/[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 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 579 by ph10, Wed Nov 24 17:39:25 2010 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  #define PCRE_SPY        /* For Win32 build, import data, not export */  #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
61    input and output without "b"; then I was told that "b" was needed in some
62    environments, so it was added for release 5.0 to both the input and output. (It
63    makes no difference on Unix-like systems.) Later I was told that it is wrong
64    for the input on Windows. I've now abstracted the modes into two macros that
65    are set here, to make it easier to fiddle with them, and removed "b" from the
66    input mode under Windows. */
67    
68    #if defined(_WIN32) || defined(WIN32)
69    #include <io.h>                /* For _setmode() */
70    #include <fcntl.h>             /* For _O_BINARY */
71    #define INPUT_MODE   "r"
72    #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
83    #include <sys/time.h>          /* These two includes are needed */
84    #include <sys/resource.h>      /* for setrlimit(). */
85    #define INPUT_MODE   "rb"
86    #define OUTPUT_MODE  "wb"
87    #endif
88    
89    
90    /* We have to include pcre_internal.h because we need the internal info for
91    displaying the results of pcre_study() and we also need to know about the
92    internal macros, structures, and other internal data values; pcretest has
93    "inside information" compared to a program that strictly follows the PCRE API.
94    
95  /* We include pcre_internal.h because we need the internal info for displaying  Although pcre_internal.h does itself include pcre.h, we explicitly include it
96  the results of pcre_study() and we also need to know about the internal  here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
97  macros, structures, and other internal data values; pcretest has "inside  appropriately for an application, not for building PCRE. */
 information" compared to a program that strictly follows the PCRE API. */  
98    
99    #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 64  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. */
123    
124    #define COMPILING_PCRETEST
125  #include "pcre_printint.src"  #include "pcre_printint.src"
126    
127    /* 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 83  Makefile. */ Line 140  Makefile. */
140  #include "pcreposix.h"  #include "pcreposix.h"
141  #endif  #endif
142    
143  /* It is also possible, for the benefit of the version imported into Exim, to  /* It is also possible, for the benefit of the version currently imported into
144  build pcretest without support for UTF8 (define NOUTF8), without the interface  Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
145  to the DFA matcher (NODFA), and without the doublecheck of the old "info"  interface to the DFA matcher (NODFA), and without the doublecheck of the old
146  function (define NOINFOCHECK). */  "info" function (define NOINFOCHECK). In fact, we automatically cut out the
147    UTF8 support if PCRE is built without it. */
148    
149    #ifndef SUPPORT_UTF8
150    #ifndef NOUTF8
151    #define NOUTF8
152    #endif
153    #endif
154    
155    
156  /* Other parameters */  /* Other parameters */
# Line 99  function (define NOINFOCHECK). */ Line 163  function (define NOINFOCHECK). */
163  #endif  #endif
164  #endif  #endif
165    
166  #define LOOPREPEAT 500000  /* This is the default loop count for timing. */
   
 #define BUFFER_SIZE 30000  
 #define PBUFFER_SIZE BUFFER_SIZE  
 #define DBUFFER_SIZE BUFFER_SIZE  
167    
168    #define LOOPREPEAT 500000
169    
170  /* Static variables */  /* Static variables */
171    
# Line 114  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;
181  static int show_malloc;  static int show_malloc;
182  static int use_utf8;  static int use_utf8;
183  static size_t gotten_store;  static size_t gotten_store;
184    
185    /* The buffers grow automatically if very long input lines are encountered. */
186    
187    static int buffer_size = 50000;
188    static uschar *buffer = NULL;
189    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            *
545    *************************************************/
546    
547    /* Input lines are read into buffer, but both patterns and data lines can be
548    continued over multiple input lines. In addition, if the buffer fills up, we
549    want to automatically expand it so as to be able to handle extremely large
550    lines that are needed for certain stress tests. When the input buffer is
551    expanded, the other two buffers must also be expanded likewise, and the
552    contents of pbuffer, which are a copy of the input for callouts, must be
553    preserved (for when expansion happens for a data line). This is not the most
554    optimal way of handling this, but hey, this is just a test program!
555    
556    Arguments:
557      f            the file to read
558      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
562                   could be a copy of start, or could be moved
563                   NULL if no data read and EOF reached
564    */
565    
566    static uschar *
567    extend_inputline(FILE *f, uschar *start, const char *prompt)
568    {
569    uschar *here = start;
570    
571    for (;;)
572      {
573      int rlen = (int)(buffer_size - (here - buffer));
574    
575      if (rlen > 1000)
576        {
577        int dlen;
578    
579        /* 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);
609        if (dlen > 0 && here[dlen - 1] == '\n') return start;
610        here += dlen;
611        }
612    
613      else
614        {
615        int new_buffer_size = 2*buffer_size;
616        uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);
617        uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
618        uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);
619    
620        if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
621          {
622          fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
623          exit(1);
624          }
625    
626        memcpy(new_buffer, buffer, buffer_size);
627        memcpy(new_pbuffer, pbuffer, buffer_size);
628    
629        buffer_size = new_buffer_size;
630    
631        start = new_buffer + (start - buffer);
632        here = new_buffer + (here - buffer);
633    
634        free(buffer);
635        free(dbuffer);
636        free(pbuffer);
637    
638        buffer = new_buffer;
639        dbuffer = new_dbuffer;
640        pbuffer = new_pbuffer;
641        }
642      }
643    
644    return NULL;  /* Control never gets here */
645    }
646    
647    
648    
649    
650    
651    
652    
653  /*************************************************  /*************************************************
654  *          Read number from string               *  *          Read number from string               *
# Line 129  static uschar *pbuffer = NULL; Line 656  static uschar *pbuffer = NULL;
656    
657  /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess  /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess
658  around with conditional compilation, just do the job by hand. It is only used  around with conditional compilation, just do the job by hand. It is only used
659  for unpicking the -o argument, so just keep it simple.  for unpicking arguments, so just keep it simple.
