/[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 91 by nigel, Sat Feb 24 21:41:34 2007 UTC revision 606 by ph10, Mon Jun 6 17:46:22 2011 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 44  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51  #ifndef _WIN32  #ifdef SUPPORT_LIBREADLINE
52  #include <sys/resource.h>  #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57  #endif  #endif
58    
 #define PCRE_SPY        /* For Win32 build, import data, not export */  
59    
60  /* We include pcre_internal.h because we need the internal info for displaying  /* A number of things vary for Windows builds. Originally, pcretest opened its
61  the results of pcre_study() and we also need to know about the internal  input and output without "b"; then I was told that "b" was needed in some
62  macros, structures, and other internal data values; pcretest has "inside  environments, so it was added for release 5.0 to both the input and output. (It
63  information" compared to a program that strictly follows the PCRE API. */  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    /* A user sent this fix for Borland Builder 5 under Windows. */
83    
84    #ifdef __BORLANDC__
85    #define _setmode(handle, mode) setmode(handle, mode)
86    #endif
87    
88    /* Not Windows */
89    
90    #else
91    #include <sys/time.h>          /* These two includes are needed */
92    #include <sys/resource.h>      /* for setrlimit(). */
93    #define INPUT_MODE   "rb"
94    #define OUTPUT_MODE  "wb"
95    #endif
96    
97    
98    /* We have to include pcre_internal.h because we need the internal info for
99    displaying the results of pcre_study() and we also need to know about the
100    internal macros, structures, and other internal data values; pcretest has
101    "inside information" compared to a program that strictly follows the PCRE API.
102    
103    Although pcre_internal.h does itself include pcre.h, we explicitly include it
104    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
105    appropriately for an application, not for building PCRE. */
106    
107    #include "pcre.h"
108  #include "pcre_internal.h"  #include "pcre_internal.h"
109    
110  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* We need access to some of the data tables that PCRE uses. So as not to have
111  two copies, we include the source file here, changing the names of the external  to keep two copies, we include the source file here, changing the names of the
112  symbols to prevent clashes. */  external symbols to prevent clashes. */
113    
114    #define _pcre_ucp_gentype      ucp_gentype
115  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
116  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
117  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 68  symbols to prevent clashes. */ Line 119  symbols to prevent clashes. */
119  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
120  #define _pcre_utt              utt  #define _pcre_utt              utt
121  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
122    #define _pcre_utt_names        utt_names
123  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
124    
125  #include "pcre_tables.c"  #include "pcre_tables.c"
126    
127  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
128  patterns. This function is in a separate file so that it can be included in  patterns. This function is in a separate file so that it can be included in
129  pcre_compile.c when that module is compiled with debugging enabled. */  pcre_compile.c when that module is compiled with debugging enabled. It needs to
130    know which case is being compiled. */
131    
132    #define COMPILING_PCRETEST
133  #include "pcre_printint.src"  #include "pcre_printint.src"
134    
135    /* The definition of the macro PRINTABLE, which determines whether to print an
136    output character as-is or as a hex value when showing compiled patterns, is
137    contained in the printint.src file. We uses it here also, in cases when the
138    locale has not been explicitly changed, so as to get consistent output from
139    systems that differ in their output from isprint() even in the "C" locale. */
140    
141    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
142    
143  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
144  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 87  Makefile. */ Line 148  Makefile. */
148  #include "pcreposix.h"  #include "pcreposix.h"
149  #endif  #endif
150    
151  /* 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
152  build pcretest without support for UTF8 (define NOUTF8), without the interface  Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
153  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
154  function (define NOINFOCHECK). */  "info" function (define NOINFOCHECK). In fact, we automatically cut out the
155    UTF8 support if PCRE is built without it. */
156    
157    #ifndef SUPPORT_UTF8
158    #ifndef NOUTF8
159    #define NOUTF8
160    #endif
161    #endif
162    
163    
164  /* Other parameters */  /* Other parameters */
# Line 103  function (define NOINFOCHECK). */ Line 171  function (define NOINFOCHECK). */
171  #endif  #endif
172  #endif  #endif
173    
174    /* This is the default loop count for timing. */
175    
176  #define LOOPREPEAT 500000  #define LOOPREPEAT 500000
177    
178  /* Static variables */  /* Static variables */
# Line 113  static int callout_count; Line 183  static int callout_count;
183  static int callout_extra;  static int callout_extra;
184  static int callout_fail_count;  static int callout_fail_count;
185  static int callout_fail_id;  static int callout_fail_id;
186    static int debug_lengths;
187  static int first_callout;  static int first_callout;
188    static int locale_set = 0;
189  static int show_malloc;  static int show_malloc;
190  static int use_utf8;  static int use_utf8;
191  static size_t gotten_store;  static size_t gotten_store;
# Line 125  static uschar *buffer = NULL; Line 197  static uschar *buffer = NULL;
197  static uschar *dbuffer = NULL;  static uschar *dbuffer = NULL;
198  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
199    
200    /* Textual explanations for runtime error codes */
201    
202    static const char *errtexts[] = {
203      NULL,  /* 0 is no error */
204      NULL,  /* NOMATCH is handled specially */
205      "NULL argument passed",
206      "bad option value",
207      "magic number missing",
208      "unknown opcode - pattern overwritten?",
209      "no more memory",
210      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
211      "match limit exceeded",
212      "callout error code",
213      NULL,  /* BADUTF8 is handled specially */
214      "bad UTF-8 offset",
215      NULL,  /* PARTIAL is handled specially */
216      "not used - internal error",
217      "internal error - pattern overwritten?",
218      "bad count value",
219      "item unsupported for DFA matching",
220      "backreference condition or recursion test not supported for DFA matching",
221      "match limit not supported for DFA matching",
222      "workspace size exceeded in DFA matching",
223      "too much recursion for DFA matching",
224      "recursion limit exceeded",
225      "not used - internal error",
226      "invalid combination of newline options",
227      "bad offset value",
228      NULL  /* SHORTUTF8 is handled specially */
229    };
230    
231    
232    /*************************************************
233    *         Alternate character tables             *
234    *************************************************/
235    
236    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
237    using the default tables of the library. However, the T option can be used to
238    select alternate sets of tables, for different kinds of testing. Note also that
239    the L (locale) option also adjusts the tables. */
