/[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 289 by ph10, Sun Dec 23 12:17:20 2007 UTC revision 616 by ph10, Mon Jul 11 15:55:25 2011 UTC
# Line 49  POSSIBILITY OF SUCH DAMAGE. Line 49  POSSIBILITY OF SUCH DAMAGE.
49  #include <errno.h>  #include <errno.h>
50    
51  #ifdef SUPPORT_LIBREADLINE  #ifdef SUPPORT_LIBREADLINE
52    #ifdef HAVE_UNISTD_H
53  #include <unistd.h>  #include <unistd.h>
54    #endif
55  #include <readline/readline.h>  #include <readline/readline.h>
56  #include <readline/history.h>  #include <readline/history.h>
57  #endif  #endif
# Line 69  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    #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
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 */
92  #include <sys/resource.h>      /* for setrlimit(). */  #include <sys/resource.h>      /* for setrlimit(). */
# Line 89  appropriately for an application, not fo Line 107  appropriately for an application, not fo
107  #include "pcre.h"  #include "pcre.h"
108  #include "pcre_internal.h"  #include "pcre_internal.h"
109    
110  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* We need access to some of the data tables that PCRE uses. So as not to have
111  two copies, we include the source file here, changing the names of the external  to keep two copies, we include the source file here, changing the names of the
112  symbols to prevent clashes. */  external symbols to prevent clashes. */
113    
114    #define _pcre_ucp_gentype      ucp_gentype
115  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
116  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
117  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 107  symbols to prevent clashes. */ Line 126  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 177  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    };
230    
231    
232    /*************************************************
233    *         Alternate character tables             *
234    *************************************************/
235    
236    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
237    using the default tables of the library. However, the T option can be used to
238    select alternate sets of tables, for different kinds of testing. Note also that
239    the L (locale) option also adjusts the tables. */
240    
241    /* This is the set of tables distributed as default with PCRE. It recognizes
242    only ASCII characters. */
243    
244    static const unsigned char tables0[] = {
245    
246    /* This table is a lower casing table. */
247    
248        0,  1,  2,  3,  4,  5,  6,  7,
249        8,  9, 10, 11, 12, 13, 14, 15,
250       16, 17, 18, 19, 20, 21, 22, 23,
251       24, 25, 26, 27, 28, 29, 30, 31,
252       32, 33, 34, 35, 36, 37, 38, 39,
253       40, 41, 42, 43, 44, 45, 46, 47,
254       48, 49, 50, 51, 52, 53, 54, 55,
255       56, 57, 58, 59, 60, 61, 62, 63,
256       64, 97, 98, 99,100,101,102,103,
257      104,105,106,107,108,109,110,111,
258      112,113,114,115,116,117,118,119,
259      120,121,122, 91, 92, 93, 94, 95,
260       96, 97, 98, 99,100,101,102,103,
261      104,105,106,107,108,109,110,111,
262      112,113,114,115,116,117,118,119,
263      120,121,122,123,124,125,126,127,
264      128,129,130,131,132,133,134,135,
265      136,137,138,139,140,141,142,143,
266      144,145,146,147,148,149,150,151,
267      152,153,154,155,156,157,158,159,
268      160,161,162,163,164,165,166,167,
269      168,169,170,171,172,173,174,175,
270      176,177,178,179,180,181,182,183,
271      184,185,186,187,188,189,190,191,
272      192,193,194,195,196,197,198,199,
273      200,201,202,203,204,205,206,207,
274      208,209,210,211,212,213,214,215,
275      216,217,218,219,220,221,222,223,
276      224,225,226,227,228,229,230,231,
277      232,233,234,235,236,237,238,239,
278      240,241,242,243,244,245,246,247,
279      248,249,250,251,252,253,254,255,
280    
281    /* This table is a case flipping table. */
282    
283        0,  1,  2,  3,  4,  5,  6,  7,
284        8,  9, 10, 11, 12, 13, 14, 15,
285       16, 17, 18, 19, 20, 21, 22, 23,
286       24, 25, 26, 27, 28, 29, 30, 31,
287       32, 33, 34, 35, 36, 37, 38, 39,
288       40, 41, 42, 43, 44, 45, 46, 47,
289       48, 49, 50, 51, 52, 53, 54, 55,
290       56, 57, 58, 59, 60, 61, 62, 63,
291       64, 97, 98, 99,100,101,102,103,
292      104,105,106,107,108,109,110,111,
293      112,113,114,115,116,117,118,119,
294      120,121,122, 91, 92, 93, 94, 95,
295       96, 65, 66, 67, 68, 69, 70, 71,
296       72, 73, 74, 75, 76, 77, 78, 79,
297       80, 81, 82, 83, 84, 85, 86, 87,
298       88, 89, 90,123,124,125,126,127,
299      128,129,130,131,132,133,134,135,
300      136,137,138,139,140,141,142,143,
301      144,145,146,147,148,149,150,151,
302      152,153,154,155,156,157,158,159,
303      160,161,162,163,164,165,166,167,
304      168,169,170,171,172,173,174,175,
305      176,177,178,179,180,181,182,183,
306      184,185,186,187,188,189,190,191,
307      192,193,194,195,196,197,198,199,
308      200,201,202,203,204,205,206,207,
309      208,209,210,211,212,213,214,215,
310      216,217,218,219,220,221,222,223,
311      224,225,226,227,228,229,230,231,
312      232,233,234,235,236,237,238,239,
313      240,241,242,243,244,245,246,247,
314      248,249,250,251,252,253,254,255,
315    
316    /* This table contains bit maps for various character classes. Each map is 32
317    bytes long and the bits run from the least significant end of each byte. The
318    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
319    graph, print, punct, and cntrl. Other classes are built from combinations. */
320    
321      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
322      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
323      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
324      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
325    
326      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
327      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
328      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
329      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
330    
