| 4 |
|
|
| 5 |
/* This program was hacked up as a tester for PCRE. I really should have |
/* This program was hacked up as a tester for PCRE. I really should have |
| 6 |
written it more tidily in the first place. Will I ever learn? It has grown and |
written it more tidily in the first place. Will I ever learn? It has grown and |
| 7 |
been extended and consequently is now rather untidy in places. */ |
been extended and consequently is now rather untidy in places. |
| 8 |
|
|
| 9 |
|
----------------------------------------------------------------------------- |
| 10 |
|
Redistribution and use in source and binary forms, with or without |
| 11 |
|
modification, are permitted provided that the following conditions are met: |
| 12 |
|
|
| 13 |
|
* Redistributions of source code must retain the above copyright notice, |
| 14 |
|
this list of conditions and the following disclaimer. |
| 15 |
|
|
| 16 |
|
* Redistributions in binary form must reproduce the above copyright |
| 17 |
|
notice, this list of conditions and the following disclaimer in the |
| 18 |
|
documentation and/or other materials provided with the distribution. |
| 19 |
|
|
| 20 |
|
* Neither the name of the University of Cambridge nor the names of its |
| 21 |
|
contributors may be used to endorse or promote products derived from |
| 22 |
|
this software without specific prior written permission. |
| 23 |
|
|
| 24 |
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
| 25 |
|
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 26 |
|
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 27 |
|
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE |
| 28 |
|
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR |
| 29 |
|
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF |
| 30 |
|
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS |
| 31 |
|
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN |
| 32 |
|
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) |
| 33 |
|
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE |
| 34 |
|
POSSIBILITY OF SUCH DAMAGE. |
| 35 |
|
----------------------------------------------------------------------------- |
| 36 |
|
*/ |
| 37 |
|
|
| 38 |
|
|
| 39 |
#include <ctype.h> |
#include <ctype.h> |
| 40 |
#include <stdio.h> |
#include <stdio.h> |
| 42 |
#include <stdlib.h> |
#include <stdlib.h> |
| 43 |
#include <time.h> |
#include <time.h> |
| 44 |
#include <locale.h> |
#include <locale.h> |
| 45 |
|
#include <errno.h> |
| 46 |
|
|
| 47 |
/* We need the internal info for displaying the results of pcre_study(). Also |
/* We need the internal info for displaying the results of pcre_study(). Also |
| 48 |
for getting the opcodes for showing compiled code. */ |
for getting the opcodes for showing compiled code. */ |
| 66 |
#endif |
#endif |
| 67 |
#endif |
#endif |
| 68 |
|
|
| 69 |
#define LOOPREPEAT 50000 |
#define LOOPREPEAT 500000 |
| 70 |
|
|
| 71 |
#define BUFFER_SIZE 30000 |
#define BUFFER_SIZE 30000 |
| 72 |
|
#define PBUFFER_SIZE BUFFER_SIZE |
| 73 |
#define DBUFFER_SIZE BUFFER_SIZE |
#define DBUFFER_SIZE BUFFER_SIZE |
| 74 |
|
|
| 75 |
|
|
| 84 |
static int use_utf8; |
static int use_utf8; |
| 85 |
static size_t gotten_store; |
static size_t gotten_store; |
| 86 |
|
|
| 87 |
|
static uschar *pbuffer = NULL; |
| 88 |
|
|
| 89 |
|
|
| 90 |
static const int utf8_table1[] = { |
static const int utf8_table1[] = { |
| 91 |
0x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff}; |
0x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff}; |
| 105 |
/* The code for doing this is held in a separate file that is also included in |
/* The code for doing this is held in a separate file that is also included in |
| 106 |
pcre.c when it is compiled with the debug switch. It defines a function called |
