12 |
/* Use the internal info for displaying the results of pcre_study(). */ |
/* Use the internal info for displaying the results of pcre_study(). */ |
13 |
|
|
14 |
#include "internal.h" |
#include "internal.h" |
15 |
|
|
16 |
|
/* It is possible to compile this test program without including support for |
17 |
|
testing the POSIX interface, though this is not available via the standard |
18 |
|
Makefile. */ |
19 |
|
|
20 |
|
#if !defined NOPOSIX |
21 |
#include "pcreposix.h" |
#include "pcreposix.h" |
22 |
|
#endif |
23 |
|
|
24 |
#ifndef CLOCKS_PER_SEC |
#ifndef CLOCKS_PER_SEC |
25 |
#ifdef CLK_TCK |
#ifdef CLK_TCK |
29 |
#endif |
#endif |
30 |
#endif |
#endif |
31 |
|
|
32 |
#define LOOPREPEAT 10000 |
#define LOOPREPEAT 20000 |
33 |
|
|
34 |
|
|
35 |
static FILE *outfile; |
static FILE *outfile; |
36 |
static int log_store = 0; |
static int log_store = 0; |
37 |
|
static size_t gotten_store; |
38 |
|
|
39 |
|
|
40 |
|
|
41 |
|
static int utf8_table1[] = { |
42 |
|
0x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff}; |
43 |
|
|
44 |
|
static int utf8_table2[] = { |
45 |
|
0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc}; |
46 |
|
|
47 |
|
static int utf8_table3[] = { |
48 |
|
0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01}; |
49 |
|
|
50 |
|
|
51 |
|
/************************************************* |
52 |
|
* Convert character value to UTF-8 * |
53 |
|
*************************************************/ |
54 |
|
|
55 |
|
/* This function takes an integer value in the range 0 - 0x7fffffff |
56 |
|
and encodes it as a UTF-8 character in 0 to 6 bytes. |
57 |
|
|
58 |
|
Arguments: |
59 |
|
cvalue the character value |
60 |
|
buffer pointer to buffer for result - at least 6 bytes long |
61 |
|
|
62 |
|
Returns: number of characters placed in the buffer |
63 |
|
-1 if input character is negative |
64 |
|
0 if input character is positive but too big (only when |
65 |
|
int is longer than 32 bits) |
66 |
|
*/ |
67 |
|
|
68 |
|
static int |
69 |
|
ord2utf8(int cvalue, unsigned char *buffer) |
70 |
|
{ |
71 |
|
register int i, j; |
72 |
|
for (i = 0; i < sizeof(utf8_table1)/sizeof(int); i++) |
73 |
|
if (cvalue <= utf8_table1[i]) break; |
74 |
|
if (i >= sizeof(utf8_table1)/sizeof(int)) return 0; |
75 |
|
if (cvalue < 0) return -1; |
76 |
|
*buffer++ = utf8_table2[i] | (cvalue & utf8_table3[i]); |
77 |
|
cvalue >>= 6 - i; |
78 |
|
for (j = 0; j < i; j++) |
79 |
|
{ |
80 |
|
*buffer++ = 0x80 | (cvalue & 0x3f); |
81 |
|
cvalue >>= 6; |
82 |
|
} |
83 |
|
return i + 1; |
84 |
|
} |
85 |
|
|
86 |
|
|
87 |
|
/************************************************* |
88 |
|
* Convert UTF-8 string to value * |
89 |
|
*************************************************/ |
90 |
|
|
91 |
|
/* This function takes one or more bytes that represents a UTF-8 character, |
92 |
|
and returns the value of the character. |
93 |
|
|
94 |
|
Argument: |
95 |
|
buffer a pointer to the byte vector |
96 |
|
vptr a pointer to an int to receive the value |
97 |
|
|
98 |
|
Returns: > 0 => the number of bytes consumed |
99 |
|
-6 to 0 => malformed UTF-8 character at offset = (-return) |
100 |
|
*/ |
101 |
|
|
102 |
|
int |
103 |
|
utf82ord(unsigned char *buffer, int *vptr) |
104 |
|
{ |
105 |
|
int c = *buffer++; |
106 |
|
int d = c; |
107 |
|
int i, j, s; |
108 |
|
|
109 |
|
for (i = -1; i < 6; i++) /* i is number of additional bytes */ |
110 |
|
{ |
111 |
|
if ((d & 0x80) == 0) break; |
112 |
|
d <<= 1; |
113 |
|
} |
114 |
|
|
115 |
|
if (i == -1) { *vptr = c; return 1; } /* ascii character */ |
116 |
|
if (i == 0 || i == 6) return 0; /* invalid UTF-8 */ |
117 |
|
|
118 |
|
/* i now has a value in the range 1-5 */ |
119 |
|
|
120 |
|
d = c & utf8_table3[i]; |
121 |
|
s = 6 - i; |
122 |
|
|
123 |
|
for (j = 0; j < i; j++) |
124 |
|
{ |
125 |
|
c = *buffer++; |
126 |
|
if ((c & 0xc0) != 0x80) return -(j+1); |
127 |
|
d |= (c & 0x3f) << s; |
