/[pcre]/code/trunk/pcretest.c
ViewVC logotype

Diff of /code/trunk/pcretest.c

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

revision 230 by ph10, Mon Sep 10 13:23:56 2007 UTC revision 667 by ph10, Mon Aug 22 14:57:32 2011 UTC
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37    
38    
39  #ifdef HAVE_CONFIG_H  #ifdef HAVE_CONFIG_H
40  #include <config.h>  #include "config.h"
41  #endif  #endif
42    
43  #include <ctype.h>  #include <ctype.h>
# Line 48  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 63  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 83  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
119  #define _pcre_utf8_table3      utf8_table3  #define _pcre_utf8_table3      utf8_table3
120  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
121    #define _pcre_utf8_char_sizes  utf8_char_sizes
122  #define _pcre_utt              utt  #define _pcre_utt              utt
123  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
124    #define _pcre_utt_names        utt_names
125  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
126    
127  #include "pcre_tables.c"  #include "pcre_tables.c"
128    
129  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
130  patterns. This function is in a separate file so that it can be included in  patterns. This function is in a separate file so that it can be included in
131  pcre_compile.c when that module is compiled with debugging enabled.  pcre_compile.c when that module is compiled with debugging enabled. It needs to
132    know which case is being compiled. */
 The definition of the macro PRINTABLE, which determines whether to print an  
 output character as-is or as a hex value when showing compiled patterns, is  
 contained in this file. We uses it here also, in cases when the locale has not  
 been explicitly changed, so as to get consistent output from systems that  
 differ in their output from isprint() even in the "C" locale. */  
133    
134    #define COMPILING_PCRETEST
135  #include "pcre_printint.src"  #include "pcre_printint.src"
136    
137  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  /* The definition of the macro PRINTABLE, which determines whether to print an
138    output character as-is or as a hex value when showing compiled patterns, is
139    contained in the printint.src file. We uses it here also, in cases when the
140    locale has not been explicitly changed, so as to get consistent output from
141    systems that differ in their output from isprint() even in the "C" locale. */
142    
143    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
144    
145  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
146  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 162  static int locale_set = 0; Line 191  static int locale_set = 0;
191  static int show_malloc;  static int show_malloc;
192  static int use_utf8;  static int use_utf8;
193  static size_t gotten_store;  static size_t gotten_store;
194    static const unsigned char *last_callout_mark = NULL;
195    
196  /* The buffers grow automatically if very long input lines are encountered. */  /* The buffers grow automatically if very long input lines are encountered. */
197    
# Line 170  static uschar *buffer = NULL; Line 200  static uschar *buffer = NULL;
200  static uschar *dbuffer = NULL;  static uschar *dbuffer = NULL;
201  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
202    
203    /* Textual explanations for runtime error codes */
204    
205    static const char *errtexts[] = {
206      NULL,  /* 0 is no error */
207      NULL,  /* NOMATCH is handled specially */
208      "NULL argument passed",
209      "bad option value",
210      "magic number missing",
211      "unknown opcode - pattern overwritten?",
212      "no more memory",
213      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
214      "match limit exceeded",
215      "callout error code",
216      NULL,  /* BADUTF8 is handled specially */
217      "bad UTF-8 offset",
218      NULL,  /* PARTIAL is handled specially */
219      "not used - internal error",
220      "internal error - pattern overwritten?",
221      "bad count value",
222      "item unsupported for DFA matching",
223      "backreference condition or recursion test not supported for DFA matching",
224      "match limit not supported for DFA matching",
225      "workspace size exceeded in DFA matching",
226      "too much recursion for DFA matching",
227      "recursion limit exceeded",
228      "not used - internal error",
229      "invalid combination of newline options",
230      "bad offset value",
231      NULL,  /* SHORTUTF8 is handled specially */
232      "nested recursion at the same subject position"
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  /*************************************************  /*************************************************
# Line 188  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
# Line 195  Returns:       pointer to the start of n Line 615  Returns:       pointer to the start of n
615  */  */
616    
617  static uschar *  static uschar *
618  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
619  {  {
620  uschar *here = start;  uschar *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 523  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 542  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 561  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 693  return 0; Line 1147  return 0;
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, ANYCRLF, or ANY. Print a message and return 0 if there is  a message and return 0 if there is no match.
 no match.  
