| 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) 1997-2007 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 |
| 38 |
*/ |
*/ |
| 39 |
|
|
| 40 |
#ifdef HAVE_CONFIG_H |
#ifdef HAVE_CONFIG_H |
| 41 |
#include <config.h> |
#include "config.h" |
| 42 |
#endif |
#endif |
| 43 |
|
|
| 44 |
#include <ctype.h> |
#include <ctype.h> |
| 55 |
#include <unistd.h> |
#include <unistd.h> |
| 56 |
#endif |
#endif |
| 57 |
|
|
| 58 |
#include <pcre.h> |
#ifdef SUPPORT_LIBZ |
| 59 |
|
#include <zlib.h> |
| 60 |
|
#endif |
| 61 |
|
|
| 62 |
|
#ifdef SUPPORT_LIBBZ2 |
| 63 |
|
#include <bzlib.h> |
| 64 |
|
#endif |
| 65 |
|
|
| 66 |
|
#include "pcre.h" |
| 67 |
|
|
| 68 |
#define FALSE 0 |
#define FALSE 0 |
| 69 |
#define TRUE 1 |
#define TRUE 1 |
| 71 |
typedef int BOOL; |
typedef int BOOL; |
| 72 |
|
|
| 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 |
| 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 |
|
|
| 104 |
|
|
| 105 |
enum { EL_LF, EL_CR, EL_CRLF, EL_ANY, EL_ANYCRLF }; |
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 |
/************************************************* |
/************************************************* |
| 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; |
| 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; |
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; |
| 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_STRING, 'N', &newline, "newline=type", "specify newline type (CR, LF, CRLF, ANYCRLF or ANY)" }, |
{ OP_NODATA, N_LOFFSETS, NULL, "line-offsets", "output line numbers and offsets, not text" }, |
| 241 |
{ OP_NODATA, 'n', NULL, "line-number", "print line number with output lines" }, |
{ OP_STRING, N_LOCALE, &locale, "locale=locale", "use the named locale" }, |
| 242 |
{ OP_NODATA, 'o', NULL, "only-matching", "show only the part of the line that matched" }, |
{ OP_LONGNUMBER, N_M_LIMIT, &match_limit, "match-limit=number", "set PCRE match limit option" }, |
| 243 |
{ OP_NODATA, 'q', NULL, "quiet", "suppress output, just set return code" }, |
{ OP_LONGNUMBER, N_M_LIMIT_REC, &match_limit_recursion, "recursion-limit=number", "set PCRE match recursion limit option" }, |
| 244 |
{ OP_NODATA, 'r', NULL, "recursive", "recursively scan sub-directories" }, |
{ OP_NODATA, 'M', NULL, "multiline", "run in multiline mode" }, |
| 245 |
{ OP_STRING, N_EXCLUDE,&exclude_pattern, "exclude=pattern","exclude matching files when recursing" }, |
{ OP_STRING, 'N', &newline, "newline=type", "set newline type (CR, LF, CRLF, ANYCRLF or ANY)" }, |
| 246 |
{ OP_STRING, N_INCLUDE,&include_pattern, "include=pattern","include matching files when recursing" }, |
{ 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 |
| 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 |
|
/************************************************* |
| 323 |
* OS-specific functions * |
* OS-specific functions * |
| 324 |
*************************************************/ |
*************************************************/ |
| 325 |
|
|
| 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) |
| 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 defined HAVE_WINDOWS_H && HAVE_WINDOWS_H |
|
#elif HAVE_WINDOWS_H |
|
| 411 |
|
|
| 412 |
#ifndef STRICT |
#ifndef STRICT |
| 413 |
# define STRICT |
# define STRICT |
| 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; |
| 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); |