pcre.c when it is compiled with the debug switch. It defines a function called |
| 107 |
print_internals(), which uses a table of opcode lengths defined by the macro |
print_internals(), which uses a table of opcode lengths defined by the macro |
| 108 |
OP_LENGTHS, whose name must be OP_lengths. */ |
OP_LENGTHS, whose name must be OP_lengths. It also uses a table that translates |
| 109 |
|
Unicode property names to numbers; this is kept in a separate file. */ |
| 110 |
|
|
| 111 |
static uschar OP_lengths[] = { OP_LENGTHS }; |
static uschar OP_lengths[] = { OP_LENGTHS }; |
| 112 |
|
|
| 113 |
|
#include "ucp.h" |
| 114 |
|
#include "ucptypetable.c" |
| 115 |
#include "printint.c" |
#include "printint.c" |
| 116 |
|
|
| 117 |
|
|
| 306 |
static int callout(pcre_callout_block *cb) |
static int callout(pcre_callout_block *cb) |
| 307 |
{ |
{ |
| 308 |
FILE *f = (first_callout | callout_extra)? outfile : NULL; |
FILE *f = (first_callout | callout_extra)? outfile : NULL; |
| 309 |
int i, pre_start, post_start; |
int i, pre_start, post_start, subject_length; |
| 310 |
|
|
| 311 |
if (callout_extra) |
if (callout_extra) |
| 312 |
{ |
{ |
| 337 |
post_start = pchars((unsigned char *)(cb->subject + cb->start_match), |
post_start = pchars((unsigned char *)(cb->subject + cb->start_match), |
| 338 |
cb->current_position - cb->start_match, f); |
cb->current_position - cb->start_match, f); |
| 339 |
|
|
| 340 |
|
subject_length = pchars((unsigned char *)cb->subject, cb->subject_length, NULL); |
| 341 |
|
|
| 342 |
(void)pchars((unsigned char *)(cb->subject + cb->current_position), |
(void)pchars((unsigned char *)(cb->subject + cb->current_position), |
| 343 |
cb->subject_length - cb->current_position, f); |
cb->subject_length - cb->current_position, f); |
| 344 |
|
|
| 345 |
if (f != NULL) fprintf(f, "\n"); |
if (f != NULL) fprintf(f, "\n"); |
| 346 |
|
|
| 347 |
/* Always print appropriate indicators, with callout number if not already |
/* Always print appropriate indicators, with callout number if not already |
| 348 |
shown */ |
shown. For automatic callouts, show the pattern offset. */ |
| 349 |
|
|
| 350 |
if (callout_extra) fprintf(outfile, " "); |
if (cb->callout_number == 255) |
| 351 |
else fprintf(outfile, "%3d ", cb->callout_number); |
{ |
| 352 |
|
fprintf(outfile, "%+3d ", cb->pattern_position); |
| 353 |
|
if (cb->pattern_position > 99) fprintf(outfile, "\n "); |
| 354 |
|
} |
| 355 |
|
else |
| 356 |
|
{ |
| 357 |
|
if (callout_extra) fprintf(outfile, " "); |
| 358 |
|
else fprintf(outfile, "%3d ", cb->callout_number); |
| 359 |
|
} |
| 360 |
|
|
| 361 |
for (i = 0; i < pre_start; i++) fprintf(outfile, " "); |
for (i = 0; i < pre_start; i++) fprintf(outfile, " "); |
| 362 |
fprintf(outfile, "^"); |
fprintf(outfile, "^"); |
| 367 |
fprintf(outfile, "^"); |
fprintf(outfile, "^"); |
| 368 |
} |
} |
| 369 |
|
|
| 370 |
|
for (i = 0; i < subject_length - pre_start - post_start + 4; i++) |
| 371 |
|
fprintf(outfile, " "); |
| 372 |
|
|
| 373 |
|
fprintf(outfile, "%.*s", (cb->next_item_length == 0)? 1 : cb->next_item_length, |
| 374 |
|
pbuffer + cb->pattern_position); |
| 375 |
|
|
| 376 |
fprintf(outfile, "\n"); |
fprintf(outfile, "\n"); |
| 377 |
first_callout = 0; |
first_callout = 0; |
| 378 |
|
|
| 449 |
|
|
| 450 |
|
|
| 451 |
/************************************************* |
/************************************************* |
| 452 |
|
* Byte flipping function * |
| 453 |
|
*************************************************/ |
| 454 |
|
|
| 455 |
|
static long int |
| 456 |
|
byteflip(long int value, int n) |
| 457 |
|
{ |
| 458 |
|
if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8); |
| 459 |
|
return ((value & 0x000000ff) << 24) | |
| 460 |
|
((value & 0x0000ff00) << 8) | |
| 461 |
|
((value & 0x00ff0000) >> 8) | |
| 462 |
