/[pcre]/code/trunk/pcretest.c
ViewVC logotype

Diff of /code/trunk/pcretest.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 411 by ph10, Fri Apr 10 15:40:21 2009 UTC revision 642 by ph10, Thu Jul 28 18:59:40 2011 UTC
# Line 79  input mode under Windows. */ Line 79  input mode under Windows. */
79  #define fileno _fileno  #define fileno _fileno
80  #endif  #endif
81    
82    /* A user sent this fix for Borland Builder 5 under Windows. */
83    
84    #ifdef __BORLANDC__
85    #define _setmode(handle, mode) setmode(handle, mode)
86    #endif
87    
88    /* Not Windows */
89    
90  #else  #else
91  #include <sys/time.h>          /* These two includes are needed */  #include <sys/time.h>          /* These two includes are needed */
92  #include <sys/resource.h>      /* for setrlimit(). */  #include <sys/resource.h>      /* for setrlimit(). */
# Line 118  external symbols to prevent clashes. */ Line 126  external symbols to prevent clashes. */
126    
127  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
128  patterns. This function is in a separate file so that it can be included in  patterns. This function is in a separate file so that it can be included in
129  pcre_compile.c when that module is compiled with debugging enabled.  pcre_compile.c when that module is compiled with debugging enabled. It needs to
130    know which case is being compiled. */
 The definition of the macro PRINTABLE, which determines whether to print an  
 output character as-is or as a hex value when showing compiled patterns, is  
 contained in this file. We uses it here also, in cases when the locale has not  
 been explicitly changed, so as to get consistent output from systems that  
 differ in their output from isprint() even in the "C" locale. */  
131    
132    #define COMPILING_PCRETEST
133  #include "pcre_printint.src"  #include "pcre_printint.src"
134    
135  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  /* The definition of the macro PRINTABLE, which determines whether to print an
136    output character as-is or as a hex value when showing compiled patterns, is
137    contained in the printint.src file. We uses it here also, in cases when the
138    locale has not been explicitly changed, so as to get consistent output from
139    systems that differ in their output from isprint() even in the "C" locale. */
140    
141    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
142    
143  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
144  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 188  static uschar *buffer = NULL; Line 197  static uschar *buffer = NULL;
197  static uschar *dbuffer = NULL;  static uschar *dbuffer = NULL;
198  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
199    
200    /* Textual explanations for runtime error codes */
201    
202    static const char *errtexts[] = {
203      NULL,  /* 0 is no error */
204      NULL,  /* NOMATCH is handled specially */
205      "NULL argument passed",
206      "bad option value",
207      "magic number missing",
208      "unknown opcode - pattern overwritten?",
209      "no more memory",
210      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
211      "match limit exceeded",
212      "callout error code",
213      NULL,  /* BADUTF8 is handled specially */
214      "bad UTF-8 offset",
215      NULL,  /* PARTIAL is handled specially */
216      "not used - internal error",
217      "internal error - pattern overwritten?",
218      "bad count value",
219      "item unsupported for DFA matching",
220      "backreference condition or recursion test not supported for DFA matching",
221      "match limit not supported for DFA matching",
222      "workspace size exceeded in DFA matching",
223      "too much recursion for DFA matching",
224      "recursion limit exceeded",
225      "not used - internal error",
226      "invalid combination of newline options",
227      "bad offset value",
228      NULL,  /* SHORTUTF8 is handled specially */
229      "nested recursion at the same subject position"
230    };
231    
232    
233    /*************************************************
234    *         Alternate character tables             *
235    *************************************************/
236    
237    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
238    using the default tables of the library. However, the T option can be used to
239    select alternate sets of tables, for different kinds of testing. Note also that
240    the L (locale) option also adjusts the tables. */
241    
242    /* This is the set of tables distributed as default with PCRE. It recognizes
243    only ASCII characters. */
244    
245    static const unsigned char tables0[] = {
246    
