/[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 200 by ph10, Wed Aug 1 09:10:40 2007 UTC revision 566 by ph10, Wed Nov 3 18:32:55 2010 UTC
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37    
38    
39  #ifdef HAVE_CONFIG_H  #ifdef HAVE_CONFIG_H
40  #include <config.h>  #include "config.h"
41  #endif  #endif
42    
43  #include <ctype.h>  #include <ctype.h>
# Line 48  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51    #ifdef SUPPORT_LIBREADLINE
52    #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57    #endif
58    
59    
60  /* A number of things vary for Windows builds. Originally, pcretest opened its  /* A number of things vary for Windows builds. Originally, pcretest opened its
61  input and output without "b"; then I was told that "b" was needed in some  input and output without "b"; then I was told that "b" was needed in some
# Line 63  input mode under Windows. */ Line 71  input mode under Windows. */
71  #define INPUT_MODE   "r"  #define INPUT_MODE   "r"
72  #define OUTPUT_MODE  "wb"  #define OUTPUT_MODE  "wb"
73    
74    #ifndef isatty
75    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
76    #endif                         /* though in some environments they seem to   */
77                                   /* be already defined, hence the #ifndefs.    */
78    #ifndef fileno
79    #define fileno _fileno
80    #endif
81    
82  #else  #else
83  #include <sys/time.h>          /* These two includes are needed */  #include <sys/time.h>          /* These two includes are needed */
84  #include <sys/resource.h>      /* for setrlimit(). */  #include <sys/resource.h>      /* for setrlimit(). */
# Line 83  appropriately for an application, not fo Line 99  appropriately for an application, not fo
99  #include "pcre.h"  #include "pcre.h"
100  #include "pcre_internal.h"  #include "pcre_internal.h"
101    
102  /* 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
103  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
104  symbols to prevent clashes. */  external symbols to prevent clashes. */
105    
106    #define _pcre_ucp_gentype      ucp_gentype
107  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
108  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
109  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 94  symbols to prevent clashes. */ Line 111  symbols to prevent clashes. */
111  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
112  #define _pcre_utt              utt  #define _pcre_utt              utt
113  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
114    #define _pcre_utt_names        utt_names
115  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
116    
117  #include "pcre_tables.c"  #include "pcre_tables.c"
118    
119  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
120  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
121  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
122    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. */  
123    
124    #define COMPILING_PCRETEST
125  #include "pcre_printint.src"  #include "pcre_printint.src"
126    
127  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  /* The definition of the macro PRINTABLE, which determines whether to print an
128    output character as-is or as a hex value when showing compiled patterns, is
129    contained in the printint.src file. We uses it here also, in cases when the
130    locale has not been explicitly changed, so as to get consistent output from
131    systems that differ in their output from isprint() even in the "C" locale. */
132    
133    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
134    
135  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
136  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 175  static int callout_count;
175  static int callout_extra;  static int callout_extra;
176  static int callout_fail_count;  static int callout_fail_count;
177  static int callout_fail_id;  static int callout_fail_id;
178    static int debug_lengths;
179  static int first_callout;  static int first_callout;
180  static int locale_set = 0;  static int locale_set = 0;
181  static int show_malloc;  static int show_malloc;
# Line 170  static uschar *dbuffer = NULL; Line 190  static uschar *dbuffer = NULL;
190  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
191    
192    
193    /*************************************************
194    *         Alternate character tables             *
195    *************************************************/
196    
197    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
198    using the default tables of the library. However, the T option can be used to
199    select alternate sets of tables, for different kinds of testing. Note also that
200    the L (locale) option also adjusts the tables. */
201    
202    /* This is the set of tables distributed as default with PCRE. It recognizes
203    only ASCII characters. */
204    
205    static const unsigned char tables0[] = {
206    
207    /* This table is a lower casing table. */
208    
209        0,  1,  2,  3,  4,  5,  6,  7,
210        8,  9, 10, 11, 12, 13, 14, 15,
211       16, 17, 18, 19, 20, 21, 22, 23,
212       24, 25, 26, 27, 28, 29, 30, 31,
213       32, 33, 34, 35, 36, 37, 38, 39,
214       40, 41, 42, 43, 44, 45, 46, 47,
215       48, 49, 50, 51, 52, 53, 54, 55,
216       56, 57, 58, 59, 60, 61, 62, 63,
217       64, 97, 98, 99,100,101,102,103,
218      104,105,106,107,108,109,110,111,
219      112,113,114,115,116,117,118,119,
