/[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 121 by ph10, Mon Mar 12 12:12:47 2007 UTC revision 654 by ph10, Tue Aug 2 11:00:40 2011 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 44  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51    #ifdef SUPPORT_LIBREADLINE
52    #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57    #endif
58    
59    
60  /* A number of things vary for Windows builds. Originally, pcretest opened its  /* A number of things vary for Windows builds. Originally, pcretest opened its
61  input and output without "b"; then I was told that "b" was needed in some  input and output without "b"; then I was told that "b" was needed in some
# Line 59  input mode under Windows. */ Line 71  input mode under Windows. */
71  #define INPUT_MODE   "r"  #define INPUT_MODE   "r"
72  #define OUTPUT_MODE  "wb"  #define OUTPUT_MODE  "wb"
73    
74    #ifndef isatty
75    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
76    #endif                         /* though in some environments they seem to   */
77                                   /* be already defined, hence the #ifndefs.    */
78    #ifndef fileno
79    #define fileno _fileno
80    #endif
81    
82    /* A user sent this fix for Borland Builder 5 under Windows. */
83    
84    #ifdef __BORLANDC__
85    #define _setmode(handle, mode) setmode(handle, mode)
86    #endif
87    
88    /* Not Windows */
89    
90  #else  #else
91  #include <sys/time.h>          /* These two includes are needed */  #include <sys/time.h>          /* These two includes are needed */
92  #include <sys/resource.h>      /* for setrlimit(). */  #include <sys/resource.h>      /* for setrlimit(). */
# Line 67  input mode under Windows. */ Line 95  input mode under Windows. */
95  #endif  #endif
96    
97    
98  #define PCRE_SPY        /* For Win32 build, import data, not export */  /* We have to include pcre_internal.h because we need the internal info for
99    displaying the results of pcre_study() and we also need to know about the
100  /* We include pcre_internal.h because we need the internal info for displaying  internal macros, structures, and other internal data values; pcretest has
101  the results of pcre_study() and we also need to know about the internal  "inside information" compared to a program that strictly follows the PCRE API.
102  macros, structures, and other internal data values; pcretest has "inside  
103  information" compared to a program that strictly follows the PCRE API. */  Although pcre_internal.h does itself include pcre.h, we explicitly include it
104    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
105    appropriately for an application, not for building PCRE. */
106    
107    #include "pcre.h"
108  #include "pcre_internal.h"  #include "pcre_internal.h"
109    
110  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* We need access to some of the data tables that PCRE uses. So as not to have
111  two copies, we include the source file here, changing the names of the external  to keep two copies, we include the source file here, changing the names of the
112  symbols to prevent clashes. */  external symbols to prevent clashes. */
113    
114    #define _pcre_ucp_gentype      ucp_gentype
115  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
116  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
117  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 87  symbols to prevent clashes. */ Line 119  symbols to prevent clashes. */
119  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
120  #define _pcre_utt              utt  #define _pcre_utt              utt
121  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
122    #define _pcre_utt_names        utt_names
123  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
124    
125  #include "pcre_tables.c"  #include "pcre_tables.c"
126    
127  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
128  patterns. This function is in a separate file so that it can be included in  patterns. This function is in a separate file so that it can be included in
129  pcre_compile.c when that module is compiled with debugging enabled.  pcre_compile.c when that module is compiled with debugging enabled. It needs to
130    know which case is being compiled. */
 The definition of the macro PRINTABLE, which determines whether to print an  
 output character as-is or as a hex value when showing compiled patterns, is  
 contained in this file. We uses it here also, in cases when the locale has not  
 been explicitly changed, so as to get consistent output from systems that  
 differ in their output from isprint() even in the "C" locale. */  
131    
132    #define COMPILING_PCRETEST
133  #include "pcre_printint.src"  #include "pcre_printint.src"
134    
135  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  /* The definition of the macro PRINTABLE, which determines whether to print an
136    output character as-is or as a hex value when showing compiled patterns, is
137    contained in the printint.src file. We uses it here also, in cases when the
138    locale has not been explicitly changed, so as to get consistent output from
139    systems that differ in their output from isprint() even in the "C" locale. */
140    
141    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
142    
143  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
144  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 149  static int callout_count; Line 183  static int callout_count;
183  static int callout_extra;  static int callout_extra;
184  static int callout_fail_count;  static int callout_fail_count;
185  static int callout_fail_id;  static int callout_fail_id;
186    static int debug_lengths;
187  static int first_callout;  static int first_callout;
188  static int locale_set = 0;  static int locale_set = 0;
189  static int show_malloc;  static int show_malloc;
190  static int use_utf8;  static int use_utf8;
191  static size_t gotten_store;  static size_t gotten_store;
192    static const unsigned char *last_callout_mark = NULL;
193    
194  /* The buffers grow automatically if very long input lines are encountered. */  /* The buffers grow automatically if very long input lines are encountered. */
195    
# Line 162  static uschar *buffer = NULL; Line 198  static uschar *buffer = NULL;
198  static uschar *dbuffer = NULL;  static uschar *dbuffer = NULL;