1152    
1153  Arguments:  Arguments:
1154    p           points after the leading '<'    p           points after the leading '<'
# Line 712  if (strncmpic(p, (uschar *)"lf>", 3) == Line 1165  if (strncmpic(p, (uschar *)"lf>", 3) ==
1165  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1166  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1167  if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1168    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1169    if (strncmpic(p, (uschar *)"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 725  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 734  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 741  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 763  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 781  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    
# Line 816  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 850  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 874  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        printf("  %sJust-in-time compiler support\n", rc? "" : "No ");
1362      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1363      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1364        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1365        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1366          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1367        (rc == -2)? "ANYCRLF" :        (rc == -2)? "ANYCRLF" :
1368        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1369        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1370        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1371                                         "all Unicode newlines");
1372      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1373      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1374      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1375      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1376      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1377      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1378      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1379      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1380      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1381      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1382      goto EXIT;      goto EXIT;
# Line 974  while (!done) Line 1458  while (!done)
1458  #endif  #endif
1459    
1460    const char *error;    const char *error;
1461      unsigned char *markptr;
1462    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1463    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1464    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1465    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1466    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1467      int do_allcaps = 0;
1468      int do_mark = 0;
1469    int do_study = 0;    int do_study = 0;
1470      int no_force_study = 0;
1471    int do_debug = debug;    int do_debug = debug;
1472    int do_G = 0;    int do_G = 0;
1473    int do_g = 0;    int do_g = 0;
1474    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1475    int do_showrest = 0;    int do_showrest = 0;
1476      int do_showcaprest = 0;
1477    int do_flip = 0;    int do_flip = 0;
1478    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1479    
1480    use_utf8 = 0;    use_utf8 = 0;
1481    debug_lengths = 1;    debug_lengths = 1;
1482    
1483    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1484    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1485    fflush(outfile);    fflush(outfile);
1486    
# Line 1047  while (!done) Line 1535  while (!done)
1535          }          }
1536        }        }
1537    
1538      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1539        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1540    
1541      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1055  while (!done) Line 1543  while (!done)
1543      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1544      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1545    
1546      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1547    
1548      if (true_study_size != 0)      if (true_study_size != 0)
1549        {        {
# Line 1071  while (!done) Line 1559  while (!done)
1559          {          {
1560          FAIL_READ:          FAIL_READ:
1561          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
1562          if (extra != NULL) new_free(extra);          if (extra != NULL) pcre_free_study(extra);
1563          if (re != NULL) new_free(re);          if (re != NULL) new_free(re);
1564          fclose(f);          fclose(f);
1565          continue;          continue;
# Line 1092  while (!done) Line 1580  while (!done)
1580    
1581    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1582      {      {
1583      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1584      goto SKIP_DATA;      goto SKIP_DATA;
1585      }      }
1586    
1587    pp = p;    pp = p;
1588    poffset = p - buffer;    poffset = (int)(p - buffer);
1589    
1590    for(;;)    for(;;)
1591      {      {
# Line 1108  while (!done) Line 1596  while (!done)
1596        pp++;        pp++;
1597        }        }
1598      if (*pp != 0) break;      if (*pp != 0) break;
1599      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1600        {        {
1601        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1602        done = 1;        done = 1;
# Line 1138  while (!done) Line 1625  while (!done)
1625    /* Look for options after final delimiter */    /* Look for options after final delimiter */
1626    
1627    options = 0;    options = 0;
   study_options = 0;  
1628    log_store = showstore;  /* default from command line */    log_store = showstore;  /* default from command line */
1629    
1630    while (*pp != 0)    while (*pp != 0)
# Line 1152  while (!done) Line 1638  while (!done)
1638        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1639        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1640    
1641        case '+': do_showrest = 1; break;        case '+':
1642          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1643          break;
1644    
1645          case '=': do_allcaps = 1; break;
1646        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1647        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
1648        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1162  while (!done) Line 1652  while (!done)
1652        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1653        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1654        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1655          case 'K': do_mark = 1; break;
1656        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1657        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1658    
# Line 1169  while (!done) Line 1660  while (!done)
1660        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1661  #endif  #endif
1662    
1663        case 'S': do_study = 1; break;        case 'S':
1664          if (do_study == 0)
1665            {
1666            do_study = 1;
1667            if (*pp == '+')
1668              {
1669              study_options |= PCRE_STUDY_JIT_COMPILE;
1670              pp++;
1671              }
1672            }
1673          else
1674            {
1675            do_study = 0;