247    /* This table is a lower casing table. */
248    
249        0,  1,  2,  3,  4,  5,  6,  7,
250        8,  9, 10, 11, 12, 13, 14, 15,
251       16, 17, 18, 19, 20, 21, 22, 23,
252       24, 25, 26, 27, 28, 29, 30, 31,
253       32, 33, 34, 35, 36, 37, 38, 39,
254       40, 41, 42, 43, 44, 45, 46, 47,
255       48, 49, 50, 51, 52, 53, 54, 55,
256       56, 57, 58, 59, 60, 61, 62, 63,
257       64, 97, 98, 99,100,101,102,103,
258      104,105,106,107,108,109,110,111,
259      112,113,114,115,116,117,118,119,
260      120,121,122, 91, 92, 93, 94, 95,
261       96, 97, 98, 99,100,101,102,103,
262      104,105,106,107,108,109,110,111,
263      112,113,114,115,116,117,118,119,
264      120,121,122,123,124,125,126,127,
265      128,129,130,131,132,133,134,135,
266      136,137,138,139,140,141,142,143,
267      144,145,146,147,148,149,150,151,
268      152,153,154,155,156,157,158,159,
269      160,161,162,163,164,165,166,167,
270      168,169,170,171,172,173,174,175,
271      176,177,178,179,180,181,182,183,
272      184,185,186,187,188,189,190,191,
273      192,193,194,195,196,197,198,199,
274      200,201,202,203,204,205,206,207,
275      208,209,210,211,212,213,214,215,
276      216,217,218,219,220,221,222,223,
277      224,225,226,227,228,229,230,231,
278      232,233,234,235,236,237,238,239,
279      240,241,242,243,244,245,246,247,
280      248,249,250,251,252,253,254,255,
281    
282    /* This table is a case flipping table. */
283    
284        0,  1,  2,  3,  4,  5,  6,  7,
285        8,  9, 10, 11, 12, 13, 14, 15,
286       16, 17, 18, 19, 20, 21, 22, 23,
287       24, 25, 26, 27, 28, 29, 30, 31,
288       32, 33, 34, 35, 36, 37, 38, 39,
289       40, 41, 42, 43, 44, 45, 46, 47,
290       48, 49, 50, 51, 52, 53, 54, 55,
291       56, 57, 58, 59, 60, 61, 62, 63,
292       64, 97, 98, 99,100,101,102,103,
293      104,105,106,107,108,109,110,111,
294      112,113,114,115,116,117,118,119,
295      120,121,122, 91, 92, 93, 94, 95,
296       96, 65, 66, 67, 68, 69, 70, 71,
297       72, 73, 74, 75, 76, 77, 78, 79,
298       80, 81, 82, 83, 84, 85, 86, 87,
299       88, 89, 90,123,124,125,126,127,
300      128,129,130,131,132,133,134,135,
301      136,137,138,139,140,141,142,143,
302      144,145,146,147,148,149,150,151,
303      152,153,154,155,156,157,158,159,
304      160,161,162,163,164,165,166,167,
305      168,169,170,171,172,173,174,175,
306      176,177,178,179,180,181,182,183,
307      184,185,186,187,188,189,190,191,
308      192,193,194,195,196,197,198,199,
309      200,201,202,203,204,205,206,207,
310      208,209,210,211,212,213,214,215,
311      216,217,218,219,220,221,222,223,
312      224,225,226,227,228,229,230,231,
313      232,233,234,235,236,237,238,239,
314      240,241,242,243,244,245,246,247,
315      248,249,250,251,252,253,254,255,
316    
317    /* This table contains bit maps for various character classes. Each map is 32
318    bytes long and the bits run from the least significant end of each byte. The
319    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
320    graph, print, punct, and cntrl. Other classes are built from combinations. */
321    
322      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
323      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
324      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
325      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
326    
327      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
328      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
329      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
330      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
331    
332      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
333      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
334      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
335      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
336    
337      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
338      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
339      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
340      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
341    
342      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
343      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
344      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
345      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
346    
347      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
348      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
349      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
350      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
351    
352      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
