action = dee_SKIP;
2170      else
2171      {      {
2172      char *s = rdstart + (int)strlen(rdstart);      fprintf(stderr, "pcregrep: Invalid value \"%s\" for -d\n", dee_option);
2173      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;  
       }  
2174      }      }
   fclose(f);  
2175    }    }
2176    
2177  /* If no file name, a single regex must be given inline. */  if (DEE_option != NULL)
2178      {
2179      if (strcmp(DEE_option, "read") == 0) DEE_action = DEE_READ;
2180      else if (strcmp(DEE_option, "skip") == 0) DEE_action = DEE_SKIP;
2181      else
2182        {
2183        fprintf(stderr, "pcregrep: Invalid value \"%s\" for -D\n", DEE_option);
2184        return 2;
2185        }
2186      }
2187    
2188  else  /* Check the values for Jeffrey Friedl's debugging options. */
2189    
2190    #ifdef JFRIEDL_DEBUG
2191    if (S_arg > 9)
2192    {    {
2193    char buffer[MBUFTHIRD + 16];    fprintf(stderr, "pcregrep: bad value for -S option\n");
2194    char *pat;    return 2;
2195    int adjust = 0;    }
2196    if (jfriedl_XT != 0 || jfriedl_XR != 0)
2197      {
2198      if (jfriedl_XT == 0) jfriedl_XT = 1;
2199      if (jfriedl_XR == 0) jfriedl_XR = 1;
2200      }
2201    #endif
2202    
2203    /* Get memory to store the pattern and hints lists. */
2204    
2205    pattern_list = (pcre **)malloc(MAX_PATTERN_COUNT * sizeof(pcre *));
2206    hints_list = (pcre_extra **)malloc(MAX_PATTERN_COUNT * sizeof(pcre_extra *));
2207    
2208    if (pattern_list == NULL || hints_list == NULL)
2209      {
2210      fprintf(stderr, "pcregrep: malloc failed\n");
2211      goto EXIT2;
2212      }
2213    
2214    /* If no patterns were provided by -e, and there is no file provided by -f,
2215    the first argument is the one and only pattern, and it must exist. */
2216    
2217    if (cmd_pattern_count == 0 && pattern_filename == NULL)
2218      {
2219    if (i >= argc) return usage(2);    if (i >= argc) return usage(2);
2220      patterns[cmd_pattern_count++] = argv[i++];
2221      }
2222    
2223    /* Compile the patterns that were provided on the command line, either by
2224    multiple uses of -e or as a single unkeyed pattern. */
2225    
2226    if (whole_lines)  for (j = 0; j < cmd_pattern_count; j++)
2227      {
2228      if (!compile_pattern(patterns[j], pcre_options, NULL,
2229           (j == 0 && cmd_pattern_count == 1)? 0 : j + 1))
2230        goto EXIT2;
2231      }
2232    
2233    /* Compile the regular expressions that are provided in a file. */
2234    
2235    if (pattern_filename != NULL)
2236      {
2237      int linenumber = 0;
2238      FILE *f;
2239      char *filename;
2240      char buffer[MBUFTHIRD];
2241    
2242      if (strcmp(pattern_filename, "-") == 0)
2243      {      {
2244      sprintf(buffer, "^(?:%.*s)$", MBUFTHIRD, argv[i++]);      f = stdin;
2245      pat = buffer;      filename = stdin_name;
     adjust = 4;  
2246      }      }
2247    else if (word_match)    else
2248      {      {
2249      sprintf(buffer, "\\b%.*s\\b", MBUFTHIRD, argv[i++]);      f = fopen(pattern_filename, "r");
2250      pat = buffer;      if (f == NULL)
2251      adjust = 2;        {
2252          fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pattern_filename,
2253            strerror(errno));
2254          goto EXIT2;
2255          }
2256        filename = pattern_filename;
2257      }      }
   else pat = argv[i++];  
2258    
2259    pattern_list[0] = pcre_compile(pat, options, &error, &errptr, NULL);    while (fgets(buffer, MBUFTHIRD, f) != NULL)
   