1676            no_force_study = 1;
1677            }
1678          break;
1679    
1680        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1681          case 'W': options |= PCRE_UCP; break;
1682        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1683          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1684        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1685        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1686        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1687    
1688          case 'T':
1689          switch (*pp++)
1690            {
1691            case '0': tables = tables0; break;
1692            case '1': tables = tables1; break;
1693    
1694            case '\r':
1695            case '\n':
1696            case ' ':
1697            case 0:
1698            fprintf(outfile, "** Missing table number after /T\n");
1699            goto SKIP_DATA;
1700    
1701            default:
1702            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1703            goto SKIP_DATA;
1704            }
1705          break;
1706    
1707        case 'L':        case 'L':
1708        ppp = pp;        ppp = pp;
1709        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1201  while (!done) Line 1729  while (!done)
1729    
1730        case '<':        case '<':
1731          {          {
1732          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1733          if (x == 0) goto SKIP_DATA;            {
1734          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1735          while (*pp++ != '>');            pp += 3;
1736              }
1737            else
1738              {
1739              int x = check_newline(pp, outfile);
1740              if (x == 0) goto SKIP_DATA;
1741              options |= x;
1742              while (*pp++ != '>');
1743              }
1744          }          }
1745        break;        break;
1746    
# Line 1234  while (!done) Line 1770  while (!done)
1770      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1771      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1772      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1773        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1774        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1775    
1776      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1777    
# Line 1254  while (!done) Line 1792  while (!done)
1792  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1793    
1794      {      {
1795        unsigned long int get_options;
1796    
1797      if (timeit > 0)      if (timeit > 0)
1798        {        {
1799        register int i;        register int i;
# Line 1283  while (!done) Line 1823  while (!done)
1823          {          {
1824          for (;;)          for (;;)
1825            {            {
1826            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1827              {              {
1828              done = 1;              done = 1;
1829              goto CONTINUE;              goto CONTINUE;
# Line 1297  while (!done) Line 1837  while (!done)
1837        goto CONTINUE;        goto CONTINUE;
1838        }        }
1839    
1840      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1841      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
1842      returns only limited data. Check that it agrees with the newer one. */      lines. */
1843    
1844        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1845        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1846    
1847        /* Print information if required. There are now two info-returning
1848        functions. The old one has a limited interface and returns only limited
1849        data. Check that it agrees with the newer one. */
1850    
1851      if (log_store)      if (log_store)
1852        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1313  while (!done) Line 1860  while (!done)
1860      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1861      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1862    
1863      /* 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
1864      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1865        suppresses the effect of /S (used for a few test patterns where studying is
1866        never sensible). */
1867    
1868      if (do_study)      if (do_study || (force_study >= 0 && !no_force_study))
1869        {        {
1870        if (timeit > 0)        if (timeit > 0)
1871          {          {
# Line 1324  while (!done) Line 1873  while (!done)
1873          clock_t time_taken;          clock_t time_taken;
1874          clock_t start_time = clock();          clock_t start_time = clock();
1875          for (i = 0; i < timeit; i++)          for (i = 0; i < timeit; i++)
1876            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options | force_study_options, &error);
1877          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1878          if (extra != NULL) free(extra);          if (extra != NULL) pcre_free_study(extra);
1879          fprintf(outfile, "  Study time %.4f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1880            (((double)time_taken * 1000.0) / (double)timeit) /            (((double)time_taken * 1000.0) / (double)timeit) /
1881              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1882          }          }
1883        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options | force_study_options, &error);
1884        if (error != NULL)        if (error != NULL)
1885          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
1886        else if (extra != NULL)        else if (extra != NULL)
1887          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1888        }        }
1889    
1890        /* If /K was present, we set up for handling MARK data. */
1891    
1892        if (do_mark)
1893          {
1894          if (extra == NULL)
1895            {
1896            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1897            extra->flags = 0;
1898            }
1899          extra->mark = &markptr;
1900          extra->flags |= PCRE_EXTRA_MARK;
1901          }
1902    
1903      /* 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
1904      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
1905      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 1346  while (!done) Line 1908  while (!done)
1908      if (do_flip)      if (do_flip)
1909        {        {
1910        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1911        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1912            byteflip(rre->magic_number, sizeof(rre->magic_number));
1913        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1914        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1915        rre->flags = byteflip(rre->flags, sizeof(rre->flags));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1916        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->top_bracket =
1917        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1918        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));        rre->top_backref =
1919        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1920        rre->name_table_offset = byteflip(rre->name_table_offset,        rre->first_byte =
1921            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1922          rre->req_byte =
