/[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 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 587 by ph10, Fri Jan 14 19:01:25 2011 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) 1997-2006 University of Cambridge             Copyright (c) 1997-2011 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>  #include <locale.h>
46  #include <stdio.h>  #include <stdio.h>
# Line 46  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 56  POSSIBILITY OF SUCH DAMAGE. Line 70  POSSIBILITY OF SUCH DAMAGE.
70    
71  typedef int BOOL;  typedef int BOOL;
72    
 #define VERSION "4.2 09-Jan-2006"  
73  #define MAX_PATTERN_COUNT 100  #define MAX_PATTERN_COUNT 100
74    #define OFFSET_SIZE 99
75    
76  #if BUFSIZ > 8192  #if BUFSIZ > 8192
77  #define MBUFTHIRD BUFSIZ  #define MBUFTHIRD BUFSIZ
# Line 65  typedef int BOOL; Line 79  typedef int BOOL;
79  #define MBUFTHIRD 8192  #define MBUFTHIRD 8192
80  #endif  #endif
81    
   
82  /* Values for the "filenames" variable, which specifies options for file name  /* Values for the "filenames" variable, which specifies options for file name
83  output. The order is important; it is assumed that a file name is wanted for  output. The order is important; it is assumed that a file name is wanted for
84  all values greater than FN_DEFAULT. */  all values greater than FN_DEFAULT. */
85    
86  enum { FN_NONE, FN_DEFAULT, FN_ONLY, FN_NOMATCH_ONLY, FN_FORCE };  enum { FN_NONE, FN_DEFAULT, FN_MATCH_ONLY, FN_NOMATCH_ONLY, FN_FORCE };
87    
88    /* File reading styles */
89    
90    enum { FR_PLAIN, FR_LIBZ, FR_LIBBZ2 };
91    
92  /* Actions for the -d and -D options */  /* Actions for the -d and -D options */
93    
# Line 83  enum { DEE_READ, DEE_SKIP }; Line 100  enum { DEE_READ, DEE_SKIP };
100  #define PO_LINE_MATCH     0x0002  #define PO_LINE_MATCH     0x0002
101  #define PO_FIXED_STRINGS  0x0004  #define PO_FIXED_STRINGS  0x0004
102    
103    /* Line ending types */
104    
105    enum { EL_LF, EL_CR, EL_CRLF, EL_ANY, EL_ANYCRLF };
106    
107    /* In newer versions of gcc, with FORTIFY_SOURCE set (the default in some
108    environments), a warning is issued if the value of fwrite() is ignored.
109    Unfortunately, casting to (void) does not suppress the warning. To get round
110    this, we use a macro that compiles a fudge. Oddly, this does not also seem to
111    apply to fprintf(). */
112    
113    #define FWRITE(a,b,c,d) if (fwrite(a,b,c,d)) {}
114    
115    
116    
117  /*************************************************  /*************************************************
# Line 94  regular code. */ Line 123  regular code. */
123    
124  #ifdef JFRIEDL_DEBUG  #ifdef JFRIEDL_DEBUG
125  static int S_arg = -1;  static int S_arg = -1;
126    static unsigned int jfriedl_XR = 0; /* repeat regex attempt this many times */
127    static unsigned int jfriedl_XT = 0; /* replicate text this many times */
128    static const char *jfriedl_prefix = "";
129    static const char *jfriedl_postfix = "";
130  #endif  #endif
131    
132    static int  endlinetype;
133    
134  static char *colour_string = (char *)"1;31";  static char *colour_string = (char *)"1;31";
135  static char *colour_option = NULL;  static char *colour_option = NULL;
136  static char *dee_option = NULL;  static char *dee_option = NULL;
137  static char *DEE_option = NULL;  static char *DEE_option = NULL;
138    static char *newline = NULL;
139  static char *pattern_filename = NULL;  static char *pattern_filename = NULL;
140  static char *stdin_name = (char *)"(standard input)";  static char *stdin_name = (char *)"(standard input)";
141  static char *locale = NULL;  static char *locale = NULL;
# Line 107  static char *locale = NULL; Line 143  static char *locale = NULL;
143  static const unsigned char *pcretables = NULL;  static const unsigned char *pcretables = NULL;
144    
145  static int  pattern_count = 0;  static int  pattern_count = 0;
146  static pcre **pattern_list;  static pcre **pattern_list = NULL;
147  static pcre_extra **hints_list;  static pcre_extra **hints_list = NULL;
148    
149  static char *include_pattern = NULL;  static char *include_pattern = NULL;
150  static char *exclude_pattern = NULL;  static char *exclude_pattern = NULL;
151    static char *include_dir_pattern = NULL;
152    static char *exclude_dir_pattern = NULL;
153    
154  static pcre *include_compiled = NULL;  static pcre *include_compiled = NULL;
155  static pcre *exclude_compiled = NULL;  static pcre *exclude_compiled = NULL;
156    static pcre *include_dir_compiled = NULL;
157    static pcre *exclude_dir_compiled = NULL;
158    
159  static int after_context = 0;  static int after_context = 0;
160  static int before_context = 0;  static int before_context = 0;
# Line 123  static int dee_action = dee_READ; Line 163  static int dee_action = dee_READ;
163  static int DEE_action = DEE_READ;  static int DEE_action = DEE_READ;
164  static int error_count = 0;  static int error_count = 0;
165  static int filenames = FN_DEFAULT;  static int filenames = FN_DEFAULT;
166    static int only_matching = -1;
167  static int process_options = 0;  static int process_options = 0;
168    
169    static unsigned long int match_limit = 0;
170    static unsigned long int match_limit_recursion = 0;
171    
172  static BOOL count_only = FALSE;  static BOOL count_only = FALSE;
173  static BOOL do_colour = FALSE;  static BOOL do_colour = FALSE;
174    static BOOL file_offsets = FALSE;
175  static BOOL hyphenpending = FALSE;  static BOOL hyphenpending = FALSE;
176  static BOOL invert = FALSE;  static BOOL invert = FALSE;
177    static BOOL line_buffered = FALSE;
178    static BOOL line_offsets = FALSE;
179  static BOOL multiline = FALSE;  static BOOL multiline = FALSE;
180  static BOOL number = FALSE;  static BOOL number = FALSE;
181  static BOOL only_matching = FALSE;  static BOOL omit_zero_count = FALSE;
182    static BOOL resource_error = FALSE;
183  static BOOL quiet = FALSE;  static BOOL quiet = FALSE;
184  static BOOL silent = FALSE;  static BOOL silent = FALSE;
185    static BOOL utf8 = FALSE;
186    
187  /* Structure for options and list of them */  /* Structure for options and list of them */
188    
189  enum { OP_NODATA, OP_STRING, OP_OP_STRING, OP_NUMBER, OP_OP_NUMBER,  enum { OP_NODATA, OP_STRING, OP_OP_STRING, OP_NUMBER, OP_LONGNUMBER,
190         OP_PATLIST };         OP_OP_NUMBER, OP_PATLIST };
191    
192  typedef struct option_item {  typedef struct option_item {
193    int type;    int type;
# Line 151  typedef struct option_item { Line 200  typedef struct option_item {
200  /* Options without a single-letter equivalent get a negative value. This can be  /* Options without a single-letter equivalent get a negative value. This can be
201  used to identify them. */  used to identify them. */
202    
203  #define N_COLOUR    (-1)  #define N_COLOUR       (-1)
204  #define N_EXCLUDE   (-2)  #define N_EXCLUDE      (-2)
205  #define N_HELP      (-3)  #define N_EXCLUDE_DIR  (-3)
206  #define N_INCLUDE   (-4)  #define N_HELP         (-4)
207  #define N_LABEL     (-5)  #define N_INCLUDE      (-5)
208  #define N_LOCALE    (-6)  #define N_INCLUDE_DIR  (-6)
209  #define N_NULL      (-7)  #define N_LABEL        (-7)
210    #define N_LOCALE       (-8)
211    #define N_NULL         (-9)
212    #define N_LOFFSETS     (-10)
213    #define N_FOFFSETS     (-11)
214    #define N_LBUFFER      (-12)
215    #define N_M_LIMIT      (-13)
216    #define N_M_LIMIT_REC  (-14)
217    
218  static option_item optionlist[] = {  static option_item optionlist[] = {
219    { OP_NODATA,    N_NULL,   NULL,              "",              "  terminate options" },    { OP_NODATA,     N_NULL,   NULL,              "",              "  terminate options" },
220    { OP_NODATA,    N_HELP,   NULL,              "help",          "display this help and exit" },    { OP_NODATA,     N_HELP,   NULL,              "help",          "display this help and exit" },
221    { 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" },
222    { 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" },
223    { OP_OP_STRING, N_COLOUR, &colour_option,    "color=option",  "matched text color option" },    { OP_OP_STRING,  N_COLOUR, &colour_option,    "color=option",  "matched text color option" },
224    { OP_NUMBER,    'C',      &both_context,     "context=number", "set number of context lines, before & after" },    { OP_OP_STRING,  N_COLOUR, &colour_option,    "colour=option", "matched text colour option" },
225    { 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" },
226    { OP_OP_STRING, N_COLOUR, &colour_option,    "colour=option", "matched text colour option" },    { OP_NODATA,     'c',      NULL,              "count",         "print only a count of matching lines per FILE" },
227    { OP_STRING,    'D',      &DEE_option,       "devices=action","how to handle devices, FIFOs, and sockets" },    { OP_STRING,     'D',      &DEE_option,       "devices=action","how to handle devices, FIFOs, and sockets" },
228    { OP_STRING,    'd',      &dee_option,       "directories=action", "how to handle directories" },    { OP_STRING,     'd',      &dee_option,       "directories=action", "how to handle directories" },
229    { OP_PATLIST,   'e',      NULL,              "regex(p)",      "specify pattern (may be used more than once)" },    { OP_PATLIST,    'e',      NULL,              "regex(p)=pattern", "specify pattern (may be used more than once)" },
230    { OP_NODATA,    'F',      NULL,              "fixed_strings", "patterns are sets of newline-separated strings" },    { OP_NODATA,     'F',      NULL,              "fixed-strings", "patterns are sets of newline-separated strings" },
231    { OP_STRING,    'f',      &pattern_filename, "file=path",     "read patterns from file" },    { OP_STRING,     'f',      &pattern_filename, "file=path",     "read patterns from file" },
232    { OP_NODATA,    'H',      NULL,              "with-filename", "force the prefixing filename on output" },    { OP_NODATA,     N_FOFFSETS, NULL,            "file-offsets",  "output file offsets, not text" },
233    { OP_NODATA,    'h',      NULL,              "no-filename",   "suppress the prefixing filename on output" },    { OP_NODATA,     'H',      NULL,              "with-filename", "force the prefixing filename on output" },
234    { OP_NODATA,    'i',      NULL,              "ignore-case",   "ignore case distinctions" },    { OP_NODATA,     'h',      NULL,              "no-filename",   "suppress the prefixing filename on output" },
235    { OP_NODATA,    'l',      NULL,              "files-with-matches", "print only FILE names containing matches" },    { OP_NODATA,     'i',      NULL,              "ignore-case",   "ignore case distinctions" },
236    { OP_NODATA,    'L',      NULL,              "files-without-match","print only FILE names not containing matches" },    { OP_NODATA,     'l',      NULL,              "files-with-matches", "print only FILE names containing matches" },
237    { OP_STRING,    N_LABEL,  &stdin_name,       "label=name",    "set name for standard input" },    { OP_NODATA,     'L',      NULL,              "files-without-match","print only FILE names not containing matches" },
238    { OP_STRING,    N_LOCALE, &locale,           "locale=locale", "use the named locale" },    { OP_STRING,     N_LABEL,  &stdin_name,       "label=name",    "set name for standard input" },
239    { OP_NODATA,    'M',      NULL,              "multiline",     "run in multiline mode" },    { OP_NODATA,     N_LBUFFER, NULL,             "line-buffered", "use line buffering" },
240    { OP_NODATA,    'n',      NULL,              "line-number",   "print line number with output lines" },    { OP_NODATA,     N_LOFFSETS, NULL,            "line-offsets",  "output line numbers and offsets, not text" },
241    { OP_NODATA,    'o',      NULL,              "only-matching", "show only the part of the line that matched" },    { OP_STRING,     N_LOCALE, &locale,           "locale=locale", "use the named locale" },
