/[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 376 by ph10, Sun Mar 1 12:00:59 2009 UTC revision 780 by ph10, Fri Dec 2 11:39:21 2011 UTC
# Line 71  input mode under Windows. */ Line 71  input mode under Windows. */
71  #define INPUT_MODE   "r"  #define INPUT_MODE   "r"
72  #define OUTPUT_MODE  "wb"  #define OUTPUT_MODE  "wb"
73    
74  #define isatty _isatty         /* This is what Windows calls them, I'm told */  #ifndef isatty
75    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
76    #endif                         /* though in some environments they seem to   */
77                                   /* be already defined, hence the #ifndefs.    */
78    #ifndef fileno
79  #define fileno _fileno  #define fileno _fileno
80    #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 */
# Line 99  to keep two copies, we include the sourc Line 112  to keep two copies, we include the sourc
112  external symbols to prevent clashes. */  external symbols to prevent clashes. */
113    
114  #define _pcre_ucp_gentype      ucp_gentype  #define _pcre_ucp_gentype      ucp_gentype
115    #define _pcre_ucp_typerange    ucp_typerange
116  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
117  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
118  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
119  #define _pcre_utf8_table3      utf8_table3  #define _pcre_utf8_table3      utf8_table3
120  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
121    #define _pcre_utf8_char_sizes  utf8_char_sizes
122  #define _pcre_utt              utt  #define _pcre_utt              utt
123  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
124  #define _pcre_utt_names        utt_names  #define _pcre_utt_names        utt_names
# Line 113  external symbols to prevent clashes. */ Line 128  external symbols to prevent clashes. */
128    
129  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
130  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
131  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
132    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. */  
133    
134    #define COMPILING_PCRETEST
135  #include "pcre_printint.src"  #include "pcre_printint.src"
136    
137  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  /* The definition of the macro PRINTABLE, which determines whether to print an
138    output character as-is or as a hex value when showing compiled patterns, is
139    contained in the printint.src file. We uses it here also, in cases when the
140    locale has not been explicitly changed, so as to get consistent output from
141    systems that differ in their output from isprint() even in the "C" locale. */
142    
143    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
144    
145  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
146  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 175  static int locale_set = 0; Line 191  static int locale_set = 0;
191  static int show_malloc;  static int show_malloc;
192  static int use_utf8;  static int use_utf8;
193  static size_t gotten_store;  static size_t gotten_store;
194    static size_t first_gotten_store = 0;
195    static const unsigned char *last_callout_mark = NULL;
196    
197  /* The buffers grow automatically if very long input lines are encountered. */  /* The buffers grow automatically if very long input lines are encountered. */
198    
# Line 183  static uschar *buffer = NULL; Line 201  static uschar *buffer = NULL;
201  static uschar *dbuffer = NULL;  static uschar *dbuffer = NULL;
202  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
203    
204    /* Textual explanations for runtime error codes */
205    
206    static const char *errtexts[] = {
207      NULL,  /* 0 is no error */
208      NULL,  /* NOMATCH is handled specially */
209      "NULL argument passed",
210      "bad option value",
211      "magic number missing",
212      "unknown opcode - pattern overwritten?",
213      "no more memory",
214      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
215      "match limit exceeded",
216      "callout error code",
217      NULL,  /* BADUTF8 is handled specially */
218      "bad UTF-8 offset",
219      NULL,  /* PARTIAL is handled specially */
220      "not used - internal error",
221      "internal error - pattern overwritten?",
222      "bad count value",
223      "item unsupported for DFA matching",
224      "backreference condition or recursion test not supported for DFA matching",
225      "match limit not supported for DFA matching",
226      "workspace size exceeded in DFA matching",
227      "too much recursion for DFA matching",
228      "recursion limit exceeded",
229      "not used - internal error",
230      "invalid combination of newline options",
231      "bad offset value",
232      NULL,  /* SHORTUTF8 is handled specially */
233      "nested recursion at the same subject position",
234      "JIT stack limit reached"
235    };
236    
237    
238    /*************************************************
239    *         Alternate character tables             *
240    *************************************************/
241    
242    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
243    using the default tables of the library. However, the T option can be used to
244    select alternate sets of tables, for different kinds of testing. Note also that
245    the L (locale) option also adjusts the tables. */
246    
247    /* This is the set of tables distributed as default with PCRE. It recognizes
248    only ASCII characters. */
249    
250    static const unsigned char tables0[] = {
251    
252    /* This table is a lower casing table. */
253    
254        0,  1,  2,  3,  4,  5,  6,  7,
255        8,  9, 10, 11, 12, 13, 14, 15,
256       16, 17, 18, 19, 20, 21, 22, 23,
257       24, 25, 26, 27, 28, 29, 30, 31,
258       32, 33, 34, 35, 36, 37, 38, 39,
259       40, 41, 42, 43, 44, 45, 46, 47,
260       48, 49, 50, 51, 52, 53, 54, 55,
261       56, 57, 58, 59, 60, 61, 62, 63,
262       64, 97, 98, 99,100,101,102,103,
263      104,105,106,107,108,109,110,111,
264      112,113,114,115,116,117,118,119,
265      120,121,122, 91, 92, 93, 94, 95,
266       96, 97, 98, 99,100,101,102,103,
267      104,105,106,107,108,109,110,111,
268      112,113,114,115,116,117,118,119,
269      120,121,122,123,124,125,126,127,
270      128,129,130,131,132,133,134,135,
271      136,137,138,139,140,141,142,143,
272      144,145,146,147,148,149,150,151,
273      152,153,154,155,156,157,158,159,
274      160,161,162,163,164,165,166,167,
275      168,169,170,171,172,173,174,175,
276      176,177,178,179,180,181,182,183,
277      184,185,186,187,188,189,190,191,
278      192,193,194,195,196,197,198,199,
279      200,201,202,203,204,205,206,207,
280      208,209,210,211,212,213,214,215,
281      216,217,218,219,220,221,222,223,
282      224,225,226,227,228,229,230,231,
283      232,233,234,235,236,237,238,239,
284      240,241,242,243,244,245,246,247,
285      248,249,250,251,252,253,254,255,
286    
287    /* This table is a case flipping table. */
288    
289        0,  1,  2,  3,  4,  5,  6,  7,
290        8,  9, 10, 11, 12, 13, 14, 15,
