/[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 199 by ph10, Tue Jul 31 14:39:09 2007 UTC revision 676 by ph10, Sat Aug 27 15:53:04 2011 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
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>
44  #include <stdio.h>  #include <stdio.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 156  static int callout_count; Line 185  static int callout_count;
185  static int callout_extra;  static int callout_extra;
186  static int callout_fail_count;  static int callout_fail_count;
187  static int callout_fail_id;  static int callout_fail_id;
188    static int debug_lengths;
189  static int first_callout;  static int first_callout;
190  static int locale_set = 0;  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 169  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      "JIT stack limit reached"
234    };
235    
236    
237    /*************************************************
238    *         Alternate character tables             *
239    *************************************************/
240    
241    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
242    using the default tables of the library. However, the T option can be used to
243    select alternate sets of tables, for different kinds of testing. Note also that
244    the L (locale) option also adjusts the tables. */
245    
246    /* This is the set of tables distributed as default with PCRE. It recognizes
247    only ASCII characters. */
248    
249    static const unsigned char tables0[] = {
250    
251    /* This table is a lower casing table. */
252    
253        0,  1,  2,  3,  4,  5,  6,  7,
254        8,  9, 10, 11, 12, 13, 14, 15,
255       16, 17, 18, 19, 20, 21, 22, 23,
256       24, 25, 26, 27, 28, 29, 30, 31,
257       32, 33, 34, 35, 36, 37, 38, 39,
258       40, 41, 42, 43, 44, 45, 46, 47,
259       48, 49, 50, 51, 52, 53, 54, 55,
260       56, 57, 58, 59, 60, 61, 62, 63,
261       64, 97, 98, 99,100,101,102,103,
262      104,105,106,107,108,109,110,111,
263      112,113,114,115,116,117,118,119,
264      120,121,122, 91, 92, 93, 94, 95,
265       96, 97, 98, 99,100,101,102,103,
266      104,105,106,107,108,109,110,111,
267      112,113,114,115,116,117,118,119,
268      120,121,122,123,124,125,126,127,
269      128,129,130,131,132,133,134,135,
270      136,137,138,139,140,141,142,143,
271      144,145,146,147,148,149,150,151,
272      152,153,154,155,156,157,158,159,
273      160,161,162,163,164,165,166,167,
274      168,169,170,171,172,173,174,175,
275      176,177,178,179,180,181,182,183,
276      184,185,186,187,188,189,190,191,
277      192,193,194,195,196,197,198,199,
278      200,201,202,203,204,205,206,207,
279      208,209,210,211,212,213,214,215,
280      216,217,218,219,220,221,222,223,
281      224,225,226,227,228,229,230,231,
282      232,233,234,235,236,237,238,239,
283      240,241,242,243,244,245,246,247,
284      248,249,250,251,252,253,254,255,
285    
286    /* This table is a case flipping table. */
287    
288        0,  1,  2,  3,  4,  5,  6,  7,
289        8,  9, 10, 11, 12, 13, 14, 15,
290       16, 17, 18, 19, 20, 21, 22, 23,
291       24, 25, 26, 27, 28, 29, 30, 31,
292       32, 33, 34, 35, 36, 37, 38, 39,
293       40, 41, 42, 43, 44, 45, 46, 47,
294       48, 49, 50, 51, 52, 53, 54, 55,
295       56, 57, 58, 59, 60, 61, 62, 63,
296       64, 97, 98, 99,100,101,102,103,
297      104,105,106,107,108,109,110,111,
298      112,113,114,115,116,117,118,119,
299      120,121,122, 91, 92, 93, 94, 95,
300       96, 65, 66, 67, 68, 69, 70, 71,
301       72, 73, 74, 75, 76, 77, 78, 79,
302       80, 81, 82, 83, 84, 85, 86, 87,
303       88, 89, 90,123,124,125,126,127,
304      128,129,130,131,132,133,134,135,
305      136,137,138,139,140,141,142,143,
306      144,145,146,147,148,149,150,151,
307      152,153,154,155,156,157,158,159,
308      160,161,162,163,164,165,166,167,
309      168,169,170,171,172,173,174,175,
310      176,177,178,179,180,181,182,183,
311      184,185,186,187,188,189,190,191,
312      192,193,194,195,196,197,198,199,
313      200,201,202,203,204,205,206,207,
314      208,209,210,211,212,213,214,215,
315      216,217,218,219,220,221,222,223,
316      224,225,226,227,228,229,230,231,
317      232,233,234,235,236,237,238,239,
318      240,241,242,243,244,245,246,247,
319      248,249,250,251,252,253,254,255,
320    
321    /* This table contains bit maps for various character classes. Each map is 32
322    bytes long and the bits run from the least significant end of each byte. The
323    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
324    graph, print, punct, and cntrl. Other classes are built from combinations. */
325    
326      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
327      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
328      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
329      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
330    
331      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
332      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
333      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
334      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
335    
336      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
337      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
338      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
339      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
340    
341      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
342      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
343      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
344      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
345    
346      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
347      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
348      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
349      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
350    
351      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
352      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
353      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
354      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
355    
356      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
357      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
358      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
359      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
360    
361      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
362      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
363      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
364      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
365    
366      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
367      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
368      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
369      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
370    