|
((value & 0xff000000) >> 24); |
| 463 |
|
} |
| 464 |
|
|
| 465 |
|
|
| 466 |
|
|
| 467 |
|
|
| 468 |
|
/************************************************* |
| 469 |
* Main Program * |
* Main Program * |
| 470 |
*************************************************/ |
*************************************************/ |
| 471 |
|
|
| 499 |
|
|
| 500 |
buffer = (unsigned char *)malloc(BUFFER_SIZE); |
buffer = (unsigned char *)malloc(BUFFER_SIZE); |
| 501 |
dbuffer = (unsigned char *)malloc(DBUFFER_SIZE); |
dbuffer = (unsigned char *)malloc(DBUFFER_SIZE); |
| 502 |
|
pbuffer = (unsigned char *)malloc(PBUFFER_SIZE); |
| 503 |
|
|
| 504 |
/* Static so that new_malloc can use it. */ |
/* The outfile variable is static so that new_malloc can use it. The _setmode() |
| 505 |
|
stuff is some magic that I don't understand, but which apparently does good |
| 506 |
|
things in Windows. It's related to line terminations. */ |
| 507 |
|
|
| 508 |
|
#if defined(_WIN32) || defined(WIN32) |
| 509 |
|
_setmode( _fileno( stdout ), 0x8000 ); |
| 510 |
|
#endif /* defined(_WIN32) || defined(WIN32) */ |
| 511 |
|
|
| 512 |
outfile = stdout; |
outfile = stdout; |
| 513 |
|
|
| 539 |
printf("Compiled with\n"); |
printf("Compiled with\n"); |
| 540 |
(void)pcre_config(PCRE_CONFIG_UTF8, &rc); |
(void)pcre_config(PCRE_CONFIG_UTF8, &rc); |
| 541 |
printf(" %sUTF-8 support\n", rc? "" : "No "); |
printf(" %sUTF-8 support\n", rc? "" : "No "); |
| 542 |
|
(void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc); |
| 543 |
|
printf(" %sUnicode properties support\n", rc? "" : "No "); |
| 544 |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); |
| 545 |
printf(" Newline character is %s\n", (rc == '\r')? "CR" : "LF"); |
printf(" Newline character is %s\n", (rc == '\r')? "CR" : "LF"); |
| 546 |
(void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); |
(void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); |
| 560 |
printf(" -C show PCRE compile-time options and exit\n"); |
printf(" -C show PCRE compile-time options and exit\n"); |
| 561 |
printf(" -d debug: show compiled code; implies -i\n" |
printf(" -d debug: show compiled code; implies -i\n" |
| 562 |
" -i show information about compiled pattern\n" |
" -i show information about compiled pattern\n" |
| 563 |
|
" -m output memory used information\n" |
| 564 |
" -o <n> set size of offsets vector to <n>\n"); |
" -o <n> set size of offsets vector to <n>\n"); |
| 565 |
#if !defined NOPOSIX |
#if !defined NOPOSIX |
| 566 |
printf(" -p use POSIX interface\n"); |
printf(" -p use POSIX interface\n"); |
| 567 |
#endif |
#endif |
| 568 |
printf(" -s output store information\n" |
printf(" -s output store (memory) used information\n" |
| 569 |
" -t time compilation and execution\n"); |
" -t time compilation and execution\n"); |
| 570 |
return 1; |
return 1; |
| 571 |
} |
} |
| 588 |
|
|
| 589 |
if (argc > 1) |
if (argc > 1) |
| 590 |
{ |
{ |
| 591 |
infile = fopen(argv[op], "r"); |
infile = fopen(argv[op], "rb"); |
| 592 |
if (infile == NULL) |
if (infile == NULL) |
| 593 |
{ |
{ |
| 594 |
printf("** Failed to open %s\n", argv[op]); |
printf("** Failed to open %s\n", argv[op]); |
| 598 |
|
|
| 599 |
if (argc > 2) |
if (argc > 2) |
| 600 |
{ |
{ |
| 601 |
outfile = fopen(argv[op+1], "w"); |
outfile = fopen(argv[op+1], "wb"); |
| 602 |
if (outfile == NULL) |
if (outfile == NULL) |
| 603 |
{ |
{ |
| 604 |
printf("** Failed to open %s\n", argv[op+1]); |
printf("** Failed to open %s\n", argv[op+1]); |
| 631 |
|
|
| 632 |
const char *error; |
const char *error; |
| 633 |
unsigned char *p, *pp, *ppp; |
unsigned char *p, *pp, *ppp; |
| 634 |
|
unsigned char *to_file = NULL; |
| 635 |
const unsigned char *tables = NULL; |
const unsigned char *tables = NULL; |
| 636 |
|
unsigned long int true_size, true_study_size = 0; |