128 |
|
s += 6; |
129 |
|
} |
130 |
|
|
131 |
|
/* Check that encoding was the correct unique one */ |
132 |
|
|
133 |
|
for (j = 0; j < sizeof(utf8_table1)/sizeof(int); j++) |
134 |
|
if (d <= utf8_table1[j]) break; |
135 |
|
if (j != i) return -(i+1); |
136 |
|
|
137 |
|
/* Valid value */ |
138 |
|
|
139 |
|
*vptr = d; |
140 |
|
return i+1; |
141 |
|
} |
142 |
|
|
143 |
|
|
144 |
|
|
145 |
|
|
146 |
|
|
147 |
|
|
156 |
"*", "*?", "+", "+?", "?", "??", "{", "{", "{", |
"*", "*?", "+", "+?", "?", "??", "{", "{", "{", |
157 |
"*", "*?", "+", "+?", "?", "??", "{", "{", "{", |
"*", "*?", "+", "+?", "?", "??", "{", "{", "{", |
158 |
"*", "*?", "+", "+?", "?", "??", "{", "{", |
"*", "*?", "+", "+?", "?", "??", "{", "{", |
159 |
"class", "Ref", |
"class", "Ref", "Recurse", |
160 |
"Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", |
"Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", |
161 |
"AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cref", |
"AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cref", |
162 |
"Brazero", "Braminzero", "Bra" |
"Brazero", "Braminzero", "Bra" |
163 |
}; |
}; |
164 |
|
|
165 |
|
|
166 |
static void print_internals(pcre *re, FILE *outfile) |
static void print_internals(pcre *re) |
167 |
{ |
{ |
168 |
unsigned char *code = ((real_pcre *)re)->code; |
unsigned char *code = ((real_pcre *)re)->code; |
169 |
|
|
372 |
|
|
373 |
|
|
374 |
|
|
375 |
/* Character string printing function. */ |
/* Character string printing function. A "normal" and a UTF-8 version. */ |
376 |
|
|
377 |
static void pchars(unsigned char *p, int length) |
static void pchars(unsigned char *p, int length, int utf8) |
378 |
{ |
{ |
379 |
int c; |
int c; |
380 |
while (length-- > 0) |
while (length-- > 0) |
381 |
|
{ |
382 |
|
if (utf8) |
383 |
|
{ |
384 |
|
int rc = utf82ord(p, &c); |
385 |
|
if (rc > 0) |
386 |
|
{ |
387 |
|
length -= rc - 1; |
388 |
|
p += rc; |
389 |
|
if (c < 256 && isprint(c)) fprintf(outfile, "%c", c); |
390 |
|
else fprintf(outfile, "\\x{%02x}", c); |
391 |
|
continue; |
392 |
|
} |
393 |
|
} |
394 |
|
|
395 |
|
/* Not UTF-8, or malformed UTF-8 */ |
396 |
|
|
397 |
if (isprint(c = *(p++))) fprintf(outfile, "%c", c); |
if (isprint(c = *(p++))) fprintf(outfile, "%c", c); |
398 |
else fprintf(outfile, "\\x%02x", c); |
else fprintf(outfile, "\\x%02x", c); |
399 |
|
} |
400 |
} |
} |
401 |
|
|
402 |
|
|
406 |
|
|
407 |
static void *new_malloc(size_t size) |
static void *new_malloc(size_t size) |
408 |
{ |
{ |
409 |
if (log_store) fprintf(outfile, "Store size request: %d\n", (int)size); |
gotten_store = size; |
410 |
|
if (log_store) |
411 |
|
fprintf(outfile, "Memory allocation (code space): %d\n", |
412 |
|
(int)((int)size - offsetof(real_pcre, code[0]))); |
413 |
return malloc(size); |
return malloc(size); |
414 |
} |
} |
415 |
|
|
416 |
|
|
417 |
|
|
418 |
|
|
419 |
|
/* Get one piece of information from the pcre_fullinfo() function */ |
420 |
|
|
421 |
|
static void new_info(pcre *re, pcre_extra *study, int option, void *ptr) |
422 |
|
{ |
423 |
|
int rc; |
424 |
|
if ((rc = pcre_fullinfo(re, study, option, ptr)) < 0) |
425 |
|
fprintf(outfile, "Error %d from pcre_fullinfo(%d)\n", rc, option); |
426 |
|
} |
427 |
|
|
428 |
|
|
429 |
|
|
430 |
|
|
431 |
/* Read lines from named file or stdin and write to named file or stdout; lines |
/* Read lines from named file or stdin and write to named file or stdout; lines |
432 |
consist of a regular expression, in delimiters and optionally followed by |
consist of a regular expression, in delimiters and optionally followed by |
433 |
options, followed by a set of test data, terminated by an empty line. */ |
options, followed by a set of test data, terminated by an empty line. */ |
440 |
int op = 1; |
int op = 1; |
441 |
int timeit = 0; |
int timeit = 0; |
442 |
int showinfo = 0; |
int showinfo = 0; |
443 |
|
int showstore = 0; |
444 |
