| 36 |
*/ |
*/ |
| 37 |
|
|
| 38 |
|
|
| 39 |
|
#ifdef HAVE_CONFIG_H |
| 40 |
|
#include <config.h> |
| 41 |
|
#endif |
| 42 |
|
|
| 43 |
#include <ctype.h> |
#include <ctype.h> |
| 44 |
#include <stdio.h> |
#include <stdio.h> |
| 45 |
#include <string.h> |
#include <string.h> |
| 71 |
#endif |
#endif |
| 72 |
|
|
| 73 |
|
|
| 74 |
#define PCRE_SPY /* For Win32 build, import data, not export */ |
/* We have to include pcre_internal.h because we need the internal info for |
| 75 |
|
displaying the results of pcre_study() and we also need to know about the |
| 76 |
/* We include pcre_internal.h because we need the internal info for displaying |
internal macros, structures, and other internal data values; pcretest has |
| 77 |
the results of pcre_study() and we also need to know about the internal |
"inside information" compared to a program that strictly follows the PCRE API. |
| 78 |
macros, structures, and other internal data values; pcretest has "inside |
|
| 79 |
information" compared to a program that strictly follows the PCRE API. */ |
Although pcre_internal.h does itself include pcre.h, we explicitly include it |
| 80 |
|
here before pcre_internal.h so that the PCRE_EXP_xxx macros get set |
| 81 |
|
appropriately for an application, not for building PCRE. */ |
| 82 |
|
|
| 83 |
|
#include "pcre.h" |
| 84 |
#include "pcre_internal.h" |
#include "pcre_internal.h" |
| 85 |
|
|
| 86 |
/* We need access to the data tables that PCRE uses. So as not to have to keep |
/* We need access to the data tables that PCRE uses. So as not to have to keep |
| 156 |
static int callout_extra; |
static int callout_extra; |
| 157 |
static int callout_fail_count; |
static int callout_fail_count; |
| 158 |
static int callout_fail_id; |
static int callout_fail_id; |
| 159 |
|
static int debug_lengths; |
| 160 |
static int first_callout; |
static int first_callout; |
| 161 |
static int locale_set = 0; |
static int locale_set = 0; |
| 162 |
static int show_malloc; |
static int show_malloc; |
| 664 |
|
|
| 665 |
|
|
| 666 |
/************************************************* |
/************************************************* |
| 667 |
|
* Case-independent strncmp() function * |
| 668 |
|
*************************************************/ |
| 669 |
|
|
| 670 |
|
/* |
| 671 |
|
Arguments: |
| 672 |
|
s first string |
| 673 |
|
t second string |
| 674 |
|
n number of characters to compare |
| 675 |
|
|
| 676 |
|
Returns: < 0, = 0, or > 0, according to the comparison |
| 677 |
|
*/ |
| 678 |
|
|
| 679 |
|
static int |
| 680 |
|
strncmpic(uschar *s, uschar *t, int n) |
| 681 |
|
{ |
| 682 |
|
while (n--) |
| 683 |
|
{ |
| 684 |
|
int c = tolower(*s++) - tolower(*t++); |
| 685 |
|
if (c) return c; |
| 686 |
|
} |
| 687 |
|
return 0; |
| 688 |
|
} |
| 689 |
|
|
| 690 |
|
|
| 691 |
|
|
| 692 |
|
/************************************************* |
| 693 |
* Check newline indicator * |
* Check newline indicator * |
| 694 |
*************************************************/ |
*************************************************/ |
| 695 |
|
|
| 696 |
/* This is used both at compile and run-time to check for <xxx> escapes, where |
/* This is used both at compile and run-time to check for <xxx> escapes, where |
| 697 |
xxx is LF, CR, CRLF, or ANY. Print a message and return 0 if there is no match. |
xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is |
| 698 |
|
no match. |
| 699 |
|
|
| 700 |
Arguments: |
Arguments: |
| 701 |
p points after the leading '<' |
p points after the leading '<' |
| 707 |
static int |
static int |
| 708 |
check_newline(uschar *p, FILE *f) |
check_newline(uschar *p, FILE *f) |