| 637 |
|
size_t size, regex_gotten_store; |
| 638 |
int do_study = 0; |
int do_study = 0; |
| 639 |
int do_debug = debug; |
int do_debug = debug; |
| 640 |
int do_G = 0; |
int do_G = 0; |
| 641 |
int do_g = 0; |
int do_g = 0; |
| 642 |
int do_showinfo = showinfo; |
int do_showinfo = showinfo; |
| 643 |
int do_showrest = 0; |
int do_showrest = 0; |
| 644 |
|
int do_flip = 0; |
| 645 |
int erroroffset, len, delimiter; |
int erroroffset, len, delimiter; |
| 646 |
|
|
| 647 |
use_utf8 = 0; |
use_utf8 = 0; |
| 655 |
while (isspace(*p)) p++; |
while (isspace(*p)) p++; |
| 656 |
if (*p == 0) continue; |
if (*p == 0) continue; |
| 657 |
|
|
| 658 |
/* Get the delimiter and seek the end of the pattern; if is isn't |
/* See if the pattern is to be loaded pre-compiled from a file. */ |
| 659 |
complete, read more. */ |
|
| 660 |
|
if (*p == '<' && strchr((char *)(p+1), '<') == NULL) |
| 661 |
|
{ |
| 662 |
|
unsigned long int magic; |
| 663 |
|
uschar sbuf[8]; |
| 664 |
|
FILE *f; |
| 665 |
|
|
| 666 |
|
p++; |
| 667 |
|
pp = p + (int)strlen((char *)p); |
| 668 |
|
while (isspace(pp[-1])) pp--; |
| 669 |
|
*pp = 0; |
| 670 |
|
|
| 671 |
|
f = fopen((char *)p, "rb"); |
| 672 |
|
if (f == NULL) |
| 673 |
|
{ |
| 674 |
|
fprintf(outfile, "Failed to open %s: %s\n", p, strerror(errno)); |
| 675 |
|
continue; |
| 676 |
|
} |
| 677 |
|
|
| 678 |
|
if (fread(sbuf, 1, 8, f) != 8) goto FAIL_READ; |
| 679 |
|
|
| 680 |
|
true_size = |
| 681 |
|
(sbuf[0] << 24) | (sbuf[1] << 16) | (sbuf[2] << 8) | sbuf[3]; |
| 682 |
|
true_study_size = |
| 683 |
|
(sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7]; |
| 684 |
|
|
| 685 |
|
re = (real_pcre *)new_malloc(true_size); |
| 686 |
|
regex_gotten_store = gotten_store; |
| 687 |
|
|
| 688 |
|
if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ; |
| 689 |
|
|
| 690 |
|
magic = ((real_pcre *)re)->magic_number; |
| 691 |
|
if (magic != MAGIC_NUMBER) |
| 692 |
|
{ |
| 693 |
|
if (byteflip(magic, sizeof(magic)) == MAGIC_NUMBER) |
| 694 |
|
{ |
| 695 |
|
do_flip = 1; |
| 696 |
|
} |
| 697 |
|
else |
| 698 |
|
{ |
| 699 |
|
fprintf(outfile, "Data in %s is not a compiled PCRE regex\n", p); |
| 700 |
|
fclose(f); |
| 701 |
|
continue; |
| 702 |
|
} |
| 703 |
|
} |
| 704 |
|
|
| 705 |
|
fprintf(outfile, "Compiled regex%s loaded from %s\n", |
| 706 |
|
do_flip? " (byte-inverted)" : "", p); |
| 707 |
|
|
| 708 |
|
/* Need to know if UTF-8 for printing data strings */ |
| 709 |
|
|
| 710 |
|
new_info(re, NULL, PCRE_INFO_OPTIONS, &options); |
| 711 |
|
use_utf8 = (options & PCRE_UTF8) != 0; |
| 712 |
|
|
| 713 |
|
/* Now see if there is any following study data */ |
| 714 |
|
|
| 715 |
|
if (true_study_size != 0) |
| 716 |
|
{ |
| 717 |
|
pcre_study_data *psd; |
| 718 |
|
|
| 719 |
|
extra = (pcre_extra *)new_malloc(sizeof(pcre_extra) + true_study_size); |
| 720 |
|
extra->flags = PCRE_EXTRA_STUDY_DATA; |
| 721 |
|
|
| 722 |
|
psd = (pcre_study_data *)(((char *)extra) + sizeof(pcre_extra)); |
| 723 |
|
extra->study_data = psd; |
| 724 |
|
|
| 725 |
|
if (fread(psd, 1, true_study_size, f) != true_study_size) |
| 726 |
|
{ |
| 727 |
|
FAIL_READ: |
| 728 |
|
fprintf(outfile, "Failed to read data from %s\n", p); |
| 729 |
|
if (extra != NULL) new_free(extra); |
| 730 |
|
if (re != NULL) new_free(re); |
| 731 |
|
fclose(f); |
| 732 |
|
continue; |
| 733 |
|
} |
| 734 |
|
fprintf(outfile, "Study data loaded from %s\n", p); |
| 735 |
|
do_study = 1; /* To get the data output if requested */ |
| 736 |
|
} |
| 737 |
|
else fprintf(outfile, "No study data\n"); |
| 738 |
|
|
| 739 |
|
fclose(f); |
| 740 |
|
goto SHOW_INFO; |
| 741 |
|
} |
| 742 |
|
|
| 743 |
|
/* In-line pattern (the usual case). Get the delimiter and seek the end of |
| 744 |
|
the pattern; if is isn't complete, read more. */ |
| 745 |
|
|
| 746 |