242    { OP_NODATA,    'q',      NULL,              "quiet",         "suppress output, just set return code" },    { OP_LONGNUMBER, N_M_LIMIT, &match_limit,     "match-limit=number", "set PCRE match limit option" },
243    { OP_NODATA,    'r',      NULL,              "recursive",     "recursively scan sub-directories" },    { OP_LONGNUMBER, N_M_LIMIT_REC, &match_limit_recursion, "recursion-limit=number", "set PCRE match recursion limit option" },
244    { OP_STRING,    N_EXCLUDE,&exclude_pattern,  "exclude=pattern","exclude matching files when recursing" },    { OP_NODATA,     'M',      NULL,              "multiline",     "run in multiline mode" },
245    { OP_STRING,    N_INCLUDE,&include_pattern,  "include=pattern","include matching files when recursing" },    { OP_STRING,     'N',      &newline,          "newline=type",  "set newline type (CR, LF, CRLF, ANYCRLF or ANY)" },
246      { OP_NODATA,     'n',      NULL,              "line-number",   "print line number with output lines" },
247      { OP_OP_NUMBER,  'o',      &only_matching,    "only-matching=n", "show only the part of the line that matched" },
248      { OP_NODATA,     'q',      NULL,              "quiet",         "suppress output, just set return code" },
249      { OP_NODATA,     'r',      NULL,              "recursive",     "recursively scan sub-directories" },
250      { OP_STRING,     N_EXCLUDE,&exclude_pattern,  "exclude=pattern","exclude matching files when recursing" },
251      { OP_STRING,     N_INCLUDE,&include_pattern,  "include=pattern","include matching files when recursing" },
252      { OP_STRING,     N_EXCLUDE_DIR,&exclude_dir_pattern, "exclude-dir=pattern","exclude matching directories when recursing" },
253      { OP_STRING,     N_INCLUDE_DIR,&include_dir_pattern, "include-dir=pattern","include matching directories when recursing" },
254    
255      /* These two were accidentally implemented with underscores instead of
256      hyphens in the option names. As this was not discovered for several releases,
257      the incorrect versions are left in the table for compatibility. However, the
258      --help function misses out any option that has an underscore in its name. */
259    
260      { OP_STRING,    N_EXCLUDE_DIR,&exclude_dir_pattern, "exclude_dir=pattern","exclude matching directories when recursing" },
261      { OP_STRING,    N_INCLUDE_DIR,&include_dir_pattern, "include_dir=pattern","include matching directories when recursing" },
262    
263  #ifdef JFRIEDL_DEBUG  #ifdef JFRIEDL_DEBUG
264    { OP_OP_NUMBER, 'S',      &S_arg,            "jeffS",         "replace matched (sub)string with X" },    { OP_OP_NUMBER, 'S',      &S_arg,            "jeffS",         "replace matched (sub)string with X" },
265  #endif  #endif
# Line 210  static const char *prefix[] = { Line 283  static const char *prefix[] = {
283  static const char *suffix[] = {  static const char *suffix[] = {
284    "", "\\b", ")$",   ")$",   "\\E", "\\E\\b", "\\E)$",   "\\E)$" };    "", "\\b", ")$",   ")$",   "\\E", "\\E\\b", "\\E)$",   "\\E)$" };
285    
286    /* UTF-8 tables - used only when the newline setting is "any". */
287    
288    const int utf8_table3[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};
289    
290    const char utf8_table4[] = {
291      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
292      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
293      2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
294      3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 };
295    
296    
297    
298    /*************************************************
299    *         Exit from the program                  *
300    *************************************************/
301    
302    /* If there has been a resource error, give a suitable message.
303    
304    Argument:  the return code
305    Returns:   does not return
306    */
307    
308    static void
309    pcregrep_exit(int rc)
310    {
311    if (resource_error)
312      {
313      fprintf(stderr, "pcregrep: Error %d or %d means that a resource limit "
314        "was exceeded.\n", PCRE_ERROR_MATCHLIMIT, PCRE_ERROR_RECURSIONLIMIT);
315      fprintf(stderr, "pcregrep: Check your regex for nested unlimited loops.\n");
316      }
317    
318    exit(rc);
319    }
320    
321    
322  /*************************************************  /*************************************************
# Line 222  although at present the only ones are fo Line 329  although at present the only ones are fo
329    
330  /************* Directory scanning in Unix ***********/  /************* Directory scanning in Unix ***********/
331    
332  #if IS_UNIX  #if defined HAVE_SYS_STAT_H && defined HAVE_DIRENT_H && defined HAVE_SYS_TYPES_H
333  #include <sys/types.h>  #include <sys/types.h>
334  #include <sys/stat.h>  #include <sys/stat.h>
335  #include <dirent.h>  #include <dirent.h>
# Line 254  for (;;) Line 361  for (;;)
361    if (strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0)    if (strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0)
362      return dent->d_name;      return dent->d_name;
363    }    }
364  return NULL;   /* Keep compiler happy; never executed */  /* Control never reaches here */
365  }  }
366    
367  static void  static void
# Line 276  return (statbuf.st_mode & S_IFMT) == S_I Line 383  return (statbuf.st_mode & S_IFMT) == S_I
383  }  }
384    
385    
386  /************* Test stdout for being a terminal in Unix **********/  /************* Test for a terminal in Unix **********/
387    
388  static BOOL  static BOOL
389  is_stdout_tty(void)  is_stdout_tty(void)
# Line 284  is_stdout_tty(void) Line 391  is_stdout_tty(void)
391  return isatty(fileno(stdout));  return isatty(fileno(stdout));
392  }  }
393    
394    static BOOL
395    is_file_tty(FILE *f)
396    {
397    return isatty(fileno(f));
398    }
399    
400    
401  /************* Directory scanning in Win32 ***********/  /************* Directory scanning in Win32 ***********/
402    
403  /* 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
404  Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES  Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES
405  when it did not exist. */  when it did not exist. David Byron added a patch that moved the #include of
406    <windows.h> to before the INVALID_FILE_ATTRIBUTES definition rather than after.
407    The double test below stops gcc 4.4.4 grumbling that HAVE_WINDOWS_H is
408    undefined when it is indeed undefined. */
409    
410  #elif HAVE_WIN32API  #elif defined HAVE_WINDOWS_H && HAVE_WINDOWS_H
411    
412  #ifndef STRICT  #ifndef STRICT
413  # define STRICT  # define STRICT
# Line 300  when it did not exist. */ Line 415  when it did not exist. */
415  #ifndef WIN32_LEAN_AND_MEAN  #ifndef WIN32_LEAN_AND_MEAN
416  # define WIN32_LEAN_AND_MEAN  # define WIN32_LEAN_AND_MEAN
417  #endif  #endif
418    
419    #include <windows.h>
420    
421  #ifndef INVALID_FILE_ATTRIBUTES  #ifndef INVALID_FILE_ATTRIBUTES
422  #define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF  #define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF
423  #endif  #endif
424    
 #include <windows.h>  
   
425  typedef struct directory_type  typedef struct directory_type
426  {  {
427  HANDLE handle;  HANDLE handle;
# Line 335  dir = (directory_type *) malloc(sizeof(* Line 451  dir = (directory_type *) malloc(sizeof(*
451  if ((pattern == NULL) || (dir == NULL))  if ((pattern == NULL) || (dir == NULL))
452    {    {
453    fprintf(stderr, "pcregrep: malloc failed\n");    fprintf(stderr, "pcregrep: malloc failed\n");
454    exit(2);    pcregrep_exit(2);
455    }    }
456  memcpy(pattern, filename, len);  memcpy(pattern, filename, len);
457  memcpy(&(pattern[len]), "\\*", 3);  memcpy(&(pattern[len]), "\\*", 3);
# Line 390  regular if they are not directories. */ Line 506  regular if they are not directories. */
506    
507  int isregfile(char *filename)  int isregfile(char *filename)
508  {  {
509  return !isdirectory(filename)  return !isdirectory(filename);
510  }  }
511    
512    
513  /************* Test stdout for being a terminal in Win32 **********/  /************* Test for a terminal in Win32 **********/
514    
515  /* I don't know how to do this; assume never */  /* I don't know how to do this; assume never */
516    
517  static BOOL  static BOOL
518  is_stdout_tty(void)  is_stdout_tty(void)
519  {  {
520  FALSE;  return FALSE;
521    }
522    
523    static BOOL
524    is_file_tty(FILE *f)
525    {
526    return FALSE;
527  }  }
528    
529    
# Line 414  FALSE; Line 536  FALSE;
536  typedef void directory_type;  typedef void directory_type;
537    
538  int isdirectory(char *filename) { return 0; }  int isdirectory(char *filename) { return 0; }
539  directory_type * opendirectory(char *filename) {}  directory_type * opendirectory(char *filename) { return (directory_type*)0;}
540  char *readdirectory(directory_type *dir) {}  char *readdirectory(directory_type *dir) { return (char*)0;}
541  void closedirectory(directory_type *dir) {}  void closedirectory(directory_type *dir) {}
542    
543    
# Line 426  void closedirectory(directory_type *dir) Line 548  void closedirectory(directory_type *dir)
548  int isregfile(char *filename) { return 1; }  int isregfile(char *filename) { return 1; }
549    
550    
551  /************* Test stdout for being a terminal when we can't do it **********/  /************* Test for a terminal when we can't do it **********/
552    
553  static BOOL  static BOOL
554  is_stdout_tty(void)  is_stdout_tty(void)
# Line 434  is_stdout_tty(void) Line 556  is_stdout_tty(void)
556  return FALSE;  return FALSE;
557  }  }
558    
559    static BOOL
560    is_file_tty(FILE *f)
561    {
562    return FALSE;
563    }
564    
565  #endif  #endif
566    
567    
568    
569  #if ! HAVE_STRERROR  #ifndef HAVE_STRERROR
570  /*************************************************  /*************************************************
571  *     Provide strerror() for non-ANSI libraries  *  *     Provide strerror() for non-ANSI libraries  *
572  *************************************************/  *************************************************/
# Line 462  return sys_errlist[n]; Line 589  return sys_errlist[n];
589    
590    
591  /*************************************************  /*************************************************
592    *            Read one line of input              *
593    *************************************************/
594    
595    /* Normally, input is read using fread() into a large buffer, so many lines may
596    be read at once. However, doing this for tty input means that no output appears
597    until a lot of input has been typed. Instead, tty input is handled line by
598    line. We cannot use fgets() for this, because it does not stop at a binary
599    zero, and therefore there is no way of telling how many characters it has read,
600    because there may be binary zeros embedded in the data.
601    
602    Arguments:
603      buffer     the buffer to read into
604      length     the maximum number of characters to read
605      f          the file
606    
607    Returns:     the number of characters read, zero at end of file
608    */
609    
610    static int
611    read_one_line(char *buffer, int length, FILE *f)
612    {
613    int c;
614    int yield = 0;
615    while ((c = fgetc(f)) != EOF)
616      {
617      buffer[yield++] = c;
618      if (c == '\n' || yield >= length) break;
619      }
620    return yield;
621    }
622    
623    
624    
625    /*************************************************
626    *             Find end of line                   *
627    *************************************************/
628    
629    /* The length of the endline sequence that is found is set via lenptr. This may
630    be zero at the very end of the file if there is no line-ending sequence there.