240    
241    /* This is the set of tables distributed as default with PCRE. It recognizes
242    only ASCII characters. */
243    
244    static const unsigned char tables0[] = {
245    
246    /* This table is a lower casing table. */
247    
248        0,  1,  2,  3,  4,  5,  6,  7,
249        8,  9, 10, 11, 12, 13, 14, 15,
250       16, 17, 18, 19, 20, 21, 22, 23,
251       24, 25, 26, 27, 28, 29, 30, 31,
252       32, 33, 34, 35, 36, 37, 38, 39,
253       40, 41, 42, 43, 44, 45, 46, 47,
254       48, 49, 50, 51, 52, 53, 54, 55,
255       56, 57, 58, 59, 60, 61, 62, 63,
256       64, 97, 98, 99,100,101,102,103,
257      104,105,106,107,108,109,110,111,
258      112,113,114,115,116,117,118,119,
259      120,121,122, 91, 92, 93, 94, 95,
260       96, 97, 98, 99,100,101,102,103,
261      104,105,106,107,108,109,110,111,
262      112,113,114,115,116,117,118,119,
263      120,121,122,123,124,125,126,127,
264      128,129,130,131,132,133,134,135,
265      136,137,138,139,140,141,142,143,
266      144,145,146,147,148,149,150,151,
267      152,153,154,155,156,157,158,159,
268      160,161,162,163,164,165,166,167,
269      168,169,170,171,172,173,174,175,
270      176,177,178,179,180,181,182,183,
271      184,185,186,187,188,189,190,191,
272      192,193,194,195,196,197,198,199,
273      200,201,202,203,204,205,206,207,
274      208,209,210,211,212,213,214,215,
275      216,217,218,219,220,221,222,223,
276      224,225,226,227,228,229,230,231,
277      232,233,234,235,236,237,238,239,
278      240,241,242,243,244,245,246,247,
279      248,249,250,251,252,253,254,255,
280    
281    /* This table is a case flipping table. */
282    
283        0,  1,  2,  3,  4,  5,  6,  7,
284        8,  9, 10, 11, 12, 13, 14, 15,
285       16, 17, 18, 19, 20, 21, 22, 23,
286       24, 25, 26, 27, 28, 29, 30, 31,
287       32, 33, 34, 35, 36, 37, 38, 39,
288       40, 41, 42, 43, 44, 45, 46, 47,
289       48, 49, 50, 51, 52, 53, 54, 55,
290       56, 57, 58, 59, 60, 61, 62, 63,
291       64, 97, 98, 99,100,101,102,103,
292      104,105,106,107,108,109,110,111,
293      112,113,114,115,116,117,118,119,
294      120,121,122, 91, 92, 93, 94, 95,
295       96, 65, 66, 67, 68, 69, 70, 71,
296       72, 73, 74, 75, 76, 77, 78, 79,
297       80, 81, 82, 83, 84, 85, 86, 87,
298       88, 89, 90,123,124,125,126,127,
299      128,129,130,131,132,133,134,135,
300      136,137,138,139,140,141,142,143,
301      144,145,146,147,148,149,150,151,
302      152,153,154,155,156,157,158,159,
303      160,161,162,163,164,165,166,167,
304      168,169,170,171,172,173,174,175,
305      176,177,178,179,180,181,182,183,
306      184,185,186,187,188,189,190,191,
307      192,193,194,195,196,197,198,199,
308      200,201,202,203,204,205,206,207,
309      208,209,210,211,212,213,214,215,
310      216,217,218,219,220,221,222,223,
311      224,225,226,227,228,229,230,231,
312      232,233,234,235,236,237,238,239,
313      240,241,242,243,244,245,246,247,
314      248,249,250,251,252,253,254,255,
315    
316    /* This table contains bit maps for various character classes. Each map is 32
317    bytes long and the bits run from the least significant end of each byte. The
318    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
319    graph, print, punct, and cntrl. Other classes are built from combinations. */
320    
321      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
322      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
323      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
324      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
325    
326      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
327      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
328      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
329      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
330    
331      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
332      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
333      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
334      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
335    
336      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
337      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
338      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
339      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
340    
341      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
342      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
343      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
344      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
345    
346      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
347      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
348      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
349      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
350    
351      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
352      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
353      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
354      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
355    
356      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
357      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
358      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
359      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
360    
361      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
362      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
363      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
364      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
365    
366      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
367      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
368      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
369      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
370    
371    /* This table identifies various classes of character by individual bits:
372      0x01   white space character
373      0x02   letter
374      0x04   decimal digit
375      0x08   hexadecimal digit
376      0x10   alphanumeric or '_'
377      0x80   regular expression metacharacter or binary zero
378    */
379    
380      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
381      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
382      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
383      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
384      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
385      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
386      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
387      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
388      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
389      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
390      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
391      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
392      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
393      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
394      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
395      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
396      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
397      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
398      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
399      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
400      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
401      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
412    
413    /* This is a set of tables that came orginally from a Windows user. It seems to
414    be at least an approximation of ISO 8859. In particular, there are characters
415    greater than 128 that are marked as spaces, letters, etc. */
416    