199  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
200    
201    /* Textual explanations for runtime error codes */
202    
203    static const char *errtexts[] = {
204      NULL,  /* 0 is no error */
205      NULL,  /* NOMATCH is handled specially */
206      "NULL argument passed",
207      "bad option value",
208      "magic number missing",
209      "unknown opcode - pattern overwritten?",
210      "no more memory",
211      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
212      "match limit exceeded",
213      "callout error code",
214      NULL,  /* BADUTF8 is handled specially */
215      "bad UTF-8 offset",
216      NULL,  /* PARTIAL is handled specially */
217      "not used - internal error",
218      "internal error - pattern overwritten?",
219      "bad count value",
220      "item unsupported for DFA matching",
221      "backreference condition or recursion test not supported for DFA matching",
222      "match limit not supported for DFA matching",
223      "workspace size exceeded in DFA matching",
224      "too much recursion for DFA matching",
225      "recursion limit exceeded",
226      "not used - internal error",
227      "invalid combination of newline options",
228      "bad offset value",
229      NULL,  /* SHORTUTF8 is handled specially */
230      "nested recursion at the same subject position"
231    };
232    
233    
234    /*************************************************
235    *         Alternate character tables             *
236    *************************************************/
237    
238    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
239    using the default tables of the library. However, the T option can be used to
240    select alternate sets of tables, for different kinds of testing. Note also that
241    the L (locale) option also adjusts the tables. */
242    
243    /* This is the set of tables distributed as default with PCRE. It recognizes
244    only ASCII characters. */
245    
246    static const unsigned char tables0[] = {
247    
248    /* This table is a lower casing table. */
249    
250        0,  1,  2,  3,  4,  5,  6,  7,
251        8,  9, 10, 11, 12, 13, 14, 15,
252       16, 17, 18, 19, 20, 21, 22, 23,
253       24, 25, 26, 27, 28, 29, 30, 31,
254       32, 33, 34, 35, 36, 37, 38, 39,
255       40, 41, 42, 43, 44, 45, 46, 47,
256       48, 49, 50, 51, 52, 53, 54, 55,
257       56, 57, 58, 59, 60, 61, 62, 63,
258       64, 97, 98, 99,100,101,102,103,
259      104,105,106,107,108,109,110,111,
260      112,113,114,115,116,117,118,119,
261      120,121,122, 91, 92, 93, 94, 95,
262       96, 97, 98, 99,100,101,102,103,
263      104,105,106,107,108,109,110,111,
264      112,113,114,115,116,117,118,119,
265      120,121,122,123,124,125,126,127,
266      128,129,130,131,132,133,134,135,
267      136,137,138,139,140,141,142,143,
268      144,145,146,147,148,149,150,151,
269      152,153,154,155,156,157,158,159,
270      160,161,162,163,164,165,166,167,
271      168,169,170,171,172,173,174,175,
272      176,177,178,179,180,181,182,183,
273      184,185,186,187,188,189,190,191,
274      192,193,194,195,196,197,198,199,
275      200,201,202,203,204,205,206,207,
276      208,209,210,211,212,213,214,215,
277      216,217,218,219,220,221,222,223,
278      224,225,226,227,228,229,230,231,
279      232,233,234,235,236,237,238,239,
280      240,241,242,243,244,245,246,247,
281      248,249,250,251,252,253,254,255,
282    
283    /* This table is a case flipping table. */
284    
285        0,  1,  2,  3,  4,  5,  6,  7,
286        8,  9, 10, 11, 12, 13, 14, 15,
287       16, 17, 18, 19, 20, 21, 22, 23,
288       24, 25, 26, 27, 28, 29, 30, 31,
289       32, 33, 34, 35, 36, 37, 38, 39,
290       40, 41, 42, 43, 44, 45, 46, 47,
291       48, 49, 50, 51, 52, 53, 54, 55,
292       56, 57, 58, 59, 60, 61, 62, 63,
293       64, 97, 98, 99,100,101,102,103,
294      104,105,106,107,108,109,110,111,
295      112,113,114,115,116,117,118,119,
296      120,121,122, 91, 92, 93, 94, 95,
297       96, 65, 66, 67, 68, 69, 70, 71,
298       72, 73, 74, 75, 76, 77, 78, 79,
299       80, 81, 82, 83, 84, 85, 86, 87,
300       88, 89, 90,123,124,125,126,127,
301      128,129,130,131,132,133,134,135,
302      136,137,138,139,140,141,142,143,
303      144,145,146,147,148,149,150,151,
304      152,153,154,155,156,157,158,159,
305      160,161,162,163,164,165,166,167,
306      168,169,170,171,172,173,174,175,
307      176,177,178,179,180,181,182,183,
308      184,185,186,187,188,189,190,191,
309      192,193,194,195,196,197,198,199,
310      200,201,202,203,204,205,206,207,
311      208,209,210,211,212,213,214,215,
312      216,217,218,219,220,221,222,223,
313      224,225,226,227,228,229,230,231,
314      232,233,234,235,236,237,238,239,
315      240,241,242,243,244,245,246,247,
316      248,249,250,251,252,253,254,255,
317    
318    /* This table contains bit maps for various character classes. Each map is 32
319    bytes long and the bits run from the least significant end of each byte. The
320    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
321    graph, print, punct, and cntrl. Other classes are built from combinations. */
322    
323      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
324      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
325      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
326      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
327    
328      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
329      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
330      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
331      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
332    
333      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
334      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
335      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
336      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
337    
338      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
339      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
340      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
341      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
342    
343      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
344      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
345      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