631    
632    Arguments:
633      p         current position in line
634      endptr    end of available data
635      lenptr    where to put the length of the eol sequence
636    
637    Returns:    pointer to the last byte of the line, including the newline byte(s)
638    */
639    
640    static char *
641    end_of_line(char *p, char *endptr, int *lenptr)
642    {
643    switch(endlinetype)
644      {
645      default:      /* Just in case */
646      case EL_LF:
647      while (p < endptr && *p != '\n') p++;
648      if (p < endptr)
649        {
650        *lenptr = 1;
651        return p + 1;
652        }
653      *lenptr = 0;
654      return endptr;
655    
656      case EL_CR:
657      while (p < endptr && *p != '\r') p++;
658      if (p < endptr)
659        {
660        *lenptr = 1;
661        return p + 1;
662        }
663      *lenptr = 0;
664      return endptr;
665    
666      case EL_CRLF:
667      for (;;)
668        {
669        while (p < endptr && *p != '\r') p++;
670        if (++p >= endptr)
671          {
672          *lenptr = 0;
673          return endptr;
674          }
675        if (*p == '\n')
676          {
677          *lenptr = 2;
678          return p + 1;
679          }
680        }
681      break;
682    
683      case EL_ANYCRLF:
684      while (p < endptr)
685        {
686        int extra = 0;
687        register int c = *((unsigned char *)p);
688    
689        if (utf8 && c >= 0xc0)
690          {
691          int gcii, gcss;
692          extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
693          gcss = 6*extra;
694          c = (c & utf8_table3[extra]) << gcss;
695          for (gcii = 1; gcii <= extra; gcii++)
696            {
697            gcss -= 6;
698            c |= (p[gcii] & 0x3f) << gcss;
699            }
700          }
701    
702        p += 1 + extra;
703    
704        switch (c)
705          {
706          case 0x0a:    /* LF */
707          *lenptr = 1;
708          return p;
709    
710          case 0x0d:    /* CR */
711          if (p < endptr && *p == 0x0a)
712            {
713            *lenptr = 2;
714            p++;
715            }
716          else *lenptr = 1;
717          return p;
718    
719          default:
720          break;
721          }
722        }   /* End of loop for ANYCRLF case */
723    
724      *lenptr = 0;  /* Must have hit the end */
725      return endptr;
726    
727      case EL_ANY:
728      while (p < endptr)
729        {
730        int extra = 0;
731        register int c = *((unsigned char *)p);
732    
733        if (utf8 && 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 |= (p[gcii] & 0x3f) << gcss;
743            }
744          }
745    
746        p += 1 + extra;
747    
748        switch (c)
749          {
750          case 0x0a:    /* LF */
751          case 0x0b:    /* VT */
752          case 0x0c:    /* FF */
753          *lenptr = 1;
754          return p;
755    
756          case 0x0d:    /* CR */
757          if (p < endptr && *p == 0x0a)
758            {
759            *lenptr = 2;
760            p++;
761            }
762          else *lenptr = 1;
763          return p;
764    
765          case 0x85:    /* NEL */
766          *lenptr = utf8? 2 : 1;
767          return p;
768    
769          case 0x2028:  /* LS */
770          case 0x2029:  /* PS */
771          *lenptr = 3;
772          return p;
773    
774          default:
775          break;
776          }
777        }   /* End of loop for ANY case */
778    
779      *lenptr = 0;  /* Must have hit the end */
780      return endptr;
781      }     /* End of overall switch */
782    }
783    
784    
785    
786    /*************************************************
787    *         Find start of previous line            *
788    *************************************************/
789    
790    /* This is called when looking back for before lines to print.
791    
792    Arguments:
793      p         start of the subsequent line
794      startptr  start of available data
795    
796    Returns:    pointer to the start of the previous line
797    */
798    
799    static char *
800    previous_line(char *p, char *startptr)
801    {
802    switch(endlinetype)
803      {
804      default:      /* Just in case */
805      case EL_LF:
806      p--;
807      while (p > startptr && p[-1] != '\n') p--;
808      return p;
809    
810      case EL_CR:
811      p--;
812      while (p > startptr && p[-1] != '\n') p--;
813      return p;
814    
815      case EL_CRLF:
816      for (;;)
817        {
818        p -= 2;
819        while (p > startptr && p[-1] != '\n') p--;
820        if (p <= startptr + 1 || p[-2] == '\r') return p;
821        }
822      return p;   /* But control should never get here */
823    
824      case EL_ANY:
825      case EL_ANYCRLF:
826      if (*(--p) == '\n' && p > startptr && p[-1] == '\r') p--;
827      if (utf8) while ((*p & 0xc0) == 0x80) p--;
828    
829      while (p > startptr)
830        {
831        register int c;
832        char *pp = p - 1;
833    
834        if (utf8)
835          {
836          int extra = 0;
837          while ((*pp & 0xc0) == 0x80) pp--;
838          c = *((unsigned char *)pp);
839          if (c >= 0xc0)
840            {
841            int gcii, gcss;
842            extra = utf8_table4[c & 0x3f];  /* Number of additional bytes */
843            gcss = 6*extra;
844            c = (c & utf8_table3[extra]) << gcss;
845            for (gcii = 1; gcii <= extra; gcii++)
846              {
847              gcss -= 6;
848              c |= (pp[gcii] & 0x3f) << gcss;
849              }
850            }
851          }
852        else c = *((unsigned char *)pp);
853    
854        if (endlinetype == EL_ANYCRLF) switch (c)
855          {
856          case 0x0a:    /* LF */
857          case 0x0d:    /* CR */
858          return p;
859    
860          default:
861          break;
862          }
863    
864        else switch (c)
865          {
866          case 0x0a:    /* LF */
867          case 0x0b:    /* VT */
868          case 0x0c:    /* FF */
869          case 0x0d:    /* CR */
870          case 0x85:    /* NEL */
871          case 0x2028:  /* LS */
872          case 0x2029:  /* PS */
873          return p;
874    
875          default:
876          break;
877          }
878    
879        p = pp;  /* Back one character */
880        }        /* End of loop for ANY case */
881    
882      return startptr;  /* Hit start of data */
883      }     /* End of overall switch */
884    }
885    
886    
887    
888    
889    
890    /*************************************************
891  *       Print the previous "after" lines         *  *       Print the previous "after" lines         *
892  *************************************************/  *************************************************/
893    
# Line 486  if (after_context > 0 && lastmatchnumber Line 912  if (after_context > 0 && lastmatchnumber
912    int count = 0;    int count = 0;
913    while (lastmatchrestart < endptr && count++ < after_context)    while (lastmatchrestart < endptr && count++ < after_context)
914      {      {
915        int ellength;
916      char *pp = lastmatchrestart;      char *pp = lastmatchrestart;
917      if (printname != NULL) fprintf(stdout, "%s-", printname);      if (printname != NULL) fprintf(stdout, "%s-", printname);
918      if (number) fprintf(stdout, "%d-", lastmatchnumber++);      if (number) fprintf(stdout, "%d-", lastmatchnumber++);
919      while (*pp != '\n') pp++;      pp = end_of_line(pp, endptr, &ellength);
920      fwrite(lastmatchrestart, 1, pp - lastmatchrestart + 1, stdout);      FWRITE(lastmatchrestart, 1, pp - lastmatchrestart, stdout);
921      lastmatchrestart = pp + 1;      lastmatchrestart = pp;
922      }      }
923    hyphenpending = TRUE;    hyphenpending = TRUE;
924    }    }
# Line 500  if (after_context > 0 && lastmatchnumber Line 927  if (after_context > 0 && lastmatchnumber
927    
928    
929  /*************************************************  /*************************************************
930    *   Apply patterns to subject till one matches   *
931    *************************************************/
932    
933    /* This function is called to run through all patterns, looking for a match. It
934    is used multiple times for the same subject when colouring is enabled, in order
935    to find all possible matches.
936    
937    Arguments:
938      matchptr    the start of the subject
939      length      the length of the subject to match
940      offsets     the offets vector to fill in
941      mrc         address of where to put the result of pcre_exec()
942    
943    Returns:      TRUE if there was a match
944                  FALSE if there was no match
945                  invert if there was a non-fatal error
946    */
947    
948    static BOOL
949    match_patterns(char *matchptr, size_t length, int *offsets, int *mrc)
950    {
951    int i;
952    size_t slen = length;
953    const char *msg = "this text:\n\n";
954    if (slen > 200)
955      {
956      slen = 200;
957      msg = "text that starts:\n\n";
958      }
959    for (i = 0; i < pattern_count; i++)
960      {
961      *mrc = pcre_exec(pattern_list[i], hints_list[i], matchptr, (int)length, 0,
962        PCRE_NOTEMPTY, offsets, OFFSET_SIZE);
963      if (*mrc >= 0) return TRUE;
964      if (*mrc == PCRE_ERROR_NOMATCH) continue;
965      fprintf(stderr, "pcregrep: pcre_exec() gave error %d while matching ", *mrc);
966      if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1);
967      fprintf(stderr, "%s", msg);
968      FWRITE(matchptr, 1, slen, stderr);   /* In case binary zero included */
969      fprintf(stderr, "\n\n");
970      if (*mrc == PCRE_ERROR_MATCHLIMIT || *mrc == PCRE_ERROR_RECURSIONLIMIT)
971        resource_error = TRUE;
972      if (error_count++ > 20)
973        {
974        fprintf(stderr, "pcregrep: Too many errors - abandoned.\n");
975        pcregrep_exit(2);
976        }
977      return invert;    /* No more matching; don't show the line again */
978      }
979    
980    return FALSE;  /* No match, no errors */
981    }
982    
983    
984    
985    /*************************************************
986  *            Grep an individual file             *  *            Grep an individual file             *
987  *************************************************/  *************************************************/
988    
# Line 511  be in the middle third most of the time, Line 994  be in the middle third most of the time,
994  "before" context printing.  "before" context printing.
995    
996  Arguments:  Arguments:
997    in           the fopened FILE stream    handle       the fopened FILE stream for a normal file
998                   the gzFile pointer when reading is via libz
999                   the BZFILE pointer when reading is via libbz2
1000      frtype       FR_PLAIN, FR_LIBZ, or FR_LIBBZ2
1001    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
1002                 or NULL if the file name is not to be printed                 or NULL if the file name is not to be printed
1003                 it cannot be NULL if filenames[_nomatch]_only is set                 it cannot be NULL if filenames[_nomatch]_only is set
1004    
1005  Returns:       0 if there was at least one match  Returns:       0 if there was at least one match
1006                 1 otherwise (no matches)                 1 otherwise (no matches)
1007                   2 if there is a read error on a .bz2 file
1008  */  */
1009    
1010  static int  static int
1011  pcregrep(FILE *in, char *printname)  pcregrep(void *handle, int frtype, char *printname)
1012  {  {
1013  int rc = 1;  int rc = 1;
1014  int linenumber = 1;  int linenumber = 1;
1015  int lastmatchnumber = 0;  int lastmatchnumber = 0;
1016  int count = 0;  int count = 0;
1017  int offsets[99];  int filepos = 0;
1018    int offsets[OFFSET_SIZE];
1019  char *lastmatchrestart = NULL;  char *lastmatchrestart = NULL;
1020  char buffer[3*MBUFTHIRD];  char buffer[3*MBUFTHIRD];
1021  char *ptr = buffer;  char *ptr = buffer;
1022  char *endptr;  char *endptr;
1023  size_t bufflength;  size_t bufflength;
1024  BOOL endhyphenpending = FALSE;  BOOL endhyphenpending = FALSE;
1025    BOOL input_line_buffered = line_buffered;
1026    FILE *in = NULL;                    /* Ensure initialized */
1027    
1028    #ifdef SUPPORT_LIBZ
1029    gzFile ingz = NULL;
1030    #endif
1031    
1032    #ifdef SUPPORT_LIBBZ2
1033    BZFILE *inbz2 = NULL;
1034    #endif
1035    
1036    
1037  /* Do the first read into the start of the buffer and set up the pointer to  /* Do the first read into the start of the buffer and set up the pointer to end
1038  end of what we have. */  of what we have. In the case of libz, a non-zipped .gz file will be read as a
1039    plain file. However, if a .bz2 file isn't actually bzipped, the first read will
1040    fail. */
1041    
1042    #ifdef SUPPORT_LIBZ
1043    if (frtype == FR_LIBZ)
1044      {
1045      ingz = (gzFile)handle;
1046      bufflength = gzread (ingz, buffer, 3*MBUFTHIRD);
1047      }
1048    else
1049    #endif
1050    
1051    #ifdef SUPPORT_LIBBZ2
1052    if (frtype == FR_LIBBZ2)
1053      {
1054      inbz2 = (BZFILE *)handle;
1055      bufflength = BZ2_bzread(inbz2, buffer, 3*MBUFTHIRD);
1056      if ((int)bufflength < 0) return 2;   /* Gotcha: bufflength is size_t; */
1057      }                                    /* without the cast it is unsigned. */
1058    else
1059    #endif
1060    
1061      {
1062      in = (FILE *)handle;
1063      if (is_file_tty(in)) input_line_buffered = TRUE;
1064      bufflength = input_line_buffered?
