/[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 107 by ph10, Wed Mar 7 11:02:28 2007 UTC revision 612 by ph10, Sat Jul 2 15:20:59 2011 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 44  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51    #ifdef SUPPORT_LIBREADLINE
52    #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57    #endif
58    
59    
60  /* A number of things vary for Windows builds. Originally, pcretest opened its  /* A number of things vary for Windows builds. Originally, pcretest opened its
61  input and output without "b"; then I was told that "b" was needed in some  input and output without "b"; then I was told that "b" was needed in some
# Line 59  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 67  input mode under Windows. */ Line 95  input mode under Windows. */
95  #endif  #endif
96    
97    
98  #define PCRE_SPY        /* For Win32 build, import data, not export */  /* We have to include pcre_internal.h because we need the internal info for
99    displaying the results of pcre_study() and we also need to know about the
100  /* We include pcre_internal.h because we need the internal info for displaying  internal macros, structures, and other internal data values; pcretest has
101  the results of pcre_study() and we also need to know about the internal  "inside information" compared to a program that strictly follows the PCRE API.
102  macros, structures, and other internal data values; pcretest has "inside  
103  information" compared to a program that strictly follows the PCRE API. */  Although pcre_internal.h does itself include pcre.h, we explicitly include it
104    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
105    appropriately for an application, not for building PCRE. */
106    
107    #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 87  symbols to prevent clashes. */ Line 119  symbols to prevent clashes. */
119  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
120  #define _pcre_utt              utt  #define _pcre_utt              utt
121  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
122    #define _pcre_utt_names        utt_names
123  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
124    
125  #include "pcre_tables.c"  #include "pcre_tables.c"
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 149  static int callout_count; Line 183  static int callout_count;
183  static int callout_extra;  static int callout_extra;
184  static int callout_fail_count;  static int callout_fail_count;
185  static int callout_fail_id;  static int callout_fail_id;
186    static int debug_lengths;
187  static int first_callout;  static int first_callout;
188  static int locale_set = 0;  static int locale_set = 0;
189  static int show_malloc;  static int show_malloc;
# Line 162  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 180  optimal way of handling this, but hey, t Line 595  optimal way of handling this, but hey, t
595  Arguments:  Arguments:
596    f            the file to read    f            the file to read
597    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
598      prompt       for stdin or readline()
599    
600  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
601                 could be a copy of start, or could be moved                 could be a copy of start, or could be moved
# Line 187  Returns:       pointer to the start of n Line 603  Returns:       pointer to the start of n
603  */  */
604    
605  static uschar *  static uschar *
606  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
607  {  {
608  uschar *here = start;  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      {      {
616      int dlen;      int dlen;
617      if (fgets((char *)here, rlen,  f) == NULL)  
618        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
619        input is a terminal. Note that readline() removes the trailing newline, so
620        we must put it back again, to be compatible with fgets(). */
621    
622    #ifdef SUPPORT_LIBREADLINE
623        if (isatty(fileno(f)))
624          {
625          size_t len;
626          char *s = readline(prompt);
627          if (s == NULL) return (here == start)? NULL : start;
628          len = strlen(s);
629          if (len > 0) add_history(s);
630          if (len > rlen - 1) len = rlen - 1;
631          memcpy(here, s, len);
632          here[len] = '\n';
633          here[len+1] = 0;
634          free(s);
635          }
636        else
637    #endif
638    
639        /* Read the next line by normal means, prompting if the file is stdin. */
640    
641          {
642          if (f == stdin) printf("%s", prompt);
643          if (fgets((char *)here, rlen,  f) == NULL)
644            return (here == start)? NULL : start;
645          }
646    
647      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
648      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
649      here += dlen;      here += dlen;
# Line 656  return count; Line 1100  return count;
1100    
1101    
1102  /*************************************************  /*************************************************
1103    *         Case-independent strncmp() function    *
1104    *************************************************/
1105    
1106    /*
1107    Arguments:
1108      s         first string
1109      t         second string
1110      n         number of characters to compare
1111    
1112    Returns:    < 0, = 0, or > 0, according to the comparison
1113    */
1114    
1115    static int
1116    strncmpic(uschar *s, uschar *t, int n)
1117    {
1118    while (n--)
1119      {
1120      int c = tolower(*s++) - tolower(*t++);
1121      if (c) return c;
1122      }
1123    return 0;
1124    }
1125    
1126    
1127    
1128    /*************************************************
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, or ANY. Print a message and return 0 if there is no match.  a message and return 0 if there is no match.