353      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
354      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
355      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
356    
357      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
358      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
359      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
360      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
361    
362      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
363      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
364      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
365      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
366    
367      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
368      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
369      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
370      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
371    
372    /* This table identifies various classes of character by individual bits:
373      0x01   white space character
374      0x02   letter
375      0x04   decimal digit
376      0x08   hexadecimal digit
377      0x10   alphanumeric or '_'
378      0x80   regular expression metacharacter or binary zero
379    */
380    
381      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
382      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
383      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
384      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
385      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
386      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
387      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
388      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
389      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
390      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
391      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
392      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
393      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
394      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
395      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
396      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
397      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
398      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
399      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
400      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
401      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
412      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
413    
414    /* This is a set of tables that came orginally from a Windows user. It seems to
415    be at least an approximation of ISO 8859. In particular, there are characters
416    greater than 128 that are marked as spaces, letters, etc. */
417    
418    static const unsigned char tables1[] = {
419    0,1,2,3,4,5,6,7,
420    8,9,10,11,12,13,14,15,
421    16,17,18,19,20,21,22,23,
422    24,25,26,27,28,29,30,31,
423    32,33,34,35,36,37,38,39,
424    40,41,42,43,44,45,46,47,
425    48,49,50,51,52,53,54,55,
426    56,57,58,59,60,61,62,63,
427    64,97,98,99,100,101,102,103,
428    104,105,106,107,108,109,110,111,
429    112,113,114,115,116,117,118,119,
430    120,121,122,91,92,93,94,95,
431    96,97,98,99,100,101,102,103,
432    104,105,106,107,108,109,110,111,
433    112,113,114,115,116,117,118,119,
434    120,121,122,123,124,125,126,127,
435    128,129,130,131,132,133,134,135,
436    136,137,138,139,140,141,142,143,
437    144,145,146,147,148,149,150,151,
438    152,153,154,155,156,157,158,159,
439    160,161,162,163,164,165,166,167,
440    168,169,170,171,172,173,174,175,
441    176,177,178,179,180,181,182,183,
442    184,185,186,187,188,189,190,191,
443    224,225,226,227,228,229,230,231,
444    232,233,234,235,236,237,238,239,
445    240,241,242,243,244,245,246,215,
446    248,249,250,251,252,253,254,223,
447    224,225,226,227,228,229,230,231,
448    232,233,234,235,236,237,238,239,
449    240,241,242,243,244,245,246,247,
450    248,249,250,251,252,253,254,255,
451    0,1,2,3,4,5,6,7,
452    8,9,10,11,12,13,14,15,
453    16,17,18,19,20,21,22,23,
454    24,25,26,27,28,29,30,31,
455    32,33,34,35,36,37,38,39,
456    40,41,42,43,44,45,46,47,
457    48,49,50,51,52,53,54,55,
458    56,57,58,59,60,61,62,63,
459    64,97,98,99,100,101,102,103,
460    104,105,106,107,108,109,110,111,
461    112,113,114,115,116,117,118,119,
462    120,121,122,91,92,93,94,95,
463    96,65,66,67,68,69,70,71,
464    72,73,74,75,76,77,78,79,
465    80,81,82,83,84,85,86,87,
466    88,89,90,123,124,125,126,127,
467    128,129,130,131,132,133,134,135,