220      120,121,122, 91, 92, 93, 94, 95,
221       96, 97, 98, 99,100,101,102,103,
222      104,105,106,107,108,109,110,111,
223      112,113,114,115,116,117,118,119,
224      120,121,122,123,124,125,126,127,
225      128,129,130,131,132,133,134,135,
226      136,137,138,139,140,141,142,143,
227      144,145,146,147,148,149,150,151,
228      152,153,154,155,156,157,158,159,
229      160,161,162,163,164,165,166,167,
230      168,169,170,171,172,173,174,175,
231      176,177,178,179,180,181,182,183,
232      184,185,186,187,188,189,190,191,
233      192,193,194,195,196,197,198,199,
234      200,201,202,203,204,205,206,207,
235      208,209,210,211,212,213,214,215,
236      216,217,218,219,220,221,222,223,
237      224,225,226,227,228,229,230,231,
238      232,233,234,235,236,237,238,239,
239      240,241,242,243,244,245,246,247,
240      248,249,250,251,252,253,254,255,
241    
242    /* This table is a case flipping table. */
243    
244        0,  1,  2,  3,  4,  5,  6,  7,
245        8,  9, 10, 11, 12, 13, 14, 15,
246       16, 17, 18, 19, 20, 21, 22, 23,
247       24, 25, 26, 27, 28, 29, 30, 31,
248       32, 33, 34, 35, 36, 37, 38, 39,
249       40, 41, 42, 43, 44, 45, 46, 47,
250       48, 49, 50, 51, 52, 53, 54, 55,
251       56, 57, 58, 59, 60, 61, 62, 63,
252       64, 97, 98, 99,100,101,102,103,
253      104,105,106,107,108,109,110,111,
254      112,113,114,115,116,117,118,119,
255      120,121,122, 91, 92, 93, 94, 95,
256       96, 65, 66, 67, 68, 69, 70, 71,
257       72, 73, 74, 75, 76, 77, 78, 79,
258       80, 81, 82, 83, 84, 85, 86, 87,
259       88, 89, 90,123,124,125,126,127,
260      128,129,130,131,132,133,134,135,
261      136,137,138,139,140,141,142,143,
262      144,145,146,147,148,149,150,151,
263      152,153,154,155,156,157,158,159,
264      160,161,162,163,164,165,166,167,
265      168,169,170,171,172,173,174,175,
266      176,177,178,179,180,181,182,183,
267      184,185,186,187,188,189,190,191,
268      192,193,194,195,196,197,198,199,
269      200,201,202,203,204,205,206,207,
270      208,209,210,211,212,213,214,215,
271      216,217,218,219,220,221,222,223,
272      224,225,226,227,228,229,230,231,
273      232,233,234,235,236,237,238,239,
274      240,241,242,243,244,245,246,247,
275      248,249,250,251,252,253,254,255,
276    
277    /* This table contains bit maps for various character classes. Each map is 32
278    bytes long and the bits run from the least significant end of each byte. The
279    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
280    graph, print, punct, and cntrl. Other classes are built from combinations. */
281    
282      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
283      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
284      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
285      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
286    
287      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
288      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
289      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
290      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
291    
292      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
293      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
294      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
295      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
296    
297      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
298      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
299      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
300      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
301    
302      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
303      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
304      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
305      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
306    
307      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
308      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
309      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
310      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
311    
312      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
313      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
314      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
315      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
316    
317      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
318      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
319      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
320      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
321    
322      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
323      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
324      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
325      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
326    
327      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
328      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
329      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
330      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
331    
332    /* This table identifies various classes of character by individual bits:
333      0x01   white space character
334      0x02   letter
335      0x04   decimal digit
336      0x08   hexadecimal digit
337      0x10   alphanumeric or '_'
338      0x80   regular expression metacharacter or binary zero