1065        read_one_line(buffer, 3*MBUFTHIRD, in) :
1066        fread(buffer, 1, 3*MBUFTHIRD, in);
1067      }
1068    
 bufflength = fread(buffer, 1, 3*MBUFTHIRD, in);  
1069  endptr = buffer + bufflength;  endptr = buffer + bufflength;
1070    
1071  /* 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 548  way, the buffer is shifted left and re-f Line 1075  way, the buffer is shifted left and re-f
1075    
1076  while (ptr < endptr)  while (ptr < endptr)
1077    {    {
1078    int i;    int endlinelength;
1079    int mrc = 0;    int mrc = 0;
1080    BOOL match = FALSE;    BOOL match;
1081      char *matchptr = ptr;
1082    char *t = ptr;    char *t = ptr;
1083    size_t length, linelength;    size_t length, linelength;
1084    
1085    /* At this point, ptr is at the start of a line. We need to find the length    /* At this point, ptr is at the start of a line. We need to find the length
1086    of the subject string to pass to pcre_exec(). In multiline mode, it is the    of the subject string to pass to pcre_exec(). In multiline mode, it is the
1087    length remainder of the data in the buffer. Otherwise, it is the length of    length remainder of the data in the buffer. Otherwise, it is the length of
1088    the next line. After matching, we always advance by the length of the next    the next line, excluding the terminating newline. After matching, we always
1089    line. In multiline mode the PCRE_FIRSTLINE option is used for compiling, so    advance by the length of the next line. In multiline mode the PCRE_FIRSTLINE
1090    that any match is constrained to be in the first line. */    option is used for compiling, so that any match is constrained to be in the
1091      first line. */
1092    linelength = 0;  
1093    while (t < endptr && *t++ != '\n') linelength++;    t = end_of_line(t, endptr, &endlinelength);
1094    length = multiline? endptr - ptr : linelength;    linelength = t - ptr - endlinelength;
1095      length = multiline? (size_t)(endptr - ptr) : linelength;
1096    /* Run through all the patterns until one matches. Note that we don't include  
1097    the final newline in the subject string. */    /* Extra processing for Jeffrey Friedl's debugging. */
1098    
1099    for (i = 0; i < pattern_count; i++)  #ifdef JFRIEDL_DEBUG
1100      {    if (jfriedl_XT || jfriedl_XR)
1101      mrc = pcre_exec(pattern_list[i], hints_list[i], ptr, length, 0, 0,    {
1102        offsets, 99);        #include <sys/time.h>
1103      if (mrc >= 0) { match = TRUE; break; }        #include <time.h>
1104      if (mrc != PCRE_ERROR_NOMATCH)        struct timeval start_time, end_time;
1105        {        struct timezone dummy;
1106        fprintf(stderr, "pcregrep: pcre_exec() error %d while matching ", mrc);        int i;
1107        if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1);  
1108        fprintf(stderr, "this line:\n");        if (jfriedl_XT)
1109        fwrite(ptr, 1, linelength, stderr);   /* In case binary zero included */        {
1110        fprintf(stderr, "\n");            unsigned long newlen = length * jfriedl_XT + strlen(jfriedl_prefix) + strlen(jfriedl_postfix);
1111        if (error_count == 0 &&            const char *orig = ptr;
1112            (mrc == PCRE_ERROR_MATCHLIMIT || mrc == PCRE_ERROR_RECURSIONLIMIT))            ptr = malloc(newlen + 1);
1113          {            if (!ptr) {
1114          fprintf(stderr, "pcregrep: error %d means that a resource limit "                    printf("out of memory");
1115            "was exceeded\n", mrc);                    pcregrep_exit(2);
1116          fprintf(stderr, "pcregrep: check your regex for nested unlimited loops\n");            }
1117          }            endptr = ptr;
1118        if (error_count++ > 20)            strcpy(endptr, jfriedl_prefix); endptr += strlen(jfriedl_prefix);
1119          {            for (i = 0; i < jfriedl_XT; i++) {
1120          fprintf(stderr, "pcregrep: too many errors - abandoned\n");                    strncpy(endptr, orig,  length);
1121          exit(2);                    endptr += length;
1122          }            }
1123        match = invert;    /* No more matching; don't show the line again */            strcpy(endptr, jfriedl_postfix); endptr += strlen(jfriedl_postfix);
1124        break;            length = newlen;
1125        }        }
1126      }  
1127          if (gettimeofday(&start_time, &dummy) != 0)
1128                  perror("bad gettimeofday");
1129    
1130    
1131          for (i = 0; i < jfriedl_XR; i++)
1132              match = (pcre_exec(pattern_list[0], hints_list[0], ptr, length, 0,
1133                  PCRE_NOTEMPTY, offsets, OFFSET_SIZE) >= 0);
1134    
1135          if (gettimeofday(&end_time, &dummy) != 0)
1136                  perror("bad gettimeofday");
1137    
1138          double delta = ((end_time.tv_sec + (end_time.tv_usec / 1000000.0))
1139                          -
1140                          (start_time.tv_sec + (start_time.tv_usec / 1000000.0)));
1141    
1142          printf("%s TIMER[%.4f]\n", match ? "MATCH" : "FAIL", delta);
1143          return 0;
1144      }
1145    #endif
1146    
1147      /* We come back here after a match when the -o option (only_matching) is set,
1148      in order to find any further matches in the same line. */
1149    
1150      ONLY_MATCHING_RESTART:
1151    
1152      /* Run through all the patterns until one matches or there is an error other
1153      than NOMATCH. This code is in a subroutine so that it can be re-used for
1154      finding subsequent matches when colouring matched lines. */
1155    
1156      match = match_patterns(matchptr, length, offsets, &mrc);
1157    
1158    /* If it's a match or a not-match (as required), do what's wanted. */    /* If it's a match or a not-match (as required), do what's wanted. */
1159    
# Line 614  while (ptr < endptr) Line 1172  while (ptr < endptr)
1172      /* If all we want is a file name, there is no need to scan any more lines      /* If all we want is a file name, there is no need to scan any more lines
1173      in the file. */      in the file. */
1174    
1175      else if (filenames == FN_ONLY)      else if (filenames == FN_MATCH_ONLY)
1176        {        {
1177        fprintf(stdout, "%s\n", printname);        fprintf(stdout, "%s\n", printname);
1178        return 0;        return 0;
# Line 624  while (ptr < endptr) Line 1182  while (ptr < endptr)
1182    
1183      else if (quiet) return 0;      else if (quiet) return 0;
1184    
1185      /* The --only-matching option prints just the substring that matched, and      /* The --only-matching option prints just the substring that matched, or a
1186      does not pring any context. */      captured portion of it, as long as this string is not empty, and the
1187        --file-offsets and --line-offsets options output offsets for the matching
1188        substring (they both force --only-matching = 0). None of these options
1189        prints any context. Afterwards, adjust the start and length, and then jump
1190        back to look for further matches in the same line. If we are in invert
1191        mode, however, nothing is printed and we do not restart - this could still
1192        be useful because the return code is set. */
1193    
1194      else if (only_matching)      else if (only_matching >= 0)
1195        {        {
1196        if (printname != NULL) fprintf(stdout, "%s:", printname);        if (!invert)
1197        if (number) fprintf(stdout, "%d:", linenumber);          {
1198        fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout);          if (printname != NULL) fprintf(stdout, "%s:", printname);
1199        fprintf(stdout, "\n");          if (number) fprintf(stdout, "%d:", linenumber);
1200            if (line_offsets)
1201              fprintf(stdout, "%d,%d\n", (int)(matchptr + offsets[0] - ptr),
1202                offsets[1] - offsets[0]);
1203            else if (file_offsets)
1204              fprintf(stdout, "%d,%d\n",
1205                (int)(filepos + matchptr + offsets[0] - ptr),
1206                offsets[1] - offsets[0]);
1207            else if (only_matching < mrc)
1208              {
1209              int plen = offsets[2*only_matching + 1] - offsets[2*only_matching];
1210              if (plen > 0)
1211                {
1212                if (do_colour) fprintf(stdout, "%c[%sm", 0x1b, colour_string);
1213                FWRITE(matchptr + offsets[only_matching*2], 1, plen, stdout);
1214                if (do_colour) fprintf(stdout, "%c[00m", 0x1b);
1215                fprintf(stdout, "\n");
1216                }
1217              }
1218            else if (printname != NULL || number) fprintf(stdout, "\n");
1219            matchptr += offsets[1];
1220            length -= offsets[1];
1221            match = FALSE;
1222            if (line_buffered) fflush(stdout);
1223            rc = 0;    /* Had some success */
1224            goto ONLY_MATCHING_RESTART;
1225            }
1226        }        }
1227    
1228      /* This is the default case when none of the above options is set. We print      /* This is the default case when none of the above options is set. We print
# Line 646  while (ptr < endptr) Line 1236  while (ptr < endptr)
1236    
1237        if (after_context > 0 && lastmatchnumber > 0)        if (after_context > 0 && lastmatchnumber > 0)
1238          {          {
1239            int ellength;
1240          int linecount = 0;          int linecount = 0;
1241          char *p = lastmatchrestart;          char *p = lastmatchrestart;
1242    
1243          while (p < ptr && linecount < after_context)          while (p < ptr && linecount < after_context)
1244            {            {
1245            while (*p != '\n') p++;            p = end_of_line(p, ptr, &ellength);
           p++;  
1246            linecount++;            linecount++;
1247            }            }
1248    
# Line 665  while (ptr < endptr) Line 1255  while (ptr < endptr)
1255            char *pp = lastmatchrestart;            char *pp = lastmatchrestart;
1256            if (printname != NULL) fprintf(stdout, "%s-", printname);            if (printname != NULL) fprintf(stdout, "%s-", printname);
1257            if (number) fprintf(stdout, "%d-", lastmatchnumber++);            if (number) fprintf(stdout, "%d-", lastmatchnumber++);
1258            while (*pp != '\n') pp++;            pp = end_of_line(pp, endptr, &ellength);
1259            fwrite(lastmatchrestart, 1, pp - lastmatchrestart + 1, stdout);            FWRITE(lastmatchrestart, 1, pp - lastmatchrestart, stdout);
1260            lastmatchrestart = pp + 1;            lastmatchrestart = pp;
1261            }            }
1262          if (lastmatchrestart != ptr) hyphenpending = TRUE;          if (lastmatchrestart != ptr) hyphenpending = TRUE;
1263          }          }
# Line 693  while (ptr < endptr) Line 1283  while (ptr < endptr)
1283                 linecount < before_context)                 linecount < before_context)
1284            {            {
1285            linecount++;            linecount++;
