1 |
/*************************************************
|
2 |
* PCRE testing program *
|
3 |
*************************************************/
|
4 |
|
5 |
#include <ctype.h>
|
6 |
#include <stdio.h>
|
7 |
#include <string.h>
|
8 |
#include <stdlib.h>
|
9 |
#include <time.h>
|
10 |
#include <locale.h>
|
11 |
|
12 |
/* Use the internal info for displaying the results of pcre_study(). */
|
13 |
|
14 |
#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"
|
22 |
#endif
|
23 |
|
24 |
#ifndef CLOCKS_PER_SEC
|
25 |
#ifdef CLK_TCK
|
26 |
#define CLOCKS_PER_SEC CLK_TCK
|
27 |
#else
|
28 |
#define CLOCKS_PER_SEC 100
|
29 |
#endif
|
30 |
#endif
|
31 |
|
32 |
#define LOOPREPEAT 20000
|
33 |
|
34 |
|
35 |
static FILE *outfile;
|
36 |
static int log_store = 0;
|
37 |
|
38 |
|
39 |
|
40 |
/* Debugging function to print the internal form of the regex. This is the same
|
41 |
code as contained in pcre.c under the DEBUG macro. */
|
42 |
|
43 |
static const char *OP_names[] = {
|
44 |
"End", "\\A", "\\B", "\\b", "\\D", "\\d",
|
45 |
"\\S", "\\s", "\\W", "\\w", "\\Z", "\\z",
|
46 |
"Opt", "^", "$", "Any", "chars", "not",
|
47 |
"*", "*?", "+", "+?", "?", "??", "{", "{", "{",
|
48 |
"*", "*?", "+", "+?", "?", "??", "{", "{", "{",
|
49 |
"*", "*?", "+", "+?", "?", "??", "{", "{", "{",
|
50 |
"*", "*?", "+", "+?", "?", "??", "{", "{",
|
51 |
"class", "Ref",
|
52 |
"Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",
|
53 |
"AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cref",
|
54 |
"Brazero", "Braminzero", "Bra"
|
55 |
};
|
56 |
|
57 |
|
58 |
static void print_internals(pcre *re)
|
59 |
{
|
60 |
unsigned char *code = ((real_pcre *)re)->code;
|
61 |
|
62 |
fprintf(outfile, "------------------------------------------------------------------\n");
|
63 |
|
64 |
for(;;)
|
65 |
{
|
66 |
int c;
|
67 |
int charlength;
|
68 |
|
69 |
fprintf(outfile, "%3d ", (int)(code - ((real_pcre *)re)->code));
|
70 |
|
71 |
if (*code >= OP_BRA)
|
72 |
{
|
73 |
fprintf(outfile, "%3d Bra %d", (code[1] << 8) + code[2], *code - OP_BRA);
|
74 |
code += 2;
|
75 |
}
|
76 |
|
77 |
else switch(*code)
|
78 |
{
|
79 |
case OP_END:
|
80 |
fprintf(outfile, " %s\n", OP_names[*code]);
|
81 |
fprintf(outfile, "------------------------------------------------------------------\n");
|
82 |
return;
|
83 |
|
84 |
case OP_OPT:
|
85 |
fprintf(outfile, " %.2x %s", code[1], OP_names[*code]);
|
86 |
code++;
|
87 |
break;
|
88 |
|
89 |
case OP_COND:
|
90 |
fprintf(outfile, "%3d Cond", (code[1] << 8) + code[2]);
|
91 |
code += 2;
|
92 |
break;
|
93 |
|
94 |
case OP_CREF:
|
95 |
fprintf(outfile, " %.2d %s", code[1], OP_names[*code]);
|
96 |
code++;
|
97 |
break;
|
98 |
|
99 |
case OP_CHARS:
|
100 |
charlength = *(++code);
|
101 |
fprintf(outfile, "%3d ", charlength);
|
102 |
while (charlength-- > 0)
|
103 |
if (isprint(c = *(++code))) fprintf(outfile, "%c", c);
|
104 |
else fprintf(outfile, "\\x%02x", c);
|
105 |
break;
|
106 |
|
107 |
case OP_KETRMAX:
|
108 |
case OP_KETRMIN:
|
109 |
case OP_ALT:
|
110 |
case OP_KET:
|
111 |
case OP_ASSERT:
|
112 |
case OP_ASSERT_NOT:
|
113 |
case OP_ASSERTBACK:
|
114 |
case OP_ASSERTBACK_NOT:
|
115 |
case OP_ONCE:
|
116 |
fprintf(outfile, "%3d %s", (code[1] << 8) + code[2], OP_names[*code]);
|
117 |
code += 2;
|
118 |
break;
|
119 |
|
120 |
case OP_REVERSE:
|
121 |
fprintf(outfile, "%3d %s", (code[1] << 8) + code[2], OP_names[*code]);
|
122 |
code += 2;
|
123 |
break;
|
124 |
|
125 |
case OP_STAR:
|
126 |
case OP_MINSTAR:
|
127 |
case OP_PLUS:
|
128 |
case OP_MINPLUS:
|
129 |
case OP_QUERY:
|
130 |
case OP_MINQUERY:
|
131 |
case OP_TYPESTAR:
|
132 |
case OP_TYPEMINSTAR:
|
133 |
case OP_TYPEPLUS:
|
134 |
case OP_TYPEMINPLUS:
|
135 |
case OP_TYPEQUERY:
|
136 |