int posix = 0; |
int posix = 0; |
445 |
int debug = 0; |
int debug = 0; |
446 |
int done = 0; |
int done = 0; |
455 |
|
|
456 |
while (argc > 1 && argv[op][0] == '-') |
while (argc > 1 && argv[op][0] == '-') |
457 |
{ |
{ |
458 |
if (strcmp(argv[op], "-s") == 0) log_store = 1; |
if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0) |
459 |
|
showstore = 1; |
460 |
else if (strcmp(argv[op], "-t") == 0) timeit = 1; |
else if (strcmp(argv[op], "-t") == 0) timeit = 1; |
461 |
else if (strcmp(argv[op], "-i") == 0) showinfo = 1; |
else if (strcmp(argv[op], "-i") == 0) showinfo = 1; |
462 |
else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1; |
else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1; |
512 |
{ |
{ |
513 |
pcre *re = NULL; |
pcre *re = NULL; |
514 |
pcre_extra *extra = NULL; |
pcre_extra *extra = NULL; |
515 |
|
|
516 |
|
#if !defined NOPOSIX /* There are still compilers that require no indent */ |
517 |
regex_t preg; |
regex_t preg; |
518 |
|
int do_posix = 0; |
519 |
|
#endif |
520 |
|
|
521 |
const char *error; |
const char *error; |
522 |
unsigned char *p, *pp, *ppp; |
unsigned char *p, *pp, *ppp; |
523 |
unsigned const char *tables = NULL; |
unsigned const char *tables = NULL; |
524 |
int do_study = 0; |
int do_study = 0; |
525 |
int do_debug = debug; |
int do_debug = debug; |
526 |
|
int do_G = 0; |
527 |
|
int do_g = 0; |
528 |
int do_showinfo = showinfo; |
int do_showinfo = showinfo; |
529 |
int do_posix = 0; |
int do_showrest = 0; |
530 |
|
int utf8 = 0; |
531 |
int erroroffset, len, delimiter; |
int erroroffset, len, delimiter; |
532 |
|
|
533 |
if (infile == stdin) printf(" re> "); |
if (infile == stdin) printf(" re> "); |
543 |
|
|
544 |
delimiter = *p++; |
delimiter = *p++; |
545 |
|
|
546 |
if (isalnum(delimiter)) |
if (isalnum(delimiter) || delimiter == '\\') |
547 |
{ |
{ |
548 |
fprintf(outfile, "** Delimiter must not be alphameric\n"); |
fprintf(outfile, "** Delimiter must not be alphameric or \\\n"); |
549 |
goto SKIP_DATA; |
goto SKIP_DATA; |
550 |
} |
} |
551 |
|
|
553 |
|
|
554 |
for(;;) |
for(;;) |
555 |
{ |
{ |
556 |
while (*pp != 0 && *pp != delimiter) pp++; |
while (*pp != 0) |
557 |
|
{ |
558 |
|
if (*pp == '\\' && pp[1] != 0) pp++; |
559 |
|
else if (*pp == delimiter) break; |
560 |
|
pp++; |
561 |
|
} |
562 |
if (*pp != 0) break; |
if (*pp != 0) break; |
563 |
|
|
564 |
len = sizeof(buffer) - (pp - buffer); |
len = sizeof(buffer) - (pp - buffer); |
578 |
if (infile != stdin) fprintf(outfile, "%s", (char *)pp); |
if (infile != stdin) fprintf(outfile, "%s", (char *)pp); |
579 |
} |
} |
580 |
|
|
581 |
|
/* If the first character after the delimiter is backslash, make |
582 |
|
the pattern end with backslash. This is purely to provide a way |
583 |
|
of testing for the error message when a pattern ends with backslash. */ |
584 |
|
|
585 |
|
if (pp[1] == '\\') *pp++ = '\\'; |
586 |
|
|
587 |
/* Terminate the pattern at the delimiter */ |
/* Terminate the pattern at the delimiter */ |
588 |
|
|
589 |
*pp++ = 0; |
*pp++ = 0; |
592 |
|
|
593 |
options = 0; |
options = 0; |
594 |
study_options = 0; |
study_options = 0; |
595 |
|
log_store = showstore; /* default from command line */ |
596 |
|
|
597 |
while (*pp != 0) |
while (*pp != 0) |
598 |
{ |
{ |
599 |
switch (*pp++) |
switch (*pp++) |
600 |
{ |
{ |
601 |
|
case 'g': do_g = 1; break; |
602 |
case 'i': options |= PCRE_CASELESS; break; |
case 'i': options |= PCRE_CASELESS; break; |
603 |
case 'm': options |= PCRE_MULTILINE; break; |
case 'm': options |= PCRE_MULTILINE; break; |
604 |
case 's': options |= PCRE_DOTALL; break; |
case 's': options |= PCRE_DOTALL; break; |
605 |
case 'x': options |= PCRE_EXTENDED; break; |
case 'x': options |= PCRE_EXTENDED; break; |
606 |
|
|
607 |
|
case '+': do_showrest = 1; break; |
608 |
case 'A': options |= PCRE_ANCHORED; break; |
case 'A': options |= PCRE_ANCHORED; break; |
609 |