371      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
372      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
373      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
374      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
375    
376    /* This table identifies various classes of character by individual bits:
377      0x01   white space character
378      0x02   letter
379      0x04   decimal digit
380      0x08   hexadecimal digit
381      0x10   alphanumeric or '_'
382      0x80   regular expression metacharacter or binary zero
383    */
384    
385      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
386      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
387      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
388      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
389      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
390      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
391      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
392      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
393      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
394      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
395      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
396      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
397      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
398      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
399      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
400      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
401      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
412      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
413      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
414      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
415      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
416      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
417    
418    /* This is a set of tables that came orginally from a Windows user. It seems to
419    be at least an approximation of ISO 8859. In particular, there are characters
420    greater than 128 that are marked as spaces, letters, etc. */
421    
422    static const unsigned char tables1[] = {
423    0,1,2,3,4,5,6,7,
424    8,9,10,11,12,13,14,15,
425    16,17,18,19,20,21,22,23,
426    24,25,26,27,28,29,30,31,
427    32,33,34,35,36,37,38,39,
428    40,41,42,43,44,45,46,47,
429    48,49,50,51,52,53,54,55,
430    56,57,58,59,60,61,62,63,
431    64,97,98,99,100,101,102,103,
432    104,105,106,107,108,109,110,111,
433    112,113,114,115,116,117,118,119,
434    120,121,122,91,92,93,94,95,
435    96,97,98,99,100,101,102,103,
436    104,105,106,107,108,109,110,111,
437    112,113,114,115,116,117,118,119,
438    120,121,122,123,124,125,126,127,
439    128,129,130,131,132,133,134,135,
440    136,137,138,139,140,141,142,143,
441    144,145,146,147,148,149,150,151,
442    152,153,154,155,156,157,158,159,
443    160,161,162,163,164,165,166,167,
444    168,169,170,171,172,173,174,175,
445    176,177,178,179,180,181,182,183,
446    184,185,186,187,188,189,190,191,
447    224,225,226,227,228,229,230,231,
448    232,233,234,235,236,237,238,239,
449    240,241,242,243,244,245,246,215,
450    248,249,250,251,252,253,254,223,
451    224,225,226,227,228,229,230,231,
452    232,233,234,235,236,237,238,239,
453    240,241,242,243,244,245,246,247,
454    248,249,250,251,252,253,254,255,
455    0,1,2,3,4,5,6,7,
456    8,9,10,11,12,13,14,15,
457    16,17,18,19,20,21,22,23,
458    24,25,26,27,28,29,30,31,
459    32,33,34,35,36,37,38,39,
460    40,41,42,43,44,45,46,47,
461    48,49,50,51,52,53,54,55,
462    56,57,58,59,60,61,62,63,
463    64,97,98,99,100,101,102,103,
464    104,105,106,107,108,109,110,111,
465    112,113,114,115,116,117,118,119,
466    120,121,122,91,92,93,94,95,
467    96,65,66,67,68,69,70,71,
468    72,73,74,75,76,77,78,79,
469    80,81,82,83,84,85,86,87,
470    88,89,90,123,124,125,126,127,
471    128,129,130,131,132,133,134,135,
472    136,137,138,139,140,141,142,143,
473    144,145,146,147,148,149,150,151,
474    152,153,154,155,156,157,158,159,
475    160,161,162,163,164,165,166,167,
476    168,169,170,171,172,173,174,175,
477    176,177,178,179,180,181,182,183,
478    184,185,186,187,188,189,190,191,
479    224,225,226,227,228,229,230,231,
480    232,233,234,235,236,237,238,239,
481    240,241,242,243,244,245,246,215,
482    248,249,250,251,252,253,254,223,
483    192,193,194,195,196,197,198,199,
484    200,201,202,203,204,205,206,207,
485    208,209,210,211,212,213,214,247,
486    216,217,218,219,220,221,222,255,
487    0,62,0,0,1,0,0,0,
488    0,0,0,0,0,0,0,0,
489    32,0,0,0,1,0,0,0,
490    0,0,0,0,0,0,0,0,
491    0,0,0,0,0,0,255,3,
492    126,0,0,0,126,0,0,0,
493    0,0,0,0,0,0,0,0,
494    0,0,0,0,0,0,0,0,
495    0,0,0,0,0,0,255,3,
496    0,0,0,0,0,0,0,0,
497    0,0,0,0,0,0,12,2,
498    0,0,0,0,0,0,0,0,
499    0,0,0,0,0,0,0,0,
500    254,255,255,7,0,0,0,0,
501    0,0,0,0,0,0,0,0,
502    255,255,127,127,0,0,0,0,
503    0,0,0,0,0,0,0,0,
504    0,0,0,0,254,255,255,7,
505    0,0,0,0,0,4,32,4,
506    0,0,0,128,255,255,127,255,
507    0,0,0,0,0,0,255,3,
508    254,255,255,135,254,255,255,7,
509    0,0,0,0,0,4,44,6,
510    255,255,127,255,255,255,127,255,
511    0,0,0,0,254,255,255,255,
512    255,255,255,255,255,255,255,127,
513    0,0,0,0,254,255,255,255,
514    255,255,255,255,255,255,255,255,
515    0,2,0,0,255,255,255,255,
516    255,255,255,255,255,255,255,127,
517    0,0,0,0,255,255,255,255,
518    255,255,255,255,255,255,255,255,
519    0,0,0,0,254,255,0,252,
520    1,0,0,248,1,0,0,120,
521    0,0,0,0,254,255,255,255,
522    0,0,128,0,0,0,128,0,
523    255,255,255,255,0,0,0,0,
524    0,0,0,0,0,0,0,128,
525    255,255,255,255,0,0,0,0,
526    0,0,0,0,0,0,0,0,
527    128,0,0,0,0,0,0,0,
528    0,1,1,0,1,1,0,0,
529    0,0,0,0,0,0,0,0,
530    0,0,0,0,0,0,0,0,
531    1,0,0,0,128,0,0,0,
532    128,128,128,128,0,0,128,0,
533    28,28,28,28,28,28,28,28,
534    28,28,0,0,0,0,0,128,
535    0,26,26,26,26,26,26,18,
536    18,18,18,18,18,18,18,18,
537    18,18,18,18,18,18,18,18,
538    18,18,18,128,128,0,128,16,
539    0,26,26,26,26,26,26,18,
540    18,18,18,18,18,18,18,18,
541    18,18,18,18,18,18,18,18,
542    18,18,18,128,128,0,0,0,
543    0,0,0,0,0,1,0,0,
544    0,0,0,0,0,0,0,0,
545    0,0,0,0,0,0,0,0,
546    0,0,0,0,0,0,0,0,
547    1,0,0,0,0,0,0,0,
548    0,0,18,0,0,0,0,0,
549    0,0,20,20,0,18,0,0,
550    0,20,18,0,0,0,0,0,
551    18,18,18,18,18,18,18,18,
552    18,18,18,18,18,18,18,18,
553    18,18,18,18,18,18,18,0,
554    18,18,18,18,18,18,18,18,
555    18,18,18,18,18,18,18,18,
556    18,18,18,18,18,18,18,18,
557    18,18,18,18,18,18,18,0,
558    18,18,18,18,18,18,18,18
559    };
560    
561    
562    
563    
564    #ifndef HAVE_STRERROR
565    /*************************************************
566    *     Provide strerror() for non-ANSI libraries  *
567    *************************************************/
568    
569    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