339    */
340    
341      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
342      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
343      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
344      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
345      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
346      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
347      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
348      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
349      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
350      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
351      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
352      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
353      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
354      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
355      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
356      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
357      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
358      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
359      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
360      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
361      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
362      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
363      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
364      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
365      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
366      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
367      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
368      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
369      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
370      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
371      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
372      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
373    
374    /* This is a set of tables that came orginally from a Windows user. It seems to
375    be at least an approximation of ISO 8859. In particular, there are characters
376    greater than 128 that are marked as spaces, letters, etc. */
377    
378    static const unsigned char tables1[] = {
379    0,1,2,3,4,5,6,7,
380    8,9,10,11,12,13,14,15,
381    16,17,18,19,20,21,22,23,
382    24,25,26,27,28,29,30,31,
383    32,33,34,35,36,37,38,39,
384    40,41,42,43,44,45,46,47,
385    48,49,50,51,52,53,54,55,
386    56,57,58,59,60,61,62,63,
387    64,97,98,99,100,101,102,103,
388    104,105,106,107,108,109,110,111,
389    112,113,114,115,116,117,118,119,
390    120,121,122,91,92,93,94,95,
391    96,97,98,99,100,101,102,103,
392    104,105,106,107,108,109,110,111,
393    112,113,114,115,116,117,118,119,
394    120,121,122,123,124,125,126,127,
395    128,129,130,131,132,133,134,135,
396    136,137,138,139,140,141,142,143,
397    144,145,146,147,148,149,150,151,
398    152,153,154,155,156,157,158,159,
399    160,161,162,163,164,165,166,167,
400    168,169,170,171,172,173,174,175,
401    176,177,178,179,180,181,182,183,
402    184,185,186,187,188,189,190,191,
403    224,225,226,227,228,229,230,231,
404    232,233,234,235,236,237,238,239,
405    240,241,242,243,244,245,246,215,
406    248,249,250,251,252,253,254,223,
407    224,225,226,227,228,229,230,231,
408    232,233,234,235,236,237,238,239,
409    240,241,242,243,244,245,246,247,
410    248,249,250,251,252,253,254,255,
411    0,1,2,3,4,5,6,7,
412    8,9,10,11,12,13,14,15,
413    16,17,18,19,20,21,22,23,
414    24,25,26,27,28,29,30,31,
415    32,33,34,35,36,37,38,39,
416    40,41,42,43,44,45,46,47,
417    48,49,50,51,52,53,54,55,
418    56,57,58,59,60,61,62,63,
419    64,97,98,99,100,101,102,103,
420    104,105,106,107,108,109,110,111,
421    112,113,114,115,116,117,118,119,
422    120,121,122,91,92,93,94,95,
423    96,65,66,67,68,69,70,71,
424    72,73,74,75,76,77,78,79,
425    80,81,82,83,84,85,86,87,
426    88,89,90,123,124,125,126,127,
427    128,129,130,131,132,133,134,135,
428    136,137,138,139,140,141,142,143,
429    144,145,146,147,148,149,150,151,
430    152,153,154,155,156,157,158,159,
431    160,161,162,163,164,165,166,167,
432    168,169,170,171,172,173,174,175,
433    176,177,178,179,180,181,182,183,
434    184,185,186,187,188,189,190,191,
435    224,225,226,227,228,229,230,231,
436    232,233,234,235,236,237,238,239,
437    240,241,242,243,244,245,246,215,
438    248,249,250,251,252,253,254,223,
439    192,193,194,195,196,197,198,199,
440    200,201,202,203,204,205,206,207,
441    208,209,210,211,212,213,214,247,
442    216,217,218,219,220,221,222,255,
443    0,62,0,0,1,0,0,0,
444    0,0,0,0,0,0,0,0,
445    32,0,0,0,1,0,0,0,
446    0,0,0,0,0,0,0,0,
447    0,0,0,0,0,0,255,3,
448    126,0,0,0,126,0,0,0,
449    0,0,0,0,0,0,0,0,
450    0,0,0,0,0,0,0,0,
451    0,0,0,0,0,0,255,3,
452    0,0,0,0,0,0,0,0,
453    0,0,0,0,0,0,12,2,
454    0,0,0,0,0,0,0,0,
455    0,0,0,0,0,0,0,0,
456    254,255,255,7,0,0,0,0,
457    0,0,0,0,0,0,0,0,
458    255,255,127,127,0,0,0,0,
459    0,0,0,0,0,0,0,0,
460    0,0,0,0,254,255,255,7,
461    0,0,0,0,0,4,32,4,
462    0,0,0,128,255,255,127,255,
463    0,0,0,0,0,0,255,3,
464    254,255,255,135,254,255,255,7,
465    0,0,0,0,0,4,44,6,
466    255,255,127,255,255,255,127,255,
467    0,0,0,0,254,255,255,255,
468    255,255,255,255,255,255,255,127,
469    0,0,0,0,254,255,255,255,
470    255,255,255,255,255,255,255,255,
471    0,2,0,0,255,255,255,255,