case 'D': do_debug = do_showinfo = 1; break; |
case 'D': do_debug = do_showinfo = 1; break; |
610 |
case 'E': options |= PCRE_DOLLAR_ENDONLY; break; |
case 'E': options |= PCRE_DOLLAR_ENDONLY; break; |
611 |
|
case 'G': do_G = 1; break; |
612 |
case 'I': do_showinfo = 1; break; |
case 'I': do_showinfo = 1; break; |
613 |
|
case 'M': log_store = 1; break; |
614 |
|
|
615 |
|
#if !defined NOPOSIX |
616 |
case 'P': do_posix = 1; break; |
case 'P': do_posix = 1; break; |
617 |
|
#endif |
618 |
|
|
619 |
case 'S': do_study = 1; break; |
case 'S': do_study = 1; break; |
620 |
case 'U': options |= PCRE_UNGREEDY; break; |
case 'U': options |= PCRE_UNGREEDY; break; |
621 |
case 'X': options |= PCRE_EXTRA; break; |
case 'X': options |= PCRE_EXTRA; break; |
622 |
|
case '8': options |= PCRE_UTF8; utf8 = 1; break; |
623 |
|
|
624 |
case 'L': |
case 'L': |
625 |
ppp = pp; |
ppp = pp; |
645 |
timing, showing, or debugging options, nor the ability to pass over |
timing, showing, or debugging options, nor the ability to pass over |
646 |
local character tables. */ |
local character tables. */ |
647 |
|
|
648 |
|
#if !defined NOPOSIX |
649 |
if (posix || do_posix) |
if (posix || do_posix) |
650 |
{ |
{ |
651 |
int rc; |
int rc; |
668 |
/* Handle compiling via the native interface */ |
/* Handle compiling via the native interface */ |
669 |
|
|
670 |
else |
else |
671 |
|
#endif /* !defined NOPOSIX */ |
672 |
|
|
673 |
{ |
{ |
674 |
if (timeit) |
if (timeit) |
675 |
{ |
{ |
682 |
if (re != NULL) free(re); |
if (re != NULL) free(re); |
683 |
} |
} |
684 |
time_taken = clock() - start_time; |
time_taken = clock() - start_time; |
685 |
fprintf(outfile, "Compile time %.2f milliseconds\n", |
fprintf(outfile, "Compile time %.3f milliseconds\n", |
686 |
((double)time_taken)/(4 * CLOCKS_PER_SEC)); |
((double)time_taken * 1000.0) / |
687 |
|
((double)LOOPREPEAT * (double)CLOCKS_PER_SEC)); |
688 |
} |
} |
689 |
|
|
690 |
re = pcre_compile((char *)p, options, &error, &erroroffset, tables); |
re = pcre_compile((char *)p, options, &error, &erroroffset, tables); |
714 |
goto CONTINUE; |
goto CONTINUE; |
715 |
} |
} |
716 |
|
|
717 |
/* Compilation succeeded; print data if required */ |
/* Compilation succeeded; print data if required. There are now two |
718 |
|
info-returning functions. The old one has a limited interface and |
719 |
|
returns only limited data. Check that it agrees with the newer one. */ |
720 |
|
|
721 |
if (do_showinfo) |
if (do_showinfo) |
722 |
{ |
{ |
723 |
int first_char, count; |
int old_first_char, old_options, old_count; |
724 |
|
int count, backrefmax, first_char, need_char; |
725 |
|
size_t size; |
726 |
|
|
727 |
|
if (do_debug) print_internals(re); |
728 |
|
|
729 |
|
new_info(re, NULL, PCRE_INFO_OPTIONS, &options); |
730 |
|
new_info(re, NULL, PCRE_INFO_SIZE, &size); |
731 |
|
new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count); |
732 |
|
new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax); |
733 |
|
new_info(re, NULL, PCRE_INFO_FIRSTCHAR, &first_char); |
734 |
|
new_info(re, NULL, PCRE_INFO_LASTLITERAL, &need_char); |
735 |
|
|
736 |
if (do_debug) print_internals(re, outfile); |
old_count = pcre_info(re, &old_options, &old_first_char); |
|
|
|
|
count = pcre_info(re, &options, &first_char); |
|
737 |
if (count < 0) fprintf(outfile, |
if (count < 0) fprintf(outfile, |
738 |
"Error %d while reading info\n", count); |
"Error %d from pcre_info()\n", count); |
739 |
else |
else |
740 |
{ |
{ |
741 |
fprintf(outfile, "Identifying subpattern count = %d\n", count); |
if (old_count != count) fprintf(outfile, |
742 |
if (options == 0) fprintf(outfile, "No options\n"); |
"Count disagreement: pcre_fullinfo=%d pcre_info=%d\n", count, |
743 |
else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s\n", |
old_count); |
744 |
((options & PCRE_ANCHORED) != 0)? " anchored" : "", |
|
745 |
((options & PCRE_CASELESS) != 0)? " caseless" : "", |