346      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
347    
348      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
349      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
350      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
351      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
352    
353      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
354      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
355      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
356      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
357    
358      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
359      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
360      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
361      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
362    
363      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
364      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
365      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
366      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
367    
368      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
369      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
370      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
371      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
372    
373    /* This table identifies various classes of character by individual bits:
374      0x01   white space character
375      0x02   letter
376      0x04   decimal digit
377      0x08   hexadecimal digit
378      0x10   alphanumeric or '_'
379      0x80   regular expression metacharacter or binary zero
380    */
381    
382      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
383      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
384      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
385      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
386      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
387      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
388      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
389      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
390      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
391      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
392      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
393      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
394      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
395      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
396      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
397      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
398      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
399      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
400      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
401      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
412      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
413      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
414    
415    /* This is a set of tables that came orginally from a Windows user. It seems to
416    be at least an approximation of ISO 8859. In particular, there are characters
417    greater than 128 that are marked as spaces, letters, etc. */
418    
419    static const unsigned char tables1[] = {
420    0,1,2,3,4,5,6,7,
421    8,9,10,11,12,13,14,15,
422    16,17,18,19,20,21,22,23,
423    24,25,26,27,28,29,30,31,
424    32,33,34,35,36,37,38,39,
425    40,41,42,43,44,45,46,47,
426    48,49,50,51,52,53,54,55,
427    56,57,58,59,60,61,62,63,
428    64,97,98,99,100,101,102,103,
429    104,105,106,107,108,109,110,111,
430    112,113,114,115,116,117,118,119,
431    120,121,122,91,92,93,94,95,
432    96,97,98,99,100,101,102,103,
433    104,105,106,107,108,109,110,111,
434    112,113,114,115,116,117,118,119,
435    120,121,122,123,124,125,126,127,
436    128,129,130,131,132,133,134,135,
437    136,137,138,139,140,141,142,143,
438    144,145,146,147,148,149,150,151,
439    152,153,154,155,156,157,158,159,
440    160,161,162,163,164,165,166,167,
441    168,169,170,171,172,173,174,175,
442    176,177,178,179,180,181,182,183,
443    184,185,186,187,188,189,190,191,
444    224,225,226,227,228,229,230,231,
445    232,233,234,235,236,237,238,239,
446    240,241,242,243,244,245,246,215,
447    248,249,250,251,252,253,254,223,
448    224,225,226,227,228,229,230,231,
449    232,233,234,235,236,237,238,239,
450    240,241,242,243,244,245,246,247,
451    248,249,250,251,252,253,254,255,
452    0,1,2,3,4,5,6,7,
453    8,9,10,11,12,13,14,15,
454    16,17,18,19,20,21,22,23,
455    24,25,26,27,28,29,30,31,
456    32,33,34,35,36,37,38,39,
457    40,41,42,43,44,45,46,47,
458    48,49,50,51,52,53,54,55,
459    56,57,58,59,60,61,62,63,
460    64,97,98,99,100,101,102,103,
461    104,105,106,107,108,109,110,111,
462    112,113,114,115,116,117,118,119,
463    120,121,122,91,92,93,94,95,
464    96,65,66,67,68,69,70,71,
465    72,73,74,75,76,77,78,79,
466    80,81,82,83,84,85,86,87,
467    88,89,90,123,124,125,126,127,
468    128,129,130,131,132,133,134,135,
469    136,137,138,139,140,141,142,143,
470    144,145,146,147,148,149,150,151,
471    152,153,154,155,156,157,158,159,
472    160,161,162,163,164,165,166,167,
473    168,169,170,171,172,173,174,175,
474    176,177,178,179,180,181,182,183,
475    184,185,186,187,188,189,190,191,
476    224,225,226,227,228,229,230,231,
477    232,233,234,235,236,237,238,239,
478    240,241,242,243,244,245,246,215,
479    248,249,250,251,252,253,254,223,
480    192,193,194,195,196,197,198,199,
481    200,201,202,203,204,205,206,207,
482    208,209,210,211,212,213,214,247,
483    216,217,218,219,220,221,222,255,
484    0,62,0,0,1,0,0,0,
485    0,0,0,0,0,0,0,0,
486    32,0,0,0,1,0,0,0,
487    0,0,0,0,0,0,0,0,
488    0,0,0,0,0,0,255,3,
489    126,0,0,0,126,0,0,0,
490    0,0,0,0,0,0,0,0,
491    0,0,0,0,0,0,0,0,
492    0,0,0,0,0,0,255,3,
493    0,0,0,0,0,0,0,0,
494    0,0,0,0,0,0,12,2,
495    0,0,0,0,0,0,0,0,
496    0,0,0,0,0,0,0,0,
497    254,255,255,7,0,0,0,0,
498    0,0,0,0,0,0,0,0,
499    255,255,127,127,0,0,0,0,
500    0,0,0,0,0,0,0,0,
501    0,0,0,0,254,255,255,7,
502    0,0,0,0,0,4,32,4,
503    0,0,0,128,255,255,127,255,
504    0,0,0,0,0,0,255,3,
505    254,255,255,135,254,255,255,7,
506    0,0,0,0,0,4,44,6,
507    255,255,127,255,255,255,127,255,
