| 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) 1997-2010 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 |
| 55 |
#include <unistd.h> |
#include <unistd.h> |
| 56 |
#endif |
#endif |
| 57 |
|
|
| 58 |
|
#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" |
#include "pcre.h" |
| 67 |
|
|
| 68 |
#define FALSE 0 |
#define FALSE 0 |
| 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; |
| 170 |
static BOOL file_offsets = FALSE; |
static BOOL file_offsets = FALSE; |
| 171 |
static BOOL hyphenpending = FALSE; |
static BOOL hyphenpending = FALSE; |
| 172 |
static BOOL invert = FALSE; |
static BOOL invert = FALSE; |
| 173 |
|
static BOOL line_buffered = FALSE; |
| 174 |
static BOOL line_offsets = FALSE; |
static BOOL line_offsets = FALSE; |
| 175 |
static BOOL multiline = FALSE; |
static BOOL multiline = FALSE; |
| 176 |
static BOOL number = FALSE; |
static BOOL number = FALSE; |
| 177 |
|
static BOOL omit_zero_count = FALSE; |
| 178 |
static BOOL only_matching = FALSE; |
static BOOL only_matching = FALSE; |
| 179 |
static BOOL quiet = FALSE; |
static BOOL quiet = FALSE; |
| 180 |
static BOOL silent = FALSE; |
static BOOL silent = FALSE; |
| 196 |
/* 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 |
| 197 |
used to identify them. */ |
used to identify them. */ |
| 198 |
|
|
| 199 |
#define N_COLOUR (-1) |
#define N_COLOUR (-1) |
| 200 |
#define N_EXCLUDE (-2) |
#define N_EXCLUDE (-2) |
| 201 |
#define N_HELP (-3) |
#define N_EXCLUDE_DIR (-3) |
| 202 |
#define N_INCLUDE (-4) |
#define N_HELP (-4) |
| 203 |
#define N_LABEL (-5) |
#define N_INCLUDE (-5) |
| 204 |
#define N_LOCALE (-6) |
#define N_INCLUDE_DIR (-6) |
| 205 |
#define N_NULL (-7) |
#define N_LABEL (-7) |
| 206 |
#define N_LOFFSETS (-8) |
#define N_LOCALE (-8) |
| 207 |
#define N_FOFFSETS (-9) |
#define N_NULL (-9) |
| 208 |
|
#define N_LOFFSETS (-10) |
| 209 |
|
#define N_FOFFSETS (-11) |
| 210 |
|
#define N_LBUFFER (-12) |
| 211 |
|
|
| 212 |
static option_item optionlist[] = { |
static option_item optionlist[] = { |
| 213 |
{ OP_NODATA, N_NULL, NULL, "", " terminate options" }, |
{ OP_NODATA, N_NULL, NULL, "", " terminate options" }, |
| 220 |
{ OP_OP_STRING, N_COLOUR, &colour_option, "colour=option", "matched text colour option" }, |
{ OP_OP_STRING, N_COLOUR, &colour_option, "colour=option", "matched text colour option" }, |
| 221 |
{ 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" }, |
| 222 |
{ OP_STRING, 'd', &dee_option, "directories=action", "how to handle directories" }, |
{ OP_STRING, 'd', &dee_option, "directories=action", "how to handle directories" }, |
| 223 |
{ 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)" }, |
| 224 |
{ 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" }, |
| 225 |
{ OP_STRING, 'f', &pattern_filename, "file=path", "read patterns from file" }, |
{ OP_STRING, 'f', &pattern_filename, "file=path", "read patterns from file" }, |
| 226 |
{ OP_NODATA, N_FOFFSETS, NULL, "file-offsets", "output file offsets, not text" }, |
{ OP_NODATA, N_FOFFSETS, NULL, "file-offsets", "output file offsets, not text" }, |
| 227 |
{ OP_NODATA, 'H', NULL, "with-filename", "force the prefixing filename on output" }, |
{ OP_NODATA, 'H', NULL, "with-filename", "force the prefixing filename on output" }, |
| 230 |
{ OP_NODATA, 'l', NULL, "files-with-matches", "print only FILE names containing matches" }, |
{ OP_NODATA, 'l', NULL, "files-with-matches", "print only FILE names containing matches" }, |
| 231 |
{ OP_NODATA, 'L', NULL, "files-without-match","print only FILE names not containing matches" }, |
{ OP_NODATA, 'L', NULL, "files-without-match","print only FILE names not containing matches" }, |
| 232 |
{ OP_STRING, N_LABEL, &stdin_name, "label=name", "set name for standard input" }, |
{ OP_STRING, N_LABEL, &stdin_name, "label=name", "set name for standard input" }, |
| 233 |
|
{ OP_NODATA, N_LBUFFER, NULL, "line-buffered", "use line buffering" }, |
| 234 |
{ OP_NODATA, N_LOFFSETS, NULL, "line-offsets", "output line numbers and offsets, not text" }, |
{ OP_NODATA, N_LOFFSETS, NULL, "line-offsets", "output line numbers and offsets, not text" }, |
| 235 |
{ OP_STRING, N_LOCALE, &locale, "locale=locale", "use the named locale" }, |
{ OP_STRING, N_LOCALE, &locale, "locale=locale", "use the named locale" }, |
| 236 |
{ OP_NODATA, 'M', NULL, "multiline", "run in multiline mode" }, |
{ OP_NODATA, 'M', NULL, "multiline", "run in multiline mode" }, |
| 241 |
{ OP_NODATA, 'r', NULL, "recursive", "recursively scan sub-directories" }, |
{ OP_NODATA, 'r', NULL, "recursive", "recursively scan sub-directories" }, |
| 242 |
{ OP_STRING, N_EXCLUDE,&exclude_pattern, "exclude=pattern","exclude matching files when recursing" }, |
{ OP_STRING, N_EXCLUDE,&exclude_pattern, "exclude=pattern","exclude matching files when recursing" }, |
| 243 |
{ OP_STRING, N_INCLUDE,&include_pattern, "include=pattern","include matching files when recursing" }, |
{ OP_STRING, N_INCLUDE,&include_pattern, "include=pattern","include matching files when recursing" }, |
| 244 |
|
{ OP_STRING, N_EXCLUDE_DIR,&exclude_dir_pattern, "exclude_dir=pattern","exclude matching directories when recursing" }, |
| 245 |
|
{ OP_STRING, N_INCLUDE_DIR,&include_dir_pattern, "include_dir=pattern","include matching directories when recursing" }, |
| 246 |
#ifdef JFRIEDL_DEBUG |
#ifdef JFRIEDL_DEBUG |
| 247 |
{ 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" }, |
| 248 |
#endif |
#endif |
| 342 |
} |
} |
| 343 |
|
|
| 344 |
|
|
| 345 |
/************* Test stdout for being a terminal in Unix **********/ |
/************* Test for a terminal in Unix **********/ |
| 346 |
|
|
| 347 |
static BOOL |
static BOOL |
| 348 |
is_stdout_tty(void) |
is_stdout_tty(void) |
| 350 |
return isatty(fileno(stdout)); |
return isatty(fileno(stdout)); |
| 351 |
} |
} |
| 352 |
|
|
| 353 |
|
static BOOL |
| 354 |
|
is_file_tty(FILE *f) |
| 355 |
|
{ |
| 356 |
|
return isatty(fileno(f)); |
| 357 |
|
} |
| 358 |