if (old_first_char != first_char) fprintf(outfile, |
746 |
((options & PCRE_EXTENDED) != 0)? " extended" : "", |
"First char disagreement: pcre_fullinfo=%d pcre_info=%d\n", |
747 |
((options & PCRE_MULTILINE) != 0)? " multiline" : "", |
first_char, old_first_char); |
748 |
((options & PCRE_DOTALL) != 0)? " dotall" : "", |
|
749 |
((options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "", |
if (old_options != options) fprintf(outfile, |
750 |
((options & PCRE_EXTRA) != 0)? " extra" : "", |
"Options disagreement: pcre_fullinfo=%d pcre_info=%d\n", options, |
751 |
((options & PCRE_UNGREEDY) != 0)? " ungreedy" : ""); |
old_options); |
752 |
if (first_char == -1) |
} |
753 |
{ |
|
754 |
fprintf(outfile, "First char at start or follows \\n\n"); |
if (size != gotten_store) fprintf(outfile, |
755 |
} |
"Size disagreement: pcre_fullinfo=%d call to malloc for %d\n", |
756 |
else if (first_char < 0) |
size, gotten_store); |
757 |
{ |
|
758 |
fprintf(outfile, "No first char\n"); |
fprintf(outfile, "Capturing subpattern count = %d\n", count); |
759 |
} |
if (backrefmax > 0) |
760 |
|
fprintf(outfile, "Max back reference = %d\n", backrefmax); |
761 |
|
if (options == 0) fprintf(outfile, "No options\n"); |
762 |
|
else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s\n", |
763 |
|
((options & PCRE_ANCHORED) != 0)? " anchored" : "", |
764 |
|
((options & PCRE_CASELESS) != 0)? " caseless" : "", |
765 |
|
((options & PCRE_EXTENDED) != 0)? " extended" : "", |
766 |
|
((options & PCRE_MULTILINE) != 0)? " multiline" : "", |
767 |
|
((options & PCRE_DOTALL) != 0)? " dotall" : "", |
768 |
|
((options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "", |
769 |
|
((options & PCRE_EXTRA) != 0)? " extra" : "", |
770 |
|
((options & PCRE_UNGREEDY) != 0)? " ungreedy" : "", |
771 |
|
((options & PCRE_UTF8) != 0)? " utf8" : ""); |
772 |
|
|
773 |
|
if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0) |
774 |
|
fprintf(outfile, "Case state changes\n"); |
775 |
|
|
776 |
|
if (first_char == -1) |
777 |
|
{ |
778 |
|
fprintf(outfile, "First char at start or follows \\n\n"); |
779 |
|
} |
780 |
|
else if (first_char < 0) |
781 |
|
{ |
782 |
|
fprintf(outfile, "No first char\n"); |
783 |
|
} |
784 |
|
else |
785 |
|
{ |
786 |
|
if (isprint(first_char)) |
787 |
|
fprintf(outfile, "First char = \'%c\'\n", first_char); |
788 |
else |
else |
789 |
{ |
fprintf(outfile, "First char = %d\n", first_char); |
790 |
if (isprint(first_char)) |
} |
791 |
fprintf(outfile, "First char = \'%c\'\n", first_char); |
|
792 |
else |
if (need_char < 0) |
793 |
fprintf(outfile, "First char = %d\n", first_char); |
{ |
794 |
} |
fprintf(outfile, "No need char\n"); |
795 |
|
} |
796 |
|
else |
797 |
|
{ |
798 |
|
if (isprint(need_char)) |
799 |
|
fprintf(outfile, "Need char = \'%c\'\n", need_char); |
800 |
|
else |
801 |
|
fprintf(outfile, "Need char = %d\n", need_char); |
802 |
} |
} |
803 |
} |
} |
804 |
|
|
816 |
extra = pcre_study(re, study_options, &error); |
extra = pcre_study(re, study_options, &error); |
817 |
time_taken = clock() - start_time; |
time_taken = clock() - start_time; |
818 |
if (extra != NULL) free(extra); |
if (extra != NULL) free(extra); |
819 |
fprintf(outfile, " Study time %.2f milliseconds\n", |
fprintf(outfile, " Study time %.3f milliseconds\n", |
820 |
((double)time_taken)/(4 * CLOCKS_PER_SEC)); |
((double)time_taken * 1000.0)/ |
821 |
|
((double)LOOPREPEAT * (double)CLOCKS_PER_SEC)); |
822 |
} |
} |
823 |
|
|
824 |
extra = pcre_study(re, study_options, &error); |
extra = pcre_study(re, study_options, &error); |
827 |
else if (extra == NULL) |
else if (extra == NULL) |
828 |
fprintf(outfile, "Study returned NULL\n"); |
fprintf(outfile, "Study returned NULL\n"); |
829 |
|
|
|
/* This looks at internal information. A bit kludgy to do it this |
|
|
way, but it is useful for testing. */ |
|
|
|
|
830 |
else if (do_showinfo) |
