/[pcre]/code/branches/pcre16/pcretest.c
ViewVC logotype

Diff of /code/branches/pcre16/pcretest.c

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

code/trunk/pcretest.c revision 146 by ph10, Thu Apr 5 09:17:28 2007 UTC code/branches/pcre16/pcretest.c revision 756 by ph10, Mon Nov 21 10:48:42 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_ucp_typerange    ucp_typerange
116  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
117  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
118  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 90  symbols to prevent clashes. */ Line 120  symbols to prevent clashes. */
120  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
121  #define _pcre_utt              utt  #define _pcre_utt              utt
122  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
123    #define _pcre_utt_names        utt_names
124  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
125    
126  #include "pcre_tables.c"  #include "pcre_tables.c"
127    
128  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
129  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
130  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
131    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. */  
132    
133    #define COMPILING_PCRETEST
134  #include "pcre_printint.src"  #include "pcre_printint.src"
135    
136  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  /* The definition of the macro PRINTABLE, which determines whether to print an
137    output character as-is or as a hex value when showing compiled patterns, is
138    contained in the printint.src file. We uses it here also, in cases when the
139    locale has not been explicitly changed, so as to get consistent output from
140    systems that differ in their output from isprint() even in the "C" locale. */
141    
142    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
143    
144  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
145  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 184  static int callout_count;
184  static int callout_extra;  static int callout_extra;
185  static int callout_fail_count;  static int callout_fail_count;
186  static int callout_fail_id;  static int callout_fail_id;
187    static int debug_lengths;
188  static int first_callout;  static int first_callout;
189  static int locale_set = 0;  static int locale_set = 0;
190  static int show_malloc;  static int show_malloc;
191  static int use_utf8;  static int use_utf8;
192  static size_t gotten_store;  static size_t gotten_store;
193    static const unsigned char *last_callout_mark = NULL;
194    
195  /* The buffers grow automatically if very long input lines are encountered. */  /* The buffers grow automatically if very long input lines are encountered. */
196    
197  static int buffer_size = 50000;  static int buffer_size = 50000;
198  static uschar *buffer = NULL;  static pcre_uint8 *buffer = NULL;
199  static uschar *dbuffer = NULL;  static pcre_uint8 *dbuffer = NULL;
200  static uschar *pbuffer = NULL;  static pcre_uint8 *pbuffer = NULL;
201    
202    /* Textual explanations for runtime error codes */
203    
204    static const char *errtexts[] = {
205      NULL,  /* 0 is no error */
206      NULL,  /* NOMATCH is handled specially */
207      "NULL argument passed",
208      "bad option value",
209      "magic number missing",
210      "unknown opcode - pattern overwritten?",
211      "no more memory",
212      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
213      "match limit exceeded",
214      "callout error code",
215      NULL,  /* BADUTF8 is handled specially */
216      "bad UTF-8 offset",
217      NULL,  /* PARTIAL is handled specially */
218      "not used - internal error",
219      "internal error - pattern overwritten?",
220      "bad count value",
221      "item unsupported for DFA matching",
222      "backreference condition or recursion test not supported for DFA matching",
223      "match limit not supported for DFA matching",
224      "workspace size exceeded in DFA matching",
225      "too much recursion for DFA matching",
226      "recursion limit exceeded",
227      "not used - internal error",
228      "invalid combination of newline options",
229      "bad offset value",
230      NULL,  /* SHORTUTF8 is handled specially */
231      "nested recursion at the same subject position",
232      "JIT stack limit reached"
233    };
234    
235    
236    /*************************************************
237    *         Alternate character tables             *
238    *************************************************/
239    
240    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
241    using the default tables of the library. However, the T option can be used to
242    select alternate sets of tables, for different kinds of testing. Note also that
243    the L (locale) option also adjusts the tables. */
244    
245    /* This is the set of tables distributed as default with PCRE. It recognizes
246    only ASCII characters. */
247    
248    static const unsigned char tables0[] = {
249    
250    /* This table is a lower casing table. */
251    
252        0,  1,  2,  3,  4,  5,  6,  7,
253        8,  9, 10, 11, 12, 13, 14, 15,
254       16, 17, 18, 19, 20, 21, 22, 23,
255       24, 25, 26, 27, 28, 29, 30, 31,
256       32, 33, 34, 35, 36, 37, 38, 39,
257       40, 41, 42, 43, 44, 45, 46, 47,
258       48, 49, 50, 51, 52, 53, 54, 55,
259       56, 57, 58, 59, 60, 61, 62, 63,
260       64, 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, 91, 92, 93, 94, 95,
264       96, 97, 98, 99,100,101,102,103,
265      104,105,106,107,108,109,110,111,
266      112,113,114,115,116,117,118,119,
267      120,121,122,123,124,125,126,127,
268      128,129,130,131,132,133,134,135,
269      136,137,138,139,140,141,142,143,
270      144,145,146,147,148,149,150,151,
271      152,153,154,155,156,157,158,159,
272      160,161,162,163,164,165,166,167,
273      168,169,170,171,172,173,174,175,
274      176,177,178,179,180,181,182,183,
275      184,185,186,187,188,189,190,191,
276      192,193,194,195,196,197,198,199,
277      200,201,202,203,204,205,206,207,
278      208,209,210,211,212,213,214,215,
279      216,217,218,219,220,221,222,223,
280      224,225,226,227,228,229,230,231,
281      232,233,234,235,236,237,238,239,
282      240,241,242,243,244,245,246,247,
283      248,249,250,251,252,253,254,255,
284    
285    /* This table is a case flipping table. */
286    
287        0,  1,  2,  3,  4,  5,  6,  7,
288        8,  9, 10, 11, 12, 13, 14, 15,
289       16, 17, 18, 19, 20, 21, 22, 23,
290       24, 25, 26, 27, 28, 29, 30, 31,
291       32, 33, 34, 35, 36, 37, 38, 39,
292       40, 41, 42, 43, 44, 45, 46, 47,
293       48, 49, 50, 51, 52, 53, 54, 55,
294       56, 57, 58, 59, 60, 61, 62, 63,
295       64, 97, 98, 99,100,101,102,103,
296      104,105,106,107,108,109,110,111,
297      112,113,114,115,116,117,118,119,
298      120,121,122, 91, 92, 93, 94, 95,
299       96, 65, 66, 67, 68, 69, 70, 71,
300       72, 73, 74, 75, 76, 77, 78, 79,
301       80, 81, 82, 83, 84, 85, 86, 87,
302       88, 89, 90,123,124,125,126,127,
303      128,129,130,131,132,133,134,135,
304      136,137,138,139,140,141,142,143,
305      144,145,146,147,148,149,150,151,
306      152,153,154,155,156,157,158,159,
307      160,161,162,163,164,165,166,167,
308      168,169,170,171,172,173,174,175,
309      176,177,178,179,180,181,182,183,
310      184,185,186,187,188,189,190,191,
311      192,193,194,195,196,197,198,199,
312      200,201,202,203,204,205,206,207,
313      208,209,210,211,212,213,214,215,
314      216,217,218,219,220,221,222,223,
315      224,225,226,227,228,229,230,231,
316      232,233,234,235,236,237,238,239,
317      240,241,242,243,244,245,246,247,
318      248,249,250,251,252,253,254,255,
319    
320    /* This table contains bit maps for various character classes. Each map is 32
321    bytes long and the bits run from the least significant end of each byte. The
322    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
323    graph, print, punct, and cntrl. Other classes are built from combinations. */
324    
325      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
326      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
327      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
328      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
329    