472    255,255,255,255,255,255,255,127,
473    0,0,0,0,255,255,255,255,
474    255,255,255,255,255,255,255,255,
475    0,0,0,0,254,255,0,252,
476    1,0,0,248,1,0,0,120,
477    0,0,0,0,254,255,255,255,
478    0,0,128,0,0,0,128,0,
479    255,255,255,255,0,0,0,0,
480    0,0,0,0,0,0,0,128,
481    255,255,255,255,0,0,0,0,
482    0,0,0,0,0,0,0,0,
483    128,0,0,0,0,0,0,0,
484    0,1,1,0,1,1,0,0,
485    0,0,0,0,0,0,0,0,
486    0,0,0,0,0,0,0,0,
487    1,0,0,0,128,0,0,0,
488    128,128,128,128,0,0,128,0,
489    28,28,28,28,28,28,28,28,
490    28,28,0,0,0,0,0,128,
491    0,26,26,26,26,26,26,18,
492    18,18,18,18,18,18,18,18,
493    18,18,18,18,18,18,18,18,
494    18,18,18,128,128,0,128,16,
495    0,26,26,26,26,26,26,18,
496    18,18,18,18,18,18,18,18,
497    18,18,18,18,18,18,18,18,
498    18,18,18,128,128,0,0,0,
499    0,0,0,0,0,1,0,0,
500    0,0,0,0,0,0,0,0,
501    0,0,0,0,0,0,0,0,
502    0,0,0,0,0,0,0,0,
503    1,0,0,0,0,0,0,0,
504    0,0,18,0,0,0,0,0,
505    0,0,20,20,0,18,0,0,
506    0,20,18,0,0,0,0,0,
507    18,18,18,18,18,18,18,18,
508    18,18,18,18,18,18,18,18,
509    18,18,18,18,18,18,18,0,
510    18,18,18,18,18,18,18,18,
511    18,18,18,18,18,18,18,18,
512    18,18,18,18,18,18,18,18,
513    18,18,18,18,18,18,18,0,
514    18,18,18,18,18,18,18,18
515    };
516    
517    
518    
519    
520    #ifndef HAVE_STRERROR
521    /*************************************************
522    *     Provide strerror() for non-ANSI libraries  *
523    *************************************************/
524    
525    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
526    in their libraries, but can provide the same facility by this simple
527    alternative function. */
528    
529    extern int   sys_nerr;
530    extern char *sys_errlist[];
531    
532    char *
533    strerror(int n)
534    {
535    if (n < 0 || n >= sys_nerr) return "unknown error number";
536    return sys_errlist[n];
537    }
538    #endif /* HAVE_STRERROR */
539    
540    
541    
542    
543  /*************************************************  /*************************************************
544  *        Read or extend an input line            *  *        Read or extend an input line            *
# Line 187  optimal way of handling this, but hey, t Line 556  optimal way of handling this, but hey, t
556  Arguments:  Arguments:
557    f            the file to read    f            the file to read
558    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
559      prompt       for stdin or readline()
560    
561  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
562                 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 564  Returns:       pointer to the start of n
564  */  */
565    
566  static uschar *  static uschar *
567  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
568  {  {
569  uschar *here = start;  uschar *here = start;
570    
571  for (;;)  for (;;)
572    {    {
573    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
574    
575    if (rlen > 1000)    if (rlen > 1000)
576      {      {
577      int dlen;      int dlen;
578      if (fgets((char *)here, rlen,  f) == NULL)  
579        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
580        input is a terminal. Note that readline() removes the trailing newline, so
581        we must put it back again, to be compatible with fgets(). */
582    
583    #ifdef SUPPORT_LIBREADLINE
584        if (isatty(fileno(f)))
585          {
586          size_t len;
587          char *s = readline(prompt);
588          if (s == NULL) return (here == start)? NULL : start;
589          len = strlen(s);
590          if (len > 0) add_history(s);
591          if (len > rlen - 1) len = rlen - 1;
592          memcpy(here, s, len);
593          here[len] = '\n';
594          here[len+1] = 0;
595          free(s);
596          }
597        else
598    #endif
599    
600        /* Read the next line by normal means, prompting if the file is stdin. */
601    
602          {
603          if (f == stdin) printf("%s", prompt);
604          if (fgets((char *)here, rlen,  f) == NULL)
605            return (here == start)? NULL : start;
606          }
607    
608      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
609      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
610      here += dlen;      here += dlen;
# Line 663  return count; Line 1061  return count;
1061    
1062    
1063  /*************************************************  /*************************************************
1064    *         Case-independent strncmp() function    *
1065    *************************************************/
1066    
1067    /*
1068    Arguments:
1069      s         first string
1070      t         second string
1071      n         number of characters to compare
1072    
1073    Returns:    < 0, = 0, or > 0, according to the comparison
1074    */
1075    
1076    static int
1077    strncmpic(uschar *s, uschar *t, int n)
1078    {
1079    while (n--)
1080      {
1081      int c = tolower(*s++) - tolower(*t++);
1082      if (c) return c;
1083      }
1084    return 0;
1085    }
1086    
1087    
1088    
1089    /*************************************************
1090  *         Check newline indicator                *  *         Check newline indicator                *
1091  *************************************************/  *************************************************/
1092    
1093  /* 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
1094  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.  