| 709 |
{ |
{ |
| 710 |
if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR; |
if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR; |
| 711 |
if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF; |
if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF; |
| 712 |
if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; |
if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; |
| 713 |
if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY; |
if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF; |
| 714 |
|
if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY; |
| 715 |
|
if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF; |
| 716 |
|
if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE; |
| 717 |
fprintf(f, "Unknown newline type at: <%s\n", p); |
fprintf(f, "Unknown newline type at: <%s\n", p); |
| 718 |
return 0; |
return 0; |
| 719 |
} |
} |
| 885 |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); |
| 886 |
printf(" Newline sequence is %s\n", (rc == '\r')? "CR" : |
printf(" Newline sequence is %s\n", (rc == '\r')? "CR" : |
| 887 |
(rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" : |
(rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" : |
| 888 |
|
(rc == -2)? "ANYCRLF" : |
| 889 |
(rc == -1)? "ANY" : "???"); |
(rc == -1)? "ANY" : "???"); |
| 890 |
|
(void)pcre_config(PCRE_CONFIG_BSR, &rc); |
| 891 |
|
printf(" \\R matches %s\n", rc? "CR, LF, or CRLF only" : |
| 892 |
|
"all Unicode newlines"); |
| 893 |
(void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); |
(void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); |
| 894 |
printf(" Internal link size = %d\n", rc); |
printf(" Internal link size = %d\n", rc); |
| 895 |
(void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc); |
(void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc); |
| 926 |
if (offsets == NULL) |
if (offsets == NULL) |
| 927 |
{ |
{ |
| 928 |
printf("** Failed to get %d bytes of memory for offsets vector\n", |
printf("** Failed to get %d bytes of memory for offsets vector\n", |
| 929 |
size_offsets_max * sizeof(int)); |
(int)(size_offsets_max * sizeof(int))); |
| 930 |
yield = 1; |
yield = 1; |
| 931 |
goto EXIT; |
goto EXIT; |
| 932 |
} |
} |
| 986 |
size_t size, regex_gotten_store; |
size_t size, regex_gotten_store; |
| 987 |
int do_study = 0; |
int do_study = 0; |
| 988 |
int do_debug = debug; |
int do_debug = debug; |
|
int debug_lengths = 1; |
|
| 989 |
int do_G = 0; |
int do_G = 0; |
| 990 |
int do_g = 0; |
int do_g = 0; |
| 991 |
int do_showinfo = showinfo; |
int do_showinfo = showinfo; |
| 994 |
int erroroffset, len, delimiter, poffset; |
int erroroffset, len, delimiter, poffset; |
| 995 |
|
|
| 996 |
use_utf8 = 0; |
use_utf8 = 0; |
| 997 |
|
debug_lengths = 1; |
| 998 |
|
|
| 999 |
if (infile == stdin) printf(" re> "); |
if (infile == stdin) printf(" re> "); |
| 1000 |
if (extend_inputline(infile, buffer) == NULL) break; |
if (extend_inputline(infile, buffer) == NULL) break; |
| 1354 |
rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number)); |
rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number)); |
| 1355 |
rre->size = byteflip(rre->size, sizeof(rre->size)); |
rre->size = byteflip(rre->size, sizeof(rre->size)); |
| 1356 |
rre->options = byteflip(rre->options, sizeof(rre->options)); |
rre->options = byteflip(rre->options, sizeof(rre->options)); |
| 1357 |
|
rre->flags = byteflip(rre->flags, sizeof(rre->flags)); |
| 1358 |
rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket)); |
rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket)); |
| 1359 |
rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref)); |
rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref)); |
| 1360 |
rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte)); |
rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte)); |
| 1389 |
#if !defined NOINFOCHECK |
#if !defined NOINFOCHECK |
| 1390 |
int old_first_char, old_options, old_count; |
int old_first_char, old_options, old_count; |
| 1391 |
#endif |
#endif |
| 1392 |
int count, backrefmax, first_char, need_char; |
int count, backrefmax, first_char, need_char, okpartial, jchanged, |
| 1393 |
|
hascrorlf; |
| 1394 |
int nameentrysize, namecount; |
int nameentrysize, namecount; |
| 1395 |
const uschar *nametable; |
const uschar *nametable; |
| 1396 |
|
|
| 1403 |
new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize); |
new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize); |
| 1404 |
new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount); |
new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount); |
| 1405 |
new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable); |
new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable); |
| 1406 |
|
new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial); |
| 1407 |
|
new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged); |
| 1408 |
|
new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf); |
| 1409 |
|
|
| 1410 |
#if !defined NOINFOCHECK |
#if !defined NOINFOCHECK |
| 1411 |
old_count = pcre_info(re, &old_options, &old_first_char); |
old_count = pcre_info(re, &old_options, &old_first_char); |
| 1447 |
} |
} |
| 1448 |
} |
} |
| 1449 |
|
|
| 1450 |
/* The NOPARTIAL bit is a private bit in the options, so we have |
if (!okpartial) fprintf(outfile, "Partial matching not supported\n"); |
| 1451 |
to fish it out via out back door */ |
if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n"); |
| 1452 |
|
|
| 1453 |
all_options = ((real_pcre *)re)->options; |
all_options = ((real_pcre *)re)->options; |
| 1454 |
if (do_flip) |
if (do_flip) all_options = byteflip(all_options, sizeof(all_options)); |
|
{ |
|
|
all_options = byteflip(all_options, sizeof(all_options)); |
|
|
} |
|
|
|
|
|
if ((all_options & PCRE_NOPARTIAL) != 0) |
|
|
fprintf(outfile, "Partial matching not supported\n"); |
|
| 1455 |
|
|
| 1456 |
if (get_options == 0) fprintf(outfile, "No options\n"); |
if (get_options == 0) fprintf(outfile, "No options\n"); |
| 1457 |
else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s\n", |
else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", |
| 1458 |
((get_options & PCRE_ANCHORED) != 0)? " anchored" : "", |
((get_options & PCRE_ANCHORED) != 0)? " anchored" : "", |
| 1459 |
((get_options & PCRE_CASELESS) != 0)? " caseless" : "", |
((get_options & PCRE_CASELESS) != 0)? " caseless" : "", |
| 1460 |
((get_options & PCRE_EXTENDED) != 0)? " extended" : "", |
((get_options & PCRE_EXTENDED) != 0)? " extended" : "", |
| 1461 |
((get_options & PCRE_MULTILINE) != 0)? " multiline" : "", |
((get_options & PCRE_MULTILINE) != 0)? " multiline" : "", |
| 1462 |
((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "", |
((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "", |
| 1463 |
((get_options & PCRE_DOTALL) != 0)? " dotall" : "", |
((get_options & PCRE_DOTALL) != 0)? " dotall" : "", |
| 1464 |
|
((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "", |
| 1465 |
|
((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "", |
| 1466 |
((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "", |
((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "", |
| 1467 |
((get_options & PCRE_EXTRA) != 0)? " extra" : "", |