case OP_TYPEMINQUERY:
|
137 |
if (*code >= OP_TYPESTAR)
|
138 |
fprintf(outfile, " %s", OP_names[code[1]]);
|
139 |
else if (isprint(c = code[1])) fprintf(outfile, " %c", c);
|
140 |
else fprintf(outfile, " \\x%02x", c);
|
141 |
fprintf(outfile, "%s", OP_names[*code++]);
|
142 |
break;
|
143 |
|
144 |
case OP_EXACT:
|
145 |
case OP_UPTO:
|
146 |
case OP_MINUPTO:
|
147 |
if (isprint(c = code[3])) fprintf(outfile, " %c{", c);
|
148 |
else fprintf(outfile, " \\x%02x{", c);
|
149 |
if (*code != OP_EXACT) fprintf(outfile, ",");
|
150 |
fprintf(outfile, "%d}", (code[1] << 8) + code[2]);
|
151 |
if (*code == OP_MINUPTO) fprintf(outfile, "?");
|
152 |
code += 3;
|
153 |
break;
|
154 |
|
155 |
case OP_TYPEEXACT:
|
156 |
case OP_TYPEUPTO:
|
157 |
case OP_TYPEMINUPTO:
|
158 |
fprintf(outfile, " %s{", OP_names[code[3]]);
|
159 |
if (*code != OP_TYPEEXACT) fprintf(outfile, "0,");
|
160 |
fprintf(outfile, "%d}", (code[1] << 8) + code[2]);
|
161 |
if (*code == OP_TYPEMINUPTO) fprintf(outfile, "?");
|
162 |
code += 3;
|
163 |
break;
|
164 |
|
165 |
case OP_NOT:
|
166 |
if (isprint(c = *(++code))) fprintf(outfile, " [^%c]", c);
|
167 |
else fprintf(outfile, " [^\\x%02x]", c);
|
168 |
break;
|
169 |
|
170 |
case OP_NOTSTAR:
|
171 |
case OP_NOTMINSTAR:
|
172 |
case OP_NOTPLUS:
|
173 |
case OP_NOTMINPLUS:
|
174 |
case OP_NOTQUERY:
|
175 |
case OP_NOTMINQUERY:
|
176 |
if (isprint(c = code[1])) fprintf(outfile, " [^%c]", c);
|
177 |
else fprintf(outfile, " [^\\x%02x]", c);
|
178 |
fprintf(outfile, "%s", OP_names[*code++]);
|
179 |
break;
|
180 |
|
181 |
case OP_NOTEXACT:
|
182 |
case OP_NOTUPTO:
|
183 |
case OP_NOTMINUPTO:
|
184 |
if (isprint(c = code[3])) fprintf(outfile, " [^%c]{", c);
|
185 |
else fprintf(outfile, " [^\\x%02x]{", c);
|
186 |
if (*code != OP_NOTEXACT) fprintf(outfile, ",");
|
187 |
fprintf(outfile, "%d}", (code[1] << 8) + code[2]);
|
188 |
if (*code == OP_NOTMINUPTO) fprintf(outfile, "?");
|
189 |
code += 3;
|
190 |
break;
|
191 |
|
192 |
case OP_REF:
|
193 |
fprintf(outfile, " \\%d", *(++code));
|
194 |
code++;
|
195 |
goto CLASS_REF_REPEAT;
|
196 |
|
197 |
case OP_CLASS:
|
198 |
{
|
199 |
int i, min, max;
|
200 |
code++;
|
201 |
fprintf(outfile, " [");
|
202 |
|
203 |
for (i = 0; i < 256; i++)
|
204 |
{
|
205 |
if ((code[i/8] & (1 << (i&7))) != 0)
|
206 |
{
|
207 |
int j;
|
208 |
for (j = i+1; j < 256; j++)
|
209 |
if ((code[j/8] & (1 << (j&7))) == 0) break;
|
210 |
if (i == '-' || i == ']') fprintf(outfile, "\\");
|
211 |
if (isprint(i)) fprintf(outfile, "%c", i); else fprintf(outfile, "\\x%02x", i);
|
212 |
if (--j > i)
|
213 |
{
|
214 |
fprintf(outfile, "-");
|
215 |
if (j == '-' || j == ']') fprintf(outfile, "\\");
|
216 |
if (isprint(j)) fprintf(outfile, "%c", j); else fprintf(outfile, "\\x%02x", j);
|
217 |
}
|
218 |
i = j;
|
219 |
}
|
220 |
}
|
221 |
fprintf(outfile, "]");
|
222 |
code += 32;
|
223 |
|
224 |
CLASS_REF_REPEAT:
|
225 |
|
226 |
switch(*code)
|
227 |
{
|
228 |
case OP_CRSTAR:
|
229 |
case OP_CRMINSTAR:
|
230 |
case OP_CRPLUS:
|
231 |
case OP_CRMINPLUS:
|
232 |
case OP_CRQUERY:
|
233 |
case OP_CRMINQUERY:
|
234 |
fprintf(outfile, "%s", OP_names[*code]);
|
235 |
break;
|
236 |
|
237 |
case OP_CRRANGE:
|
238 |
case OP_CRMINRANGE:
|
239 |
min = (code[1] << 8) + code[2];
|
240 |
max = (code[3] << 8) + code[4];
|
241 |
if (max == 0) fprintf(outfile, "{%d,}", min);
|
242 |
else fprintf(outfile, "{%d,%d}", min, max);
|
243 |
if (*code == OP_CRMINRANGE) fprintf(outfile, "?");
|
244 |
code += 4;
|
245 |
break;
|
246 |
|
247 |
default:
|
248 |
code--;
|
249 |
}
|
250 |
}
|
251 |
break;
|
252 |
|
253 |
/* Anything else is just a one-node item */
|
254 |
|
255 |
default:
|
256 |
fprintf(outfile, " %s", OP_names[*code]);
|
257 |
break;