468    136,137,138,139,140,141,142,143,
469    144,145,146,147,148,149,150,151,
470    152,153,154,155,156,157,158,159,
471    160,161,162,163,164,165,166,167,
472    168,169,170,171,172,173,174,175,
473    176,177,178,179,180,181,182,183,
474    184,185,186,187,188,189,190,191,
475    224,225,226,227,228,229,230,231,
476    232,233,234,235,236,237,238,239,
477    240,241,242,243,244,245,246,215,
478    248,249,250,251,252,253,254,223,
479    192,193,194,195,196,197,198,199,
480    200,201,202,203,204,205,206,207,
481    208,209,210,211,212,213,214,247,
482    216,217,218,219,220,221,222,255,
483    0,62,0,0,1,0,0,0,
484    0,0,0,0,0,0,0,0,
485    32,0,0,0,1,0,0,0,
486    0,0,0,0,0,0,0,0,
487    0,0,0,0,0,0,255,3,
488    126,0,0,0,126,0,0,0,
489    0,0,0,0,0,0,0,0,
490    0,0,0,0,0,0,0,0,
491    0,0,0,0,0,0,255,3,
492    0,0,0,0,0,0,0,0,
493    0,0,0,0,0,0,12,2,
494    0,0,0,0,0,0,0,0,
495    0,0,0,0,0,0,0,0,
496    254,255,255,7,0,0,0,0,
497    0,0,0,0,0,0,0,0,
498    255,255,127,127,0,0,0,0,
499    0,0,0,0,0,0,0,0,
500    0,0,0,0,254,255,255,7,
501    0,0,0,0,0,4,32,4,
502    0,0,0,128,255,255,127,255,
503    0,0,0,0,0,0,255,3,
504    254,255,255,135,254,255,255,7,
505    0,0,0,0,0,4,44,6,
506    255,255,127,255,255,255,127,255,
507    0,0,0,0,254,255,255,255,
508    255,255,255,255,255,255,255,127,
509    0,0,0,0,254,255,255,255,
510    255,255,255,255,255,255,255,255,
511    0,2,0,0,255,255,255,255,
512    255,255,255,255,255,255,255,127,
513    0,0,0,0,255,255,255,255,
514    255,255,255,255,255,255,255,255,
515    0,0,0,0,254,255,0,252,
516    1,0,0,248,1,0,0,120,
517    0,0,0,0,254,255,255,255,
518    0,0,128,0,0,0,128,0,
519    255,255,255,255,0,0,0,0,
520    0,0,0,0,0,0,0,128,
521    255,255,255,255,0,0,0,0,
522    0,0,0,0,0,0,0,0,
523    128,0,0,0,0,0,0,0,
524    0,1,1,0,1,1,0,0,
525    0,0,0,0,0,0,0,0,
526    0,0,0,0,0,0,0,0,
527    1,0,0,0,128,0,0,0,
528    128,128,128,128,0,0,128,0,
529    28,28,28,28,28,28,28,28,
530    28,28,0,0,0,0,0,128,
531    0,26,26,26,26,26,26,18,
532    18,18,18,18,18,18,18,18,
533    18,18,18,18,18,18,18,18,
534    18,18,18,128,128,0,128,16,
535    0,26,26,26,26,26,26,18,
536    18,18,18,18,18,18,18,18,
537    18,18,18,18,18,18,18,18,
538    18,18,18,128,128,0,0,0,
539    0,0,0,0,0,1,0,0,
540    0,0,0,0,0,0,0,0,
541    0,0,0,0,0,0,0,0,
542    0,0,0,0,0,0,0,0,
543    1,0,0,0,0,0,0,0,
544    0,0,18,0,0,0,0,0,
545    0,0,20,20,0,18,0,0,
546    0,20,18,0,0,0,0,0,
547    18,18,18,18,18,18,18,18,
548    18,18,18,18,18,18,18,18,
549    18,18,18,18,18,18,18,0,
550    18,18,18,18,18,18,18,18,
551    18,18,18,18,18,18,18,18,
552    18,18,18,18,18,18,18,18,
553    18,18,18,18,18,18,18,0,
554    18,18,18,18,18,18,18,18
555    };
556    
557    
558    
559    
560    #ifndef HAVE_STRERROR
561    /*************************************************
562    *     Provide strerror() for non-ANSI libraries  *
563    *************************************************/
564    
565    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
566    in their libraries, but can provide the same facility by this simple
567    alternative function. */
568    
569    extern int   sys_nerr;
570    extern char *sys_errlist[];
571    
572    char *
573    strerror(int n)
574    {
575    if (n < 0 || n >= sys_nerr) return "unknown error number";
576    return sys_errlist[n];
577    }
578    #endif /* HAVE_STRERROR */
579    
580    
581    
582    
583  /*************************************************  /*************************************************
# Line 220  uschar *here = start; Line 610  uschar *here = start;
610    
611  for (;;)  for (;;)
612    {    {
613    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
614    
615    if (rlen > 1000)    if (rlen > 1000)
616      {      {
# Line 250  for (;;) Line 640  for (;;)
640      /* Read the next line by normal means, prompting if the file is stdin. */      /* Read the next line by normal means, prompting if the file is stdin. */
641    
642        {        {
643        if (f == stdin) printf(prompt);        if (f == stdin) printf("%s", prompt);
644        if (fgets((char *)here, rlen,  f) == NULL)        if (fgets((char *)here, rlen,  f) == NULL)
645          return (here == start)? NULL : start;          return (here == start)? NULL : start;
646        }        }
# Line 740  return 0; Line 1130  return 0;
1130  *         Check newline indicator                *  *         Check newline indicator                *
1131  *************************************************/  *************************************************/
1132    
1133  /* 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. Print
1134  xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is  a message and return 0 if there is no match.
 no match.  