else if (do_showinfo) |
831 |
{ |
{ |
832 |
real_pcre_extra *xx = (real_pcre_extra *)extra; |
uschar *start_bits = NULL; |
833 |
if ((xx->options & PCRE_STUDY_MAPPED) == 0) |
new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits); |
834 |
|
if (start_bits == NULL) |
835 |
fprintf(outfile, "No starting character set\n"); |
fprintf(outfile, "No starting character set\n"); |
836 |
else |
else |
837 |
{ |
{ |
840 |
fprintf(outfile, "Starting character set: "); |
fprintf(outfile, "Starting character set: "); |
841 |
for (i = 0; i < 256; i++) |
for (i = 0; i < 256; i++) |
842 |
{ |
{ |
843 |
if ((xx->start_bits[i/8] & (1<<(i%8))) != 0) |
if ((start_bits[i/8] & (1<<(i%8))) != 0) |
844 |
{ |
{ |
845 |
if (c > 75) |
if (c > 75) |
846 |
{ |
{ |
870 |
for (;;) |
for (;;) |
871 |
{ |
{ |
872 |
unsigned char *q; |
unsigned char *q; |
873 |
|
unsigned char *bptr = dbuffer; |
874 |
int count, c; |
int count, c; |
875 |
|
int copystrings = 0; |
876 |
|
int getstrings = 0; |
877 |
|
int getlist = 0; |
878 |
|
int gmatched = 0; |
879 |
|
int start_offset = 0; |
880 |
|
int g_notempty = 0; |
881 |
int offsets[45]; |
int offsets[45]; |
882 |
int size_offsets = sizeof(offsets)/sizeof(int); |
int size_offsets = sizeof(offsets)/sizeof(int); |
883 |
|
|
884 |
options = 0; |
options = 0; |
885 |
|
|
886 |
if (infile == stdin) printf(" data> "); |
if (infile == stdin) printf("data> "); |
887 |
if (fgets((char *)buffer, sizeof(buffer), infile) == NULL) |
if (fgets((char *)buffer, sizeof(buffer), infile) == NULL) |
888 |
{ |
{ |
889 |
done = 1; |
done = 1; |
923 |
break; |
break; |
924 |
|
|
925 |
case 'x': |
case 'x': |
926 |
|
|
927 |
|
/* Handle \x{..} specially - new Perl thing for utf8 */ |
928 |
|
|
929 |
|
if (*p == '{') |
930 |
|
{ |
931 |
|
unsigned char *pt = p; |
932 |
|
c = 0; |
933 |
|
while (isxdigit(*(++pt))) |
934 |
|
c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W'); |
935 |
|
if (*pt == '}') |
936 |
|
{ |
937 |
|
unsigned char buffer[8]; |
938 |
|
int ii, utn; |
939 |
|
utn = ord2utf8(c, buffer); |
940 |
|
for (ii = 0; ii < utn - 1; ii++) *q++ = buffer[ii]; |
941 |
|
c = buffer[ii]; /* Last byte */ |
942 |
|
p = pt + 1; |
943 |
|
break; |
944 |
|
} |
945 |
|
/* Not correct form; fall through */ |
946 |
|
} |
947 |
|
|
948 |
|
/* Ordinary \x */ |
949 |
|
|
950 |
c = 0; |
c = 0; |
951 |
while (i++ < 2 && isxdigit(*p)) |
while (i++ < 2 && isxdigit(*p)) |
952 |
{ |
{ |
967 |
options |= PCRE_NOTBOL; |
options |= PCRE_NOTBOL; |
968 |
continue; |
continue; |
969 |
|
|
970 |
|
case 'C': |
971 |
|
while(isdigit(*p)) n = n * 10 + *p++ - '0'; |
972 |
|
copystrings |= 1 << n; |
973 |
|
continue; |
974 |
|
|
975 |
|
case 'G': |
976 |
|
while(isdigit(*p)) n = n * 10 + *p++ - '0'; |
977 |
|
getstrings |= 1 << n; |
978 |
|
continue; |
979 |
|
|
980 |
|
case 'L': |
981 |
|
getlist = 1; |
982 |
|
continue; |
983 |
|
|
984 |
|
case 'N': |
985 |
|
options |= PCRE_NOTEMPTY; |
986 |
|
continue; |
987 |
|
|
988 |
case 'O': |
case 'O': |
989 |
while(isdigit(*p)) n = n * 10 + *p++ - '0'; |
while(isdigit(*p)) n = n * 10 + *p++ - '0'; |
990 |
if (n <= (int)(sizeof(offsets)/sizeof(int))) size_offsets = n; |
if (n <= (int)(sizeof(offsets)/sizeof(int))) size_offsets = n; |
1002 |
/* Handle matching via the POSIX interface, which does not |
/* Handle matching via the POSIX interface, which does not |
1003 |
support timing. */ |
support timing. */ |
1004 |
|
|
1005 |
|
#if !defined NOPOSIX |
1006 |
if (posix || do_posix) |
if (posix || do_posix) |
1007 |
{ |
{ |
1008 |
int rc; |
int rc; |
1009 |
int eflags = 0; |
int eflags = 0; |
1010 |
regmatch_t pmatch[30]; |
regmatch_t pmatch[sizeof(offsets)/sizeof(int)]; |
1011 |
if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL; |
if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL; |
1012 |
if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL; |
if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL; |