417    static const unsigned char tables1[] = {
418    0,1,2,3,4,5,6,7,
419    8,9,10,11,12,13,14,15,
420    16,17,18,19,20,21,22,23,
421    24,25,26,27,28,29,30,31,
422    32,33,34,35,36,37,38,39,
423    40,41,42,43,44,45,46,47,
424    48,49,50,51,52,53,54,55,
425    56,57,58,59,60,61,62,63,
426    64,97,98,99,100,101,102,103,
427    104,105,106,107,108,109,110,111,
428    112,113,114,115,116,117,118,119,
429    120,121,122,91,92,93,94,95,
430    96,97,98,99,100,101,102,103,
431    104,105,106,107,108,109,110,111,
432    112,113,114,115,116,117,118,119,
433    120,121,122,123,124,125,126,127,
434    128,129,130,131,132,133,134,135,
435    136,137,138,139,140,141,142,143,
436    144,145,146,147,148,149,150,151,
437    152,153,154,155,156,157,158,159,
438    160,161,162,163,164,165,166,167,
439    168,169,170,171,172,173,174,175,
440    176,177,178,179,180,181,182,183,
441    184,185,186,187,188,189,190,191,
442    224,225,226,227,228,229,230,231,
443    232,233,234,235,236,237,238,239,
444    240,241,242,243,244,245,246,215,
445    248,249,250,251,252,253,254,223,
446    224,225,226,227,228,229,230,231,
447    232,233,234,235,236,237,238,239,
448    240,241,242,243,244,245,246,247,
449    248,249,250,251,252,253,254,255,
450    0,1,2,3,4,5,6,7,
451    8,9,10,11,12,13,14,15,
452    16,17,18,19,20,21,22,23,
453    24,25,26,27,28,29,30,31,
454    32,33,34,35,36,37,38,39,
455    40,41,42,43,44,45,46,47,
456    48,49,50,51,52,53,54,55,
457    56,57,58,59,60,61,62,63,
458    64,97,98,99,100,101,102,103,
459    104,105,106,107,108,109,110,111,
460    112,113,114,115,116,117,118,119,
461    120,121,122,91,92,93,94,95,
462    96,65,66,67,68,69,70,71,
463    72,73,74,75,76,77,78,79,
464    80,81,82,83,84,85,86,87,
465    88,89,90,123,124,125,126,127,
466    128,129,130,131,132,133,134,135,
467    136,137,138,139,140,141,142,143,
468    144,145,146,147,148,149,150,151,
469    152,153,154,155,156,157,158,159,
470    160,161,162,163,164,165,166,167,
471    168,169,170,171,172,173,174,175,
472    176,177,178,179,180,181,182,183,
473    184,185,186,187,188,189,190,191,
474    224,225,226,227,228,229,230,231,
475    232,233,234,235,236,237,238,239,
476    240,241,242,243,244,245,246,215,
477    248,249,250,251,252,253,254,223,
478    192,193,194,195,196,197,198,199,
479    200,201,202,203,204,205,206,207,
480    208,209,210,211,212,213,214,247,
481    216,217,218,219,220,221,222,255,
482    0,62,0,0,1,0,0,0,
483    0,0,0,0,0,0,0,0,
484    32,0,0,0,1,0,0,0,
485    0,0,0,0,0,0,0,0,
486    0,0,0,0,0,0,255,3,
487    126,0,0,0,126,0,0,0,
488    0,0,0,0,0,0,0,0,
489    0,0,0,0,0,0,0,0,
490    0,0,0,0,0,0,255,3,
491    0,0,0,0,0,0,0,0,
492    0,0,0,0,0,0,12,2,
493    0,0,0,0,0,0,0,0,
494    0,0,0,0,0,0,0,0,
495    254,255,255,7,0,0,0,0,
496    0,0,0,0,0,0,0,0,
497    255,255,127,127,0,0,0,0,
498    0,0,0,0,0,0,0,0,
499    0,0,0,0,254,255,255,7,
500    0,0,0,0,0,4,32,4,
501    0,0,0,128,255,255,127,255,
502    0,0,0,0,0,0,255,3,
503    254,255,255,135,254,255,255,7,
504    0,0,0,0,0,4,44,6,
505    255,255,127,255,255,255,127,255,
506    0,0,0,0,254,255,255,255,
507    255,255,255,255,255,255,255,127,
508    0,0,0,0,254,255,255,255,
509    255,255,255,255,255,255,255,255,
510    0,2,0,0,255,255,255,255,
511    255,255,255,255,255,255,255,127,
512    0,0,0,0,255,255,255,255,
513    255,255,255,255,255,255,255,255,
514    0,0,0,0,254,255,0,252,
515    1,0,0,248,1,0,0,120,
516    0,0,0,0,254,255,255,255,
517    0,0,128,0,0,0,128,0,
518    255,255,255,255,0,0,0,0,
519    0,0,0,0,0,0,0,128,
520    255,255,255,255,0,0,0,0,
521    0,0,0,0,0,0,0,0,
522    128,0,0,0,0,0,0,0,
523    0,1,1,0,1,1,0,0,
524    0,0,0,0,0,0,0,0,
525    0,0,0,0,0,0,0,0,
526    1,0,0,0,128,0,0,0,
527    128,128,128,128,0,0,128,0,
528    28,28,28,28,28,28,28,28,
529    28,28,0,0,0,0,0,128,
530    0,26,26,26,26,26,26,18,
531    18,18,18,18,18,18,18,18,
532    18,18,18,18,18,18,18,18,
533    18,18,18,128,128,0,128,16,
534    0,26,26,26,26,26,26,18,
535    18,18,18,18,18,18,18,18,
536    18,18,18,18,18,18,18,18,
537    18,18,18,128,128,0,0,0,
538    0,0,0,0,0,1,0,0,
539    0,0,0,0,0,0,0,0,
540    0,0,0,0,0,0,0,0,
541    0,0,0,0,0,0,0,0,
542    1,0,0,0,0,0,0,0,
543    0,0,18,0,0,0,0,0,
544    0,0,20,20,0,18,0,0,
545    0,20,18,0,0,0,0,0,
546    18,18,18,18,18,18,18,18,
547    18,18,18,18,18,18,18,18,
548    18,18,18,18,18,18,18,0,
549    18,18,18,18,18,18,18,18,
550    18,18,18,18,18,18,18,18,
551    18,18,18,18,18,18,18,18,
552    18,18,18,18,18,18,18,0,
553    18,18,18,18,18,18,18,18
554    };
555    
556    
557    
558    
559    #ifndef HAVE_STRERROR
560    /*************************************************
561    *     Provide strerror() for non-ANSI libraries  *
562    *************************************************/
563    
564    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
565    in their libraries, but can provide the same facility by this simple
566    alternative function. */
567    
568    extern int   sys_nerr;
569    extern char *sys_errlist[];
570    
571    char *
572    strerror(int n)
573    {
574    if (n < 0 || n >= sys_nerr) return "unknown error number";
575    return sys_errlist[n];
576    }
577    #endif /* HAVE_STRERROR */
578    
579    
580    
581    
582  /*************************************************  /*************************************************
# Line 143  optimal way of handling this, but hey, t Line 595  optimal way of handling this, but hey, t
595  Arguments:  Arguments:
596    f            the file to read    f            the file to read
597    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
598      prompt       for stdin or readline()
599    
600  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
601                 could be a copy of start, or could be moved                 could be a copy of start, or could be moved
# Line 150  Returns:       pointer to the start of n Line 603  Returns:       pointer to the start of n
603  */  */
604    
605  static uschar *  static uschar *
606  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
607  {  {
608  uschar *here = start;  uschar *here = start;
609    
610  for (;;)  for (;;)
611    {    {
612    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
613    
614    if (rlen > 1000)    if (rlen > 1000)
615      {      {
616      int dlen;      int dlen;
617      if (fgets((char *)here, rlen,  f) == NULL)  
618        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
619        input is a terminal. Note that readline() removes the trailing newline, so
620        we must put it back again, to be compatible with fgets(). */
621    
622    #ifdef SUPPORT_LIBREADLINE
623        if (isatty(fileno(f)))
624          {
625          size_t len;
626          char *s = readline(prompt);
627          if (s == NULL) return (here == start)? NULL : start;
628          len = strlen(s);
629          if (len > 0) add_history(s);
630          if (len > rlen - 1) len = rlen - 1;
631          memcpy(here, s, len);
632          here[len] = '\n';
633          here[len+1] = 0;
634          free(s);
635          }
636        else
637    #endif
638    
639        /* Read the next line by normal means, prompting if the file is stdin. */
640    
641          {
642          if (f == stdin) printf("%s", prompt);
643          if (fgets((char *)here, rlen,  f) == NULL)
644            return (here == start)? NULL : start;
645          }
646    
647      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
648      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
649      here += dlen;      here += dlen;
# Line 213  return NULL;  /* Control never gets here Line 695  return NULL;  /* Control never gets here
695    
696  /* 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
697  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
698  for unpicking the -o argument, so just keep it simple.  for unpicking arguments, so just keep it simple.