660    
661  Arguments:  Arguments:
662    str           string to be converted    str           string to be converted
# Line 159  return(result); Line 686  return(result);
686  and returns the value of the character.  and returns the value of the character.
687    
688  Argument:  Argument:
689    buffer   a pointer to the byte vector    utf8bytes   a pointer to the byte vector
690    vptr     a pointer to an int to receive the value    vptr        a pointer to an int to receive the value
691    
692  Returns:   >  0 => the number of bytes consumed  Returns:      >  0 => the number of bytes consumed
693             -6 to 0 => malformed UTF-8 character at offset = (-return)                -6 to 0 => malformed UTF-8 character at offset = (-return)
694  */  */
695    
696  #if !defined NOUTF8  #if !defined NOUTF8
697    
698  static int  static int
699  utf82ord(unsigned char *buffer, int *vptr)  utf82ord(unsigned char *utf8bytes, int *vptr)
700  {  {
701  int c = *buffer++;  int c = *utf8bytes++;
702  int d = c;  int d = c;
703  int i, j, s;  int i, j, s;
704    
# Line 191  d = (c & utf8_table3[i]) << s; Line 718  d = (c & utf8_table3[i]) << s;
718    
719  for (j = 0; j < i; j++)  for (j = 0; j < i; j++)
720    {    {
721    c = *buffer++;    c = *utf8bytes++;
722    if ((c & 0xc0) != 0x80) return -(j+1);    if ((c & 0xc0) != 0x80) return -(j+1);
723    s -= 6;    s -= 6;
724    d |= (c & 0x3f) << s;    d |= (c & 0x3f) << s;
# Line 222  and encodes it as a UTF-8 character in 0 Line 749  and encodes it as a UTF-8 character in 0
749    
750  Arguments:  Arguments:
751    cvalue     the character value    cvalue     the character value
752    buffer     pointer to buffer for result - at least 6 bytes long    utf8bytes  pointer to buffer for result - at least 6 bytes long
753    
754  Returns:     number of characters placed in the buffer  Returns:     number of characters placed in the buffer
755  */  */
756    
757    #if !defined NOUTF8
758    
759  static int  static int
760  ord2utf8(int cvalue, uschar *buffer)  ord2utf8(int cvalue, uschar *utf8bytes)
761  {  {
762  register int i, j;  register int i, j;
763  for (i = 0; i < utf8_table1_size; i++)  for (i = 0; i < utf8_table1_size; i++)
764    if (cvalue <= utf8_table1[i]) break;    if (cvalue <= utf8_table1[i]) break;
765  buffer += i;  utf8bytes += i;
766  for (j = i; j > 0; j--)  for (j = i; j > 0; j--)
767   {   {
768   *buffer-- = 0x80 | (cvalue & 0x3f);   *utf8bytes-- = 0x80 | (cvalue & 0x3f);
769   cvalue >>= 6;   cvalue >>= 6;
770   }   }
771  *buffer = utf8_table2[i] | cvalue;  *utf8bytes = utf8_table2[i] | cvalue;
772  return i + 1;  return i + 1;
773  }  }
774    
775    #endif
776    
777    
778    
779  /*************************************************  /*************************************************
# Line 269  while (length-- > 0) Line 800  while (length-- > 0)
800        {        {
801        length -= rc - 1;        length -= rc - 1;
802        p += rc;        p += rc;
803        if (c < 256 && isprint(c))        if (PRINTHEX(c))
804          {          {
805          if (f != NULL) fprintf(f, "%c", c);          if (f != NULL) fprintf(f, "%c", c);
806          yield++;          yield++;
807          }          }
808        else        else
809          {          {
810          int n;          int n = 4;
811          if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);          if (f != NULL) fprintf(f, "\\x{%02x}", c);
812          yield += n;          yield += (n <= 0x000000ff)? 2 :
813                     (n <= 0x00000fff)? 3 :
814                     (n <= 0x0000ffff)? 4 :
815                     (n <= 0x000fffff)? 5 : 6;
816          }          }
817        continue;        continue;
818        }        }
# Line 287  while (length-- > 0) Line 821  while (length-- > 0)
821    
822     /* Not UTF-8, or malformed UTF-8  */     /* Not UTF-8, or malformed UTF-8  */
823    
824    if (isprint(c = *(p++)))    c = *p++;
825      if (PRINTHEX(c))
826      {      {
827      if (f != NULL) fprintf(f, "%c", c);      if (f != NULL) fprintf(f, "%c", c);
828      yield++;      yield++;
# Line 461  if ((rc = pcre_fullinfo(re, study, optio Line 996  if ((rc = pcre_fullinfo(re, study, optio
996  *         Byte flipping function                 *  *         Byte flipping function                 *
997  *************************************************/  *************************************************/
998    
999  static long int  static unsigned long int
1000  byteflip(long int value, int n)  byteflip(unsigned long int value, int n)
1001  {  {
1002  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
1003  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
# Line 526  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                *
1091    *************************************************/
1092    
1093    /* This is used both at compile and run-time to check for <xxx> escapes. Print
1094    a message and return 0 if there is no match.