1013 |
|
|
1014 |
rc = regexec(&preg, (char *)dbuffer, sizeof(pmatch)/sizeof(regmatch_t), |
rc = regexec(&preg, (const char *)bptr, size_offsets, pmatch, eflags); |
|
pmatch, eflags); |
|
1015 |
|
|
1016 |
if (rc != 0) |
if (rc != 0) |
1017 |
{ |
{ |
1021 |
else |
else |
1022 |
{ |
{ |
1023 |
size_t i; |
size_t i; |
1024 |
for (i = 0; i < sizeof(pmatch)/sizeof(regmatch_t); i++) |
for (i = 0; i < size_offsets; i++) |
1025 |
{ |
{ |
1026 |
if (pmatch[i].rm_so >= 0) |
if (pmatch[i].rm_so >= 0) |
1027 |
{ |
{ |
1028 |
fprintf(outfile, "%2d: ", (int)i); |
fprintf(outfile, "%2d: ", (int)i); |
1029 |
pchars(dbuffer + pmatch[i].rm_so, |
pchars(dbuffer + pmatch[i].rm_so, |
1030 |
pmatch[i].rm_eo - pmatch[i].rm_so); |
pmatch[i].rm_eo - pmatch[i].rm_so, utf8); |
1031 |
fprintf(outfile, "\n"); |
fprintf(outfile, "\n"); |
1032 |
|
if (i == 0 && do_showrest) |
1033 |
|
{ |
1034 |
|
fprintf(outfile, " 0+ "); |
1035 |
|
pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo, utf8); |
1036 |
|
fprintf(outfile, "\n"); |
1037 |
|
} |
1038 |
} |
} |
1039 |
} |
} |
1040 |
} |
} |
1041 |
} |
} |
1042 |
|
|
1043 |
/* Handle matching via the native interface */ |
/* Handle matching via the native interface - repeats for /g and /G */ |
1044 |
|
|
1045 |
else |
else |
1046 |
|
#endif /* !defined NOPOSIX */ |
1047 |
|
|
1048 |
|
for (;; gmatched++) /* Loop for /g or /G */ |
1049 |
{ |
{ |
1050 |
if (timeit) |
if (timeit) |
1051 |
{ |
{ |
1052 |
register int i; |
register int i; |
1053 |
clock_t time_taken; |
clock_t time_taken; |
1054 |
clock_t start_time = clock(); |
clock_t start_time = clock(); |
1055 |
for (i = 0; i < 4000; i++) |
for (i = 0; i < LOOPREPEAT; i++) |
1056 |
count = pcre_exec(re, extra, (char *)dbuffer, len, options, offsets, |
count = pcre_exec(re, extra, (char *)bptr, len, |
1057 |
size_offsets); |
start_offset, options | g_notempty, offsets, size_offsets); |
1058 |
time_taken = clock() - start_time; |
time_taken = clock() - start_time; |
1059 |
fprintf(outfile, "Execute time %.2f milliseconds\n", |
fprintf(outfile, "Execute time %.3f milliseconds\n", |
1060 |
((double)time_taken)/(4 * CLOCKS_PER_SEC)); |
((double)time_taken * 1000.0)/ |
1061 |
|
((double)LOOPREPEAT * (double)CLOCKS_PER_SEC)); |
1062 |
} |
} |
1063 |
|
|
1064 |
count = pcre_exec(re, extra, (char *)dbuffer, len, options, offsets, |
count = pcre_exec(re, extra, (char *)bptr, len, |
1065 |
size_offsets); |
start_offset, options | g_notempty, offsets, size_offsets); |
1066 |
|
|
1067 |
if (count == 0) |
if (count == 0) |
1068 |
{ |
{ |
1070 |
count = size_offsets/3; |
count = size_offsets/3; |
1071 |
} |
} |
1072 |
|
|
1073 |
|
/* Matched */ |
1074 |
|
|
1075 |
if (count >= 0) |
if (count >= 0) |
1076 |
{ |
{ |
1077 |
int i; |
int i; |
1078 |
count *= 2; |
for (i = 0; i < count * 2; i += 2) |
|
for (i = 0; i < count; i += 2) |
|
1079 |
{ |
{ |
1080 |
if (offsets[i] < 0) |
if (offsets[i] < 0) |
1081 |
fprintf(outfile, "%2d: <unset>\n", i/2); |
fprintf(outfile, "%2d: <unset>\n", i/2); |
1082 |
else |
else |
1083 |
{ |
{ |
1084 |
fprintf(outfile, "%2d: ", i/2); |
fprintf(outfile, "%2d: ", i/2); |
1085 |
pchars(dbuffer + offsets[i], offsets[i+1] - offsets[i]); |
pchars(bptr + offsets[i], offsets[i+1] - offsets[i], utf8); |
1086 |
fprintf(outfile, "\n"); |
fprintf(outfile, "\n"); |
1087 |
|
if (i == 0) |
1088 |
|
{ |
1089 |
|
if (do_showrest) |
1090 |
|
{ |
1091 |
|
fprintf(outfile, " 0+ "); |
1092 |
|
pchars(bptr + offsets[i+1], len - offsets[i+1], utf8); |
1093 |
|
fprintf(outfile, "\n"); |
1094 |
|
} |
1095 |
|
} |
1096 |
|
} |
1097 |
|
} |
1098 |
|
|
1099 |
|
for (i = 0; i < 32; i++) |
1100 |
|
{ |
1101 |
|
if ((copystrings & (1 << i)) != 0) |
1102 |
|
{ |
1103 |
|
char copybuffer[16]; |
1104 |
|
int rc = pcre_copy_substring((char *)bptr, offsets, count, |
1105 |
|
i, copybuffer, sizeof(copybuffer)); |
1106 |
|
if (rc < 0) |
1107 |
|