1286            p--;            p = previous_line(p, buffer);
           while (p > buffer && p[-1] != '\n') p--;  
1287            }            }
1288    
1289          if (lastmatchnumber > 0 && p > lastmatchrestart && !hyphenprinted)          if (lastmatchnumber > 0 && p > lastmatchrestart && !hyphenprinted)
# Line 702  while (ptr < endptr) Line 1291  while (ptr < endptr)
1291    
1292          while (p < ptr)          while (p < ptr)
1293            {            {
1294              int ellength;
1295            char *pp = p;            char *pp = p;
1296            if (printname != NULL) fprintf(stdout, "%s-", printname);            if (printname != NULL) fprintf(stdout, "%s-", printname);
1297            if (number) fprintf(stdout, "%d-", linenumber - linecount--);            if (number) fprintf(stdout, "%d-", linenumber - linecount--);
1298            while (*pp != '\n') pp++;            pp = end_of_line(pp, endptr, &ellength);
1299            fwrite(p, 1, pp - p + 1, stdout);   /* In case binary zero */            FWRITE(p, 1, pp - p, stdout);
1300            p = pp + 1;            p = pp;
1301            }            }
1302          }          }
1303    
# Line 722  while (ptr < endptr) Line 1312  while (ptr < endptr)
1312    
1313        /* 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
1314        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
1315        line number appropriately. Because the PCRE_FIRSTLINE option is set, the        line number appropriately, but only when there actually was a match
1316        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
1317          the match will always be before the first newline sequence. */
1318    
1319        if (multiline)        if (multiline & !invert)
1320          {          {
1321          char *endmatch = ptr + offsets[1];          char *endmatch = ptr + offsets[1];
1322          t = ptr;          t = ptr;
1323          while (t < endmatch) { if (*t++ == '\n') linenumber++; }          while (t < endmatch)
1324          while (endmatch < endptr && *endmatch != '\n') endmatch++;            {
1325          linelength = endmatch - ptr;            t = end_of_line(t, endptr, &endlinelength);
1326              if (t < endmatch) linenumber++; else break;
1327              }
1328            linelength = t - ptr - endlinelength;
1329          }          }
1330    
1331        /*** NOTE: Use only fwrite() to output the data line, so that binary        /*** NOTE: Use only fwrite() to output the data line, so that binary
# Line 746  while (ptr < endptr) Line 1340  while (ptr < endptr)
1340          {          {
1341          int first = S_arg * 2;          int first = S_arg * 2;
1342          int last  = first + 1;          int last  = first + 1;
1343          fwrite(ptr, 1, offsets[first], stdout);          FWRITE(ptr, 1, offsets[first], stdout);
1344          fprintf(stdout, "X");          fprintf(stdout, "X");
1345          fwrite(ptr + offsets[last], 1, linelength - offsets[last], stdout);          FWRITE(ptr + offsets[last], 1, linelength - offsets[last], stdout);
1346          }          }
1347        else        else
1348  #endif  #endif
1349    
1350        /* We have to split the line(s) up if colouring. */        /* We have to split the line(s) up if colouring, and search for further
1351          matches, but not of course if the line is a non-match. */
1352        if (do_colour)  
1353          if (do_colour && !invert)
1354          {          {
1355          fwrite(ptr, 1, offsets[0], stdout);          int plength;
1356            int last_offset = 0;
1357            FWRITE(ptr, 1, offsets[0], stdout);
1358          fprintf(stdout, "%c[%sm", 0x1b, colour_string);          fprintf(stdout, "%c[%sm", 0x1b, colour_string);
1359          fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout);          FWRITE(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
1360          fprintf(stdout, "%c[00m", 0x1b);          fprintf(stdout, "%c[00m", 0x1b);
1361          fwrite(ptr + offsets[1], 1, linelength - offsets[1], stdout);          for (;;)
1362              {
1363              last_offset += offsets[1];
1364              matchptr += offsets[1];
1365              length -= offsets[1];
1366              if (!match_patterns(matchptr, length, offsets, &mrc)) break;
1367              FWRITE(matchptr, 1, offsets[0], stdout);
1368              fprintf(stdout, "%c[%sm", 0x1b, colour_string);
1369              FWRITE(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout);
1370              fprintf(stdout, "%c[00m", 0x1b);
1371              }
1372    
1373            /* In multiline mode, we may have already printed the complete line
1374            and its line-ending characters (if they matched the pattern), so there
1375            may be no more to print. */
1376    
1377            plength = (linelength + endlinelength) - last_offset;
1378            if (plength > 0)
1379              FWRITE(ptr + last_offset, 1, plength, stdout);
1380          }          }
       else fwrite(ptr, 1, linelength, stdout);  
1381    
1382        fprintf(stdout, "\n");        /* Not colouring; no need to search for further matches */
1383    
1384          else FWRITE(ptr, 1, linelength + endlinelength, stdout);
1385        }        }
1386    
1387      /* End of doing what has to be done for a match */      /* End of doing what has to be done for a match. If --line-buffered was
1388        given, flush the output. */
1389    
1390        if (line_buffered) fflush(stdout);
1391      rc = 0;    /* Had some success */      rc = 0;    /* Had some success */
1392    
1393      /* Remember where the last match happened for after_context. We remember      /* Remember where the last match happened for after_context. We remember
1394      where we are about to restart, and that line's number. */      where we are about to restart, and that line's number. */
1395    
1396      lastmatchrestart = ptr + linelength + 1;      lastmatchrestart = ptr + linelength + endlinelength;
1397      lastmatchnumber = linenumber + 1;      lastmatchnumber = linenumber + 1;
1398      }      }
1399    
1400    /* Advance to after the newline and increment the line number. */    /* For a match in multiline inverted mode (which of course did not cause
1401      anything to be printed), we have to move on to the end of the match before
1402      proceeding. */
1403    
1404      if (multiline && invert && match)
1405        {
1406        int ellength;
1407        char *endmatch = ptr + offsets[1];
1408        t = ptr;
1409        while (t < endmatch)
1410          {
1411          t = end_of_line(t, endptr, &ellength);
1412          if (t <= endmatch) linenumber++; else break;
1413          }
1414        endmatch = end_of_line(endmatch, endptr, &ellength);
1415        linelength = endmatch - ptr - ellength;
1416        }
1417    
1418      /* Advance to after the newline and increment the line number. The file
1419      offset to the current line is maintained in filepos. */
1420    
1421    ptr += linelength + 1;    ptr += linelength + endlinelength;
1422      filepos += (int)(linelength + endlinelength);
1423    linenumber++;    linenumber++;
1424    
1425      /* If input is line buffered, and the buffer is not yet full, read another
1426      line and add it into the buffer. */
1427    
1428      if (input_line_buffered && bufflength < sizeof(buffer))
1429        {
1430        int add = read_one_line(ptr, sizeof(buffer) - (ptr - buffer), in);
1431        bufflength += add;
1432        endptr += add;
1433        }
1434    
1435    /* 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
1436    the current point is in the top 1/3 of the buffer, slide the buffer down by    the current point is in the top 1/3 of the buffer, slide the buffer down by
1437    1/3 and refill it. Before we do this, if some unprinted "after" lines are    1/3 and refill it. Before we do this, if some unprinted "after" lines are
# Line 803  while (ptr < endptr) Line 1451  while (ptr < endptr)
1451    
1452      memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD);      memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD);
1453      ptr -= MBUFTHIRD;      ptr -= MBUFTHIRD;
1454      bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in);  
1455    #ifdef SUPPORT_LIBZ
1456        if (frtype == FR_LIBZ)
1457          bufflength = 2*MBUFTHIRD +
1458            gzread (ingz, buffer + 2*MBUFTHIRD, MBUFTHIRD);
1459        else
1460    #endif
1461    
1462    #ifdef SUPPORT_LIBBZ2
1463        if (frtype == FR_LIBBZ2)
1464          bufflength = 2*MBUFTHIRD +
1465            BZ2_bzread(inbz2, buffer + 2*MBUFTHIRD, MBUFTHIRD);
1466        else
1467    #endif
1468    
1469        bufflength = 2*MBUFTHIRD +
1470          (input_line_buffered?
1471           read_one_line(buffer + 2*MBUFTHIRD, MBUFTHIRD, in) :
1472           fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in));
1473      endptr = buffer + bufflength;      endptr = buffer + bufflength;
1474    
1475      /* Adjust any last match point */      /* Adjust any last match point */
# Line 815  while (ptr < endptr) Line 1481  while (ptr < endptr)
1481  /* 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
1482  hyphenpending if it prints something. */  hyphenpending if it prints something. */
1483    
1484  if (!only_matching && !count_only)  if (only_matching < 0 && !count_only)
1485    {    {
1486    do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname);    do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname);
1487    hyphenpending |= endhyphenpending;    hyphenpending |= endhyphenpending;
# Line 834  if (filenames == FN_NOMATCH_ONLY) Line 1500  if (filenames == FN_NOMATCH_ONLY)
1500    
1501  if (count_only)  if (count_only)
1502    {    {
1503    if (printname != NULL) fprintf(stdout, "%s:", printname);    if (count > 0 || !omit_zero_count)
1504    fprintf(stdout, "%d\n", count);      {
1505        if (printname != NULL && filenames != FN_NONE)
1506          fprintf(stdout, "%s:", printname);
1507        fprintf(stdout, "%d\n", count);
1508        }
1509    }    }
1510    
1511  return rc;  return rc;
# Line 867  grep_or_recurse(char *pathname, BOOL dir Line 1537  grep_or_recurse(char *pathname, BOOL dir
1537  {  {
1538  int rc = 1;  int rc = 1;
1539  int sep;  int sep;
1540  FILE *in;  int frtype;
1541    int pathlen;
1542    void *handle;
1543    FILE *in = NULL;           /* Ensure initialized */
1544    
1545    #ifdef SUPPORT_LIBZ
1546    gzFile ingz = NULL;
1547    #endif
1548    
1549    #ifdef SUPPORT_LIBBZ2
1550    BZFILE *inbz2 = NULL;
1551    #endif
1552    
1553  /* If the file name is "-" we scan stdin */  /* If the file name is "-" we scan stdin */
1554    
1555  if (strcmp(pathname, "-") == 0)  if (strcmp(pathname, "-") == 0)
1556    {    {
1557    return pcregrep(stdin,    return pcregrep(stdin, FR_PLAIN,
1558      (filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))?      (filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))?