| 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 |
|
|
| 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) |
| 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 |
|
|
| 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 * |
* Find end of line * |
| 627 |
*************************************************/ |
*************************************************/ |
| 628 |
|
|
| 634 |
endptr end of available data |
endptr end of available data |
| 635 |
lenptr where to put the length of the eol sequence |
lenptr where to put the length of the eol sequence |
| 636 |
|
|
| 637 |
Returns: pointer to the last byte of the line |
Returns: pointer to the last byte of the line, including the newline byte(s) |
| 638 |
*/ |
*/ |
| 639 |
|
|
| 640 |
static char * |
static char * |
| 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 |
pp = end_of_line(pp, endptr, &ellength); |
pp = end_of_line(pp, endptr, &ellength); |
| 920 |
fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
FWRITE(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
| 921 |
lastmatchrestart = pp; |
lastmatchrestart = pp; |
| 922 |
} |
} |
| 923 |
hyphenpending = TRUE; |
hyphenpending = TRUE; |
| 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 |
|
startoffset where to start matching |
| 941 |
|
offsets the offets vector to fill in |
| 942 |
|
mrc address of where to put the result of pcre_exec() |
| 943 |
|
|
| 944 |
|
Returns: TRUE if there was a match |
| 945 |
|
FALSE if there was no match |
| 946 |
|
invert if there was a non-fatal error |
| 947 |
|
*/ |
| 948 |
|
|
| 949 |
|
static BOOL |
| 950 |
|
match_patterns(char *matchptr, size_t length, int startoffset, int *offsets, |
| 951 |
|
int *mrc) |
| 952 |
|
{ |
| 953 |
|
int i; |
| 954 |
|
size_t slen = length; |
| 955 |
|
const char *msg = "this text:\n\n"; |
| 956 |
|
if (slen > 200) |
| 957 |
|
{ |
| 958 |
|
slen = 200; |
| 959 |
|
msg = "text that starts:\n\n"; |
| 960 |
|
} |
| 961 |
|
for (i = 0; i < pattern_count; i++) |
| 962 |
|
{ |
| 963 |
|
*mrc = pcre_exec(pattern_list[i], hints_list[i], matchptr, (int)length, |
| 964 |
|
startoffset, PCRE_NOTEMPTY, offsets, OFFSET_SIZE); |
| 965 |
|
if (*mrc >= 0) return TRUE; |
| 966 |
|
if (*mrc == PCRE_ERROR_NOMATCH) continue; |
| 967 |
|
fprintf(stderr, "pcregrep: pcre_exec() gave error %d while matching ", *mrc); |
| 968 |
|
if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1); |
| 969 |
|
fprintf(stderr, "%s", msg); |
| 970 |
|
FWRITE(matchptr, 1, slen, stderr); /* In case binary zero included */ |
| 971 |
|
fprintf(stderr, "\n\n"); |
| 972 |
|
if (*mrc == PCRE_ERROR_MATCHLIMIT || *mrc == PCRE_ERROR_RECURSIONLIMIT) |
| 973 |
|
resource_error = TRUE; |
| 974 |
|
if (error_count++ > 20) |
| 975 |
|
{ |
| 976 |
|
fprintf(stderr, "pcregrep: Too many errors - abandoned.\n"); |
| 977 |
|
pcregrep_exit(2); |
| 978 |
|
} |
| 979 |
|
return invert; /* No more matching; don't show the line again */ |
| 980 |
|
} |
| 981 |
|
|
| 982 |
|
return FALSE; /* No match, no errors */ |
| 983 |
|
} |
| 984 |
|
|
| 985 |
|
|
| 986 |
|
|
| 987 |
|
/************************************************* |
| 988 |
* Grep an individual file * |
* Grep an individual file * |
| 989 |
*************************************************/ |
*************************************************/ |
| 990 |
|
|
| 996 |
"before" context printing. |
"before" context printing. |
| 997 |
|
|
| 998 |
Arguments: |
Arguments: |