1095    
1096  Arguments:  Arguments:
1097    p           points after the leading '<'    p           points after the leading '<'
# Line 680  Returns:      appropriate PCRE_NEWLINE_x Line 1103  Returns:      appropriate PCRE_NEWLINE_x
1103  static int  static int
1104  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
1105  {  {
1106  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1107  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1108  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1109  if (strncmp((char *)p, "anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1110  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1111    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1112    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1113  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
1114  return 0;  return 0;
1115  }  }
# Line 698  return 0; Line 1123  return 0;
1123  static void  static void
1124  usage(void)  usage(void)
1125  {  {
1126  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1127    printf("Input and output default to stdin and stdout.\n");
1128    #ifdef SUPPORT_LIBREADLINE
1129    printf("If input is a terminal, readline() is used to read from it.\n");
1130    #else
1131    printf("This version of pcretest is not linked with readline().\n");
1132    #endif
1133    printf("\nOptions:\n");
1134  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
1135  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
1136  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 1139  printf("  -dfa     force DFA matching fo
1139  #endif  #endif
1140  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
1141  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
1142           "  -M       find MATCH_LIMIT minimum for each subject\n"
1143         "  -m       output memory used information\n"         "  -m       output memory used information\n"
1144         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
1145  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 736  int main(int argc, char **argv) Line 1169  int main(int argc, char **argv)
1169  FILE *infile = stdin;  FILE *infile = stdin;
1170  int options = 0;  int options = 0;
1171  int study_options = 0;  int study_options = 0;
1172    int default_find_match_limit = FALSE;
1173  int op = 1;  int op = 1;
1174  int timeit = 0;  int timeit = 0;
1175  int timeitm = 0;  int timeitm = 0;
# Line 795  while (argc > 1 && argv[op][0] == '-') Line 1229  while (argc > 1 && argv[op][0] == '-')
1229    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
1230    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1231    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1232      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1233  #if !defined NODFA  #if !defined NODFA
1234    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1235  #endif  #endif
# Line 847  while (argc > 1 && argv[op][0] == '-') Line 1282  while (argc > 1 && argv[op][0] == '-')
1282    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1283      {      {
1284      int rc;      int rc;
1285        unsigned long int lrc;
1286      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1287      printf("Compiled with\n");      printf("Compiled with\n");
1288      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 854  while (argc > 1 && argv[op][0] == '-') Line 1290  while (argc > 1 && argv[op][0] == '-')
1290      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1291      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1292      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1293      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1294        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
1295        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1296          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1297        (rc == -2)? "ANYCRLF" :        (rc == -2)? "ANYCRLF" :
1298        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
1299        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1300        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1301                                         "all Unicode newlines");
1302      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1303      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1304      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1305      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1306      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1307      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1308      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1309      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1310      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1311      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1312      goto EXIT;      goto EXIT;
# Line 947  while (!done) Line 1388  while (!done)
1388  #endif  #endif
1389    
1390    const char *error;    const char *error;
1391      unsigned char *markptr;
1392    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1393    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1394    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1395    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1396    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1397      int do_mark = 0;
1398    int do_study = 0;    int do_study = 0;
1399    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1400    int do_G = 0;    int do_G = 0;
1401    int do_g = 0;    int do_g = 0;
1402    int do_showinfo = showinfo;    int do_showinfo = showinfo;
# Line 963  while (!done) Line 1405  while (!done)
1405    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1406    
1407    use_utf8 = 0;    use_utf8 = 0;
1408      debug_lengths = 1;
1409    
1410    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1411    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1412    fflush(outfile);    fflush(outfile);
1413    
# Line 1065  while (!done) Line 1507  while (!done)
1507    
1508    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1509      {      {
1510      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1511      goto SKIP_DATA;      goto SKIP_DATA;
1512      }      }
1513    
1514    pp = p;    pp = p;
1515    poffset = p - buffer;    poffset = (int)(p - buffer);
1516    
1517    for(;;)    for(;;)
1518      {      {
# Line 1081  while (!done) Line 1523  while (!done)
1523        pp++;        pp++;
1524        }        }
1525      if (*pp != 0) break;      if (*pp != 0) break;
1526      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1527        {        {
1528        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1529        done = 1;        done = 1;
# Line 1135  while (!done) Line 1576  while (!done)
1576        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1577        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1578        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1579          case 'K': do_mark = 1; break;
1580        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1581        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1582    
# Line 1144  while (!done) Line 1586  while (!done)
1586    
1587        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1588        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1589          case 'W': options |= PCRE_UCP; break;
1590        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1591        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
1592        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1593        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1594    
1595          case 'T':
1596          switch (*pp++)
1597            {
1598            case '0': tables = tables0; break;
