/[pcre]/code/branches/pcre16/pcretest.c
ViewVC logotype

Diff of /code/branches/pcre16/pcretest.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

code/trunk/pcretest.c revision 75 by nigel, Sat Feb 24 21:40:37 2007 UTC code/branches/pcre16/pcretest.c revision 774 by zherczeg, Thu Dec 1 06:08:45 2011 UTC
# Line 4  Line 4 
4    
5  /* This program was hacked up as a tester for PCRE. I really should have  /* This program was hacked up as a tester for PCRE. I really should have
6  written it more tidily in the first place. Will I ever learn? It has grown and  written it more tidily in the first place. Will I ever learn? It has grown and
7  been extended and consequently is now rather untidy in places.  been extended and consequently is now rather, er, *very* untidy in places.
8    
9  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
10  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# 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  /* We need the internal info for displaying the results of pcre_study(). Also  #ifdef SUPPORT_LIBREADLINE
52  for getting the opcodes for showing compiled code. */  #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57    #endif
58    
59    
60    /* A number of things vary for Windows builds. Originally, pcretest opened its
61    input and output without "b"; then I was told that "b" was needed in some
62    environments, so it was added for release 5.0 to both the input and output. (It
63    makes no difference on Unix-like systems.) Later I was told that it is wrong
64    for the input on Windows. I've now abstracted the modes into two macros that
65    are set here, to make it easier to fiddle with them, and removed "b" from the
66    input mode under Windows. */
67    
68    #if defined(_WIN32) || defined(WIN32)
69    #include <io.h>                /* For _setmode() */
70    #include <fcntl.h>             /* For _O_BINARY */
71    #define INPUT_MODE   "r"
72    #define OUTPUT_MODE  "wb"
73    
74    #ifndef isatty
75    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
76    #endif                         /* though in some environments they seem to   */
77                                   /* be already defined, hence the #ifndefs.    */
78    #ifndef fileno
79    #define fileno _fileno
80    #endif
81    
82    /* 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"
109    
110    /* We need access to some of the data tables that PCRE uses. So as not to have
111    to keep two copies, we include the source file here, changing the names of the
112    external symbols to prevent clashes. */
113    
114    #define _pcre_ucp_gentype      ucp_gentype
115    #define _pcre_ucp_typerange    ucp_typerange
116    #define _pcre_utf8_table1      utf8_table1
117    #define _pcre_utf8_table1_size utf8_table1_size
118    #define _pcre_utf8_table2      utf8_table2
119    #define _pcre_utf8_table3      utf8_table3
120    #define _pcre_utf8_table4      utf8_table4
121    #define _pcre_utt              utt
122    #define _pcre_utt_size         utt_size
123    #define _pcre_utt_names        utt_names
124    #define _pcre_OP_lengths       OP_lengths
125    
126    #include "pcre_tables.c"
127    
128    /* We also need the pcre_printint() function for printing out compiled
129    patterns. This function is in a separate file so that it can be included in
130    pcre_compile.c when that module is compiled with debugging enabled. It needs to
131    know which case is being compiled. */
132    
133    #define COMPILING_PCRETEST
134    #include "pcre_printint.src"
135    
136    /* The definition of the macro PRINTABLE, which determines whether to print an
137    output character as-is or as a hex value when showing compiled patterns, is
138    contained in the printint.src file. We uses it here also, in cases when the
139    locale has not been explicitly changed, so as to get consistent output from
140    systems that differ in their output from isprint() even in the "C" locale. */
141    
142  #define PCRE_SPY        /* For Win32 build, import data, not export */  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
 #include "internal.h"  
143    
144  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
145  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 58  Makefile. */ Line 149  Makefile. */
149  #include "pcreposix.h"  #include "pcreposix.h"
150  #endif  #endif
151    
152    /* It is also possible, for the benefit of the version currently imported into
153    Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
154    interface to the DFA matcher (NODFA), and without the doublecheck of the old
155    "info" function (define NOINFOCHECK). In fact, we automatically cut out the
156    UTF8 support if PCRE is built without it. */
157    
158    #ifndef SUPPORT_UTF8
159    #ifndef NOUTF8
160    #define NOUTF8
161    #endif
162    #endif
163    
164    
165    /* Other parameters */
166    
167  #ifndef CLOCKS_PER_SEC  #ifndef CLOCKS_PER_SEC
168  #ifdef CLK_TCK  #ifdef CLK_TCK
169  #define CLOCKS_PER_SEC CLK_TCK  #define CLOCKS_PER_SEC CLK_TCK