699    
700  Arguments:  Arguments:
701    str           string to be converted    str           string to be converted
# Line 311  Arguments: Line 793  Arguments:
793  Returns:     number of characters placed in the buffer  Returns:     number of characters placed in the buffer
794  */  */
795    
796    #if !defined NOUTF8
797    
798  static int  static int
799  ord2utf8(int cvalue, uschar *utf8bytes)  ord2utf8(int cvalue, uschar *utf8bytes)
800  {  {
# Line 327  for (j = i; j > 0; j--) Line 811  for (j = i; j > 0; j--)
811  return i + 1;  return i + 1;
812  }  }
813    
814    #endif
815    
816    
817    
818  /*************************************************  /*************************************************
# Line 353  while (length-- > 0) Line 839  while (length-- > 0)
839        {        {
840        length -= rc - 1;        length -= rc - 1;
841        p += rc;        p += rc;
842        if (c < 256 && isprint(c))        if (PRINTHEX(c))
843          {          {
844          if (f != NULL) fprintf(f, "%c", c);          if (f != NULL) fprintf(f, "%c", c);
845          yield++;          yield++;
846          }          }
847        else        else
848          {          {
849          int n;          int n = 4;
850          if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);          if (f != NULL) fprintf(f, "\\x{%02x}", c);
851          yield += n;          yield += (n <= 0x000000ff)? 2 :
852                     (n <= 0x00000fff)? 3 :
853                     (n <= 0x0000ffff)? 4 :
854                     (n <= 0x000fffff)? 5 : 6;
855          }          }
856        continue;        continue;
857        }        }
# Line 371  while (length-- > 0) Line 860  while (length-- > 0)
860    
861     /* Not UTF-8, or malformed UTF-8  */     /* Not UTF-8, or malformed UTF-8  */
862    
863    if (isprint(c = *(p++)))    c = *p++;
864      if (PRINTHEX(c))
865      {      {
866      if (f != NULL) fprintf(f, "%c", c);      if (f != NULL) fprintf(f, "%c", c);
867      yield++;      yield++;
# Line 610  return count; Line 1100  return count;
1100    
1101    
1102  /*************************************************  /*************************************************
1103    *         Case-independent strncmp() function    *
1104    *************************************************/
1105    
1106    /*
1107    Arguments:
1108      s         first string
1109      t         second string
1110      n         number of characters to compare
1111    
1112    Returns:    < 0, = 0, or > 0, according to the comparison
1113    */
1114    
1115    static int
1116    strncmpic(uschar *s, uschar *t, int n)
1117    {
1118    while (n--)
1119      {
1120      int c = tolower(*s++) - tolower(*t++);
1121      if (c) return c;
1122      }
1123    return 0;
1124    }
1125    
1126    
1127    
1128    /*************************************************
1129  *         Check newline indicator                *  *         Check newline indicator                *
1130  *************************************************/  *************************************************/
1131    
1132  /* This is used both at compile and run-time to check for <xxx> escapes, where  /* This is used both at compile and run-time to check for <xxx> escapes. Print
1133  xxx is LF, CR, or CRLF. Print a message and return 0 if there is no match.  a message and return 0 if there is no match.
1134    
1135  Arguments:  Arguments:
1136    p           points after the leading '<'    p           points after the leading '<'
# Line 626  Returns:      appropriate PCRE_NEWLINE_x Line 1142  Returns:      appropriate PCRE_NEWLINE_x
1142  static int  static int
1143  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
1144  {  {
1145  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1146  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1147  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1148    if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1149    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1150    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1151    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1152  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
1153  return 0;  return 0;
1154  }  }
# Line 636  return 0; Line 1156  return 0;
1156    
1157    
1158  /*************************************************  /*************************************************
1159    *             Usage function                     *
1160    *************************************************/
1161    
1162    static void
1163    usage(void)
1164    {
1165    printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1166    printf("Input and output default to stdin and stdout.\n");
1167    #ifdef SUPPORT_LIBREADLINE
1168    printf("If input is a terminal, readline() is used to read from it.\n");
1169    #else
1170    printf("This version of pcretest is not linked with readline().\n");
1171    #endif
1172    printf("\nOptions:\n");
1173    printf("  -b       show compiled code (bytecode)\n");
1174    printf("  -C       show PCRE compile-time options and exit\n");
1175    printf("  -d       debug: show compiled code and information (-b and -i)\n");
1176    #if !defined NODFA
1177    printf("  -dfa     force DFA matching for all subjects\n");
1178    #endif
1179    printf("  -help    show usage information\n");
1180    printf("  -i       show information about compiled patterns\n"
1181           "  -M       find MATCH_LIMIT minimum for each subject\n"
1182           "  -m       output memory used information\n"
1183           "  -o <n>   set size of offsets vector to <n>\n");
1184    #if !defined NOPOSIX
1185    printf("  -p       use POSIX interface\n");
1186    #endif
1187    printf("  -q       quiet: do not output PCRE version number at start\n");
1188    printf("  -S <n>   set stack size to <n> megabytes\n");
1189    printf("  -s       force each pattern to be studied\n"
1190           "  -t       time compilation and execution\n");
1191    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1192    printf("  -tm      time execution (matching) only\n");
1193    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
1194    }
1195    
1196    
1197    
1198    /*************************************************
1199  *                Main Program                    *  *                Main Program                    *
1200  *************************************************/  *************************************************/
1201    
# Line 648  int main(int argc, char **argv) Line 1208  int main(int argc, char **argv)
1208  FILE *infile = stdin;  FILE *infile = stdin;
1209  int options = 0;  int options = 0;
1210  int study_options = 0;  int study_options = 0;
1211    int default_find_match_limit = FALSE;
1212  int op = 1;  int op = 1;
1213  int timeit = 0;  int timeit = 0;
1214    int timeitm = 0;
1215  int showinfo = 0;  int showinfo = 0;
1216  int showstore = 0;  int showstore = 0;
1217    int force_study = 0;
1218  int quiet = 0;  int quiet = 0;
1219  int size_offsets = 45;  int size_offsets = 45;
1220  int size_offsets_max;  int size_offsets_max;
# Line 681  buffer = (unsigned char *)malloc(buffer_ Line 1244  buffer = (unsigned char *)malloc(buffer_
1244  dbuffer = (unsigned char *)malloc(buffer_size);  dbuffer = (unsigned char *)malloc(buffer_size);
1245  pbuffer = (unsigned char *)malloc(buffer_size);  pbuffer = (unsigned char *)malloc(buffer_size);
1246    
1247  /* The outfile variable is static so that new_malloc can use it. The _setmode()  /* The outfile variable is static so that new_malloc can use it. */
 stuff is some magic that I don't understand, but which apparently does good  
 things in Windows. It's related to line terminations.  */  
   
 #if defined(_WIN32) || defined(WIN32)  
 _setmode( _fileno( stdout ), 0x8000 );  
 #endif  /* defined(_WIN32) || defined(WIN32) */  
1248    
1249  outfile = stdout;  outfile = stdout;
1250    
1251    /* The following  _setmode() stuff is some Windows magic that tells its runtime
1252    library to translate CRLF into a single LF character. At least, that's what
1253    I've been told: never having used Windows I take this all on trust. Originally
1254    it set 0x8000, but then I was advised that _O_BINARY was better. */
1255    
1256    #if defined(_WIN32) || defined(WIN32)
1257    _setmode( _fileno( stdout ), _O_BINARY );
1258    #endif
1259    
1260  /* Scan options */  /* Scan options */
1261    
1262  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
1263    {    {
1264    unsigned char *endptr;    unsigned char *endptr;
1265    
1266    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1267      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 1;
   else if (strcmp(argv[op], "-t") == 0) timeit = 1;  
1268    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1269      else if (strcmp(argv[op], "-b") == 0) debug = 1;
1270    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1271    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1272      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1273  #if !defined NODFA  #if !defined NODFA
1274    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1275  #endif  #endif
# Line 713  while (argc > 1 && argv[op][0] == '-') Line 1280  while (argc > 1 && argv[op][0] == '-')
1280      op++;      op++;
1281      argc--;      argc--;
1282      }      }
1283      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
1284        {
1285        int both = argv[op][2] == 0;
1286        int temp;
1287        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
1288                         *endptr == 0))
1289          {
1290          timeitm = temp;
1291          op++;
1292          argc--;
1293          }
1294        else timeitm = LOOPREPEAT;
1295        if (both) timeit = timeitm;
1296        }
1297    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
1298        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1299          *endptr == 0))          *endptr == 0))
1300      {      {
1301  #ifdef _WIN32  #if defined(_WIN32) || defined(WIN32)
1302      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
1303      exit(1);      exit(1);
1304  #else  #else
# Line 741  while (argc > 1 && argv[op][0] == '-') Line 1322  while (argc > 1 && argv[op][0] == '-')
1322    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1323      {      {
1324      int rc;      int rc;
1325        unsigned long int lrc;
1326      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1327      printf("Compiled with\n");      printf("Compiled with\n");