570    in their libraries, but can provide the same facility by this simple
571    alternative function. */
572    
573    extern int   sys_nerr;
574    extern char *sys_errlist[];
575    
576    char *
577    strerror(int n)
578    {
579    if (n < 0 || n >= sys_nerr) return "unknown error number";
580    return sys_errlist[n];
581    }
582    #endif /* HAVE_STRERROR */
583    
584    
585    /*************************************************
586    *         JIT memory callback                    *
587    *************************************************/
588    
589    static pcre_jit_stack* jit_callback(void *arg)
590    {
591    return (pcre_jit_stack *)arg;
592    }
593    
594    
595  /*************************************************  /*************************************************
# Line 187  optimal way of handling this, but hey, t Line 608  optimal way of handling this, but hey, t
608  Arguments:  Arguments:
609    f            the file to read    f            the file to read
610    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
611      prompt       for stdin or readline()
612    
613  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
614                 could be a copy of start, or could be moved                 could be a copy of start, or could be moved
# Line 194  Returns: pointer to the start of n Line 616  Returns: pointer to the start of n
616  */  */
617    
618  static uschar *  static uschar *
619  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
620  {  {
621  uschar *here = start;  uschar *here = start;
622    
623  for (;;)  for (;;)
624    {    {
625    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
626    
627    if (rlen > 1000)    if (rlen > 1000)
628      {      {
629      int dlen;      int dlen;
630      if (fgets((char *)here, rlen,  f) == NULL)  
631        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
632        input is a terminal. Note that readline() removes the trailing newline, so
633        we must put it back again, to be compatible with fgets(). */
634    
635    #ifdef SUPPORT_LIBREADLINE
636        if (isatty(fileno(f)))
637          {
638          size_t len;
639          char *s = readline(prompt);
640          if (s == NULL) return (here == start)? NULL : start;
641          len = strlen(s);
642          if (len > 0) add_history(s);
643          if (len > rlen - 1) len = rlen - 1;
644          memcpy(here, s, len);
645          here[len] = '\n';
646          here[len+1] = 0;
647          free(s);
648          }
649        else
650    #endif
651    
652        /* Read the next line by normal means, prompting if the file is stdin. */
653    
654          {
655          if (f == stdin) printf("%s", prompt);
656          if (fgets((char *)here, rlen,  f) == NULL)
657            return (here == start)? NULL : start;
658          }
659    
660      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
661      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
662      here += dlen;      here += dlen;
# Line 522  fprintf(outfile, "%.*s", (cb->next_item_ Line 972  fprintf(outfile, "%.*s", (cb->next_item_
972  fprintf(outfile, "\n");  fprintf(outfile, "\n");
973  first_callout = 0;  first_callout = 0;
974    
975    if (cb->mark != last_callout_mark)
976      {
977      fprintf(outfile, "Latest Mark: %s\n",
978        (cb->mark == NULL)? "<unset>" : (char *)(cb->mark));
979      last_callout_mark = cb->mark;
980      }
981    
982  if (cb->callout_data != NULL)  if (cb->callout_data != NULL)
983    {    {
984    int callout_data = *((int *)(cb->callout_data));    int callout_data = *((int *)(cb->callout_data));
# Line 541  return (cb->callout_number != callout_fa Line 998  return (cb->callout_number != callout_fa
998  *            Local malloc functions              *  *            Local malloc functions              *
999  *************************************************/  *************************************************/
1000    
1001  /* Alternative malloc function, to test functionality and show the size of the  /* Alternative malloc function, to test functionality and save the size of a
1002  compiled re. */  compiled re. The show_malloc variable is set only during matching. */
1003    
1004  static void *new_malloc(size_t size)  static void *new_malloc(size_t size)
1005  {  {
# Line 560  if (show_malloc) Line 1017  if (show_malloc)
1017  free(block);  free(block);
1018  }  }
1019    
   
1020  /* For recursion malloc/free, to test stacking calls */  /* For recursion malloc/free, to test stacking calls */
1021    
1022  static void *stack_malloc(size_t size)  static void *stack_malloc(size_t size)
# Line 595  if ((rc = pcre_fullinfo(re, study, optio Line 1051  if ((rc = pcre_fullinfo(re, study, optio
1051    
1052    
1053  /*************************************************  /*************************************************
1054    *      Check for supported JIT architecture      *
1055    *************************************************/
1056    
1057    /* If it won't JIT-compile a very simple regex, return FALSE. */
1058    
1059    static int check_jit_arch(void)
1060    {
1061    const char *error;
1062    int erroffset, rc;
1063    pcre *re = pcre_compile("abc", 0, &error, &erroffset, NULL);
1064    pcre_extra *extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
1065    rc = extra != NULL && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 &&
1066      extra->executable_jit != NULL;
1067    pcre_free_study(extra);
1068    free(re);
1069    return rc;
1070    }
1071    
1072    
1073    /*************************************************
1074  *         Byte flipping function                 *  *         Byte flipping function                 *
1075  *************************************************/  *************************************************/
1076    
# Line 663  return count; Line 1139  return count;
1139    
1140    
1141  /*************************************************  /*************************************************
1142    *         Case-independent strncmp() function    *
1143    *************************************************/
1144    
1145    /*
1146    Arguments:
1147      s         first string
1148      t         second string
1149      n         number of characters to compare
1150    
1151    Returns:    < 0, = 0, or > 0, according to the comparison
1152    */
1153    
1154    static int
1155    strncmpic(uschar *s, uschar *t, int n)
1156    {
1157    while (n--)
1158      {
1159      int c = tolower(*s++) - tolower(*t++);
1160      if (c) return c;
1161      }
1162    return 0;
1163    }
1164    
1165    
1166    
1167    /*************************************************
1168  *         Check newline indicator                *  *         Check newline indicator                *
1169  *************************************************/  *************************************************/
1170    
1171  /* 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
1172  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.  