330      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
331      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
332      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
333      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
334    
335      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
336      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
337      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
338      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
339    
340      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
341      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
342      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
343      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
344    
345      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
346      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
347      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
348      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
349    
350      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
351      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
352      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
353      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
354    
355      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
356      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
357      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
358      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
359    
360      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
361      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
362      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
363      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
364    
365      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
366      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
367      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
368      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
369    
370      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
371      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
372      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
373      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
374    
375    /* This table identifies various classes of character by individual bits:
376      0x01   white space character
377      0x02   letter
378      0x04   decimal digit
379      0x08   hexadecimal digit
380      0x10   alphanumeric or '_'
381      0x80   regular expression metacharacter or binary zero
382    */
383    
384      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
385      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
386      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
387      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
388      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
389      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
390      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
391      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
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      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
396      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
397      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
398      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
399      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
400      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
401      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
412      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
413      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
414      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
415      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
416    
417    /* This is a set of tables that came orginally from a Windows user. It seems to
418    be at least an approximation of ISO 8859. In particular, there are characters
419    greater than 128 that are marked as spaces, letters, etc. */
420    
421    static const unsigned char tables1[] = {
422    0,1,2,3,4,5,6,7,
423    8,9,10,11,12,13,14,15,
424    16,17,18,19,20,21,22,23,
425    24,25,26,27,28,29,30,31,
426    32,33,34,35,36,37,38,39,
427    40,41,42,43,44,45,46,47,
428    48,49,50,51,52,53,54,55,
429    56,57,58,59,60,61,62,63,
430    64,97,98,99,100,101,102,103,
431    104,105,106,107,108,109,110,111,
432    112,113,114,115,116,117,118,119,
433    120,121,122,91,92,93,94,95,
434    96,97,98,99,100,101,102,103,
435    104,105,106,107,108,109,110,111,
436    112,113,114,115,116,117,118,119,
437    120,121,122,123,124,125,126,127,
438    128,129,130,131,132,133,134,135,
439    136,137,138,139,140,141,142,143,
440    144,145,146,147,148,149,150,151,
441    152,153,154,155,156,157,158,159,
442    160,161,162,163,164,165,166,167,
443    168,169,170,171,172,173,174,175,
444    176,177,178,179,180,181,182,183,
445    184,185,186,187,188,189,190,191,
446    224,225,226,227,228,229,230,231,
447    232,233,234,235,236,237,238,239,
448    240,241,242,243,244,245,246,215,
449    248,249,250,251,252,253,254,223,
450    224,225,226,227,228,229,230,231,
451    232,233,234,235,236,237,238,239,
452    240,241,242,243,244,245,246,247,
453    248,249,250,251,252,253,254,255,
454    0,1,2,3,4,5,6,7,
455    8,9,10,11,12,13,14,15,
456    16,17,18,19,20,21,22,23,
457    24,25,26,27,28,29,30,31,
458    32,33,34,35,36,37,38,39,
459    40,41,42,43,44,45,46,47,
460    48,49,50,51,52,53,54,55,
461    56,57,58,59,60,61,62,63,
462    64,97,98,99,100,101,102,103,
463    104,105,106,107,108,109,110,111,
464    112,113,114,115,116,117,118,119,
465    120,121,122,91,92,93,94,95,
466    96,65,66,67,68,69,70,71,
467    72,73,74,75,76,77,78,79,
468    80,81,82,83,84,85,86,87,
469    88,89,90,123,124,125,126,127,
470    128,129,130,131,132,133,134,135,
471    136,137,138,139,140,141,142,143,
472    144,145,146,147,148,149,150,151,
473    152,153,154,155,156,157,158,159,
474    160,161,162,163,164,165,166,167,
475    168,169,170,171,172,173,174,175,
476    176,177,178,179,180,181,182,183,
477    184,185,186,187,188,189,190,191,
478    224,225,226,227,228,229,230,231,
479    232,233,234,235,236,237,238,239,
480    240,241,242,243,244,245,246,215,
481    248,249,250,251,252,253,254,223,
482    192,193,194,195,196,197,198,199,
483    200,201,202,203,204,205,206,207,
484    208,209,210,211,212,213,214,247,
485    216,217,218,219,220,221,222,255,
486    0,62,0,0,1,0,0,0,
487    0,0,0,0,0,0,0,0,
488    32,0,0,0,1,0,0,0,
489    0,0,0,0,0,0,0,0,
490    0,0,0,0,0,0,255,3,
491    126,0,0,0,126,0,0,0,
492    0,0,0,0,0,0,0,0,
493    0,0,0,0,0,0,0,0,
494    0,0,0,0,0,0,255,3,
495    0,0,0,0,0,0,0,0,
496    0,0,0,0,0,0,12,2,
497    0,0,0,0,0,0,0,0,
498    0,0,0,0,0,0,0,0,
499    254,255,255,7,0,0,0,0,
500    0,0,0,0,0,0,0,0,
501    255,255,127,127,0,0,0,0,
502    0,0,0,0,0,0,0,0,
503    0,0,0,0,254,255,255,7,
504    0,0,0,0,0,4,32,4,
505    0,0,0,128,255,255,127,255,
506    0,0,0,0,0,0,255,3,
507    254,255,255,135,254,255,255,7,
508    0,0,0,0,0,4,44,6,
509    255,255,127,255,255,255,127,255,
510    0,0,0,0,254,255,255,255,
511    255,255,255,255,255,255,255,127,
512    0,0,0,0,254,255,255,255,
513    255,255,255,255,255,255,255,255,
514    0,2,0,0,255,255,255,255,
515    255,255,255,255,255,255,255,127,
516    0,0,0,0,255,255,255,255,
517    255,255,255,255,255,255,255,255,
518    0,0,0,0,254,255,0,252,
519    1,0,0,248,1,0,0,120,
520    0,0,0,0,254,255,255,255,
521    0,0,128,0,0,0,128,0,
522    255,255,255,255,0,0,0,0,
523    0,0,0,0,0,0,0,128,
524    255,255,255,255,0,0,0,0,
525    0,0,0,0,0,0,0,0,
526    128,0,0,0,0,0,0,0,
527    0,1,1,0,1,1,0,0,
528    0,0,0,0,0,0,0,0,
529    0,0,0,0,0,0,0,0,
530    1,0,0,0,128,0,0,0,
531    128,128,128,128,0,0,128,0,
532    28,28,28,28,28,28,28,28,
533    28,28,0,0,0,0,0,128,
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    18,18,18,128,128,0,128,16,
538    0,26,26,26,26,26,26,18,
539    18,18,18,18,18,18,18,18,
540    18,18,18,18,18,18,18,18,
541    18,18,18,128,128,0,0,0,
542    0,0,0,0,0,1,0,0,
543    0,0,0,0,0,0,0,0,
544    0,0,0,0,0,0,0,0,
545    0,0,0,0,0,0,0,0,
546    1,0,0,0,0,0,0,0,
547    0,0,18,0,0,0,0,0,
548    0,0,20,20,0,18,0,0,
549    0,20,18,0,0,0,0,0,
550    18,18,18,18,18,18,18,18,
551    18,18,18,18,18,18,18,18,
552    18,18,18,18,18,18,18,0,
553    18,18,18,18,18,18,18,18,