331      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
332      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
333      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
334      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
335    
336      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
337      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
338      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
339      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
340    
341      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
342      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
343      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
344      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
345    
346      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
347      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
348      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
349      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
350    
351      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
352      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
353      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
354      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
355    
356      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
357      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
358      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
359      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
360    
361      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
362      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
363      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
364      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
365    
366      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
367      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
368      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
369      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
370    
371    /* This table identifies various classes of character by individual bits:
372      0x01   white space character
373      0x02   letter
374      0x04   decimal digit
375      0x08   hexadecimal digit
376      0x10   alphanumeric or '_'
377      0x80   regular expression metacharacter or binary zero
378    */
379    
380      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
381      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
382      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
383      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
384      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
385      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
386      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
387      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
388      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
389      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
390      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
391      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
392      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
393      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
394      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
395      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
396      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
397      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
398      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
399      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
400      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
401      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
412    
413    /* This is a set of tables that came orginally from a Windows user. It seems to
414    be at least an approximation of ISO 8859. In particular, there are characters
415    greater than 128 that are marked as spaces, letters, etc. */
416    
417    static const unsigned char tables1[] = {
418    0,1,2,3,4,5,6,7,
419    8,9,10,11,12,13,14,15,
420    16,17,18,19,20,21,22,23,
421    24,25,26,27,28,29,30,31,
422    32,33,34,35,36,37,38,39,
423    40,41,42,43,44,45,46,47,
424    48,49,50,51,52,53,54,55,
425    56,57,58,59,60,61,62,63,
426    64,97,98,99,100,101,102,103,
427    104,105,106,107,108,109,110,111,
428    112,113,114,115,116,117,118,119,
429    120,121,122,91,92,93,94,95,
430    96,97,98,99,100,101,102,103,
431    104,105,106,107,108,109,110,111,
432    112,113,114,115,116,117,118,119,
433    120,121,122,123,124,125,126,127,
434    128,129,130,131,132,133,134,135,
435    136,137,138,139,140,141,142,143,
436    144,145,146,147,148,149,150,151,
437    152,153,154,155,156,157,158,159,
438    160,161,162,163,164,165,166,167,
439    168,169,170,171,172,173,174,175,
440    176,177,178,179,180,181,182,183,
441    184,185,186,187,188,189,190,191,
442    224,225,226,227,228,229,230,231,
443    232,233,234,235,236,237,238,239,
444    240,241,242,243,244,245,246,215,
445    248,249,250,251,252,253,254,223,
446    224,225,226,227,228,229,230,231,
447    232,233,234,235,236,237,238,239,
448    240,241,242,243,244,245,246,247,
449    248,249,250,251,252,253,254,255,
450    0,1,2,3,4,5,6,7,
451    8,9,10,11,12,13,14,15,
452    16,17,18,19,20,21,22,23,
453    24,25,26,27,28,29,30,31,
454    32,33,34,35,36,37,38,39,
455    40,41,42,43,44,45,46,47,
456    48,49,50,51,52,53,54,55,
457    56,57,58,59,60,61,62,63,
458    64,97,98,99,100,101,102,103,
459    104,105,106,107,108,109,110,111,
460    112,113,114,115,116,117,118,119,
461    120,121,122,91,92,93,94,95,
462    96,65,66,67,68,69,70,71,
463    72,73,74,75,76,77,78,79,
464    80,81,82,83,84,85,86,87,
465    88,89,90,123,124,125,126,127,
466    128,129,130,131,132,133,134,135,
467    136,137,138,139,140,141,142,143,
468    144,145,146,147,148,149,150,151,
469    152,153,154,155,156,157,158,159,
470    160,161,162,163,164,165,166,167,
471    168,169,170,171,172,173,174,175,
472    176,177,178,179,180,181,182,183,
473    184,185,186,187,188,189,190,191,
474    224,225,226,227,228,229,230,231,
475    232,233,234,235,236,237,238,239,