291       16, 17, 18, 19, 20, 21, 22, 23,
292       24, 25, 26, 27, 28, 29, 30, 31,
293       32, 33, 34, 35, 36, 37, 38, 39,
294       40, 41, 42, 43, 44, 45, 46, 47,
295       48, 49, 50, 51, 52, 53, 54, 55,
296       56, 57, 58, 59, 60, 61, 62, 63,
297       64, 97, 98, 99,100,101,102,103,
298      104,105,106,107,108,109,110,111,
299      112,113,114,115,116,117,118,119,
300      120,121,122, 91, 92, 93, 94, 95,
301       96, 65, 66, 67, 68, 69, 70, 71,
302       72, 73, 74, 75, 76, 77, 78, 79,
303       80, 81, 82, 83, 84, 85, 86, 87,
304       88, 89, 90,123,124,125,126,127,
305      128,129,130,131,132,133,134,135,
306      136,137,138,139,140,141,142,143,
307      144,145,146,147,148,149,150,151,
308      152,153,154,155,156,157,158,159,
309      160,161,162,163,164,165,166,167,
310      168,169,170,171,172,173,174,175,
311      176,177,178,179,180,181,182,183,
312      184,185,186,187,188,189,190,191,
313      192,193,194,195,196,197,198,199,
314      200,201,202,203,204,205,206,207,
315      208,209,210,211,212,213,214,215,
316      216,217,218,219,220,221,222,223,
317      224,225,226,227,228,229,230,231,
318      232,233,234,235,236,237,238,239,
319      240,241,242,243,244,245,246,247,
320      248,249,250,251,252,253,254,255,
321    
322    /* This table contains bit maps for various character classes. Each map is 32
323    bytes long and the bits run from the least significant end of each byte. The
324    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
325    graph, print, punct, and cntrl. Other classes are built from combinations. */
326    
327      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
328      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
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      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
334      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
335      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
336    
337      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
338      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
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      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
344      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
345      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
346    
347      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
348      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
349      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
350      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
351    
352      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
353      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
354      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
355      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
356    
357      0x00,0x00,0x00,0x00,0xfe,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      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
363      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
364      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
365      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
366    
367      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
368      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
369      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
370      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
371    
372      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
373      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
374      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
375      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
376    
377    /* This table identifies various classes of character by individual bits:
378      0x01   white space character
379      0x02   letter
380      0x04   decimal digit
381      0x08   hexadecimal digit
382      0x10   alphanumeric or '_'
383      0x80   regular expression metacharacter or binary zero
384    */
385    
386      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
387      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
388      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
389      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
390      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
391      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
392      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
393      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
394      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
395      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
396      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
397      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
398      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
399      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
400      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
401      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
412      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
413      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
414      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
415      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
416      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
417      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
418    
419    /* This is a set of tables that came orginally from a Windows user. It seems to
420    be at least an approximation of ISO 8859. In particular, there are characters
421    greater than 128 that are marked as spaces, letters, etc. */
422    
423    static const unsigned char tables1[] = {
424    0,1,2,3,4,5,6,7,
425    8,9,10,11,12,13,14,15,
426    16,17,18,19,20,21,22,23,
427    24,25,26,27,28,29,30,31,
428    32,33,34,35,36,37,38,39,
429    40,41,42,43,44,45,46,47,
430    48,49,50,51,52,53,54,55,
431    56,57,58,59,60,61,62,63,
432    64,97,98,99,100,101,102,103,
433    104,105,106,107,108,109,110,111,
434    112,113,114,115,116,117,118,119,
435    120,121,122,91,92,93,94,95,
436    96,97,98,99,100,101,102,103,
437    104,105,106,107,108,109,110,111,
438    112,113,114,115,116,117,118,119,
439    120,121,122,123,124,125,126,127,
440    128,129,130,131,132,133,134,135,
441    136,137,138,139,140,141,142,143,
442    144,145,146,147,148,149,150,151,
443    152,153,154,155,156,157,158,159,
444    160,161,162,163,164,165,166,167,
445    168,169,170,171,172,173,174,175,
446    176,177,178,179,180,181,182,183,
447    184,185,186,187,188,189,190,191,
448    224,225,226,227,228,229,230,231,
449    232,233,234,235,236,237,238,239,
450    240,241,242,243,244,245,246,215,
451    248,249,250,251,252,253,254,223,
452    224,225,226,227,228,229,230,231,
453    232,233,234,235,236,237,238,239,
454    240,241,242,243,244,245,246,247,
455    248,249,250,251,252,253,254,255,
456    0,1,2,3,4,5,6,7,
457    8,9,10,11,12,13,14,15,
458    16,17,18,19,20,21,22,23,
459    24,25,26,27,28,29,30,31,
460    32,33,34,35,36,37,38,39,
461    40,41,42,43,44,45,46,47,
462    48,49,50,51,52,53,54,55,
463    56,57,58,59,60,61,62,63,
464    64,97,98,99,100,101,102,103,