1328      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 748  while (argc > 1 && argv[op][0] == '-') Line 1330  while (argc > 1 && argv[op][0] == '-')
1330      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1331      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1332      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1333      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
1334        (rc == '\n')? "LF" : "CRLF");      in EBCDIC environments. CR is 13 and NL is 10. */
1335        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1336          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1337          (rc == -2)? "ANYCRLF" :
1338          (rc == -1)? "ANY" : "???");
1339        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1340        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1341                                         "all Unicode newlines");
1342      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1343      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1344      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1345      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1346      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1347      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1348      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1349      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
1350      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1351      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1352      exit(0);      goto EXIT;
1353        }
1354      else if (strcmp(argv[op], "-help") == 0 ||
1355               strcmp(argv[op], "--help") == 0)
1356        {
1357        usage();
1358        goto EXIT;
1359      }      }
1360    else    else
1361      {      {
1362      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
1363      printf("Usage:   pcretest [options] [<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 <n> set stack size to <n> megabytes\n");  
     printf("  -s     output store (memory) used information\n"  
            "  -t     time compilation and execution\n");  
1364      yield = 1;      yield = 1;
1365      goto EXIT;      goto EXIT;
1366      }      }
# Line 794  offsets = (int *)malloc(size_offsets_max Line 1375  offsets = (int *)malloc(size_offsets_max
1375  if (offsets == NULL)  if (offsets == NULL)
1376    {    {
1377    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1378      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1379    yield = 1;    yield = 1;
1380    goto EXIT;    goto EXIT;
1381    }    }
# Line 803  if (offsets == NULL) Line 1384  if (offsets == NULL)
1384    
1385  if (argc > 1)  if (argc > 1)
1386    {    {
1387    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
1388    if (infile == NULL)    if (infile == NULL)
1389      {      {
1390      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
# Line 814  if (argc > 1) Line 1395  if (argc > 1)
1395    
1396  if (argc > 2)  if (argc > 2)
1397    {    {
1398    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
1399    if (outfile == NULL)    if (outfile == NULL)
1400      {      {
1401      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
# Line 847  while (!done) Line 1428  while (!done)
1428  #endif  #endif
1429    
1430    const char *error;    const char *error;
1431      unsigned char *markptr;
1432    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1433    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1434    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1435    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1436    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1437      int do_mark = 0;
1438    int do_study = 0;    int do_study = 0;
1439    int do_debug = debug;    int do_debug = debug;
1440    int do_G = 0;    int do_G = 0;
# Line 859  while (!done) Line 1442  while (!done)
1442    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1443    int do_showrest = 0;    int do_showrest = 0;
1444    int do_flip = 0;    int do_flip = 0;
1445    int erroroffset, len, delimiter;    int erroroffset, len, delimiter, poffset;
1446    
1447    use_utf8 = 0;    use_utf8 = 0;
1448      debug_lengths = 1;
1449    
1450    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1451    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1452    fflush(outfile);    fflush(outfile);
1453    
# Line 964  while (!done) Line 1547  while (!done)
1547    
1548    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1549      {      {
1550      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1551      goto SKIP_DATA;      goto SKIP_DATA;
1552      }      }
1553    
1554    pp = p;    pp = p;
1555      poffset = (int)(p - buffer);
1556    
1557    for(;;)    for(;;)
1558      {      {
# Line 979  while (!done) Line 1563  while (!done)
1563        pp++;        pp++;
1564        }        }
1565      if (*pp != 0) break;      if (*pp != 0) break;
1566      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1567        {        {
1568        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1569        done = 1;        done = 1;
# Line 989  while (!done) Line 1572  while (!done)
1572      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1573      }      }
1574    
1575      /* The buffer may have moved while being extended; reset the start of data
1576      pointer to the correct relative point in the buffer. */
1577    
1578      p = buffer + poffset;
1579    
1580    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1581    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1582    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 1020  while (!done) Line 1608  while (!done)
1608    
1609        case '+': do_showrest = 1; break;        case '+': do_showrest = 1; break;
1610        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1611          case 'B': do_debug = 1; break;
1612        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1613        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1614        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
# Line 1027  while (!done) Line 1616  while (!done)
1616        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1617        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1618        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
1619          case 'K': do_mark = 1; break;
1620        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1621        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1622    
# Line 1036  while (!done) Line 1626  while (!done)
1626    
1627        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1628        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1629          case 'W': options |= PCRE_UCP; break;
1630        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1631          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1632          case 'Z': debug_lengths = 0; break;
1633        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1634        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1635    
1636          case 'T':
1637          switch (*pp++)
1638            {
1639            case '0': tables = tables0; break;
1640            case '1': tables = tables1; break;
1641    
1642            case '\r':
1643            case '\n':
1644            case ' ':
1645            case 0:
1646            fprintf(outfile, "** Missing table number after /T\n");
1647            goto SKIP_DATA;
1648    
1649            default:
1650            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1651            goto SKIP_DATA;
1652            }
1653          break;
1654    
1655        case 'L':        case 'L':
1656        ppp = pp;        ppp = pp;
1657        /* The '\r' test here is so that it works on Windows */        /* The '\r' test here is so that it works on Windows. */
1658        while (*ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;        /* The '0' test is just in case this is an unterminated line. */
1659          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1660        *ppp = 0;        *ppp = 0;
1661        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1662          {          {
1663          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1664          goto SKIP_DATA;          goto SKIP_DATA;
1665          }          }
1666          locale_set = 1;
1667        tables = pcre_maketables();        tables = pcre_maketables();
1668        pp = ppp;        pp = ppp;
1669        break;        break;
# Line 1063  while (!done) Line 1677  while (!done)
1677    
1678        case '<':        case '<':
1679          {          {
1680          int x = check_newline(pp, outfile);          if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1681          if (x == 0) goto SKIP_DATA;            {
1682          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1683          while (*pp++ != '>');            pp += 3;
1684              }
1685            else
1686              {
1687              int x = check_newline(pp, outfile);
1688              if (x == 0) goto SKIP_DATA;
1689              options |= x;
1690              while (*pp++ != '>');
1691              }
1692          }          }
1693        break;        break;
1694    
# Line 1096  while (!done) Line 1718  while (!done)
1718      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1719      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1720      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1721        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1722        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1723    
1724      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1725    
# Line 1116  while (!done) Line 1740  while (!done)
1740  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1741    
1742      {      {
1743      if (timeit)      unsigned long int get_options;
1744    
1745        if (timeit > 0)
1746        {        {
1747        register int i;        register int i;
1748        clock_t time_taken;        clock_t time_taken;
1749        clock_t start_time = clock();        clock_t start_time = clock();
1750        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1751          {          {
1752          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1753          if (re != NULL) free(re);          if (re != NULL) free(re);
1754          }          }