1173    
1174  Arguments:  Arguments:
1175    p           points after the leading '<'    p           points after the leading '<'
# Line 680  Returns: appropriate PCRE_NEWLINE_x Line 1181  Returns: appropriate PCRE_NEWLINE_x
1181  static int  static int
1182  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
1183  {  {
1184  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1185  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1186  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1187  if (strncmp((char *)p, "anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1188  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1189    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1190    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1191  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
1192  return 0;  return 0;
1193  }  }
# Line 698  return 0; Line 1201  return 0;
1201  static void  static void
1202  usage(void)  usage(void)
1203  {  {
1204  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1205    printf("Input and output default to stdin and stdout.\n");
1206    #ifdef SUPPORT_LIBREADLINE
1207    printf("If input is a terminal, readline() is used to read from it.\n");
1208    #else
1209    printf("This version of pcretest is not linked with readline().\n");
1210    #endif
1211    printf("\nOptions:\n");
1212  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
1213  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
1214  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 707  printf(" -dfa force DFA matching fo Line 1217  printf(" -dfa force DFA matching fo
1217  #endif  #endif
1218  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
1219  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
1220           "  -M       find MATCH_LIMIT minimum for each subject\n"
1221         "  -m       output memory used information\n"         "  -m       output memory used information\n"
1222         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
1223  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 714  printf(" -p use POSIX interface\n Line 1225  printf(" -p use POSIX interface\n
1225  #endif  #endif
1226  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
1227  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
1228  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied at basic level\n"
1229           "  -s+      force each pattern to be studied, using JIT if available\n"
1230         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
1231  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1232  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 736  int main(int argc, char **argv) Line 1248  int main(int argc, char **argv)
1248  FILE *infile = stdin;  FILE *infile = stdin;
1249  int options = 0;  int options = 0;
1250  int study_options = 0;  int study_options = 0;
1251    int default_find_match_limit = FALSE;
1252  int op = 1;  int op = 1;
1253  int timeit = 0;  int timeit = 0;
1254  int timeitm = 0;  int timeitm = 0;
1255  int showinfo = 0;  int showinfo = 0;
1256  int showstore = 0;  int showstore = 0;
1257    int force_study = -1;
1258    int force_study_options = 0;
1259  int quiet = 0;  int quiet = 0;
1260  int size_offsets = 45;  int size_offsets = 45;
1261  int size_offsets_max;  int size_offsets_max;
# Line 754  int all_use_dfa = 0; Line 1269  int all_use_dfa = 0;
1269  int yield = 0;  int yield = 0;
1270  int stack_size;  int stack_size;
1271    
1272    pcre_jit_stack *jit_stack = NULL;
1273    
1274    
1275  /* 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
1276  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. */
1277    
# Line 789  while (argc > 1 && argv[op][0] == '-') Line 1307  while (argc > 1 && argv[op][0] == '-')
1307    {    {
1308    unsigned char *endptr;    unsigned char *endptr;
1309    
1310    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1311      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 0;
1312      else if (strcmp(argv[op], "-s+") == 0)
1313        {
1314        force_study = 1;
1315        force_study_options = PCRE_STUDY_JIT_COMPILE;
1316        }
1317    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1318    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1319    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1320    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1321      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1322  #if !defined NODFA  #if !defined NODFA
1323    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1324  #endif  #endif
# Line 823  while (argc > 1 && argv[op][0] == '-') Line 1347  while (argc > 1 && argv[op][0] == '-')
1347        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1348          *endptr == 0))          *endptr == 0))
1349      {      {
1350  #if defined(_WIN32) || defined(WIN32)  #if defined(_WIN32) || defined(WIN32) || defined(__minix)
1351      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
1352      exit(1);      exit(1);
1353  #else  #else
# Line 847  while (argc > 1 && argv[op][0] == '-') Line 1371  while (argc > 1 && argv[op][0] == '-')
1371    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1372      {      {
1373      int rc;      int rc;
1374        unsigned long int lrc;
1375      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1376      printf("Compiled with\n");      printf("Compiled with\n");
1377      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
1378      printf("  %sUTF-8 support\n", rc? "" : "No ");      printf("  %sUTF-8 support\n", rc? "" : "No ");
1379      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1380      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1381        (void)pcre_config(PCRE_CONFIG_JIT, &rc);
1382        if (rc)
1383          printf("  Just-in-time compiler support%s\n", check_jit_arch()?