((get_options & PCRE_EXTRA) != 0)? " extra" : "", |
| 1468 |
((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "", |
((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "", |
| 1471 |
((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "", |
((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "", |
| 1472 |
((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : ""); |
((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : ""); |
| 1473 |
|
|
| 1474 |
|
if (jchanged) fprintf(outfile, "Duplicate name status changes\n"); |
| 1475 |
|
|
| 1476 |
switch (get_options & PCRE_NEWLINE_BITS) |
switch (get_options & PCRE_NEWLINE_BITS) |
| 1477 |
{ |
{ |
| 1478 |
case PCRE_NEWLINE_CR: |
case PCRE_NEWLINE_CR: |
| 1487 |
fprintf(outfile, "Forced newline sequence: CRLF\n"); |
fprintf(outfile, "Forced newline sequence: CRLF\n"); |
| 1488 |
break; |
break; |
| 1489 |
|
|
| 1490 |
|
case PCRE_NEWLINE_ANYCRLF: |
| 1491 |
|
fprintf(outfile, "Forced newline sequence: ANYCRLF\n"); |
| 1492 |
|
break; |
| 1493 |
|
|
| 1494 |
case PCRE_NEWLINE_ANY: |
case PCRE_NEWLINE_ANY: |
| 1495 |
fprintf(outfile, "Forced newline sequence: ANY\n"); |
fprintf(outfile, "Forced newline sequence: ANY\n"); |
| 1496 |
break; |
break; |
| 1640 |
for (;;) |
for (;;) |
| 1641 |
{ |
{ |
| 1642 |
uschar *q; |
uschar *q; |
| 1643 |
uschar *bptr = dbuffer; |
uschar *bptr; |
| 1644 |
int *use_offsets = offsets; |
int *use_offsets = offsets; |
| 1645 |
int use_size_offsets = size_offsets; |
int use_size_offsets = size_offsets; |
| 1646 |
int callout_data = 0; |
int callout_data = 0; |
| 1696 |
p = buffer; |
p = buffer; |
| 1697 |
while (isspace(*p)) p++; |
while (isspace(*p)) p++; |
| 1698 |
|
|
| 1699 |
q = dbuffer; |
bptr = q = dbuffer; |
| 1700 |
while ((c = *p++) != 0) |
while ((c = *p++) != 0) |
| 1701 |
{ |
{ |
| 1702 |
int i = 0; |
int i = 0; |
| 1891 |
if (offsets == NULL) |
if (offsets == NULL) |
| 1892 |
{ |
{ |
| 1893 |
printf("** Failed to get %d bytes of memory for offsets vector\n", |
printf("** Failed to get %d bytes of memory for offsets vector\n", |
| 1894 |
size_offsets_max * sizeof(int)); |
(int)(size_offsets_max * sizeof(int))); |
| 1895 |
yield = 1; |
yield = 1; |
| 1896 |
goto EXIT; |
goto EXIT; |
| 1897 |
} |
} |
| 2021 |
|
|
| 2022 |
for (;; gmatched++) /* Loop for /g or /G */ |
for (;; gmatched++) /* Loop for /g or /G */ |
| 2023 |
{ |
{ |
|
int gany_fudge; |
|
| 2024 |
if (timeitm > 0) |
if (timeitm > 0) |
| 2025 |
{ |
{ |
| 2026 |
register int i; |
register int i; |
| 2260 |
} |
} |
| 2261 |
|
|
| 2262 |
/* Failed to match. If this is a /g or /G loop and we previously set |
/* Failed to match. If this is a /g or /G loop and we previously set |
| 2263 |
g_notempty after a null match, this is not necessarily the end. |
g_notempty after a null match, this is not necessarily the end. We want |
| 2264 |
We want to advance the start offset, and continue. In the case of UTF-8 |
to advance the start offset, and continue. We won't be at the end of the |
| 2265 |
matching, the advance must be one character, not one byte. Fudge the |
string - that was checked before setting g_notempty. |
| 2266 |
offset values to achieve this. We won't be at the end of the string - |
|
| 2267 |
that was checked before setting g_notempty. */ |
Complication arises in the case when the newline option is "any" or |
| 2268 |
|
"anycrlf". If the previous match was at the end of a line terminated by |
| 2269 |
|
CRLF, an advance of one character just passes the \r, whereas we should |
| 2270 |
|
prefer the longer newline sequence, as does the code in pcre_exec(). |
| 2271 |
|