554    18,18,18,18,18,18,18,18,
555    18,18,18,18,18,18,18,18,
556    18,18,18,18,18,18,18,0,
557    18,18,18,18,18,18,18,18
558    };
559    
560    
561    
562    
563    #ifndef HAVE_STRERROR
564    /*************************************************
565    *     Provide strerror() for non-ANSI libraries  *
566    *************************************************/
567    
568    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
569    in their libraries, but can provide the same facility by this simple
570    alternative function. */
571    
572    extern int   sys_nerr;
573    extern char *sys_errlist[];
574    
575    char *
576    strerror(int n)
577    {
578    if (n < 0 || n >= sys_nerr) return "unknown error number";
579    return sys_errlist[n];
580    }
581    #endif /* HAVE_STRERROR */
582    
583    
584    /*************************************************
585    *         JIT memory callback                    *
586    *************************************************/
587    
588    static pcre_jit_stack* jit_callback(void *arg)
589    {
590    return (pcre_jit_stack *)arg;
591    }
592    
593    
594  /*************************************************  /*************************************************
595  *        Read or extend an input line            *  *        Read or extend an input line            *
596  *************************************************/  *************************************************/
# Line 183  optimal way of handling this, but hey, t Line 607  optimal way of handling this, but hey, t
607  Arguments:  Arguments:
608    f            the file to read    f            the file to read
609    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
610      prompt       for stdin or readline()
611    
612  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
613                 could be a copy of start, or could be moved                 could be a copy of start, or could be moved
614                 NULL if no data read and EOF reached                 NULL if no data read and EOF reached
615  */  */
616    
617  static uschar *  static pcre_uint8 *
618  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, pcre_uint8 *start, const char *prompt)
619  {  {
620  uschar *here = start;  pcre_uint8 *here = start;
621    
622  for (;;)  for (;;)
623    {    {
624    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
625    
626    if (rlen > 1000)    if (rlen > 1000)
627      {      {
628      int dlen;      int dlen;
629      if (fgets((char *)here, rlen,  f) == NULL)  
630        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
631        input is a terminal. Note that readline() removes the trailing newline, so
632        we must put it back again, to be compatible with fgets(). */
633    
634    #ifdef SUPPORT_LIBREADLINE
635        if (isatty(fileno(f)))
636          {
637          size_t len;
638          char *s = readline(prompt);
639          if (s == NULL) return (here == start)? NULL : start;
640          len = strlen(s);
641          if (len > 0) add_history(s);
642          if (len > rlen - 1) len = rlen - 1;
643          memcpy(here, s, len);
644          here[len] = '\n';
645          here[len+1] = 0;
646          free(s);
647          }
648        else
649    #endif
650    
651        /* Read the next line by normal means, prompting if the file is stdin. */
652    
653          {
654          if (f == stdin) printf("%s", prompt);
655          if (fgets((char *)here, rlen,  f) == NULL)
656            return (here == start)? NULL : start;
657          }
658    
659      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
660      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
661      here += dlen;      here += dlen;
# Line 211  for (;;) Line 664  for (;;)
664    else    else
665      {      {
666      int new_buffer_size = 2*buffer_size;      int new_buffer_size = 2*buffer_size;
667      uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);      pcre_uint8 *new_buffer = (unsigned char *)malloc(new_buffer_size);
668      uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);      pcre_uint8 *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
669      uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);      pcre_uint8 *new_pbuffer = (unsigned char *)malloc(new_buffer_size);
670    
671      if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)      if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
672        {        {
# Line 355  Returns:     number of characters placed Line 808  Returns:     number of characters placed
808  #if !defined NOUTF8  #if !defined NOUTF8
809    
810  static int  static int
811  ord2utf8(int cvalue, uschar *utf8bytes)  ord2utf8(int cvalue, pcre_uint8 *utf8bytes)
812  {  {
813  register int i, j;  register int i, j;
814  for (i = 0; i < utf8_table1_size; i++)  for (i = 0; i < utf8_table1_size; i++)
# Line 518  fprintf(outfile, "%.*s", (cb->next_item_ Line 971  fprintf(outfile, "%.*s", (cb->next_item_
971  fprintf(outfile, "\n");  fprintf(outfile, "\n");
972  first_callout = 0;  first_callout = 0;
973    
974    if (cb->mark != last_callout_mark)
975      {
976      fprintf(outfile, "Latest Mark: %s\n",
977        (cb->mark == NULL)? "<unset>" : (char *)(cb->mark));
978      last_callout_mark = cb->mark;
979      }
980    
981  if (cb->callout_data != NULL)  if (cb->callout_data != NULL)
982    {    {
983    int callout_data = *((int *)(cb->callout_data));    int callout_data = *((int *)(cb->callout_data));
# Line 537  return (cb->callout_number != callout_fa Line 997  return (cb->callout_number != callout_fa
997  *            Local malloc functions              *  *            Local malloc functions              *
998  *************************************************/  *************************************************/
999    
1000  /* Alternative malloc function, to test functionality and show the size of the  /* Alternative malloc function, to test functionality and save the size of a
1001  compiled re. */  compiled re. The show_malloc variable is set only during matching. */
1002    
1003  static void *new_malloc(size_t size)  static void *new_malloc(size_t size)
1004  {  {
# Line 556  if (show_malloc) Line 1016  if (show_malloc)
1016  free(block);  free(block);
1017  }  }
1018    
   
1019  /* For recursion malloc/free, to test stacking calls */  /* For recursion malloc/free, to test stacking calls */
1020    
1021  static void *stack_malloc(size_t size)  static void *stack_malloc(size_t size)
# Line 612  return ((value & 0x000000ff) << 24) | Line 1071  return ((value & 0x000000ff) << 24) |
1071  *************************************************/  *************************************************/
1072    
1073  static int  static int
1074  check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,  check_match_limit(pcre *re, pcre_extra *extra, pcre_uint8 *bptr, int len,
1075    int start_offset, int options, int *use_offsets, int use_size_offsets,    int start_offset, int options, int *use_offsets, int use_size_offsets,
1076    int flag, unsigned long int *limit, int errnumber, const char *msg)    int flag, unsigned long int *limit, int errnumber, const char *msg)
1077  {  {
# Line 659  return count; Line 1118  return count;
1118    
1119    
1120  /*************************************************  /*************************************************
1121    *         Case-independent strncmp() function    *
1122    *************************************************/
1123    
1124    /*
1125    Arguments:
1126      s         first string
1127      t         second string
1128      n         number of characters to compare
1129    
1130    Returns:    < 0, = 0, or > 0, according to the comparison
1131    */
1132    
1133    static int
1134    strncmpic(pcre_uint8 *s, pcre_uint8 *t, int n)
1135    {
1136    while (n--)
1137      {
1138      int c = tolower(*s++) - tolower(*t++);
1139      if (c) return c;
1140      }
1141    return 0;
1142    }
1143    
1144    
1145    
1146    /*************************************************
1147  *         Check newline indicator                *  *         Check newline indicator                *
1148  *************************************************/  *************************************************/
1149    
1150  /* 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
1151  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.