1599            case '1': tables = tables1; break;
1600    
1601            case '\r':
1602            case '\n':
1603            case ' ':
1604            case 0:
1605            fprintf(outfile, "** Missing table number after /T\n");
1606            goto SKIP_DATA;
1607    
1608            default:
1609            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1610            goto SKIP_DATA;
1611            }
1612          break;
1613    
1614        case 'L':        case 'L':
1615        ppp = pp;        ppp = pp;
1616        /* 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 1636  while (!done)
1636    
1637        case '<':        case '<':
1638          {          {
1639          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1640          if (x == 0) goto SKIP_DATA;            {
1641          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1642          while (*pp++ != '>');            pp += 3;
1643              }
1644            else
1645              {
1646              int x = check_newline(pp, outfile);
1647              if (x == 0) goto SKIP_DATA;
1648              options |= x;
1649              while (*pp++ != '>');
1650              }
1651          }          }
1652        break;        break;
1653    
# Line 1207  while (!done) Line 1677  while (!done)
1677      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1678      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1679      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1680        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1681        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1682    
1683      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1684    
# Line 1227  while (!done) Line 1699  while (!done)
1699  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1700    
1701      {      {
1702        unsigned long int get_options;
1703    
1704      if (timeit > 0)      if (timeit > 0)
1705        {        {
1706        register int i;        register int i;
# Line 1256  while (!done) Line 1730  while (!done)
1730          {          {
1731          for (;;)          for (;;)
1732            {            {
1733            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1734              {              {
1735              done = 1;              done = 1;
1736              goto CONTINUE;              goto CONTINUE;
# Line 1270  while (!done) Line 1744  while (!done)
1744        goto CONTINUE;        goto CONTINUE;
1745        }        }
1746    
1747      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1748      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
1749      returns only limited data. Check that it agrees with the newer one. */      lines. */
1750    
1751        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1752        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1753    
1754        /* Print information if required. There are now two info-returning
1755        functions. The old one has a limited interface and returns only limited
1756        data. Check that it agrees with the newer one. */
1757    
1758      if (log_store)      if (log_store)
1759        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1311  while (!done) Line 1792  while (!done)
1792          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1793        }        }
1794    
1795        /* If /K was present, we set up for handling MARK data. */
1796    
1797        if (do_mark)
1798          {
1799          if (extra == NULL)
1800            {
1801            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1802            extra->flags = 0;
1803            }
1804          extra->mark = &markptr;
1805          extra->flags |= PCRE_EXTRA_MARK;
1806          }
1807    
1808      /* 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
1809      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
1810      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 1813  while (!done)
1813      if (do_flip)      if (do_flip)
1814        {        {
1815        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1816        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1817            byteflip(rre->magic_number, sizeof(rre->magic_number));
1818        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1819        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1820        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1821        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1822        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1823        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1824        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1825          rre->first_byte =
1826            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1827          rre->req_byte =
1828            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1829          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1830          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1831        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1832          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1833        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1834            sizeof(rre->name_count));
1835    
1836        if (extra != NULL)        if (extra != NULL)
1837          {          {
1838          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1839          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1840          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1841            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1842          }          }
1843        }        }
1844    
# Line 1350  while (!done) Line 1852  while (!done)
1852        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1853        }        }
1854    
1855        /* We already have the options in get_options (see above) */
1856    
1857      if (do_showinfo)      if (do_showinfo)
1858        {        {
1859        unsigned long int get_options, all_options;        unsigned long int all_options;
1860  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1861        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1862  #endif  #endif
1863        int count, backrefmax, first_char, need_char, okpartial, jchanged;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1864            hascrorlf;
1865        int nameentrysize, namecount;        int nameentrysize, namecount;
1866        const uschar *nametable;        const uschar *nametable;
1867    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1868        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1869        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1870        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1371  while (!done) Line 1875  while (!done)
1875        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1876        new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);        new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1877        new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);        new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1878          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1879    
1880  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1881        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 1918  while (!done)
1918          }          }
1919    
1920        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1921          if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1922    
1923        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1924        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
1925    
1926        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1927          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\n",