1384            "" : " (but this architecture is unsupported)");
1385        else
1386          printf("  No just-in-time compiler support\n");
1387      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1388      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1389        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1390        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1391          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1392        (rc == -2)? "ANYCRLF" :        (rc == -2)? "ANYCRLF" :
1393        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1394        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1395        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1396                                         "all Unicode newlines");
1397      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1398      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1399      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1400      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1401      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1402      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1403      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1404      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1405      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1406      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1407      goto EXIT;      goto EXIT;
# Line 947  while (!done) Line 1483  while (!done)
1483  #endif  #endif
1484    
1485    const char *error;    const char *error;
1486      unsigned char *markptr;
1487    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1488    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1489    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1490    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1491    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1492      int do_allcaps = 0;
1493      int do_mark = 0;
1494    int do_study = 0;    int do_study = 0;
1495      int no_force_study = 0;
1496    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1497    int do_G = 0;    int do_G = 0;
1498    int do_g = 0;    int do_g = 0;
1499    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1500    int do_showrest = 0;    int do_showrest = 0;
1501      int do_showcaprest = 0;
1502    int do_flip = 0;    int do_flip = 0;
1503    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1504    
1505    use_utf8 = 0;    use_utf8 = 0;
1506      debug_lengths = 1;
1507    
1508    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1509    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1510    fflush(outfile);    fflush(outfile);
1511    
# Line 1020  while (!done) Line 1560  while (!done)
1560          }          }
1561        }        }
1562    
1563      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1564        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1565    
1566      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1028  while (!done) Line 1568  while (!done)
1568      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1569      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1570    
1571      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1572    
1573      if (true_study_size != 0)      if (true_study_size != 0)
1574        {        {
# Line 1044  while (!done) Line 1584  while (!done)
1584          {          {
1585          FAIL_READ:          FAIL_READ:
1586          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
1587          if (extra != NULL) new_free(extra);          if (extra != NULL) pcre_free_study(extra);
1588          if (re != NULL) new_free(re);          if (re != NULL) new_free(re);
1589          fclose(f);          fclose(f);
1590          continue;          continue;
# Line 1065  while (!done) Line 1605  while (!done)
1605    
1606    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1607      {      {
1608      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1609      goto SKIP_DATA;      goto SKIP_DATA;
1610      }      }
1611    
1612    pp = p;    pp = p;
1613    poffset = p - buffer;    poffset = (int)(p - buffer);
1614    
1615    for(;;)    for(;;)
1616      {      {
# Line 1081  while (!done) Line 1621  while (!done)
1621        pp++;        pp++;
1622        }        }
1623      if (*pp != 0) break;      if (*pp != 0) break;
1624      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1625        {        {
1626        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1627        done = 1;        done = 1;
# Line 1111  while (!done) Line 1650  while (!done)
1650    /* Look for options after final delimiter */    /* Look for options after final delimiter */
1651    
1652    options = 0;    options = 0;
   study_options = 0;  
1653    log_store = showstore;  /* default from command line */    log_store = showstore;  /* default from command line */
1654    
1655    while (*pp != 0)    while (*pp != 0)
# Line 1125  while (!done) Line 1663  while (!done)
1663        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1664        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1665    
1666        case '+': do_showrest = 1; break;        case '+':
1667          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1668          break;
1669    
1670          case '=': do_allcaps = 1; break;
1671        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1672        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
1673        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1135  while (!done) Line 1677  while (!done)
1677        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1678        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1679        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1680          case 'K': do_mark = 1; break;
1681        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1682        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1683    
# Line 1142  while (!done) Line 1685  while (!done)
1685        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1686  #endif  #endif
1687    
1688        case 'S': do_study = 1; break;        case 'S':
1689          if (do_study == 0)
1690            {
1691            do_study = 1;
1692            if (*pp == '+')
1693              {
1694              study_options |= PCRE_STUDY_JIT_COMPILE;
1695              pp++;
1696              }
1697            }
1698          else
1699            {
1700            do_study = 0;
1701            no_force_study = 1;
1702            }
1703          break;
1704    
1705        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1706          case 'W': options |= PCRE_UCP; break;
1707        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1708          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1709        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1710        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1711        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1712    
1713          case 'T':
1714          switch (*pp++)
1715            {
1716            case '0': tables = tables0; break;
1717            case '1': tables = tables1; break;
1718    
1719            case '\r':
1720            case '\n':
1721            case ' ':
1722            case 0:
1723            fprintf(outfile, "** Missing table number after /T\n");
1724            goto SKIP_DATA;
1725    
1726            default:
1727            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1728            goto SKIP_DATA;
1729            }
1730          break;
1731    
1732        case 'L':        case 'L':
1733        ppp = pp;        ppp = pp;
1734        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1174  while (!done) Line 1754  while (!done)
1754    
1755        case '<':        case '<':
1756          {          {
1757          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1758          if (x == 0) goto SKIP_DATA;            {
1759          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1760          while (*pp++ != '>');            pp += 3;
1761              }
1762            else
1763              {
1764              int x = check_newline(pp, outfile);
1765              if (x == 0) goto SKIP_DATA;
1766              options |= x;
1767              while (*pp++ != '>');
1768              }
1769          }          }
1770        break;        break;
1771    
# Line 1207  while (!done) Line 1795  while (!done)
1795      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1796      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1797      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1798        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1799        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1800    
1801      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1802    
# Line 1227  while (!done) Line 1817  while (!done)
1817  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1818    
1819      {      {
1820        unsigned long int get_options;
1821    
1822      if (timeit > 0)      if (timeit > 0)
1823        {        {
1824        register int i;        register int i;
# Line 1256  while (!done) Line 1848  while (!done)
1848          {          {
1849          for (;;)          for (;;)
1850            {            {
1851            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1852              {              {
1853              done = 1;              done = 1;
1854              goto CONTINUE;              goto CONTINUE;
# Line 1270  while (!done) Line 1862  while (!done)
1862        goto CONTINUE;        goto CONTINUE;
1863        }        }
1864    
1865      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1866      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
1867      returns only limited data. Check that it agrees with the newer one. */      lines. */
1868    
1869        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1870        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1871    
1872        /* Print information if required. There are now two info-returning
1873        functions. The old one has a limited interface and returns only limited
1874        data. Check that it agrees with the newer one. */
1875    
1876      if (log_store)      if (log_store)
1877        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1286  while (!done) Line 1885  while (!done)
1885      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1886      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1887    