delimiter = *p++; |
delimiter = *p++; |
| 747 |
|
|
| 786 |
|
|
| 787 |
if (pp[1] == '\\') *pp++ = '\\'; |
if (pp[1] == '\\') *pp++ = '\\'; |
| 788 |
|
|
| 789 |
/* Terminate the pattern at the delimiter */ |
/* Terminate the pattern at the delimiter, and save a copy of the pattern |
| 790 |
|
for callouts. */ |
| 791 |
|
|
| 792 |
*pp++ = 0; |
*pp++ = 0; |
| 793 |
|
strcpy((char *)pbuffer, (char *)p); |
| 794 |
|
|
| 795 |
/* Look for options after final delimiter */ |
/* Look for options after final delimiter */ |
| 796 |
|
|
| 810 |
|
|
| 811 |
case '+': do_showrest = 1; break; |
case '+': do_showrest = 1; break; |
| 812 |
case 'A': options |= PCRE_ANCHORED; break; |
case 'A': options |= PCRE_ANCHORED; break; |
| 813 |
|
case 'C': options |= PCRE_AUTO_CALLOUT; break; |
| 814 |
case 'D': do_debug = do_showinfo = 1; break; |
case 'D': do_debug = do_showinfo = 1; break; |
| 815 |
case 'E': options |= PCRE_DOLLAR_ENDONLY; break; |
case 'E': options |= PCRE_DOLLAR_ENDONLY; break; |
| 816 |
|
case 'F': do_flip = 1; break; |
| 817 |
case 'G': do_G = 1; break; |
case 'G': do_G = 1; break; |
| 818 |
case 'I': do_showinfo = 1; break; |
case 'I': do_showinfo = 1; break; |
| 819 |
case 'M': log_store = 1; break; |
case 'M': log_store = 1; break; |
| 842 |
pp = ppp; |
pp = ppp; |
| 843 |
break; |
break; |
| 844 |
|
|
| 845 |
|
case '>': |
| 846 |
|
to_file = pp; |
| 847 |
|
while (*pp != 0) pp++; |
| 848 |
|
while (isspace(pp[-1])) pp--; |
| 849 |
|
*pp = 0; |
| 850 |
|
break; |
| 851 |
|
|
| 852 |
case '\n': case ' ': break; |
case '\n': case ' ': break; |
| 853 |
|
|
| 854 |
default: |
default: |
| 855 |
fprintf(outfile, "** Unknown option '%c'\n", pp[-1]); |
fprintf(outfile, "** Unknown option '%c'\n", pp[-1]); |
| 856 |
goto SKIP_DATA; |
goto SKIP_DATA; |
| 866 |
{ |
{ |
| 867 |
int rc; |
int rc; |
| 868 |
int cflags = 0; |
int cflags = 0; |
| 869 |
|
|
| 870 |
if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE; |
if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE; |
| 871 |
if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE; |
if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE; |
| 872 |
rc = regcomp(&preg, (char *)p, cflags); |
rc = regcomp(&preg, (char *)p, cflags); |
| 941 |
sizeof(real_pcre) - |
sizeof(real_pcre) - |
| 942 |
((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size)); |
((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size)); |
| 943 |
|
|
| 944 |
|
/* Extract the size for possible writing before possibly flipping it, |
| 945 |
|
and remember the store that was got. */ |
| 946 |
|
|
| 947 |
|
true_size = ((real_pcre *)re)->size; |
| 948 |
|
regex_gotten_store = gotten_store; |
| 949 |
|
|
| 950 |
|
/* If /S was present, study the regexp to generate additional info to |
| 951 |
|
help with the matching. */ |
| 952 |
|
|
| 953 |
|
if (do_study) |
| 954 |
|
{ |
| 955 |
|
if (timeit) |
| 956 |
|
{ |
| 957 |
|
register int i; |
| 958 |
|
clock_t time_taken; |
| 959 |
|
clock_t start_time = clock(); |
| 960 |
|
for (i = 0; i < LOOPREPEAT; i++) |
| 961 |
|
extra = pcre_study(re, study_options, &error); |
| 962 |
|
time_taken = clock() - start_time; |
| 963 |
|
if (extra != NULL) free(extra); |
| 964 |
|
fprintf(outfile, " Study time %.3f milliseconds\n", |
| 965 |
|
(((double)time_taken * 1000.0) / (double)LOOPREPEAT) / |
| 966 |
|
(double)CLOCKS_PER_SEC); |
| 967 |
|
} |
| 968 |
|
extra = pcre_study(re, study_options, &error); |
| 969 |
|
if (error != NULL) |
| 970 |
|
fprintf(outfile, "Failed to study: %s\n", error); |
| 971 |
|
else if (extra != NULL) |
| 972 |
|
true_study_size = ((pcre_study_data *)(extra->study_data))->size; |
| 973 |
|
} |
| 974 |
|
|
| 975 |
|
/* If the 'F' option was present, we flip the bytes of all the integer |
| 976 |
|
fields in the regex data block and the study block. This is to make it |