508    0,0,0,0,254,255,255,255,
509    255,255,255,255,255,255,255,127,
510    0,0,0,0,254,255,255,255,
511    255,255,255,255,255,255,255,255,
512    0,2,0,0,255,255,255,255,
513    255,255,255,255,255,255,255,127,
514    0,0,0,0,255,255,255,255,
515    255,255,255,255,255,255,255,255,
516    0,0,0,0,254,255,0,252,
517    1,0,0,248,1,0,0,120,
518    0,0,0,0,254,255,255,255,
519    0,0,128,0,0,0,128,0,
520    255,255,255,255,0,0,0,0,
521    0,0,0,0,0,0,0,128,
522    255,255,255,255,0,0,0,0,
523    0,0,0,0,0,0,0,0,
524    128,0,0,0,0,0,0,0,
525    0,1,1,0,1,1,0,0,
526    0,0,0,0,0,0,0,0,
527    0,0,0,0,0,0,0,0,
528    1,0,0,0,128,0,0,0,
529    128,128,128,128,0,0,128,0,
530    28,28,28,28,28,28,28,28,
531    28,28,0,0,0,0,0,128,
532    0,26,26,26,26,26,26,18,
533    18,18,18,18,18,18,18,18,
534    18,18,18,18,18,18,18,18,
535    18,18,18,128,128,0,128,16,
536    0,26,26,26,26,26,26,18,
537    18,18,18,18,18,18,18,18,
538    18,18,18,18,18,18,18,18,
539    18,18,18,128,128,0,0,0,
540    0,0,0,0,0,1,0,0,
541    0,0,0,0,0,0,0,0,
542    0,0,0,0,0,0,0,0,
543    0,0,0,0,0,0,0,0,
544    1,0,0,0,0,0,0,0,
545    0,0,18,0,0,0,0,0,
546    0,0,20,20,0,18,0,0,
547    0,20,18,0,0,0,0,0,
548    18,18,18,18,18,18,18,18,
549    18,18,18,18,18,18,18,18,
550    18,18,18,18,18,18,18,0,
551    18,18,18,18,18,18,18,18,
552    18,18,18,18,18,18,18,18,
553    18,18,18,18,18,18,18,18,
554    18,18,18,18,18,18,18,0,
555    18,18,18,18,18,18,18,18
556    };
557    
558    
559    
560    
561    #ifndef HAVE_STRERROR
562    /*************************************************
563    *     Provide strerror() for non-ANSI libraries  *
564    *************************************************/
565    
566    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
567    in their libraries, but can provide the same facility by this simple
568    alternative function. */
569    
570    extern int   sys_nerr;
571    extern char *sys_errlist[];
572    
573    char *
574    strerror(int n)
575    {
576    if (n < 0 || n >= sys_nerr) return "unknown error number";
577    return sys_errlist[n];
578    }
579    #endif /* HAVE_STRERROR */
580    
581    
582    
583    
584  /*************************************************  /*************************************************
# Line 180  optimal way of handling this, but hey, t Line 597  optimal way of handling this, but hey, t
597  Arguments:  Arguments:
598    f            the file to read    f            the file to read
599    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
600      prompt       for stdin or readline()
601    
602  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
603                 could be a copy of start, or could be moved                 could be a copy of start, or could be moved
# Line 187  Returns:       pointer to the start of n Line 605  Returns:       pointer to the start of n
605  */  */
606    
607  static uschar *  static uschar *
608  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
609  {  {
610  uschar *here = start;  uschar *here = start;
611    
612  for (;;)  for (;;)
613    {    {
614    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
615    
616    if (rlen > 1000)    if (rlen > 1000)
617      {      {
618      int dlen;      int dlen;
619      if (fgets((char *)here, rlen,  f) == NULL)  
620        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
621        input is a terminal. Note that readline() removes the trailing newline, so
622        we must put it back again, to be compatible with fgets(). */
623    
624    #ifdef SUPPORT_LIBREADLINE
625        if (isatty(fileno(f)))
626          {
627          size_t len;
628          char *s = readline(prompt);
629          if (s == NULL) return (here == start)? NULL : start;
630          len = strlen(s);
631          if (len > 0) add_history(s);
632          if (len > rlen - 1) len = rlen - 1;
633          memcpy(here, s, len);
634          here[len] = '\n';
635          here[len+1] = 0;
636          free(s);
637          }
638        else
639    #endif
640    
641        /* Read the next line by normal means, prompting if the file is stdin. */
642    
643          {
644          if (f == stdin) printf("%s", prompt);
645          if (fgets((char *)here, rlen,  f) == NULL)
646            return (here == start)? NULL : start;
647          }
648    
649      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
650      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
651      here += dlen;      here += dlen;
# Line 515  fprintf(outfile, "%.*s", (cb->next_item_ Line 961  fprintf(outfile, "%.*s", (cb->next_item_
961  fprintf(outfile, "\n");  fprintf(outfile, "\n");
962  first_callout = 0;  first_callout = 0;
963    
964    if (cb->mark != last_callout_mark)
965      {
966      fprintf(outfile, "Latest Mark: %s\n",
967        (cb->mark == NULL)? "<unset>" : (char *)(cb->mark));
968      last_callout_mark = cb->mark;
969      }
970    
971  if (cb->callout_data != NULL)  if (cb->callout_data != NULL)
972    {    {
973    int callout_data = *((int *)(cb->callout_data));    int callout_data = *((int *)(cb->callout_data));
# Line 656  return count; Line 1109  return count;
1109    
1110    
1111  /*************************************************  /*************************************************
1112    *         Case-independent strncmp() function    *
1113    *************************************************/
1114    
1115    /*
1116    Arguments:
1117      s         first string
1118      t         second string
1119      n         number of characters to compare
1120    
1121    Returns:    < 0, = 0, or > 0, according to the comparison
1122    */
1123    
1124    static int
1125    strncmpic(uschar *s, uschar *t, int n)
1126    {
1127    while (n--)
1128      {
1129      int c = tolower(*s++) - tolower(*t++);
1130      if (c) return c;
1131      }
1132    return 0;
1133    }
1134    
1135    
1136    
1137    /*************************************************
1138  *         Check newline indicator                *  *         Check newline indicator                *
1139  *************************************************/  *************************************************/
1140    
1141  /* 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
1142  xxx is LF, CR, CRLF, or ANY. Print a message and return 0 if there is no match.  a message and return 0 if there is no match.