# Line 66  Makefile. */ Line 172  Makefile. */
172  #endif  #endif
173  #endif  #endif
174    
175  #define LOOPREPEAT 500000  /* This is the default loop count for timing. */
176    
177  #define BUFFER_SIZE 30000  #define LOOPREPEAT 500000
 #define PBUFFER_SIZE BUFFER_SIZE  
 #define DBUFFER_SIZE BUFFER_SIZE  
178    
179    /* Static variables */
180    
181  static FILE *outfile;  static FILE *outfile;
182  static int log_store = 0;  static int log_store = 0;
# Line 79  static int callout_count; Line 184  static int callout_count;
184  static int callout_extra;  static int callout_extra;
185  static int callout_fail_count;  static int callout_fail_count;
186  static int callout_fail_id;  static int callout_fail_id;
187    static int debug_lengths;
188  static int first_callout;  static int first_callout;
189    static int locale_set = 0;
190  static int show_malloc;  static int show_malloc;
191  static int use_utf8;  static int use_utf8;
192  static size_t gotten_store;  static size_t gotten_store;
193    static const unsigned char *last_callout_mark = NULL;
194    
195    /* The buffers grow automatically if very long input lines are encountered. */
196    
197  static uschar *pbuffer = NULL;  static int buffer_size = 50000;
198    static pcre_uint8 *buffer = NULL;
199    static pcre_uint8 *dbuffer = NULL;
200    static pcre_uint8 *pbuffer = NULL;
201    
202    /* Textual explanations for runtime error codes */
203    
204    static const char *errtexts[] = {
205      NULL,  /* 0 is no error */
206      NULL,  /* NOMATCH is handled specially */
207      "NULL argument passed",
208      "bad option value",
209      "magic number missing",
210      "unknown opcode - pattern overwritten?",
211      "no more memory",
212      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
213      "match limit exceeded",
214      "callout error code",
215      NULL,  /* BADUTF8 is handled specially */
216      "bad UTF-8 offset",
217      NULL,  /* PARTIAL is handled specially */
218      "not used - internal error",
219      "internal error - pattern overwritten?",
220      "bad count value",
221      "item unsupported for DFA matching",
222      "backreference condition or recursion test not supported for DFA matching",
223      "match limit not supported for DFA matching",
224      "workspace size exceeded in DFA matching",
225      "too much recursion for DFA matching",
226      "recursion limit exceeded",
227      "not used - internal error",
228      "invalid combination of newline options",
229      "bad offset value",
230      NULL,  /* SHORTUTF8 is handled specially */
231      "nested recursion at the same subject position",
232      "JIT stack limit reached"
233    };
234    
235    
236  static const int utf8_table1[] = {  /*************************************************
237    0x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff};  *         Alternate character tables             *
238    *************************************************/
239    
240    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
241    using the default tables of the library. However, the T option can be used to
242    select alternate sets of tables, for different kinds of testing. Note also that
243    the L (locale) option also adjusts the tables. */
244    
245    /* This is the set of tables distributed as default with PCRE. It recognizes
246    only ASCII characters. */
247    
248    static const unsigned char tables0[] = {
249    
250    /* This table is a lower casing table. */
251    
252        0,  1,  2,  3,  4,  5,  6,  7,
253        8,  9, 10, 11, 12, 13, 14, 15,
254       16, 17, 18, 19, 20, 21, 22, 23,
255       24, 25, 26, 27, 28, 29, 30, 31,
256       32, 33, 34, 35, 36, 37, 38, 39,
257       40, 41, 42, 43, 44, 45, 46, 47,
258       48, 49, 50, 51, 52, 53, 54, 55,
259       56, 57, 58, 59, 60, 61, 62, 63,
260       64, 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, 91, 92, 93, 94, 95,
264       96, 97, 98, 99,100,101,102,103,
265      104,105,106,107,108,109,110,111,
266      112,113,114,115,116,117,118,119,
267      120,121,122,123,124,125,126,127,
268      128,129,130,131,132,133,134,135,
269      136,137,138,139,140,141,142,143,
270      144,145,146,147,148,149,150,151,
271      152,153,154,155,156,157,158,159,
272      160,161,162,163,164,165,166,167,
273      168,169,170,171,172,173,174,175,
274      176,177,178,179,180,181,182,183,
275      184,185,186,187,188,189,190,191,
276      192,193,194,195,196,197,198,199,
277      200,201,202,203,204,205,206,207,
278      208,209,210,211,212,213,214,215,
279      216,217,218,219,220,221,222,223,
280      224,225,226,227,228,229,230,231,
281      232,233,234,235,236,237,238,239,
282      240,241,242,243,244,245,246,247,
283      248,249,250,251,252,253,254,255,
284    
285    /* This table is a case flipping table. */
286    
287        0,  1,  2,  3,  4,  5,  6,  7,
288        8,  9, 10, 11, 12, 13, 14, 15,
289       16, 17, 18, 19, 20, 21, 22, 23,
290       24, 25, 26, 27, 28, 29, 30, 31,
291       32, 33, 34, 35, 36, 37, 38, 39,