1135    
1136  Arguments:  Arguments:
1137    p           points after the leading '<'    p           points after the leading '<'
# Line 798  printf("  -p       use POSIX interface\n Line 1187  printf("  -p       use POSIX interface\n
1187  #endif  #endif
1188  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
1189  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
1190  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied\n"
1191         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
1192  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1193  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 826  int timeit = 0; Line 1215  int timeit = 0;
1215  int timeitm = 0;  int timeitm = 0;
1216  int showinfo = 0;  int showinfo = 0;
1217  int showstore = 0;  int showstore = 0;
1218    int force_study = 0;
1219  int quiet = 0;  int quiet = 0;
1220  int size_offsets = 45;  int size_offsets = 45;
1221  int size_offsets_max;  int size_offsets_max;
# Line 874  while (argc > 1 && argv[op][0] == '-') Line 1264  while (argc > 1 && argv[op][0] == '-')
1264    {    {
1265    unsigned char *endptr;    unsigned char *endptr;
1266    
1267    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1268      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 1;
1269    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1270    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1271    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
# Line 909  while (argc > 1 && argv[op][0] == '-') Line 1299  while (argc > 1 && argv[op][0] == '-')
1299        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1300          *endptr == 0))          *endptr == 0))
1301      {      {
1302  #if defined(_WIN32) || defined(WIN32)  #if defined(_WIN32) || defined(WIN32) || defined(__minix)
1303      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
1304      exit(1);      exit(1);
1305  #else  #else
# Line 1039  while (!done) Line 1429  while (!done)
1429  #endif  #endif
1430    
1431    const char *error;    const char *error;
1432      unsigned char *markptr;
1433    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1434    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1435    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1436    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1437    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1438      int do_allcaps = 0;
1439      int do_mark = 0;
1440    int do_study = 0;    int do_study = 0;
1441      int no_force_study = 0;
1442    int do_debug = debug;    int do_debug = debug;
1443    int do_G = 0;    int do_G = 0;
1444    int do_g = 0;    int do_g = 0;
1445    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1446    int do_showrest = 0;    int do_showrest = 0;
1447      int do_showcaprest = 0;
1448    int do_flip = 0;    int do_flip = 0;
1449    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1450    
# Line 1111  while (!done) Line 1506  while (!done)
1506          }          }
1507        }        }
1508    
1509      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1510        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1511    
1512      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1119  while (!done) Line 1514  while (!done)
1514      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1515      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1516    
1517      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1518    
1519      if (true_study_size != 0)      if (true_study_size != 0)
1520        {        {
# Line 1161  while (!done) Line 1556  while (!done)
1556      }      }
1557    
1558    pp = p;    pp = p;
1559    poffset = p - buffer;    poffset = (int)(p - buffer);
1560    
1561    for(;;)    for(;;)
1562      {      {
# Line 1215  while (!done) Line 1610  while (!done)
1610        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1611        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1612    
1613        case '+': do_showrest = 1; break;        case '+':
1614          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1615          break;
1616    
1617          case '=': do_allcaps = 1; break;
1618        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1619        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
1620        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1225  while (!done) Line 1624  while (!done)
1624        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1625        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1626        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1627          case 'K': do_mark = 1; break;
1628        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1629        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1630    
# Line 1232  while (!done) Line 1632  while (!done)
1632        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1633  #endif  #endif
1634    
1635        case 'S': do_study = 1; break;        case 'S':
1636          if (do_study == 0) do_study = 1; else
1637            {
1638            do_study = 0;
1639            no_force_study = 1;
1640            }
1641          break;
1642    
1643        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1644          case 'W': options |= PCRE_UCP; break;
1645        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1646          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1647        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1648        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1649        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1650    
1651          case 'T':
1652          switch (*pp++)
1653            {
1654            case '0': tables = tables0; break;
1655            case '1': tables = tables1; break;
1656    
1657            case '\r':
1658            case '\n':
1659            case ' ':
1660            case 0:
1661            fprintf(outfile, "** Missing table number after /T\n");
1662            goto SKIP_DATA;
1663    
1664            default:
1665            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1666            goto SKIP_DATA;
1667            }
1668          break;
1669    
1670        case 'L':        case 'L':
1671        ppp = pp;        ppp = pp;
1672        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1264  while (!done) Line 1692  while (!done)
1692    
1693        case '<':        case '<':
1694          {          {
1695          if (strncmp((char *)pp, "JS>", 3) == 0)          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1696            {            {
1697            options |= PCRE_JAVASCRIPT_COMPAT;            options |= PCRE_JAVASCRIPT_COMPAT;
1698            pp += 3;            pp += 3;
# Line 1305  while (!done) Line 1733  while (!done)
1733      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1734      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1735      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1736        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1737        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1738    
1739      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1740    
# Line 1325  while (!done) Line 1755  while (!done)
1755  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1756    
1757      {      {
1758        unsigned long int get_options;
1759    