   if (pattern_list[0] == NULL)  
2260      {      {
2261      fprintf(stderr, "pcregrep: Error in regex at offset %d: %s\n",      char *s = buffer + (int)strlen(buffer);
2262        errptr - adjust, error);      while (s > buffer && isspace((unsigned char)(s[-1]))) s--;
2263      return 2;      *s = 0;
2264        linenumber++;
2265        if (buffer[0] == 0) continue;   /* Skip blank lines */
2266        if (!compile_pattern(buffer, pcre_options, filename, linenumber))
2267          goto EXIT2;
2268      }      }
2269    pattern_count++;  
2270      if (f != stdin) fclose(f);
2271    }    }
2272    
2273  /* 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 1111  for (j = 0; j < pattern_count; j++) Line 2280  for (j = 0; j < pattern_count; j++)
2280      char s[16];      char s[16];
2281      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);
2282      fprintf(stderr, "pcregrep: Error while studying regex%s: %s\n", s, error);      fprintf(stderr, "pcregrep: Error while studying regex%s: %s\n", s, error);
2283      return 2;      goto EXIT2;
2284      }      }
2285      hint_count++;
2286    }    }
2287    
2288  /* If there are include or exclude patterns, compile them. */  /* If there are include or exclude patterns, compile them. */
2289    
2290  if (exclude_pattern != NULL)  if (exclude_pattern != NULL)
2291    {    {
2292    exclude_compiled = pcre_compile(exclude_pattern, 0, &error, &errptr, NULL);    exclude_compiled = pcre_compile(exclude_pattern, 0, &error, &errptr,
2293        pcretables);
2294    if (exclude_compiled == NULL)    if (exclude_compiled == NULL)
2295      {      {
2296      fprintf(stderr, "pcregrep: Error in 'exclude' regex at offset %d: %s\n",      fprintf(stderr, "pcregrep: Error in 'exclude' regex at offset %d: %s\n",
2297        errptr, error);        errptr, error);
2298      return 2;      goto EXIT2;
2299      }      }
2300    }    }
2301    
2302  if (include_pattern != NULL)  if (include_pattern != NULL)
2303    {    {
2304    include_compiled = pcre_compile(include_pattern, 0, &error, &errptr, NULL);    include_compiled = pcre_compile(include_pattern, 0, &error, &errptr,
2305        pcretables);
2306    if (include_compiled == NULL)    if (include_compiled == NULL)
2307      {      {
2308      fprintf(stderr, "pcregrep: Error in 'include' regex at offset %d: %s\n",      fprintf(stderr, "pcregrep: Error in 'include' regex at offset %d: %s\n",
2309        errptr, error);        errptr, error);
2310      return 2;      goto EXIT2;
2311        }
2312      }
2313    
2314    if (exclude_dir_pattern != NULL)
2315      {
2316      exclude_dir_compiled = pcre_compile(exclude_dir_pattern, 0, &error, &errptr,
2317        pcretables);
2318      if (exclude_dir_compiled == NULL)
2319        {
2320        fprintf(stderr, "pcregrep: Error in 'exclude_dir' regex at offset %d: %s\n",
2321          errptr, error);
2322        goto EXIT2;
2323        }
2324      }
2325    
2326    if (include_dir_pattern != NULL)
2327      {
2328      include_dir_compiled = pcre_compile(include_dir_pattern, 0, &error, &errptr,
2329        pcretables);
2330      if (include_dir_compiled == NULL)
2331        {
2332        fprintf(stderr, "pcregrep: Error in 'include_dir' regex at offset %d: %s\n",
2333          errptr, error);
2334        goto EXIT2;
2335      }      }
2336    }    }
2337    
2338  /* If there are no further arguments, do the business on stdin and exit */  /* If there are no further arguments, do the business on stdin and exit. */
2339    
2340  if (i >= argc) return pcregrep(stdin,  if (i >= argc)
2341    (filenames_only || filenames_nomatch_only)? stdin_name : NULL);    {
2342      rc = pcregrep(stdin, FR_PLAIN, (filenames > FN_DEFAULT)? stdin_name : NULL);
2343      goto EXIT;
2344      }
2345    
2346  /* Otherwise, work through the remaining arguments as files or directories.  /* Otherwise, work through the remaining arguments as files or directories.
2347  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
2348  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
2349  */  otherwise forced. */
2350    
2351  only_one_at_top = (i == argc - 1);  only_one_at_top = i == argc - 1;   /* Catch initial value of i */
2352    
2353  for (; i < argc; i++)  for (; i < argc; i++)
2354    {    {
2355    int frc = grep_or_recurse(argv[i], recurse, filenames, only_one_at_top);    int frc = grep_or_recurse(argv[i], dee_action == dee_RECURSE,
2356        only_one_at_top);
2357    if (frc > 1) rc = frc;    if (frc > 1) rc = frc;
2358      else if (frc == 0 && rc == 1) rc = 0;      else if (frc == 0 && rc == 1) rc = 0;
2359    }    }
2360    
2361    EXIT:
2362    if (pattern_list != NULL)
2363      {
2364      for (i = 0; i < pattern_count; i++) free(pattern_list[i]);
2365      free(pattern_list);
2366      }
2367    if (hints_list != NULL)
2368      {
2369      for (i = 0; i < hint_count; i++) free(hints_list[i]);
2370      free(hints_list);
2371      }
2372  return rc;  return rc;
2373    
2374    EXIT2:
2375    rc = 2;
2376    goto EXIT;
2377  }  }
2378    
2379  /* End of pcregrep */  /* End of pcregrep */

Legend:
Removed from v.85  
changed lines
  Added in v.345

  ViewVC Help
Powered by ViewVC 1.1.5