1152    
1153  Arguments:  Arguments:
1154    p           points after the leading '<'    p           points after the leading '<'
# Line 673  Returns:      appropriate PCRE_NEWLINE_x Line 1158  Returns:      appropriate PCRE_NEWLINE_x
1158  */  */
1159    
1160  static int  static int
1161  check_newline(uschar *p, FILE *f)  check_newline(pcre_uint8 *p, FILE *f)
1162  {  {
1163  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (pcre_uint8 *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1164  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (pcre_uint8 *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1165  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (pcre_uint8 *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1166  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (pcre_uint8 *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1167    if (strncmpic(p, (pcre_uint8 *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1168    if (strncmpic(p, (pcre_uint8 *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1169    if (strncmpic(p, (pcre_uint8 *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1170  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
1171  return 0;  return 0;
1172  }  }
# Line 692  return 0; Line 1180  return 0;
1180  static void  static void
1181  usage(void)  usage(void)
1182  {  {
1183  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1184    printf("Input and output default to stdin and stdout.\n");
1185    #ifdef SUPPORT_LIBREADLINE
1186    printf("If input is a terminal, readline() is used to read from it.\n");
1187    #else
1188    printf("This version of pcretest is not linked with readline().\n");
1189    #endif
1190    printf("\nOptions:\n");
1191  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
1192  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
1193  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 1196  printf("  -dfa     force DFA matching fo
1196  #endif  #endif
1197  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
1198  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
1199           "  -M       find MATCH_LIMIT minimum for each subject\n"
1200         "  -m       output memory used information\n"         "  -m       output memory used information\n"
1201         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
1202  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 708  printf("  -p       use POSIX interface\n Line 1204  printf("  -p       use POSIX interface\n
1204  #endif  #endif
1205  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
1206  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
1207  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied at basic level\n"
1208           "  -s+      force each pattern to be studied, using JIT if available\n"
1209         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
1210  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1211  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 730  int main(int argc, char **argv) Line 1227  int main(int argc, char **argv)
1227  FILE *infile = stdin;  FILE *infile = stdin;
1228  int options = 0;  int options = 0;
1229  int study_options = 0;  int study_options = 0;
1230    int default_find_match_limit = FALSE;
1231  int op = 1;  int op = 1;
1232  int timeit = 0;  int timeit = 0;
1233  int timeitm = 0;  int timeitm = 0;
1234  int showinfo = 0;  int showinfo = 0;
1235  int showstore = 0;  int showstore = 0;
1236    int force_study = -1;
1237    int force_study_options = 0;
1238  int quiet = 0;  int quiet = 0;
1239  int size_offsets = 45;  int size_offsets = 45;
1240  int size_offsets_max;  int size_offsets_max;
# Line 748  int all_use_dfa = 0; Line 1248  int all_use_dfa = 0;
1248  int yield = 0;  int yield = 0;
1249  int stack_size;  int stack_size;
1250    
1251    pcre_jit_stack *jit_stack = NULL;
1252    
1253    
1254  /* These vectors store, end-to-end, a list of captured substring names. Assume  /* These vectors store, end-to-end, a list of captured substring names. Assume
1255  that 1024 is plenty long enough for the few names we'll be testing. */  that 1024 is plenty long enough for the few names we'll be testing. */
1256    
1257  uschar copynames[1024];  pcre_uchar copynames[1024];
1258  uschar getnames[1024];  pcre_uchar getnames[1024];
1259    
1260  uschar *copynamesptr;  pcre_uchar *copynamesptr;
1261  uschar *getnamesptr;  pcre_uchar *getnamesptr;
1262    
1263  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* Get buffers from malloc() so that Electric Fence will check their misuse
1264  when I am debugging. They grow automatically when very long lines are read. */  when I am debugging. They grow automatically when very long lines are read. */
1265    
1266  buffer = (unsigned char *)malloc(buffer_size);  buffer = (pcre_uint8 *)malloc(buffer_size);
1267  dbuffer = (unsigned char *)malloc(buffer_size);  dbuffer = (pcre_uint8 *)malloc(buffer_size);
1268  pbuffer = (unsigned char *)malloc(buffer_size);  pbuffer = (pcre_uint8 *)malloc(buffer_size);
1269    
1270  /* The outfile variable is static so that new_malloc can use it. */  /* The outfile variable is static so that new_malloc can use it. */
1271    
# Line 783  while (argc > 1 && argv[op][0] == '-') Line 1286  while (argc > 1 && argv[op][0] == '-')
1286    {    {
1287    unsigned char *endptr;    unsigned char *endptr;
1288    
1289    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1290      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 0;
1291      else if (strcmp(argv[op], "-s+") == 0)
1292        {
1293        force_study = 1;
1294        force_study_options = PCRE_STUDY_JIT_COMPILE;
1295        }
1296    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1297    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1298    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1299    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1300      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1301  #if !defined NODFA  #if !defined NODFA
1302    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1303  #endif  #endif
# Line 817  while (argc > 1 && argv[op][0] == '-') Line 1326  while (argc > 1 && argv[op][0] == '-')
1326        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1327          *endptr == 0))          *endptr == 0))
1328      {      {
1329  #if defined(_WIN32) || defined(WIN32)  #if defined(_WIN32) || defined(WIN32) || defined(__minix)
1330      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
1331      exit(1);      exit(1);
1332  #else  #else
# Line 841  while (argc > 1 && argv[op][0] == '-') Line 1350  while (argc > 1 && argv[op][0] == '-')
1350    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1351      {      {
1352      int rc;      int rc;
1353        unsigned long int lrc;
1354      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1355      printf("Compiled with\n");      printf("Compiled with\n");
1356      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
1357      printf("  %sUTF-8 support\n", rc? "" : "No ");      printf("  %sUTF-8 support\n", rc? "" : "No ");
1358      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1359      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1360        (void)pcre_config(PCRE_CONFIG_JIT, &rc);
1361        if (rc)
1362          printf("  Just-in-time compiler support\n");
1363        else
1364          printf("  No just-in-time compiler support\n");
1365      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1366      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1367        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1368        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1369          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1370          (rc == -2)? "ANYCRLF" :
1371        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1372        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1373        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1374                                         "all Unicode newlines");
1375      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1376      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1377      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1378      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1379      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1380      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1381      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1382      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1383      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1384      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1385      goto EXIT;      goto EXIT;
# Line 887  offsets = (int *)malloc(size_offsets_max Line 1408  offsets = (int *)malloc(size_offsets_max
1408  if (offsets == NULL)  if (offsets == NULL)
1409    {    {
1410    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1411      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1412    yield = 1;    yield = 1;
1413    goto EXIT;    goto EXIT;
1414    }    }
# Line 940  while (!done) Line 1461  while (!done)
1461  #endif  #endif
1462    
1463    const char *error;    const char *error;
1464      unsigned char *markptr;
1465    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1466    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1467    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1468    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1469    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1470      int do_allcaps = 0;
1471      int do_mark = 0;
1472    int do_study = 0;    int do_study = 0;
1473      int no_force_study = 0;
1474    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1475    int do_G = 0;    int do_G = 0;
1476    int do_g = 0;    int do_g = 0;
1477    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1478    int do_showrest = 0;    int do_showrest = 0;
1479      int do_showcaprest = 0;
1480    int do_flip = 0;    int do_flip = 0;
1481    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1482    
1483    use_utf8 = 0;    use_utf8 = 0;
1484      debug_lengths = 1;
1485    
1486    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1487    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1488    fflush(outfile);    fflush(outfile);
1489    
# Line 971  while (!done) Line 1496  while (!done)
1496    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
1497      {      {
1498      unsigned long int magic, get_options;      unsigned long int magic, get_options;
1499      uschar sbuf[8];      pcre_uint8 sbuf[8];
1500      FILE *f;      FILE *f;
1501    
1502      p++;      p++;
# Line 1013  while (!done) Line 1538  while (!done)
1538          }          }
1539        }        }
1540    
1541      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1542        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1543    
