/[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 145 by ph10, Wed Apr 4 14:06:52 2007 UTC revision 604 by ph10, Thu Jun 2 19:04:54 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 79  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 90  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 152  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 165  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 183  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 190  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 659  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 675  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 692  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 701  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 730  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;
# Line 789  while (argc > 1 && argv[op][0] == '-') Line 1268  while (argc > 1 && argv[op][0] == '-')
1268    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1269    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1270    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1271      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1272  #if !defined NODFA  #if !defined NODFA
1273    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1274  #endif  #endif
# Line 841  while (argc > 1 && argv[op][0] == '-') Line 1321  while (argc > 1 && argv[op][0] == '-')
1321    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1322      {      {
1323      int rc;      int rc;
1324        unsigned long int lrc;
1325      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1326      printf("Compiled with\n");      printf("Compiled with\n");
1327      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 848  while (argc > 1 && argv[op][0] == '-') Line 1329  while (argc > 1 && argv[op][0] == '-')
1329      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1330      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1331      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1332      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1333        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1334        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1335          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1336          (rc == -2)? "ANYCRLF" :
1337        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1338        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1339        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1340                                         "all Unicode newlines");
1341      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1342      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1343      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1344      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1345      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1346      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1347      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1348      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1349      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1350      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1351      goto EXIT;      goto EXIT;
# Line 887  offsets = (int *)malloc(size_offsets_max Line 1374  offsets = (int *)malloc(size_offsets_max
1374  if (offsets == NULL)  if (offsets == NULL)
1375    {    {
1376    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1377      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1378    yield = 1;    yield = 1;
1379    goto EXIT;    goto EXIT;
1380    }    }
# Line 940  while (!done) Line 1427  while (!done)
1427  #endif  #endif
1428    
1429    const char *error;    const char *error;
1430      unsigned char *markptr;
1431    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1432    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1433    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1434    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1435    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1436      int do_mark = 0;
1437    int do_study = 0;    int do_study = 0;
1438    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1439    int do_G = 0;    int do_G = 0;
1440    int do_g = 0;    int do_g = 0;
1441    int do_showinfo = showinfo;    int do_showinfo = showinfo;
# Line 956  while (!done) Line 1444  while (!done)
1444    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1445    
1446    use_utf8 = 0;    use_utf8 = 0;
1447      debug_lengths = 1;
1448    
1449    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1450    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1451    fflush(outfile);    fflush(outfile);
1452    
# Line 1058  while (!done) Line 1546  while (!done)
1546    
1547    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1548      {      {
1549      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1550      goto SKIP_DATA;      goto SKIP_DATA;
1551      }      }
1552    
1553    pp = p;    pp = p;
1554    poffset = p - buffer;    poffset = (int)(p - buffer);
1555    
1556    for(;;)    for(;;)
1557      {      {
# Line 1074  while (!done) Line 1562  while (!done)
1562        pp++;        pp++;
1563        }        }
1564      if (*pp != 0) break;      if (*pp != 0) break;
1565      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1566        {        {
1567        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1568        done = 1;        done = 1;
# Line 1128  while (!done) Line 1615  while (!done)
1615        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1616        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1617        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1618          case 'K': do_mark = 1; break;
1619        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1620        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1621    
# Line 1137  while (!done) Line 1625  while (!done)
1625    
1626        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1627        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1628          case 'W': options |= PCRE_UCP; break;
1629        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1630          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1631        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1632        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1633        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1634    
1635          case 'T':
1636          switch (*pp++)
1637            {
1638            case '0': tables = tables0; break;
1639            case '1': tables = tables1; break;
1640    
1641            case '\r':
1642            case '\n':
1643            case ' ':
1644            case 0:
1645            fprintf(outfile, "** Missing table number after /T\n");
1646            goto SKIP_DATA;
1647    
1648            default:
1649            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1650            goto SKIP_DATA;
1651            }
1652          break;
1653    
1654        case 'L':        case 'L':
1655        ppp = pp;        ppp = pp;
1656        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1167  while (!done) Line 1676  while (!done)
1676    
1677        case '<':        case '<':
1678          {          {
1679          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1680          if (x == 0) goto SKIP_DATA;            {
1681          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1682          while (*pp++ != '>');            pp += 3;
1683              }
1684            else
1685              {
1686              int x = check_newline(pp, outfile);
1687              if (x == 0) goto SKIP_DATA;
1688              options |= x;
1689              while (*pp++ != '>');
1690              }
1691          }          }
1692        break;        break;
1693    
# Line 1200  while (!done) Line 1717  while (!done)
1717      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1718      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1719      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1720        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1721        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1722    
1723      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1724    
# Line 1220  while (!done) Line 1739  while (!done)
1739  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1740    