1928            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1929            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1930            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1931            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1932            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1933            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1934              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1935              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1936            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1937            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1938            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1939            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1940            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1941              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
1942            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1943            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1944    
# Line 1507  while (!done) Line 2016  while (!done)
2016          else          else
2017            {            {
2018            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2019            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2020    
2021              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2022              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2023    
2024              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2025            if (start_bits == NULL)            if (start_bits == NULL)
2026              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2027            else            else
2028              {              {
2029              int i;              int i;
# Line 1557  while (!done) Line 2070  while (!done)
2070        else        else
2071          {          {
2072          uschar sbuf[8];          uschar sbuf[8];
2073          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2074          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2075          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2076          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2077    
2078          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2079          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2080          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2081          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2082    
2083          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2084              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1592  while (!done) Line 2105  while (!done)
2105    
2106        new_free(re);        new_free(re);
2107        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2108        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2109            {
2110            new_free((void *)tables);
2111            setlocale(LC_CTYPE, "C");
2112            locale_set = 0;
2113            }
2114        continue;  /* With next regex */        continue;  /* With next regex */
2115        }        }
2116      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1609  while (!done) Line 2127  while (!done)
2127      int callout_data_set = 0;      int callout_data_set = 0;
2128      int count, c;      int count, c;
2129      int copystrings = 0;      int copystrings = 0;
2130      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2131      int getstrings = 0;      int getstrings = 0;
2132      int getlist = 0;      int getlist = 0;
2133      int gmatched = 0;      int gmatched = 0;
# Line 1639  while (!done) Line 2157  while (!done)
2157      len = 0;      len = 0;
2158      for (;;)      for (;;)
2159        {        {
2160        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2161          {          {
2162          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2163              {
2164              fprintf(outfile, "\n");
2165              break;
2166              }
2167          done = 1;          done = 1;
2168          goto CONTINUE;          goto CONTINUE;
2169          }          }
# Line 1708  while (!done) Line 2229  while (!done)
2229              {              {
2230              unsigned char buff8[8];              unsigned char buff8[8];
2231              int ii, utn;              int ii, utn;
2232              utn = ord2utf8(c, buff8);              if (use_utf8)
2233              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2234              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2235                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2236                  c = buff8[ii];   /* Last byte */
2237                  }
2238                else
2239                 {
2240                 if (c > 255)
2241                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2242                     "UTF-8 mode is not enabled.\n"
2243                     "** Truncation will probably give the wrong result.\n", c);
2244                 }
2245              p = pt + 1;              p = pt + 1;
2246              break;              break;
2247              }              }
# Line 1806  while (!done) Line 2337  while (!done)
2337  #endif  #endif
2338            use_dfa = 1;            use_dfa = 1;
2339          continue;          continue;
2340    #endif
2341    
2342    #if !defined NODFA
2343          case 'F':          case 'F':
2344          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2345          continue;          continue;
# Line 1840  while (!done) Line 2373  while (!done)
2373          continue;          continue;
2374    
2375          case 'N':          case 'N':
2376          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2377              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2378            else
2379              options |= PCRE_NOTEMPTY;
2380          continue;          continue;
2381    
2382          case 'O':          case 'O':
# Line 1863  while (!done) Line 2399  while (!done)
2399          continue;          continue;
2400    
2401          case 'P':          case 'P':
2402          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2403              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2404          continue;          continue;
2405    
2406          case 'Q':          case 'Q':
# Line 1898  while (!done) Line 2435  while (!done)
2435          show_malloc = 1;          show_malloc = 1;
2436          continue;          continue;
2437    
2438            case 'Y':
2439            options |= PCRE_NO_START_OPTIMIZE;
2440            continue;
2441    
2442          case 'Z':          case 'Z':
2443          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2444          continue;          continue;
# Line 1918  while (!done) Line 2459  while (!done)
2459        *q++ = c;        *q++ = c;
2460        }        }
2461      *q = 0;      *q = 0;
2462      len = q - dbuffer;      len = (int)(q - dbuffer);
2463    
2464        /* Move the data to the end of the buffer so that a read over the end of
2465        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2466        we are using the POSIX interface, we must include the terminating zero. */
2467    
2468    #if !defined NOPOSIX
2469        if (posix || do_posix)
2470          {
2471          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2472          bptr += buffer_size - len - 1;
2473          }
2474        else
2475    #endif
2476          {
2477          memmove(bptr + buffer_size - len, bptr, len);
2478          bptr += buffer_size - len;
2479          }
2480    
2481      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2482        {        {