1544      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1021  while (!done) Line 1546  while (!done)
1546      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1547      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1548    
1549      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1550    
1551      if (true_study_size != 0)      if (true_study_size != 0)
1552        {        {
# Line 1037  while (!done) Line 1562  while (!done)
1562          {          {
1563          FAIL_READ:          FAIL_READ:
1564          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
1565          if (extra != NULL) new_free(extra);          if (extra != NULL) pcre_free_study(extra);
1566          if (re != NULL) new_free(re);          if (re != NULL) new_free(re);
1567          fclose(f);          fclose(f);
1568          continue;          continue;
# Line 1058  while (!done) Line 1583  while (!done)
1583    
1584    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1585      {      {
1586      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1587      goto SKIP_DATA;      goto SKIP_DATA;
1588      }      }
1589    
1590    pp = p;    pp = p;
1591    poffset = p - buffer;    poffset = (int)(p - buffer);
1592    
1593    for(;;)    for(;;)
1594      {      {
# Line 1074  while (!done) Line 1599  while (!done)
1599        pp++;        pp++;
1600        }        }
1601      if (*pp != 0) break;      if (*pp != 0) break;
1602      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1603        {        {
1604        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1605        done = 1;        done = 1;
# Line 1104  while (!done) Line 1628  while (!done)
1628    /* Look for options after final delimiter */    /* Look for options after final delimiter */
1629    
1630    options = 0;    options = 0;
   study_options = 0;  
1631    log_store = showstore;  /* default from command line */    log_store = showstore;  /* default from command line */
1632    
1633    while (*pp != 0)    while (*pp != 0)
# Line 1118  while (!done) Line 1641  while (!done)
1641        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1642        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1643    
1644        case '+': do_showrest = 1; break;        case '+':
1645          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1646          break;
1647    
1648          case '=': do_allcaps = 1; break;
1649        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1650        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
1651        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1128  while (!done) Line 1655  while (!done)
1655        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1656        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1657        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1658          case 'K': do_mark = 1; break;
1659        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1660        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1661    
# Line 1135  while (!done) Line 1663  while (!done)
1663        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1664  #endif  #endif
1665    
1666        case 'S': do_study = 1; break;        case 'S':
1667          if (do_study == 0)
1668            {
1669            do_study = 1;
1670            if (*pp == '+')
1671              {
1672              study_options |= PCRE_STUDY_JIT_COMPILE;
1673              pp++;
1674              }
1675            }
1676          else
1677            {
1678            do_study = 0;
1679            no_force_study = 1;
1680            }
1681          break;
1682    
1683        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1684          case 'W': options |= PCRE_UCP; break;
1685        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1686          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1687        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1688        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1689        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1690    
1691          case 'T':
1692          switch (*pp++)
1693            {
1694            case '0': tables = tables0; break;
1695            case '1': tables = tables1; break;
1696    
1697            case '\r':
1698            case '\n':
1699            case ' ':
1700            case 0:
1701            fprintf(outfile, "** Missing table number after /T\n");
1702            goto SKIP_DATA;
1703    
1704            default:
1705            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1706            goto SKIP_DATA;
1707            }
1708          break;
1709    
1710        case 'L':        case 'L':
1711        ppp = pp;        ppp = pp;
1712        /* 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 1732  while (!done)
1732    
1733        case '<':        case '<':
1734          {          {
1735          int x = check_newline(pp, outfile);          if (strncmpic(pp, (pcre_uint8 *)"JS>", 3) == 0)
1736          if (x == 0) goto SKIP_DATA;            {
1737          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1738          while (*pp++ != '>');            pp += 3;
1739              }
1740            else
1741              {
1742              int x = check_newline(pp, outfile);
1743              if (x == 0) goto SKIP_DATA;
1744              options |= x;
1745              while (*pp++ != '>');
1746              }
1747          }          }
1748        break;        break;
1749    
# Line 1200  while (!done) Line 1773  while (!done)
1773      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1774      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1775      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1776        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1777        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1778    
1779      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1780    
# Line 1220  while (!done) Line 1795  while (!done)
1795  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1796    
1797      {      {
1798        unsigned long int get_options;
1799    
1800      if (timeit > 0)      if (timeit > 0)
1801        {        {
1802        register int i;        register int i;
# Line 1249  while (!done) Line 1826  while (!done)
1826          {          {
1827          for (;;)          for (;;)
1828            {            {
1829            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1830              {              {
1831              done = 1;              done = 1;
1832              goto CONTINUE;              goto CONTINUE;
# Line 1263  while (!done) Line 1840  while (!done)
1840        goto CONTINUE;        goto CONTINUE;
1841        }        }
1842    
1843      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1844      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
1845      returns only limited data. Check that it agrees with the newer one. */      lines. */
1846    
1847        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1848        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1849    
1850        /* Print information if required. There are now two info-returning
1851        functions. The old one has a limited interface and returns only limited
1852        data. Check that it agrees with the newer one. */
1853    
1854      if (log_store)      if (log_store)
1855        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1279  while (!done) Line 1863  while (!done)
1863      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1864      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1865    
1866      /* 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
1867      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1868        suppresses the effect of /S (used for a few test patterns where studying is
1869        never sensible). */
1870    
1871      if (do_study)      if (do_study || (force_study >= 0 && !no_force_study))
1872        {        {
1873        if (timeit > 0)        if (timeit > 0)
1874          {          {
# Line 1290  while (!done) Line 1876  while (!done)
1876          clock_t time_taken;          clock_t time_taken;
1877          clock_t start_time = clock();          clock_t start_time = clock();
1878          for (i = 0; i < timeit; i++)          for (i = 0; i < timeit; i++)
1879            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options | force_study_options, &error);
1880          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1881          if (extra != NULL) free(extra);          if (extra != NULL) pcre_free_study(extra);
1882          fprintf(outfile, "  Study time %.4f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1883            (((double)time_taken * 1000.0) / (double)timeit) /            (((double)time_taken * 1000.0) / (double)timeit) /
1884              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1885          }          }
1886        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options | force_study_options, &error);
1887        if (error != NULL)        if (error != NULL)
1888          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
1889        else if (extra != NULL)        else if (extra != NULL)
1890          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1891        }        }
1892    
1893        /* If /K was present, we set up for handling MARK data. */
1894    
1895        if (do_mark)
1896          {
1897          if (extra == NULL)
1898            {
1899            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1900            extra->flags = 0;
1901            }
1902          extra->mark = &markptr;
1903          extra->flags |= PCRE_EXTRA_MARK;
1904          }
1905    
1906      /* 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
1907      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
1908      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 1911  while (!done)
1911      if (do_flip)      if (do_flip)
1912        {        {
1913        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1914        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1915            byteflip(rre->magic_number, sizeof(rre->magic_number));
1916        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1917        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1918        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1919        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1920        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1921        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1922        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1923          rre->first_byte =
1924            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1925          rre->req_byte =