1134    
1135  Arguments:  Arguments:
1136    p           points after the leading '<'    p           points after the leading '<'
# Line 672  Returns:      appropriate PCRE_NEWLINE_x Line 1142  Returns:      appropriate PCRE_NEWLINE_x
1142  static int  static int
1143  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
1144  {  {
1145  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1146  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1147  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1148  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1149    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1150    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1151    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1152  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
1153  return 0;  return 0;
1154  }  }
# Line 689  return 0; Line 1162  return 0;
1162  static void  static void
1163  usage(void)  usage(void)
1164  {  {
1165  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1166    printf("Input and output default to stdin and stdout.\n");
1167    #ifdef SUPPORT_LIBREADLINE
1168    printf("If input is a terminal, readline() is used to read from it.\n");
1169    #else
1170    printf("This version of pcretest is not linked with readline().\n");
1171    #endif
1172    printf("\nOptions:\n");
1173  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
1174  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
1175  printf("  -d       debug: show compiled code and information (-b and -i)\n");  printf("  -d       debug: show compiled code and information (-b and -i)\n");
# Line 698  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 705  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 727  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 780  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 838  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 845  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" :
1338        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1339        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1340        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1341                                         "all Unicode newlines");
1342      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
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      exit(0);      goto EXIT;
1353      }      }
1354    else if (strcmp(argv[op], "-help") == 0 ||    else if (strcmp(argv[op], "-help") == 0 ||
1355             strcmp(argv[op], "--help") == 0)             strcmp(argv[op], "--help") == 0)
# Line 884  offsets = (int *)malloc(size_offsets_max Line 1375  offsets = (int *)malloc(size_offsets_max
1375  if (offsets == NULL)  if (offsets == NULL)
1376    {    {
1377    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1378      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1379    yield = 1;    yield = 1;
1380    goto EXIT;    goto EXIT;
1381    }    }
# Line 937  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;
# Line 952  while (!done) Line 1446  while (!done)
1446    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1447    
1448    use_utf8 = 0;    use_utf8 = 0;
1449      debug_lengths = 1;
1450    
1451    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1452    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1453    fflush(outfile);    fflush(outfile);
1454    
# Line 1009  while (!done) Line 1503  while (!done)
1503          }          }
1504        }        }
1505    
1506      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1507        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1508    
1509      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1017  while (!done) Line 1511  while (!done)
1511      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1512      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1513    
1514      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1515    
1516      if (true_study_size != 0)      if (true_study_size != 0)
1517        {        {
# Line 1054  while (!done) Line 1548  while (!done)
1548    
1549    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1550      {      {
1551      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1552      goto SKIP_DATA;      goto SKIP_DATA;
1553      }      }
1554    
1555    pp = p;    pp = p;
1556    poffset = p - buffer;    poffset = (int)(p - buffer);
1557    
1558    for(;;)    for(;;)
1559      {      {
# Line 1070  while (!done) Line 1564  while (!done)
1564        pp++;        pp++;
1565        }        }
1566      if (*pp != 0) break;      if (*pp != 0) break;
1567      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1568        {        {
1569        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1570        done = 1;        done = 1;
# Line 1124  while (!done) Line 1617  while (!done)
1617        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1618        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1619        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1620          case 'K': do_mark = 1; break;
1621        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1622        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1623    
# Line 1131  while (!done) Line 1625  while (!done)
1625        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1626  #endif  #endif
1627    
1628        case 'S': do_study = 1; break;        case 'S':
1629          if (do_study == 0) do_study = 1; else
1630            {
1631            do_study = 0;
1632            no_force_study = 1;
1633            }
1634          break;
1635    
1636        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1637          case 'W': options |= PCRE_UCP; break;
1638        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1639          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1640          case 'Z': debug_lengths = 0; break;
1641        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1642        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1643    
1644          case 'T':
1645          switch (*pp++)
1646            {
1647            case '0': tables = tables0; break;
1648            case '1': tables = tables1; break;
1649    
1650            case '\r':
1651            case '\n':
1652            case ' ':
1653            case 0:
1654            fprintf(outfile, "** Missing table number after /T\n");
1655            goto SKIP_DATA;
1656    
1657            default:
1658            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1659            goto SKIP_DATA;
1660            }
1661          break;
1662    
1663        case 'L':        case 'L':
1664        ppp = pp;        ppp = pp;
1665        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1162  while (!done) Line 1685  while (!done)
1685    
1686        case '<':        case '<':
1687          {          {
1688          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1689          if (x == 0) goto SKIP_DATA;            {
1690          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1691          while (*pp++ != '>');            pp += 3;
1692              }
1693            else
1694              {