1760      if (timeit > 0)      if (timeit > 0)
1761        {        {
1762        register int i;        register int i;
# Line 1368  while (!done) Line 1800  while (!done)
1800        goto CONTINUE;        goto CONTINUE;
1801        }        }
1802    
1803      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1804      info-returning functions. The old one has a limited interface and      within the regex; check for this so that we know how to process the data
1805      returns only limited data. Check that it agrees with the newer one. */      lines. */
1806    
1807        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1808        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1809    
1810        /* Print information if required. There are now two info-returning
1811        functions. The old one has a limited interface and returns only limited
1812        data. Check that it agrees with the newer one. */
1813    
1814      if (log_store)      if (log_store)
1815        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1384  while (!done) Line 1823  while (!done)
1823      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1824      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1825    
1826      /* If /S was present, study the regexp to generate additional info to      /* If -s or /S was present, study the regex to generate additional info to
1827      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1828        suppresses the effect of /S (used for a few test patterns where studying is
1829        never sensible). */
1830    
1831      if (do_study)      if (do_study || (force_study && !no_force_study))
1832        {        {
1833        if (timeit > 0)        if (timeit > 0)
1834          {          {
# Line 1409  while (!done) Line 1850  while (!done)
1850          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1851        }        }
1852    
1853        /* If /K was present, we set up for handling MARK data. */
1854    
1855        if (do_mark)
1856          {
1857          if (extra == NULL)
1858            {
1859            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1860            extra->flags = 0;
1861            }
1862          extra->mark = &markptr;
1863          extra->flags |= PCRE_EXTRA_MARK;
1864          }
1865    
1866      /* If the 'F' option was present, we flip the bytes of all the integer      /* If the 'F' option was present, we flip the bytes of all the integer
1867      fields in the regex data block and the study block. This is to make it      fields in the regex data block and the study block. This is to make it
1868      possible to test PCRE's handling of byte-flipped patterns, e.g. those      possible to test PCRE's handling of byte-flipped patterns, e.g. those
# Line 1441  while (!done) Line 1895  while (!done)
1895          {          {
1896          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1897          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1898          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1899            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1900          }          }
1901        }        }
1902    
# Line 1455  while (!done) Line 1910  while (!done)
1910        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1911        }        }
1912    
1913        /* We already have the options in get_options (see above) */
1914    
1915      if (do_showinfo)      if (do_showinfo)
1916        {        {
1917        unsigned long int get_options, all_options;        unsigned long int all_options;
1918  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1919        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1920  #endif  #endif
# Line 1466  while (!done) Line 1923  while (!done)
1923        int nameentrysize, namecount;        int nameentrysize, namecount;
1924        const uschar *nametable;        const uschar *nametable;
1925    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1926        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1927        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1928        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1526  while (!done) Line 1982  while (!done)
1982        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
1983    
1984        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1985          else fprintf(outfile, "Options:%s%s%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%s%s\n",
1986            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1987            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1988            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
# Line 1540  while (!done) Line 1996  while (!done)
1996            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1997            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1998            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1999              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2000            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2001              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2002            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2003    
2004        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
# Line 1608  while (!done) Line 2066  while (!done)
2066        /* 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
2067        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2068        so messes up the test suite. (And with the /F option, it might be        so messes up the test suite. (And with the /F option, it might be
2069        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2070          information unless -i or -d was also present. This means that, except
2071          when auto-callouts are involved, the output from runs with and without
2072          -s should be identical. */
2073    
2074        if (do_study)        if (do_study || (force_study && showinfo && !no_force_study))
2075          {          {
2076          if (extra == NULL)          if (extra == NULL)
2077            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2078          else          else
2079            {            {
2080            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2081            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2082    
2083              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2084              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2085    
2086              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2087            if (start_bits == NULL)            if (start_bits == NULL)
2088              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2089            else            else
2090              {              {
2091              int i;              int i;
# Line 1684  while (!done) Line 2149  while (!done)
2149            }            }
2150          else          else
2151            {            {
2152            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2153    
2154              /* If there is study data, write it, but verify the writing only
2155              if the studying was requested by /S, not just by -s. */
2156    
2157            if (extra != NULL)            if (extra != NULL)
2158              {              {
2159              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1694  while (!done) Line 2163  while (!done)