fprintf(outfile, "copy substring %d failed %d\n", i, rc); |
1108 |
|
else |
1109 |
|
fprintf(outfile, "%2dC %s (%d)\n", i, copybuffer, rc); |
1110 |
|
} |
1111 |
|
} |
1112 |
|
|
1113 |
|
for (i = 0; i < 32; i++) |
1114 |
|
{ |
1115 |
|
if ((getstrings & (1 << i)) != 0) |
1116 |
|
{ |
1117 |
|
const char *substring; |
1118 |
|
int rc = pcre_get_substring((char *)bptr, offsets, count, |
1119 |
|
i, &substring); |
1120 |
|
if (rc < 0) |
1121 |
|
fprintf(outfile, "get substring %d failed %d\n", i, rc); |
1122 |
|
else |
1123 |
|
{ |
1124 |
|
fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc); |
1125 |
|
/* free((void *)substring); */ |
1126 |
|
pcre_free_substring(substring); |
1127 |
|
} |
1128 |
|
} |
1129 |
|
} |
1130 |
|
|
1131 |
|
if (getlist) |
1132 |
|
{ |
1133 |
|
const char **stringlist; |
1134 |
|
int rc = pcre_get_substring_list((char *)bptr, offsets, count, |
1135 |
|
&stringlist); |
1136 |
|
if (rc < 0) |
1137 |
|
fprintf(outfile, "get substring list failed %d\n", rc); |
1138 |
|
else |
1139 |
|
{ |
1140 |
|
for (i = 0; i < count; i++) |
1141 |
|
fprintf(outfile, "%2dL %s\n", i, stringlist[i]); |
1142 |
|
if (stringlist[i] != NULL) |
1143 |
|
fprintf(outfile, "string list not terminated by NULL\n"); |
1144 |
|
/* free((void *)stringlist); */ |
1145 |
|
pcre_free_substring_list(stringlist); |
1146 |
} |
} |
1147 |
} |
} |
1148 |
} |
} |
1149 |
|
|
1150 |
|
/* Failed to match. If this is a /g or /G loop and we previously set |
1151 |
|
g_notempty after a null match, this is not necessarily the end. |
1152 |
|
We want to advance the start offset, and continue. Fudge the offset |
1153 |
|
values to achieve this. We won't be at the end of the string - that |
1154 |
|
was checked before setting g_notempty. */ |
1155 |
|
|
1156 |
else |
else |
1157 |
{ |
{ |
1158 |
if (count == -1) fprintf(outfile, "No match\n"); |
if (g_notempty != 0) |
1159 |
else fprintf(outfile, "Error %d\n", count); |
{ |
1160 |
|
offsets[0] = start_offset; |
1161 |
|
offsets[1] = start_offset + 1; |
1162 |
|
} |
1163 |
|
else |
1164 |
|
{ |
1165 |
|
if (gmatched == 0) /* Error if no previous matches */ |
1166 |
|
{ |
1167 |
|
if (count == -1) fprintf(outfile, "No match\n"); |
1168 |
|
else fprintf(outfile, "Error %d\n", count); |
1169 |
|
} |
1170 |
|
break; /* Out of the /g loop */ |
1171 |
|
} |
1172 |
|
} |
1173 |
|
|
1174 |
|
/* If not /g or /G we are done */ |
1175 |
|
|
1176 |
|
if (!do_g && !do_G) break; |
1177 |
|
|
1178 |
|
/* If we have matched an empty string, first check to see if we are at |
1179 |
|
the end of the subject. If so, the /g loop is over. Otherwise, mimic |
1180 |
|
what Perl's /g options does. This turns out to be rather cunning. First |
1181 |
|
we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the |
1182 |
|
same point. If this fails (picked up above) we advance to the next |
1183 |
|
character. */ |
1184 |
|
|
1185 |
|
g_notempty = 0; |
1186 |
|
if (offsets[0] == offsets[1]) |
1187 |
|
{ |
1188 |
|
if (offsets[0] == len) break; |
1189 |
|
g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED; |
1190 |
} |
} |
1191 |
} |
|
1192 |
} |
/* For /g, update the start offset, leaving the rest alone */ |
1193 |
|
|
1194 |
|
if (do_g) start_offset = offsets[1]; |
1195 |
|
|
1196 |
|
/* For /G, update the pointer and length */ |
1197 |
|
|
1198 |
|
else |
1199 |
|
{ |
1200 |
|
bptr += offsets[1]; |
1201 |
|
len -= offsets[1]; |
1202 |
|
} |
1203 |
|
} /* End of loop for /g and /G */ |
1204 |
|
} /* End of loop for data lines */ |
1205 |
|
|
1206 |
CONTINUE: |
CONTINUE: |
1207 |
|
|
1208 |
|
#if !defined NOPOSIX |
1209 |
if (posix || do_posix) regfree(&preg); |
if (posix || do_posix) regfree(&preg); |
1210 |
|
#endif |
1211 |
|
|
1212 |
if (re != NULL) free(re); |
if (re != NULL) free(re); |
1213 |
if (extra != NULL) free(extra); |
if (extra != NULL) free(extra); |
1214 |
if (tables != NULL) |
if (tables != NULL) |