1095    
1096    Arguments:
1097      p           points after the leading '<'
1098      f           file for error message
1099    
1100    Returns:      appropriate PCRE_NEWLINE_xxx flags, or 0
1101    */
1102    
1103    static int
1104    check_newline(uschar *p, FILE *f)
1105    {
1106    if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1107    if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1108    if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1109    if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1110    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);
1114    return 0;
1115    }
1116    
1117    
1118    
1119    /*************************************************
1120    *             Usage function                     *
1121    *************************************************/
1122    
1123    static void
1124    usage(void)
1125    {
1126    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");
1135    printf("  -C       show PCRE compile-time options and exit\n");
1136    printf("  -d       debug: show compiled code and information (-b and -i)\n");
1137    #if !defined NODFA
1138    printf("  -dfa     force DFA matching for all subjects\n");
1139    #endif
1140    printf("  -help    show usage information\n");
1141    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"
1144           "  -o <n>   set size of offsets vector to <n>\n");
1145    #if !defined NOPOSIX
1146    printf("  -p       use POSIX interface\n");
1147    #endif
1148    printf("  -q       quiet: do not output PCRE version number at start\n");
1149    printf("  -S <n>   set stack size to <n> megabytes\n");
1150    printf("  -s       output store (memory) used information\n"
1151           "  -t       time compilation and execution\n");
1152    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1153    printf("  -tm      time execution (matching) only\n");
1154    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
1155    }
1156    
1157    
1158    
1159    /*************************************************
1160  *                Main Program                    *  *                Main Program                    *
1161  *************************************************/  *************************************************/
1162    
# Line 538  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;
1176  int showinfo = 0;  int showinfo = 0;
1177  int showstore = 0;  int showstore = 0;
1178  int quiet = 0;  int quiet = 0;
# Line 553  int debug = 0; Line 1186  int debug = 0;
1186  int done = 0;  int done = 0;
1187  int all_use_dfa = 0;  int all_use_dfa = 0;
1188  int yield = 0;  int yield = 0;
1189    int stack_size;
1190    
1191    /* These vectors store, end-to-end, a list of captured substring names. Assume
1192    that 1024 is plenty long enough for the few names we'll be testing. */
1193    
1194  unsigned char *buffer;  uschar copynames[1024];
1195  unsigned char *dbuffer;  uschar getnames[1024];
1196    
1197    uschar *copynamesptr;
1198    uschar *getnamesptr;
1199    
1200  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* Get buffers from malloc() so that Electric Fence will check their misuse
1201  when I am debugging. */  when I am debugging. They grow automatically when very long lines are read. */
1202    
1203  buffer = (unsigned char *)malloc(BUFFER_SIZE);  buffer = (unsigned char *)malloc(buffer_size);
1204  dbuffer = (unsigned char *)malloc(DBUFFER_SIZE);  dbuffer = (unsigned char *)malloc(buffer_size);
1205  pbuffer = (unsigned char *)malloc(PBUFFER_SIZE);  pbuffer = (unsigned char *)malloc(buffer_size);
   
 /* The outfile variable is static so that new_malloc can use it. The _setmode()  
 stuff is some magic that I don't understand, but which apparently does good  
 things in Windows. It's related to line terminations.  */  
1206    
1207  #if defined(_WIN32) || defined(WIN32)  /* The outfile variable is static so that new_malloc can use it. */
 _setmode( _fileno( stdout ), 0x8000 );  
 #endif  /* defined(_WIN32) || defined(WIN32) */  
1208    
1209  outfile = stdout;  outfile = stdout;
1210    
1211    /* The following  _setmode() stuff is some Windows magic that tells its runtime
1212    library to translate CRLF into a single LF character. At least, that's what
1213    I've been told: never having used Windows I take this all on trust. Originally
1214    it set 0x8000, but then I was advised that _O_BINARY was better. */
1215    
1216    #if defined(_WIN32) || defined(WIN32)
1217    _setmode( _fileno( stdout ), _O_BINARY );
1218    #endif
1219    
1220  /* Scan options */  /* Scan options */
1221    
1222  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
# Line 582  while (argc > 1 && argv[op][0] == '-') Line 1225  while (argc > 1 && argv[op][0] == '-')
1225    
1226    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
1227      showstore = 1;      showstore = 1;
   else if (strcmp(argv[op], "-t") == 0) timeit = 1;  
1228    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1229      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 596  while (argc > 1 && argv[op][0] == '-') Line 1240  while (argc > 1 && argv[op][0] == '-')
1240      op++;      op++;
1241      argc--;      argc--;
1242      }      }
1243      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
1244        {
1245        int both = argv[op][2] == 0;
1246        int temp;
1247        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
1248                         *endptr == 0))
1249          {
1250          timeitm = temp;
1251          op++;
1252          argc--;
1253          }
1254        else timeitm = LOOPREPEAT;
1255        if (both) timeit = timeitm;
1256        }
1257      else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
1258          ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1259            *endptr == 0))
1260        {
1261    #if defined(_WIN32) || defined(WIN32)
1262        printf("PCRE: -S not supported on this OS\n");
1263        exit(1);
1264    #else
1265        int rc;
1266        struct rlimit rlim;
1267        getrlimit(RLIMIT_STACK, &rlim);
1268        rlim.rlim_cur = stack_size * 1024 * 1024;
1269        rc = setrlimit(RLIMIT_STACK, &rlim);
1270        if (rc != 0)
1271          {
1272        printf("PCRE: setrlimit() failed with error %d\n", rc);
1273        exit(1);
1274          }
1275        op++;
1276        argc--;
1277    #endif
1278        }
1279  #if !defined NOPOSIX  #if !defined NOPOSIX
1280    else if (strcmp(argv[op], "-p") == 0) posix = 1;    else if (strcmp(argv[op], "-p") == 0) posix = 1;
1281  #endif  #endif
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 609  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 character is %s\n", (rc == '\r')? "CR" : "LF");      /* Note that these values are always the ASCII values, even
1294        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" :
1298          (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      exit(0);      goto EXIT;
1313        }
1314      else if (strcmp(argv[op], "-help") == 0 ||
1315               strcmp(argv[op], "--help") == 0)
1316        {
1317        usage();
1318        goto EXIT;
1319      }      }
1320    else    else
1321      {      {
1322      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
1323      printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");      usage();
     printf("  -C     show PCRE compile-time options and exit\n");  
     printf("  -d     debug: show compiled code; implies -i\n");  
 #if !defined NODFA  
     printf("  -dfa   force DFA matching for all subjects\n");  
 #endif  
     printf("  -i     show information about compiled pattern\n"  
            "  -m     output memory used information\n"  
            "  -o <n> set size of offsets vector to <n>\n");  
 #if !defined NOPOSIX  
     printf("  -p     use POSIX interface\n");  
 #endif  
     printf("  -s     output store (memory) used information\n"  
            "  -t     time compilation and execution\n");  
1324      yield = 1;      yield = 1;
1325      goto EXIT;      goto EXIT;
1326      }      }
# Line 653  offsets = (int *)malloc(size_offsets_max Line 1335  offsets = (int *)malloc(size_offsets_max
1335  if (offsets == NULL)  if (offsets == NULL)
1336    {    {
1337    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1338      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1339    yield = 1;    yield = 1;
1340    goto EXIT;    goto EXIT;
1341    }    }
# Line 662  if (offsets == NULL) Line 1344  if (offsets == NULL)
1344    
1345  if (argc > 1)  if (argc > 1)
1346    {    {
1347    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
1348    if (infile == NULL)    if (infile == NULL)
1349      {      {
1350      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
# Line 673  if (argc > 1) Line 1355  if (argc > 1)
1355    
1356  if (argc > 2)  if (argc > 2)
1357    {    {
1358    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
1359    if (outfile == NULL)    if (outfile == NULL)
1360      {      {
1361      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
# Line 706  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;
1400    int do_G = 0;    int do_G = 0;