476    240,241,242,243,244,245,246,215,
477    248,249,250,251,252,253,254,223,
478    192,193,194,195,196,197,198,199,
479    200,201,202,203,204,205,206,207,
480    208,209,210,211,212,213,214,247,
481    216,217,218,219,220,221,222,255,
482    0,62,0,0,1,0,0,0,
483    0,0,0,0,0,0,0,0,
484    32,0,0,0,1,0,0,0,
485    0,0,0,0,0,0,0,0,
486    0,0,0,0,0,0,255,3,
487    126,0,0,0,126,0,0,0,
488    0,0,0,0,0,0,0,0,
489    0,0,0,0,0,0,0,0,
490    0,0,0,0,0,0,255,3,
491    0,0,0,0,0,0,0,0,
492    0,0,0,0,0,0,12,2,
493    0,0,0,0,0,0,0,0,
494    0,0,0,0,0,0,0,0,
495    254,255,255,7,0,0,0,0,
496    0,0,0,0,0,0,0,0,
497    255,255,127,127,0,0,0,0,
498    0,0,0,0,0,0,0,0,
499    0,0,0,0,254,255,255,7,
500    0,0,0,0,0,4,32,4,
501    0,0,0,128,255,255,127,255,
502    0,0,0,0,0,0,255,3,
503    254,255,255,135,254,255,255,7,
504    0,0,0,0,0,4,44,6,
505    255,255,127,255,255,255,127,255,
506    0,0,0,0,254,255,255,255,
507    255,255,255,255,255,255,255,127,
508    0,0,0,0,254,255,255,255,
509    255,255,255,255,255,255,255,255,
510    0,2,0,0,255,255,255,255,
511    255,255,255,255,255,255,255,127,
512    0,0,0,0,255,255,255,255,
513    255,255,255,255,255,255,255,255,
514    0,0,0,0,254,255,0,252,
515    1,0,0,248,1,0,0,120,
516    0,0,0,0,254,255,255,255,
517    0,0,128,0,0,0,128,0,
518    255,255,255,255,0,0,0,0,
519    0,0,0,0,0,0,0,128,
520    255,255,255,255,0,0,0,0,
521    0,0,0,0,0,0,0,0,
522    128,0,0,0,0,0,0,0,
523    0,1,1,0,1,1,0,0,
524    0,0,0,0,0,0,0,0,
525    0,0,0,0,0,0,0,0,
526    1,0,0,0,128,0,0,0,
527    128,128,128,128,0,0,128,0,
528    28,28,28,28,28,28,28,28,
529    28,28,0,0,0,0,0,128,
530    0,26,26,26,26,26,26,18,
531    18,18,18,18,18,18,18,18,
532    18,18,18,18,18,18,18,18,
533    18,18,18,128,128,0,128,16,
534    0,26,26,26,26,26,26,18,
535    18,18,18,18,18,18,18,18,
536    18,18,18,18,18,18,18,18,
537    18,18,18,128,128,0,0,0,
538    0,0,0,0,0,1,0,0,
539    0,0,0,0,0,0,0,0,
540    0,0,0,0,0,0,0,0,
541    0,0,0,0,0,0,0,0,
542    1,0,0,0,0,0,0,0,
543    0,0,18,0,0,0,0,0,
544    0,0,20,20,0,18,0,0,
545    0,20,18,0,0,0,0,0,
546    18,18,18,18,18,18,18,18,
547    18,18,18,18,18,18,18,18,
548    18,18,18,18,18,18,18,0,
549    18,18,18,18,18,18,18,18,
550    18,18,18,18,18,18,18,18,
551    18,18,18,18,18,18,18,18,
552    18,18,18,18,18,18,18,0,
553    18,18,18,18,18,18,18,18
554    };
555    
556    
557    
558    
559    #ifndef HAVE_STRERROR
560    /*************************************************
561    *     Provide strerror() for non-ANSI libraries  *
562    *************************************************/
563    
564    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
565    in their libraries, but can provide the same facility by this simple
566    alternative function. */
567    
568    extern int   sys_nerr;
569    extern char *sys_errlist[];
570    
571    char *
572    strerror(int n)
573    {
574    if (n < 0 || n >= sys_nerr) return "unknown error number";
575    return sys_errlist[n];
576    }
577    #endif /* HAVE_STRERROR */
578    
579    
580    
581    
582  /*************************************************  /*************************************************
# Line 209  uschar *here = start; Line 609  uschar *here = start;
609    
610  for (;;)  for (;;)
611    {    {
612    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
613    
614    if (rlen > 1000)    if (rlen > 1000)
615      {      {
# Line 239  for (;;) Line 639  for (;;)
639      /* 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. */
640    
641        {        {
642        if (f == stdin) printf(prompt);        if (f == stdin) printf("%s", prompt);
643        if (fgets((char *)here, rlen,  f) == NULL)        if (fgets((char *)here, rlen,  f) == NULL)
644          return (here == start)? NULL : start;          return (here == start)? NULL : start;
645        }        }
# Line 729  return 0; Line 1129  return 0;
1129  *         Check newline indicator                *  *         Check newline indicator                *
1130  *************************************************/  *************************************************/
1131    
1132  /* 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
1133  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.  
1134    
1135  Arguments:  Arguments:
1136    p           points after the leading '<'    p           points after the leading '<'
# Line 779  printf("  -dfa     force DFA matching fo Line 1178  printf("  -dfa     force DFA matching fo
1178  #endif  #endif
1179  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
1180  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
1181           "  -M       find MATCH_LIMIT minimum for each subject\n"
1182         "  -m       output memory used information\n"         "  -m       output memory used information\n"
1183         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
1184  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 786  printf("  -p       use POSIX interface\n Line 1186  printf("  -p       use POSIX interface\n
1186  #endif  #endif
1187  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
1188  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
1189  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied\n"
1190         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
1191  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1192  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 808  int main(int argc, char **argv) Line 1208  int main(int argc, char **argv)
1208  FILE *infile = stdin;  FILE *infile = stdin;
1209  int options = 0;  int options = 0;
1210  int study_options = 0;  int study_options = 0;
1211    int default_find_match_limit = FALSE;
1212  int op = 1;  int op = 1;
1213  int timeit = 0;  int timeit = 0;
1214  int timeitm = 0;  int timeitm = 0;
1215  int showinfo = 0;  int showinfo = 0;
1216  int showstore = 0;  int showstore = 0;
1217    int force_study = 0;
1218  int quiet = 0;  int quiet = 0;
1219  int size_offsets = 45;  int size_offsets = 45;
1220  int size_offsets_max;  int size_offsets_max;
# Line 861  while (argc > 1 && argv[op][0] == '-') Line 1263  while (argc > 1 && argv[op][0] == '-')
1263    {    {