# Line 1939  while (!done) Line 2497  while (!done)
2497          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2498        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2499        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2500          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2501    
2502        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2503    
# Line 1983  while (!done) Line 2542  while (!done)
2542    
2543      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2544        {        {
2545          markptr = NULL;
2546    
2547        if (timeitm > 0)        if (timeitm > 0)
2548          {          {
2549          register int i;          register int i;
# Line 1994  while (!done) Line 2555  while (!done)
2555            {            {
2556            int workspace[1000];            int workspace[1000];
2557            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
2558              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2559                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2560                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2561            }            }
# Line 2057  while (!done) Line 2618  while (!done)
2618        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2619          {          {
2620          int workspace[1000];          int workspace[1000];
2621          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2622            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2623            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2624          if (count == 0)          if (count == 0)
# Line 2128  while (!done) Line 2689  while (!done)
2689              }              }
2690            }            }
2691    
2692            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2693    
2694          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2695            {            {
2696            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2211  while (!done) Line 2774  while (!done)
2774    
2775        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2776          {          {
2777          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2778  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2779          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2780            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2781              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2782  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2783                outfile);
2784              }
2785          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2786          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2787          }          }
# Line 2226  while (!done) Line 2791  while (!done)
2791        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
2792        string - that was checked before setting g_notempty.        string - that was checked before setting g_notempty.
2793    
2794        Complication arises in the case when the newline option is "any" or        Complication arises in the case when the newline convention is "any",
2795        "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
2796        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,
2797        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
2798        Fudge the offset value to achieve this.        pcre_exec(). Fudge the offset value to achieve this. We check for a
2799          newline setting in the pattern; if none was set, use pcre_config() to
2800          find the default.
2801    
2802        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
2803        character, not one byte. */        character, not one byte. */
# Line 2246  while (!done) Line 2813  while (!done)
2813              {              {
2814              int d;              int d;
2815              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);              (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2816              obits = (d == '\r')? PCRE_NEWLINE_CR :              /* Note that these values are always the ASCII ones, even in
2817                      (d == '\n')? PCRE_NEWLINE_LF :              EBCDIC environments. CR = 13, NL = 10. */
2818                      (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :              obits = (d == 13)? PCRE_NEWLINE_CR :
2819                        (d == 10)? PCRE_NEWLINE_LF :
2820                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2821                      (d == -2)? PCRE_NEWLINE_ANYCRLF :                      (d == -2)? PCRE_NEWLINE_ANYCRLF :
2822                      (d == -1)? PCRE_NEWLINE_ANY : 0;                      (d == -1)? PCRE_NEWLINE_ANY : 0;
2823              }              }
2824            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||            if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2825                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2826                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)                 (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2827                &&                &&
2828                start_offset < len - 1 &&                start_offset < len - 1 &&
# Line 2263  while (!done) Line 2833  while (!done)
2833              {              {
2834              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2835                {                {
2836                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2837                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2838                }                }
2839              }              }
2840            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
# Line 2275  while (!done) Line 2843  while (!done)
2843            {            {
2844            if (count == PCRE_ERROR_NOMATCH)            if (count == PCRE_ERROR_NOMATCH)
2845              {              {
2846              if (gmatched == 0) fprintf(outfile, "No match\n");              if (gmatched == 0)
2847                  {
2848                  if (markptr == NULL) fprintf(outfile, "No match\n");
2849                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2850                  }
2851              }              }
2852            else fprintf(outfile, "Error %d\n", count);            else fprintf(outfile, "Error %d\n", count);
2853            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
# Line 2287  while (!done) Line 2859  while (!done)
2859        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2860    
2861        /* 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
2862        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
2863        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
2864        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
2865        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
2866        character. */        character. */
2867    
# Line 2298  while (!done) Line 2870  while (!done)
2870        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2871          {          {
2872          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2873          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2874          }          }
2875    
2876        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2325  while (!done) Line 2897  while (!done)
2897    
2898    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
2899    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
2900    if (tables != NULL)    if (locale_set)
2901      {      {
2902      new_free((void *)tables);      new_free((void *)tables);
2903      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
Legend:
Removed from v.200  
changed lines
  Added in v.566
  ViewVC Help
Powered by ViewVC 1.1.5