2163                  strerror(errno));                  strerror(errno));
2164                }                }
2165              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2166              }              }
2167            }            }
2168          fclose(f);          fclose(f);
# Line 1702  while (!done) Line 2170  while (!done)
2170    
2171        new_free(re);        new_free(re);
2172        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2173        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2174            {
2175            new_free((void *)tables);
2176            setlocale(LC_CTYPE, "C");
2177            locale_set = 0;
2178            }
2179        continue;  /* With next regex */        continue;  /* With next regex */
2180        }        }
2181      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1724  while (!done) Line 2197  while (!done)
2197      int getlist = 0;      int getlist = 0;
2198      int gmatched = 0;      int gmatched = 0;
2199      int start_offset = 0;      int start_offset = 0;
2200        int start_offset_sign = 1;
2201      int g_notempty = 0;      int g_notempty = 0;
2202      int use_dfa = 0;      int use_dfa = 0;
2203    
# Line 1751  while (!done) Line 2225  while (!done)
2225        {        {
2226        if (extend_inputline(infile, buffer + len, "data> ") == NULL)        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
2227          {          {
2228          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2229              {
2230              fprintf(outfile, "\n");
2231              break;
2232              }
2233          done = 1;          done = 1;
2234          goto CONTINUE;          goto CONTINUE;
2235          }          }
# Line 1852  while (!done) Line 2330  while (!done)
2330          continue;          continue;
2331    
2332          case '>':          case '>':
2333            if (*p == '-')
2334              {
2335              start_offset_sign = -1;
2336              p++;
2337              }
2338          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2339            start_offset *= start_offset_sign;
2340          continue;          continue;
2341    
2342          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1925  while (!done) Line 2409  while (!done)
2409  #endif  #endif
2410            use_dfa = 1;            use_dfa = 1;
2411          continue;          continue;
2412    #endif
2413    
2414    #if !defined NODFA
2415          case 'F':          case 'F':
2416          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2417          continue;          continue;
# Line 1959  while (!done) Line 2445  while (!done)
2445          continue;          continue;
2446    
2447          case 'N':          case 'N':
2448          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2449              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2450            else
2451              options |= PCRE_NOTEMPTY;
2452          continue;          continue;
2453    
2454          case 'O':          case 'O':
# Line 1982  while (!done) Line 2471  while (!done)
2471          continue;          continue;
2472    
2473          case 'P':          case 'P':
2474          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2475              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2476          continue;          continue;
2477    
2478          case 'Q':          case 'Q':
# Line 2041  while (!done) Line 2531  while (!done)
2531        *q++ = c;        *q++ = c;
2532        }        }
2533      *q = 0;      *q = 0;
2534      len = q - dbuffer;      len = (int)(q - dbuffer);
2535    
2536      /* Move the data to the end of the buffer so that a read over the end of      /* Move the data to the end of the buffer so that a read over the end of
2537      the buffer will be seen by valgrind, even if it doesn't cause a crash. If      the buffer will be seen by valgrind, even if it doesn't cause a crash. If
# Line 2104  while (!done) Line 2594  while (!done)
2594              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2595                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2596              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2597              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2598                {                {
2599                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2600                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2601                  outfile);                  outfile);
2602                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 2124  while (!done) Line 2614  while (!done)
2614    
2615      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2616        {        {
2617          markptr = NULL;
2618    
2619        if (timeitm > 0)        if (timeitm > 0)
2620          {          {
2621          register int i;          register int i;
# Line 2135  while (!done) Line 2627  while (!done)
2627            {            {
2628            int workspace[1000];            int workspace[1000];
2629            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2630              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2631                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2632                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2633            }            }
# Line 2198  while (!done) Line 2690  while (!done)
2690        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2691          {          {
2692          int workspace[1000];          int workspace[1000];
2693          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2694            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2695            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2696          if (count == 0)          if (count == 0)
# Line 2245  while (!done) Line 2737  while (!done)
2737              do_g = do_G = FALSE;        /* Break g/G loop */              do_g = do_G = FALSE;        /* Break g/G loop */
2738              }              }
2739            }            }
2740    
2741            /* do_allcaps requests showing of all captures in the pattern, to check
2742            unset ones at the end. */
2743    
2744            if (do_allcaps)
2745              {
2746              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2747              count++;   /* Allow for full match */
2748              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2749              }
2750    
2751            /* Output the captured substrings */
2752    
2753          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2754            {            {
2755            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2756                {
2757                if (use_offsets[i] != -1)
2758                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2759                    use_offsets[i], i);
2760                if (use_offsets[i+1] != -1)
2761                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2762                    use_offsets[i+1], i+1);
2763              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2764                }
2765            else            else
2766              {              {