1695              int x = check_newline(pp, outfile);
1696              if (x == 0) goto SKIP_DATA;
1697              options |= x;
1698              while (*pp++ != '>');
1699              }
1700          }          }
1701        break;        break;
1702    
# Line 1195  while (!done) Line 1726  while (!done)
1726      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1727      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1728      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1729        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1730        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1731    
1732      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1733    
# Line 1215  while (!done) Line 1748  while (!done)
1748  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1749    
1750      {      {
1751        unsigned long int get_options;
1752    
1753      if (timeit > 0)      if (timeit > 0)
1754        {        {
1755        register int i;        register int i;
# Line 1244  while (!done) Line 1779  while (!done)
1779          {          {
1780          for (;;)          for (;;)
1781            {            {
1782            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1783              {              {
1784              done = 1;              done = 1;
1785              goto CONTINUE;              goto CONTINUE;
# Line 1258  while (!done) Line 1793  while (!done)
1793        goto CONTINUE;        goto CONTINUE;
1794        }        }
1795    
1796      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1797      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
1798      returns only limited data. Check that it agrees with the newer one. */      lines. */
1799    
1800        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1801        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1802    
1803        /* Print information if required. There are now two info-returning
1804        functions. The old one has a limited interface and returns only limited
1805        data. Check that it agrees with the newer one. */
1806    
1807      if (log_store)      if (log_store)
1808        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1274  while (!done) Line 1816  while (!done)
1816      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1817      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1818    
1819      /* 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
1820      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1821        suppresses the effect of /S (used for a few test patterns where studying is
1822        never sensible). */
1823    
1824      if (do_study)      if (do_study || (force_study && !no_force_study))
1825        {        {
1826        if (timeit > 0)        if (timeit > 0)
1827          {          {
# Line 1299  while (!done) Line 1843  while (!done)
1843          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1844        }        }
1845    
1846        /* If /K was present, we set up for handling MARK data. */
1847    
1848        if (do_mark)
1849          {
1850          if (extra == NULL)
1851            {
1852            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1853            extra->flags = 0;
1854            }
1855          extra->mark = &markptr;
1856          extra->flags |= PCRE_EXTRA_MARK;
1857          }
1858    
1859      /* 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
1860      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
1861      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 1307  while (!done) Line 1864  while (!done)
1864      if (do_flip)      if (do_flip)
1865        {        {
1866        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1867        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1868            byteflip(rre->magic_number, sizeof(rre->magic_number));
1869        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1870        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1871        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1872        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1873        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1874        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1875        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1876          rre->first_byte =
1877            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1878          rre->req_byte =
1879            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1880          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1881          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1882        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1883          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1884        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1885            sizeof(rre->name_count));
1886    
1887        if (extra != NULL)        if (extra != NULL)
1888          {          {
1889          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1890          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1891          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1892            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1893          }          }
1894        }        }
1895    
# Line 1335  while (!done) Line 1900  while (!done)
1900      if (do_debug)      if (do_debug)
1901        {        {
1902        fprintf(outfile, "------------------------------------------------------------------\n");        fprintf(outfile, "------------------------------------------------------------------\n");
1903        pcre_printint(re, outfile);        pcre_printint(re, outfile, debug_lengths);
1904        }        }
1905    
1906        /* We already have the options in get_options (see above) */
1907    
1908      if (do_showinfo)      if (do_showinfo)
1909        {        {
1910        unsigned long int get_options, all_options;        unsigned long int all_options;
1911  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1912        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1913  #endif  #endif
1914        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1915            hascrorlf;
1916        int nameentrysize, namecount;        int nameentrysize, namecount;
1917        const uschar *nametable;        const uschar *nametable;
1918    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1919        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1920        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1921        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1357  while (!done) Line 1924  while (!done)
1924        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1925        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1926        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1927          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1928          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1929          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1930    
1931  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1932        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1398  while (!done) Line 1968  while (!done)