1741      {      {
1742        unsigned long int get_options;
1743    
1744      if (timeit > 0)      if (timeit > 0)
1745        {        {
1746        register int i;        register int i;
# Line 1249  while (!done) Line 1770  while (!done)
1770          {          {
1771          for (;;)          for (;;)
1772            {            {
1773            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1774              {              {
1775              done = 1;              done = 1;
1776              goto CONTINUE;              goto CONTINUE;
# Line 1263  while (!done) Line 1784  while (!done)
1784        goto CONTINUE;        goto CONTINUE;
1785        }        }
1786    
1787      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1788      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
1789      returns only limited data. Check that it agrees with the newer one. */      lines. */
1790    
1791        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1792        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1793    
1794        /* Print information if required. There are now two info-returning
1795        functions. The old one has a limited interface and returns only limited
1796        data. Check that it agrees with the newer one. */
1797    
1798      if (log_store)      if (log_store)
1799        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1304  while (!done) Line 1832  while (!done)
1832          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1833        }        }
1834    
1835        /* If /K was present, we set up for handling MARK data. */
1836    
1837        if (do_mark)
1838          {
1839          if (extra == NULL)
1840            {
1841            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1842            extra->flags = 0;
1843            }
1844          extra->mark = &markptr;
1845          extra->flags |= PCRE_EXTRA_MARK;
1846          }
1847    
1848      /* 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
1849      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
1850      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 1312  while (!done) Line 1853  while (!done)
1853      if (do_flip)      if (do_flip)
1854        {        {
1855        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1856        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1857            byteflip(rre->magic_number, sizeof(rre->magic_number));
1858        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1859        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1860        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1861        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1862        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1863        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1864        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1865          rre->first_byte =
1866            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1867          rre->req_byte =
1868            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1869          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1870          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1871        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1872          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1873        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1874            sizeof(rre->name_count));
1875    
1876        if (extra != NULL)        if (extra != NULL)
1877          {          {
1878          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1879          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1880          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1881            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1882          }          }
1883        }        }
1884    
# Line 1343  while (!done) Line 1892  while (!done)
1892        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1893        }        }
1894    
1895        /* We already have the options in get_options (see above) */
1896    
1897      if (do_showinfo)      if (do_showinfo)
1898        {        {
1899        unsigned long int get_options, all_options;        unsigned long int all_options;
1900  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1901        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1902  #endif  #endif
1903        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1904            hascrorlf;
1905        int nameentrysize, namecount;        int nameentrysize, namecount;
1906        const uschar *nametable;        const uschar *nametable;
1907    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1908        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1909        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1910        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1362  while (!done) Line 1913  while (!done)
1913        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1914        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1915        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1916          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1917          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1918          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1919    
1920  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1921        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1403  while (!done) Line 1957  while (!done)
1957            }            }
1958          }          }
1959    
1960        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1961        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1962    
1963        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1964        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");  
1965    
1966        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1967          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",
1968            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1969            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1970            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1971            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1972            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1973            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1974              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1975              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1976            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1977            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1978            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1979            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1980            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1981              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
1982            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1983              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
1984            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1985    
1986          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1987    
1988        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
1989          {          {
1990          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1445  while (!done) Line 1999  while (!done)
1999          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
2000          break;          break;
2001    
2002            case PCRE_NEWLINE_ANYCRLF:
2003            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
2004            break;
2005    
2006          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
2007          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
2008          break;          break;
# Line 1499  while (!done) Line 2057  while (!done)
2057          else          else
2058            {            {
2059            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2060            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2061    
2062              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2063              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2064    
2065              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2066            if (start_bits == NULL)            if (start_bits == NULL)