465    104,105,106,107,108,109,110,111,
466    112,113,114,115,116,117,118,119,
467    120,121,122,91,92,93,94,95,
468    96,65,66,67,68,69,70,71,
469    72,73,74,75,76,77,78,79,
470    80,81,82,83,84,85,86,87,
471    88,89,90,123,124,125,126,127,
472    128,129,130,131,132,133,134,135,
473    136,137,138,139,140,141,142,143,
474    144,145,146,147,148,149,150,151,
475    152,153,154,155,156,157,158,159,
476    160,161,162,163,164,165,166,167,
477    168,169,170,171,172,173,174,175,
478    176,177,178,179,180,181,182,183,
479    184,185,186,187,188,189,190,191,
480    224,225,226,227,228,229,230,231,
481    232,233,234,235,236,237,238,239,
482    240,241,242,243,244,245,246,215,
483    248,249,250,251,252,253,254,223,
484    192,193,194,195,196,197,198,199,
485    200,201,202,203,204,205,206,207,
486    208,209,210,211,212,213,214,247,
487    216,217,218,219,220,221,222,255,
488    0,62,0,0,1,0,0,0,
489    0,0,0,0,0,0,0,0,
490    32,0,0,0,1,0,0,0,
491    0,0,0,0,0,0,0,0,
492    0,0,0,0,0,0,255,3,
493    126,0,0,0,126,0,0,0,
494    0,0,0,0,0,0,0,0,
495    0,0,0,0,0,0,0,0,
496    0,0,0,0,0,0,255,3,
497    0,0,0,0,0,0,0,0,
498    0,0,0,0,0,0,12,2,
499    0,0,0,0,0,0,0,0,
500    0,0,0,0,0,0,0,0,
501    254,255,255,7,0,0,0,0,
502    0,0,0,0,0,0,0,0,
503    255,255,127,127,0,0,0,0,
504    0,0,0,0,0,0,0,0,
505    0,0,0,0,254,255,255,7,
506    0,0,0,0,0,4,32,4,
507    0,0,0,128,255,255,127,255,
508    0,0,0,0,0,0,255,3,
509    254,255,255,135,254,255,255,7,
510    0,0,0,0,0,4,44,6,
511    255,255,127,255,255,255,127,255,
512    0,0,0,0,254,255,255,255,
513    255,255,255,255,255,255,255,127,
514    0,0,0,0,254,255,255,255,
515    255,255,255,255,255,255,255,255,
516    0,2,0,0,255,255,255,255,
517    255,255,255,255,255,255,255,127,
518    0,0,0,0,255,255,255,255,
519    255,255,255,255,255,255,255,255,
520    0,0,0,0,254,255,0,252,
521    1,0,0,248,1,0,0,120,
522    0,0,0,0,254,255,255,255,
523    0,0,128,0,0,0,128,0,
524    255,255,255,255,0,0,0,0,
525    0,0,0,0,0,0,0,128,
526    255,255,255,255,0,0,0,0,
527    0,0,0,0,0,0,0,0,
528    128,0,0,0,0,0,0,0,
529    0,1,1,0,1,1,0,0,
530    0,0,0,0,0,0,0,0,
531    0,0,0,0,0,0,0,0,
532    1,0,0,0,128,0,0,0,
533    128,128,128,128,0,0,128,0,
534    28,28,28,28,28,28,28,28,
535    28,28,0,0,0,0,0,128,
536    0,26,26,26,26,26,26,18,
537    18,18,18,18,18,18,18,18,
538    18,18,18,18,18,18,18,18,
539    18,18,18,128,128,0,128,16,
540    0,26,26,26,26,26,26,18,
541    18,18,18,18,18,18,18,18,
542    18,18,18,18,18,18,18,18,
543    18,18,18,128,128,0,0,0,
544    0,0,0,0,0,1,0,0,
545    0,0,0,0,0,0,0,0,
546    0,0,0,0,0,0,0,0,
547    0,0,0,0,0,0,0,0,
548    1,0,0,0,0,0,0,0,
549    0,0,18,0,0,0,0,0,
550    0,0,20,20,0,18,0,0,
551    0,20,18,0,0,0,0,0,
552    18,18,18,18,18,18,18,18,
553    18,18,18,18,18,18,18,18,
554    18,18,18,18,18,18,18,0,
555    18,18,18,18,18,18,18,18,
556    18,18,18,18,18,18,18,18,
557    18,18,18,18,18,18,18,18,
558    18,18,18,18,18,18,18,0,
559    18,18,18,18,18,18,18,18
560    };
561    
562    
563    
564    
565    #ifndef HAVE_STRERROR
566    /*************************************************
567    *     Provide strerror() for non-ANSI libraries  *
568    *************************************************/
569    
570    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
571    in their libraries, but can provide the same facility by this simple
572    alternative function. */
573    
574    extern int   sys_nerr;
575    extern char *sys_errlist[];
576    
577    char *
578    strerror(int n)
579    {
580    if (n < 0 || n >= sys_nerr) return "unknown error number";
581    return sys_errlist[n];
582    }
583    #endif /* HAVE_STRERROR */
584    
585    
586    /*************************************************
587    *         JIT memory callback                    *
588    *************************************************/
589    
590    static pcre_jit_stack* jit_callback(void *arg)
591    {
592    return (pcre_jit_stack *)arg;
593    }
594    
595    
596  /*************************************************  /*************************************************
# Line 215  uschar *here = start; Line 623  uschar *here = start;
623    
624  for (;;)  for (;;)
625    {    {
626    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
627    
628    if (rlen > 1000)    if (rlen > 1000)
629      {      {
# Line 245  for (;;) Line 653  for (;;)
653      /* 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. */
654    
655        {        {
656        if (f == stdin) printf(prompt);        if (f == stdin) printf("%s", prompt);
657        if (fgets((char *)here, rlen,  f) == NULL)        if (fgets((char *)here, rlen,  f) == NULL)
658          return (here == start)? NULL : start;          return (here == start)? NULL : start;
659        }        }
# Line 565  fprintf(outfile, "%.*s", (cb->next_item_ Line 973  fprintf(outfile, "%.*s", (cb->next_item_
973  fprintf(outfile, "\n");  fprintf(outfile, "\n");
974  first_callout = 0;  first_callout = 0;
975    
976    if (cb->mark != last_callout_mark)
977      {
978      fprintf(outfile, "Latest Mark: %s\n",
979        (cb->mark == NULL)? "<unset>" : (char *)(cb->mark));
980      last_callout_mark = cb->mark;
981      }
982    
983  if (cb->callout_data != NULL)  if (cb->callout_data != NULL)
984    {    {
985    int callout_data = *((int *)(cb->callout_data));    int callout_data = *((int *)(cb->callout_data));
# Line 584  return (cb->callout_number != callout_fa Line 999  return (cb->callout_number != callout_fa
999  *            Local malloc functions              *  *            Local malloc functions              *
1000  *************************************************/  *************************************************/
1001    
1002  /* Alternative malloc function, to test functionality and show the size of the  /* Alternative malloc function, to test functionality and save the size of a
1003  compiled re. */  compiled re, which is the first store request that pcre_compile() makes. The
1004    show_malloc variable is set only during matching. */
1005    
1006  static void *new_malloc(size_t size)  static void *new_malloc(size_t size)
1007  {  {
1008  void *block = malloc(size);  void *block = malloc(size);
1009  gotten_store = size;  gotten_store = size;
1010    if (first_gotten_store == 0) first_gotten_store = size;
1011  if (show_malloc)  if (show_malloc)
1012    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);
1013  return block;  return block;
# Line 603  if (show_malloc) Line 1020  if (show_malloc)
1020  free(block);  free(block);
1021  }  }
1022    
   
1023  /* For recursion malloc/free, to test stacking calls */  /* For recursion malloc/free, to test stacking calls */
1024    
1025  static void *stack_malloc(size_t size)  static void *stack_malloc(size_t size)
# Line 735  return 0; Line 1151  return 0;
1151  *         Check newline indicator                *  *         Check newline indicator                *
1152  *************************************************/  *************************************************/
1153    
1154  /* 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
1155  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.  