|
258 |
}
|
259 |
|
260 |
code++;
|
261 |
fprintf(outfile, "\n");
|
262 |
}
|
263 |
}
|
264 |
|
265 |
|
266 |
|
267 |
/* Character string printing function. */
|
268 |
|
269 |
static void pchars(unsigned char *p, int length)
|
270 |
{
|
271 |
int c;
|
272 |
while (length-- > 0)
|
273 |
if (isprint(c = *(p++))) fprintf(outfile, "%c", c);
|
274 |
else fprintf(outfile, "\\x%02x", c);
|
275 |
}
|
276 |
|
277 |
|
278 |
|
279 |
/* Alternative malloc function, to test functionality and show the size of the
|
280 |
compiled re. */
|
281 |
|
282 |
static void *new_malloc(size_t size)
|
283 |
{
|
284 |
if (log_store)
|
285 |
fprintf(outfile, "Memory allocation (code space): %d\n",
|
286 |
(int)((int)size - offsetof(real_pcre, code[0])));
|
287 |
return malloc(size);
|
288 |
}
|
289 |
|
290 |
|
291 |
|
292 |
/* Read lines from named file or stdin and write to named file or stdout; lines
|
293 |
consist of a regular expression, in delimiters and optionally followed by
|
294 |
options, followed by a set of test data, terminated by an empty line. */
|
295 |
|
296 |
int main(int argc, char **argv)
|
297 |
{
|
298 |
FILE *infile = stdin;
|
299 |
int options = 0;
|
300 |
int study_options = 0;
|
301 |
int op = 1;
|
302 |
int timeit = 0;
|
303 |
int showinfo = 0;
|
304 |
int showstore = 0;
|
305 |
int posix = 0;
|
306 |
int debug = 0;
|
307 |
int done = 0;
|
308 |
unsigned char buffer[30000];
|
309 |
unsigned char dbuffer[1024];
|
310 |
|
311 |
/* Static so that new_malloc can use it. */
|
312 |
|
313 |
outfile = stdout;
|
314 |
|
315 |
/* Scan options */
|
316 |
|
317 |
while (argc > 1 && argv[op][0] == '-')
|
318 |
{
|
319 |
if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
|
320 |
showstore = 1;
|
321 |
else if (strcmp(argv[op], "-t") == 0) timeit = 1;
|
322 |
else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
|
323 |
else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
|
324 |
else if (strcmp(argv[op], "-p") == 0) posix = 1;
|
325 |
else
|
326 |
{
|
327 |
printf("*** Unknown option %s\n", argv[op]);
|
328 |
printf("Usage: pcretest [-d] [-i] [-p] [-s] [-t] [<input> [<output>]]\n");
|
329 |
printf(" -d debug: show compiled code; implies -i\n"
|
330 |
" -i show information about compiled pattern\n"
|
331 |
" -p use POSIX interface\n"
|
332 |
" -s output store information\n"
|
333 |
" -t time compilation and execution\n");
|
334 |
return 1;
|
335 |
}
|
336 |
op++;
|
337 |
argc--;
|
338 |
}
|
339 |
|
340 |
/* Sort out the input and output files */
|
341 |
|
342 |
if (argc > 1)
|
343 |
{
|
344 |
infile = fopen(argv[op], "r");
|
345 |
if (infile == NULL)
|
346 |
{
|
347 |
printf("** Failed to open %s\n", argv[op]);
|
348 |
return 1;
|
349 |
}
|
350 |
}
|
351 |
|
352 |
if (argc > 2)
|
353 |
{
|
354 |
outfile = fopen(argv[op+1], "w");
|
355 |
if (outfile == NULL)
|
356 |
{
|
357 |
printf("** Failed to open %s\n", argv[op+1]);
|
358 |
return 1;
|
359 |
}
|
360 |
}
|
361 |
|
362 |
/* Set alternative malloc function */
|
363 |
|
364 |
pcre_malloc = new_malloc;
|
365 |
|
366 |
/* Heading line, then prompt for first regex if stdin */
|
367 |
|
368 |
fprintf(outfile, "PCRE version %s\n\n", pcre_version());
|
369 |
|
370 |
/* Main loop */
|
371 |
|
372 |
while (!done)
|
373 |
{
|
374 |
pcre *re = NULL;
|
375 |
pcre_extra *extra = NULL;
|
376 |
|
377 |
#if !defined NOPOSIX /* There are still compilers that require no indent */
|
378 |
regex_t preg;
|
379 |
#endif
|
380 |
|
381 |
const char *error;
|
382 |
unsigned char *p, *pp, *ppp;
|
383 |
unsigned const char *tables = NULL;
|
384 |
int do_study = 0;
|
385 |
int do_debug = debug;
|
386 |
int do_G = 0;
|
387 |
int do_g = 0;
|
388 |
int do_showinfo = showinfo;
|
389 |
int do_showrest = 0;
|
390 |