| 999 |
in the fopened FILE stream |
handle the fopened FILE stream for a normal file |
| 1000 |
|
the gzFile pointer when reading is via libz |
| 1001 |
|
the BZFILE pointer when reading is via libbz2 |
| 1002 |
|
frtype FR_PLAIN, FR_LIBZ, or FR_LIBBZ2 |
| 1003 |
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 |
| 1004 |
or NULL if the file name is not to be printed |
or NULL if the file name is not to be printed |
| 1005 |
it cannot be NULL if filenames[_nomatch]_only is set |
it cannot be NULL if filenames[_nomatch]_only is set |
| 1006 |
|
|
| 1007 |
Returns: 0 if there was at least one match |
Returns: 0 if there was at least one match |
| 1008 |
1 otherwise (no matches) |
1 otherwise (no matches) |
| 1009 |
|
2 if there is a read error on a .bz2 file |
| 1010 |
*/ |
*/ |
| 1011 |
|
|
| 1012 |
static int |
static int |
| 1013 |
pcregrep(FILE *in, char *printname) |
pcregrep(void *handle, int frtype, char *printname) |
| 1014 |
{ |
{ |
| 1015 |
int rc = 1; |
int rc = 1; |
| 1016 |
int linenumber = 1; |
int linenumber = 1; |
| 1017 |
int lastmatchnumber = 0; |
int lastmatchnumber = 0; |
| 1018 |
int count = 0; |
int count = 0; |
| 1019 |
int offsets[99]; |
int filepos = 0; |
| 1020 |
|
int offsets[OFFSET_SIZE]; |
| 1021 |
char *lastmatchrestart = NULL; |
char *lastmatchrestart = NULL; |
| 1022 |
char buffer[3*MBUFTHIRD]; |
char buffer[3*MBUFTHIRD]; |
| 1023 |
char *ptr = buffer; |
char *ptr = buffer; |
| 1024 |
char *endptr; |
char *endptr; |
| 1025 |
size_t bufflength; |
size_t bufflength; |
| 1026 |
BOOL endhyphenpending = FALSE; |
BOOL endhyphenpending = FALSE; |
| 1027 |
|
BOOL input_line_buffered = line_buffered; |
| 1028 |
|
FILE *in = NULL; /* Ensure initialized */ |
| 1029 |
|
|
| 1030 |
|
#ifdef SUPPORT_LIBZ |
| 1031 |
|
gzFile ingz = NULL; |
| 1032 |
|
#endif |
| 1033 |
|
|
| 1034 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1035 |
|
BZFILE *inbz2 = NULL; |
| 1036 |
|
#endif |
| 1037 |
|
|
| 1038 |
|
|
| 1039 |
|
/* Do the first read into the start of the buffer and set up the pointer to end |
| 1040 |
|
of what we have. In the case of libz, a non-zipped .gz file will be read as a |
| 1041 |
|
plain file. However, if a .bz2 file isn't actually bzipped, the first read will |
| 1042 |
|
fail. */ |
| 1043 |
|
|
| 1044 |
/* Do the first read into the start of the buffer and set up the pointer to |
#ifdef SUPPORT_LIBZ |
| 1045 |
end of what we have. */ |
if (frtype == FR_LIBZ) |
| 1046 |
|
{ |
| 1047 |
|
ingz = (gzFile)handle; |
| 1048 |
|
bufflength = gzread (ingz, buffer, 3*MBUFTHIRD); |
| 1049 |
|
} |
| 1050 |
|
else |
| 1051 |
|
#endif |
| 1052 |
|
|
| 1053 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1054 |
|
if (frtype == FR_LIBBZ2) |
| 1055 |
|
{ |
| 1056 |
|
inbz2 = (BZFILE *)handle; |
| 1057 |
|
bufflength = BZ2_bzread(inbz2, buffer, 3*MBUFTHIRD); |
| 1058 |
|
if ((int)bufflength < 0) return 2; /* Gotcha: bufflength is size_t; */ |
| 1059 |
|
} /* without the cast it is unsigned. */ |
| 1060 |
|
else |
| 1061 |
|
#endif |
| 1062 |
|
|
| 1063 |
|
{ |
| 1064 |
|
in = (FILE *)handle; |
| 1065 |
|
if (is_file_tty(in)) input_line_buffered = TRUE; |
| 1066 |
|
bufflength = input_line_buffered? |
| 1067 |
|
read_one_line(buffer, 3*MBUFTHIRD, in) : |
| 1068 |
|
fread(buffer, 1, 3*MBUFTHIRD, in); |
| 1069 |
|
} |
| 1070 |
|
|
|
bufflength = fread(buffer, 1, 3*MBUFTHIRD, in); |
|
| 1071 |
endptr = buffer + bufflength; |