1156    
1157  Arguments:  Arguments:
1158    p           points after the leading '<'    p           points after the leading '<'
# Line 785  printf("  -dfa     force DFA matching fo Line 1200  printf("  -dfa     force DFA matching fo
1200  #endif  #endif
1201  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
1202  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
1203           "  -M       find MATCH_LIMIT minimum for each subject\n"
1204         "  -m       output memory used information\n"         "  -m       output memory used information\n"
1205         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
1206  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 792  printf("  -p       use POSIX interface\n Line 1208  printf("  -p       use POSIX interface\n
1208  #endif  #endif
1209  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
1210  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
1211  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied at basic level\n"
1212           "  -s+      force each pattern to be studied, using JIT if available\n"
1213         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
1214  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1215  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 814  int main(int argc, char **argv) Line 1231  int main(int argc, char **argv)
1231  FILE *infile = stdin;  FILE *infile = stdin;
1232  int options = 0;  int options = 0;
1233  int study_options = 0;  int study_options = 0;
1234    int default_find_match_limit = FALSE;
1235  int op = 1;  int op = 1;
1236  int timeit = 0;  int timeit = 0;
1237  int timeitm = 0;  int timeitm = 0;
1238  int showinfo = 0;  int showinfo = 0;
1239  int showstore = 0;  int showstore = 0;
1240    int force_study = -1;
1241    int force_study_options = 0;
1242  int quiet = 0;  int quiet = 0;
1243  int size_offsets = 45;  int size_offsets = 45;
1244  int size_offsets_max;  int size_offsets_max;
# Line 832  int all_use_dfa = 0; Line 1252  int all_use_dfa = 0;
1252  int yield = 0;  int yield = 0;
1253  int stack_size;  int stack_size;
1254    
1255    pcre_jit_stack *jit_stack = NULL;
1256    
1257    
1258  /* These vectors store, end-to-end, a list of captured substring names. Assume  /* These vectors store, end-to-end, a list of captured substring names. Assume
1259  that 1024 is plenty long enough for the few names we'll be testing. */  that 1024 is plenty long enough for the few names we'll be testing. */
1260    
# Line 867  while (argc > 1 && argv[op][0] == '-') Line 1290  while (argc > 1 && argv[op][0] == '-')
1290    {    {
1291    unsigned char *endptr;    unsigned char *endptr;
1292    
1293    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1294      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 0;
1295      else if (strcmp(argv[op], "-s+") == 0)
1296        {
1297        force_study = 1;
1298        force_study_options = PCRE_STUDY_JIT_COMPILE;
1299        }
1300    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1301    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1302    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1303    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1304      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1305  #if !defined NODFA  #if !defined NODFA
1306    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1307  #endif  #endif
# Line 901  while (argc > 1 && argv[op][0] == '-') Line 1330  while (argc > 1 && argv[op][0] == '-')
1330        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1331          *endptr == 0))          *endptr == 0))
1332      {      {
1333  #if defined(_WIN32) || defined(WIN32)  #if defined(_WIN32) || defined(WIN32) || defined(__minix)
1334      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
1335      exit(1);      exit(1);
1336  #else  #else
# Line 925  while (argc > 1 && argv[op][0] == '-') Line 1354  while (argc > 1 && argv[op][0] == '-')
1354    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1355      {      {
1356      int rc;      int rc;
1357      unsigned long int lrc;      unsigned long int lrc;
1358      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1359      printf("Compiled with\n");      printf("Compiled with\n");
1360      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
1361      printf("  %sUTF-8 support\n", rc? "" : "No ");      printf("  %sUTF-8 support\n", rc? "" : "No ");
1362      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1363      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1364        (void)pcre_config(PCRE_CONFIG_JIT, &rc);
1365        if (rc)
1366          printf("  Just-in-time compiler support\n");
1367        else
1368          printf("  No just-in-time compiler support\n");
1369      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1370      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1371        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1372        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1373          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1374        (rc == -2)? "ANYCRLF" :        (rc == -2)? "ANYCRLF" :
1375        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1376      (void)pcre_config(PCRE_CONFIG_BSR, &rc);      (void)pcre_config(PCRE_CONFIG_BSR, &rc);
# Line 1029  while (!done) Line 1465  while (!done)
1465  #endif  #endif
1466    
1467    const char *error;    const char *error;
1468      unsigned char *markptr;
1469    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1470    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1471    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1472    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1473    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1474      int do_allcaps = 0;
1475      int do_mark = 0;
1476    int do_study = 0;    int do_study = 0;
1477      int no_force_study = 0;
1478    int do_debug = debug;    int do_debug = debug;
1479    int do_G = 0;    int do_G = 0;
1480    int do_g = 0;    int do_g = 0;
1481    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1482    int do_showrest = 0;    int do_showrest = 0;
1483      int do_showcaprest = 0;
1484    int do_flip = 0;    int do_flip = 0;
1485    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1486    
# Line 1082  while (!done) Line 1523  while (!done)
1523        (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7];        (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7];
1524    
1525      re = (real_pcre *)new_malloc(true_size);      re = (real_pcre *)new_malloc(true_size);
1526      regex_gotten_store = gotten_store;      regex_gotten_store = first_gotten_store;
1527    
1528      if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ;      if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ;
1529    
# Line 1101  while (!done) Line 1542  while (!done)
1542          }          }
1543        }        }
1544    
1545      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1546        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1547    
1548      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1109  while (!done) Line 1550  while (!done)
1550      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1551      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1552    
1553      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1554    
1555      if (true_study_size != 0)      if (true_study_size != 0)
1556        {        {
# Line 1125  while (!done) Line 1566  while (!done)
1566          {          {
1567          FAIL_READ:          FAIL_READ:
1568          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
1569          if (extra != NULL) new_free(extra);          if (extra != NULL) pcre_free_study(extra);
1570          if (re != NULL) new_free(re);          if (re != NULL) new_free(re);
1571          fclose(f);          fclose(f);
1572          continue;          continue;
# Line 1151  while (!done) Line 1592  while (!done)
1592      }      }
1593    
1594    pp = p;    pp = p;
1595    poffset = p - buffer;    poffset = (int)(p - buffer);
1596    
1597    for(;;)    for(;;)
1598      {      {
# Line 1191  while (!done) Line 1632  while (!done)
1632    /* Look for options after final delimiter */    /* Look for options after final delimiter */
1633    
1634    options = 0;    options = 0;
1635    study_options = 0;    study_options = 0;
1636    log_store = showstore;  /* default from command line */    log_store = showstore;  /* default from command line */
1637    
1638    while (*pp != 0)    while (*pp != 0)