1968            }            }
1969          }          }
1970    
1971        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1972        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1973    
1974        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1975        if (do_flip)        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
         {  
         all_options = byteflip(all_options, sizeof(all_options));  
          }  
   
       if ((all_options & PCRE_NOPARTIAL) != 0)  
         fprintf(outfile, "Partial matching not supported\n");  
1976    
1977        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1978          else fprintf(outfile, "Options:%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",
1979            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1980            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1981            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1982            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1983            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1984            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1985              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1986              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1987            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1988            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1989            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1990            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1991            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1992              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
1993            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1994              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
1995            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1996    
1997          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1998    
1999        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
2000          {          {
2001          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1440  while (!done) Line 2010  while (!done)
2010          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
2011          break;          break;
2012    
2013            case PCRE_NEWLINE_ANYCRLF:
2014            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
2015            break;
2016    
2017          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
2018          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
2019          break;          break;
# Line 1485  while (!done) Line 2059  while (!done)
2059        /* 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
2060        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2061        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
2062        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2063          information unless -i or -d was also present. This means that, except
2064          when auto-callouts are involved, the output from runs with and without
2065          -s should be identical. */
2066    
2067        if (do_study)        if (do_study || (force_study && showinfo && !no_force_study))
2068          {          {
2069          if (extra == NULL)          if (extra == NULL)
2070            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2071          else          else
2072            {            {
2073            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2074            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2075    
2076              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2077              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2078    
2079              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2080            if (start_bits == NULL)            if (start_bits == NULL)
2081              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2082            else            else
2083              {              {
2084              int i;              int i;
# Line 1544  while (!done) Line 2125  while (!done)
2125        else        else
2126          {          {
2127          uschar sbuf[8];          uschar sbuf[8];
2128          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2129          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2130          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2131          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2132    
2133          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2134          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2135          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2136          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2137    
2138          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2139              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1561  while (!done) Line 2142  while (!done)
2142            }            }
2143          else          else
2144            {            {
2145            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2146    
2147              /* If there is study data, write it, but verify the writing only
2148              if the studying was requested by /S, not just by -s. */
2149    
2150            if (extra != NULL)            if (extra != NULL)
2151              {              {
2152              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1571  while (!done) Line 2156  while (!done)
2156                  strerror(errno));                  strerror(errno));
2157                }                }
2158              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2159              }              }
2160            }            }
2161          fclose(f);          fclose(f);
# Line 1579  while (!done) Line 2163  while (!done)
2163    
2164        new_free(re);        new_free(re);
2165        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2166        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2167            {
2168            new_free((void *)tables);
2169            setlocale(LC_CTYPE, "C");
2170            locale_set = 0;
2171            }
2172        continue;  /* With next regex */        continue;  /* With next regex */
2173        }        }
2174      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1589  while (!done) Line 2178  while (!done)
2178    for (;;)    for (;;)
2179      {      {
2180      uschar *q;      uschar *q;
2181      uschar *bptr = dbuffer;      uschar *bptr;
2182      int *use_offsets = offsets;      int *use_offsets = offsets;
2183      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2184      int callout_data = 0;      int callout_data = 0;
2185      int callout_data_set = 0;      int callout_data_set = 0;
2186      int count, c;      int count, c;
2187      int copystrings = 0;      int copystrings = 0;
2188      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2189      int getstrings = 0;      int getstrings = 0;
2190      int getlist = 0;      int getlist = 0;
2191      int gmatched = 0;      int gmatched = 0;
2192      int start_offset = 0;      int start_offset = 0;