int do_posix = 0;
|
391 |
int erroroffset, len, delimiter;
|
392 |
|
393 |
if (infile == stdin) printf(" re> ");
|
394 |
if (fgets((char *)buffer, sizeof(buffer), infile) == NULL) break;
|
395 |
if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
|
396 |
|
397 |
p = buffer;
|
398 |
while (isspace(*p)) p++;
|
399 |
if (*p == 0) continue;
|
400 |
|
401 |
/* Get the delimiter and seek the end of the pattern; if is isn't
|
402 |
complete, read more. */
|
403 |
|
404 |
delimiter = *p++;
|
405 |
|
406 |
if (isalnum(delimiter) || delimiter == '\\')
|
407 |
{
|
408 |
fprintf(outfile, "** Delimiter must not be alphameric or \\\n");
|
409 |
goto SKIP_DATA;
|
410 |
}
|
411 |
|
412 |
pp = p;
|
413 |
|
414 |
for(;;)
|
415 |
{
|
416 |
while (*pp != 0)
|
417 |
{
|
418 |
if (*pp == '\\' && pp[1] != 0) pp++;
|
419 |
else if (*pp == delimiter) break;
|
420 |
pp++;
|
421 |
}
|
422 |
if (*pp != 0) break;
|
423 |
|
424 |
len = sizeof(buffer) - (pp - buffer);
|
425 |
if (len < 256)
|
426 |
{
|
427 |
fprintf(outfile, "** Expression too long - missing delimiter?\n");
|
428 |
goto SKIP_DATA;
|
429 |
}
|
430 |
|
431 |
if (infile == stdin) printf(" > ");
|
432 |
if (fgets((char *)pp, len, infile) == NULL)
|
433 |
{
|
434 |
fprintf(outfile, "** Unexpected EOF\n");
|
435 |
done = 1;
|
436 |
goto CONTINUE;
|
437 |
}
|
438 |
if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
|
439 |
}
|
440 |
|
441 |
/* If the first character after the delimiter is backslash, make
|
442 |
the pattern end with backslash. This is purely to provide a way
|
443 |
of testing for the error message when a pattern ends with backslash. */
|
444 |
|
445 |
if (pp[1] == '\\') *pp++ = '\\';
|
446 |
|
447 |
/* Terminate the pattern at the delimiter */
|
448 |
|
449 |
*pp++ = 0;
|
450 |
|
451 |
/* Look for options after final delimiter */
|
452 |
|
453 |
options = 0;
|
454 |
study_options = 0;
|
455 |
log_store = showstore; /* default from command line */
|
456 |
|
457 |
while (*pp != 0)
|
458 |
{
|
459 |
switch (*pp++)
|
460 |
{
|
461 |
case 'g': do_g = 1; break;
|
462 |
case 'i': options |= PCRE_CASELESS; break;
|
463 |
case 'm': options |= PCRE_MULTILINE; break;
|
464 |
case 's': options |= PCRE_DOTALL; break;
|
465 |
case 'x': options |= PCRE_EXTENDED; break;
|
466 |
|
467 |
case '+': do_showrest = 1; break;
|
468 |
case 'A': options |= PCRE_ANCHORED; break;
|
469 |
case 'D': do_debug = do_showinfo = 1; break;
|
470 |
case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
|
471 |
case 'G': do_G = 1; break;
|
472 |
case 'I': do_showinfo = 1; break;
|
473 |
case 'M': log_store = 1; break;
|
474 |
|
475 |
#if !defined NOPOSIX
|
476 |
case 'P': do_posix = 1; break;
|
477 |
#endif
|
478 |
|
479 |
case 'S': do_study = 1; break;
|
480 |
case 'U': options |= PCRE_UNGREEDY; break;
|
481 |
case 'X': options |= PCRE_EXTRA; break;
|
482 |
|
483 |
case 'L':
|
484 |
ppp = pp;
|
485 |
while (*ppp != '\n' && *ppp != ' ') ppp++;
|
486 |
*ppp = 0;
|
487 |
if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
|
488 |
{
|
489 |
fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
|
490 |
goto SKIP_DATA;
|
491 |
}
|
492 |
tables = pcre_maketables();
|
493 |
pp = ppp;
|
494 |
break;
|
495 |
|
496 |
case '\n': case ' ': break;
|
497 |
default:
|
498 |
fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);
|
499 |
goto SKIP_DATA;
|
500 |
}
|
501 |
}
|
502 |
|
503 |
/* Handle compiling via the POSIX interface, which doesn't support the
|
504 |
timing, showing, or debugging options, nor the ability to pass over
|
505 |
local character tables. */
|
506 |
|
507 |
#if !defined NOPOSIX
|
508 |
if (posix || do_posix)
|
509 |
{
|
510 |
int rc;
|
511 |
int cflags = 0;
|
512 |
if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