2067              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2068            else            else
2069              {              {
2070              int i;              int i;
# Line 1549  while (!done) Line 2111  while (!done)
2111        else        else
2112          {          {
2113          uschar sbuf[8];          uschar sbuf[8];
2114          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2115          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2116          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2117          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2118    
2119          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2120          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2121          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2122          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2123    
2124          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2125              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1584  while (!done) Line 2146  while (!done)
2146    
2147        new_free(re);        new_free(re);
2148        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2149        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2150            {
2151            new_free((void *)tables);
2152            setlocale(LC_CTYPE, "C");
2153            locale_set = 0;
2154            }
2155        continue;  /* With next regex */        continue;  /* With next regex */
2156        }        }
2157      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1594  while (!done) Line 2161  while (!done)
2161    for (;;)    for (;;)
2162      {      {
2163      uschar *q;      uschar *q;
2164      uschar *bptr = dbuffer;      uschar *bptr;
2165      int *use_offsets = offsets;      int *use_offsets = offsets;
2166      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2167      int callout_data = 0;      int callout_data = 0;
2168      int callout_data_set = 0;      int callout_data_set = 0;
2169      int count, c;      int count, c;
2170      int copystrings = 0;      int copystrings = 0;
2171      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2172      int getstrings = 0;      int getstrings = 0;
2173      int getlist = 0;      int getlist = 0;
2174      int gmatched = 0;      int gmatched = 0;
2175      int start_offset = 0;      int start_offset = 0;
2176        int start_offset_sign = 1;
2177      int g_notempty = 0;      int g_notempty = 0;
2178      int use_dfa = 0;      int use_dfa = 0;
2179    
# Line 1631  while (!done) Line 2199  while (!done)
2199      len = 0;      len = 0;
2200      for (;;)      for (;;)
2201        {        {
2202        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2203          {          {
2204          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2205              {
2206              fprintf(outfile, "\n");
2207              break;
2208              }
2209          done = 1;          done = 1;
2210          goto CONTINUE;          goto CONTINUE;
2211          }          }
# Line 1650  while (!done) Line 2221  while (!done)
2221      p = buffer;      p = buffer;
2222      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2223    
2224      q = dbuffer;      bptr = q = dbuffer;
2225      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2226        {        {
2227        int i = 0;        int i = 0;
# Line 1700  while (!done) Line 2271  while (!done)
2271              {              {
2272              unsigned char buff8[8];              unsigned char buff8[8];
2273              int ii, utn;              int ii, utn;
2274              utn = ord2utf8(c, buff8);              if (use_utf8)
2275              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2276              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2277                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2278                  c = buff8[ii];   /* Last byte */
2279                  }
2280                else
2281                 {
2282                 if (c > 255)
2283                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2284                     "UTF-8 mode is not enabled.\n"
2285                     "** Truncation will probably give the wrong result.\n", c);
2286                 }
2287              p = pt + 1;              p = pt + 1;
2288              break;              break;
2289              }              }
# Line 1725  while (!done) Line 2306  while (!done)
2306          continue;          continue;
2307    
2308          case '>':          case '>':
2309            if (*p == '-')
2310              {
2311              start_offset_sign = -1;
2312              p++;
2313              }
2314          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2315            start_offset *= start_offset_sign;
2316          continue;          continue;
2317    
2318          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1798  while (!done) Line 2385  while (!done)
2385  #endif  #endif
2386            use_dfa = 1;            use_dfa = 1;
2387          continue;          continue;
2388    #endif
2389    
2390    #if !defined NODFA
2391          case 'F':          case 'F':
2392          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2393          continue;          continue;
# Line 1832  while (!done) Line 2421  while (!done)
2421          continue;          continue;
2422    
2423          case 'N':          case 'N':
2424          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2425              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2426            else
2427              options |= PCRE_NOTEMPTY;
2428          continue;          continue;
2429    
2430          case 'O':          case 'O':
# Line 1845  while (!done) Line 2437  while (!done)
2437            if (offsets == NULL)            if (offsets == NULL)
2438              {              {
2439              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2440                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2441              yield = 1;              yield = 1;
2442              goto EXIT;              goto EXIT;
2443              }              }
# Line 1855  while (!done) Line 2447  while (!done)
2447          continue;          continue;
2448    
2449          case 'P':          case 'P':
2450          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2451              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2452          continue;          continue;
2453    
2454          case 'Q':          case 'Q':
# Line 1890  while (!done) Line 2483  while (!done)
2483          show_malloc = 1;          show_malloc = 1;
2484          continue;          continue;
2485    
2486            case 'Y':
2487            options |= PCRE_NO_START_OPTIMIZE;
2488            continue;
2489    
2490          case 'Z':          case 'Z':
2491          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2492          continue;          continue;
# Line 1910  while (!done) Line 2507  while (!done)
2507        *q++ = c;        *q++ = c;
2508        }        }
2509      *q = 0;      *q = 0;
2510      len = q - dbuffer;      len = (int)(q - dbuffer);
2511    
2512        /* Move the data to the end of the buffer so that a read over the end of
2513        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2514        we are using the POSIX interface, we must include the terminating zero. */
2515    
2516    #if !defined NOPOSIX
2517        if (posix || do_posix)
2518          {
2519          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2520          bptr += buffer_size - len - 1;
2521          }
2522        else
2523    #endif
2524          {
2525          memmove(bptr + buffer_size - len, bptr, len);
2526          bptr += buffer_size - len;
2527          }
2528    
2529      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2530        {        {
# Line 1931  while (!done) Line 2545  while (!done)
2545          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2546        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2547        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2548          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2549    
2550        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2551    
# Line 1975  while (!done) Line 2590  while (!done)
2590    
2591      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2592        {        {