1755        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1756        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1757          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1758            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1759        }        }
1760    
# Line 1145  while (!done) Line 1771  while (!done)
1771          {          {
1772          for (;;)          for (;;)
1773            {            {
1774            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1775              {              {
1776              done = 1;              done = 1;
1777              goto CONTINUE;              goto CONTINUE;
# Line 1159  while (!done) Line 1785  while (!done)
1785        goto CONTINUE;        goto CONTINUE;
1786        }        }
1787    
1788      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1789      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
1790      returns only limited data. Check that it agrees with the newer one. */      lines. */
1791    
1792        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1793        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1794    
1795        /* Print information if required. There are now two info-returning
1796        functions. The old one has a limited interface and returns only limited
1797        data. Check that it agrees with the newer one. */
1798    
1799      if (log_store)      if (log_store)
1800        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1175  while (!done) Line 1808  while (!done)
1808      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1809      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1810    
1811      /* If /S was present, study the regexp to generate additional info to      /* If -s or /S was present, study the regexp to generate additional info to
1812      help with the matching. */      help with the matching. */
1813    
1814      if (do_study)      if (do_study || force_study)
1815        {        {
1816        if (timeit)        if (timeit > 0)
1817          {          {
1818          register int i;          register int i;
1819          clock_t time_taken;          clock_t time_taken;
1820          clock_t start_time = clock();          clock_t start_time = clock();
1821          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1822            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options, &error);
1823          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1824          if (extra != NULL) free(extra);          if (extra != NULL) free(extra);
1825          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1826            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1827              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1828          }          }
1829        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options, &error);
# Line 1200  while (!done) Line 1833  while (!done)
1833          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1834        }        }
1835    
1836        /* If /K was present, we set up for handling MARK data. */
1837    
1838        if (do_mark)
1839          {
1840          if (extra == NULL)
1841            {
1842            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1843            extra->flags = 0;
1844            }
1845          extra->mark = &markptr;
1846          extra->flags |= PCRE_EXTRA_MARK;
1847          }
1848    
1849      /* 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
1850      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
1851      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 1208  while (!done) Line 1854  while (!done)
1854      if (do_flip)      if (do_flip)
1855        {        {
1856        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1857        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1858            byteflip(rre->magic_number, sizeof(rre->magic_number));
1859        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1860        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1861        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1862        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1863        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1864        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1865        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1866          rre->first_byte =
1867            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1868          rre->req_byte =
1869            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1870          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1871          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1872        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1873          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1874        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1875            sizeof(rre->name_count));
1876    
1877        if (extra != NULL)        if (extra != NULL)
1878          {          {
1879          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1880          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1881          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1882            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1883          }          }
1884        }        }
1885    
# Line 1233  while (!done) Line 1887  while (!done)
1887    
1888      SHOW_INFO:      SHOW_INFO:
1889    
1890        if (do_debug)
1891          {
1892          fprintf(outfile, "------------------------------------------------------------------\n");
1893          pcre_printint(re, outfile, debug_lengths);
1894          }
1895    
1896        /* We already have the options in get_options (see above) */
1897    
1898      if (do_showinfo)      if (do_showinfo)
1899        {        {
1900        unsigned long int get_options, all_options;        unsigned long int all_options;
1901  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1902        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1903  #endif  #endif
1904        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1905            hascrorlf;
1906        int nameentrysize, namecount;        int nameentrysize, namecount;
1907        const uschar *nametable;        const uschar *nametable;
1908    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         pcre_printint(re, outfile);  
         }  
   
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1909        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1910        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1911        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1258  while (!done) Line 1914  while (!done)
1914        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1915        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1916        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1917          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1918          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1919          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1920    
1921  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1922        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1299  while (!done) Line 1958  while (!done)
1958            }            }
1959          }          }
1960    
1961        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1962        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1963    
1964        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1965        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");  
1966    
1967        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1968          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s\n",          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
1969            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1970            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1971            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1972            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1973            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1974            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1975              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1976              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1977            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1978            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1979            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1980            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1981            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1982              ((get_options & PCRE_UCP) != 0)? " ucp" : "",
1983            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1984              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
1985            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1986    
1987        switch (get_options & PCRE_NEWLINE_CRLF)        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1988    
1989          switch (get_options & PCRE_NEWLINE_BITS)
1990          {          {
1991          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
1992          fprintf(outfile, "Forced newline sequence: CR\n");          fprintf(outfile, "Forced newline sequence: CR\n");
# Line 1341  while (!done) Line 2000  while (!done)
2000          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
2001          break;          break;
2002    
2003            case PCRE_NEWLINE_ANYCRLF:
2004            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
2005            break;
2006    
2007            case PCRE_NEWLINE_ANY:
2008            fprintf(outfile, "Forced newline sequence: ANY\n");
2009            break;
2010    
2011          default:          default:
2012          break;          break;
2013          }          }
# Line 1358  while (!done) Line 2025  while (!done)
2025          int ch = first_char & 255;          int ch = first_char & 255;
2026          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
2027            "" : " (caseless)";            "" : " (caseless)";
2028          if (isprint(ch))          if (PRINTHEX(ch))
2029            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
2030          else          else
2031            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1373  while (!done) Line 2040  while (!done)