1923            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1924          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1925          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1926        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1927          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1928        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1929            sizeof(rre->name_count));
1930    
1931        if (extra != NULL)        if (extra != NULL)
1932          {          {
1933          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1934          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1935          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1936            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1937          }          }
1938        }        }
1939    
# Line 1378  while (!done) Line 1947  while (!done)
1947        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1948        }        }
1949    
1950        /* We already have the options in get_options (see above) */
1951    
1952      if (do_showinfo)      if (do_showinfo)
1953        {        {
1954        unsigned long int get_options, all_options;        unsigned long int all_options;
1955  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1956        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1957  #endif  #endif
# Line 1389  while (!done) Line 1960  while (!done)
1960        int nameentrysize, namecount;        int nameentrysize, namecount;
1961        const uschar *nametable;        const uschar *nametable;
1962    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1963        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1964        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1965        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1449  while (!done) Line 2019  while (!done)
2019        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
2020    
2021        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
2022          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",
2023            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
2024            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
2025            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
2026            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
2027            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
2028            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
2029              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
2030              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
2031            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
2032            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
2033            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
2034            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
2035            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
2036              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2037            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2038              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2039            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2040    
2041        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
# Line 1529  while (!done) Line 2103  while (!done)
2103        /* 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
2104        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2105        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
2106        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2107          information unless -i or -d was also present. This means that, except
2108          when auto-callouts are involved, the output from runs with and without
2109          -s should be identical. */
2110    
2111        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
2112          {          {
2113          if (extra == NULL)          if (extra == NULL)
2114            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2115          else          else
2116            {            {
2117            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2118            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2119    
2120              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2121              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2122    
2123              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2124            if (start_bits == NULL)            if (start_bits == NULL)
2125              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2126            else            else
2127              {              {
2128              int i;              int i;
# Line 1571  while (!done) Line 2152  while (!done)
2152              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2153              }              }
2154            }            }
2155    
2156            /* Show this only if the JIT was set by /S, not by -s. */
2157    
2158            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
2159              {
2160              int jit;
2161              new_info(re, extra, PCRE_INFO_JIT, &jit);
2162              if (jit)
2163                fprintf(outfile, "JIT study was successful\n");
2164              else
2165    #ifdef SUPPORT_JIT
2166                fprintf(outfile, "JIT study was not successful\n");
2167    #else
2168                fprintf(outfile, "JIT support is not available in this version of PCRE\n");
2169    #endif
2170              }
2171          }          }
2172        }        }
2173    
# Line 1588  while (!done) Line 2185  while (!done)
2185        else        else
2186          {          {
2187          uschar sbuf[8];          uschar sbuf[8];
2188          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2189          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2190          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2191          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2192    
2193          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2194          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2195          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2196          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2197    
2198          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2199              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1605  while (!done) Line 2202  while (!done)
2202            }            }
2203          else          else
2204            {            {
2205            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2206    
2207              /* If there is study data, write it. */
2208    
2209            if (extra != NULL)            if (extra != NULL)
2210              {              {
2211              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1615  while (!done) Line 2215  while (!done)
2215                  strerror(errno));                  strerror(errno));
2216                }                }
2217              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2218              }              }
2219            }            }
2220          fclose(f);          fclose(f);
2221          }          }
2222    
2223        new_free(re);        new_free(re);
2224        if (extra != NULL) new_free(extra);        if (extra != NULL) pcre_free_study(extra);
2225        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2226            {