| 977 |
|
possible to test PCRE's handling of byte-flipped patterns, e.g. those |
| 978 |
|
compiled on a different architecture. */ |
| 979 |
|
|
| 980 |
|
if (do_flip) |
| 981 |
|
{ |
| 982 |
|
real_pcre *rre = (real_pcre *)re; |
| 983 |
|
rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number)); |
| 984 |
|
rre->size = byteflip(rre->size, sizeof(rre->size)); |
| 985 |
|
rre->options = byteflip(rre->options, sizeof(rre->options)); |
| 986 |
|
rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket)); |
| 987 |
|
rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref)); |
| 988 |
|
rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte)); |
| 989 |
|
rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte)); |
| 990 |
|
rre->name_table_offset = byteflip(rre->name_table_offset, |
| 991 |
|
sizeof(rre->name_table_offset)); |
| 992 |
|
rre->name_entry_size = byteflip(rre->name_entry_size, |
| 993 |
|
sizeof(rre->name_entry_size)); |
| 994 |
|
rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count)); |
| 995 |
|
|
| 996 |
|
if (extra != NULL) |
| 997 |
|
{ |
| 998 |
|
pcre_study_data *rsd = (pcre_study_data *)(extra->study_data); |
| 999 |
|
rsd->size = byteflip(rsd->size, sizeof(rsd->size)); |
| 1000 |
|
rsd->options = byteflip(rsd->options, sizeof(rsd->options)); |
| 1001 |
|
} |
| 1002 |
|
} |
| 1003 |
|
|
| 1004 |
|
/* Extract information from the compiled data if required */ |
| 1005 |
|
|
| 1006 |
|
SHOW_INFO: |
| 1007 |
|
|
| 1008 |
if (do_showinfo) |
if (do_showinfo) |
| 1009 |
{ |
{ |
| 1010 |
unsigned long int get_options; |
unsigned long int get_options, all_options; |
| 1011 |
int old_first_char, old_options, old_count; |
int old_first_char, old_options, old_count; |
| 1012 |
int count, backrefmax, first_char, need_char; |
int count, backrefmax, first_char, need_char; |
| 1013 |
int nameentrysize, namecount; |
int nameentrysize, namecount; |
| 1014 |
const uschar *nametable; |
const uschar *nametable; |
|
size_t size; |
|
| 1015 |
|
|
| 1016 |
if (do_debug) |
if (do_debug) |
| 1017 |
{ |
{ |
| 1047 |
get_options, old_options); |
get_options, old_options); |
| 1048 |
} |
} |
| 1049 |
|
|
| 1050 |
if (size != gotten_store) fprintf(outfile, |
if (size != regex_gotten_store) fprintf(outfile, |
| 1051 |
"Size disagreement: pcre_fullinfo=%d call to malloc for %d\n", |
"Size disagreement: pcre_fullinfo=%d call to malloc for %d\n", |
| 1052 |
size, gotten_store); |
size, regex_gotten_store); |
| 1053 |
|
|
| 1054 |
fprintf(outfile, "Capturing subpattern count = %d\n", count); |
fprintf(outfile, "Capturing subpattern count = %d\n", count); |
| 1055 |
if (backrefmax > 0) |
if (backrefmax > 0) |
| 1067 |
} |
} |
| 1068 |
} |
} |
| 1069 |
|
|
| 1070 |
|
/* The NOPARTIAL bit is a private bit in the options, so we have |
| 1071 |
|
to fish it out via out back door */ |
| 1072 |
|
|
| 1073 |
|
all_options = ((real_pcre *)re)->options; |
| 1074 |
|
if (do_flip) |
| 1075 |
|
{ |
| 1076 |
|
all_options = byteflip(all_options, sizeof(all_options)); |
| 1077 |
|
} |
| 1078 |
|
|
| 1079 |
|
if ((all_options & PCRE_NOPARTIAL) != 0) |
| 1080 |
|
fprintf(outfile, "Partial matching not supported\n"); |
| 1081 |
|
|
| 1082 |
if (get_options == 0) fprintf(outfile, "No options\n"); |
if (get_options == 0) fprintf(outfile, "No options\n"); |
| 1083 |
else fprintf(outfile, "Options:%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\n", |
| 1084 |
((get_options & PCRE_ANCHORED) != 0)? " anchored" : "", |
((get_options & PCRE_ANCHORED) != 0)? " anchored" : "", |
| 1128 |
else |
else |
| 1129 |
fprintf(outfile, "Need char = %d%s\n", ch, caseless); |
fprintf(outfile, "Need char = %d%s\n", ch, caseless); |
| 1130 |
} |
} |
|
} |
|
|
|
|
|
/* If /S was present, study the regexp to generate additional info to |