endptr = buffer + bufflength; |
| 1072 |
|
|
| 1073 |
/* 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 |
| 1077 |
|
|
| 1078 |
while (ptr < endptr) |
while (ptr < endptr) |
| 1079 |
{ |
{ |
| 1080 |
int i, endlinelength; |
int endlinelength; |
| 1081 |
int mrc = 0; |
int mrc = 0; |
| 1082 |
BOOL match = FALSE; |
int startoffset = 0; |
| 1083 |
|
BOOL match; |
| 1084 |
|
char *matchptr = ptr; |
| 1085 |
char *t = ptr; |
char *t = ptr; |
| 1086 |
size_t length, linelength; |
size_t length, linelength; |
| 1087 |
|
|
| 1088 |
/* 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 |
| 1089 |
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 |
| 1090 |
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 |
| 1091 |
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 |
| 1092 |
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 |
| 1093 |
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 |
| 1094 |
|
first line. */ |
| 1095 |
|
|
| 1096 |
t = end_of_line(t, endptr, &endlinelength); |
t = end_of_line(t, endptr, &endlinelength); |
| 1097 |
linelength = t - ptr - endlinelength; |
linelength = t - ptr - endlinelength; |
| 1106 |
#include <time.h> |
#include <time.h> |
| 1107 |
struct timeval start_time, end_time; |
struct timeval start_time, end_time; |
| 1108 |
struct timezone dummy; |
struct timezone dummy; |
| 1109 |
|
int i; |
| 1110 |
|
|
| 1111 |
if (jfriedl_XT) |
if (jfriedl_XT) |
| 1112 |
{ |
{ |
| 1115 |
ptr = malloc(newlen + 1); |
ptr = malloc(newlen + 1); |
| 1116 |
if (!ptr) { |
if (!ptr) { |
| 1117 |
printf("out of memory"); |
printf("out of memory"); |
| 1118 |
exit(2); |
pcregrep_exit(2); |
| 1119 |
} |
} |
| 1120 |
endptr = ptr; |
endptr = ptr; |
| 1121 |
strcpy(endptr, jfriedl_prefix); endptr += strlen(jfriedl_prefix); |
strcpy(endptr, jfriedl_prefix); endptr += strlen(jfriedl_prefix); |
| 1132 |
|
|
| 1133 |
|
|
| 1134 |
for (i = 0; i < jfriedl_XR; i++) |
for (i = 0; i < jfriedl_XR; i++) |
| 1135 |
match = (pcre_exec(pattern_list[0], hints_list[0], ptr, length, 0, 0, offsets, 99) >= 0); |
match = (pcre_exec(pattern_list[0], hints_list[0], ptr, length, 0, |
| 1136 |
|
PCRE_NOTEMPTY, offsets, OFFSET_SIZE) >= 0); |
| 1137 |
|
|
| 1138 |
if (gettimeofday(&end_time, &dummy) != 0) |
if (gettimeofday(&end_time, &dummy) != 0) |
| 1139 |
perror("bad gettimeofday"); |
perror("bad gettimeofday"); |
| 1147 |
} |
} |
| 1148 |
#endif |
#endif |
| 1149 |
|
|
| 1150 |
|
/* We come back here after a match when the -o option (only_matching) is set, |
| 1151 |
|
in order to find any further matches in the same line. */ |
| 1152 |
|
|
| 1153 |
/* Run through all the patterns until one matches. Note that we don't include |
ONLY_MATCHING_RESTART: |
|
the final newline in the subject string. */ |
|
| 1154 |
|
|
| 1155 |
for (i = 0; i < pattern_count; i++) |
/* Run through all the patterns until one matches or there is an error other |
| 1156 |
{ |
than NOMATCH. This code is in a subroutine so that it can be re-used for |
| 1157 |
mrc = pcre_exec(pattern_list[i], hints_list[i], ptr, length, 0, 0, |
finding subsequent matches when colouring matched lines. */ |
| 1158 |
offsets, 99); |
|
| 1159 |
if (mrc >= 0) { match = TRUE; break; } |
match = match_patterns(matchptr, length, startoffset, offsets, &mrc); |
|
if (mrc != PCRE_ERROR_NOMATCH) |
|
|
{ |
|
|
fprintf(stderr, "pcregrep: pcre_exec() error %d while matching ", mrc); |
|
|