1559        stdin_name : NULL);        stdin_name : NULL);
1560    }    }
1561    
   
1562  /* If the file is a directory, skip if skipping or if we are recursing, scan  /* If the file is a directory, skip if skipping or if we are recursing, scan
1563  each file within it, subject to any include or exclude patterns that were set.  each file and directory within it, subject to any include or exclude patterns
1564  The scanning code is localized so it can be made system-specific. */  that were set. The scanning code is localized so it can be made
1565    system-specific. */
1566    
1567  if ((sep = isdirectory(pathname)) != 0)  if ((sep = isdirectory(pathname)) != 0)
1568    {    {
# Line 902  if ((sep = isdirectory(pathname)) != 0) Line 1583  if ((sep = isdirectory(pathname)) != 0)
1583    
1584      while ((nextfile = readdirectory(dir)) != NULL)      while ((nextfile = readdirectory(dir)) != NULL)
1585        {        {
1586        int frc, blen;        int frc, nflen;
1587        sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile);        sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile);
1588        blen = strlen(buffer);        nflen = (int)(strlen(nextfile));
1589    
1590        if (exclude_compiled != NULL &&        if (isdirectory(buffer))
1591            pcre_exec(exclude_compiled, NULL, buffer, blen, 0, 0, NULL, 0) >= 0)          {
1592          continue;          if (exclude_dir_compiled != NULL &&
1593                pcre_exec(exclude_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0)
1594        if (include_compiled != NULL &&            continue;
1595            pcre_exec(include_compiled, NULL, buffer, blen, 0, 0, NULL, 0) < 0)  
1596          continue;          if (include_dir_compiled != NULL &&
1597                pcre_exec(include_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0)
1598              continue;
1599            }
1600          else
1601            {
1602            if (exclude_compiled != NULL &&
1603                pcre_exec(exclude_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0)
1604              continue;
1605    
1606            if (include_compiled != NULL &&
1607                pcre_exec(include_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0)
1608              continue;
1609            }
1610    
1611        frc = grep_or_recurse(buffer, dir_recurse, FALSE);        frc = grep_or_recurse(buffer, dir_recurse, FALSE);
1612        if (frc > 1) rc = frc;        if (frc > 1) rc = frc;
# Line 935  skipping was not requested. The scan pro Line 1629  skipping was not requested. The scan pro
1629  argument at top level, we don't show the file name, unless we are only showing  argument at top level, we don't show the file name, unless we are only showing
1630  the file name, or the filename was forced (-H). */  the file name, or the filename was forced (-H). */
1631    
1632  in = fopen(pathname, "r");  pathlen = (int)(strlen(pathname));
1633  if (in == NULL)  
1634    /* Open using zlib if it is supported and the file name ends with .gz. */
1635    
1636    #ifdef SUPPORT_LIBZ
1637    if (pathlen > 3 && strcmp(pathname + pathlen - 3, ".gz") == 0)
1638      {
1639      ingz = gzopen(pathname, "rb");
1640      if (ingz == NULL)
1641        {
1642        if (!silent)
1643          fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
1644            strerror(errno));
1645        return 2;
1646        }
1647      handle = (void *)ingz;
1648      frtype = FR_LIBZ;
1649      }
1650    else
1651    #endif
1652    
1653    /* Otherwise open with bz2lib if it is supported and the name ends with .bz2. */
1654    
1655    #ifdef SUPPORT_LIBBZ2
1656    if (pathlen > 4 && strcmp(pathname + pathlen - 4, ".bz2") == 0)
1657      {
1658      inbz2 = BZ2_bzopen(pathname, "rb");
1659      handle = (void *)inbz2;
1660      frtype = FR_LIBBZ2;
1661      }
1662    else
1663    #endif
1664    
1665    /* Otherwise use plain fopen(). The label is so that we can come back here if
1666    an attempt to read a .bz2 file indicates that it really is a plain file. */
1667    
1668    #ifdef SUPPORT_LIBBZ2
1669    PLAIN_FILE:
1670    #endif
1671      {
1672      in = fopen(pathname, "rb");
1673      handle = (void *)in;
1674      frtype = FR_PLAIN;
1675      }
1676    
1677    /* All the opening methods return errno when they fail. */
1678    
1679    if (handle == NULL)
1680    {    {
1681    if (!silent)    if (!silent)
1682      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,      fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname,
# Line 944  if (in == NULL) Line 1684  if (in == NULL)
1684    return 2;    return 2;
1685    }    }
1686    
1687  rc = pcregrep(in, (filenames > FN_DEFAULT ||  /* Now grep the file */
1688    
1689    rc = pcregrep(handle, frtype, (filenames > FN_DEFAULT ||
1690    (filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL);    (filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL);
1691    
1692    /* Close in an appropriate manner. */
1693    
1694    #ifdef SUPPORT_LIBZ
1695    if (frtype == FR_LIBZ)
1696      gzclose(ingz);
1697    else
1698    #endif
1699    
1700    /* If it is a .bz2 file and the result is 2, it means that the first attempt to
1701    read failed. If the error indicates that the file isn't in fact bzipped, try
1702    again as a normal file. */
1703    
1704    #ifdef SUPPORT_LIBBZ2
1705    if (frtype == FR_LIBBZ2)
1706      {
1707      if (rc == 2)
1708        {
1709        int errnum;
1710        const char *err = BZ2_bzerror(inbz2, &errnum);
1711        if (errnum == BZ_DATA_ERROR_MAGIC)
1712          {
1713          BZ2_bzclose(inbz2);
1714          goto PLAIN_FILE;
1715          }
1716        else if (!silent)
1717          fprintf(stderr, "pcregrep: Failed to read %s using bzlib: %s\n",
1718            pathname, err);
1719        }
1720      BZ2_bzclose(inbz2);
1721      }
1722    else
1723    #endif
1724    
1725    /* Normal file close */
1726    
1727  fclose(in);  fclose(in);
1728    
1729    /* Pass back the yield from pcregrep(). */
1730    
1731  return rc;  return rc;
1732  }  }
1733    
# Line 968  for (op = optionlist; op->one_char != 0; Line 1748  for (op = optionlist; op->one_char != 0;
1748    if (op->one_char > 0) fprintf(stderr, "%c", op->one_char);    if (op->one_char > 0) fprintf(stderr, "%c", op->one_char);
1749    }    }
1750  fprintf(stderr, "] [long options] [pattern] [files]\n");  fprintf(stderr, "] [long options] [pattern] [files]\n");
1751  fprintf(stderr, "Type `pcregrep --help' for more information.\n");  fprintf(stderr, "Type `pcregrep --help' for more information and the long "
1752      "options.\n");
1753  return rc;  return rc;
1754  }  }
1755    
# Line 987  option_item *op; Line 1768  option_item *op;
1768  printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n");  printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n");
1769  printf("Search for PATTERN in each FILE or standard input.\n");  printf("Search for PATTERN in each FILE or standard input.\n");
1770  printf("PATTERN must be present if neither -e nor -f is used.\n");  printf("PATTERN must be present if neither -e nor -f is used.\n");
1771  printf("\"-\" can be used as a file name to mean STDIN.\n\n");  printf("\"-\" can be used as a file name to mean STDIN.\n");
1772  printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n");  
1773    #ifdef SUPPORT_LIBZ
1774    printf("Files whose names end in .gz are read using zlib.\n");
1775    #endif
1776    
1777    #ifdef SUPPORT_LIBBZ2
1778    printf("Files whose names end in .bz2 are read using bzlib2.\n");
1779    #endif
1780    
1781    #if defined SUPPORT_LIBZ || defined SUPPORT_LIBBZ2
1782    printf("Other files and the standard input are read as plain files.\n\n");
1783    #else
1784    printf("All files are read as plain files, without any interpretation.\n\n");
1785    #endif
1786    
1787    printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n");
1788  printf("Options:\n");  printf("Options:\n");
1789    
1790  for (op = optionlist; op->one_char != 0; op++)  for (op = optionlist; op->one_char != 0; op++)
1791    {    {
1792    int n;    int n;
1793    char s[4];    char s[4];
1794    
1795      /* Two options were accidentally implemented and documented with underscores
1796      instead of hyphens in their names, something that was not noticed for quite a
1797      few releases. When fixing this, I left the underscored versions in the list
1798      in case people were using them. However, we don't want to display them in the
1799      help data. There are no other options that contain underscores, and we do not
1800      expect ever to implement such options. Therefore, just omit any option that
1801      contains an underscore. */
1802    
1803      if (strchr(op->long_name, '_') != NULL) continue;
1804    
1805    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, "   ");
1806    printf("  %s --%s%n", s, op->long_name, &n);    n = 31 - printf("  %s --%s", s, op->long_name);
   n = 30 - n;  
1807    if (n < 1) n = 1;    if (n < 1) n = 1;
1808    printf("%.*s%s\n", n, "                    ", op->help_text);    printf("%.*s%s\n", n, "                     ", op->help_text);
1809    }    }
1810    
1811  printf("\nWhen reading patterns from a file instead of using a command line option,\n");  printf("\nWhen reading patterns from a file instead of using a command line option,\n");
# Line 1023  handle_option(int letter, int options) Line 1828  handle_option(int letter, int options)
1828  {  {
1829  switch(letter)  switch(letter)
1830    {    {
1831    case N_HELP: help(); exit(0);    case N_FOFFSETS: file_offsets = TRUE; break;
1832      case N_HELP: help(); pcregrep_exit(0);
1833      case N_LOFFSETS: line_offsets = number = TRUE; break;
1834      case N_LBUFFER: line_buffered = TRUE; break;
1835    case 'c': count_only = TRUE; break;    case 'c': count_only = TRUE; break;
1836    case 'F': process_options |= PO_FIXED_STRINGS; break;    case 'F': process_options |= PO_FIXED_STRINGS; break;
1837    case 'H': filenames = FN_FORCE; break;    case 'H': filenames = FN_FORCE; break;
1838    case 'h': filenames = FN_NONE; break;    case 'h': filenames = FN_NONE; break;
1839    case 'i': options |= PCRE_CASELESS; break;    case 'i': options |= PCRE_CASELESS; break;
1840    case 'l': filenames = FN_ONLY; break;    case 'l': omit_zero_count = TRUE; filenames = FN_MATCH_ONLY; break;
1841    case 'L': filenames = FN_NOMATCH_ONLY; break;    case 'L': filenames = FN_NOMATCH_ONLY; break;
1842    case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break;    case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break;
1843    case 'n': number = TRUE; break;    case 'n': number = TRUE; break;
1844    case 'o': only_matching = TRUE; break;    case 'o': only_matching = 0; break;
1845    case 'q': quiet = TRUE; break;    case 'q': quiet = TRUE; break;
1846    case 'r': dee_action = dee_RECURSE; break;    case 'r': dee_action = dee_RECURSE; break;
1847    case 's': silent = TRUE; break;    case 's': silent = TRUE; break;
1848    case 'u': options |= PCRE_UTF8; break;    case 'u': options |= PCRE_UTF8; utf8 = TRUE; break;
1849    case 'v': invert = TRUE; break;    case 'v': invert = TRUE; break;
1850    case 'w': process_options |= PO_WORD_MATCH; break;    case 'w': process_options |= PO_WORD_MATCH; break;
1851    case 'x': process_options |= PO_LINE_MATCH; break;    case 'x': process_options |= PO_LINE_MATCH; break;
1852    
1853    case 'V':    case 'V':
1854    fprintf(stderr, "pcregrep version %s using ", VERSION);    fprintf(stderr, "pcregrep version %s\n", pcre_version());
1855    fprintf(stderr, "PCRE version %s\n", pcre_version());    pcregrep_exit(0);
   exit(0);  
1856    break;    break;
1857    
1858    default:    default:
1859    fprintf(stderr, "pcregrep: Unknown option -%c\n", letter);    fprintf(stderr, "pcregrep: Unknown option -%c\n", letter);
1860    exit(usage(2));    pcregrep_exit(usage(2));
1861    }    }
1862    
1863  return options;  return options;
# Line 1120  sprintf(buffer, "%s%.*s%s", prefix[proce Line 1927  sprintf(buffer, "%s%.*s%s", prefix[proce
1927    suffix[process_options]);    suffix[process_options]);
1928  pattern_list[pattern_count] =  pattern_list[pattern_count] =
1929    pcre_compile(buffer, options, &error, &errptr, pcretables);    pcre_compile(buffer, options, &error, &errptr, pcretables);
1930  if (pattern_list[pattern_count++] != NULL) return TRUE;  if (pattern_list[pattern_count] != NULL)
1931      {
1932      pattern_count++;
1933      return TRUE;
1934      }
1935    
1936  /* Handle compile errors */  /* Handle compile errors */
1937    
# Line 1152  return FALSE; Line 1963  return FALSE;
1963  *************************************************/  *************************************************/
1964    
1965  /* When the -F option has been used, each string may be a list of strings,  /* When the -F option has been used, each string may be a list of strings,
1966  separated by newlines. They will be matched literally.  separated by line breaks. They will be matched literally.
1967    
1968  Arguments:  Arguments:
1969    pattern        the pattern string    pattern        the pattern string
# Line 1170  compile_pattern(char *pattern, int optio Line 1981  compile_pattern(char *pattern, int optio
1981  {  {
1982  if ((process_options & PO_FIXED_STRINGS) != 0)  if ((process_options & PO_FIXED_STRINGS) != 0)
1983    {    {
1984      char *eop = pattern + strlen(pattern);
1985    char buffer[MBUFTHIRD];    char buffer[MBUFTHIRD];
1986    for(;;)    for(;;)
1987      {      {
1988      char *p = strchr(pattern, '\n');      int ellength;
1989      if (p == NULL)      char *p = end_of_line(pattern, eop, &ellength);
1990        if (ellength == 0)
1991        return compile_single_pattern(pattern, options, filename, count);        return compile_single_pattern(pattern, options, filename, count);
1992      sprintf(buffer, "%.*s", p - pattern, pattern);      sprintf(buffer, "%.*s", (int)(p - pattern - ellength), pattern);
1993      pattern = p + 1;      pattern = p;
1994      if (!compile_single_pattern(buffer, options, filename, count))      if (!compile_single_pattern(buffer, options, filename, count))
1995        return FALSE;        return FALSE;
1996      }      }
# Line 1200  int i, j; Line 2013  int i, j;
2013  int rc = 1;  int rc = 1;
2014  int pcre_options = 0;  int pcre_options = 0;
2015  int cmd_pattern_count = 0;  int cmd_pattern_count = 0;
2016    int hint_count = 0;
2017  int errptr;  int errptr;
2018  BOOL only_one_at_top;  BOOL only_one_at_top;
2019  char *patterns[MAX_PATTERN_COUNT];  char *patterns[MAX_PATTERN_COUNT];
2020  const char *locale_from = "--locale";  const char *locale_from = "--locale";
2021  const char *error;  const char *error;
2022    
2023    /* Set the default line ending value from the default in the PCRE library;
2024    "lf", "cr", "crlf", and "any" are supported. Anything else is treated as "lf".