# Line 718  while (!done) Line 1402  while (!done)
1402    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1403    int do_showrest = 0;    int do_showrest = 0;
1404    int do_flip = 0;    int do_flip = 0;
1405    int erroroffset, len, delimiter;    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 (fgets((char *)buffer, BUFFER_SIZE, infile) == 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 735  while (!done) Line 1419  while (!done)
1419    
1420    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
1421      {      {
1422      unsigned long int magic;      unsigned long int magic, get_options;
1423      uschar sbuf[8];      uschar sbuf[8];
1424      FILE *f;      FILE *f;
1425    
# Line 783  while (!done) Line 1467  while (!done)
1467    
1468      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
1469    
1470      new_info(re, NULL, PCRE_INFO_OPTIONS, &options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1471      use_utf8 = (options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1472    
1473      /* Now see if there is any following study data */      /* Now see if there is any following study data */
1474    
# Line 823  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 = (int)(p - buffer);
1516    
1517    for(;;)    for(;;)
1518      {      {
# Line 838  while (!done) Line 1523  while (!done)
1523        pp++;        pp++;
1524        }        }
1525      if (*pp != 0) break;      if (*pp != 0) break;
1526        if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     len = BUFFER_SIZE - (pp - buffer);  
     if (len < 256)  
       {  
       fprintf(outfile, "** Expression too long - missing delimiter?\n");  
       goto SKIP_DATA;  
       }  
   
     if (infile == stdin) printf("    > ");  
     if (fgets((char *)pp, len, infile) == NULL)  
1527        {        {
1528        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1529        done = 1;        done = 1;
# Line 856  while (!done) Line 1532  while (!done)
1532      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1533      }      }
1534    
1535      /* The buffer may have moved while being extended; reset the start of data
1536      pointer to the correct relative point in the buffer. */
1537    
1538      p = buffer + poffset;
1539    
1540    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1541    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1542    of testing for the error message when a pattern ends with backslash. */    of testing for the error message when a pattern ends with backslash. */
# Line 887  while (!done) Line 1568  while (!done)
1568    
1569        case '+': do_showrest = 1; break;        case '+': do_showrest = 1; break;
1570        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1571          case 'B': do_debug = 1; break;
1572        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1573        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1574        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
1575        case 'F': do_flip = 1; break;        case 'F': do_flip = 1; break;
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;
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 902  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 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1592          case 'Z': debug_lengths = 0; break;
1593        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1594        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1595    
1596          case 'T':
1597          switch (*pp++)
1598            {
1599            case '0': tables = tables0; break;
1600            case '1': tables = tables1; break;
1601    
1602            case '\r':
1603            case '\n':
1604            case ' ':
1605            case 0:
1606            fprintf(outfile, "** Missing table number after /T\n");
1607            goto SKIP_DATA;
1608    
1609            default:
1610            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1611            goto SKIP_DATA;
1612            }
1613          break;
1614    
1615        case 'L':        case 'L':
1616        ppp = pp;        ppp = pp;
1617        /* The '\r' test here is so that it works on Windows */        /* The '\r' test here is so that it works on Windows. */
1618        while (*ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;        /* The '0' test is just in case this is an unterminated line. */
1619          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1620        *ppp = 0;        *ppp = 0;
1621        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1622          {          {
1623          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1624          goto SKIP_DATA;          goto SKIP_DATA;
1625          }          }
1626          locale_set = 1;
1627        tables = pcre_maketables();        tables = pcre_maketables();
1628        pp = ppp;        pp = ppp;
1629        break;        break;
# Line 927  while (!done) Line 1635  while (!done)
1635        *pp = 0;        *pp = 0;
1636        break;        break;
1637    
1638          case '<':
1639            {
1640            if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1641              {
1642              options |= PCRE_JAVASCRIPT_COMPAT;
1643              pp += 3;
1644              }
1645            else
1646              {
1647              int x = check_newline(pp, outfile);
1648              if (x == 0) goto SKIP_DATA;
1649              options |= x;
1650              while (*pp++ != '>');
1651              }
1652            }
1653          break;
1654    
1655        case '\r':                      /* So that it works in Windows */        case '\r':                      /* So that it works in Windows */
1656        case '\n':        case '\n':
1657        case ' ':        case ' ':
# Line 953  while (!done) Line 1678  while (!done)
1678      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1679      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1680      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1681        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1682        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1683    
1684      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1685    
# Line 961  while (!done) Line 1688  while (!done)
1688    
1689      if (rc != 0)      if (rc != 0)
1690        {        {
1691        (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);        (void)regerror(rc, &preg, (char *)buffer, buffer_size);
1692        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
1693        goto SKIP_DATA;        goto SKIP_DATA;
1694        }        }
# Line 973  while (!done) Line 1700  while (!done)
1700  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1701    
1702      {      {
1703      if (timeit)      unsigned long int get_options;
1704    
1705        if (timeit > 0)
1706        {        {
1707        register int i;        register int i;
1708        clock_t time_taken;        clock_t time_taken;
1709        clock_t start_time = clock();        clock_t start_time = clock();
1710        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1711          {          {
1712          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1713          if (re != NULL) free(re);          if (re != NULL) free(re);
1714          }          }
1715        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1716        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1717          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1718            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1719        }        }
1720    
# Line 1002  while (!done) Line 1731  while (!done)
1731          {          {
1732          for (;;)          for (;;)
1733            {            {
1734            if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1735              {              {
1736              done = 1;              done = 1;
1737              goto CONTINUE;              goto CONTINUE;
# Line 1016  while (!done) Line 1745  while (!done)
1745        goto CONTINUE;        goto CONTINUE;
1746        }        }
1747    
1748      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1749      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
1750      returns only limited data. Check that it agrees with the newer one. */      lines. */
1751    
1752        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1753        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1754    
1755        /* Print information if required. There are now two info-returning
1756        functions. The old one has a limited interface and returns only limited
1757        data. Check that it agrees with the newer one. */
1758    
1759      if (log_store)      if (log_store)
1760        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1037  while (!done) Line 1773  while (!done)
1773    
1774      if (do_study)      if (do_study)
1775        {        {
1776        if (timeit)        if (timeit > 0)
1777          {          {
1778          register int i;          register int i;
1779          clock_t time_taken;          clock_t time_taken;
1780          clock_t start_time = clock();          clock_t start_time = clock();