1926            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1927          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1928          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1929        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1930          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1931        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1932            sizeof(rre->name_count));
1933    
1934        if (extra != NULL)        if (extra != NULL)
1935          {          {
1936          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1937          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1938          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1939            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1940          }          }
1941        }        }
1942    
# Line 1343  while (!done) Line 1950  while (!done)
1950        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1951        }        }
1952    
1953        /* We already have the options in get_options (see above) */
1954    
1955      if (do_showinfo)      if (do_showinfo)
1956        {        {
1957        unsigned long int get_options, all_options;        unsigned long int all_options;
1958  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1959        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1960  #endif  #endif
1961        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1962            hascrorlf;
1963        int nameentrysize, namecount;        int nameentrysize, namecount;
1964        const uschar *nametable;        const pcre_uchar *nametable;
1965    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1966        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1967        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1968        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1362  while (!done) Line 1971  while (!done)
1971        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1972        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1973        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1974          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1975          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1976          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1977    
1978  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1979        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 2015  while (!done)
2015            }            }
2016          }          }
2017    
2018        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
2019        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
2020    
2021        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
2022        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");  
2023    
2024        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
2025          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",
2026            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
2027            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
2028            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
2029            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
2030            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
2031            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
2032              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
2033              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
2034            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
2035            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
2036            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
2037            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
2038            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
2039              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2040            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2041              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2042            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2043    
2044          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
2045    
2046        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
2047          {          {
2048          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1445  while (!done) Line 2057  while (!done)
2057          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
2058          break;          break;
2059    
2060            case PCRE_NEWLINE_ANYCRLF:
2061            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
2062            break;
2063    
2064          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
2065          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
2066          break;          break;
# Line 1490  while (!done) Line 2106  while (!done)
2106        /* 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
2107        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2108        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
2109        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2110          information unless -i or -d was also present. This means that, except
2111          when auto-callouts are involved, the output from runs with and without
2112          -s should be identical. */
2113    
2114        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
2115          {          {
2116          if (extra == NULL)          if (extra == NULL)
2117            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2118          else          else
2119            {            {
2120            uschar *start_bits = NULL;            pcre_uint8 *start_bits = NULL;
2121            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2122    
2123              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2124              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2125    
2126              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2127            if (start_bits == NULL)            if (start_bits == NULL)
2128              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2129            else            else
2130              {              {
2131              int i;              int i;
# Line 1532  while (!done) Line 2155  while (!done)
2155              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2156              }              }
2157            }            }
2158    
2159            /* Show this only if the JIT was set by /S, not by -s. */
2160    
2161            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
2162              {
2163              int jit;
2164              new_info(re, extra, PCRE_INFO_JIT, &jit);
2165              if (jit)
2166                fprintf(outfile, "JIT study was successful\n");
2167              else
2168    #ifdef SUPPORT_JIT
2169                fprintf(outfile, "JIT study was not successful\n");
2170    #else
2171                fprintf(outfile, "JIT support is not available in this version of PCRE\n");
2172    #endif
2173              }
2174          }          }
2175        }        }
2176    
# Line 1548  while (!done) Line 2187  while (!done)
2187          }          }
2188        else        else
2189          {          {
2190          uschar sbuf[8];          pcre_uint8 sbuf[8];
2191          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (pcre_uint8)((true_size >> 24) & 255);
2192          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (pcre_uint8)((true_size >> 16) & 255);
2193          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (pcre_uint8)((true_size >>  8) & 255);
2194          sbuf[3] = (true_size)  & 255;          sbuf[3] = (pcre_uint8)((true_size) & 255);
2195    
2196          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (pcre_uint8)((true_study_size >> 24) & 255);
2197          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (pcre_uint8)((true_study_size >> 16) & 255);
2198          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (pcre_uint8)((true_study_size >>  8) & 255);
2199          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (pcre_uint8)((true_study_size) & 255);
2200    
2201          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2202              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1566  while (!done) Line 2205  while (!done)
2205            }            }
2206          else          else
2207            {            {
2208            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2209    
2210              /* If there is study data, write it. */
2211    
2212            if (extra != NULL)            if (extra != NULL)
2213              {              {
2214              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1576  while (!done) Line 2218  while (!done)
2218                  strerror(errno));                  strerror(errno));
2219                }                }
2220              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2221              }              }
2222            }            }
2223          fclose(f);          fclose(f);
2224          }          }
2225    
2226        new_free(re);        new_free(re);
2227        if (extra != NULL) new_free(extra);        if (extra != NULL) pcre_free_study(extra);
2228        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2229            {
2230            new_free((void *)tables);
2231            setlocale(LC_CTYPE, "C");
2232            locale_set = 0;
2233            }
2234        continue;  /* With next regex */        continue;  /* With next regex */
2235        }        }
2236      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1593  while (!done) Line 2239  while (!done)
2239    
2240    for (;;)    for (;;)
2241      {      {
2242      uschar *q;      pcre_uint8 *q;
2243      uschar *bptr = dbuffer;      pcre_uint8 *bptr;
2244      int *use_offsets = offsets;      int *use_offsets = offsets;
2245      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2246      int callout_data = 0;      int callout_data = 0;
2247      int callout_data_set = 0;      int callout_data_set = 0;
2248      int count, c;      int count, c;
2249      int copystrings = 0;      int copystrings = 0;
2250      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2251      int getstrings = 0;      int getstrings = 0;
2252      int getlist = 0;      int getlist = 0;
2253      int gmatched = 0;      int gmatched = 0;
2254      int start_offset = 0;      int start_offset = 0;
2255        int start_offset_sign = 1;
2256      int g_notempty = 0;      int g_notempty = 0;
2257      int use_dfa = 0;      int use_dfa = 0;
2258    
# Line 1619  while (!done) Line 2266  while (!done)
2266    
2267      pcre_callout = callout;      pcre_callout = callout;
2268      first_callout = 1;      first_callout = 1;
2269        last_callout_mark = NULL;
2270      callout_extra = 0;      callout_extra = 0;