# Line 1205  while (!done) Line 1646  while (!done)
1646        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1647        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1648    
1649        case '+': do_showrest = 1; break;        case '+':
1650          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1651          break;
1652    
1653          case '=': do_allcaps = 1; break;
1654        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1655        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
1656        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1215  while (!done) Line 1660  while (!done)
1660        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1661        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1662        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1663          case 'K': do_mark = 1; break;
1664        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1665        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1666    
# Line 1222  while (!done) Line 1668  while (!done)
1668        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1669  #endif  #endif
1670    
1671        case 'S': do_study = 1; break;        case 'S':
1672          if (do_study == 0)
1673            {
1674            do_study = 1;
1675            if (*pp == '+')
1676              {
1677              study_options |= PCRE_STUDY_JIT_COMPILE;
1678              pp++;
1679              }
1680            }
1681          else
1682            {
1683            do_study = 0;
1684            no_force_study = 1;
1685            }
1686          break;
1687    
1688        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1689          case 'W': options |= PCRE_UCP; break;
1690        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1691          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1692        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1693        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1694        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1695    
1696          case 'T':
1697          switch (*pp++)
1698            {
1699            case '0': tables = tables0; break;
1700            case '1': tables = tables1; break;
1701    
1702            case '\r':
1703            case '\n':
1704            case ' ':
1705            case 0:
1706            fprintf(outfile, "** Missing table number after /T\n");
1707            goto SKIP_DATA;
1708    
1709            default:
1710            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1711            goto SKIP_DATA;
1712            }
1713          break;
1714    
1715        case 'L':        case 'L':
1716        ppp = pp;        ppp = pp;
1717        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1254  while (!done) Line 1737  while (!done)
1737    
1738        case '<':        case '<':
1739          {          {
1740          if (strncmp((char *)pp, "JS>", 3) == 0)          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1741            {            {
1742            options |= PCRE_JAVASCRIPT_COMPAT;            options |= PCRE_JAVASCRIPT_COMPAT;
1743            pp += 3;            pp += 3;
# Line 1295  while (!done) Line 1778  while (!done)
1778      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1779      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1780      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1781        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1782        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1783    
1784        first_gotten_store = 0;
1785      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1786    
1787      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
# Line 1315  while (!done) Line 1801  while (!done)
1801  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1802    
1803      {      {
1804        unsigned long int get_options;
1805    
1806      if (timeit > 0)      if (timeit > 0)
1807        {        {
1808        register int i;        register int i;
# Line 1331  while (!done) Line 1819  while (!done)
1819            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1820        }        }
1821    
1822        first_gotten_store = 0;
1823      re = pcre_compile((char *)p, options, &error, &erroroffset, tables);      re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1824    
1825      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
# Line 1358  while (!done) Line 1847  while (!done)
1847        goto CONTINUE;        goto CONTINUE;
1848        }        }
1849    
1850      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1851      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
1852      returns only limited data. Check that it agrees with the newer one. */      lines. */
1853    
1854      if (log_store)      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1855        fprintf(outfile, "Memory allocation (code space): %d\n",      if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
         (int)(gotten_store -  
               sizeof(real_pcre) -  
               ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size));  
1856    
1857      /* Extract the size for possible writing before possibly flipping it,      /* Extract the size for possible writing before possibly flipping it,
1858      and remember the store that was got. */      and remember the store that was got. */
1859    
1860      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1861      regex_gotten_store = gotten_store;      regex_gotten_store = first_gotten_store;
1862    
1863      /* If /S was present, study the regexp to generate additional info to      /* Output code size information if requested */
     help with the matching. */  
1864    
1865      if (do_study)      if (log_store)
1866          fprintf(outfile, "Memory allocation (code space): %d\n",
1867            (int)(first_gotten_store -
1868                  sizeof(real_pcre) -
1869                  ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size));
1870    
1871        /* If -s or /S was present, study the regex to generate additional info to
1872        help with the matching, unless the pattern has the SS option, which
1873        suppresses the effect of /S (used for a few test patterns where studying is
1874        never sensible). */
1875    
1876        if (do_study || (force_study >= 0 && !no_force_study))
1877        {        {
1878        if (timeit > 0)        if (timeit > 0)
1879          {          {
# Line 1385  while (!done) Line 1881  while (!done)
1881          clock_t time_taken;          clock_t time_taken;
1882          clock_t start_time = clock();          clock_t start_time = clock();
1883          for (i = 0; i < timeit; i++)          for (i = 0; i < timeit; i++)
1884            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options | force_study_options, &error);
1885          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1886          if (extra != NULL) free(extra);          if (extra != NULL) pcre_free_study(extra);
1887          fprintf(outfile, "  Study time %.4f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1888            (((double)time_taken * 1000.0) / (double)timeit) /            (((double)time_taken * 1000.0) / (double)timeit) /
1889              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1890          }          }
1891        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options | force_study_options, &error);
1892        if (error != NULL)        if (error != NULL)
1893          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
1894        else if (extra != NULL)        else if (extra != NULL)
1895            {
1896          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1897            if (log_store)
1898              {
1899              size_t jitsize;
1900              new_info(re, extra, PCRE_INFO_JITSIZE, &jitsize);
1901              if (jitsize != 0)
1902                fprintf(outfile, "Memory allocation (JIT code): %d\n", jitsize);
1903              }
1904            }
1905          }
1906    
1907        /* If /K was present, we set up for handling MARK data. */
1908    
1909        if (do_mark)
1910          {
1911          if (extra == NULL)
1912            {
1913            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1914            extra->flags = 0;
1915            }
1916          extra->mark = &markptr;
1917          extra->flags |= PCRE_EXTRA_MARK;
1918        }        }
1919    
1920      /* 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
# Line 1431  while (!done) Line 1949  while (!done)
1949          {          {