2193        int start_offset_sign = 1;
2194      int g_notempty = 0;      int g_notempty = 0;
2195      int use_dfa = 0;      int use_dfa = 0;
2196    
# Line 1626  while (!done) Line 2216  while (!done)
2216      len = 0;      len = 0;
2217      for (;;)      for (;;)
2218        {        {
2219        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2220          {          {
2221          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2222              {
2223              fprintf(outfile, "\n");
2224              break;
2225              }
2226          done = 1;          done = 1;
2227          goto CONTINUE;          goto CONTINUE;
2228          }          }
# Line 1645  while (!done) Line 2238  while (!done)
2238      p = buffer;      p = buffer;
2239      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2240    
2241      q = dbuffer;      bptr = q = dbuffer;
2242      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2243        {        {
2244        int i = 0;        int i = 0;
# Line 1695  while (!done) Line 2288  while (!done)
2288              {              {
2289              unsigned char buff8[8];              unsigned char buff8[8];
2290              int ii, utn;              int ii, utn;
2291              utn = ord2utf8(c, buff8);              if (use_utf8)
2292              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2293              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2294                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2295                  c = buff8[ii];   /* Last byte */
2296                  }
2297                else
2298                 {
2299                 if (c > 255)
2300                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2301                     "UTF-8 mode is not enabled.\n"
2302                     "** Truncation will probably give the wrong result.\n", c);
2303                 }
2304              p = pt + 1;              p = pt + 1;
2305              break;              break;
2306              }              }
# Line 1720  while (!done) Line 2323  while (!done)
2323          continue;          continue;
2324    
2325          case '>':          case '>':
2326            if (*p == '-')
2327              {
2328              start_offset_sign = -1;
2329              p++;
2330              }
2331          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2332            start_offset *= start_offset_sign;
2333          continue;          continue;
2334    
2335          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1793  while (!done) Line 2402  while (!done)
2402  #endif  #endif
2403            use_dfa = 1;            use_dfa = 1;
2404          continue;          continue;
2405    #endif
2406    
2407    #if !defined NODFA
2408          case 'F':          case 'F':
2409          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2410          continue;          continue;
# Line 1827  while (!done) Line 2438  while (!done)
2438          continue;          continue;
2439    
2440          case 'N':          case 'N':
2441          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2442              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2443            else
2444              options |= PCRE_NOTEMPTY;
2445          continue;          continue;
2446    
2447          case 'O':          case 'O':
# Line 1840  while (!done) Line 2454  while (!done)
2454            if (offsets == NULL)            if (offsets == NULL)
2455              {              {
2456              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2457                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2458              yield = 1;              yield = 1;
2459              goto EXIT;              goto EXIT;
2460              }              }
# Line 1850  while (!done) Line 2464  while (!done)
2464          continue;          continue;
2465    
2466          case 'P':          case 'P':
2467          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2468              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2469          continue;          continue;
2470    
2471          case 'Q':          case 'Q':
# Line 1885  while (!done) Line 2500  while (!done)
2500          show_malloc = 1;          show_malloc = 1;
2501          continue;          continue;
2502    
2503            case 'Y':
2504            options |= PCRE_NO_START_OPTIMIZE;
2505            continue;
2506    
2507          case 'Z':          case 'Z':
2508          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2509          continue;          continue;
# Line 1905  while (!done) Line 2524  while (!done)
2524        *q++ = c;        *q++ = c;
2525        }        }
2526      *q = 0;      *q = 0;
2527      len = q - dbuffer;      len = (int)(q - dbuffer);
2528    
2529        /* Move the data to the end of the buffer so that a read over the end of
2530        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2531        we are using the POSIX interface, we must include the terminating zero. */
2532    
2533    #if !defined NOPOSIX
2534        if (posix || do_posix)
2535          {
2536          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2537          bptr += buffer_size - len - 1;
2538          }
2539        else
2540    #endif
2541          {
2542          memmove(bptr + buffer_size - len, bptr, len);
2543          bptr += buffer_size - len;
2544          }
2545    
2546      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2547        {        {
# Line 1926  while (!done) Line 2562  while (!done)
2562          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2563        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2564        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2565          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2566    
2567        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2568    
# Line 1970  while (!done) Line 2607  while (!done)
2607    
2608      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2609        {        {
2610          markptr = NULL;
2611    
2612        if (timeitm > 0)        if (timeitm > 0)
2613          {          {
2614          register int i;          register int i;
# Line 1981  while (!done) Line 2620  while (!done)
2620            {            {
2621            int workspace[1000];            int workspace[1000];
2622            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2623              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2624                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2625                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2626            }            }
# Line 2044  while (!done) Line 2683  while (!done)
2683        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2684          {          {
2685          int workspace[1000];          int workspace[1000];
2686          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2687            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2688            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2689          if (count == 0)          if (count == 0)
# Line 2115  while (!done) Line 2754  while (!done)
2754              }              }