1781          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1782            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options, &error);
1783          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1784          if (extra != NULL) free(extra);          if (extra != NULL) free(extra);
1785          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1786            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1787              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1788          }          }
1789        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options, &error);
# Line 1057  while (!done) Line 1793  while (!done)
1793          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1794        }        }
1795    
1796        /* If /K was present, we set up for handling MARK data. */
1797    
1798        if (do_mark)
1799          {
1800          if (extra == NULL)
1801            {
1802            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1803            extra->flags = 0;
1804            }
1805          extra->mark = &markptr;
1806          extra->flags |= PCRE_EXTRA_MARK;
1807          }
1808    
1809      /* 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
1810      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
1811      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 1065  while (!done) Line 1814  while (!done)
1814      if (do_flip)      if (do_flip)
1815        {        {
1816        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1817        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1818            byteflip(rre->magic_number, sizeof(rre->magic_number));
1819        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1820        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1821        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1822        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1823        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1824        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1825        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1826          rre->first_byte =
1827            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1828          rre->req_byte =
1829            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1830          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1831          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1832        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1833          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1834        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1835            sizeof(rre->name_count));
1836    
1837        if (extra != NULL)        if (extra != NULL)
1838          {          {
1839          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1840          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1841          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1842            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1843          }          }
1844        }        }
1845    
# Line 1090  while (!done) Line 1847  while (!done)
1847    
1848      SHOW_INFO:      SHOW_INFO:
1849    
1850        if (do_debug)
1851          {
1852          fprintf(outfile, "------------------------------------------------------------------\n");
1853          pcre_printint(re, outfile, debug_lengths);
1854          }
1855    
1856        /* We already have the options in get_options (see above) */
1857    
1858      if (do_showinfo)      if (do_showinfo)
1859        {        {
1860        unsigned long int get_options, all_options;        unsigned long int all_options;
1861  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1862        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1863  #endif  #endif
1864        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1865            hascrorlf;
1866        int nameentrysize, namecount;        int nameentrysize, namecount;
1867        const uschar *nametable;        const uschar *nametable;
1868    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         pcre_printint(re, outfile);  
         }  
   
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1869        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1870        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1871        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1115  while (!done) Line 1874  while (!done)
1874        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1875        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1876        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1877          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1878          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1879          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1880    
1881  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1882        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1156  while (!done) Line 1918  while (!done)
1918            }            }
1919          }          }
1920    
1921        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1922        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1923    
1924        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1925        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");  
1926    
1927        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1928          else fprintf(outfile, "Options:%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",
1929            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1930            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1931            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1932            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1933            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1934            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1935              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1936              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1937            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1938            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1939            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1940            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1941            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1942            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");            ((get_options & PCRE_UCP) != 0)? " ucp" : "",
1943              ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1944              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
1945              ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1946    
1947          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1948    
1949          switch (get_options & PCRE_NEWLINE_BITS)
1950            {
1951            case PCRE_NEWLINE_CR:
1952            fprintf(outfile, "Forced newline sequence: CR\n");
1953            break;
1954    
1955            case PCRE_NEWLINE_LF:
1956            fprintf(outfile, "Forced newline sequence: LF\n");
1957            break;
1958    
1959            case PCRE_NEWLINE_CRLF:
1960            fprintf(outfile, "Forced newline sequence: CRLF\n");
1961            break;
1962    
1963        if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)          case PCRE_NEWLINE_ANYCRLF:
1964          fprintf(outfile, "Case state changes\n");          fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1965            break;
1966    
1967            case PCRE_NEWLINE_ANY:
1968            fprintf(outfile, "Forced newline sequence: ANY\n");
1969            break;
1970    
1971            default:
1972            break;
1973            }
1974    
1975        if (first_char == -1)        if (first_char == -1)
1976          {          {
1977          fprintf(outfile, "First char at start or follows \\n\n");          fprintf(outfile, "First char at start or follows newline\n");
1978          }          }
1979        else if (first_char < 0)        else if (first_char < 0)
1980          {          {
# Line 1199  while (!done) Line 1985  while (!done)
1985          int ch = first_char & 255;          int ch = first_char & 255;
1986          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
1987            "" : " (caseless)";            "" : " (caseless)";
1988          if (isprint(ch))          if (PRINTHEX(ch))
1989            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
1990          else          else
1991            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1214  while (!done) Line 2000  while (!done)
2000          int ch = need_char & 255;          int ch = need_char & 255;
2001          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
2002            "" : " (caseless)";            "" : " (caseless)";
2003          if (isprint(ch))          if (PRINTHEX(ch))
2004            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
2005          else          else
2006            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1232  while (!done) Line 2018  while (!done)
2018          else          else
2019            {            {