1950          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1951          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1952          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1953            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1954          }          }
1955        }        }
1956    
1957      /* Extract information from the compiled data if required */      /* Extract information from the compiled data if required. There are now
1958        two info-returning functions. The old one has a limited interface and
1959        returns only limited data. Check that it agrees with the newer one. */
1960    
1961      SHOW_INFO:      SHOW_INFO:
1962    
# Line 1445  while (!done) Line 1966  while (!done)
1966        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1967        }        }
1968    
1969        /* We already have the options in get_options (see above) */
1970    
1971      if (do_showinfo)      if (do_showinfo)
1972        {        {
1973        unsigned long int get_options, all_options;        unsigned long int all_options;
1974  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1975        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1976  #endif  #endif
# Line 1456  while (!done) Line 1979  while (!done)
1979        int nameentrysize, namecount;        int nameentrysize, namecount;
1980        const uschar *nametable;        const uschar *nametable;
1981    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1982        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1983        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1984        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1516  while (!done) Line 2038  while (!done)
2038        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
2039    
2040        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
2041          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",
2042            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
2043            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
2044            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
# Line 1530  while (!done) Line 2052  while (!done)
2052            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
2053            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
2054            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
2055              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2056            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2057              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2058            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2059    
2060        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
# Line 1598  while (!done) Line 2122  while (!done)
2122        /* 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
2123        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2124        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
2125        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2126          information unless -i or -d was also present. This means that, except
2127          when auto-callouts are involved, the output from runs with and without
2128          -s should be identical. */
2129    
2130        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
2131          {          {
2132          if (extra == NULL)          if (extra == NULL)
2133            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2134          else          else
2135            {            {
2136            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2137            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2138    
2139              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2140              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2141    
2142              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2143            if (start_bits == NULL)            if (start_bits == NULL)
2144              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2145            else            else
2146              {              {
2147              int i;              int i;
# Line 1640  while (!done) Line 2171  while (!done)
2171              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2172              }              }
2173            }            }
2174    
2175            /* Show this only if the JIT was set by /S, not by -s. */
2176    
2177            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
2178              {
2179              int jit;
2180              new_info(re, extra, PCRE_INFO_JIT, &jit);
2181              if (jit)
2182                fprintf(outfile, "JIT study was successful\n");
2183              else
2184    #ifdef SUPPORT_JIT
2185                fprintf(outfile, "JIT study was not successful\n");
2186    #else
2187                fprintf(outfile, "JIT support is not available in this version of PCRE\n");
2188    #endif
2189              }
2190          }          }
2191        }        }
2192    
# Line 1674  while (!done) Line 2221  while (!done)
2221            }            }
2222          else          else
2223            {            {
2224            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2225    
2226              /* If there is study data, write it. */
2227    
2228            if (extra != NULL)            if (extra != NULL)
2229              {              {
2230              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1684  while (!done) Line 2234  while (!done)
2234                  strerror(errno));                  strerror(errno));
2235                }                }
2236              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2237              }              }
2238            }            }
2239          fclose(f);          fclose(f);
2240          }          }
2241    
2242        new_free(re);        new_free(re);
2243        if (extra != NULL) new_free(extra);        if (extra != NULL) pcre_free_study(extra);
2244        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2245            {
2246            new_free((void *)tables);
2247            setlocale(LC_CTYPE, "C");
2248            locale_set = 0;
2249            }
2250        continue;  /* With next regex */        continue;  /* With next regex */
2251        }        }
2252      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1709  while (!done) Line 2263  while (!done)
2263      int callout_data_set = 0;      int callout_data_set = 0;
2264      int count, c;      int count, c;
2265      int copystrings = 0;      int copystrings = 0;
2266      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2267      int getstrings = 0;      int getstrings = 0;
2268      int getlist = 0;      int getlist = 0;
2269      int gmatched = 0;      int gmatched = 0;
2270      int start_offset = 0;      int start_offset = 0;
2271        int start_offset_sign = 1;
2272      int g_notempty = 0;      int g_notempty = 0;
2273      int use_dfa = 0;      int use_dfa = 0;
2274    
# Line 1727  while (!done) Line 2282  while (!done)
2282    
2283      pcre_callout = callout;      pcre_callout = callout;
2284      first_callout = 1;      first_callout = 1;
2285        last_callout_mark = NULL;
2286      callout_extra = 0;      callout_extra = 0;
2287      callout_count = 0;      callout_count = 0;
2288      callout_fail_count = 999999;      callout_fail_count = 999999;
# Line 1741  while (!done) Line 2297  while (!done)
2297        {        {
2298        if (extend_inputline(infile, buffer + len, "data> ") == NULL)        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
2299          {          {
2300          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2301              {
2302              fprintf(outfile, "\n");
2303              break;
2304              }
2305          done = 1;          done = 1;
2306          goto CONTINUE;          goto CONTINUE;
2307          }          }
# Line 1801  while (!done) Line 2361  while (!done)
2361            {            {
2362            unsigned char *pt = p;            unsigned char *pt = p;
2363            c = 0;            c = 0;
2364            while (isxdigit(*(++pt)))  
2365              c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W');            /* We used to have "while (isxdigit(*(++pt)))" here, but it fails
2366              when isxdigit() is a macro that refers to its argument more than
2367              once. This is banned by the C Standard, but apparently happens in at
2368              least one MacOS environment. */
2369    
2370              for (pt++; isxdigit(*pt); pt++)