if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1); |
|
|
fprintf(stderr, "this line:\n"); |
|
|
fwrite(ptr, 1, linelength, stderr); /* In case binary zero included */ |
|
|
fprintf(stderr, "\n"); |
|
|
if (error_count == 0 && |
|
|
(mrc == PCRE_ERROR_MATCHLIMIT || mrc == PCRE_ERROR_RECURSIONLIMIT)) |
|
|
{ |
|
|
fprintf(stderr, "pcregrep: error %d means that a resource limit " |
|
|
"was exceeded\n", mrc); |
|
|
fprintf(stderr, "pcregrep: check your regex for nested unlimited loops\n"); |
|
|
} |
|
|
if (error_count++ > 20) |
|
|
{ |
|
|
fprintf(stderr, "pcregrep: too many errors - abandoned\n"); |
|
|
exit(2); |
|
|
} |
|
|
match = invert; /* No more matching; don't show the line again */ |
|
|
break; |
|
|
} |
|
|
} |
|
| 1160 |
|
|
| 1161 |
/* 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. */ |
| 1162 |
|
|
| 1175 |
/* 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 |
| 1176 |
in the file. */ |
in the file. */ |
| 1177 |
|
|
| 1178 |
else if (filenames == FN_ONLY) |
else if (filenames == FN_MATCH_ONLY) |
| 1179 |
{ |
{ |
| 1180 |
fprintf(stdout, "%s\n", printname); |
fprintf(stdout, "%s\n", printname); |
| 1181 |
return 0; |
return 0; |
| 1185 |
|
|
| 1186 |
else if (quiet) return 0; |
else if (quiet) return 0; |
| 1187 |
|
|
| 1188 |
/* The --only-matching option prints just the substring that matched, and |
/* The --only-matching option prints just the substring that matched, or a |
| 1189 |
does not pring any context. */ |
captured portion of it, as long as this string is not empty, and the |
| 1190 |
|
--file-offsets and --line-offsets options output offsets for the matching |
| 1191 |
|
substring (they both force --only-matching = 0). None of these options |
| 1192 |
|
prints any context. Afterwards, adjust the start and then jump back to look |
| 1193 |
|
for further matches in the same line. If we are in invert mode, however, |
| 1194 |
|
nothing is printed and we do not restart - this could still be useful |
| 1195 |
|
because the return code is set. */ |
| 1196 |
|
|
| 1197 |
else if (only_matching) |
else if (only_matching >= 0) |
| 1198 |
{ |
{ |
| 1199 |
if (printname != NULL) fprintf(stdout, "%s:", printname); |
if (!invert) |
| 1200 |
if (number) fprintf(stdout, "%d:", linenumber); |
{ |
| 1201 |
fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
if (printname != NULL) fprintf(stdout, "%s:", printname); |
| 1202 |
fprintf(stdout, "\n"); |
if (number) fprintf(stdout, "%d:", linenumber); |
| 1203 |
|
if (line_offsets) |
| 1204 |
|
fprintf(stdout, "%d,%d\n", (int)(matchptr + offsets[0] - ptr), |
| 1205 |
|
offsets[1] - offsets[0]); |
| 1206 |
|
else if (file_offsets) |
| 1207 |
|
fprintf(stdout, "%d,%d\n", |
| 1208 |
|
(int)(filepos + matchptr + offsets[0] - ptr), |
| 1209 |
|
offsets[1] - offsets[0]); |
| 1210 |
|
else if (only_matching < mrc) |
| 1211 |
|
{ |
| 1212 |
|
int plen = offsets[2*only_matching + 1] - offsets[2*only_matching]; |
| 1213 |
|
if (plen > 0) |
| 1214 |
|
{ |
| 1215 |
|
if (do_colour) fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
| 1216 |
|
FWRITE(matchptr + offsets[only_matching*2], 1, plen, stdout); |
| 1217 |
|
if (do_colour) fprintf(stdout, "%c[00m", 0x1b); |
| 1218 |
|
fprintf(stdout, "\n"); |
| 1219 |
|
} |
| 1220 |
|
} |
| 1221 |
|
else if (printname != NULL || number) fprintf(stdout, "\n"); |
| 1222 |
|
match = FALSE; |
| 1223 |
|
if (line_buffered) fflush(stdout); |
| 1224 |
|
rc = 0; /* Had some success */ |