292       40, 41, 42, 43, 44, 45, 46, 47,
293       48, 49, 50, 51, 52, 53, 54, 55,
294       56, 57, 58, 59, 60, 61, 62, 63,
295       64, 97, 98, 99,100,101,102,103,
296      104,105,106,107,108,109,110,111,
297      112,113,114,115,116,117,118,119,
298      120,121,122, 91, 92, 93, 94, 95,
299       96, 65, 66, 67, 68, 69, 70, 71,
300       72, 73, 74, 75, 76, 77, 78, 79,
301       80, 81, 82, 83, 84, 85, 86, 87,
302       88, 89, 90,123,124,125,126,127,
303      128,129,130,131,132,133,134,135,
304      136,137,138,139,140,141,142,143,
305      144,145,146,147,148,149,150,151,
306      152,153,154,155,156,157,158,159,
307      160,161,162,163,164,165,166,167,
308      168,169,170,171,172,173,174,175,
309      176,177,178,179,180,181,182,183,
310      184,185,186,187,188,189,190,191,
311      192,193,194,195,196,197,198,199,
312      200,201,202,203,204,205,206,207,
313      208,209,210,211,212,213,214,215,
314      216,217,218,219,220,221,222,223,
315      224,225,226,227,228,229,230,231,
316      232,233,234,235,236,237,238,239,
317      240,241,242,243,244,245,246,247,
318      248,249,250,251,252,253,254,255,
319    
320    /* This table contains bit maps for various character classes. Each map is 32
321    bytes long and the bits run from the least significant end of each byte. The
322    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
323    graph, print, punct, and cntrl. Other classes are built from combinations. */
324    
325      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
326      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
327      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
328      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
329    
330      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
331      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
332      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
333      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
334    
335      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
336      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
337      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
338      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
339    
340      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
341      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
342      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
343      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
344    
345      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
346      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
347      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
348      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
349    
350      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
351      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
352      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
353      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
354    
355      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
356      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
357      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
358      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
359    
360      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
361      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
362      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
363      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
364    
365      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
366      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
367      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
368      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
369    
370      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
371      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
372      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
373      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
374    
375    /* This table identifies various classes of character by individual bits:
376      0x01   white space character
377      0x02   letter
378      0x04   decimal digit
379      0x08   hexadecimal digit
380      0x10   alphanumeric or '_'
381      0x80   regular expression metacharacter or binary zero
382    */
383    
384  static const int utf8_table2[] = {    0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
385    0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc};    0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
386      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
387      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
388      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
389      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
390      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
391      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
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      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