|
|
help with the matching. */ |
|
|
|
|
|
if (do_study) |
|
|
{ |
|
|
if (timeit) |
|
|
{ |
|
|
register int i; |
|
|
clock_t time_taken; |
|
|
clock_t start_time = clock(); |
|
|
for (i = 0; i < LOOPREPEAT; i++) |
|
|
extra = pcre_study(re, study_options, &error); |
|
|
time_taken = clock() - start_time; |
|
|
if (extra != NULL) free(extra); |
|
|
fprintf(outfile, " Study time %.3f milliseconds\n", |
|
|
(((double)time_taken * 1000.0) / (double)LOOPREPEAT) / |
|
|
(double)CLOCKS_PER_SEC); |
|
|
} |
|
|
|
|
|
extra = pcre_study(re, study_options, &error); |
|
|
if (error != NULL) |
|
|
fprintf(outfile, "Failed to study: %s\n", error); |
|
|
else if (extra == NULL) |
|
|
fprintf(outfile, "Study returned NULL\n"); |
|
| 1131 |
|
|
| 1132 |
/* Don't output study size; at present it is in any case a fixed |
/* Don't output study size; at present it is in any case a fixed |
| 1133 |
value, but it varies, depending on the computer architecture, and |
value, but it varies, depending on the computer architecture, and |
| 1134 |
so messes up the test suite. */ |
so messes up the test suite. (And with the /F option, it might be |
| 1135 |
|
flipped.) */ |
| 1136 |
|
|
| 1137 |
else if (do_showinfo) |
if (do_study) |
| 1138 |
{ |
{ |
| 1139 |
size_t size; |
if (extra == NULL) |
| 1140 |
uschar *start_bits = NULL; |
fprintf(outfile, "Study returned NULL\n"); |
|
new_info(re, extra, PCRE_INFO_STUDYSIZE, &size); |
|
|
new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits); |
|
|
/* fprintf(outfile, "Study size = %d\n", size); */ |
|
|
if (start_bits == NULL) |
|
|
fprintf(outfile, "No starting character set\n"); |
|
| 1141 |
else |
else |
| 1142 |
{ |
{ |
| 1143 |
int i; |
uschar *start_bits = NULL; |
| 1144 |
int c = 24; |
new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits); |
| 1145 |
fprintf(outfile, "Starting character set: "); |
|
| 1146 |
for (i = 0; i < 256; i++) |
if (start_bits == NULL) |
| 1147 |
|
fprintf(outfile, "No starting byte set\n"); |
| 1148 |
|
else |
| 1149 |
{ |
{ |
| 1150 |
if ((start_bits[i/8] & (1<<(i%8))) != 0) |
int i; |
| 1151 |
|
int c = 24; |
| 1152 |
|
fprintf(outfile, "Starting byte set: "); |
| 1153 |
|
for (i = 0; i < 256; i++) |
| 1154 |
{ |
{ |
| 1155 |
if (c > 75) |
if ((start_bits[i/8] & (1<<(i&7))) != 0) |
|
{ |
|
|
fprintf(outfile, "\n "); |
|
|
c = 2; |
|
|
} |
|
|
if (isprint(i) && i != ' ') |
|
| 1156 |
{ |
{ |
| 1157 |
fprintf(outfile, "%c ", i); |
if (c > 75) |
| 1158 |
c += 2; |
{ |
| 1159 |
} |
fprintf(outfile, "\n "); |
| 1160 |
else |
c = 2; |
| 1161 |
{ |
} |
| 1162 |
fprintf(outfile, "\\x%02x ", i); |
if (isprint(i) && i != ' ') |
| 1163 |
c += 5; |
{ |
| 1164 |
|
fprintf(outfile, "%c ", i); |
| 1165 |
|
c += 2; |
| 1166 |
|
} |
| 1167 |
|
else |
| 1168 |
|
{ |
| 1169 |
|
fprintf(outfile, "\\x%02x ", i); |
| 1170 |
|
c += 5; |
| 1171 |
|
} |
| 1172 |
} |
} |
| 1173 |
} |
} |
| 1174 |
|
fprintf(outfile, "\n"); |
| 1175 |
} |
} |
|
fprintf(outfile, "\n"); |
|
| 1176 |
} |
} |
| 1177 |
} |
} |
| 1178 |
} |
} |
| 1179 |
} |
|
| 1180 |
|
/* If the '>' option was present, we write out the regex to a file, and |
| 1181 |
|
that is all. The first 8 bytes of the file are the regex length and then |
| 1182 |
|
the study length, in big-endian order. */ |
| 1183 |
|
|
| 1184 |
|
if (to_file != NULL) |
| 1185 |
|
{ |
| 1186 |
|
FILE *f = fopen((char *)to_file, "wb"); |
| 1187 |
|
if (f == NULL) |
| 1188 |
|
{ |
| 1189 |
|
fprintf(outfile, "Unable to open %s: %s\n", to_file, strerror(errno)); |
| 1190 |
|
} |
| 1191 |
|
else |
| 1192 |
|
{ |
| 1193 |
|
uschar sbuf[8]; |
| 1194 |
|
sbuf[0] = (true_size >> 24) & 255; |
| 1195 |
|
sbuf[1] = (true_size >> 16) & 255; |
| 1196 |
|
sbuf[2] = (true_size >> 8) & 255; |
| 1197 |
|
sbuf[3] = (true_size) & 255; |
| 1198 |
|
|
| 1199 |
|