2593          markptr = NULL;
2594    
2595        if (timeitm > 0)        if (timeitm > 0)
2596          {          {
2597          register int i;          register int i;
# Line 1986  while (!done) Line 2603  while (!done)
2603            {            {
2604            int workspace[1000];            int workspace[1000];
2605            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2606              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2607                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2608                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2609            }            }
# Line 2049  while (!done) Line 2666  while (!done)
2666        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2667          {          {
2668          int workspace[1000];          int workspace[1000];
2669          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2670            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2671            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2672          if (count == 0)          if (count == 0)
# Line 2120  while (!done) Line 2737  while (!done)
2737              }              }
2738            }            }
2739    
2740            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2741    
2742          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2743            {            {
2744            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2203  while (!done) Line 2822  while (!done)
2822    
2823        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2824          {          {
2825          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2826  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2827          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2828            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2829              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2830  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2831                outfile);
2832              }
2833          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2834          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2835          }          }
# Line 2218  while (!done) Line 2839  while (!done)
2839        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
2840        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2841    
2842        Complication arises in the case when the newline option is "any".        Complication arises in the case when the newline convention is "any",
2843        If the previous match was at the end of a line terminated by CRLF, an        "crlf", or "anycrlf". If the previous match was at the end of a line
2844        advance of one character just passes the \r, whereas we should prefer the        terminated by CRLF, an advance of one character just passes the \r,
2845        longer newline sequence, as does the code in pcre_exec(). Fudge the        whereas we should prefer the longer newline sequence, as does the code in
2846        offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2847          newline setting in the pattern; if none was set, use pcre_config() to
2848          find the default.
2849    
2850        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
2851        character, not one byte. */        character, not one byte. */
# Line 2232  while (!done) Line 2855  while (!done)
2855          if (g_notempty != 0)          if (g_notempty != 0)
2856            {            {
2857            int onechar = 1;            int onechar = 1;
2858              unsigned int obits = ((real_pcre *)re)->options;
2859            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2860            if ((((real_pcre *)re)->options & PCRE_NEWLINE_BITS) ==            if ((obits & PCRE_NEWLINE_BITS) == 0)
2861                    PCRE_NEWLINE_ANY &&              {
2862                int d;
2863                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2864                /* Note that these values are always the ASCII ones, even in
2865                EBCDIC environments. CR = 13, NL = 10. */
2866                obits = (d == 13)? PCRE_NEWLINE_CR :
2867                        (d == 10)? PCRE_NEWLINE_LF :
2868                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2869                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2870                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2871                }
2872              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2873                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2874                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2875                  &&
2876                start_offset < len - 1 &&                start_offset < len - 1 &&
2877                bptr[start_offset] == '\r' &&                bptr[start_offset] == '\r' &&
2878                bptr[start_offset+1] == '\n')                bptr[start_offset+1] == '\n')
# Line 2243  while (!done) Line 2881  while (!done)
2881              {              {
2882              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2883                {                {
2884                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2885                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2886                }                }
2887              }              }
2888            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
2889            }            }
2890          else          else
2891            {            {
2892            if (count == PCRE_ERROR_NOMATCH)            switch(count)
2893              {              {
2894              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
2895                if (gmatched == 0)
2896                  {
2897                  if (markptr == NULL) fprintf(outfile, "No match\n");
2898                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2899                  }
2900                break;
2901    
2902                case PCRE_ERROR_BADUTF8:
2903                case PCRE_ERROR_SHORTUTF8:
2904                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
2905                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
2906                if (use_size_offsets >= 2)
2907                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
2908                    use_offsets[1]);
2909                fprintf(outfile, "\n");
2910                break;
2911    
2912                default:
2913                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
2914                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
2915                else
2916                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
2917                break;
2918              }              }
2919            else fprintf(outfile, "Error %d\n", count);  
2920            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
2921            }            }
2922          }          }
# Line 2267  while (!done) Line 2926  while (!done)
2926        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2927    
2928        /* 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
2929        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
2930        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
2931        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
2932        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
2933        character. */        character. */
2934    
# Line 2278  while (!done) Line 2937  while (!done)
2937        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2938          {          {
2939          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2940          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2941          }          }
2942    
2943        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2305  while (!done) Line 2964  while (!done)
2964    
2965    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
2966    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
2967    if (tables != NULL)    if (locale_set)
2968      {      {
2969      new_free((void *)tables);      new_free((void *)tables);
2970      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");

Legend:
Removed from v.145  
changed lines
  Added in v.604

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12