Fudge the offset value to achieve this. |
| 2272 |
|
|
| 2273 |
|
Otherwise, in the case of UTF-8 matching, the advance must be one |
| 2274 |
|
character, not one byte. */ |
| 2275 |
|
|
| 2276 |
else |
else |
| 2277 |
{ |
{ |
| 2278 |
if (g_notempty != 0) |
if (g_notempty != 0) |
| 2279 |
{ |
{ |
| 2280 |
int onechar = 1; |
int onechar = 1; |
| 2281 |
|
unsigned int obits = ((real_pcre *)re)->options; |
| 2282 |
use_offsets[0] = start_offset; |
use_offsets[0] = start_offset; |
| 2283 |
if (use_utf8) |
if ((obits & PCRE_NEWLINE_BITS) == 0) |
| 2284 |
|
{ |
| 2285 |
|
int d; |
| 2286 |
|
(void)pcre_config(PCRE_CONFIG_NEWLINE, &d); |
| 2287 |
|
obits = (d == '\r')? PCRE_NEWLINE_CR : |
| 2288 |
|
(d == '\n')? PCRE_NEWLINE_LF : |
| 2289 |
|
(d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF : |
| 2290 |
|
(d == -2)? PCRE_NEWLINE_ANYCRLF : |
| 2291 |
|
(d == -1)? PCRE_NEWLINE_ANY : 0; |
| 2292 |
|
} |
| 2293 |
|
if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY || |
| 2294 |
|
(obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF) |
| 2295 |
|
&& |
| 2296 |
|
start_offset < len - 1 && |
| 2297 |
|
bptr[start_offset] == '\r' && |
| 2298 |
|
bptr[start_offset+1] == '\n') |
| 2299 |
|
onechar++; |
| 2300 |
|
else if (use_utf8) |
| 2301 |
{ |
{ |
| 2302 |
while (start_offset + onechar < len) |
while (start_offset + onechar < len) |
| 2303 |
{ |
{ |
| 2329 |
what Perl's /g options does. This turns out to be rather cunning. First |
what Perl's /g options does. This turns out to be rather cunning. First |
| 2330 |
we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the |
we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the |
| 2331 |
same point. If this fails (picked up above) we advance to the next |
same point. If this fails (picked up above) we advance to the next |
| 2332 |
character. |
character. */ |
|
|
|
|
Yet more complication arises in the case when the newline option is |
|
|
"any" and a pattern in multiline mode has to match at the start of a |
|
|
line. If a previous match was at the end of a line, and advance of one |
|
|
character just passes the \r, whereas we should prefer the longer newline |
|
|
sequence, as does the code in pcre_exec(). So we fudge it. */ |
|
| 2333 |
|
|
| 2334 |
g_notempty = 0; |
g_notempty = 0; |
|
gany_fudge = 0; |
|
| 2335 |
|
|
| 2336 |
if (use_offsets[0] == use_offsets[1]) |
if (use_offsets[0] == use_offsets[1]) |
| 2337 |
{ |
{ |
| 2338 |
if (use_offsets[0] == len) break; |
if (use_offsets[0] == len) break; |
| 2339 |
g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED; |
g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED; |
|
if ((((real_pcre *)re)->options & PCRE_STARTLINE) != 0 && |
|
|
(((real_pcre *)re)->options & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY && |
|
|
use_offsets[0] < len - 1 && |
|
|
bptr[use_offsets[0]] == '\r' && |
|
|
bptr[use_offsets[0]+1] == '\n') |
|
|
gany_fudge = 1; |
|
| 2340 |
} |
} |
| 2341 |
|
|
| 2342 |
/* For /g, update the start offset, leaving the rest alone */ |
/* For /g, update the start offset, leaving the rest alone */ |
| 2343 |
|
|
| 2344 |
if (do_g) start_offset = use_offsets[1] + gany_fudge; |
if (do_g) start_offset = use_offsets[1]; |
| 2345 |
|
|
| 2346 |
/* For /G, update the pointer and length */ |
/* For /G, update the pointer and length */ |
| 2347 |
|
|
| 2348 |
else |
else |
| 2349 |
{ |
{ |
| 2350 |
bptr += use_offsets[1] + gany_fudge; |
bptr += use_offsets[1]; |
| 2351 |
len -= use_offsets[1] + gany_fudge; |
len -= use_offsets[1]; |
| 2352 |
} |
} |
| 2353 |
} /* End of loop for /g and /G */ |
} /* End of loop for /g and /G */ |
| 2354 |
|
|