2271      callout_count = 0;      callout_count = 0;
2272      callout_fail_count = 999999;      callout_fail_count = 999999;
# Line 1631  while (!done) Line 2279  while (!done)
2279      len = 0;      len = 0;
2280      for (;;)      for (;;)
2281        {        {
2282        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2283          {          {
2284          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2285              {
2286              fprintf(outfile, "\n");
2287              break;
2288              }
2289          done = 1;          done = 1;
2290          goto CONTINUE;          goto CONTINUE;
2291          }          }
# Line 1650  while (!done) Line 2301  while (!done)
2301      p = buffer;      p = buffer;
2302      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2303    
2304      q = dbuffer;      bptr = q = dbuffer;
2305      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2306        {        {
2307        int i = 0;        int i = 0;
# Line 1694  while (!done) Line 2345  while (!done)
2345            {            {
2346            unsigned char *pt = p;            unsigned char *pt = p;
2347            c = 0;            c = 0;
2348            while (isxdigit(*(++pt)))  
2349              c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W');            /* We used to have "while (isxdigit(*(++pt)))" here, but it fails
2350              when isxdigit() is a macro that refers to its argument more than
2351              once. This is banned by the C Standard, but apparently happens in at
2352              least one MacOS environment. */
2353    
2354              for (pt++; isxdigit(*pt); pt++)
2355                c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'a' - 10);
2356            if (*pt == '}')            if (*pt == '}')
2357              {              {
2358              unsigned char buff8[8];              unsigned char buff8[8];
2359              int ii, utn;              int ii, utn;
2360              utn = ord2utf8(c, buff8);              if (use_utf8)
2361              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2362              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2363                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2364                  c = buff8[ii];   /* Last byte */
2365                  }
2366                else
2367                 {
2368                 if (c > 255)
2369                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2370                     "UTF-8 mode is not enabled.\n"
2371                     "** Truncation will probably give the wrong result.\n", c);
2372                 }
2373              p = pt + 1;              p = pt + 1;
2374              break;              break;
2375              }              }
# Line 1715  while (!done) Line 2382  while (!done)
2382          c = 0;          c = 0;
2383          while (i++ < 2 && isxdigit(*p))          while (i++ < 2 && isxdigit(*p))
2384            {            {
2385            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W');            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'a' - 10);
2386            p++;            p++;
2387            }            }
2388          break;          break;
# Line 1725  while (!done) Line 2392  while (!done)
2392          continue;          continue;
2393    
2394          case '>':          case '>':
2395            if (*p == '-')
2396              {
2397              start_offset_sign = -1;
2398              p++;
2399              }
2400          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2401            start_offset *= start_offset_sign;
2402          continue;          continue;
2403    
2404          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1744  while (!done) Line 2417  while (!done)
2417            }            }
2418          else if (isalnum(*p))          else if (isalnum(*p))
2419            {            {
2420            uschar *npp = copynamesptr;            pcre_uchar *npp = copynamesptr;
2421            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2422            *npp++ = 0;            *npp++ = 0;
2423            *npp = 0;            *npp = 0;
# Line 1798  while (!done) Line 2471  while (!done)
2471  #endif  #endif
2472            use_dfa = 1;            use_dfa = 1;
2473          continue;          continue;
2474    #endif
2475    
2476    #if !defined NODFA
2477          case 'F':          case 'F':
2478          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2479          continue;          continue;
# Line 1812  while (!done) Line 2487  while (!done)
2487            }            }
2488          else if (isalnum(*p))          else if (isalnum(*p))
2489            {            {
2490            uschar *npp = getnamesptr;            pcre_uchar *npp = getnamesptr;
2491            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2492            *npp++ = 0;            *npp++ = 0;
2493            *npp = 0;            *npp = 0;
# Line 1823  while (!done) Line 2498  while (!done)
2498            }            }
2499          continue;          continue;
2500    
2501            case 'J':
2502            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2503            if (extra != NULL
2504                && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
2505                && extra->executable_jit != NULL)
2506              {
2507              if (jit_stack != NULL) pcre_jit_stack_free(jit_stack);
2508              jit_stack = pcre_jit_stack_alloc(1, n * 1024);
2509              pcre_assign_jit_stack(extra, jit_callback, jit_stack);
2510              }
2511            continue;
2512    
2513          case 'L':          case 'L':
2514          getlist = 1;          getlist = 1;
2515          continue;          continue;
# Line 1832  while (!done) Line 2519  while (!done)
2519          continue;          continue;
2520    
2521          case 'N':          case 'N':
2522          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2523              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2524            else
2525              options |= PCRE_NOTEMPTY;
2526          continue;          continue;
2527    
2528          case 'O':          case 'O':
# Line 1845  while (!done) Line 2535  while (!done)
2535            if (offsets == NULL)            if (offsets == NULL)
2536              {              {
2537              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2538                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2539              yield = 1;              yield = 1;
2540              goto EXIT;              goto EXIT;
2541              }              }
# Line 1855  while (!done) Line 2545  while (!done)
2545          continue;          continue;
2546    
2547          case 'P':          case 'P':
2548          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2549              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2550          continue;          continue;
2551    
2552          case 'Q':          case 'Q':
# Line 1890  while (!done) Line 2581  while (!done)
2581          show_malloc = 1;          show_malloc = 1;
2582          continue;          continue;
2583    
2584            case 'Y':
2585            options |= PCRE_NO_START_OPTIMIZE;
2586            continue;
2587    
2588          case 'Z':          case 'Z':
2589          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2590          continue;          continue;
# Line 1910  while (!done) Line 2605  while (!done)
2605        *q++ = c;        *q++ = c;
2606        }        }
2607      *q = 0;      *q = 0;
2608      len = q - dbuffer;      len = (int)(q - dbuffer);
2609    
2610        /* Move the data to the end of the buffer so that a read over the end of
2611        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2612        we are using the POSIX interface, we must include the terminating zero. */
2613    
2614    #if !defined NOPOSIX
2615        if (posix || do_posix)
2616          {
2617          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2618          bptr += buffer_size - len - 1;
2619          }
2620        else
2621    #endif
2622          {
2623          memmove(bptr + buffer_size - len, bptr, len);
2624          bptr += buffer_size - len;
2625          }
2626    
2627      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2628        {        {
# Line 1931  while (!done) Line 2643  while (!done)
2643          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2644        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2645        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2646          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2647    
2648        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2649    
# Line 1955  while (!done) Line 2668  while (!done)
2668              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2669                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2670              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2671              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2672                {                {
2673                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2674                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2675                  outfile);                  outfile);
2676                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 1975  while (!done) Line 2688  while (!done)
2688    
2689      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2690        {        {
2691          markptr = NULL;
2692    
2693        if (timeitm > 0)        if (timeitm > 0)
2694          {          {
2695          register int i;          register int i;
# Line 1986  while (!done) Line 2701  while (!done)
2701            {            {
2702            int workspace[1000];            int workspace[1000];
2703            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2704              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2705                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2706                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2707            }            }
# Line 2005  while (!done) Line 2720  while (!done)
2720    
2721        /* If find_match_limit is set, we want to do repeated matches with        /* If find_match_limit is set, we want to do repeated matches with
2722        varying limits in order to find the minimum value for the match limit and        varying limits in order to find the minimum value for the match limit and
2723        for the recursion limit. */        for the recursion limit. The match limits are relevant only to the normal
2724          running of pcre_exec(), so disable the JIT optimization. This makes it
2725          possible to run the same set of tests with and without JIT externally
2726          requested. */
2727    