| 1225 |
|
startoffset = offsets[1]; /* Restart after the match */ |
| 1226 |
|
goto ONLY_MATCHING_RESTART; |
| 1227 |
|
} |
| 1228 |
} |
} |
| 1229 |
|
|
| 1230 |
/* 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 |
| 1258 |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
| 1259 |
if (number) fprintf(stdout, "%d-", lastmatchnumber++); |
if (number) fprintf(stdout, "%d-", lastmatchnumber++); |
| 1260 |
pp = end_of_line(pp, endptr, &ellength); |
pp = end_of_line(pp, endptr, &ellength); |
| 1261 |
fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
FWRITE(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
| 1262 |
lastmatchrestart = pp; |
lastmatchrestart = pp; |
| 1263 |
} |
} |
| 1264 |
if (lastmatchrestart != ptr) hyphenpending = TRUE; |
if (lastmatchrestart != ptr) hyphenpending = TRUE; |
| 1298 |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
| 1299 |
if (number) fprintf(stdout, "%d-", linenumber - linecount--); |
if (number) fprintf(stdout, "%d-", linenumber - linecount--); |
| 1300 |
pp = end_of_line(pp, endptr, &ellength); |
pp = end_of_line(pp, endptr, &ellength); |
| 1301 |
fwrite(p, 1, pp - p, stdout); |
FWRITE(p, 1, pp - p, stdout); |
| 1302 |
p = pp; |
p = pp; |
| 1303 |
} |
} |
| 1304 |
} |
} |
| 1314 |
|
|
| 1315 |
/* 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 |
| 1316 |
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 |
| 1317 |
line number appropriately. Because the PCRE_FIRSTLINE option is set, the |
line number appropriately, but only when there actually was a match |
| 1318 |
start of the match will always be before the first newline sequence. */ |
(invert not set). Because the PCRE_FIRSTLINE option is set, the start of |
| 1319 |
|
the match will always be before the first newline sequence. */ |
| 1320 |
|
|
| 1321 |
if (multiline) |
if (multiline & !invert) |
| 1322 |
{ |
{ |
|
int ellength; |
|
| 1323 |
char *endmatch = ptr + offsets[1]; |
char *endmatch = ptr + offsets[1]; |
| 1324 |
t = ptr; |
t = ptr; |
| 1325 |
while (t < endmatch) |
while (t < endmatch) |
| 1326 |
{ |
{ |
| 1327 |
t = end_of_line(t, endptr, &ellength); |
t = end_of_line(t, endptr, &endlinelength); |
| 1328 |
if (t <= endmatch) linenumber++; else break; |
if (t < endmatch) linenumber++; else break; |
| 1329 |
} |
} |
| 1330 |
endmatch = end_of_line(endmatch, endptr, &ellength); |
linelength = t - ptr - endlinelength; |
|
linelength = endmatch - ptr - ellength; |
|
| 1331 |
} |
} |
| 1332 |
|
|
| 1333 |
/*** NOTE: Use only fwrite() to output the data line, so that binary |
/*** NOTE: Use only fwrite() to output the data line, so that binary |
| 1342 |
{ |
{ |
| 1343 |
int first = S_arg * 2; |
int first = S_arg * 2; |
| 1344 |
int last = first + 1; |
int last = first + 1; |
| 1345 |
fwrite(ptr, 1, offsets[first], stdout); |
FWRITE(ptr, 1, offsets[first], stdout); |
| 1346 |
fprintf(stdout, "X"); |
fprintf(stdout, "X"); |
| 1347 |
fwrite(ptr + offsets[last], 1, linelength - offsets[last], stdout); |
FWRITE(ptr + offsets[last], 1, linelength - offsets[last], stdout); |
| 1348 |
} |
} |
| 1349 |
else |
else |
| 1350 |
#endif |
#endif |
| 1351 |
|
|
| 1352 |
/* We have to split the line(s) up if colouring. */ |
/* We have to split the line(s) up if colouring, and search for further |
| 1353 |
|
matches, but not of course if the line is a non-match. */ |
| 1354 |
|
|
| 1355 |
if (do_colour) |
if (do_colour && !invert) |
| 1356 |