1143    
1144  Arguments:  Arguments:
1145    p           points after the leading '<'    p           points after the leading '<'
# Line 672  Returns:      appropriate PCRE_NEWLINE_x Line 1151  Returns:      appropriate PCRE_NEWLINE_x
1151  static int  static int
1152  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
1153  {  {
1154  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1155  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1156  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1157  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1158    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1159    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1160    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1161  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
1162  return 0;  return 0;
1163  }  }
# Line 689  return 0; Line 1171  return 0;
1171  static void  static void
1172  usage(void)  usage(void)
1173  {  {
1174  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1175    printf("Input and output default to stdin and stdout.\n");
1176    #ifdef SUPPORT_LIBREADLINE
1177    printf("If input is a terminal, readline() is used to read from it.\n");
1178    #else
1179    printf("This version of pcretest is not linked with readline().\n");
1180    #endif
1181    printf("\nOptions:\n");
1182  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
1183  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
1184  printf("  -d       debug: show compiled code and information (-b and -i)\n");  printf("  -d       debug: show compiled code and information (-b and -i)\n");
# Line 698  printf("  -dfa     force DFA matching fo Line 1187  printf("  -dfa     force DFA matching fo
1187  #endif  #endif
1188  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
1189  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
1190           "  -M       find MATCH_LIMIT minimum for each subject\n"
1191         "  -m       output memory used information\n"         "  -m       output memory used information\n"
1192         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
1193  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 705  printf("  -p       use POSIX interface\n Line 1195  printf("  -p       use POSIX interface\n
1195  #endif  #endif
1196  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
1197  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
1198  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied\n"
1199         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
1200  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1201  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 727  int main(int argc, char **argv) Line 1217  int main(int argc, char **argv)
1217  FILE *infile = stdin;  FILE *infile = stdin;
1218  int options = 0;  int options = 0;
1219  int study_options = 0;  int study_options = 0;
1220    int default_find_match_limit = FALSE;
1221  int op = 1;  int op = 1;
1222  int timeit = 0;  int timeit = 0;
1223  int timeitm = 0;  int timeitm = 0;
1224  int showinfo = 0;  int showinfo = 0;
1225  int showstore = 0;  int showstore = 0;
1226    int force_study = 0;
1227  int quiet = 0;  int quiet = 0;
1228  int size_offsets = 45;  int size_offsets = 45;
1229  int size_offsets_max;  int size_offsets_max;
# Line 780  while (argc > 1 && argv[op][0] == '-') Line 1272  while (argc > 1 && argv[op][0] == '-')
1272    {    {
1273    unsigned char *endptr;    unsigned char *endptr;
1274    
1275    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1276      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 1;
1277    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1278    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1279    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1280    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1281      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1282  #if !defined NODFA  #if !defined NODFA
1283    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1284  #endif  #endif
# Line 814  while (argc > 1 && argv[op][0] == '-') Line 1307  while (argc > 1 && argv[op][0] == '-')
1307        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1308          *endptr == 0))          *endptr == 0))
1309      {      {
1310  #if defined(_WIN32) || defined(WIN32)  #if defined(_WIN32) || defined(WIN32) || defined(__minix)
1311      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
1312      exit(1);      exit(1);
1313  #else  #else
# Line 838  while (argc > 1 && argv[op][0] == '-') Line 1331  while (argc > 1 && argv[op][0] == '-')
1331    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1332      {      {
1333      int rc;      int rc;
1334        unsigned long int lrc;
1335      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1336      printf("Compiled with\n");      printf("Compiled with\n");
1337      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 845  while (argc > 1 && argv[op][0] == '-') Line 1339  while (argc > 1 && argv[op][0] == '-')
1339      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1340      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1341      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1342      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1343        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1344        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1345          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1346          (rc == -2)? "ANYCRLF" :
1347        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1348        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1349        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1350                                         "all Unicode newlines");
1351      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1352      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1353      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1354      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1355      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1356      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1357      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1358      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1359      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1360      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1361      goto EXIT;      goto EXIT;
# Line 884  offsets = (int *)malloc(size_offsets_max Line 1384  offsets = (int *)malloc(size_offsets_max
1384  if (offsets == NULL)  if (offsets == NULL)
1385    {    {
1386    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1387      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1388    yield = 1;    yield = 1;
1389    goto EXIT;    goto EXIT;
1390    }    }
# Line 937  while (!done) Line 1437  while (!done)
1437  #endif  #endif
1438    
1439    const char *error;    const char *error;
1440      unsigned char *markptr;
1441    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1442    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1443    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1444    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1445    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1446      int do_allcaps = 0;
1447      int do_mark = 0;
1448    int do_study = 0;    int do_study = 0;
1449      int no_force_study = 0;
1450    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1451    int do_G = 0;    int do_G = 0;
1452    int do_g = 0;    int do_g = 0;
1453    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1454    int do_showrest = 0;    int do_showrest = 0;
1455      int do_showcaprest = 0;
1456    int do_flip = 0;    int do_flip = 0;
1457    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1458    
1459    use_utf8 = 0;    use_utf8 = 0;
1460      debug_lengths = 1;
1461    
1462    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1463    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1464    fflush(outfile);    fflush(outfile);
1465    
# Line 1010  while (!done) Line 1514  while (!done)
1514          }          }
1515        }        }
1516    
1517      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1518        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1519    
1520      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
# Line 1018  while (!done) Line 1522  while (!done)
1522      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1523      use_utf8 = (get_options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1524    
1525      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1526    
1527      if (true_study_size != 0)      if (true_study_size != 0)
1528        {        {
# Line 1055  while (!done) Line 1559  while (!done)
1559    
1560    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1561      {      {
1562      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1563      goto SKIP_DATA;      goto SKIP_DATA;
1564      }      }
1565    
1566    pp = p;    pp = p;
1567    poffset = p - buffer;    poffset = (int)(p - buffer);
1568    
1569    for(;;)    for(;;)
1570      {      {
# Line 1071  while (!done) Line 1575  while (!done)
1575        pp++;        pp++;
1576        }        }
1577      if (*pp != 0) break;      if (*pp != 0) break;
1578      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1579        {        {
1580        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1581        done = 1;        done = 1;
# Line 1115  while (!done) Line 1618  while (!done)
1618        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1619        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1620    
1621        case '+': do_showrest = 1; break;        case '+':
1622          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1623          break;
1624    
1625          case '=': do_allcaps = 1; break;
1626        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1627        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
1628        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1125  while (!done) Line 1632  while (!done)
1632        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1633        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1634        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1635          case 'K': do_mark = 1; break;
1636        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1637        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1638    
# Line 1132  while (!done) Line 1640  while (!done)
1640        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1641  #endif  #endif
1642    
1643        case 'S': do_study = 1; break;        case 'S':
1644          if (do_study == 0) do_study = 1; else
1645            {
1646            do_study = 0;
1647            no_force_study = 1;
1648            }
1649          break;
1650    
1651        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1652          case 'W': options |= PCRE_UCP; break;