1264    unsigned char *endptr;    unsigned char *endptr;
1265    
1266    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1267      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 1;
1268    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1269    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1270    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1271    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1272      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1273  #if !defined NODFA  #if !defined NODFA
1274    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1275  #endif  #endif
# Line 919  while (argc > 1 && argv[op][0] == '-') Line 1322  while (argc > 1 && argv[op][0] == '-')
1322    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1323      {      {
1324      int rc;      int rc;
1325        unsigned long int lrc;
1326      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1327      printf("Compiled with\n");      printf("Compiled with\n");
1328      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 926  while (argc > 1 && argv[op][0] == '-') Line 1330  while (argc > 1 && argv[op][0] == '-')
1330      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1331      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1332      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1333      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1334        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1335        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1336          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1337        (rc == -2)? "ANYCRLF" :        (rc == -2)? "ANYCRLF" :
1338        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1339      (void)pcre_config(PCRE_CONFIG_BSR, &rc);      (void)pcre_config(PCRE_CONFIG_BSR, &rc);
# Line 937  while (argc > 1 && argv[op][0] == '-') Line 1343  while (argc > 1 && argv[op][0] == '-')
1343      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1344      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1345      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1346      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1347      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1348      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1349      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1350      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1351      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1352      goto EXIT;      goto EXIT;
# Line 1022  while (!done) Line 1428  while (!done)
1428  #endif  #endif
1429    
1430    const char *error;    const char *error;
1431      unsigned char *markptr;
1432    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1433    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1434    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1435    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1436    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1437      int do_mark = 0;
1438    int do_study = 0;    int do_study = 0;
1439      int no_force_study = 0;
1440    int do_debug = debug;    int do_debug = debug;
1441    int do_G = 0;    int do_G = 0;
1442    int do_g = 0;    int do_g = 0;
1443    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1444    int do_showrest = 0;    int do_showrest = 0;
1445      int do_showcaprest = 0;
1446    int do_flip = 0;    int do_flip = 0;
1447    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1448    
# Line 1094  while (!done) Line 1504  while (!done)
1504          }          }
1505        }        }
1506    
1507      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1508        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1509    
1510      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1102  while (!done) Line 1512  while (!done)
1512      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1513      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1514    
1515      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1516    
1517      if (true_study_size != 0)      if (true_study_size != 0)
1518        {        {
# Line 1144  while (!done) Line 1554  while (!done)
1554      }      }
1555    
1556    pp = p;    pp = p;
1557    poffset = p - buffer;    poffset = (int)(p - buffer);
1558    
1559    for(;;)    for(;;)
1560      {      {
# Line 1198  while (!done) Line 1608  while (!done)
1608        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1609        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1610    
1611        case '+': do_showrest = 1; break;        case '+':
1612          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1613          break;
1614    
1615        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1616        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
1617        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1208  while (!done) Line 1621  while (!done)
1621        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1622        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1623        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1624          case 'K': do_mark = 1; break;
1625        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1626        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1627    
# Line 1215  while (!done) Line 1629  while (!done)
1629        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1630  #endif  #endif
1631    
1632        case 'S': do_study = 1; break;        case 'S':
1633          if (do_study == 0) do_study = 1; else
1634            {
1635            do_study = 0;
1636            no_force_study = 1;
1637            }
1638          break;
1639    
1640        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1641          case 'W': options |= PCRE_UCP; break;
1642        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1643          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1644        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1645        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1646        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1647    