1888      /* 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
1889      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1890        suppresses the effect of /S (used for a few test patterns where studying is
1891        never sensible). */
1892    
1893      if (do_study)      if (do_study || (force_study >= 0 && !no_force_study))
1894        {        {
1895        if (timeit > 0)        if (timeit > 0)
1896          {          {
# Line 1297  while (!done) Line 1898  while (!done)
1898          clock_t time_taken;          clock_t time_taken;
1899          clock_t start_time = clock();          clock_t start_time = clock();
1900          for (i = 0; i < timeit; i++)          for (i = 0; i < timeit; i++)
1901            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options | force_study_options, &error);
1902          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1903          if (extra != NULL) free(extra);          if (extra != NULL) pcre_free_study(extra);
1904          fprintf(outfile, "  Study time %.4f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1905            (((double)time_taken * 1000.0) / (double)timeit) /            (((double)time_taken * 1000.0) / (double)timeit) /
1906              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1907          }          }
1908        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options | force_study_options, &error);
1909        if (error != NULL)        if (error != NULL)
1910          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
1911        else if (extra != NULL)        else if (extra != NULL)
1912          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1913        }        }
1914    
1915        /* If /K was present, we set up for handling MARK data. */
1916    
1917        if (do_mark)
1918          {
1919          if (extra == NULL)
1920            {
1921            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1922            extra->flags = 0;
1923            }
1924          extra->mark = &markptr;
1925          extra->flags |= PCRE_EXTRA_MARK;
1926          }
1927    
1928      /* 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
1929      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
1930      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 1319  while (!done) Line 1933  while (!done)
1933      if (do_flip)      if (do_flip)
1934        {        {
1935        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1936        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1937            byteflip(rre->magic_number, sizeof(rre->magic_number));
1938        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1939        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1940        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1941        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1942        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1943        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1944        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1945          rre->first_byte =
1946            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1947          rre->req_byte =
1948            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1949          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1950          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1951        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1952          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1953        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1954            sizeof(rre->name_count));
1955    
1956        if (extra != NULL)        if (extra != NULL)
1957          {          {
1958          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1959          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1960          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1961            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1962          }          }
1963        }        }
1964    
# Line 1350  while (!done) Line 1972  while (!done)
1972        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1973        }        }
1974    
1975        /* We already have the options in get_options (see above) */
1976    
1977      if (do_showinfo)      if (do_showinfo)
1978        {        {
1979        unsigned long int get_options, all_options;        unsigned long int all_options;
1980  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1981        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1982  #endif  #endif
1983        int count, backrefmax, first_char, need_char, okpartial, jchanged;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1984            hascrorlf;
1985        int nameentrysize, namecount;        int nameentrysize, namecount;
1986        const uschar *nametable;        const uschar *nametable;
1987    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1988        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1989        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1990        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1371  while (!done) Line 1995  while (!done)
1995        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1996        new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);        new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1997        new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);        new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1998          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1999    
2000  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
2001        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1413  while (!done) Line 2038  while (!done)
2038          }          }
2039    
2040        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
2041          if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
2042    
2043        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
2044        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
2045    
2046        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
2047          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",
2048            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
2049            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
2050            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
2051            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
2052            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
2053            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
2054              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
2055              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
2056            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
2057            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
2058            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
2059            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
2060            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
2061              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2062            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2063              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2064            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2065    
2066        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
# Line 1498  while (!done) Line 2128  while (!done)
2128        /* 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
2129        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2130        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
2131        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2132          information unless -i or -d was also present. This means that, except
2133          when auto-callouts are involved, the output from runs with and without
2134          -s should be identical. */
2135    
2136        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
2137          {          {
2138          if (extra == NULL)          if (extra == NULL)
2139            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2140          else          else
2141            {            {
2142            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2143            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2144    
2145              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2146              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2147    
2148              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2149            if (start_bits == NULL)            if (start_bits == NULL)
2150              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2151            else            else
2152              {              {
2153              int i;              int i;
# Line 1540  while (!done) Line 2177  while (!done)
2177              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2178              }              }
2179            }            }
2180    
2181            /* Show this only if the JIT was set by /S, not by -s. */
2182    
2183            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
2184              {
2185              int jit;
2186              new_info(re, extra, PCRE_INFO_JIT, &jit);
2187              if (jit)
2188                fprintf(outfile, "JIT study was successful\n");
2189              else
2190    #ifdef SUPPORT_JIT
2191                fprintf(outfile, "JIT study was not successful\n");
2192    #else
2193                fprintf(outfile, "JIT support is not available in this version of PCRE\n");
2194    #endif
2195              }
2196          }          }
2197        }        }
2198    
# Line 1557  while (!done) Line 2210  while (!done)
2210        else        else
2211          {          {
2212          uschar sbuf[8];          uschar sbuf[8];
2213          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2214          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2215          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2216          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2217    
2218          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2219          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2220          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2221          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2222    