|
513 |
if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
|
514 |
rc = regcomp(&preg, (char *)p, cflags);
|
515 |
|
516 |
/* Compilation failed; go back for another re, skipping to blank line
|
517 |
if non-interactive. */
|
518 |
|
519 |
if (rc != 0)
|
520 |
{
|
521 |
(void)regerror(rc, &preg, (char *)buffer, sizeof(buffer));
|
522 |
fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
|
523 |
goto SKIP_DATA;
|
524 |
}
|
525 |
}
|
526 |
|
527 |
/* Handle compiling via the native interface */
|
528 |
|
529 |
else
|
530 |
#endif /* !defined NOPOSIX */
|
531 |
|
532 |
{
|
533 |
if (timeit)
|
534 |
{
|
535 |
register int i;
|
536 |
clock_t time_taken;
|
537 |
clock_t start_time = clock();
|
538 |
for (i = 0; i < LOOPREPEAT; i++)
|
539 |
{
|
540 |
re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
|
541 |
if (re != NULL) free(re);
|
542 |
}
|
543 |
time_taken = clock() - start_time;
|
544 |
fprintf(outfile, "Compile time %.3f milliseconds\n",
|
545 |
((double)time_taken * 1000.0) /
|
546 |
((double)LOOPREPEAT * (double)CLOCKS_PER_SEC));
|
547 |
}
|
548 |
|
549 |
re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
|
550 |
|
551 |
/* Compilation failed; go back for another re, skipping to blank line
|
552 |
if non-interactive. */
|
553 |
|
554 |
if (re == NULL)
|
555 |
{
|
556 |
fprintf(outfile, "Failed: %s at offset %d\n", error, erroroffset);
|
557 |
SKIP_DATA:
|
558 |
if (infile != stdin)
|
559 |
{
|
560 |
for (;;)
|
561 |
{
|
562 |
if (fgets((char *)buffer, sizeof(buffer), infile) == NULL)
|
563 |
{
|
564 |
done = 1;
|
565 |
goto CONTINUE;
|
566 |
}
|
567 |
len = (int)strlen((char *)buffer);
|
568 |
while (len > 0 && isspace(buffer[len-1])) len--;
|
569 |
if (len == 0) break;
|
570 |
}
|
571 |
fprintf(outfile, "\n");
|
572 |
}
|
573 |
goto CONTINUE;
|
574 |
}
|
575 |
|
576 |
/* Compilation succeeded; print data if required */
|
577 |
|
578 |
if (do_showinfo)
|
579 |
{
|
580 |
int first_char, count;
|
581 |
|
582 |
if (do_debug) print_internals(re);
|
583 |
|
584 |
count = pcre_info(re, &options, &first_char);
|
585 |
if (count < 0) fprintf(outfile,
|
586 |
"Error %d while reading info\n", count);
|
587 |
else
|
588 |
{
|
589 |
fprintf(outfile, "Identifying subpattern count = %d\n", count);
|
590 |
if (options == 0) fprintf(outfile, "No options\n");
|
591 |
else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s\n",
|
592 |
((options & PCRE_ANCHORED) != 0)? " anchored" : "",
|
593 |
((options & PCRE_CASELESS) != 0)? " caseless" : "",
|
594 |
((options & PCRE_EXTENDED) != 0)? " extended" : "",
|
595 |
((options & PCRE_MULTILINE) != 0)? " multiline" : "",
|
596 |
((options & PCRE_DOTALL) != 0)? " dotall" : "",
|
597 |
((options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
|
598 |
((options & PCRE_EXTRA) != 0)? " extra" : "",
|
599 |
((options & PCRE_UNGREEDY) != 0)? " ungreedy" : "");
|
600 |
|
601 |
if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)
|
602 |
fprintf(outfile, "Case state changes\n");
|
603 |
|
604 |
if (first_char == -1)
|
605 |
{
|
606 |
fprintf(outfile, "First char at start or follows \\n\n");
|
607 |
}
|
608 |
else if (first_char < 0)
|
609 |
{
|
610 |
fprintf(outfile, "No first char\n");
|
611 |
}
|
612 |
else
|
613 |
{
|
614 |
if (isprint(first_char))
|
615 |
fprintf(outfile, "First char = \'%c\'\n", first_char);
|
616 |
else
|
617 |
fprintf(outfile, "First char = %d\n", first_char);
|
618 |
}
|
619 |
|
620 |
if (((((real_pcre *)re)->options) & PCRE_REQCHSET) != 0)
|
621 |
{
|
622 |
int req_char = ((real_pcre *)re)->req_char;
|
623 |
if (isprint(req_char))
|
624 |
fprintf(outfile, "Req char = \'%c\'\n", req_char);
|
625 |
else
|
626 |
fprintf(outfile, "Req char = %d\n", req_char);