1653        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1654        case 'Z': debug_lengths = 0;        case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1655          case 'Z': debug_lengths = 0; break;
1656        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1657        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1658    
1659          case 'T':
1660          switch (*pp++)
1661            {
1662            case '0': tables = tables0; break;
1663            case '1': tables = tables1; break;
1664    
1665            case '\r':
1666            case '\n':
1667            case ' ':
1668            case 0:
1669            fprintf(outfile, "** Missing table number after /T\n");
1670            goto SKIP_DATA;
1671    
1672            default:
1673            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1674            goto SKIP_DATA;
1675            }
1676          break;
1677    
1678        case 'L':        case 'L':
1679        ppp = pp;        ppp = pp;
1680        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1164  while (!done) Line 1700  while (!done)
1700    
1701        case '<':        case '<':
1702          {          {
1703          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1704          if (x == 0) goto SKIP_DATA;            {
1705          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1706          while (*pp++ != '>');            pp += 3;
1707              }
1708            else
1709              {
1710              int x = check_newline(pp, outfile);
1711              if (x == 0) goto SKIP_DATA;
1712              options |= x;
1713              while (*pp++ != '>');
1714              }
1715          }          }
1716        break;        break;
1717    
# Line 1197  while (!done) Line 1741  while (!done)
1741      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1742      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1743      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1744        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1745        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1746    
1747      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1748    
# Line 1217  while (!done) Line 1763  while (!done)
1763  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1764    
1765      {      {
1766        unsigned long int get_options;
1767    
1768      if (timeit > 0)      if (timeit > 0)
1769        {        {
1770        register int i;        register int i;
# Line 1246  while (!done) Line 1794  while (!done)
1794          {          {
1795          for (;;)          for (;;)
1796            {            {
1797            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1798              {              {
1799              done = 1;              done = 1;
1800              goto CONTINUE;              goto CONTINUE;
# Line 1260  while (!done) Line 1808  while (!done)
1808        goto CONTINUE;        goto CONTINUE;
1809        }        }
1810    
1811      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1812      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
1813      returns only limited data. Check that it agrees with the newer one. */      lines. */
1814    
1815        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1816        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1817    
1818        /* Print information if required. There are now two info-returning
1819        functions. The old one has a limited interface and returns only limited
1820        data. Check that it agrees with the newer one. */
1821    
1822      if (log_store)      if (log_store)
1823        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1276  while (!done) Line 1831  while (!done)
1831      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1832      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1833    
1834      /* 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
1835      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1836        suppresses the effect of /S (used for a few test patterns where studying is
1837        never sensible). */
1838    
1839      if (do_study)      if (do_study || (force_study && !no_force_study))
1840        {        {
1841        if (timeit > 0)        if (timeit > 0)
1842          {          {
# Line 1301  while (!done) Line 1858  while (!done)
1858          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1859        }        }
1860    
1861        /* If /K was present, we set up for handling MARK data. */
1862    
1863        if (do_mark)
1864          {
1865          if (extra == NULL)
1866            {
1867            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1868            extra->flags = 0;
1869            }
1870          extra->mark = &markptr;
1871          extra->flags |= PCRE_EXTRA_MARK;
1872          }
1873    
1874      /* 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
1875      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
1876      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 1309  while (!done) Line 1879  while (!done)
1879      if (do_flip)      if (do_flip)
1880        {        {
1881        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1882        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1883            byteflip(rre->magic_number, sizeof(rre->magic_number));
1884        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1885        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1886        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1887        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1888        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1889        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1890        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1891          rre->first_byte =
1892            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1893          rre->req_byte =
1894            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1895          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1896          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1897        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1898          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1899        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1900            sizeof(rre->name_count));
1901    
1902        if (extra != NULL)        if (extra != NULL)
1903          {          {
1904          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1905          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1906          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1907            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1908          }          }
1909        }        }
1910    
# Line 1340  while (!done) Line 1918  while (!done)
1918        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1919        }        }
1920    
1921        /* We already have the options in get_options (see above) */
1922    
1923      if (do_showinfo)      if (do_showinfo)
1924        {        {
1925        unsigned long int get_options, all_options;        unsigned long int all_options;
1926  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1927        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1928  #endif  #endif
1929        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1930            hascrorlf;
1931        int nameentrysize, namecount;        int nameentrysize, namecount;
1932        const uschar *nametable;        const uschar *nametable;
1933    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1934        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1935        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1936        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1359  while (!done) Line 1939  while (!done)
1939        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1940        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1941        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1942          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1943          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1944          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1945    
1946  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1947        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1400  while (!done) Line 1983  while (!done)
1983            }            }
1984          }          }
1985    
1986        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1987        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1988    
1989        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1990        if (do_flip)        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
         {  
         all_options = byteflip(all_options, sizeof(all_options));  
          }  
   
       if ((all_options & PCRE_NOPARTIAL) != 0)  
         fprintf(outfile, "Partial matching not supported\n");  
1991    
1992        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1993          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",
1994            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1995            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1996            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1997            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1998            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1999            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
2000              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
2001              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
2002            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
2003            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
2004            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
2005            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
2006            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
2007              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2008            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2009              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2010            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2011    
2012          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
2013    
2014        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
2015          {          {
2016          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1442  while (!done) Line 2025  while (!done)
2025          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
2026          break;          break;
2027    
2028            case PCRE_NEWLINE_ANYCRLF:
2029            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
2030            break;
2031    
2032          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
2033          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