1648          case 'T':
1649          switch (*pp++)
1650            {
1651            case '0': tables = tables0; break;
1652            case '1': tables = tables1; break;
1653    
1654            case '\r':
1655            case '\n':
1656            case ' ':
1657            case 0:
1658            fprintf(outfile, "** Missing table number after /T\n");
1659            goto SKIP_DATA;
1660    
1661            default:
1662            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1663            goto SKIP_DATA;
1664            }
1665          break;
1666    
1667        case 'L':        case 'L':
1668        ppp = pp;        ppp = pp;
1669        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1247  while (!done) Line 1689  while (!done)
1689    
1690        case '<':        case '<':
1691          {          {
1692          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1693          if (x == 0) goto SKIP_DATA;            {
1694          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1695          while (*pp++ != '>');            pp += 3;
1696              }
1697            else
1698              {
1699              int x = check_newline(pp, outfile);
1700              if (x == 0) goto SKIP_DATA;
1701              options |= x;
1702              while (*pp++ != '>');
1703              }
1704          }          }
1705        break;        break;
1706    
# Line 1280  while (!done) Line 1730  while (!done)
1730      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1731      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1732      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1733        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1734        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1735    
1736      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1737    
# Line 1300  while (!done) Line 1752  while (!done)
1752  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1753    
1754      {      {
1755        unsigned long int get_options;
1756    
1757      if (timeit > 0)      if (timeit > 0)
1758        {        {
1759        register int i;        register int i;
# Line 1343  while (!done) Line 1797  while (!done)
1797        goto CONTINUE;        goto CONTINUE;
1798        }        }
1799    
1800      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1801      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
1802      returns only limited data. Check that it agrees with the newer one. */      lines. */
1803    
1804        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1805        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1806    
1807        /* Print information if required. There are now two info-returning
1808        functions. The old one has a limited interface and returns only limited
1809        data. Check that it agrees with the newer one. */
1810    
1811      if (log_store)      if (log_store)
1812        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1359  while (!done) Line 1820  while (!done)
1820      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1821      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1822    
1823      /* 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
1824      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1825        suppresses the effect of /S (used for a few test patterns where studying is
1826        never sensible). */
1827    
1828      if (do_study)      if (do_study || (force_study && !no_force_study))
1829        {        {
1830        if (timeit > 0)        if (timeit > 0)
1831          {          {
# Line 1384  while (!done) Line 1847  while (!done)
1847          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1848        }        }
1849    
1850        /* If /K was present, we set up for handling MARK data. */
1851    
1852        if (do_mark)
1853          {
1854          if (extra == NULL)
1855            {
1856            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1857            extra->flags = 0;
1858            }
1859          extra->mark = &markptr;
1860          extra->flags |= PCRE_EXTRA_MARK;
1861          }
1862    
1863      /* 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
1864      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
1865      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 1416  while (!done) Line 1892  while (!done)
1892          {          {
1893          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1894          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1895          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1896            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1897          }          }
1898        }        }
1899    
# Line 1430  while (!done) Line 1907  while (!done)
1907        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1908        }        }
1909    
1910        /* We already have the options in get_options (see above) */
1911    
1912      if (do_showinfo)      if (do_showinfo)
1913        {        {
1914        unsigned long int get_options, all_options;        unsigned long int all_options;
1915  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1916        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1917  #endif  #endif
# Line 1441  while (!done) Line 1920  while (!done)
1920        int nameentrysize, namecount;        int nameentrysize, namecount;
1921        const uschar *nametable;        const uschar *nametable;
1922    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1923        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1924        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1925        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1501  while (!done) Line 1979  while (!done)
1979        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
1980    
1981        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1982          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",
1983            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1984            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1985            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