2025    Note that the return values from pcre_config(), though derived from the ASCII
2026    codes, are the same in EBCDIC environments, so we must use the actual values
2027    rather than escapes such as as '\r'. */
2028    
2029    (void)pcre_config(PCRE_CONFIG_NEWLINE, &i);
2030    switch(i)
2031      {
2032      default:               newline = (char *)"lf"; break;
2033      case 13:               newline = (char *)"cr"; break;
2034      case (13 << 8) | 10:   newline = (char *)"crlf"; break;
2035      case -1:               newline = (char *)"any"; break;
2036      case -2:               newline = (char *)"anycrlf"; break;
2037      }
2038    
2039  /* Process the options */  /* Process the options */
2040    
2041  for (i = 1; i < argc; i++)  for (i = 1; i < argc; i++)
# Line 1223  for (i = 1; i < argc; i++) Line 2053  for (i = 1; i < argc; i++)
2053    if (argv[i][1] == 0)    if (argv[i][1] == 0)
2054      {      {
2055      if (pattern_filename != NULL || pattern_count > 0) break;      if (pattern_filename != NULL || pattern_count > 0) break;
2056        else exit(usage(2));        else pcregrep_exit(usage(2));
2057      }      }
2058    
2059    /* Handle a long name option, or -- to terminate the options */    /* Handle a long name option, or -- to terminate the options */
# Line 1245  for (i = 1; i < argc; i++) Line 2075  for (i = 1; i < argc; i++)
2075      Some options have variations in the long name spelling: specifically, we      Some options have variations in the long name spelling: specifically, we
2076      allow "regexp" because GNU grep allows it, though I personally go along      allow "regexp" because GNU grep allows it, though I personally go along
2077      with Jeffrey Friedl and Larry Wall in preferring "regex" without the "p".      with Jeffrey Friedl and Larry Wall in preferring "regex" without the "p".
2078      These options are entered in the table as "regex(p)". No option is in both      These options are entered in the table as "regex(p)". Options can be in
2079      these categories, fortunately. */      both these categories. */
2080    
2081      for (op = optionlist; op->one_char != 0; op++)      for (op = optionlist; op->one_char != 0; op++)
2082        {        {
2083        char *opbra = strchr(op->long_name, '(');        char *opbra = strchr(op->long_name, '(');
2084        char *equals = strchr(op->long_name, '=');        char *equals = strchr(op->long_name, '=');
2085        if (opbra == NULL)     /* Not a (p) case */  
2086          /* Handle options with only one spelling of the name */
2087    
2088          if (opbra == NULL)     /* Does not contain '(' */
2089          {          {
2090          if (equals == NULL)  /* Not thing=data case */          if (equals == NULL)  /* Not thing=data case */
2091            {            {
# Line 1260  for (i = 1; i < argc; i++) Line 2093  for (i = 1; i < argc; i++)
2093            }            }
2094          else                 /* Special case xxx=data */          else                 /* Special case xxx=data */
2095            {            {
2096            int oplen = equals - op->long_name;            int oplen = (int)(equals - op->long_name);
2097            int arglen = (argequals == NULL)? strlen(arg) : argequals - arg;            int arglen = (argequals == NULL)?
2098                (int)strlen(arg) : (int)(argequals - arg);
2099            if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0)            if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0)
2100              {              {
2101              option_data = arg + arglen;              option_data = arg + arglen;
# Line 1274  for (i = 1; i < argc; i++) Line 2108  for (i = 1; i < argc; i++)
2108              }              }
2109            }            }
2110          }          }
2111        else                   /* Special case xxxx(p) */  
2112          /* Handle options with an alternate spelling of the name */
2113    
2114          else
2115          {          {
2116          char buff1[24];          char buff1[24];
2117          char buff2[24];          char buff2[24];
2118          int baselen = opbra - op->long_name;  
2119            int baselen = (int)(opbra - op->long_name);
2120            int fulllen = (int)(strchr(op->long_name, ')') - op->long_name + 1);
2121            int arglen = (argequals == NULL || equals == NULL)?
2122              (int)strlen(arg) : (int)(argequals - arg);
2123    
2124          sprintf(buff1, "%.*s", baselen, op->long_name);          sprintf(buff1, "%.*s", baselen, op->long_name);
2125          sprintf(buff2, "%s%.*s", buff1, strlen(op->long_name) - baselen - 2,          sprintf(buff2, "%s%.*s", buff1, fulllen - baselen - 2, opbra + 1);
2126            opbra + 1);  
2127          if (strcmp(arg, buff1) == 0 || strcmp(arg, buff2) == 0)          if (strncmp(arg, buff1, arglen) == 0 ||
2128               strncmp(arg, buff2, arglen) == 0)
2129              {
2130              if (equals != NULL && argequals != NULL)
2131                {
2132                option_data = argequals;
2133                if (*option_data == '=')
2134                  {
2135                  option_data++;
2136                  longopwasequals = TRUE;
2137                  }
2138                }
2139            break;            break;
2140              }
2141          }          }
2142        }        }
2143    
2144      if (op->one_char == 0)      if (op->one_char == 0)
2145        {        {
2146        fprintf(stderr, "pcregrep: Unknown option %s\n", argv[i]);        fprintf(stderr, "pcregrep: Unknown option %s\n", argv[i]);
2147        exit(usage(2));        pcregrep_exit(usage(2));
2148        }        }
2149      }      }
2150    
2151      /* Jeffrey Friedl's debugging harness uses these additional options which
2152      are not in the right form for putting in the option table because they use
2153      only one hyphen, yet are more than one character long. By putting them
2154      separately here, they will not get displayed as part of the help() output,
2155      but I don't think Jeffrey will care about that. */
2156    
2157    #ifdef JFRIEDL_DEBUG
2158      else if (strcmp(argv[i], "-pre") == 0) {
2159              jfriedl_prefix = argv[++i];
2160              continue;
2161      } else if (strcmp(argv[i], "-post") == 0) {
2162              jfriedl_postfix = argv[++i];
2163              continue;
2164      } else if (strcmp(argv[i], "-XT") == 0) {
2165              sscanf(argv[++i], "%d", &jfriedl_XT);
2166              continue;
2167      } else if (strcmp(argv[i], "-XR") == 0) {
2168              sscanf(argv[++i], "%d", &jfriedl_XR);
2169              continue;
2170      }
2171    #endif
2172    
2173    
2174    /* 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
2175    continue till we hit the last one or one that needs data. */    continue till we hit the last one or one that needs data. */
2176    
# Line 1304  for (i = 1; i < argc; i++) Line 2181  for (i = 1; i < argc; i++)
2181      while (*s != 0)      while (*s != 0)
2182        {        {
2183        for (op = optionlist; op->one_char != 0; op++)        for (op = optionlist; op->one_char != 0; op++)
2184          { if (*s == op->one_char) break; }          {
2185            if (*s == op->one_char) break;
2186            }
2187        if (op->one_char == 0)        if (op->one_char == 0)
2188          {          {
2189          fprintf(stderr, "pcregrep: Unknown option letter '%c' in \"%s\"\n",          fprintf(stderr, "pcregrep: Unknown option letter '%c' in \"%s\"\n",
2190            *s, argv[i]);            *s, argv[i]);
2191          exit(usage(2));          pcregrep_exit(usage(2));
2192          }          }
2193        if (op->type != OP_NODATA || s[1] == 0)  
2194          /* Check for a single-character option that has data: OP_OP_NUMBER
2195          is used for one that either has a numerical number or defaults, i.e. the
2196          data is optional. If a digit follows, there is data; if not, carry on
2197          with other single-character options in the same string. */
2198    
2199          option_data = s+1;
2200          if (op->type == OP_OP_NUMBER)
2201          {          {
2202          option_data = s+1;          if (isdigit((unsigned char)s[1])) break;
         break;  
2203          }          }
2204          else   /* Check for end or a dataless option */
2205            {
2206            if (op->type != OP_NODATA || s[1] == 0) break;
2207            }
2208    
2209          /* Handle a single-character option with no data, then loop for the
2210          next character in the string. */
2211    
2212        pcre_options = handle_option(*s++, pcre_options);        pcre_options = handle_option(*s++, pcre_options);
2213        }        }
2214      }      }
# Line 1332  for (i = 1; i < argc; i++) Line 2225  for (i = 1; i < argc; i++)
2225    
2226    /* If the option type is OP_OP_STRING or OP_OP_NUMBER, it's an option that    /* If the option type is OP_OP_STRING or OP_OP_NUMBER, it's an option that
2227    either has a value or defaults to something. It cannot have data in a    either has a value or defaults to something. It cannot have data in a
2228    separate item. At the moment, the only such options are "colo(u)r" and    separate item. At the moment, the only such options are "colo(u)r",
2229    Jeffrey Friedl's special debugging option. */    "only-matching", and Jeffrey Friedl's special -S debugging option. */
2230    
2231    if (*option_data == 0 &&    if (*option_data == 0 &&
2232        (op->type == OP_OP_STRING || op->type == OP_OP_NUMBER))        (op->type == OP_OP_STRING || op->type == OP_OP_NUMBER))
# Line 1343  for (i = 1; i < argc; i++) Line 2236  for (i = 1; i < argc; i++)
2236        case N_COLOUR:        case N_COLOUR:
2237        colour_option = (char *)"auto";        colour_option = (char *)"auto";
2238        break;        break;
2239    
2240          case 'o':
2241          only_matching = 0;
2242          break;
2243    
2244  #ifdef JFRIEDL_DEBUG  #ifdef JFRIEDL_DEBUG
2245        case 'S':        case 'S':
2246        S_arg = 0;        S_arg = 0;
# Line 1359  for (i = 1; i < argc; i++) Line 2257  for (i = 1; i < argc; i++)
2257      if (i >= argc - 1 || longopwasequals)      if (i >= argc - 1 || longopwasequals)
2258        {        {
2259        fprintf(stderr, "pcregrep: Data missing after %s\n", argv[i]);        fprintf(stderr, "pcregrep: Data missing after %s\n", argv[i]);
2260        exit(usage(2));        pcregrep_exit(usage(2));
2261        }        }
2262      option_data = argv[++i];      option_data = argv[++i];
2263      }      }
# Line 1380  for (i = 1; i < argc; i++) Line 2278  for (i = 1; i < argc; i++)
2278    
2279    /* Otherwise, deal with single string or numeric data values. */    /* Otherwise, deal with single string or numeric data values. */
2280    
2281    else if (op->type != OP_NUMBER && op->type != OP_OP_NUMBER)    else if (op->type != OP_NUMBER && op->type != OP_LONGNUMBER &&
2282               op->type != OP_OP_NUMBER)
2283      {      {
2284      *((char **)op->dataptr) = option_data;      *((char **)op->dataptr) = option_data;
2285      }      }
2286    
2287      /* Avoid the use of strtoul() because SunOS4 doesn't have it. This is used
2288      only for unpicking arguments, so just keep it simple. */
2289    
2290    else    else
2291      {      {
2292      char *endptr;      unsigned long int n = 0;
2293      int n = strtoul(option_data, &endptr, 10);      char *endptr = option_data;
2294        while (*endptr != 0 && isspace((unsigned char)(*endptr))) endptr++;
2295        while (isdigit((unsigned char)(*endptr)))
2296          n = n * 10 + (int)(*endptr++ - '0');
2297      if (*endptr != 0)      if (*endptr != 0)
2298        {        {
2299        if (longop)        if (longop)
2300          {          {
2301          char *equals = strchr(op->long_name, '=');          char *equals = strchr(op->long_name, '=');
2302          int nlen = (equals == NULL)? (int)strlen(op->long_name) :          int nlen = (equals == NULL)? (int)strlen(op->long_name) :
2303            equals - op->long_name;            (int)(equals - op->long_name);
2304          fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%.*s\n",          fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%.*s\n",
2305            option_data, nlen, op->long_name);            option_data, nlen, op->long_name);
2306          }          }
2307        else        else
2308          fprintf(stderr, "pcregrep: Malformed number \"%s\" after -%c\n",          fprintf(stderr, "pcregrep: Malformed number \"%s\" after -%c\n",
2309            option_data, op->one_char);            option_data, op->one_char);
2310        exit(usage(2));        pcregrep_exit(usage(2));
2311        }        }
2312      *((int *)op->dataptr) = n;      if (op->type == OP_LONGNUMBER)
2313            *((unsigned long int *)op->dataptr) = n;
2314        else
2315            *((int *)op->dataptr) = n;
2316      }      }
2317    }    }
2318    
# Line 1416  if (both_context > 0) Line 2325  if (both_context > 0)
2325    if (before_context == 0) before_context = both_context;    if (before_context == 0) before_context = both_context;
2326    }    }
2327    
2328    /* Only one of --only-matching, --file-offsets, or --line-offsets is permitted.