2034          break;          break;
# Line 1487  while (!done) Line 2074  while (!done)
2074        /* 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
2075        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2076        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
2077        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2078          information unless -i or -d was also present. This means that, except
2079          when auto-callouts are involved, the output from runs with and without
2080          -s should be identical. */
2081    
2082        if (do_study)        if (do_study || (force_study && showinfo && !no_force_study))
2083          {          {
2084          if (extra == NULL)          if (extra == NULL)
2085            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2086          else          else
2087            {            {
2088            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2089            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2090    
2091              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2092              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2093    
2094              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2095            if (start_bits == NULL)            if (start_bits == NULL)
2096              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2097            else            else
2098              {              {
2099              int i;              int i;
# Line 1546  while (!done) Line 2140  while (!done)
2140        else        else
2141          {          {
2142          uschar sbuf[8];          uschar sbuf[8];
2143          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2144          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2145          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2146          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2147    
2148          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2149          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2150          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2151          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2152    
2153          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2154              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1563  while (!done) Line 2157  while (!done)
2157            }            }
2158          else          else
2159            {            {
2160            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2161    
2162              /* If there is study data, write it, but verify the writing only
2163              if the studying was requested by /S, not just by -s. */
2164    
2165            if (extra != NULL)            if (extra != NULL)
2166              {              {
2167              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1573  while (!done) Line 2171  while (!done)
2171                  strerror(errno));                  strerror(errno));
2172                }                }
2173              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
2174              }              }
2175            }            }
2176          fclose(f);          fclose(f);
# Line 1581  while (!done) Line 2178  while (!done)
2178    
2179        new_free(re);        new_free(re);
2180        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2181        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2182            {
2183            new_free((void *)tables);
2184            setlocale(LC_CTYPE, "C");
2185            locale_set = 0;
2186            }
2187        continue;  /* With next regex */        continue;  /* With next regex */
2188        }        }
2189      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1591  while (!done) Line 2193  while (!done)
2193    for (;;)    for (;;)
2194      {      {
2195      uschar *q;      uschar *q;
2196      uschar *bptr = dbuffer;      uschar *bptr;
2197      int *use_offsets = offsets;      int *use_offsets = offsets;
2198      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2199      int callout_data = 0;      int callout_data = 0;
2200      int callout_data_set = 0;      int callout_data_set = 0;
2201      int count, c;      int count, c;
2202      int copystrings = 0;      int copystrings = 0;
2203      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2204      int getstrings = 0;      int getstrings = 0;
2205      int getlist = 0;      int getlist = 0;
2206      int gmatched = 0;      int gmatched = 0;
2207      int start_offset = 0;      int start_offset = 0;
2208        int start_offset_sign = 1;
2209      int g_notempty = 0;      int g_notempty = 0;
2210      int use_dfa = 0;      int use_dfa = 0;
2211    
# Line 1616  while (!done) Line 2219  while (!done)
2219    
2220      pcre_callout = callout;      pcre_callout = callout;
2221      first_callout = 1;      first_callout = 1;
2222        last_callout_mark = NULL;
2223      callout_extra = 0;      callout_extra = 0;
2224      callout_count = 0;      callout_count = 0;
2225      callout_fail_count = 999999;      callout_fail_count = 999999;
# Line 1628  while (!done) Line 2232  while (!done)
2232      len = 0;      len = 0;
2233      for (;;)      for (;;)
2234        {        {
2235        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2236          {          {
2237          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2238              {
2239              fprintf(outfile, "\n");
2240              break;
2241              }
2242          done = 1;          done = 1;
2243          goto CONTINUE;          goto CONTINUE;
2244          }          }
# Line 1647  while (!done) Line 2254  while (!done)
2254      p = buffer;      p = buffer;
2255      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2256    
2257      q = dbuffer;      bptr = q = dbuffer;
2258      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2259        {        {
2260        int i = 0;        int i = 0;
# Line 1697  while (!done) Line 2304  while (!done)
2304              {              {
2305              unsigned char buff8[8];              unsigned char buff8[8];
2306              int ii, utn;              int ii, utn;
2307              utn = ord2utf8(c, buff8);              if (use_utf8)
2308              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2309              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2310                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2311                  c = buff8[ii];   /* Last byte */
2312                  }
2313                else
2314                 {
2315                 if (c > 255)
2316                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2317                     "UTF-8 mode is not enabled.\n"
2318                     "** Truncation will probably give the wrong result.\n", c);
2319                 }
2320              p = pt + 1;              p = pt + 1;
2321              break;              break;
2322              }              }
# Line 1722  while (!done) Line 2339  while (!done)
2339          continue;          continue;
2340    
2341          case '>':          case '>':
2342            if (*p == '-')
2343              {
2344              start_offset_sign = -1;
2345              p++;
2346              }
2347          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2348            start_offset *= start_offset_sign;
2349          continue;          continue;
2350    
2351          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1795  while (!done) Line 2418  while (!done)
2418  #endif  #endif
2419            use_dfa = 1;            use_dfa = 1;
2420          continue;          continue;
2421    #endif
2422    
2423    #if !defined NODFA
2424          case 'F':          case 'F':
2425          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2426          continue;          continue;
# Line 1829  while (!done) Line 2454  while (!done)
2454          continue;          continue;
2455    
2456          case 'N':          case 'N':
2457          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2458              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2459            else
2460              options |= PCRE_NOTEMPTY;
2461          continue;          continue;
2462    
2463          case 'O':          case 'O':
# Line 1842  while (!done) Line 2470  while (!done)
2470            if (offsets == NULL)            if (offsets == NULL)
2471              {              {
2472              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2473                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2474              yield = 1;              yield = 1;
2475              goto EXIT;              goto EXIT;
2476              }              }
# Line 1852  while (!done) Line 2480  while (!done)
2480          continue;          continue;
2481    
2482          case 'P':          case 'P':
2483          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2484              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2485          continue;          continue;
2486    
2487          case 'Q':          case 'Q':
# Line 1887  while (!done) Line 2516  while (!done)
2516          show_malloc = 1;          show_malloc = 1;
2517          continue;          continue;
2518    
2519            case 'Y':
2520            options |= PCRE_NO_START_OPTIMIZE;
2521            continue;
2522    
2523          case 'Z':          case 'Z':
2524          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2525          continue;          continue;
# Line 1907  while (!done) Line 2540  while (!done)
2540        *q++ = c;        *q++ = c;
2541        }        }
2542      *q = 0;      *q = 0;
2543      len = q - dbuffer;      len = (int)(q - dbuffer);
2544    
2545        /* Move the data to the end of the buffer so that a read over the end of
2546        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2547        we are using the POSIX interface, we must include the terminating zero. */
2548    
2549    #if !defined NOPOSIX
2550        if (posix || do_posix)
2551          {
2552          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2553          bptr += buffer_size - len - 1;
2554          }
2555        else
2556    #endif
2557          {
2558          memmove(bptr + buffer_size - len, bptr, len);
2559          bptr += buffer_size - len;
2560          }
2561    
2562      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2563        {        {
# Line 1928  while (!done) Line 2578  while (!done)
2578          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2579        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2580        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2581          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2582    
2583        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2584    