# Line 1515  while (!done) Line 1993  while (!done)
1993            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1994            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1995            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1996              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
1997            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1998              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
1999            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2000    
2001        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
# Line 1583  while (!done) Line 2063  while (!done)
2063        /* 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
2064        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2065        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
2066        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2067          information unless -i or -d was also present. This means that, except
2068          when auto-callouts are involved, the output from runs with and without
2069          -s should be identical. */
2070    
2071        if (do_study)        if (do_study || (force_study && showinfo && !no_force_study))
2072          {          {
2073          if (extra == NULL)          if (extra == NULL)
2074            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2075          else          else
2076            {            {
2077            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2078            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2079    
2080              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2081              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2082    
2083              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2084            if (start_bits == NULL)            if (start_bits == NULL)
2085              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2086            else            else
2087              {              {
2088              int i;              int i;
# Line 1659  while (!done) Line 2146  while (!done)
2146            }            }
2147          else          else
2148            {            {
2149            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2150    
2151              /* If there is study data, write it, but verify the writing only
2152              if the studying was requested by /S, not just by -s. */
2153    
2154            if (extra != NULL)            if (extra != NULL)
2155              {              {
2156              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1669  while (!done) Line 2160  while (!done)
2160                  strerror(errno));                  strerror(errno));
2161                }                }
2162              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2163              }              }
2164            }            }
2165          fclose(f);          fclose(f);
# Line 1677  while (!done) Line 2167  while (!done)
2167    
2168        new_free(re);        new_free(re);
2169        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2170        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2171            {
2172            new_free((void *)tables);
2173            setlocale(LC_CTYPE, "C");
2174            locale_set = 0;
2175            }
2176        continue;  /* With next regex */        continue;  /* With next regex */
2177        }        }
2178      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1694  while (!done) Line 2189  while (!done)
2189      int callout_data_set = 0;      int callout_data_set = 0;
2190      int count, c;      int count, c;
2191      int copystrings = 0;      int copystrings = 0;
2192      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2193      int getstrings = 0;      int getstrings = 0;
2194      int getlist = 0;      int getlist = 0;
2195      int gmatched = 0;      int gmatched = 0;
2196      int start_offset = 0;      int start_offset = 0;
2197        int start_offset_sign = 1;
2198      int g_notempty = 0;      int g_notempty = 0;
2199      int use_dfa = 0;      int use_dfa = 0;
2200    
# Line 1726  while (!done) Line 2222  while (!done)
2222        {        {
2223        if (extend_inputline(infile, buffer + len, "data> ") == NULL)        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
2224          {          {
2225          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2226              {
2227              fprintf(outfile, "\n");
2228              break;
2229              }
2230          done = 1;          done = 1;
2231          goto CONTINUE;          goto CONTINUE;
2232          }          }
# Line 1792  while (!done) Line 2292  while (!done)
2292              {              {
2293              unsigned char buff8[8];              unsigned char buff8[8];
2294              int ii, utn;              int ii, utn;
2295              utn = ord2utf8(c, buff8);              if (use_utf8)
2296              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2297              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2298                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2299                  c = buff8[ii];   /* Last byte */
2300                  }
2301                else
2302                 {
2303                 if (c > 255)
2304                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2305                     "UTF-8 mode is not enabled.\n"
2306                     "** Truncation will probably give the wrong result.\n", c);
2307                 }
2308              p = pt + 1;              p = pt + 1;
2309              break;              break;
2310              }              }
# Line 1817  while (!done) Line 2327  while (!done)
2327          continue;          continue;
2328    
2329          case '>':          case '>':
2330            if (*p == '-')
2331              {
2332              start_offset_sign = -1;
2333              p++;
2334              }
2335          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2336            start_offset *= start_offset_sign;