2371                c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'a' - 10);
2372            if (*pt == '}')            if (*pt == '}')
2373              {              {
2374              unsigned char buff8[8];              unsigned char buff8[8];
# Line 1832  while (!done) Line 2398  while (!done)
2398          c = 0;          c = 0;
2399          while (i++ < 2 && isxdigit(*p))          while (i++ < 2 && isxdigit(*p))
2400            {            {
2401            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W');            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'a' - 10);
2402            p++;            p++;
2403            }            }
2404          break;          break;
# Line 1842  while (!done) Line 2408  while (!done)
2408          continue;          continue;
2409    
2410          case '>':          case '>':
2411            if (*p == '-')
2412              {
2413              start_offset_sign = -1;
2414              p++;
2415              }
2416          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2417            start_offset *= start_offset_sign;
2418          continue;          continue;
2419    
2420          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1915  while (!done) Line 2487  while (!done)
2487  #endif  #endif
2488            use_dfa = 1;            use_dfa = 1;
2489          continue;          continue;
2490    #endif
2491    
2492    #if !defined NODFA
2493          case 'F':          case 'F':
2494          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2495          continue;          continue;
# Line 1940  while (!done) Line 2514  while (!done)
2514            }            }
2515          continue;          continue;
2516    
2517            case 'J':
2518            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2519            if (extra != NULL
2520                && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
2521                && extra->executable_jit != NULL)
2522              {
2523              if (jit_stack != NULL) pcre_jit_stack_free(jit_stack);
2524              jit_stack = pcre_jit_stack_alloc(1, n * 1024);
2525              pcre_assign_jit_stack(extra, jit_callback, jit_stack);
2526              }
2527            continue;
2528    
2529          case 'L':          case 'L':
2530          getlist = 1;          getlist = 1;
2531          continue;          continue;
# Line 1949  while (!done) Line 2535  while (!done)
2535          continue;          continue;
2536    
2537          case 'N':          case 'N':
2538          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2539              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2540            else
2541              options |= PCRE_NOTEMPTY;
2542          continue;          continue;
2543    
2544          case 'O':          case 'O':
# Line 1972  while (!done) Line 2561  while (!done)
2561          continue;          continue;
2562    
2563          case 'P':          case 'P':
2564          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2565              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2566          continue;          continue;
2567    
2568          case 'Q':          case 'Q':
# Line 2007  while (!done) Line 2597  while (!done)
2597          show_malloc = 1;          show_malloc = 1;
2598          continue;          continue;
2599    
2600            case 'Y':
2601            options |= PCRE_NO_START_OPTIMIZE;
2602            continue;
2603    
2604          case 'Z':          case 'Z':
2605          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2606          continue;          continue;
# Line 2027  while (!done) Line 2621  while (!done)
2621        *q++ = c;        *q++ = c;
2622        }        }
2623      *q = 0;      *q = 0;
2624      len = q - dbuffer;      len = (int)(q - dbuffer);
2625    
2626      /* 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
2627      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 2065  while (!done) Line 2659  while (!done)
2659          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2660        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2661        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2662          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2663    
2664        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2665    
# Line 2089  while (!done) Line 2684  while (!done)
2684              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2685                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2686              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2687              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2688                {                {
2689                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2690                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2691                  outfile);                  outfile);
2692                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 2109  while (!done) Line 2704  while (!done)
2704    
2705      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2706        {        {
2707          markptr = NULL;
2708    
2709        if (timeitm > 0)        if (timeitm > 0)
2710          {          {
2711          register int i;          register int i;
# Line 2120  while (!done) Line 2717  while (!done)
2717            {            {
2718            int workspace[1000];            int workspace[1000];
2719            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2720              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2721                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2722                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2723            }            }
# Line 2139  while (!done) Line 2736  while (!done)
2736    
2737        /* If find_match_limit is set, we want to do repeated matches with        /* If find_match_limit is set, we want to do repeated matches with
2738        varying limits in order to find the minimum value for the match limit and        varying limits in order to find the minimum value for the match limit and
2739        for the recursion limit. */        for the recursion limit. The match limits are relevant only to the normal
2740          running of pcre_exec(), so disable the JIT optimization. This makes it
2741          possible to run the same set of tests with and without JIT externally
2742          requested. */
2743    
2744        if (find_match_limit)        if (find_match_limit)
2745          {          {
# Line 2148  while (!done) Line 2748  while (!done)
2748            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2749            extra->flags = 0;            extra->flags = 0;
2750            }            }
2751            else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
2752    
2753          (void)check_match_limit(re, extra, bptr, len, start_offset,          (void)check_match_limit(re, extra, bptr, len, start_offset,
2754            options|g_notempty, use_offsets, use_size_offsets,            options|g_notempty, use_offsets, use_size_offsets,
# Line 2183  while (!done) Line 2784  while (!done)
2784        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2785          {          {
2786          int workspace[1000];          int workspace[1000];
2787          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2788            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2789            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2790          if (count == 0)          if (count == 0)
# Line 2231  while (!done) Line 2832  while (!done)
2832              }              }
2833            }            }
2834    
2835            /* do_allcaps requests showing of all captures in the pattern, to check
2836            unset ones at the end. */
2837    
2838            if (do_allcaps)
2839              {
2840              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2841              count++;   /* Allow for full match */
2842              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2843              }
2844    
2845            /* Output the captured substrings */
2846    
2847          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2848            {            {
2849            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2850                {
2851                if (use_offsets[i] != -1)
2852                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2853                    use_offsets[i], i);
2854                if (use_offsets[i+1] != -1)