2728        if (find_match_limit)        if (find_match_limit)
2729          {          {
# Line 2014  while (!done) Line 2732  while (!done)
2732            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2733            extra->flags = 0;            extra->flags = 0;
2734            }            }
2735            else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
2736    
2737          (void)check_match_limit(re, extra, bptr, len, start_offset,          (void)check_match_limit(re, extra, bptr, len, start_offset,
2738            options|g_notempty, use_offsets, use_size_offsets,            options|g_notempty, use_offsets, use_size_offsets,
# Line 2049  while (!done) Line 2768  while (!done)
2768        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2769          {          {
2770          int workspace[1000];          int workspace[1000];
2771          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2772            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2773            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2774          if (count == 0)          if (count == 0)
# Line 2097  while (!done) Line 2816  while (!done)
2816              }              }
2817            }            }
2818    
2819            /* do_allcaps requests showing of all captures in the pattern, to check
2820            unset ones at the end. */
2821    
2822            if (do_allcaps)
2823              {
2824              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2825              count++;   /* Allow for full match */
2826              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2827              }
2828    
2829            /* Output the captured substrings */
2830    
2831          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2832            {            {
2833            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2834                {
2835                if (use_offsets[i] != -1)
2836                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2837                    use_offsets[i], i);
2838                if (use_offsets[i+1] != -1)
2839                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2840                    use_offsets[i+1], i+1);
2841              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2842                }
2843            else            else
2844              {              {
2845              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2846              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2847                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2848              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2849              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2850                {                {
2851                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2852                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2853                  fprintf(outfile, " 0+ ");                  outfile);
2854                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2855                }                }
2856              }              }
2857            }            }
2858    
2859            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2860    
2861          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2862            {            {
2863            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2193  while (!done) Line 2931  while (!done)
2931                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);
2932              if (stringlist[i] != NULL)              if (stringlist[i] != NULL)
2933                fprintf(outfile, "string list not terminated by NULL\n");                fprintf(outfile, "string list not terminated by NULL\n");
             /* free((void *)stringlist); */  
2934              pcre_free_substring_list(stringlist);              pcre_free_substring_list(stringlist);
2935              }              }
2936            }            }
# Line 2203  while (!done) Line 2940  while (!done)
2940    
2941        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2942          {          {
2943          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2944  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2945          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2946            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2947              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2948  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2949                outfile);
2950              }
2951          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2952          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2953          }          }
# Line 2218  while (!done) Line 2957  while (!done)
2957        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
2958        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2959    
2960        Complication arises in the case when the newline option is "any".        Complication arises in the case when the newline convention is "any",
2961        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
2962        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,
2963        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
2964        offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2965          newline setting in the pattern; if none was set, use pcre_config() to
2966          find the default.
2967    
2968        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
2969        character, not one byte. */        character, not one byte. */
# Line 2238  while (!done) Line 2979  while (!done)
2979              {              {
2980              int d;              int d;
2981              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2982              obits = (d == '\r')? PCRE_NEWLINE_CR :              /* Note that these values are always the ASCII ones, even in
2983                      (d == '\n')? PCRE_NEWLINE_LF :              EBCDIC environments. CR = 13, NL = 10. */
2984                      (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :              obits = (d == 13)? PCRE_NEWLINE_CR :
2985                        (d == 10)? PCRE_NEWLINE_LF :
2986                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2987                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2988                      (d == -1)? PCRE_NEWLINE_ANY : 0;                      (d == -1)? PCRE_NEWLINE_ANY : 0;
2989              }              }
2990            if ((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY &&            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2991                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2992                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2993                  &&
2994                start_offset < len - 1 &&                start_offset < len - 1 &&
2995                bptr[start_offset] == '\r' &&                bptr[start_offset] == '\r' &&
2996                bptr[start_offset+1] == '\n')                bptr[start_offset+1] == '\n')
# Line 2252  while (!done) Line 2999  while (!done)
2999              {              {
3000              while (start_offset + onechar < len)              while (start_offset + onechar < len)
3001                {                {
3002                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
3003                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
3004                }                }
3005              }              }
3006            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
3007            }            }
3008          else          else
3009            {            {
3010            if (count == PCRE_ERROR_NOMATCH)            switch(count)
3011              {              {
3012              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
3013                if (gmatched == 0)
3014                  {
3015                  if (markptr == NULL) fprintf(outfile, "No match\n");
3016                    else fprintf(outfile, "No match, mark = %s\n", markptr);
3017                  }
3018                break;
3019    
3020                case PCRE_ERROR_BADUTF8:
3021                case PCRE_ERROR_SHORTUTF8:
3022                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
3023                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
3024                if (use_size_offsets >= 2)
3025                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
3026                    use_offsets[1]);
3027                fprintf(outfile, "\n");
3028                break;
3029    
3030                default:
3031                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
3032                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
3033                else
3034                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
3035                break;
3036              }              }
3037            else fprintf(outfile, "Error %d\n", count);  
3038            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
3039            }            }
3040          }          }
# Line 2276  while (!done) Line 3044  while (!done)
3044        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
3045    
3046        /* 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
3047        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
3048        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
3049        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
3050        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
3051        character. */        character. */
3052    
# Line 2287  while (!done) Line 3055  while (!done)
3055        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
3056          {          {
3057          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
3058          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
3059          }          }
3060    
3061        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2313  while (!done) Line 3081  while (!done)
3081  #endif  #endif
3082    
3083    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
3084    if (extra != NULL) new_free(extra);    if (extra != NULL) pcre_free_study(extra);
3085    if (tables != NULL)    if (locale_set)
3086      {      {
3087      new_free((void *)tables);      new_free((void *)tables);
3088      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
3089      locale_set = 0;      locale_set = 0;
3090      }      }
3091      if (jit_stack != NULL)
3092        {
3093        pcre_jit_stack_free(jit_stack);
3094        jit_stack = NULL;
3095        }
3096    }    }
3097    
3098  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");

Legend:
Removed from v.146  
changed lines
  Added in v.756

  ViewVC Help
Powered by ViewVC 1.1.5