{ |
{ |
| 1357 |
fwrite(ptr, 1, offsets[0], stdout); |
int plength; |
| 1358 |
|
FWRITE(ptr, 1, offsets[0], stdout); |
| 1359 |
fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
| 1360 |
fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
FWRITE(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
| 1361 |
fprintf(stdout, "%c[00m", 0x1b); |
fprintf(stdout, "%c[00m", 0x1b); |
| 1362 |
fwrite(ptr + offsets[1], 1, linelength - offsets[1], stdout); |
for (;;) |
| 1363 |
|
{ |
| 1364 |
|
startoffset = offsets[1]; |
| 1365 |
|
if (startoffset >= linelength + endlinelength || |
| 1366 |
|
!match_patterns(matchptr, length, startoffset, offsets, &mrc)) |
| 1367 |
|
break; |
| 1368 |
|
FWRITE(matchptr + startoffset, 1, offsets[0] - startoffset, stdout); |
| 1369 |
|
fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
| 1370 |
|
FWRITE(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
| 1371 |
|
fprintf(stdout, "%c[00m", 0x1b); |
| 1372 |
|
} |
| 1373 |
|
|
| 1374 |
|
/* In multiline mode, we may have already printed the complete line |
| 1375 |
|
and its line-ending characters (if they matched the pattern), so there |
| 1376 |
|
may be no more to print. */ |
| 1377 |
|
|
| 1378 |
|
plength = (linelength + endlinelength) - startoffset; |
| 1379 |
|
if (plength > 0) FWRITE(ptr + startoffset, 1, plength, stdout); |
| 1380 |
} |
} |
| 1381 |
else fwrite(ptr, 1, linelength + endlinelength, stdout); |
|
| 1382 |
|
/* 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 |
| 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 + endlinelength; |
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 |
| 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 */ |
| 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; |
| 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; |
| 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 |
{ |
{ |
| 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 (isdirectory(buffer)) |
| 1591 |
|
{ |
| 1592 |
|
if (exclude_dir_compiled != NULL && |
| 1593 |
|
pcre_exec(exclude_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0) |
| 1594 |
|
continue; |
| 1595 |
|
|
| 1596 |
|
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 (exclude_compiled != NULL && |
if (include_compiled != NULL && |
| 1607 |
pcre_exec(exclude_compiled, NULL, buffer, blen, 0, 0, NULL, 0) >= 0) |
pcre_exec(include_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0) |
| 1608 |
continue; |
continue; |
| 1609 |
|
} |
|
if (include_compiled != NULL && |
|
|
pcre_exec(include_compiled, NULL, buffer, blen, 0, 0, NULL, 0) < 0) |
|
|
continue; |
|
| 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; |
| 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, |
| 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 |
|
|
| 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 |
|
|
| 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"); |
|
printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n"); |
|
| 1772 |
|
|
| 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"); |
| 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; |
| 1852 |
|
|
| 1853 |
case 'V': |
case 'V': |
| 1854 |
fprintf(stderr, "pcregrep version %s\n", pcre_version()); |
fprintf(stderr, "pcregrep version %s\n", pcre_version()); |
| 1855 |
exit(0); |
pcregrep_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; |
| 2022 |
|
|
| 2023 |