|
|
| 359 |
|
|
| 360 |
/************* Directory scanning in Win32 ***********/ |
/************* Directory scanning in Win32 ***********/ |
| 361 |
|
|
| 362 |
/* 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 |
| 363 |
Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES |
Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES |
| 364 |
when it did not exist. David Byron added a patch that moved the #include of |
when it did not exist. David Byron added a patch that moved the #include of |
| 365 |
<windows.h> to before the INVALID_FILE_ATTRIBUTES definition rather than after. |
<windows.h> to before the INVALID_FILE_ATTRIBUTES definition rather than after. |
| 366 |
*/ |
The double test below stops gcc 4.4.4 grumbling that HAVE_WINDOWS_H is |
| 367 |
|
undefined when it is indeed undefined. */ |
| 368 |
|
|
| 369 |
#elif HAVE_WINDOWS_H |
#elif defined HAVE_WINDOWS_H && HAVE_WINDOWS_H |
| 370 |
|
|
| 371 |
#ifndef STRICT |
#ifndef STRICT |
| 372 |
# define STRICT |
# define STRICT |
| 469 |
} |
} |
| 470 |
|
|
| 471 |
|
|
| 472 |
/************* Test stdout for being a terminal in Win32 **********/ |
/************* Test for a terminal in Win32 **********/ |
| 473 |
|
|
| 474 |
/* I don't know how to do this; assume never */ |
/* I don't know how to do this; assume never */ |
| 475 |
|
|
| 479 |
return FALSE; |
return FALSE; |
| 480 |
} |
} |
| 481 |
|
|
| 482 |
|
static BOOL |
| 483 |
|
is_file_tty(FILE *f) |
| 484 |
|
{ |
| 485 |
|
return FALSE; |
| 486 |
|
} |
| 487 |
|
|
| 488 |
|
|
| 489 |
/************* Directory scanning when we can't do it ***********/ |
/************* Directory scanning when we can't do it ***********/ |
| 490 |
|
|
| 507 |
int isregfile(char *filename) { return 1; } |
int isregfile(char *filename) { return 1; } |
| 508 |
|
|
| 509 |
|
|
| 510 |
/************* Test stdout for being a terminal when we can't do it **********/ |
/************* Test for a terminal when we can't do it **********/ |
| 511 |
|
|
| 512 |
static BOOL |
static BOOL |
| 513 |
is_stdout_tty(void) |
is_stdout_tty(void) |
| 515 |
return FALSE; |
return FALSE; |
| 516 |
} |
} |
| 517 |
|
|
| 518 |
|
static BOOL |
| 519 |
|
is_file_tty(FILE *f) |
| 520 |
|
{ |
| 521 |
|
return FALSE; |
| 522 |
|
} |
| 523 |
|
|
| 524 |
#endif |
#endif |
| 525 |
|
|
| 548 |
|
|
| 549 |
|
|
| 550 |
/************************************************* |
/************************************************* |
| 551 |
|
* Read one line of input * |
| 552 |
|
*************************************************/ |
| 553 |
|
|
| 554 |
|
/* Normally, input is read using fread() into a large buffer, so many lines may |
| 555 |
|
be read at once. However, doing this for tty input means that no output appears |
| 556 |
|
until a lot of input has been typed. Instead, tty input is handled line by |
| 557 |
|
line. We cannot use fgets() for this, because it does not stop at a binary |
| 558 |
|
zero, and therefore there is no way of telling how many characters it has read, |
| 559 |
|
because there may be binary zeros embedded in the data. |
| 560 |
|
|
| 561 |
|
Arguments: |
| 562 |
|
buffer the buffer to read into |
| 563 |
|
length the maximum number of characters to read |
| 564 |
|
f the file |
| 565 |
|
|
| 566 |
|
Returns: the number of characters read, zero at end of file |
| 567 |
|
*/ |
| 568 |
|
|
| 569 |
|
static int |
| 570 |
|
read_one_line(char *buffer, int length, FILE *f) |
| 571 |
|
{ |
| 572 |
|
int c; |
| 573 |
|
int yield = 0; |
| 574 |
|
while ((c = fgetc(f)) != EOF) |
| 575 |
|
{ |
| 576 |
|
buffer[yield++] = c; |
| 577 |
|
if (c == '\n' || yield >= length) break; |
| 578 |
|
} |
| 579 |
|
return yield; |
| 580 |
|
} |
| 581 |
|
|
| 582 |
|
|
| 583 |
|
|
| 584 |
|
/************************************************* |
| 585 |
* Find end of line * |
* Find end of line * |
| 586 |
*************************************************/ |
*************************************************/ |
| 587 |
|
|
| 876 |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
| 877 |
if (number) fprintf(stdout, "%d-", lastmatchnumber++); |
if (number) fprintf(stdout, "%d-", lastmatchnumber++); |
| 878 |
pp = end_of_line(pp, endptr, &ellength); |
pp = end_of_line(pp, endptr, &ellength); |
| 879 |
fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
FWRITE(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
| 880 |
lastmatchrestart = pp; |
lastmatchrestart = pp; |
| 881 |
} |
} |
| 882 |
hyphenpending = TRUE; |
hyphenpending = TRUE; |
| 886 |
|
|
| 887 |
|
|
| 888 |
/************************************************* |
/************************************************* |
| 889 |
|
* Apply patterns to subject till one matches * |
| 890 |
|
*************************************************/ |
| 891 |
|
|
| 892 |
|
/* This function is called to run through all patterns, looking for a match. It |
| 893 |
|
is used multiple times for the same subject when colouring is enabled, in order |
| 894 |
|
to find all possible matches. |
| 895 |
|
|
| 896 |
|
Arguments: |
| 897 |
|
matchptr the start of the subject |
| 898 |
|
length the length of the subject to match |
| 899 |
|
offsets the offets vector to fill in |
| 900 |
|
mrc address of where to put the result of pcre_exec() |
| 901 |
|
|
| 902 |
|
Returns: TRUE if there was a match |
| 903 |
|
FALSE if there was no match |
| 904 |
|
invert if there was a non-fatal error |
| 905 |
|
*/ |
| 906 |
|
|
| 907 |
|
static BOOL |
| 908 |
|
match_patterns(char *matchptr, size_t length, int *offsets, int *mrc) |
| 909 |
|
{ |
| 910 |
|
int i; |
| 911 |
|
for (i = 0; i < pattern_count; i++) |
| 912 |
|
{ |
| 913 |
|
*mrc = pcre_exec(pattern_list[i], hints_list[i], matchptr, (int)length, 0, |
| 914 |
|
PCRE_NOTEMPTY, offsets, OFFSET_SIZE); |
| 915 |
|
if (*mrc >= 0) return TRUE; |
| 916 |
|
if (*mrc == PCRE_ERROR_NOMATCH) continue; |
| 917 |
|
fprintf(stderr, "pcregrep: pcre_exec() error %d while matching ", *mrc); |
| 918 |
|