|
627 |
}
|
628 |
else fprintf(outfile, "No req char\n");
|
629 |
}
|
630 |
}
|
631 |
|
632 |
/* If /S was present, study the regexp to generate additional info to
|
633 |
help with the matching. */
|
634 |
|
635 |
if (do_study)
|
636 |
{
|
637 |
if (timeit)
|
638 |
{
|
639 |
register int i;
|
640 |
clock_t time_taken;
|
641 |
clock_t start_time = clock();
|
642 |
for (i = 0; i < LOOPREPEAT; i++)
|
643 |
extra = pcre_study(re, study_options, &error);
|
644 |
time_taken = clock() - start_time;
|
645 |
if (extra != NULL) free(extra);
|
646 |
fprintf(outfile, " Study time %.3f milliseconds\n",
|
647 |
((double)time_taken * 1000.0)/
|
648 |
((double)LOOPREPEAT * (double)CLOCKS_PER_SEC));
|
649 |
}
|
650 |
|
651 |
extra = pcre_study(re, study_options, &error);
|
652 |
if (error != NULL)
|
653 |
fprintf(outfile, "Failed to study: %s\n", error);
|
654 |
else if (extra == NULL)
|
655 |
fprintf(outfile, "Study returned NULL\n");
|
656 |
|
657 |
/* This looks at internal information. A bit kludgy to do it this
|
658 |
way, but it is useful for testing. */
|
659 |
|
660 |
else if (do_showinfo)
|
661 |
{
|
662 |
real_pcre_extra *xx = (real_pcre_extra *)extra;
|
663 |
if ((xx->options & PCRE_STUDY_MAPPED) == 0)
|
664 |
fprintf(outfile, "No starting character set\n");
|
665 |
else
|
666 |
{
|
667 |
int i;
|
668 |
int c = 24;
|
669 |
fprintf(outfile, "Starting character set: ");
|
670 |
for (i = 0; i < 256; i++)
|
671 |
{
|
672 |
if ((xx->start_bits[i/8] & (1<<(i%8))) != 0)
|
673 |
{
|
674 |
if (c > 75)
|
675 |
{
|
676 |
fprintf(outfile, "\n ");
|
677 |
c = 2;
|
678 |
}
|
679 |
if (isprint(i) && i != ' ')
|
680 |
{
|
681 |
fprintf(outfile, "%c ", i);
|
682 |
c += 2;
|
683 |
}
|
684 |
else
|
685 |
{
|
686 |
fprintf(outfile, "\\x%02x ", i);
|
687 |
c += 5;
|
688 |
}
|
689 |
}
|
690 |
}
|
691 |
fprintf(outfile, "\n");
|
692 |
}
|
693 |
}
|
694 |
}
|
695 |
}
|
696 |
|
697 |
/* Read data lines and test them */
|
698 |
|
699 |
for (;;)
|
700 |
{
|
701 |
unsigned char *q;
|
702 |
unsigned char *bptr = dbuffer;
|
703 |
int count, c;
|
704 |
int copystrings = 0;
|
705 |
int getstrings = 0;
|
706 |
int getlist = 0;
|
707 |
int gmatched = 0;
|
708 |
int start_offset = 0;
|
709 |
int g_notempty = 0;
|
710 |
int offsets[45];
|
711 |
int size_offsets = sizeof(offsets)/sizeof(int);
|
712 |
|
713 |
options = 0;
|
714 |
|
715 |
if (infile == stdin) printf("data> ");
|
716 |
if (fgets((char *)buffer, sizeof(buffer), infile) == NULL)
|
717 |
{
|
718 |
done = 1;
|
719 |
goto CONTINUE;
|
720 |
}
|
721 |
if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
|
722 |
|
723 |
len = (int)strlen((char *)buffer);
|
724 |
while (len > 0 && isspace(buffer[len-1])) len--;
|
725 |
buffer[len] = 0;
|
726 |
if (len == 0) break;
|
727 |
|
728 |
p = buffer;
|
729 |
while (isspace(*p)) p++;
|
730 |
|
731 |
q = dbuffer;
|
732 |
while ((c = *p++) != 0)
|
733 |
{
|
734 |
int i = 0;
|
735 |
int n = 0;
|
736 |
if (c == '\\') switch ((c = *p++))
|
737 |
{
|
738 |
case 'a': c = 7; break;
|
739 |
case 'b': c = '\b'; break;
|
740 |
case 'e': c = 27; break;
|
741 |
case 'f': c = '\f'; break;
|
742 |
case 'n': c = '\n'; break;
|
743 |
case 'r': c = '\r'; break;
|
744 |
case 't': c = '\t'; break;
|
745 |
case 'v': c = '\v'; break;
|
746 |
|
747 |
case '0': case '1': case '2': case '3':
|
748 |
case '4': case '5': case '6': case '7':
|
749 |
c -= '0';
|
750 |
while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
|
751 |
c = c * 8 + *p++ - '0';
|
752 |
break;
|
753 |
|
754 |
case 'x':
|
755 |
c = 0;
|
756 |
while (i++ < 2 && isxdigit(*p))
|
757 |
{
|
758 |
c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W');
|
759 |
p++;
|
760 |
}
|
761 |
break;
|
762 |
|
763 |
case 0: /* Allows for an empty line */