/* Set the default line ending value from the default in the PCRE library; |
/* 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". |
"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); |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &i); |
| 2030 |
switch(i) |
switch(i) |
| 2031 |
{ |
{ |
| 2032 |
default: newline = (char *)"lf"; break; |
default: newline = (char *)"lf"; break; |
| 2033 |
case '\r': newline = (char *)"cr"; break; |
case 13: newline = (char *)"cr"; break; |
| 2034 |
case ('\r' << 8) | '\n': newline = (char *)"crlf"; break; |
case (13 << 8) | 10: newline = (char *)"crlf"; break; |
| 2035 |
case -1: newline = (char *)"any"; break; |
case -1: newline = (char *)"any"; break; |
| 2036 |
case -2: newline = (char *)"anycrlf"; break; |
case -2: newline = (char *)"anycrlf"; break; |
| 2037 |
} |
} |
| 2038 |
|
|
| 2039 |
/* Process the options */ |
/* Process the options */ |
| 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 */ |
| 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 |
{ |
{ |
| 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)? (int)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; |
| 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, |
sprintf(buff2, "%s%.*s", buff1, fulllen - baselen - 2, opbra + 1); |
| 2126 |
(int)strlen(op->long_name) - baselen - 2, 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 |
/* 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 |
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 |
only one hyphen, yet are more than one character long. By putting them |
| 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 |
|
|
| 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 |
|
if (isdigit((unsigned char)s[1])) break; |
| 2203 |
} |
} |
| 2204 |
if (op->type != OP_NODATA || s[1] == 0) |
else /* Check for end or a dataless option */ |
| 2205 |
{ |
{ |
| 2206 |
option_data = s+1; |
if (op->type != OP_NODATA || s[1] == 0) break; |
|
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 |
} |
} |
| 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 -S 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)) |
| 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; |
| 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 |
} |
} |
| 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 |
|
|
| 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 |
|
|
| 2545 |
hint_count++; |
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 |
|
|
| 2577 |
/* If there are include or exclude patterns, compile them. */ |
/* If there are include or exclude patterns, compile them. */ |
| 2578 |
|
|
| 2579 |
if (exclude_pattern != NULL) |
if (exclude_pattern != NULL) |
| 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 |
{ |
{ |
| 2631 |
rc = pcregrep(stdin, (filenames > FN_DEFAULT)? stdin_name : NULL); |
rc = pcregrep(stdin, FR_PLAIN, (filenames > FN_DEFAULT)? stdin_name : NULL); |
| 2632 |
goto EXIT; |
goto EXIT; |
| 2633 |
} |
} |
| 2634 |
|
|
| 2655 |
} |
} |
| 2656 |
if (hints_list != NULL) |
if (hints_list != NULL) |
| 2657 |
{ |
{ |
| 2658 |
for (i = 0; i < hint_count; i++) free(hints_list[i]); |
for (i = 0; i < hint_count; i++) |
| 2659 |
|
{ |
| 2660 |
|
if (hints_list[i] != NULL) free(hints_list[i]); |
| 2661 |
|
} |
| 2662 |
free(hints_list); |
free(hints_list); |
| 2663 |
} |
} |
| 2664 |
return rc; |
pcregrep_exit(rc); |
| 2665 |
|
|
| 2666 |
EXIT2: |
EXIT2: |
| 2667 |
rc = 2; |
rc = 2; |