2337          continue;          continue;
2338    
2339          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1890  while (!done) Line 2406  while (!done)
2406  #endif  #endif
2407            use_dfa = 1;            use_dfa = 1;
2408          continue;          continue;
2409    #endif
2410    
2411    #if !defined NODFA
2412          case 'F':          case 'F':
2413          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2414          continue;          continue;
# Line 1924  while (!done) Line 2442  while (!done)
2442          continue;          continue;
2443    
2444          case 'N':          case 'N':
2445          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2446              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2447            else
2448              options |= PCRE_NOTEMPTY;
2449          continue;          continue;
2450    
2451          case 'O':          case 'O':
# Line 1947  while (!done) Line 2468  while (!done)
2468          continue;          continue;
2469    
2470          case 'P':          case 'P':
2471          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2472              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2473          continue;          continue;
2474    
2475          case 'Q':          case 'Q':
# Line 1982  while (!done) Line 2504  while (!done)
2504          show_malloc = 1;          show_malloc = 1;
2505          continue;          continue;
2506    
2507            case 'Y':
2508            options |= PCRE_NO_START_OPTIMIZE;
2509            continue;
2510    
2511          case 'Z':          case 'Z':
2512          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2513          continue;          continue;
# Line 2002  while (!done) Line 2528  while (!done)
2528        *q++ = c;        *q++ = c;
2529        }        }
2530      *q = 0;      *q = 0;
2531      len = q - dbuffer;      len = (int)(q - dbuffer);
2532    
2533        /* Move the data to the end of the buffer so that a read over the end of
2534        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2535        we are using the POSIX interface, we must include the terminating zero. */
2536    
2537    #if !defined NOPOSIX
2538        if (posix || do_posix)
2539          {
2540          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2541          bptr += buffer_size - len - 1;
2542          }
2543        else
2544    #endif
2545          {
2546          memmove(bptr + buffer_size - len, bptr, len);
2547          bptr += buffer_size - len;
2548          }
2549    
2550      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2551        {        {
# Line 2023  while (!done) Line 2566  while (!done)
2566          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2567        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2568        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2569          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2570    
2571        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2572    
# Line 2047  while (!done) Line 2591  while (!done)
2591              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2592                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2593              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2594              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2595                {                {
2596                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2597                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2598                  outfile);                  outfile);
2599                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 2067  while (!done) Line 2611  while (!done)
2611    
2612      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2613        {        {
2614          markptr = NULL;
2615    
2616        if (timeitm > 0)        if (timeitm > 0)
2617          {          {
2618          register int i;          register int i;
# Line 2078  while (!done) Line 2624  while (!done)
2624            {            {
2625            int workspace[1000];            int workspace[1000];
2626            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2627              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2628                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2629                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2630            }            }
# Line 2141  while (!done) Line 2687  while (!done)
2687        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2688          {          {
2689          int workspace[1000];          int workspace[1000];
2690          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2691            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2692            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2693          if (count == 0)          if (count == 0)
# Line 2199  while (!done) Line 2745  while (!done)
2745              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2746                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2747              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2748              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2749                {                {
2750                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2751                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2752                  fprintf(outfile, " 0+ ");                  outfile);
2753                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2754                }                }
2755              }              }
2756            }            }
2757    
2758            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2759    
2760          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2761            {            {
2762            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2295  while (!done) Line 2840  while (!done)
2840    
2841        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2842          {          {