2020            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2021            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2022    
2023              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2024              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2025    
2026              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2027            if (start_bits == NULL)            if (start_bits == NULL)
2028              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2029            else            else
2030              {              {
2031              int i;              int i;
# Line 1250  while (!done) Line 2040  while (!done)
2040                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
2041                    c = 2;                    c = 2;
2042                    }                    }
2043                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
2044                    {                    {
2045                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
2046                    c += 2;                    c += 2;
# Line 1282  while (!done) Line 2072  while (!done)
2072        else        else
2073          {          {
2074          uschar sbuf[8];          uschar sbuf[8];
2075          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2076          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2077          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2078          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2079    
2080          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2081          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2082          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2083          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2084    
2085          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2086              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1309  while (!done) Line 2099  while (!done)
2099                  strerror(errno));                  strerror(errno));
2100                }                }
2101              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
2102    
2103              }              }
2104            }            }
2105          fclose(f);          fclose(f);
# Line 1316  while (!done) Line 2107  while (!done)
2107    
2108        new_free(re);        new_free(re);
2109        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2110        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2111            {
2112            new_free((void *)tables);
2113            setlocale(LC_CTYPE, "C");
2114            locale_set = 0;
2115            }
2116        continue;  /* With next regex */        continue;  /* With next regex */
2117        }        }
2118      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1326  while (!done) Line 2122  while (!done)
2122    for (;;)    for (;;)
2123      {      {
2124      uschar *q;      uschar *q;
2125      uschar *bptr = dbuffer;      uschar *bptr;
2126      int *use_offsets = offsets;      int *use_offsets = offsets;
2127      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2128      int callout_data = 0;      int callout_data = 0;
2129      int callout_data_set = 0;      int callout_data_set = 0;
2130      int count, c;      int count, c;
2131      int copystrings = 0;      int copystrings = 0;
2132      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2133      int getstrings = 0;      int getstrings = 0;
2134      int getlist = 0;      int getlist = 0;
2135      int gmatched = 0;      int gmatched = 0;
2136      int start_offset = 0;      int start_offset = 0;
2137        int start_offset_sign = 1;
2138      int g_notempty = 0;      int g_notempty = 0;
2139      int use_dfa = 0;      int use_dfa = 0;
2140    
2141      options = 0;      options = 0;
2142    
2143        *copynames = 0;
2144        *getnames = 0;
2145    
2146        copynamesptr = copynames;
2147        getnamesptr = getnames;
2148    
2149      pcre_callout = callout;      pcre_callout = callout;
2150      first_callout = 1;      first_callout = 1;
2151      callout_extra = 0;      callout_extra = 0;
# Line 1351  while (!done) Line 2154  while (!done)
2154      callout_fail_id = -1;      callout_fail_id = -1;
2155      show_malloc = 0;      show_malloc = 0;
2156    
2157      if (infile == stdin) printf("data> ");      if (extra != NULL) extra->flags &=
2158      if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
2159    
2160        len = 0;
2161        for (;;)
2162        {        {
2163        done = 1;        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
2164        goto CONTINUE;          {
2165            if (len > 0)    /* Reached EOF without hitting a newline */
2166              {
2167              fprintf(outfile, "\n");
2168              break;
2169              }
2170            done = 1;
2171            goto CONTINUE;
2172            }
2173          if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
2174          len = (int)strlen((char *)buffer);
2175          if (buffer[len-1] == '\n') break;
2176        }        }
     if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);  
2177    
     len = (int)strlen((char *)buffer);  
2178      while (len > 0 && isspace(buffer[len-1])) len--;      while (len > 0 && isspace(buffer[len-1])) len--;
2179      buffer[len] = 0;      buffer[len] = 0;
2180      if (len == 0) break;      if (len == 0) break;
# Line 1367  while (!done) Line 2182  while (!done)
2182      p = buffer;      p = buffer;
2183      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2184    
2185      q = dbuffer;      bptr = q = dbuffer;
2186      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2187        {        {
2188        int i = 0;        int i = 0;
# Line 1389  while (!done) Line 2204  while (!done)
2204          c -= '0';          c -= '0';
2205          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
2206            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
2207    
2208    #if !defined NOUTF8
2209            if (use_utf8 && c > 255)
2210              {
2211              unsigned char buff8[8];
2212              int ii, utn;
2213              utn = ord2utf8(c, buff8);
2214              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2215              c = buff8[ii];   /* Last byte */
2216              }
2217    #endif
2218          break;          break;
2219    
2220          case 'x':          case 'x':
# Line 1406  while (!done) Line 2232  while (!done)
2232              {              {
2233              unsigned char buff8[8];              unsigned char buff8[8];
2234              int ii, utn;              int ii, utn;
2235              utn = ord2utf8(c, buff8);              if (use_utf8)
2236              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2237              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2238                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2239                  c = buff8[ii];   /* Last byte */
2240                  }
2241                else
2242                 {
2243                 if (c > 255)
2244                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2245                     "UTF-8 mode is not enabled.\n"
2246                     "** Truncation will probably give the wrong result.\n", c);
2247                 }
2248              p = pt + 1;              p = pt + 1;
2249              break;              break;
2250              }              }
# Line 1431  while (!done) Line 2267  while (!done)
2267          continue;          continue;
2268    
2269          case '>':          case '>':
2270            if (*p == '-')
2271              {
2272              start_offset_sign = -1;
2273              p++;
2274              }
2275          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2276            start_offset *= start_offset_sign;
2277          continue;          continue;
2278    
2279          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1450  while (!done) Line 2292  while (!done)
2292            }            }
2293          else if (isalnum(*p))          else if (isalnum(*p))
2294            {            {
2295            uschar name[256];            uschar *npp = copynamesptr;
           uschar *npp = name;  
2296            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2297              *npp++ = 0;
2298            *npp = 0;            *npp = 0;
2299            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)copynamesptr);
2300            if (n < 0)            if (n < 0)
2301              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
2302            else copystrings |= 1 << n;            copynamesptr = npp;
2303            }            }
2304          else if (*p == '+')          else if (*p == '+')
2305            {            {
# Line 1504  while (!done) Line 2346  while (!done)
2346  #endif  #endif
2347            use_dfa = 1;            use_dfa = 1;
2348          continue;          continue;
2349    #endif
2350    
2351    #if !defined NODFA
2352          case 'F':          case 'F':
2353          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2354          continue;          continue;