2223          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2224              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1574  while (!done) Line 2227  while (!done)
2227            }            }
2228          else          else
2229            {            {
2230            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2231    
2232              /* If there is study data, write it. */
2233    
2234            if (extra != NULL)            if (extra != NULL)
2235              {              {
2236              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1584  while (!done) Line 2240  while (!done)
2240                  strerror(errno));                  strerror(errno));
2241                }                }
2242              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2243              }              }
2244            }            }
2245          fclose(f);          fclose(f);
2246          }          }
2247    
2248        new_free(re);        new_free(re);
2249        if (extra != NULL) new_free(extra);        if (extra != NULL) pcre_free_study(extra);
2250        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2251            {
2252            new_free((void *)tables);
2253            setlocale(LC_CTYPE, "C");
2254            locale_set = 0;
2255            }
2256        continue;  /* With next regex */        continue;  /* With next regex */
2257        }        }
2258      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1609  while (!done) Line 2269  while (!done)
2269      int callout_data_set = 0;      int callout_data_set = 0;
2270      int count, c;      int count, c;
2271      int copystrings = 0;      int copystrings = 0;
2272      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2273      int getstrings = 0;      int getstrings = 0;
2274      int getlist = 0;      int getlist = 0;
2275      int gmatched = 0;      int gmatched = 0;
2276      int start_offset = 0;      int start_offset = 0;
2277        int start_offset_sign = 1;
2278      int g_notempty = 0;      int g_notempty = 0;
2279      int use_dfa = 0;      int use_dfa = 0;
2280    
# Line 1627  while (!done) Line 2288  while (!done)
2288    
2289      pcre_callout = callout;      pcre_callout = callout;
2290      first_callout = 1;      first_callout = 1;
2291        last_callout_mark = NULL;
2292      callout_extra = 0;      callout_extra = 0;
2293      callout_count = 0;      callout_count = 0;
2294      callout_fail_count = 999999;      callout_fail_count = 999999;
# Line 1639  while (!done) Line 2301  while (!done)
2301      len = 0;      len = 0;
2302      for (;;)      for (;;)
2303        {        {
2304        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2305          {          {
2306          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2307              {
2308              fprintf(outfile, "\n");
2309              break;
2310              }
2311          done = 1;          done = 1;
2312          goto CONTINUE;          goto CONTINUE;
2313          }          }
# Line 1708  while (!done) Line 2373  while (!done)
2373              {              {
2374              unsigned char buff8[8];              unsigned char buff8[8];
2375              int ii, utn;              int ii, utn;
2376              utn = ord2utf8(c, buff8);              if (use_utf8)
2377              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2378              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2379                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2380                  c = buff8[ii];   /* Last byte */
2381                  }
2382                else
2383                 {
2384                 if (c > 255)
2385                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2386                     "UTF-8 mode is not enabled.\n"
2387                     "** Truncation will probably give the wrong result.\n", c);
2388                 }
2389              p = pt + 1;              p = pt + 1;
2390              break;              break;
2391              }              }
# Line 1733  while (!done) Line 2408  while (!done)
2408          continue;          continue;
2409    
2410          case '>':          case '>':
2411            if (*p == '-')
2412              {
2413              start_offset_sign = -1;
2414              p++;
2415              }
2416          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2417            start_offset *= start_offset_sign;
2418          continue;          continue;
2419    
2420          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1806  while (!done) Line 2487  while (!done)
2487  #endif  #endif
2488            use_dfa = 1;            use_dfa = 1;
2489          continue;          continue;
2490    #endif
2491    
2492    #if !defined NODFA
2493          case 'F':          case 'F':
2494          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2495          continue;          continue;
# Line 1830  while (!done) Line 2513  while (!done)
2513            getnamesptr = npp;            getnamesptr = npp;
2514            }            }
2515          continue;          continue;
2516    
2517            case 'J':
2518            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2519            if (extra != NULL
2520                && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
2521                && extra->executable_jit != NULL)
2522              {
2523              if (jit_stack != NULL) pcre_jit_stack_free(jit_stack);
2524              jit_stack = pcre_jit_stack_alloc(1, n * 1024);
2525              pcre_assign_jit_stack(extra, jit_callback, jit_stack);
2526              }
2527            continue;
2528    
2529          case 'L':          case 'L':
2530          getlist = 1;          getlist = 1;
# Line 1840  while (!done) Line 2535  while (!done)
2535          continue;          continue;
2536    
2537          case 'N':          case 'N':
2538          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2539              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2540            else
2541              options |= PCRE_NOTEMPTY;
2542          continue;          continue;
2543    
2544          case 'O':          case 'O':
# Line 1863  while (!done) Line 2561  while (!done)
2561          continue;          continue;
2562    
2563          case 'P':          case 'P':
2564          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2565              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2566          continue;          continue;
2567    
2568          case 'Q':          case 'Q':
# Line 1898  while (!done) Line 2597  while (!done)
2597          show_malloc = 1;          show_malloc = 1;
2598          continue;          continue;
2599    
2600            case 'Y':
2601            options |= PCRE_NO_START_OPTIMIZE;
2602            continue;
2603    
2604          case 'Z':          case 'Z':
2605          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2606          continue;          continue;
# Line 1918  while (!done) Line 2621  while (!done)
2621        *q++ = c;        *q++ = c;
2622        }        }
2623      *q = 0;      *q = 0;
2624      len = q - dbuffer;      len = (int)(q - dbuffer);
2625    
2626        /* Move the data to the end of the buffer so that a read over the end of
2627        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2628        we are using the POSIX interface, we must include the terminating zero. */
2629    
2630    #if !defined NOPOSIX
2631        if (posix || do_posix)
2632          {
2633          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2634          bptr += buffer_size - len - 1;
2635          }
2636        else
2637    #endif
2638          {
2639          memmove(bptr + buffer_size - len, bptr, len);
2640          bptr += buffer_size - len;
2641          }
2642    
2643      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2644        {        {
# Line 1939  while (!done) Line 2659  while (!done)
2659          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2660        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2661        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2662          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2663    
2664        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2665    
# Line 1963  while (!done) Line 2684  while (!done)
2684              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2685                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2686              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2687              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2688                {                {
2689                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2690                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2691                  outfile);                  outfile);
2692                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 1983  while (!done) Line 2704  while (!done)
2704    
2705      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2706        {        {
2707          markptr = NULL;
2708    
2709        if (timeitm > 0)        if (timeitm > 0)
2710          {          {
2711          register int i;          register int i;
# Line 1994  while (!done) Line 2717  while (!done)
2717            {            {
2718            int workspace[1000];            int workspace[1000];
2719            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2720              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2721                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2722                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2723            }            }
# Line 2013  while (!done) Line 2736  while (!done)
2736    
2737        /* 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
2738        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
2739        for the recursion limit. */        for the recursion limit. The match limits are relevant only to the normal
2740          running of pcre_exec(), so disable the JIT optimization. This makes it
2741          possible to run the same set of tests with and without JIT externally
2742          requested. */
2743    
2744        if (find_match_limit)        if (find_match_limit)
2745          {          {
# Line 2022  while (!done) Line 2748  while (!done)
2748            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2749            extra->flags = 0;            extra->flags = 0;