2227            new_free((void *)tables);
2228            setlocale(LC_CTYPE, "C");
2229            locale_set = 0;
2230            }
2231        continue;  /* With next regex */        continue;  /* With next regex */
2232        }        }
2233      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1640  while (!done) Line 2244  while (!done)
2244      int callout_data_set = 0;      int callout_data_set = 0;
2245      int count, c;      int count, c;
2246      int copystrings = 0;      int copystrings = 0;
2247      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2248      int getstrings = 0;      int getstrings = 0;
2249      int getlist = 0;      int getlist = 0;
2250      int gmatched = 0;      int gmatched = 0;
2251      int start_offset = 0;      int start_offset = 0;
2252        int start_offset_sign = 1;
2253      int g_notempty = 0;      int g_notempty = 0;
2254      int use_dfa = 0;      int use_dfa = 0;
2255    
# Line 1658  while (!done) Line 2263  while (!done)
2263    
2264      pcre_callout = callout;      pcre_callout = callout;
2265      first_callout = 1;      first_callout = 1;
2266        last_callout_mark = NULL;
2267      callout_extra = 0;      callout_extra = 0;
2268      callout_count = 0;      callout_count = 0;
2269      callout_fail_count = 999999;      callout_fail_count = 999999;
# Line 1670  while (!done) Line 2276  while (!done)
2276      len = 0;      len = 0;
2277      for (;;)      for (;;)
2278        {        {
2279        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2280          {          {
2281          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2282              {
2283              fprintf(outfile, "\n");
2284              break;
2285              }
2286          done = 1;          done = 1;
2287          goto CONTINUE;          goto CONTINUE;
2288          }          }
# Line 1739  while (!done) Line 2348  while (!done)
2348              {              {
2349              unsigned char buff8[8];              unsigned char buff8[8];
2350              int ii, utn;              int ii, utn;
2351              utn = ord2utf8(c, buff8);              if (use_utf8)
2352              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2353              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2354                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2355                  c = buff8[ii];   /* Last byte */
2356                  }
2357                else
2358                 {
2359                 if (c > 255)
2360                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2361                     "UTF-8 mode is not enabled.\n"
2362                     "** Truncation will probably give the wrong result.\n", c);
2363                 }
2364              p = pt + 1;              p = pt + 1;
2365              break;              break;
2366              }              }
# Line 1764  while (!done) Line 2383  while (!done)
2383          continue;          continue;
2384    
2385          case '>':          case '>':
2386            if (*p == '-')
2387              {
2388              start_offset_sign = -1;
2389              p++;
2390              }
2391          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2392            start_offset *= start_offset_sign;
2393          continue;          continue;
2394    
2395          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1837  while (!done) Line 2462  while (!done)
2462  #endif  #endif
2463            use_dfa = 1;            use_dfa = 1;
2464          continue;          continue;
2465    #endif
2466    
2467    #if !defined NODFA
2468          case 'F':          case 'F':
2469          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2470          continue;          continue;
# Line 1861  while (!done) Line 2488  while (!done)
2488            getnamesptr = npp;            getnamesptr = npp;
2489            }            }
2490          continue;          continue;
2491    
2492            case 'J':
2493            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2494            if (extra != NULL
2495                && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
2496                && extra->executable_jit != NULL)
2497              {
2498              if (jit_stack != NULL) pcre_jit_stack_free(jit_stack);
2499              jit_stack = pcre_jit_stack_alloc(1, n * 1024);
2500              pcre_assign_jit_callback(extra, jit_callback, jit_stack);
2501              }
2502            continue;
2503    
2504          case 'L':          case 'L':
2505          getlist = 1;          getlist = 1;
# Line 1871  while (!done) Line 2510  while (!done)
2510          continue;          continue;
2511    
2512          case 'N':          case 'N':
2513          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2514              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2515            else
2516              options |= PCRE_NOTEMPTY;
2517          continue;          continue;
2518    
2519          case 'O':          case 'O':
# Line 1894  while (!done) Line 2536  while (!done)
2536          continue;          continue;
2537    
2538          case 'P':          case 'P':
2539          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2540              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2541          continue;          continue;
2542    
2543          case 'Q':          case 'Q':
# Line 1929  while (!done) Line 2572  while (!done)
2572          show_malloc = 1;          show_malloc = 1;
2573          continue;          continue;
2574    
2575            case 'Y':
2576            options |= PCRE_NO_START_OPTIMIZE;
2577            continue;
2578    
2579          case 'Z':          case 'Z':
2580          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2581          continue;          continue;
# Line 1949  while (!done) Line 2596  while (!done)
2596        *q++ = c;        *q++ = c;
2597        }        }
2598      *q = 0;      *q = 0;
2599      len = q - dbuffer;      len = (int)(q - dbuffer);
2600    
2601        /* Move the data to the end of the buffer so that a read over the end of
2602        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2603        we are using the POSIX interface, we must include the terminating zero. */
2604    
2605    #if !defined NOPOSIX
2606        if (posix || do_posix)
2607          {
2608          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2609          bptr += buffer_size - len - 1;
2610          }
2611        else
2612    #endif
2613          {
2614          memmove(bptr + buffer_size - len, bptr, len);
2615          bptr += buffer_size - len;
2616          }
2617    
2618      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2619        {        {
# Line 1970  while (!done) Line 2634  while (!done)
2634          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2635        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2636        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2637          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2638    
2639        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2640    
# Line 1994  while (!done) Line 2659  while (!done)
2659              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2660                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2661              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2662              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2663                {                {