2755            }            }
2756    
2757            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2758    
2759          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2760            {            {
2761            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2198  while (!done) Line 2839  while (!done)
2839    
2840        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2841          {          {
2842          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2843  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2844          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2845            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2846              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2847  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2848                outfile);
2849              }
2850          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2851          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2852          }          }
2853    
2854        /* Failed to match. If this is a /g or /G loop and we previously set        /* Failed to match. If this is a /g or /G loop and we previously set
2855        g_notempty after a null match, this is not necessarily the end.        g_notempty after a null match, this is not necessarily the end. We want
2856        We want to advance the start offset, and continue. In the case of UTF-8        to advance the start offset, and continue. We won't be at the end of the
2857        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2858        offset values to achieve this. We won't be at the end of the string -  
2859        that was checked before setting g_notempty. */        Complication arises in the case when the newline convention is "any",
2860          "crlf", or "anycrlf". If the previous match was at the end of a line
2861          terminated by CRLF, an advance of one character just passes the \r,
2862          whereas we should prefer the longer newline sequence, as does the code in
2863          pcre_exec(). Fudge the offset value to achieve this. We check for a
2864          newline setting in the pattern; if none was set, use pcre_config() to
2865          find the default.
2866    
2867          Otherwise, in the case of UTF-8 matching, the advance must be one
2868          character, not one byte. */
2869    
2870        else        else
2871          {          {
2872          if (g_notempty != 0)          if (g_notempty != 0)
2873            {            {
2874            int onechar = 1;            int onechar = 1;
2875              unsigned int obits = ((real_pcre *)re)->options;
2876            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2877            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2878                {
2879                int d;
2880                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2881                /* Note that these values are always the ASCII ones, even in
2882                EBCDIC environments. CR = 13, NL = 10. */
2883                obits = (d == 13)? PCRE_NEWLINE_CR :
2884                        (d == 10)? PCRE_NEWLINE_LF :
2885                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2886                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2887                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2888                }
2889              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2890                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2891                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2892                  &&
2893                  start_offset < len - 1 &&
2894                  bptr[start_offset] == '\r' &&
2895                  bptr[start_offset+1] == '\n')
2896                onechar++;
2897              else if (use_utf8)
2898              {              {
2899              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2900                {                {
2901                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2902                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2903                }                }
2904              }              }
2905            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
2906            }            }
2907          else          else
2908            {            {
2909            if (count == PCRE_ERROR_NOMATCH)            switch(count)
2910              {              {
2911              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
2912                if (gmatched == 0)
2913                  {
2914                  if (markptr == NULL) fprintf(outfile, "No match\n");
2915                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2916                  }
2917                break;
2918    
2919                case PCRE_ERROR_BADUTF8:
2920                case PCRE_ERROR_SHORTUTF8:
2921                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
2922                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
2923                if (use_size_offsets >= 2)
2924                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
2925                    use_offsets[1]);
2926                fprintf(outfile, "\n");
2927                break;
2928    
2929                default:
2930                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
2931                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
2932                else
2933                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
2934                break;
2935              }              }
2936            else fprintf(outfile, "Error %d\n", count);  
2937            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
2938            }            }
2939          }          }
# Line 2249  while (!done) Line 2943  while (!done)
2943        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2944    
2945        /* 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
2946        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
2947        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
2948        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
2949        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
2950        character. */        character. */
2951    
2952        g_notempty = 0;        g_notempty = 0;
2953    
2954        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2955          {          {
2956          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2957          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2958          }          }
2959    
2960        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2286  while (!done) Line 2981  while (!done)
2981    
2982    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
2983    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
2984    if (tables != NULL)    if (locale_set)
2985      {      {
2986      new_free((void *)tables);      new_free((void *)tables);
2987      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");

Legend:
Removed from v.107  
changed lines
  Added in v.612

  ViewVC Help
Powered by ViewVC 1.1.5