# Line 1518  while (!done) Line 2362  while (!done)
2362            }            }
2363          else if (isalnum(*p))          else if (isalnum(*p))
2364            {            {
2365            uschar name[256];            uschar *npp = getnamesptr;
           uschar *npp = name;  
2366            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2367              *npp++ = 0;
2368            *npp = 0;            *npp = 0;
2369            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)getnamesptr);
2370            if (n < 0)            if (n < 0)
2371              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
2372            else getstrings |= 1 << n;            getnamesptr = npp;
2373            }            }
2374          continue;          continue;
2375    
# Line 1538  while (!done) Line 2382  while (!done)
2382          continue;          continue;
2383    
2384          case 'N':          case 'N':
2385          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2386              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2387            else
2388              options |= PCRE_NOTEMPTY;
2389          continue;          continue;
2390    
2391          case 'O':          case 'O':
# Line 1551  while (!done) Line 2398  while (!done)
2398            if (offsets == NULL)            if (offsets == NULL)
2399              {              {
2400              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2401                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2402              yield = 1;              yield = 1;
2403              goto EXIT;              goto EXIT;
2404              }              }
# Line 1561  while (!done) Line 2408  while (!done)
2408          continue;          continue;
2409    
2410          case 'P':          case 'P':
2411          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2412              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2413            continue;
2414    
2415            case 'Q':
2416            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2417            if (extra == NULL)
2418              {
2419              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2420              extra->flags = 0;
2421              }
2422            extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
2423            extra->match_limit_recursion = n;
2424            continue;
2425    
2426            case 'q':
2427            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2428            if (extra == NULL)
2429              {
2430              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2431              extra->flags = 0;
2432              }
2433            extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
2434            extra->match_limit = n;
2435          continue;          continue;
2436    
2437  #if !defined NODFA  #if !defined NODFA
# Line 1574  while (!done) Line 2444  while (!done)
2444          show_malloc = 1;          show_malloc = 1;
2445          continue;          continue;
2446    
2447            case 'Y':
2448            options |= PCRE_NO_START_OPTIMIZE;
2449            continue;
2450    
2451          case 'Z':          case 'Z':
2452          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2453          continue;          continue;
# Line 1581  while (!done) Line 2455  while (!done)
2455          case '?':          case '?':
2456          options |= PCRE_NO_UTF8_CHECK;          options |= PCRE_NO_UTF8_CHECK;
2457          continue;          continue;
2458    
2459            case '<':
2460              {
2461              int x = check_newline(p, outfile);
2462              if (x == 0) goto NEXT_DATA;
2463              options |= x;
2464              while (*p++ != '>');
2465              }
2466            continue;
2467          }          }
2468        *q++ = c;        *q++ = c;
2469        }        }
2470      *q = 0;      *q = 0;
2471      len = q - dbuffer;      len = (int)(q - dbuffer);
2472    
2473        /* Move the data to the end of the buffer so that a read over the end of
2474        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2475        we are using the POSIX interface, we must include the terminating zero. */
2476    
2477    #if !defined NOPOSIX
2478        if (posix || do_posix)
2479          {
2480          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2481          bptr += buffer_size - len - 1;
2482          }
2483        else
2484    #endif
2485          {
2486          memmove(bptr + buffer_size - len, bptr, len);
2487          bptr += buffer_size - len;
2488          }
2489    
2490      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2491        {        {
# Line 1606  while (!done) Line 2506  while (!done)
2506          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2507        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2508        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2509          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2510    
2511        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2512    
2513        if (rc != 0)        if (rc != 0)
2514          {          {
2515          (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);          (void)regerror(rc, &preg, (char *)buffer, buffer_size);
2516          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
2517          }          }
2518        else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)        else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
# Line 1650  while (!done) Line 2551  while (!done)
2551    
2552      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2553        {        {
2554        if (timeit)        markptr = NULL;
2555    
2556          if (timeitm > 0)
2557          {          {
2558          register int i;          register int i;
2559          clock_t time_taken;          clock_t time_taken;
# Line 1660  while (!done) Line 2563  while (!done)
2563          if (all_use_dfa || use_dfa)          if (all_use_dfa || use_dfa)
2564            {            {
2565            int workspace[1000];            int workspace[1000];
2566            for (i = 0; i < LOOPREPEAT; i++)            for (i = 0; i < timeitm; i++)
2567              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2568                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2569                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2570            }            }
2571          else          else
2572  #endif  #endif
2573    
2574          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeitm; i++)
2575            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2576              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2577    
2578          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2579          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2580            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2581              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2582          }          }
2583    
# Line 1690  while (!done) Line 2593  while (!done)
2593            extra->flags = 0;            extra->flags = 0;
2594            }            }
2595    
2596          count = check_match_limit(re, extra, bptr, len, start_offset,          (void)check_match_limit(re, extra, bptr, len, start_offset,
2597            options|g_notempty, use_offsets, use_size_offsets,            options|g_notempty, use_offsets, use_size_offsets,
2598            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
2599            PCRE_ERROR_MATCHLIMIT, "match()");            PCRE_ERROR_MATCHLIMIT, "match()");
# Line 1724  while (!done) Line 2627  while (!done)
2627        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2628          {          {
2629          int workspace[1000];          int workspace[1000];
2630          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2631            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2632            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2633          if (count == 0)          if (count == 0)
# Line 1750  while (!done) Line 2653  while (!done)
2653    
2654        if (count >= 0)        if (count >= 0)
2655          {          {
2656          int i;          int i, maxcount;
2657    
2658    #if !defined NODFA
2659            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2660    #endif
2661              maxcount = use_size_offsets/3;
2662    
2663            /* This is a check against a lunatic return value. */
2664    
2665            if (count > maxcount)
2666              {
2667              fprintf(outfile,
2668                "** PCRE error: returned count %d is too big for offset size %d\n",
2669                count, use_size_offsets);
2670              count = use_size_offsets/3;
2671              if (do_g || do_G)
2672                {
2673                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2674                do_g = do_G = FALSE;        /* Break g/G loop */
2675                }
2676              }
2677    
2678          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2679            {            {
2680            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