|
764 |
p--;
|
765 |
continue;
|
766 |
|
767 |
case 'A': /* Option setting */
|
768 |
options |= PCRE_ANCHORED;
|
769 |
continue;
|
770 |
|
771 |
case 'B':
|
772 |
options |= PCRE_NOTBOL;
|
773 |
continue;
|
774 |
|
775 |
case 'C':
|
776 |
while(isdigit(*p)) n = n * 10 + *p++ - '0';
|
777 |
copystrings |= 1 << n;
|
778 |
continue;
|
779 |
|
780 |
case 'G':
|
781 |
while(isdigit(*p)) n = n * 10 + *p++ - '0';
|
782 |
getstrings |= 1 << n;
|
783 |
continue;
|
784 |
|
785 |
case 'L':
|
786 |
getlist = 1;
|
787 |
continue;
|
788 |
|
789 |
case 'N':
|
790 |
options |= PCRE_NOTEMPTY;
|
791 |
continue;
|
792 |
|
793 |
case 'O':
|
794 |
while(isdigit(*p)) n = n * 10 + *p++ - '0';
|
795 |
if (n <= (int)(sizeof(offsets)/sizeof(int))) size_offsets = n;
|
796 |
continue;
|
797 |
|
798 |
case 'Z':
|
799 |
options |= PCRE_NOTEOL;
|
800 |
continue;
|
801 |
}
|
802 |
*q++ = c;
|
803 |
}
|
804 |
*q = 0;
|
805 |
len = q - dbuffer;
|
806 |
|
807 |
/* Handle matching via the POSIX interface, which does not
|
808 |
support timing. */
|
809 |
|
810 |
#if !defined NOPOSIX
|
811 |
if (posix || do_posix)
|
812 |
{
|
813 |
int rc;
|
814 |
int eflags = 0;
|
815 |
regmatch_t pmatch[sizeof(offsets)/sizeof(int)];
|
816 |
if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
|
817 |
if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
|
818 |
|
819 |
rc = regexec(&preg, (const char *)bptr, size_offsets, pmatch, eflags);
|
820 |
|
821 |
if (rc != 0)
|
822 |
{
|
823 |
(void)regerror(rc, &preg, (char *)buffer, sizeof(buffer));
|
824 |
fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
|
825 |
}
|
826 |
else
|
827 |
{
|
828 |
size_t i;
|
829 |
for (i = 0; i < size_offsets; i++)
|
830 |
{
|
831 |
if (pmatch[i].rm_so >= 0)
|
832 |
{
|
833 |
fprintf(outfile, "%2d: ", (int)i);
|
834 |
pchars(dbuffer + pmatch[i].rm_so,
|
835 |
pmatch[i].rm_eo - pmatch[i].rm_so);
|
836 |
fprintf(outfile, "\n");
|
837 |
if (i == 0 && do_showrest)
|
838 |
{
|
839 |
fprintf(outfile, " 0+ ");
|
840 |
pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo);
|
841 |
fprintf(outfile, "\n");
|
842 |
}
|
843 |
}
|
844 |
}
|
845 |
}
|
846 |
}
|
847 |
|
848 |
/* Handle matching via the native interface - repeats for /g and /G */
|
849 |
|
850 |
else
|
851 |
#endif /* !defined NOPOSIX */
|
852 |
|
853 |
for (;; gmatched++) /* Loop for /g or /G */
|
854 |
{
|
855 |
if (timeit)
|
856 |
{
|
857 |
register int i;
|
858 |
clock_t time_taken;
|
859 |
clock_t start_time = clock();
|
860 |
for (i = 0; i < LOOPREPEAT; i++)
|
861 |
count = pcre_exec(re, extra, (char *)bptr, len,
|
862 |
start_offset, options | g_notempty, offsets, size_offsets);
|
863 |
time_taken = clock() - start_time;
|
864 |
fprintf(outfile, "Execute time %.3f milliseconds\n",
|
865 |
((double)time_taken * 1000.0)/
|
866 |
((double)LOOPREPEAT * (double)CLOCKS_PER_SEC));
|
867 |
}
|
868 |
|
869 |
count = pcre_exec(re, extra, (char *)bptr, len,
|
870 |
start_offset, options | g_notempty, offsets, size_offsets);
|
871 |
|
872 |
if (count == 0)
|
873 |
{
|
874 |
fprintf(outfile, "Matched, but too many substrings\n");
|
875 |
count = size_offsets/3;
|
876 |
}
|
877 |
|
878 |
/* Matched */
|
879 |
|
880 |
if (count >= 0)
|
881 |
{
|
882 |
int i;
|
883 |
for (i = 0; i < count * 2; i += 2)
|
884 |
{
|
885 |
if (offsets[i] < 0)
|
886 |
fprintf(outfile, "%2d: <unset>\n", i/2);
|
887 |
else
|
888 |
{
|
889 |
fprintf(outfile, "%2d: ", i/2);
|
890 |
pchars(bptr + offsets[i], offsets[i+1] - offsets[i]);
|
891 |
fprintf(outfile, "\n");
|
892 |
if (i == 0)
|
893 |
{
|
894 |
if (do_showrest)
|
895 |
{
|
896 |
fprintf(outfile, " 0+ ");
|
897 |
pchars(bptr + offsets[i+1], len - offsets[i+1]);
|
898 |
fprintf(outfile, "\n");
|
899 |
}
|