2855                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2856                    use_offsets[i+1], i+1);
2857              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2858                }
2859            else            else
2860              {              {
2861              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2862              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2863                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2864              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2865              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2866                {                {
2867                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2868                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2869                  fprintf(outfile, " 0+ ");                  outfile);
2870                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2871                }                }
2872              }              }
2873            }            }
2874    
2875            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2876    
2877          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2878            {            {
2879            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2327  while (!done) Line 2947  while (!done)
2947                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);
2948              if (stringlist[i] != NULL)              if (stringlist[i] != NULL)
2949                fprintf(outfile, "string list not terminated by NULL\n");                fprintf(outfile, "string list not terminated by NULL\n");
             /* free((void *)stringlist); */  
2950              pcre_free_substring_list(stringlist);              pcre_free_substring_list(stringlist);
2951              }              }
2952            }            }
# Line 2337  while (!done) Line 2956  while (!done)
2956    
2957        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2958          {          {
2959          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2960  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2961          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2962            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2963              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2964  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2965                outfile);
2966              }
2967          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2968          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2969          }          }
# Line 2352  while (!done) Line 2973  while (!done)
2973        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
2974        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2975    
2976        Complication arises in the case when the newline option is "any" or        Complication arises in the case when the newline convention is "any",
2977        "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
2978        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,
2979        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
2980        Fudge the offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2981          newline setting in the pattern; if none was set, use pcre_config() to
2982          find the default.
2983    
2984        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
2985        character, not one byte. */        character, not one byte. */
# Line 2372  while (!done) Line 2995  while (!done)
2995              {              {
2996              int d;              int d;
2997              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2998              obits = (d == '\r')? PCRE_NEWLINE_CR :              /* Note that these values are always the ASCII ones, even in
2999                      (d == '\n')? PCRE_NEWLINE_LF :              EBCDIC environments. CR = 13, NL = 10. */
3000                      (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :              obits = (d == 13)? PCRE_NEWLINE_CR :
3001                        (d == 10)? PCRE_NEWLINE_LF :
3002                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
3003                      (d == -2)? PCRE_NEWLINE_ANYCRLF :                      (d == -2)? PCRE_NEWLINE_ANYCRLF :
3004                      (d == -1)? PCRE_NEWLINE_ANY : 0;                      (d == -1)? PCRE_NEWLINE_ANY : 0;
3005              }              }
3006            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
3007                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
3008                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
3009                &&                &&
3010                start_offset < len - 1 &&                start_offset < len - 1 &&
# Line 2389  while (!done) Line 3015  while (!done)
3015              {              {
3016              while (start_offset + onechar < len)              while (start_offset + onechar < len)
3017                {                {
3018                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
3019                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
3020                }                }
3021              }              }
3022            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
3023            }            }
3024          else          else
3025            {            {
3026            if (count == PCRE_ERROR_NOMATCH)            switch(count)
3027              {              {
3028              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
3029                if (gmatched == 0)
3030                  {
3031                  if (markptr == NULL) fprintf(outfile, "No match\n");
3032                    else fprintf(outfile, "No match, mark = %s\n", markptr);
3033                  }
3034                break;
3035    
3036                case PCRE_ERROR_BADUTF8:
3037                case PCRE_ERROR_SHORTUTF8:
3038                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
3039                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
3040                if (use_size_offsets >= 2)
3041                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
3042                    use_offsets[1]);
3043                fprintf(outfile, "\n");
3044                break;
3045    
3046                default:
3047                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
3048                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
3049                else
3050                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
3051                break;
3052              }              }
3053            else fprintf(outfile, "Error %d\n", count);  
3054            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
3055            }            }
3056          }          }
# Line 2413  while (!done) Line 3060  while (!done)
3060        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
3061    
3062        /* 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
3063        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
3064        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
3065        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
3066        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
3067        character. */        character. */
3068    
# Line 2424  while (!done) Line 3071  while (!done)
3071        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
3072          {          {
3073          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
3074          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
3075          }          }
3076    
3077        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2450  while (!done) Line 3097  while (!done)
3097  #endif  #endif
3098    
3099    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
3100    if (extra != NULL) new_free(extra);    if (extra != NULL) pcre_free_study(extra);
3101    if (tables != NULL)    if (locale_set)
3102      {      {
3103      new_free((void *)tables);      new_free((void *)tables);
3104      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
3105      locale_set = 0;      locale_set = 0;
3106      }      }
3107      if (jit_stack != NULL)
3108        {
3109        pcre_jit_stack_free(jit_stack);
3110        jit_stack = NULL;
3111        }
3112    }    }
3113    
3114  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");

Legend:
Removed from v.376  
changed lines
  Added in v.780

  ViewVC Help
Powered by ViewVC 1.1.5