2843          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2844  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2845          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2846            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2847              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2848  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2849                outfile);
2850              }
2851          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2852          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2853          }          }
# Line 2310  while (!done) Line 2857  while (!done)
2857        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
2858        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2859    
2860        Complication arises in the case when the newline option is "any" or        Complication arises in the case when the newline convention is "any",
2861        "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
2862        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,
2863        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
2864        Fudge the offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2865          newline setting in the pattern; if none was set, use pcre_config() to
2866          find the default.
2867    
2868        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
2869        character, not one byte. */        character, not one byte. */
# Line 2330  while (!done) Line 2879  while (!done)
2879              {              {
2880              int d;              int d;
2881              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2882              obits = (d == '\r')? PCRE_NEWLINE_CR :              /* Note that these values are always the ASCII ones, even in
2883                      (d == '\n')? PCRE_NEWLINE_LF :              EBCDIC environments. CR = 13, NL = 10. */
2884                      (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :              obits = (d == 13)? PCRE_NEWLINE_CR :
2885                        (d == 10)? PCRE_NEWLINE_LF :
2886                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2887                      (d == -2)? PCRE_NEWLINE_ANYCRLF :                      (d == -2)? PCRE_NEWLINE_ANYCRLF :
2888                      (d == -1)? PCRE_NEWLINE_ANY : 0;                      (d == -1)? PCRE_NEWLINE_ANY : 0;
2889              }              }
2890            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2891                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2892                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2893                &&                &&
2894                start_offset < len - 1 &&                start_offset < len - 1 &&
# Line 2347  while (!done) Line 2899  while (!done)
2899              {              {
2900              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2901                {                {
2902                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2903                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2904                }                }
2905              }              }
2906            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
2907            }            }
2908          else          else
2909            {            {
2910            if (count == PCRE_ERROR_NOMATCH)            switch(count)
2911              {              {
2912              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
2913                if (gmatched == 0)
2914                  {
2915                  if (markptr == NULL) fprintf(outfile, "No match\n");
2916                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2917                  }
2918                break;
2919    
2920                case PCRE_ERROR_BADUTF8:
2921                case PCRE_ERROR_SHORTUTF8:
2922                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
2923                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
2924                if (use_size_offsets >= 2)
2925                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
2926                    use_offsets[1]);
2927                fprintf(outfile, "\n");
2928                break;
2929    
2930                default:
2931                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
2932                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
2933                else
2934                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
2935                break;
2936              }              }
2937            else fprintf(outfile, "Error %d\n", count);  
2938            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
2939            }            }
2940          }          }
# Line 2371  while (!done) Line 2944  while (!done)
2944        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2945    
2946        /* 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
2947        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
2948        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
2949        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
2950        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
2951        character. */        character. */
2952    
# Line 2382  while (!done) Line 2955  while (!done)
2955        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2956          {          {
2957          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2958          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2959          }          }
2960    
2961        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2409  while (!done) Line 2982  while (!done)
2982    
2983    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
2984    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
2985    if (tables != NULL)    if (locale_set)
2986      {      {
2987      new_free((void *)tables);      new_free((void *)tables);
2988      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
Legend:
Removed from v.289  
changed lines
  Added in v.616
  ViewVC Help
Powered by ViewVC 1.1.5