# Line 1952  while (!done) Line 2603  while (!done)
2603              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2604                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2605              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2606              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2607                {                {
2608                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2609                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2610                  outfile);                  outfile);
2611                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 1972  while (!done) Line 2623  while (!done)
2623    
2624      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2625        {        {
2626          markptr = NULL;
2627    
2628        if (timeitm > 0)        if (timeitm > 0)
2629          {          {
2630          register int i;          register int i;
# Line 1983  while (!done) Line 2636  while (!done)
2636            {            {
2637            int workspace[1000];            int workspace[1000];
2638            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2639              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2640                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2641                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2642            }            }
# Line 2046  while (!done) Line 2699  while (!done)
2699        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2700          {          {
2701          int workspace[1000];          int workspace[1000];
2702          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2703            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2704            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2705          if (count == 0)          if (count == 0)
# Line 2094  while (!done) Line 2747  while (!done)
2747              }              }
2748            }            }
2749    
2750            /* do_allcaps requests showing of all captures in the pattern, to check
2751            unset ones at the end. */
2752    
2753            if (do_allcaps)
2754              {
2755              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2756              count++;   /* Allow for full match */
2757              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2758              }
2759    
2760            /* Output the captured substrings */
2761    
2762          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2763            {            {
2764            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2765                {
2766                if (use_offsets[i] != -1)
2767                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2768                    use_offsets[i], i);
2769                if (use_offsets[i+1] != -1)
2770                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2771                    use_offsets[i+1], i+1);
2772              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2773                }
2774            else            else
2775              {              {
2776              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2777              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2778                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2779              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2780              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2781                {                {
2782                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2783                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2784                  fprintf(outfile, " 0+ ");                  outfile);
2785                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2786                }                }
2787              }              }
2788            }            }
2789    
2790            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2791    
2792          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2793            {            {
2794            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2200  while (!done) Line 2872  while (!done)
2872    
2873        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2874          {          {
2875          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2876  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2877          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2878            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2879              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2880  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2881                outfile);
2882              }
2883          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2884          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2885          }          }
2886    
2887        /* Failed to match. If this is a /g or /G loop and we previously set        /* Failed to match. If this is a /g or /G loop and we previously set
2888        g_notempty after a null match, this is not necessarily the end.        g_notempty after a null match, this is not necessarily the end. We want
2889        We want to advance the start offset, and continue. In the case of UTF-8        to advance the start offset, and continue. We won't be at the end of the
2890        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2891        offset values to achieve this. We won't be at the end of the string -  
2892        that was checked before setting g_notempty. */        Complication arises in the case when the newline convention is "any",
2893          "crlf", or "anycrlf". If the previous match was at the end of a line
2894          terminated by CRLF, an advance of one character just passes the \r,
2895          whereas we should prefer the longer newline sequence, as does the code in
2896          pcre_exec(). Fudge the offset value to achieve this. We check for a
2897          newline setting in the pattern; if none was set, use pcre_config() to
2898          find the default.
2899    
2900          Otherwise, in the case of UTF-8 matching, the advance must be one
2901          character, not one byte. */
2902    
2903        else        else
2904          {          {
2905          if (g_notempty != 0)          if (g_notempty != 0)
2906            {            {
2907            int onechar = 1;            int onechar = 1;
2908              unsigned int obits = ((real_pcre *)re)->options;
2909            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2910            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2911                {
2912                int d;
2913                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2914                /* Note that these values are always the ASCII ones, even in
2915                EBCDIC environments. CR = 13, NL = 10. */
2916                obits = (d == 13)? PCRE_NEWLINE_CR :
2917                        (d == 10)? PCRE_NEWLINE_LF :
2918                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2919                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2920                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2921                }
2922              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2923                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2924                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2925                  &&
2926                  start_offset < len - 1 &&
2927                  bptr[start_offset] == '\r' &&
2928                  bptr[start_offset+1] == '\n')
2929                onechar++;
2930              else if (use_utf8)
2931              {              {
2932              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2933                {                {
2934                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2935                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2936                }                }
2937              }              }
2938            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
2939            }            }
2940          else          else
2941            {            {
2942            if (count == PCRE_ERROR_NOMATCH)            switch(count)
2943              {              {
2944              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
2945                if (gmatched == 0)
2946                  {
2947                  if (markptr == NULL) fprintf(outfile, "No match\n");
2948                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2949                  }
2950                break;
2951    
2952                case PCRE_ERROR_BADUTF8:
2953                case PCRE_ERROR_SHORTUTF8:
2954                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
2955                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
2956                if (use_size_offsets >= 2)
2957                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
2958                    use_offsets[1]);
2959                fprintf(outfile, "\n");
2960                break;
2961    
2962                default:
2963                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
2964                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
2965                else
2966                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
2967                break;
2968              }              }
2969            else fprintf(outfile, "Error %d\n", count);  
2970            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
2971            }            }
2972          }          }
# Line 2251  while (!done) Line 2976  while (!done)
2976        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2977    
2978        /* 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
2979        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
2980        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
2981        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
2982        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
2983        character. */        character. */
2984    
2985        g_notempty = 0;        g_notempty = 0;
2986    
2987        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2988          {          {
2989          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2990          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2991          }          }
2992    
2993        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2288  while (!done) Line 3014  while (!done)
3014    
3015    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
3016    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
3017    if (tables != NULL)    if (locale_set)
3018      {      {
3019      new_free((void *)tables);      new_free((void *)tables);
3020      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
Legend:
Removed from v.121  
changed lines
  Added in v.654
  ViewVC Help
Powered by ViewVC 1.1.5