900 |
}
|
901 |
}
|
902 |
}
|
903 |
|
904 |
for (i = 0; i < 32; i++)
|
905 |
{
|
906 |
if ((copystrings & (1 << i)) != 0)
|
907 |
{
|
908 |
char copybuffer[16];
|
909 |
int rc = pcre_copy_substring((char *)bptr, offsets, count,
|
910 |
i, copybuffer, sizeof(copybuffer));
|
911 |
if (rc < 0)
|
912 |
fprintf(outfile, "copy substring %d failed %d\n", i, rc);
|
913 |
else
|
914 |
fprintf(outfile, "%2dC %s (%d)\n", i, copybuffer, rc);
|
915 |
}
|
916 |
}
|
917 |
|
918 |
for (i = 0; i < 32; i++)
|
919 |
{
|
920 |
if ((getstrings & (1 << i)) != 0)
|
921 |
{
|
922 |
const char *substring;
|
923 |
int rc = pcre_get_substring((char *)bptr, offsets, count,
|
924 |
i, &substring);
|
925 |
if (rc < 0)
|
926 |
fprintf(outfile, "get substring %d failed %d\n", i, rc);
|
927 |
else
|
928 |
{
|
929 |
fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
|
930 |
free((void *)substring);
|
931 |
}
|
932 |
}
|
933 |
}
|
934 |
|
935 |
if (getlist)
|
936 |
{
|
937 |
const char **stringlist;
|
938 |
int rc = pcre_get_substring_list((char *)bptr, offsets, count,
|
939 |
&stringlist);
|
940 |
if (rc < 0)
|
941 |
fprintf(outfile, "get substring list failed %d\n", rc);
|
942 |
else
|
943 |
{
|
944 |
for (i = 0; i < count; i++)
|
945 |
fprintf(outfile, "%2dL %s\n", i, stringlist[i]);
|
946 |
if (stringlist[i] != NULL)
|
947 |
fprintf(outfile, "string list not terminated by NULL\n");
|
948 |
free((void *)stringlist);
|
949 |
}
|
950 |
}
|
951 |
}
|
952 |
|
953 |
/* Failed to match. If this is a /g or /G loop and we previously set
|
954 |
PCRE_NOTEMPTY after a null match, this is not necessarily the end.
|
955 |
We want to advance the start offset, and continue. Fudge the offset
|
956 |
values to achieve this. We won't be at the end of the string - that
|
957 |
was checked before setting PCRE_NOTEMPTY. */
|
958 |
|
959 |
else
|
960 |
{
|
961 |
if (g_notempty != 0)
|
962 |
{
|
963 |
offsets[0] = start_offset;
|
964 |
offsets[1] = start_offset + 1;
|
965 |
}
|
966 |
else
|
967 |
{
|
968 |
if (gmatched == 0) /* Error if no previous matches */
|
969 |
{
|
970 |
if (count == -1) fprintf(outfile, "No match\n");
|
971 |
else fprintf(outfile, "Error %d\n", count);
|
972 |
}
|
973 |
break; /* Out of the /g loop */
|
974 |
}
|
975 |
}
|
976 |
|
977 |
/* If not /g or /G we are done */
|
978 |
|
979 |
if (!do_g && !do_G) break;
|
980 |
|
981 |
/* If we have matched an empty string, first check to see if we are at
|
982 |
the end of the subject. If so, the /g loop is over. Otherwise, mimic
|
983 |
what Perl's /g options does. This turns out to be rather cunning. First
|
984 |
we set PCRE_NOTEMPTY and try the match again at the same point. If this
|
985 |
fails (picked up above) we advance to the next character. */
|
986 |
|
987 |
g_notempty = 0;
|
988 |
if (offsets[0] == offsets[1])
|
989 |
{
|
990 |
if (offsets[0] == len) break;
|
991 |
g_notempty = PCRE_NOTEMPTY;
|
992 |
}
|
993 |
|
994 |
/* For /g, update the start offset, leaving the rest alone */
|
995 |
|
996 |
if (do_g) start_offset = offsets[1];
|
997 |
|
998 |
/* For /G, update the pointer and length */
|
999 |
|
1000 |
else
|
1001 |
{
|
1002 |
bptr += offsets[1];
|
1003 |
len -= offsets[1];
|
1004 |
}
|
1005 |
} /* End of loop for /g and /G */
|
1006 |
} /* End of loop for data lines */
|
1007 |
|
1008 |
CONTINUE:
|
1009 |
|
1010 |
#if !defined NOPOSIX
|
1011 |
if (posix || do_posix) regfree(&preg);
|
1012 |
#endif
|
1013 |
|
1014 |
if (re != NULL) free(re);
|
1015 |
if (extra != NULL) free(extra);
|
1016 |
if (tables != NULL)
|
1017 |
{
|
1018 |
free((void *)tables);
|
1019 |
setlocale(LC_CTYPE, "C");
|
1020 |
}
|
1021 |
}
|
1022 |
|
1023 |
fprintf(outfile, "\n");
|
1024 |
return 0;
|
1025 |
}
|
1026 |
|
1027 |
/* End */
|