2664                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2665                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2666                  outfile);                  outfile);
2667                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 2014  while (!done) Line 2679  while (!done)
2679    
2680      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2681        {        {
2682          markptr = NULL;
2683    
2684        if (timeitm > 0)        if (timeitm > 0)
2685          {          {
2686          register int i;          register int i;
# Line 2025  while (!done) Line 2692  while (!done)
2692            {            {
2693            int workspace[1000];            int workspace[1000];
2694            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2695              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2696                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2697                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2698            }            }
# Line 2044  while (!done) Line 2711  while (!done)
2711    
2712        /* 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
2713        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
2714        for the recursion limit. */        for the recursion limit. The match limits are relevant only to the normal
2715          running of pcre_exec(), so disable the JIT optimization. This makes it
2716          possible to run the same set of tests with and without JIT externally
2717          requested. */
2718    
2719        if (find_match_limit)        if (find_match_limit)
2720          {          {
# Line 2053  while (!done) Line 2723  while (!done)
2723            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2724            extra->flags = 0;            extra->flags = 0;
2725            }            }
2726            else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
2727    
2728          (void)check_match_limit(re, extra, bptr, len, start_offset,          (void)check_match_limit(re, extra, bptr, len, start_offset,
2729            options|g_notempty, use_offsets, use_size_offsets,            options|g_notempty, use_offsets, use_size_offsets,
2730            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
# Line 2088  while (!done) Line 2759  while (!done)
2759        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2760          {          {
2761          int workspace[1000];          int workspace[1000];
2762          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2763            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2764            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2765          if (count == 0)          if (count == 0)
# Line 2136  while (!done) Line 2807  while (!done)
2807              }              }
2808            }            }
2809    
2810            /* do_allcaps requests showing of all captures in the pattern, to check
2811            unset ones at the end. */
2812    
2813            if (do_allcaps)
2814              {
2815              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2816              count++;   /* Allow for full match */
2817              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2818              }
2819    
2820            /* Output the captured substrings */
2821    
2822          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2823            {            {
2824            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2825                {
2826                if (use_offsets[i] != -1)
2827                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2828                    use_offsets[i], i);
2829                if (use_offsets[i+1] != -1)
2830                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2831                    use_offsets[i+1], i+1);
2832              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2833                }
2834            else            else
2835              {              {
2836              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2837              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2838                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2839              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2840              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2841                {                {
2842                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2843                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2844                  fprintf(outfile, " 0+ ");                  outfile);
2845                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2846                }                }
2847              }              }
2848            }            }
2849    
2850            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2851    
2852          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2853            {            {
2854            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2232  while (!done) Line 2922  while (!done)
2922                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);
2923              if (stringlist[i] != NULL)              if (stringlist[i] != NULL)
2924                fprintf(outfile, "string list not terminated by NULL\n");                fprintf(outfile, "string list not terminated by NULL\n");
             /* free((void *)stringlist); */  
2925              pcre_free_substring_list(stringlist);              pcre_free_substring_list(stringlist);
2926              }              }
2927            }            }
# Line 2242  while (!done) Line 2931  while (!done)
2931    
2932        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2933          {          {
2934          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2935  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2936          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2937            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2938              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2939  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2940                outfile);
2941              }
2942          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2943          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2944          }          }
# Line 2257  while (!done) Line 2948  while (!done)
2948        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
2949        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2950    
2951        Complication arises in the case when the newline option is "any" or        Complication arises in the case when the newline convention is "any",
2952        "anycrlf". If the previous match was at the end of a line terminated by        "crlf", or "anycrlf". If the previous match was at the end of a line
2953        CRLF, an advance of one character just passes the \r, whereas we should        terminated by CRLF, an advance of one character just passes the \r,
2954        prefer the longer newline sequence, as does the code in pcre_exec().        whereas we should prefer the longer newline sequence, as does the code in
2955        Fudge the offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2956          newline setting in the pattern; if none was set, use pcre_config() to
2957          find the default.