2040          int ch = need_char & 255;          int ch = need_char & 255;
2041          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
2042            "" : " (caseless)";            "" : " (caseless)";
2043          if (isprint(ch))          if (PRINTHEX(ch))
2044            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
2045          else          else
2046            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1384  while (!done) Line 2051  while (!done)
2051        so messes up the test suite. (And with the /F option, it might be        so messes up the test suite. (And with the /F option, it might be
2052        flipped.) */        flipped.) */
2053    
2054        if (do_study)        if (do_study || force_study)
2055          {          {
2056          if (extra == NULL)          if (extra == NULL)
2057            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2058          else          else
2059            {            {
2060            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2061            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2062    
2063              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2064              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2065    
2066              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2067            if (start_bits == NULL)            if (start_bits == NULL)
2068              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2069            else            else
2070              {              {
2071              int i;              int i;
# Line 1409  while (!done) Line 2080  while (!done)
2080                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
2081                    c = 2;                    c = 2;
2082                    }                    }
2083                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
2084                    {                    {
2085                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
2086                    c += 2;                    c += 2;
# Line 1441  while (!done) Line 2112  while (!done)
2112        else        else
2113          {          {
2114          uschar sbuf[8];          uschar sbuf[8];
2115          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2116          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2117          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2118          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2119    
2120          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2121          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2122          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2123          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2124    
2125          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2126              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1468  while (!done) Line 2139  while (!done)
2139                  strerror(errno));                  strerror(errno));
2140                }                }
2141              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
2142    
2143              }              }
2144            }            }
2145          fclose(f);          fclose(f);
# Line 1475  while (!done) Line 2147  while (!done)
2147    
2148        new_free(re);        new_free(re);
2149        if (extra != NULL) new_free(extra);        if (extra != NULL) new_free(extra);
2150        if (tables != NULL) new_free((void *)tables);        if (locale_set)
2151            {
2152            new_free((void *)tables);
2153            setlocale(LC_CTYPE, "C");
2154            locale_set = 0;
2155            }
2156        continue;  /* With next regex */        continue;  /* With next regex */
2157        }        }
2158      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1485  while (!done) Line 2162  while (!done)
2162    for (;;)    for (;;)
2163      {      {
2164      uschar *q;      uschar *q;
2165      uschar *bptr = dbuffer;      uschar *bptr;
2166      int *use_offsets = offsets;      int *use_offsets = offsets;
2167      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2168      int callout_data = 0;      int callout_data = 0;
2169      int callout_data_set = 0;      int callout_data_set = 0;
2170      int count, c;      int count, c;
2171      int copystrings = 0;      int copystrings = 0;
2172      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2173      int getstrings = 0;      int getstrings = 0;
2174      int getlist = 0;      int getlist = 0;
2175      int gmatched = 0;      int gmatched = 0;
2176      int start_offset = 0;      int start_offset = 0;
2177        int start_offset_sign = 1;
2178      int g_notempty = 0;      int g_notempty = 0;
2179      int use_dfa = 0;      int use_dfa = 0;
2180    
# Line 1522  while (!done) Line 2200  while (!done)
2200      len = 0;      len = 0;
2201      for (;;)      for (;;)
2202        {        {
2203        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
2204          {          {
2205          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
2206              {
2207              fprintf(outfile, "\n");
2208              break;
2209              }
2210          done = 1;          done = 1;
2211          goto CONTINUE;          goto CONTINUE;
2212          }          }
# Line 1541  while (!done) Line 2222  while (!done)
2222      p = buffer;      p = buffer;
2223      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2224    
2225      q = dbuffer;      bptr = q = dbuffer;
2226      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2227        {        {
2228        int i = 0;        int i = 0;
# Line 1591  while (!done) Line 2272  while (!done)
2272              {              {
2273              unsigned char buff8[8];              unsigned char buff8[8];
2274              int ii, utn;              int ii, utn;
2275              utn = ord2utf8(c, buff8);              if (use_utf8)
2276              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2277              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2278                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2279                  c = buff8[ii];   /* Last byte */
2280                  }
2281                else
2282                 {
2283                 if (c > 255)
2284                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2285                     "UTF-8 mode is not enabled.\n"
2286                     "** Truncation will probably give the wrong result.\n", c);
2287                 }
2288              p = pt + 1;              p = pt + 1;
2289              break;              break;
2290              }              }
# Line 1616  while (!done) Line 2307  while (!done)
2307          continue;          continue;
2308    
2309          case '>':          case '>':
2310            if (*p == '-')
2311              {
2312              start_offset_sign = -1;
2313              p++;
2314              }
2315          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2316            start_offset *= start_offset_sign;
2317          continue;          continue;
2318    
2319          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1689  while (!done) Line 2386  while (!done)
2386  #endif  #endif
2387            use_dfa = 1;            use_dfa = 1;
2388          continue;          continue;
2389    #endif
2390    
2391    #if !defined NODFA
2392          case 'F':          case 'F':
2393          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
2394          continue;          continue;
# Line 1723  while (!done) Line 2422  while (!done)
2422          continue;          continue;
2423    
2424          case 'N':          case 'N':
2425          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2426              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2427            else
2428              options |= PCRE_NOTEMPTY;
2429          continue;          continue;
2430    
2431          case 'O':          case 'O':
# Line 1736  while (!done) Line 2438  while (!done)
2438            if (offsets == NULL)            if (offsets == NULL)
2439              {              {
2440              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2441                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2442              yield = 1;              yield = 1;
2443              goto EXIT;              goto EXIT;
2444              }              }
# Line 1746  while (!done) Line 2448  while (!done)
2448          continue;          continue;
2449    
2450          case 'P':          case 'P':
2451          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2452              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2453          continue;          continue;
2454    
2455          case 'Q':          case 'Q':
# Line 1781  while (!done) Line 2484  while (!done)
2484          show_malloc = 1;          show_malloc = 1;
2485          continue;          continue;
2486    
2487            case 'Y':
2488            options |= PCRE_NO_START_OPTIMIZE;
2489            continue;
2490    
2491          case 'Z':          case 'Z':
2492          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2493          continue;          continue;
# Line 1801  while (!done) Line 2508  while (!done)
2508        *q++ = c;        *q++ = c;
2509        }        }
2510      *q = 0;      *q = 0;
2511      len = q - dbuffer;      len = (int)(q - dbuffer);
2512    
2513        /* Move the data to the end of the buffer so that a read over the end of
2514        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2515        we are using the POSIX interface, we must include the terminating zero. */
2516    
2517    #if !defined NOPOSIX
2518        if (posix || do_posix)
2519          {
2520          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2521          bptr += buffer_size - len - 1;
2522          }
2523        else
2524    #endif
2525          {
2526          memmove(bptr + buffer_size - len, bptr, len);
2527          bptr += buffer_size - len;
2528          }
2529    
2530      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2531        {        {
# Line 1822  while (!done) Line 2546  while (!done)
2546          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2547        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2548        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2549          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2550    