2767              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2768              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2769                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2770              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2771              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2772                {                {
2773                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2774                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2775                  fprintf(outfile, " 0+ ");                  outfile);
2776                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2777                }                }
2778              }              }
2779            }            }
2780    
2781            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2782    
2783          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2784            {            {
2785            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2352  while (!done) Line 2863  while (!done)
2863    
2864        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2865          {          {
2866          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2867  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2868          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2869            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2870              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2871  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2872                outfile);
2873              }
2874          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2875          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2876          }          }
# Line 2367  while (!done) Line 2880  while (!done)
2880        to advance the start offset, and continue. We won't be at the end of the        to advance the start offset, and continue. We won't be at the end of the
2881        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2882    
2883        Complication arises in the case when the newline option is "any" or        Complication arises in the case when the newline convention is "any",
2884        "anycrlf". If the previous match was at the end of a line terminated by        "crlf", or "anycrlf". If the previous match was at the end of a line
2885        CRLF, an advance of one character just passes the \r, whereas we should        terminated by CRLF, an advance of one character just passes the \r,
2886        prefer the longer newline sequence, as does the code in pcre_exec().        whereas we should prefer the longer newline sequence, as does the code in
2887        Fudge the offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2888          newline setting in the pattern; if none was set, use pcre_config() to
2889          find the default.
2890    
2891        Otherwise, in the case of UTF-8 matching, the advance must be one        Otherwise, in the case of UTF-8 matching, the advance must be one
2892        character, not one byte. */        character, not one byte. */
# Line 2396  while (!done) Line 2911  while (!done)
2911                      (d == -1)? PCRE_NEWLINE_ANY : 0;                      (d == -1)? PCRE_NEWLINE_ANY : 0;
2912              }              }
2913            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2914                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2915                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2916                &&                &&
2917                start_offset < len - 1 &&                start_offset < len - 1 &&
# Line 2406  while (!done) Line 2922  while (!done)
2922              {              {
2923              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2924                {                {
2925                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2926                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2927                }                }
2928              }              }
2929            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
2930            }            }
2931          else          else
2932            {            {
2933            if (count == PCRE_ERROR_NOMATCH)            switch(count)
2934              {              {
2935              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
2936                if (gmatched == 0)
2937                  {
2938                  if (markptr == NULL) fprintf(outfile, "No match\n");
2939                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2940                  }
2941                break;
2942    
2943                case PCRE_ERROR_BADUTF8:
2944                case PCRE_ERROR_SHORTUTF8:
2945                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
2946                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
2947                if (use_size_offsets >= 2)
2948                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
2949                    use_offsets[1]);
2950                fprintf(outfile, "\n");
2951                break;
2952    
2953                default:
2954                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
2955                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
2956                else
2957                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
2958                break;
2959              }              }
2960            else fprintf(outfile, "Error %d\n", count);  
2961            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
2962            }            }
2963          }          }
# Line 2430  while (!done) Line 2967  while (!done)
2967        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2968    
2969        /* If we have matched an empty string, first check to see if we are at        /* If we have matched an empty string, first check to see if we are at
2970        the end of the subject. If so, the /g loop is over. Otherwise, mimic        the end of the subject. If so, the /g loop is over. Otherwise, mimic what
2971        what Perl's /g options does. This turns out to be rather cunning. First        Perl's /g options does. This turns out to be rather cunning. First we set
2972        we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the        PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED and try the match again at the
2973        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
2974        character. */        character. */
2975    
# Line 2441  while (!done) Line 2978  while (!done)
2978        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2979          {          {
2980          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2981          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2982          }          }
2983    
2984        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2468  while (!done) Line 3005  while (!done)
3005    
3006    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
3007    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
3008    if (tables != NULL)    if (locale_set)
3009      {      {
3010      new_free((void *)tables);      new_free((void *)tables);
3011      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");

Legend:
Removed from v.411  
changed lines
  Added in v.642

  ViewVC Help
Powered by ViewVC 1.1.5