if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1); |
| 919 |
|
fprintf(stderr, "this text:\n"); |
| 920 |
|
FWRITE(matchptr, 1, length, stderr); /* In case binary zero included */ |
| 921 |
|
fprintf(stderr, "\n"); |
| 922 |
|
if (error_count == 0 && |
| 923 |
|
(*mrc == PCRE_ERROR_MATCHLIMIT || *mrc == PCRE_ERROR_RECURSIONLIMIT)) |
| 924 |
|
{ |
| 925 |
|
fprintf(stderr, "pcregrep: error %d means that a resource limit " |
| 926 |
|
"was exceeded\n", *mrc); |
| 927 |
|
fprintf(stderr, "pcregrep: check your regex for nested unlimited loops\n"); |
| 928 |
|
} |
| 929 |
|
if (error_count++ > 20) |
| 930 |
|
{ |
| 931 |
|
fprintf(stderr, "pcregrep: too many errors - abandoned\n"); |
| 932 |
|
exit(2); |
| 933 |
|
} |
| 934 |
|
return invert; /* No more matching; don't show the line again */ |
| 935 |
|
} |
| 936 |
|
|
| 937 |
|
return FALSE; /* No match, no errors */ |
| 938 |
|
} |
| 939 |
|
|
| 940 |
|
|
| 941 |
|
|
| 942 |
|
/************************************************* |
| 943 |
* Grep an individual file * |
* Grep an individual file * |
| 944 |
*************************************************/ |
*************************************************/ |
| 945 |
|
|
| 951 |
"before" context printing. |
"before" context printing. |
| 952 |
|
|
| 953 |
Arguments: |
Arguments: |
| 954 |
in the fopened FILE stream |
handle the fopened FILE stream for a normal file |
| 955 |
|
the gzFile pointer when reading is via libz |
| 956 |
|
the BZFILE pointer when reading is via libbz2 |
| 957 |
|
frtype FR_PLAIN, FR_LIBZ, or FR_LIBBZ2 |
| 958 |
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 |
| 959 |
or NULL if the file name is not to be printed |
or NULL if the file name is not to be printed |
| 960 |
it cannot be NULL if filenames[_nomatch]_only is set |
it cannot be NULL if filenames[_nomatch]_only is set |
| 961 |
|
|
| 962 |
Returns: 0 if there was at least one match |
Returns: 0 if there was at least one match |
| 963 |
1 otherwise (no matches) |
1 otherwise (no matches) |
| 964 |
|
2 if there is a read error on a .bz2 file |
| 965 |
*/ |
*/ |
| 966 |
|
|
| 967 |
static int |
static int |
| 968 |
pcregrep(FILE *in, char *printname) |
pcregrep(void *handle, int frtype, char *printname) |
| 969 |
{ |
{ |
| 970 |
int rc = 1; |
int rc = 1; |
| 971 |
int linenumber = 1; |
int linenumber = 1; |
| 972 |
int lastmatchnumber = 0; |
int lastmatchnumber = 0; |
| 973 |
int count = 0; |
int count = 0; |
| 974 |
int filepos = 0; |
int filepos = 0; |
| 975 |
int offsets[99]; |
int offsets[OFFSET_SIZE]; |
| 976 |
char *lastmatchrestart = NULL; |
char *lastmatchrestart = NULL; |
| 977 |
char buffer[3*MBUFTHIRD]; |
char buffer[3*MBUFTHIRD]; |
| 978 |
char *ptr = buffer; |
char *ptr = buffer; |
| 979 |
char *endptr; |
char *endptr; |
| 980 |
size_t bufflength; |
size_t bufflength; |
| 981 |
BOOL endhyphenpending = FALSE; |
BOOL endhyphenpending = FALSE; |
| 982 |
|
BOOL input_line_buffered = line_buffered; |
| 983 |
|
FILE *in = NULL; /* Ensure initialized */ |
| 984 |
|
|
| 985 |
|
#ifdef SUPPORT_LIBZ |
| 986 |
|
gzFile ingz = NULL; |
| 987 |
|
#endif |
| 988 |
|
|
| 989 |
|
#ifdef SUPPORT_LIBBZ2 |
| 990 |
|
BZFILE *inbz2 = NULL; |
| 991 |
|
#endif |
| 992 |
|
|
| 993 |
|
|
| 994 |
|
/* Do the first read into the start of the buffer and set up the pointer to end |
| 995 |
|
of what we have. In the case of libz, a non-zipped .gz file will be read as a |
| 996 |
|
plain file. However, if a .bz2 file isn't actually bzipped, the first read will |
| 997 |
|
fail. */ |
| 998 |
|
|
| 999 |
|
#ifdef SUPPORT_LIBZ |
| 1000 |
|
if (frtype == FR_LIBZ) |
| 1001 |
|
{ |
| 1002 |
|
ingz = (gzFile)handle; |
| 1003 |
|
bufflength = gzread (ingz, buffer, 3*MBUFTHIRD); |
| 1004 |
|
} |
| 1005 |
|
else |
| 1006 |
|
#endif |
| 1007 |
|
|
| 1008 |
/* Do the first read into the start of the buffer and set up the pointer to |
#ifdef SUPPORT_LIBBZ2 |
| 1009 |
end of what we have. */ |
if (frtype == FR_LIBBZ2) |
| 1010 |
|
{ |
| 1011 |
|
inbz2 = (BZFILE *)handle; |
| 1012 |
|
bufflength = BZ2_bzread(inbz2, buffer, 3*MBUFTHIRD); |
| 1013 |
|
if ((int)bufflength < 0) return 2; /* Gotcha: bufflength is size_t; */ |
| 1014 |
|
} /* without the cast it is unsigned. */ |
| 1015 |
|
else |
| 1016 |
|
#endif |
| 1017 |
|
|
| 1018 |
|
{ |
| 1019 |
|
in = (FILE *)handle; |
| 1020 |
|
if (is_file_tty(in)) input_line_buffered = TRUE; |
| 1021 |
|
bufflength = input_line_buffered? |
| 1022 |
|
read_one_line(buffer, 3*MBUFTHIRD, in) : |
| 1023 |
|
fread(buffer, 1, 3*MBUFTHIRD, in); |
| 1024 |
|
} |
| 1025 |
|
|
|
bufflength = fread(buffer, 1, 3*MBUFTHIRD, in); |
|
| 1026 |
endptr = buffer + bufflength; |
endptr = buffer + bufflength; |
| 1027 |
|
|
| 1028 |
/* 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 |
| 1032 |
|
|
| 1033 |
while (ptr < endptr) |
while (ptr < endptr) |
| 1034 |
{ |
{ |
| 1035 |
int i, endlinelength; |
int endlinelength; |
| 1036 |
int mrc = 0; |
int mrc = 0; |
| 1037 |
BOOL match = FALSE; |
BOOL match; |
| 1038 |
char *matchptr = ptr; |
char *matchptr = ptr; |
| 1039 |
char *t = ptr; |
char *t = ptr; |
| 1040 |
size_t length, linelength; |
size_t length, linelength; |
| 1041 |
|
|
| 1042 |
/* 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 |
| 1043 |
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 |
| 1044 |
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 |
| 1045 |
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 |
| 1046 |
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 |
| 1047 |
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 |
| 1048 |
|
first line. */ |
| 1049 |
|
|
| 1050 |
t = end_of_line(t, endptr, &endlinelength); |
t = end_of_line(t, endptr, &endlinelength); |
| 1051 |
linelength = t - ptr - endlinelength; |
linelength = t - ptr - endlinelength; |
| 1060 |
#include <time.h> |
#include <time.h> |
| 1061 |
struct timeval start_time, end_time; |
struct timeval start_time, end_time; |
| 1062 |
struct timezone dummy; |