2329    However, the latter two set only_matching. */
2330    
2331    if ((only_matching >= 0 && (file_offsets || line_offsets)) ||
2332        (file_offsets && line_offsets))
2333      {
2334      fprintf(stderr, "pcregrep: Cannot mix --only-matching, --file-offsets "
2335        "and/or --line-offsets\n");
2336      pcregrep_exit(usage(2));
2337      }
2338    
2339    if (file_offsets || line_offsets) only_matching = 0;
2340    
2341  /* If a locale has not been provided as an option, see if the LC_CTYPE or  /* If a locale has not been provided as an option, see if the LC_CTYPE or
2342  LC_ALL environment variable is set, and if so, use it. */  LC_ALL environment variable is set, and if so, use it. */
2343    
# Line 1465  if (colour_option != NULL && strcmp(colo Line 2387  if (colour_option != NULL && strcmp(colo
2387      }      }
2388    }    }
2389    
2390    /* Interpret the newline type; the default settings are Unix-like. */
2391    
2392    if (strcmp(newline, "cr") == 0 || strcmp(newline, "CR") == 0)
2393      {
2394      pcre_options |= PCRE_NEWLINE_CR;
2395      endlinetype = EL_CR;
2396      }
2397    else if (strcmp(newline, "lf") == 0 || strcmp(newline, "LF") == 0)
2398      {
2399      pcre_options |= PCRE_NEWLINE_LF;
2400      endlinetype = EL_LF;
2401      }
2402    else if (strcmp(newline, "crlf") == 0 || strcmp(newline, "CRLF") == 0)
2403      {
2404      pcre_options |= PCRE_NEWLINE_CRLF;
2405      endlinetype = EL_CRLF;
2406      }
2407    else if (strcmp(newline, "any") == 0 || strcmp(newline, "ANY") == 0)
2408      {
2409      pcre_options |= PCRE_NEWLINE_ANY;
2410      endlinetype = EL_ANY;
2411      }
2412    else if (strcmp(newline, "anycrlf") == 0 || strcmp(newline, "ANYCRLF") == 0)
2413      {
2414      pcre_options |= PCRE_NEWLINE_ANYCRLF;
2415      endlinetype = EL_ANYCRLF;
2416      }
2417    else
2418      {
2419      fprintf(stderr, "pcregrep: Invalid newline specifier \"%s\"\n", newline);
2420      return 2;
2421      }
2422    
2423  /* Interpret the text values for -d and -D */  /* Interpret the text values for -d and -D */
2424    
2425  if (dee_option != NULL)  if (dee_option != NULL)
# Line 1490  if (DEE_option != NULL) Line 2445  if (DEE_option != NULL)
2445      }      }
2446    }    }
2447    
2448  /* Check the value for Jeff Friedl's debugging option. */  /* Check the values for Jeffrey Friedl's debugging options. */
2449    
2450  #ifdef JFRIEDL_DEBUG  #ifdef JFRIEDL_DEBUG
2451  if (S_arg > 9)  if (S_arg > 9)
# Line 1498  if (S_arg > 9) Line 2453  if (S_arg > 9)
2453    fprintf(stderr, "pcregrep: bad value for -S option\n");    fprintf(stderr, "pcregrep: bad value for -S option\n");
2454    return 2;    return 2;
2455    }    }
2456    if (jfriedl_XT != 0 || jfriedl_XR != 0)
2457      {
2458      if (jfriedl_XT == 0) jfriedl_XT = 1;
2459      if (jfriedl_XR == 0) jfriedl_XR = 1;
2460      }
2461  #endif  #endif
2462    
2463  /* Get memory to store the pattern and hints lists. */  /* Get memory to store the pattern and hints lists. */
# Line 1508  hints_list = (pcre_extra **)malloc(MAX_P Line 2468  hints_list = (pcre_extra **)malloc(MAX_P
2468  if (pattern_list == NULL || hints_list == NULL)  if (pattern_list == NULL || hints_list == NULL)
2469    {    {
2470    fprintf(stderr, "pcregrep: malloc failed\n");    fprintf(stderr, "pcregrep: malloc failed\n");
2471    return 2;    goto EXIT2;
2472    }    }
2473    
2474  /* If no patterns were provided by -e, and there is no file provided by -f,  /* If no patterns were provided by -e, and there is no file provided by -f,
# Line 1527  for (j = 0; j < cmd_pattern_count; j++) Line 2487  for (j = 0; j < cmd_pattern_count; j++)
2487    {    {
2488    if (!compile_pattern(patterns[j], pcre_options, NULL,    if (!compile_pattern(patterns[j], pcre_options, NULL,
2489         (j == 0 && cmd_pattern_count == 1)? 0 : j + 1))         (j == 0 && cmd_pattern_count == 1)? 0 : j + 1))
2490      return 2;      goto EXIT2;
2491    }    }
2492    
2493  /* Compile the regular expressions that are provided in a file. */  /* Compile the regular expressions that are provided in a file. */
# Line 1551  if (pattern_filename != NULL) Line 2511  if (pattern_filename != NULL)
2511        {        {
2512        fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pattern_filename,        fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pattern_filename,
2513          strerror(errno));          strerror(errno));
2514        return 2;        goto EXIT2;
2515        }        }
2516      filename = pattern_filename;      filename = pattern_filename;
2517      }      }
# Line 1564  if (pattern_filename != NULL) Line 2524  if (pattern_filename != NULL)
2524      linenumber++;      linenumber++;
2525      if (buffer[0] == 0) continue;   /* Skip blank lines */      if (buffer[0] == 0) continue;   /* Skip blank lines */
2526      if (!compile_pattern(buffer, pcre_options, filename, linenumber))      if (!compile_pattern(buffer, pcre_options, filename, linenumber))
2527        return 2;        goto EXIT2;
2528      }      }
2529    
2530    if (f != stdin) fclose(f);    if (f != stdin) fclose(f);
# Line 1580  for (j = 0; j < pattern_count; j++) Line 2540  for (j = 0; j < pattern_count; j++)
2540      char s[16];      char s[16];
2541      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);
2542      fprintf(stderr, "pcregrep: Error while studying regex%s: %s\n", s, error);      fprintf(stderr, "pcregrep: Error while studying regex%s: %s\n", s, error);
2543      return 2;      goto EXIT2;
2544        }
2545      hint_count++;
2546      }
2547    
2548    /* If --match-limit or --recursion-limit was set, put the value(s) into the
2549    pcre_extra block for each pattern. */
2550    
2551    if (match_limit > 0 || match_limit_recursion > 0)
2552      {
2553      for (j = 0; j < pattern_count; j++)
2554        {
2555        if (hints_list[j] == NULL)
2556          {
2557          hints_list[j] = malloc(sizeof(pcre_extra));
2558          if (hints_list[j] == NULL)
2559            {
2560            fprintf(stderr, "pcregrep: malloc failed\n");
2561            pcregrep_exit(2);
2562            }
2563          }
2564        if (match_limit > 0)
2565          {
2566          hints_list[j]->flags |= PCRE_EXTRA_MATCH_LIMIT;
2567          hints_list[j]->match_limit = match_limit;
2568          }
2569        if (match_limit_recursion > 0)
2570          {
2571          hints_list[j]->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
2572          hints_list[j]->match_limit_recursion = match_limit_recursion;
2573          }
2574      }      }
2575    }    }
2576    
# Line 1594  if (exclude_pattern != NULL) Line 2584  if (exclude_pattern != NULL)
2584      {      {
2585      fprintf(stderr, "pcregrep: Error in 'exclude' regex at offset %d: %s\n",      fprintf(stderr, "pcregrep: Error in 'exclude' regex at offset %d: %s\n",
2586        errptr, error);        errptr, error);
2587      return 2;      goto EXIT2;
2588      }      }
2589    }    }
2590    
# Line 1606  if (include_pattern != NULL) Line 2596  if (include_pattern != NULL)
2596      {      {
2597      fprintf(stderr, "pcregrep: Error in 'include' regex at offset %d: %s\n",      fprintf(stderr, "pcregrep: Error in 'include' regex at offset %d: %s\n",
2598        errptr, error);        errptr, error);
2599      return 2;      goto EXIT2;
2600        }
2601      }
2602    
2603    if (exclude_dir_pattern != NULL)
2604      {
2605      exclude_dir_compiled = pcre_compile(exclude_dir_pattern, 0, &error, &errptr,
2606        pcretables);
2607      if (exclude_dir_compiled == NULL)
2608        {
2609        fprintf(stderr, "pcregrep: Error in 'exclude_dir' regex at offset %d: %s\n",
2610          errptr, error);
2611        goto EXIT2;
2612        }
2613      }
2614    
2615    if (include_dir_pattern != NULL)
2616      {
2617      include_dir_compiled = pcre_compile(include_dir_pattern, 0, &error, &errptr,
2618        pcretables);
2619      if (include_dir_compiled == NULL)
2620        {
2621        fprintf(stderr, "pcregrep: Error in 'include_dir' regex at offset %d: %s\n",
2622          errptr, error);
2623        goto EXIT2;
2624      }      }
2625    }    }
2626    
2627  /* 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. */
2628    
2629  if (i >= argc)  if (i >= argc)
2630    return pcregrep(stdin, (filenames > FN_DEFAULT)? stdin_name : NULL);    {
2631      rc = pcregrep(stdin, FR_PLAIN, (filenames > FN_DEFAULT)? stdin_name : NULL);
2632      goto EXIT;
2633      }
2634    
2635  /* Otherwise, work through the remaining arguments as files or directories.  /* Otherwise, work through the remaining arguments as files or directories.
2636  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
# Line 1630  for (; i < argc; i++) Line 2647  for (; i < argc; i++)
2647      else if (frc == 0 && rc == 1) rc = 0;      else if (frc == 0 && rc == 1) rc = 0;
2648    }    }
2649    
2650  return rc;  EXIT:
2651    if (pattern_list != NULL)
2652      {
2653      for (i = 0; i < pattern_count; i++) free(pattern_list[i]);
2654      free(pattern_list);
2655      }
2656    if (hints_list != NULL)
2657      {
2658      for (i = 0; i < hint_count; i++)
2659        {
2660        if (hints_list[i] != NULL) free(hints_list[i]);
2661        }
2662      free(hints_list);
2663      }
2664    pcregrep_exit(rc);
2665    
2666    EXIT2:
2667    rc = 2;
2668    goto EXIT;
2669  }  }
2670    
2671  /* End of pcregrep */  /* End of pcregrep */

Legend:
Removed from v.87  
changed lines
  Added in v.587

  ViewVC Help
Powered by ViewVC 1.1.5