2551        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2552    
# Line 1866  while (!done) Line 2591  while (!done)
2591    
2592      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2593        {        {
2594        if (timeit)        markptr = NULL;
2595    
2596          if (timeitm > 0)
2597          {          {
2598          register int i;          register int i;
2599          clock_t time_taken;          clock_t time_taken;
# Line 1876  while (!done) Line 2603  while (!done)
2603          if (all_use_dfa || use_dfa)          if (all_use_dfa || use_dfa)
2604            {            {
2605            int workspace[1000];            int workspace[1000];
2606            for (i = 0; i < LOOPREPEAT; i++)            for (i = 0; i < timeitm; i++)
2607              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2608                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2609                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2610            }            }
2611          else          else
2612  #endif  #endif
2613    
2614          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeitm; i++)
2615            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2616              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2617    
2618          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2619          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2620            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2621              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2622          }          }
2623    
# Line 1940  while (!done) Line 2667  while (!done)
2667        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2668          {          {
2669          int workspace[1000];          int workspace[1000];
2670          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2671            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2672            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2673          if (count == 0)          if (count == 0)
# Line 1966  while (!done) Line 2693  while (!done)
2693    
2694        if (count >= 0)        if (count >= 0)
2695          {          {
2696          int i;          int i, maxcount;
2697    
2698    #if !defined NODFA
2699            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2700    #endif
2701              maxcount = use_size_offsets/3;
2702    
2703            /* This is a check against a lunatic return value. */
2704    
2705            if (count > maxcount)
2706              {
2707              fprintf(outfile,
2708                "** PCRE error: returned count %d is too big for offset size %d\n",
2709                count, use_size_offsets);
2710              count = use_size_offsets/3;
2711              if (do_g || do_G)
2712                {
2713                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2714                do_g = do_G = FALSE;        /* Break g/G loop */
2715                }
2716              }
2717    
2718          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2719            {            {
2720            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
# Line 1990  while (!done) Line 2738  while (!done)
2738              }              }
2739            }            }
2740    
2741            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2742    
2743          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2744            {            {
2745            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
# Line 2073  while (!done) Line 2823  while (!done)
2823    
2824        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2825          {          {
2826          fprintf(outfile, "Partial match");          if (markptr == NULL) fprintf(outfile, "Partial match");
2827  #if !defined NODFA            else fprintf(outfile, "Partial match, mark=%s", markptr);
2828          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2829            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2830              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2831  #endif            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2832                outfile);
2833              }
2834          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2835          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2836          }          }
2837    
2838        /* 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
2839        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
2840        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
2841        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2842        offset values to achieve this. We won't be at the end of the string -  
2843        that was checked before setting g_notempty. */        Complication arises in the case when the newline convention is "any",
2844          "crlf", or "anycrlf". If the previous match was at the end of a line
2845          terminated by CRLF, an advance of one character just passes the \r,
2846          whereas we should prefer the longer newline sequence, as does the code in
2847          pcre_exec(). Fudge the offset value to achieve this. We check for a
2848          newline setting in the pattern; if none was set, use pcre_config() to
2849          find the default.
2850    
2851          Otherwise, in the case of UTF-8 matching, the advance must be one
2852          character, not one byte. */
2853    
2854        else        else
2855          {          {
2856          if (g_notempty != 0)          if (g_notempty != 0)
2857            {            {
2858            int onechar = 1;            int onechar = 1;
2859              unsigned int obits = ((real_pcre *)re)->options;
2860            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2861            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2862                {
2863                int d;
2864                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2865                /* Note that these values are always the ASCII ones, even in
2866                EBCDIC environments. CR = 13, NL = 10. */
2867                obits = (d == 13)? PCRE_NEWLINE_CR :
2868                        (d == 10)? PCRE_NEWLINE_LF :
2869                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2870                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2871                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2872                }
2873              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2874                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2875                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2876                  &&
2877                  start_offset < len - 1 &&
2878                  bptr[start_offset] == '\r' &&
2879                  bptr[start_offset+1] == '\n')
2880                onechar++;
2881              else if (use_utf8)
2882              {              {
2883              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2884                {                {
2885                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
2886                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
2887                }                }
2888              }              }
2889            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
2890            }            }
2891          else          else
2892            {            {
2893            if (count == PCRE_ERROR_NOMATCH)            switch(count)
2894              {              {
2895              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
2896                if (gmatched == 0)
2897                  {
2898                  if (markptr == NULL) fprintf(outfile, "No match\n");
2899                    else fprintf(outfile, "No match, mark = %s\n", markptr);
2900                  }
2901                break;
2902    
2903                case PCRE_ERROR_BADUTF8:
2904                case PCRE_ERROR_SHORTUTF8:
2905                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
2906                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
2907                if (use_size_offsets >= 2)
2908                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
2909                    use_offsets[1]);
2910                fprintf(outfile, "\n");
2911                break;
2912    
2913                default:
2914                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
2915                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
2916                else
2917                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
2918                break;
2919              }              }
2920            else fprintf(outfile, "Error %d\n", count);  
2921            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
2922            }            }
2923          }          }
# Line 2124  while (!done) Line 2927  while (!done)
2927        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2928    
2929        /* 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
2930        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
2931        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
2932        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
2933        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
2934        character. */        character. */
2935    
2936        g_notempty = 0;        g_notempty = 0;
2937    
2938        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2939          {          {
2940          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2941          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2942          }          }
2943    
2944        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2161  while (!done) Line 2965  while (!done)
2965    
2966    if (re != NULL) new_free(re);    if (re != NULL) new_free(re);
2967    if (extra != NULL) new_free(extra);    if (extra != NULL) new_free(extra);
2968    if (tables != NULL)    if (locale_set)
2969      {      {
2970      new_free((void *)tables);      new_free((void *)tables);
2971      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
2972        locale_set = 0;
2973      }      }
2974    }    }
2975    

Legend:
Removed from v.91  
changed lines
  Added in v.606

  ViewVC Help
Powered by ViewVC 1.1.5