struct timezone dummy; |
| 1063 |
|
int i; |
| 1064 |
|
|
| 1065 |
if (jfriedl_XT) |
if (jfriedl_XT) |
| 1066 |
{ |
{ |
| 1086 |
|
|
| 1087 |
|
|
| 1088 |
for (i = 0; i < jfriedl_XR; i++) |
for (i = 0; i < jfriedl_XR; i++) |
| 1089 |
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, |
| 1090 |
|
PCRE_NOTEMPTY, offsets, OFFSET_SIZE) >= 0); |
| 1091 |
|
|
| 1092 |
if (gettimeofday(&end_time, &dummy) != 0) |
if (gettimeofday(&end_time, &dummy) != 0) |
| 1093 |
perror("bad gettimeofday"); |
perror("bad gettimeofday"); |
| 1101 |
} |
} |
| 1102 |
#endif |
#endif |
| 1103 |
|
|
| 1104 |
/* We come back here after a match when the -o option (only_matching) is set, |
/* We come back here after a match when the -o option (only_matching) is set, |
| 1105 |
in order to find any further matches in the same line. */ |
in order to find any further matches in the same line. */ |
|
|
|
|
ONLY_MATCHING_RESTART: |
|
| 1106 |
|
|
| 1107 |
/* Run through all the patterns until one matches. Note that we don't include |
ONLY_MATCHING_RESTART: |
|
the final newline in the subject string. */ |
|
| 1108 |
|
|
| 1109 |
for (i = 0; i < pattern_count; i++) |
/* Run through all the patterns until one matches or there is an error other |
| 1110 |
{ |
than NOMATCH. This code is in a subroutine so that it can be re-used for |
| 1111 |
mrc = pcre_exec(pattern_list[i], hints_list[i], matchptr, length, 0, 0, |
finding subsequent matches when colouring matched lines. */ |
| 1112 |
offsets, 99); |
|
| 1113 |
if (mrc >= 0) { match = TRUE; break; } |
match = match_patterns(matchptr, length, 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(matchptr, 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; |
|
|
} |
|
|
} |
|
| 1114 |
|
|
| 1115 |
/* 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. */ |
| 1116 |
|
|
| 1129 |
/* 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 |
| 1130 |
in the file. */ |
in the file. */ |
| 1131 |
|
|
| 1132 |
else if (filenames == FN_ONLY) |
else if (filenames == FN_MATCH_ONLY) |
| 1133 |
{ |
{ |
| 1134 |
fprintf(stdout, "%s\n", printname); |
fprintf(stdout, "%s\n", printname); |
| 1135 |
return 0; |
return 0; |
| 1140 |
else if (quiet) return 0; |
else if (quiet) return 0; |
| 1141 |
|
|
| 1142 |
/* The --only-matching option prints just the substring that matched, and |
/* The --only-matching option prints just the substring that matched, and |
| 1143 |
the --file-offsets and --line-offsets options output offsets for the |
the --file-offsets and --line-offsets options output offsets for the |
| 1144 |
matching substring (they both force --only-matching). None of these options |
matching substring (they both force --only-matching). None of these options |
| 1145 |
prints any context. Afterwards, adjust the start and length, and then jump |
prints any context. Afterwards, adjust the start and length, and then jump |
| 1146 |
back to look for further matches in the same line. If we are in invert |
back to look for further matches in the same line. If we are in invert |
| 1150 |
else if (only_matching) |
else if (only_matching) |
| 1151 |
{ |
{ |
| 1152 |
if (!invert) |
if (!invert) |
| 1153 |
{ |
{ |
| 1154 |
if (printname != NULL) fprintf(stdout, "%s:", printname); |
if (printname != NULL) fprintf(stdout, "%s:", printname); |
| 1155 |
if (number) fprintf(stdout, "%d:", linenumber); |
if (number) fprintf(stdout, "%d:", linenumber); |
| 1156 |
if (line_offsets) |
if (line_offsets) |
| 1157 |
fprintf(stdout, "%d,%d", matchptr + offsets[0] - ptr, |
fprintf(stdout, "%d,%d", (int)(matchptr + offsets[0] - ptr), |
| 1158 |
offsets[1] - offsets[0]); |
offsets[1] - offsets[0]); |
| 1159 |
else if (file_offsets) |
else if (file_offsets) |
| 1160 |
fprintf(stdout, "%d,%d", filepos + matchptr + offsets[0] - ptr, |
fprintf(stdout, "%d,%d", (int)(filepos + matchptr + offsets[0] - ptr), |
| 1161 |
offsets[1] - offsets[0]); |
offsets[1] - offsets[0]); |
| 1162 |
else |
else |
| 1163 |
fwrite(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
{ |
| 1164 |
|
if (do_colour) fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
| 1165 |
|
FWRITE(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
| 1166 |
|
if (do_colour) fprintf(stdout, "%c[00m", 0x1b); |
| 1167 |
|
} |
| 1168 |
fprintf(stdout, "\n"); |
fprintf(stdout, "\n"); |
| 1169 |
matchptr += offsets[1]; |
matchptr += offsets[1]; |
| 1170 |
length -= offsets[1]; |
length -= offsets[1]; |
| 1171 |
match = FALSE; |
match = FALSE; |
| 1172 |
goto ONLY_MATCHING_RESTART; |
goto ONLY_MATCHING_RESTART; |
| 1173 |
} |
} |
| 1174 |
} |
} |
| 1175 |
|
|
| 1176 |
/* 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 |
| 1204 |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
| 1205 |
if (number) fprintf(stdout, "%d-", lastmatchnumber++); |
if (number) fprintf(stdout, "%d-", lastmatchnumber++); |
| 1206 |
pp = end_of_line(pp, endptr, &ellength); |
pp = end_of_line(pp, endptr, &ellength); |
| 1207 |
fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
FWRITE(lastmatchrestart, 1, pp - lastmatchrestart, stdout); |
| 1208 |
lastmatchrestart = pp; |
lastmatchrestart = pp; |
| 1209 |
} |
} |
| 1210 |
if (lastmatchrestart != ptr) hyphenpending = TRUE; |
if (lastmatchrestart != ptr) hyphenpending = TRUE; |
| 1244 |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
if (printname != NULL) fprintf(stdout, "%s-", printname); |
| 1245 |
if (number) fprintf(stdout, "%d-", linenumber - linecount--); |
if (number) fprintf(stdout, "%d-", linenumber - linecount--); |
| 1246 |
pp = end_of_line(pp, endptr, &ellength); |
pp = end_of_line(pp, endptr, &ellength); |
| 1247 |
fwrite(p, 1, pp - p, stdout); |
FWRITE(p, 1, pp - p, stdout); |
| 1248 |
p = pp; |
p = pp; |
| 1249 |
} |
} |
| 1250 |
} |
} |
| 1294 |
{ |
{ |
| 1295 |
int first = S_arg * 2; |
int first = S_arg * 2; |
| 1296 |
int last = first + 1; |
int last = first + 1; |
| 1297 |
fwrite(ptr, 1, offsets[first], stdout); |