2958    
2959        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
2960        character, not one byte. */        character, not one byte. */
# Line 2277  while (!done) Line 2970  while (!done)
2970              {              {
2971              int d;              int d;
2972              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2973              obits = (d == '\r')? PCRE_NEWLINE_CR :              /* Note that these values are always the ASCII ones, even in
2974                      (d == '\n')? PCRE_NEWLINE_LF :              EBCDIC environments. CR = 13, NL = 10. */
2975                      (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :              obits = (d == 13)? PCRE_NEWLINE_CR :
2976                        (d == 10)? PCRE_NEWLINE_LF :
2977                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2978                      (d == -2)? PCRE_NEWLINE_ANYCRLF :                      (d == -2)? PCRE_NEWLINE_ANYCRLF :
2979                      (d == -1)? PCRE_NEWLINE_ANY : 0;                      (d == -1)? PCRE_NEWLINE_ANY : 0;
2980              }              }
2981            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2982                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2983                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2984                &&                &&
2985                start_offset < len - 1 &&                start_offset < len - 1 &&
# Line 2294  while (!done) Line 2990  while (!done)
2990              {              {
2991              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2992                {                {
2993                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2994                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2995                }                }
2996              }              }
2997            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
2998            }            }
2999          else          else
3000            {            {
3001            if (count == PCRE_ERROR_NOMATCH)            switch(count)
3002              {              {
3003              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
3004                if (gmatched == 0)
3005                  {
3006                  if (markptr == NULL) fprintf(outfile, "No match\n");
3007                    else fprintf(outfile, "No match, mark = %s\n", markptr);
3008                  }
3009                break;
3010    
3011                case PCRE_ERROR_BADUTF8:
3012                case PCRE_ERROR_SHORTUTF8:
3013                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
3014                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
3015                if (use_size_offsets >= 2)
3016                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
3017                    use_offsets[1]);
3018                fprintf(outfile, "\n");
3019                break;
3020    
3021                default:
3022                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
3023                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
3024                else
3025                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
3026                break;
3027              }              }
3028            else fprintf(outfile, "Error %d\n", count);  
3029            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
3030            }            }
3031          }          }
# Line 2318  while (!done) Line 3035  while (!done)
3035        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
3036    
3037        /* 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
3038        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
3039        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
3040        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
3041        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
3042        character. */        character. */
3043    
# Line 2329  while (!done) Line 3046  while (!done)
3046        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
3047          {          {
3048          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
3049          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
3050          }          }
3051    
3052        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2355  while (!done) Line 3072  while (!done)
3072  #endif  #endif
3073    
3074    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
3075    if (extra != NULL) new_free(extra);    if (extra != NULL) pcre_free_study(extra);
3076    if (tables != NULL)    if (locale_set)
3077      {      {
3078      new_free((void *)tables);      new_free((void *)tables);
3079      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
3080      locale_set = 0;      locale_set = 0;
3081      }      }
3082      if (jit_stack != NULL)
3083        {
3084        pcre_jit_stack_free(jit_stack);
3085        jit_stack = NULL;
3086        }
3087    }    }
3088    
3089  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");

Legend:
Removed from v.230  
changed lines
  Added in v.667

  ViewVC Help
Powered by ViewVC 1.1.5