# Line 1774  while (!done) Line 2698  while (!done)
2698              }              }
2699            }            }
2700    
2701            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2702    
2703          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2704            {            {
2705            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
2706              {              {
2707              char copybuffer[16];              char copybuffer[256];
2708              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,
2709                i, copybuffer, sizeof(copybuffer));                i, copybuffer, sizeof(copybuffer));
2710              if (rc < 0)              if (rc < 0)
# Line 1788  while (!done) Line 2714  while (!done)
2714              }              }
2715            }            }
2716    
2717            for (copynamesptr = copynames;
2718                 *copynamesptr != 0;
2719                 copynamesptr += (int)strlen((char*)copynamesptr) + 1)
2720              {
2721              char copybuffer[256];
2722              int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,
2723                count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));
2724              if (rc < 0)
2725                fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);
2726              else
2727                fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);
2728              }
2729    
2730          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2731            {            {
2732            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
# Line 1800  while (!done) Line 2739  while (!done)
2739              else              else
2740                {                {
2741                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
               /* free((void *)substring); */  
2742                pcre_free_substring(substring);                pcre_free_substring(substring);
2743                }                }
2744              }              }
2745            }            }
2746    
2747            for (getnamesptr = getnames;
2748                 *getnamesptr != 0;
2749                 getnamesptr += (int)strlen((char*)getnamesptr) + 1)
2750              {
2751              const char *substring;
2752              int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,
2753                count, (char *)getnamesptr, &substring);
2754              if (rc < 0)
2755                fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);
2756              else
2757                {
2758                fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);
2759                pcre_free_substring(substring);
2760                }
2761              }
2762    
2763          if (getlist)          if (getlist)
2764            {            {
2765            const char **stringlist;            const char **stringlist;
# Line 1829  while (!done) Line 2783  while (!done)
2783    
2784        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2785          {          {
2786          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2787  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2788          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2789            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2790              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2791  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2792                outfile);
2793              }
2794          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2795          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2796          }          }
2797    
2798        /* 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
2799        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
2800        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
2801        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2802        offset values to achieve this. We won't be at the end of the string -  
2803        that was checked before setting g_notempty. */        Complication arises in the case when the newline convention is "any",
2804          "crlf", or "anycrlf". If the previous match was at the end of a line
2805          terminated by CRLF, an advance of one character just passes the \r,
2806          whereas we should prefer the longer newline sequence, as does the code in
2807          pcre_exec(). Fudge the offset value to achieve this. We check for a
2808          newline setting in the pattern; if none was set, use pcre_config() to
2809          find the default.
2810    
2811          Otherwise, in the case of UTF-8 matching, the advance must be one
2812          character, not one byte. */
2813    
2814        else        else
2815          {          {
2816          if (g_notempty != 0)          if (g_notempty != 0)
2817            {            {
2818            int onechar = 1;            int onechar = 1;
2819              unsigned int obits = ((real_pcre *)re)->options;
2820            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2821            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2822                {
2823                int d;
2824                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2825                /* Note that these values are always the ASCII ones, even in
2826                EBCDIC environments. CR = 13, NL = 10. */
2827                obits = (d == 13)? PCRE_NEWLINE_CR :
2828                        (d == 10)? PCRE_NEWLINE_LF :
2829                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2830                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2831                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2832                }
2833              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2834                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2835                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2836                  &&
2837                  start_offset < len - 1 &&
2838                  bptr[start_offset] == '\r' &&
2839                  bptr[start_offset+1] == '\n')
2840                onechar++;
2841              else if (use_utf8)
2842              {              {
2843              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2844                {                {
2845                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2846                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2847                }                }
2848              }              }
2849            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
# Line 1868  while (!done) Line 2852  while (!done)
2852            {            {
2853            if (count == PCRE_ERROR_NOMATCH)            if (count == PCRE_ERROR_NOMATCH)
2854              {              {
2855              if (gmatched == 0) fprintf(outfile, "No match\n");              if (gmatched == 0)
2856                  {
2857                  if (markptr == NULL) fprintf(outfile, "No match\n");
2858                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2859                  }
2860              }              }
2861            else fprintf(outfile, "Error %d\n", count);            else fprintf(outfile, "Error %d\n", count);
2862            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
# Line 1880  while (!done) Line 2868  while (!done)
2868        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2869    
2870        /* 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
2871        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
2872        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
2873        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
2874        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
2875        character. */        character. */
2876    
2877        g_notempty = 0;        g_notempty = 0;
2878    
2879        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2880          {          {
2881          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2882          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2883          }          }
2884    
2885        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 1905  while (!done) Line 2894  while (!done)
2894          len -= use_offsets[1];          len -= use_offsets[1];
2895          }          }
2896        }  /* End of loop for /g and /G */        }  /* End of loop for /g and /G */
2897    
2898        NEXT_DATA: continue;
2899      }    /* End of loop for data lines */      }    /* End of loop for data lines */
2900    
2901    CONTINUE:    CONTINUE:
# Line 1915  while (!done) Line 2906  while (!done)
2906    
2907    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
2908    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
2909    if (tables != NULL)    if (locale_set)
2910      {      {
2911      new_free((void *)tables);      new_free((void *)tables);
2912      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
2913        locale_set = 0;
2914      }      }
2915    }    }
2916    

Legend:
Removed from v.87  
changed lines
  Added in v.579

  ViewVC Help
Powered by ViewVC 1.1.5