2750            }            }
2751            else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
2752    
2753          (void)check_match_limit(re, extra, bptr, len, start_offset,          (void)check_match_limit(re, extra, bptr, len, start_offset,
2754            options|g_notempty, use_offsets, use_size_offsets,            options|g_notempty, use_offsets, use_size_offsets,
2755            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
# Line 2057  while (!done) Line 2784  while (!done)
2784        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2785          {          {
2786          int workspace[1000];          int workspace[1000];
2787          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2788            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2789            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2790          if (count == 0)          if (count == 0)
# Line 2105  while (!done) Line 2832  while (!done)
2832              }              }
2833            }            }
2834    
2835            /* do_allcaps requests showing of all captures in the pattern, to check
2836            unset ones at the end. */
2837    
2838            if (do_allcaps)
2839              {
2840              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2841              count++;   /* Allow for full match */
2842              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2843              }
2844    
2845            /* Output the captured substrings */
2846    
2847          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2848            {            {
2849            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2850                {
2851                if (use_offsets[i] != -1)
2852                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2853                    use_offsets[i], i);
2854                if (use_offsets[i+1] != -1)
2855                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2856                    use_offsets[i+1], i+1);
2857              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2858                }
2859            else            else
2860              {              {
2861              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2862              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2863                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2864              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2865              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2866                {                {
2867                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2868                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2869                  fprintf(outfile, " 0+ ");                  outfile);
2870                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2871                }                }
2872              }              }
2873            }            }
2874    
2875            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2876    
2877          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2878            {            {
2879            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2201  while (!done) Line 2947  while (!done)
2947                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);
2948              if (stringlist[i] != NULL)              if (stringlist[i] != NULL)
2949                fprintf(outfile, "string list not terminated by NULL\n");                fprintf(outfile, "string list not terminated by NULL\n");
             /* free((void *)stringlist); */  
2950              pcre_free_substring_list(stringlist);              pcre_free_substring_list(stringlist);
2951              }              }
2952            }            }
# Line 2211  while (!done) Line 2956  while (!done)
2956    
2957        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2958          {          {
2959          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2960  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2961          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2962            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2963              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2964  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2965                outfile);
2966              }
2967          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2968          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2969          }          }
# Line 2226  while (!done) Line 2973  while (!done)
2973        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
2974        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2975    
2976        Complication arises in the case when the newline option is "any" or        Complication arises in the case when the newline convention is "any",
2977        "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
2978        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,
2979        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
2980        Fudge the offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2981          newline setting in the pattern; if none was set, use pcre_config() to
2982          find the default.
2983    
2984        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
2985        character, not one byte. */        character, not one byte. */
# Line 2246  while (!done) Line 2995  while (!done)
2995              {              {
2996              int d;              int d;
2997              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2998              obits = (d == '\r')? PCRE_NEWLINE_CR :              /* Note that these values are always the ASCII ones, even in
2999                      (d == '\n')? PCRE_NEWLINE_LF :              EBCDIC environments. CR = 13, NL = 10. */
3000                      (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :              obits = (d == 13)? PCRE_NEWLINE_CR :
3001                        (d == 10)? PCRE_NEWLINE_LF :
3002                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
3003                      (d == -2)? PCRE_NEWLINE_ANYCRLF :                      (d == -2)? PCRE_NEWLINE_ANYCRLF :
3004                      (d == -1)? PCRE_NEWLINE_ANY : 0;                      (d == -1)? PCRE_NEWLINE_ANY : 0;
3005              }              }
3006            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
3007                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
3008                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
3009                &&                &&
3010                start_offset < len - 1 &&                start_offset < len - 1 &&
# Line 2263  while (!done) Line 3015  while (!done)
3015              {              {
3016              while (start_offset + onechar < len)              while (start_offset + onechar < len)
3017                {                {
3018                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
3019                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
3020                }                }
3021              }              }
3022            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
3023            }            }
3024          else          else
3025            {            {
3026            if (count == PCRE_ERROR_NOMATCH)            switch(count)
3027              {              {
3028              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
3029                if (gmatched == 0)
3030                  {
3031                  if (markptr == NULL) fprintf(outfile, "No match\n");
3032                    else fprintf(outfile, "No match, mark = %s\n", markptr);
3033                  }
3034                break;
3035    
3036                case PCRE_ERROR_BADUTF8:
3037                case PCRE_ERROR_SHORTUTF8:
3038                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
3039                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
3040                if (use_size_offsets >= 2)
3041                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
3042                    use_offsets[1]);
3043                fprintf(outfile, "\n");
3044                break;
3045    
3046                default:
3047                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
3048                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
3049                else
3050                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
3051                break;
3052              }              }
3053            else fprintf(outfile, "Error %d\n", count);  
3054            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
3055            }            }
3056          }          }
# Line 2287  while (!done) Line 3060  while (!done)
3060        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
3061    
3062        /* 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
3063        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
3064        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
3065        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
3066        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
3067        character. */        character. */
3068    
# Line 2298  while (!done) Line 3071  while (!done)
3071        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
3072          {          {
3073          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
3074          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
3075          }          }
3076    
3077        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2324  while (!done) Line 3097  while (!done)
3097  #endif  #endif
3098    
3099    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
3100    if (extra != NULL) new_free(extra);    if (extra != NULL) pcre_free_study(extra);
3101    if (tables != NULL)    if (locale_set)
3102      {      {
3103      new_free((void *)tables);      new_free((void *)tables);
3104      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
3105      locale_set = 0;      locale_set = 0;
3106      }      }
3107      if (jit_stack != NULL)
3108        {
3109        pcre_jit_stack_free(jit_stack);
3110        jit_stack = NULL;
3111        }
3112    }    }
3113    
3114  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");

Legend:
Removed from v.199  
changed lines
  Added in v.676

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12