sbuf[4] = (true_study_size >> 24) & 255; |
| 1200 |
|
sbuf[5] = (true_study_size >> 16) & 255; |
| 1201 |
|
sbuf[6] = (true_study_size >> 8) & 255; |
| 1202 |
|
sbuf[7] = (true_study_size) & 255; |
| 1203 |
|
|
| 1204 |
|
if (fwrite(sbuf, 1, 8, f) < 8 || |
| 1205 |
|
fwrite(re, 1, true_size, f) < true_size) |
| 1206 |
|
{ |
| 1207 |
|
fprintf(outfile, "Write error on %s: %s\n", to_file, strerror(errno)); |
| 1208 |
|
} |
| 1209 |
|
else |
| 1210 |
|
{ |
| 1211 |
|
fprintf(outfile, "Compiled regex written to %s\n", to_file); |
| 1212 |
|
if (extra != NULL) |
| 1213 |
|
{ |
| 1214 |
|
if (fwrite(extra->study_data, 1, true_study_size, f) < |
| 1215 |
|
true_study_size) |
| 1216 |
|
{ |
| 1217 |
|
fprintf(outfile, "Write error on %s: %s\n", to_file, |
| 1218 |
|
strerror(errno)); |
| 1219 |
|
} |
| 1220 |
|
else fprintf(outfile, "Study data written to %s\n", to_file); |
| 1221 |
|
} |
| 1222 |
|
} |
| 1223 |
|
fclose(f); |
| 1224 |
|
} |
| 1225 |
|
continue; /* With next regex */ |
| 1226 |
|
} |
| 1227 |
|
} /* End of non-POSIX compile */ |
| 1228 |
|
|
| 1229 |
/* Read data lines and test them */ |
/* Read data lines and test them */ |
| 1230 |
|
|
| 1328 |
} |
} |
| 1329 |
break; |
break; |
| 1330 |
|
|
| 1331 |
case 0: /* Allows for an empty line */ |
case 0: /* \ followed by EOF allows for an empty line */ |
| 1332 |
p--; |
p--; |
| 1333 |
continue; |
continue; |
| 1334 |
|
|
| 1335 |
|
case '>': |
| 1336 |
|
while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0'; |
| 1337 |
|
continue; |
| 1338 |
|
|
| 1339 |
case 'A': /* Option setting */ |
case 'A': /* Option setting */ |
| 1340 |
options |= PCRE_ANCHORED; |
options |= PCRE_ANCHORED; |
| 1341 |
continue; |
continue; |
| 1446 |
if (n == 0) use_offsets = NULL; /* Ensures it can't write to it */ |
if (n == 0) use_offsets = NULL; /* Ensures it can't write to it */ |
| 1447 |
continue; |
continue; |
| 1448 |
|
|
| 1449 |
|
case 'P': |
| 1450 |
|
options |= PCRE_PARTIAL; |
| 1451 |
|
continue; |
| 1452 |
|
|
| 1453 |
case 'S': |
case 'S': |
| 1454 |
show_malloc = 1; |
show_malloc = 1; |
| 1455 |
continue; |
continue; |
| 1560 |
min = mid; |
min = mid; |
| 1561 |
mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2; |
mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2; |
| 1562 |
} |
} |
| 1563 |
else if (count >= 0 || count == PCRE_ERROR_NOMATCH) |
else if (count >= 0 || count == PCRE_ERROR_NOMATCH || |
| 1564 |
|
count == PCRE_ERROR_PARTIAL) |
| 1565 |
{ |
{ |
| 1566 |
if (mid == min + 1) |
if (mid == min + 1) |
| 1567 |
{ |
{ |
| 1597 |
/* The normal case is just to do the match once, with the default |
/* The normal case is just to do the match once, with the default |
| 1598 |
value of match_limit. */ |
value of match_limit. */ |
| 1599 |
|
|
| 1600 |
else count = pcre_exec(re, extra, (char *)bptr, len, |
else |
| 1601 |
start_offset, options | g_notempty, use_offsets, use_size_offsets); |
{ |
| 1602 |
|
count = pcre_exec(re, extra, (char *)bptr, len, |
| 1603 |
|
start_offset, options | g_notempty, use_offsets, use_size_offsets); |
| 1604 |
|
} |
| 1605 |
|
|
| 1606 |
if (count == 0) |
if (count == 0) |
| 1607 |
{ |
{ |
| 1688 |
} |
} |
| 1689 |
} |
} |
| 1690 |
|
|
| 1691 |
|
/* There was a partial match */ |
| 1692 |
|
|
| 1693 |
|
else if (count == PCRE_ERROR_PARTIAL) |
| 1694 |
|
{ |
| 1695 |
|
fprintf(outfile, "Partial match\n"); |
| 1696 |
|
break; /* Out of the /g loop */ |
| 1697 |
|
} |
| 1698 |
|
|
| 1699 |
/* 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 |
| 1700 |
g_notempty after a null match, this is not necessarily the end. |
g_notempty after a null match, this is not necessarily the end. |
| 1701 |
We want to advance the start offset, and continue. In the case of UTF-8 |
We want to advance the start offset, and continue. In the case of UTF-8 |