FWRITE(ptr, 1, offsets[first], stdout); |
| 1298 |
fprintf(stdout, "X"); |
fprintf(stdout, "X"); |
| 1299 |
fwrite(ptr + offsets[last], 1, linelength - offsets[last], stdout); |
FWRITE(ptr + offsets[last], 1, linelength - offsets[last], stdout); |
| 1300 |
} |
} |
| 1301 |
else |
else |
| 1302 |
#endif |
#endif |
| 1303 |
|
|
| 1304 |
/* We have to split the line(s) up if colouring. */ |
/* We have to split the line(s) up if colouring, and search for further |
| 1305 |
|
matches. */ |
| 1306 |
|
|
| 1307 |
if (do_colour) |
if (do_colour) |
| 1308 |
{ |
{ |
| 1309 |
fwrite(ptr, 1, offsets[0], stdout); |
int last_offset = 0; |
| 1310 |
|
FWRITE(ptr, 1, offsets[0], stdout); |
| 1311 |
fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
| 1312 |
fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
FWRITE(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
| 1313 |
fprintf(stdout, "%c[00m", 0x1b); |
fprintf(stdout, "%c[00m", 0x1b); |
| 1314 |
fwrite(ptr + offsets[1], 1, (linelength + endlinelength) - offsets[1], |
for (;;) |
| 1315 |
stdout); |
{ |
| 1316 |
|
last_offset += offsets[1]; |
| 1317 |
|
matchptr += offsets[1]; |
| 1318 |
|
length -= offsets[1]; |
| 1319 |
|
if (!match_patterns(matchptr, length, offsets, &mrc)) break; |
| 1320 |
|
FWRITE(matchptr, 1, offsets[0], stdout); |
| 1321 |
|
fprintf(stdout, "%c[%sm", 0x1b, colour_string); |
| 1322 |
|
FWRITE(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout); |
| 1323 |
|
fprintf(stdout, "%c[00m", 0x1b); |
| 1324 |
|
} |
| 1325 |
|
FWRITE(ptr + last_offset, 1, |
| 1326 |
|
(linelength + endlinelength) - last_offset, stdout); |
| 1327 |
} |
} |
| 1328 |
else fwrite(ptr, 1, linelength + endlinelength, stdout); |
|
| 1329 |
|
/* Not colouring; no need to search for further matches */ |
| 1330 |
|
|
| 1331 |
|
else FWRITE(ptr, 1, linelength + endlinelength, stdout); |
| 1332 |
} |
} |
| 1333 |
|
|
| 1334 |
/* 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 |
| 1335 |
|
given, flush the output. */ |
| 1336 |
|
|
| 1337 |
|
if (line_buffered) fflush(stdout); |
| 1338 |
rc = 0; /* Had some success */ |
rc = 0; /* Had some success */ |
| 1339 |
|
|
| 1340 |
/* Remember where the last match happened for after_context. We remember |
/* Remember where the last match happened for after_context. We remember |
| 1362 |
linelength = endmatch - ptr - ellength; |
linelength = endmatch - ptr - ellength; |
| 1363 |
} |
} |
| 1364 |
|
|
| 1365 |
/* Advance to after the newline and increment the line number. The file |
/* Advance to after the newline and increment the line number. The file |
| 1366 |
offset to the current line is maintained in filepos. */ |
offset to the current line is maintained in filepos. */ |
| 1367 |
|
|
| 1368 |
ptr += linelength + endlinelength; |
ptr += linelength + endlinelength; |
| 1369 |
filepos += linelength + endlinelength; |
filepos += (int)(linelength + endlinelength); |
| 1370 |
linenumber++; |
linenumber++; |
| 1371 |
|
|
| 1372 |
|
/* If input is line buffered, and the buffer is not yet full, read another |
| 1373 |
|
line and add it into the buffer. */ |
| 1374 |
|
|
| 1375 |
|
if (input_line_buffered && bufflength < sizeof(buffer)) |
| 1376 |
|
{ |
| 1377 |
|
int add = read_one_line(ptr, sizeof(buffer) - (ptr - buffer), in); |
| 1378 |
|
bufflength += add; |
| 1379 |
|
endptr += add; |
| 1380 |
|
} |
| 1381 |
|
|
| 1382 |
/* 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 |
| 1383 |
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 |
| 1384 |
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 |
| 1398 |
|
|
| 1399 |
memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD); |
memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD); |
| 1400 |
ptr -= MBUFTHIRD; |
ptr -= MBUFTHIRD; |
| 1401 |
bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in); |
|
| 1402 |
|
#ifdef SUPPORT_LIBZ |
| 1403 |
|
if (frtype == FR_LIBZ) |
| 1404 |
|
bufflength = 2*MBUFTHIRD + |
| 1405 |
|
gzread (ingz, buffer + 2*MBUFTHIRD, MBUFTHIRD); |
| 1406 |
|
else |
| 1407 |
|
#endif |
| 1408 |
|
|
| 1409 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1410 |
|
if (frtype == FR_LIBBZ2) |
| 1411 |
|
bufflength = 2*MBUFTHIRD + |
| 1412 |
|
BZ2_bzread(inbz2, buffer + 2*MBUFTHIRD, MBUFTHIRD); |
| 1413 |
|
else |
| 1414 |
|
#endif |
| 1415 |
|
|
| 1416 |
|
bufflength = 2*MBUFTHIRD + |
| 1417 |
|
(input_line_buffered? |
| 1418 |
|
read_one_line(buffer + 2*MBUFTHIRD, MBUFTHIRD, in) : |
| 1419 |
|
fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in)); |
| 1420 |
endptr = buffer + bufflength; |
endptr = buffer + bufflength; |
| 1421 |
|
|
| 1422 |
/* Adjust any last match point */ |
/* Adjust any last match point */ |
| 1447 |
|
|
| 1448 |
if (count_only) |
if (count_only) |
| 1449 |
{ |
{ |
| 1450 |
if (printname != NULL) fprintf(stdout, "%s:", printname); |
if (count > 0 || !omit_zero_count) |
| 1451 |
fprintf(stdout, "%d\n", count); |
{ |
| 1452 |
|
if (printname != NULL && filenames != FN_NONE) |
| 1453 |
|
fprintf(stdout, "%s:", printname); |
| 1454 |
|
fprintf(stdout, "%d\n", count); |
| 1455 |
|
} |
| 1456 |
} |
} |
| 1457 |
|
|
| 1458 |
return rc; |
return rc; |
| 1484 |
{ |
{ |
| 1485 |
int rc = 1; |
int rc = 1; |
| 1486 |
int sep; |
int sep; |
| 1487 |
FILE *in; |
int frtype; |
| 1488 |
|
int pathlen; |
| 1489 |
|
void *handle; |
| 1490 |
|
FILE *in = NULL; /* Ensure initialized */ |
| 1491 |
|
|
| 1492 |
|
#ifdef SUPPORT_LIBZ |
| 1493 |
|
gzFile ingz = NULL; |
| 1494 |
|
#endif |
| 1495 |
|
|
| 1496 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1497 |
|
BZFILE *inbz2 = NULL; |
| 1498 |
|
#endif |
| 1499 |
|
|
| 1500 |
/* If the file name is "-" we scan stdin */ |
/* If the file name is "-" we scan stdin */ |
| 1501 |
|
|
| 1502 |
if (strcmp(pathname, "-") == 0) |
if (strcmp(pathname, "-") == 0) |
| 1503 |
{ |
{ |
| 1504 |
return pcregrep(stdin, |
return pcregrep(stdin, FR_PLAIN, |
| 1505 |
(filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))? |
(filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))? |
| 1506 |
stdin_name : NULL); |
stdin_name : NULL); |
| 1507 |
} |
} |
| 1508 |
|
|
|
|
|
| 1509 |
/* 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 |
| 1510 |
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 |
| 1511 |
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 |
| 1512 |
|
system-specific. */ |
| 1513 |
|
|
| 1514 |
if ((sep = isdirectory(pathname)) != 0) |
if ((sep = isdirectory(pathname)) != 0) |
| 1515 |
{ |
{ |
| 1530 |
|
|
| 1531 |
while ((nextfile = readdirectory(dir)) != NULL) |
while ((nextfile = readdirectory(dir)) != NULL) |
| 1532 |
{ |
{ |
| 1533 |
int frc, blen; |
int frc, nflen; |
| 1534 |
sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile); |
sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile); |
| 1535 |
blen = strlen(buffer); |
nflen = (int)(strlen(nextfile)); |
| 1536 |
|
|
| 1537 |
|
if (isdirectory(buffer)) |
| 1538 |
|
{ |
| 1539 |
|
if (exclude_dir_compiled != NULL && |
| 1540 |
|
pcre_exec(exclude_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0) |
| 1541 |
|
continue; |
| 1542 |
|
|
| 1543 |
if (exclude_compiled != NULL && |
if (include_dir_compiled != NULL && |
| 1544 |
pcre_exec(exclude_compiled, NULL, buffer, blen, 0, 0, NULL, 0) >= 0) |
pcre_exec(include_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0) |
| 1545 |
continue; |
continue; |
| 1546 |
|
} |
| 1547 |
if (include_compiled != NULL && |
else |
| 1548 |
pcre_exec(include_compiled, NULL, buffer, blen, 0, 0, NULL, 0) < 0) |
{ |
| 1549 |
continue; |
if (exclude_compiled != NULL && |
| 1550 |
|
pcre_exec(exclude_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0) |
| 1551 |
|
continue; |
| 1552 |
|
|
| 1553 |
|
if (include_compiled != NULL && |
| 1554 |
|
pcre_exec(include_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0) |
| 1555 |
|
continue; |
| 1556 |
|
} |
| 1557 |
|
|
| 1558 |
frc = grep_or_recurse(buffer, dir_recurse, FALSE); |
frc = grep_or_recurse(buffer, dir_recurse, FALSE); |
| 1559 |
if (frc > 1) rc = frc; |
if (frc > 1) rc = frc; |
| 1576 |
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 |
| 1577 |
the file name, or the filename was forced (-H). */ |
the file name, or the filename was forced (-H). */ |
| 1578 |
|
|
| 1579 |
in = fopen(pathname, "r"); |
pathlen = (int)(strlen(pathname)); |
| 1580 |
if (in == NULL) |
|
| 1581 |
|
/* Open using zlib if it is supported and the file name ends with .gz. */ |
| 1582 |
|
|
| 1583 |
|
#ifdef SUPPORT_LIBZ |
| 1584 |
|
if (pathlen > 3 && strcmp(pathname + pathlen - 3, ".gz") == 0) |
| 1585 |
|
{ |
| 1586 |
|
ingz = gzopen(pathname, "rb"); |
| 1587 |
|
if (ingz == NULL) |
| 1588 |
|
{ |
| 1589 |
|
if (!silent) |
| 1590 |
|
fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname, |
| 1591 |
|
strerror(errno)); |
| 1592 |
|
return 2; |
| 1593 |
|
} |
| 1594 |
|
handle = (void *)ingz; |
| 1595 |
|
frtype = FR_LIBZ; |
| 1596 |
|
} |
| 1597 |
|
else |
| 1598 |
|
#endif |
| 1599 |
|
|
| 1600 |
|
/* Otherwise open with bz2lib if it is supported and the name ends with .bz2. */ |
| 1601 |
|
|
| 1602 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1603 |
|
if (pathlen > 4 && strcmp(pathname + pathlen - 4, ".bz2") == 0) |
| 1604 |
|
{ |
| 1605 |
|
inbz2 = BZ2_bzopen(pathname, "rb"); |
| 1606 |
|
handle = (void *)inbz2; |
| 1607 |
|
frtype = FR_LIBBZ2; |
| 1608 |
|
} |
| 1609 |
|
else |
| 1610 |
|
#endif |
| 1611 |
|
|
| 1612 |
|
/* Otherwise use plain fopen(). The label is so that we can come back here if |
| 1613 |
|
an attempt to read a .bz2 file indicates that it really is a plain file. */ |
| 1614 |
|
|
| 1615 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1616 |
|
PLAIN_FILE: |
| 1617 |
|
#endif |
| 1618 |
|
{ |
| 1619 |
|
in = fopen(pathname, "rb"); |
| 1620 |
|
handle = (void *)in; |
| 1621 |
|
frtype = FR_PLAIN; |
| 1622 |
|
} |
| 1623 |
|
|
| 1624 |
|
/* All the opening methods return errno when they fail. */ |
| 1625 |
|
|
| 1626 |
|
if (handle == NULL) |
| 1627 |
{ |
{ |
| 1628 |
if (!silent) |
if (!silent) |
| 1629 |
fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname, |
fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname, |
| 1631 |
return 2; |
return 2; |
| 1632 |
} |
} |
| 1633 |
|
|
| 1634 |
rc = pcregrep(in, (filenames > FN_DEFAULT || |
/* Now grep the file */ |
| 1635 |
|
|
| 1636 |
|
rc = pcregrep(handle, frtype, (filenames > FN_DEFAULT || |
| 1637 |
(filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL); |
(filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL); |
| 1638 |
|
|
| 1639 |
|
/* Close in an appropriate manner. */ |
| 1640 |
|
|
| 1641 |
|
#ifdef SUPPORT_LIBZ |
| 1642 |
|
if (frtype == FR_LIBZ) |
| 1643 |
|
gzclose(ingz); |
| 1644 |
|
else |
| 1645 |
|
#endif |
| 1646 |
|
|
| 1647 |
|
/* If it is a .bz2 file and the result is 2, it means that the first attempt to |
| 1648 |
|
read failed. If the error indicates that the file isn't in fact bzipped, try |
| 1649 |
|
again as a normal file. */ |
| 1650 |
|
|
| 1651 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1652 |
|
if (frtype == FR_LIBBZ2) |
| 1653 |
|
{ |
| 1654 |
|
if (rc == 2) |
| 1655 |
|
{ |
| 1656 |
|
int errnum; |
| 1657 |
|
const char *err = BZ2_bzerror(inbz2, &errnum); |
| 1658 |
|
if (errnum == BZ_DATA_ERROR_MAGIC) |
| 1659 |
|
{ |
| 1660 |
|
BZ2_bzclose(inbz2); |
| 1661 |
|
goto PLAIN_FILE; |
| 1662 |
|
} |
| 1663 |
|
else if (!silent) |
| 1664 |
|
fprintf(stderr, "pcregrep: Failed to read %s using bzlib: %s\n", |
| 1665 |
|
pathname, err); |
| 1666 |
|
} |
| 1667 |
|
BZ2_bzclose(inbz2); |
| 1668 |
|
} |
| 1669 |
|
else |
| 1670 |
|
#endif |
| 1671 |
|
|
| 1672 |
|
/* Normal file close */ |
| 1673 |
|
|
| 1674 |
fclose(in); |
fclose(in); |
| 1675 |
|
|
| 1676 |
|
/* Pass back the yield from pcregrep(). */ |
| 1677 |
|
|
| 1678 |
return rc; |
return rc; |
| 1679 |
} |
} |
| 1680 |
|
|
| 1715 |
printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n"); |
printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n"); |
| 1716 |
printf("Search for PATTERN in each FILE or standard input.\n"); |
printf("Search for PATTERN in each FILE or standard input.\n"); |
| 1717 |
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"); |
| 1718 |
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"); |
| 1719 |
printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n"); |
|
| 1720 |
|
#ifdef SUPPORT_LIBZ |
| 1721 |
|
printf("Files whose names end in .gz are read using zlib.\n"); |
| 1722 |
|
#endif |
| 1723 |
|
|
| 1724 |
|
#ifdef SUPPORT_LIBBZ2 |
| 1725 |
|
printf("Files whose names end in .bz2 are read using bzlib2.\n"); |
| 1726 |
|
#endif |
| 1727 |
|
|
| 1728 |
|
#if defined SUPPORT_LIBZ || defined SUPPORT_LIBBZ2 |
| 1729 |
|
printf("Other files and the standard input are read as plain files.\n\n"); |
| 1730 |
|
#else |
| 1731 |
|
printf("All files are read as plain files, without any interpretation.\n\n"); |
| 1732 |
|
#endif |
| 1733 |
|
|
| 1734 |
|
printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n"); |
| 1735 |
printf("Options:\n"); |
printf("Options:\n"); |
| 1736 |
|
|
| 1737 |
for (op = optionlist; op->one_char != 0; op++) |
for (op = optionlist; op->one_char != 0; op++) |
| 1739 |
int n; |
int n; |
| 1740 |
char s[4]; |
char s[4]; |
| 1741 |
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, " "); |
| 1742 |
printf(" %s --%s%n", s, op->long_name, &n); |
n = 30 - printf(" %s --%s", s, op->long_name); |
|
n = 30 - n; |
|
| 1743 |
if (n < 1) n = 1; |
if (n < 1) n = 1; |
| 1744 |
printf("%.*s%s\n", n, " ", op->help_text); |
printf("%.*s%s\n", n, " ", op->help_text); |
| 1745 |
} |
} |
| 1764 |
{ |
{ |
| 1765 |
switch(letter) |
switch(letter) |
| 1766 |
{ |
{ |
| 1767 |
case N_FOFFSETS: file_offsets = TRUE; break; |
case N_FOFFSETS: file_offsets = TRUE; break; |
| 1768 |
case N_HELP: help(); exit(0); |
case N_HELP: help(); exit(0); |
| 1769 |
case N_LOFFSETS: line_offsets = number = TRUE; break; |
case N_LOFFSETS: line_offsets = number = TRUE; break; |
| 1770 |
|
case N_LBUFFER: line_buffered = TRUE; break; |
| 1771 |
case 'c': count_only = TRUE; break; |
case 'c': count_only = TRUE; break; |
| 1772 |
case 'F': process_options |= PO_FIXED_STRINGS; break; |
case 'F': process_options |= PO_FIXED_STRINGS; break; |
| 1773 |
case 'H': filenames = FN_FORCE; break; |
case 'H': filenames = FN_FORCE; break; |
| 1774 |
case 'h': filenames = FN_NONE; break; |
case 'h': filenames = FN_NONE; break; |
| 1775 |
case 'i': options |= PCRE_CASELESS; break; |
case 'i': options |= PCRE_CASELESS; break; |
| 1776 |
case 'l': filenames = FN_ONLY; break; |
case 'l': omit_zero_count = TRUE; filenames = FN_MATCH_ONLY; break; |
| 1777 |
case 'L': filenames = FN_NOMATCH_ONLY; break; |
case 'L': filenames = FN_NOMATCH_ONLY; break; |
| 1778 |
case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break; |
case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break; |
| 1779 |
case 'n': number = TRUE; break; |
case 'n': number = TRUE; break; |
| 1958 |
|
|
| 1959 |
/* 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; |
| 1960 |
"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". |
| 1961 |
*/ |
Note that the return values from pcre_config(), though derived from the ASCII |
| 1962 |
|
codes, are the same in EBCDIC environments, so we must use the actual values |
| 1963 |
|
rather than escapes such as as '\r'. */ |
| 1964 |
|
|
| 1965 |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &i); |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &i); |
| 1966 |
switch(i) |
switch(i) |
| 1967 |
{ |
{ |
| 1968 |
default: newline = (char *)"lf"; break; |
default: newline = (char *)"lf"; break; |
| 1969 |
case '\r': newline = (char *)"cr"; break; |
case 13: newline = (char *)"cr"; break; |
| 1970 |
case ('\r' << 8) | '\n': newline = (char *)"crlf"; break; |
case (13 << 8) | 10: newline = (char *)"crlf"; break; |
| 1971 |
case -1: newline = (char *)"any"; break; |
case -1: newline = (char *)"any"; break; |
| 1972 |
case -2: newline = (char *)"anycrlf"; break; |
case -2: newline = (char *)"anycrlf"; break; |
| 1973 |
} |
} |
| 1974 |
|
|
| 1975 |
/* Process the options */ |
/* Process the options */ |
| 2011 |
Some options have variations in the long name spelling: specifically, we |
Some options have variations in the long name spelling: specifically, we |
| 2012 |
allow "regexp" because GNU grep allows it, though I personally go along |
allow "regexp" because GNU grep allows it, though I personally go along |
| 2013 |
with Jeffrey Friedl and Larry Wall in preferring "regex" without the "p". |
with Jeffrey Friedl and Larry Wall in preferring "regex" without the "p". |
| 2014 |
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 |
| 2015 |
these categories, fortunately. */ |
both these categories. */ |
| 2016 |
|
|
| 2017 |
for (op = optionlist; op->one_char != 0; op++) |
for (op = optionlist; op->one_char != 0; op++) |
| 2018 |
{ |
{ |
| 2019 |
char *opbra = strchr(op->long_name, '('); |
char *opbra = strchr(op->long_name, '('); |
| 2020 |
char *equals = strchr(op->long_name, '='); |
char *equals = strchr(op->long_name, '='); |
| 2021 |
if (opbra == NULL) /* Not a (p) case */ |
|
| 2022 |
|
/* Handle options with only one spelling of the name */ |
| 2023 |
|
|
| 2024 |
|
if (opbra == NULL) /* Does not contain '(' */ |
| 2025 |
{ |
{ |
| 2026 |
if (equals == NULL) /* Not thing=data case */ |
if (equals == NULL) /* Not thing=data case */ |
| 2027 |
{ |
{ |
| 2029 |
} |
} |
| 2030 |
else /* Special case xxx=data */ |
else /* Special case xxx=data */ |
| 2031 |
{ |
{ |
| 2032 |
int oplen = equals - op->long_name; |
int oplen = (int)(equals - op->long_name); |
| 2033 |
int arglen = (argequals == NULL)? (int)strlen(arg) : argequals - arg; |
int arglen = (argequals == NULL)? |
| 2034 |
|
(int)strlen(arg) : (int)(argequals - arg); |
| 2035 |
if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0) |
if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0) |
| 2036 |
{ |
{ |
| 2037 |
option_data = arg + arglen; |
option_data = arg + arglen; |
| 2044 |
} |
} |
| 2045 |
} |
} |
| 2046 |
} |
} |
| 2047 |
else /* Special case xxxx(p) */ |
|
| 2048 |
|
/* Handle options with an alternate spelling of the name */ |
| 2049 |
|
|
| 2050 |
|
else |
| 2051 |
{ |
{ |
| 2052 |
char buff1[24]; |
char buff1[24]; |
| 2053 |
char buff2[24]; |
char buff2[24]; |
| 2054 |
int baselen = opbra - op->long_name; |
|
| 2055 |
|
int baselen = (int)(opbra - op->long_name); |
| 2056 |
|
int fulllen = (int)(strchr(op->long_name, ')') - op->long_name + 1); |
| 2057 |
|
int arglen = (argequals == NULL || equals == NULL)? |
| 2058 |
|
(int)strlen(arg) : (int)(argequals - arg); |
| 2059 |
|
|
| 2060 |
sprintf(buff1, "%.*s", baselen, op->long_name); |
sprintf(buff1, "%.*s", baselen, op->long_name); |
| 2061 |
sprintf(buff2, "%s%.*s", buff1, |
sprintf(buff2, "%s%.*s", buff1, fulllen - baselen - 2, opbra + 1); |
| 2062 |
(int)strlen(op->long_name) - baselen - 2, opbra + 1); |
|
| 2063 |
if (strcmp(arg, buff1) == 0 || strcmp(arg, buff2) == 0) |
if (strncmp(arg, buff1, arglen) == 0 || |
| 2064 |
|
strncmp(arg, buff2, arglen) == 0) |
| 2065 |
|
{ |
| 2066 |
|
if (equals != NULL && argequals != NULL) |
| 2067 |
|
{ |
| 2068 |
|
option_data = argequals; |
| 2069 |
|
if (*option_data == '=') |
| 2070 |
|
{ |
| 2071 |
|
option_data++; |
| 2072 |
|
longopwasequals = TRUE; |
| 2073 |
|
} |
| 2074 |
|
} |
| 2075 |
break; |
break; |
| 2076 |
|
} |
| 2077 |
} |
} |
| 2078 |
} |
} |
| 2079 |
|
|
| 2084 |
} |
} |
| 2085 |
} |
} |
| 2086 |
|
|
|
|
|
| 2087 |
/* Jeffrey Friedl's debugging harness uses these additional options which |
/* Jeffrey Friedl's debugging harness uses these additional options which |
| 2088 |
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 |
| 2089 |
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 |
| 2197 |
{ |
{ |
| 2198 |
*((char **)op->dataptr) = option_data; |
*((char **)op->dataptr) = option_data; |
| 2199 |
} |
} |
| 2200 |
|
|
| 2201 |
|
/* Avoid the use of strtoul() because SunOS4 doesn't have it. This is used |
| 2202 |
|
only for unpicking arguments, so just keep it simple. */ |
| 2203 |
|
|
| 2204 |
else |
else |
| 2205 |
{ |
{ |
| 2206 |
char *endptr; |
int n = 0; |
| 2207 |
int n = strtoul(option_data, &endptr, 10); |
char *endptr = option_data; |
| 2208 |
|
while (*endptr != 0 && isspace((unsigned char)(*endptr))) endptr++; |
| 2209 |
|
while (isdigit((unsigned char)(*endptr))) |
| 2210 |
|
n = n * 10 + (int)(*endptr++ - '0'); |
| 2211 |
if (*endptr != 0) |
if (*endptr != 0) |
| 2212 |
{ |
{ |
| 2213 |
if (longop) |
if (longop) |
| 2214 |
{ |
{ |
| 2215 |
char *equals = strchr(op->long_name, '='); |
char *equals = strchr(op->long_name, '='); |
| 2216 |
int nlen = (equals == NULL)? (int)strlen(op->long_name) : |
int nlen = (equals == NULL)? (int)strlen(op->long_name) : |
| 2217 |
equals - op->long_name; |
(int)(equals - op->long_name); |
| 2218 |
fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%.*s\n", |
fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%.*s\n", |
| 2219 |
option_data, nlen, op->long_name); |
option_data, nlen, op->long_name); |
| 2220 |
} |
} |
| 2235 |
if (after_context == 0) after_context = both_context; |
if (after_context == 0) after_context = both_context; |
| 2236 |
if (before_context == 0) before_context = both_context; |
if (before_context == 0) before_context = both_context; |
| 2237 |
} |
} |
| 2238 |
|
|
| 2239 |
/* Only one of --only-matching, --file-offsets, or --line-offsets is permitted. |
/* Only one of --only-matching, --file-offsets, or --line-offsets is permitted. |
| 2240 |
However, the latter two set the only_matching flag. */ |
However, the latter two set the only_matching flag. */ |
| 2241 |
|
|
| 2242 |
if ((only_matching && (file_offsets || line_offsets)) || |
if ((only_matching && (file_offsets || line_offsets)) || |
| 2243 |
(file_offsets && line_offsets)) |
(file_offsets && line_offsets)) |
| 2244 |
{ |
{ |
| 2245 |
fprintf(stderr, "pcregrep: Cannot mix --only-matching, --file-offsets " |
fprintf(stderr, "pcregrep: Cannot mix --only-matching, --file-offsets " |
| 2246 |
"and/or --line-offsets\n"); |
"and/or --line-offsets\n"); |
| 2247 |
exit(usage(2)); |
exit(usage(2)); |
| 2248 |
} |
} |
| 2249 |
|
|
| 2250 |
if (file_offsets || line_offsets) only_matching = TRUE; |
if (file_offsets || line_offsets) only_matching = TRUE; |
| 2251 |
|
|
| 2252 |
/* 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 |
| 2253 |
LC_ALL environment variable is set, and if so, use it. */ |
LC_ALL environment variable is set, and if so, use it. */ |
| 2482 |
} |
} |
| 2483 |
} |
} |
| 2484 |
|
|
| 2485 |
|
if (exclude_dir_pattern != NULL) |
| 2486 |
|
{ |
| 2487 |
|
exclude_dir_compiled = pcre_compile(exclude_dir_pattern, 0, &error, &errptr, |
| 2488 |
|
pcretables); |
| 2489 |
|
if (exclude_dir_compiled == NULL) |
| 2490 |
|
{ |
| 2491 |
|
fprintf(stderr, "pcregrep: Error in 'exclude_dir' regex at offset %d: %s\n", |
| 2492 |
|
errptr, error); |
| 2493 |
|
goto EXIT2; |
| 2494 |
|
} |
| 2495 |
|
} |
| 2496 |
|
|
| 2497 |
|
if (include_dir_pattern != NULL) |
| 2498 |
|
{ |
| 2499 |
|
include_dir_compiled = pcre_compile(include_dir_pattern, 0, &error, &errptr, |
| 2500 |
|
pcretables); |
| 2501 |
|
if (include_dir_compiled == NULL) |
| 2502 |
|
{ |
| 2503 |
|
fprintf(stderr, "pcregrep: Error in 'include_dir' regex at offset %d: %s\n", |
| 2504 |
|
errptr, error); |
| 2505 |
|
goto EXIT2; |
| 2506 |
|
} |
| 2507 |
|
} |
| 2508 |
|
|
| 2509 |
/* 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. */ |
| 2510 |
|
|
| 2511 |
if (i >= argc) |
if (i >= argc) |
| 2512 |
{ |
{ |
| 2513 |
rc = pcregrep(stdin, (filenames > FN_DEFAULT)? stdin_name : NULL); |
rc = pcregrep(stdin, FR_PLAIN, (filenames > FN_DEFAULT)? stdin_name : NULL); |
| 2514 |
goto EXIT; |
goto EXIT; |
| 2515 |
} |
} |
| 2516 |
|
|