396      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
397      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
398      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
399      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
400      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
401      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
402      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
403      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
404      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
405      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
406      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
407      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
408      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
409      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
410      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
411      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
412      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
413      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
414      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
415      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
416    
417    /* This is a set of tables that came orginally from a Windows user. It seems to
418    be at least an approximation of ISO 8859. In particular, there are characters
419    greater than 128 that are marked as spaces, letters, etc. */
420    
421    static const unsigned char tables1[] = {
422    0,1,2,3,4,5,6,7,
423    8,9,10,11,12,13,14,15,
424    16,17,18,19,20,21,22,23,
425    24,25,26,27,28,29,30,31,
426    32,33,34,35,36,37,38,39,
427    40,41,42,43,44,45,46,47,
428    48,49,50,51,52,53,54,55,
429    56,57,58,59,60,61,62,63,
430    64,97,98,99,100,101,102,103,
431    104,105,106,107,108,109,110,111,
432    112,113,114,115,116,117,118,119,
433    120,121,122,91,92,93,94,95,
434    96,97,98,99,100,101,102,103,
435    104,105,106,107,108,109,110,111,
436    112,113,114,115,116,117,118,119,
437    120,121,122,123,124,125,126,127,
438    128,129,130,131,132,133,134,135,
439    136,137,138,139,140,141,142,143,
440    144,145,146,147,148,149,150,151,
441    152,153,154,155,156,157,158,159,
442    160,161,162,163,164,165,166,167,
443    168,169,170,171,172,173,174,175,
444    176,177,178,179,180,181,182,183,
445    184,185,186,187,188,189,190,191,
446    224,225,226,227,228,229,230,231,
447    232,233,234,235,236,237,238,239,
448    240,241,242,243,244,245,246,215,
449    248,249,250,251,252,253,254,223,
450    224,225,226,227,228,229,230,231,
451    232,233,234,235,236,237,238,239,
452    240,241,242,243,244,245,246,247,
453    248,249,250,251,252,253,254,255,
454    0,1,2,3,4,5,6,7,
455    8,9,10,11,12,13,14,15,
456    16,17,18,19,20,21,22,23,
457    24,25,26,27,28,29,30,31,
458    32,33,34,35,36,37,38,39,
459    40,41,42,43,44,45,46,47,
460    48,49,50,51,52,53,54,55,
461    56,57,58,59,60,61,62,63,
462    64,97,98,99,100,101,102,103,
463    104,105,106,107,108,109,110,111,
464    112,113,114,115,116,117,118,119,
465    120,121,122,91,92,93,94,95,
466    96,65,66,67,68,69,70,71,
467    72,73,74,75,76,77,78,79,
468    80,81,82,83,84,85,86,87,
469    88,89,90,123,124,125,126,127,
470    128,129,130,131,132,133,134,135,
471    136,137,138,139,140,141,142,143,
472    144,145,146,147,148,149,150,151,
473    152,153,154,155,156,157,158,159,
474    160,161,162,163,164,165,166,167,
475    168,169,170,171,172,173,174,175,
476    176,177,178,179,180,181,182,183,
477    184,185,186,187,188,189,190,191,
478    224,225,226,227,228,229,230,231,
479    232,233,234,235,236,237,238,239,
480    240,241,242,243,244,245,246,215,
481    248,249,250,251,252,253,254,223,
482    192,193,194,195,196,197,198,199,
483    200,201,202,203,204,205,206,207,
484    208,209,210,211,212,213,214,247,
485    216,217,218,219,220,221,222,255,
486    0,62,0,0,1,0,0,0,
487    0,0,0,0,0,0,0,0,
488    32,0,0,0,1,0,0,0,
489    0,0,0,0,0,0,0,0,
490    0,0,0,0,0,0,255,3,
491    126,0,0,0,126,0,0,0,
492    0,0,0,0,0,0,0,0,
493    0,0,0,0,0,0,0,0,
494    0,0,0,0,0,0,255,3,
495    0,0,0,0,0,0,0,0,
496    0,0,0,0,0,0,12,2,
497    0,0,0,0,0,0,0,0,
498    0,0,0,0,0,0,0,0,
499    254,255,255,7,0,0,0,0,
500    0,0,0,0,0,0,0,0,
501    255,255,127,127,0,0,0,0,
502    0,0,0,0,0,0,0,0,
503    0,0,0,0,254,255,255,7,
504    0,0,0,0,0,4,32,4,
505    0,0,0,128,255,255,127,255,
506    0,0,0,0,0,0,255,3,
507    254,255,255,135,254,255,255,7,
508    0,0,0,0,0,4,44,6,
509    255,255,127,255,255,255,127,255,
510    0,0,0,0,254,255,255,255,
511    255,255,255,255,255,255,255,127,
512    0,0,0,0,254,255,255,255,
513    255,255,255,255,255,255,255,255,
514    0,2,0,0,255,255,255,255,
515    255,255,255,255,255,255,255,127,
516    0,0,0,0,255,255,255,255,
517    255,255,255,255,255,255,255,255,
518    0,0,0,0,254,255,0,252,
519    1,0,0,248,1,0,0,120,
520    0,0,0,0,254,255,255,255,
521    0,0,128,0,0,0,128,0,
522    255,255,255,255,0,0,0,0,
523    0,0,0,0,0,0,0,128,
524    255,255,255,255,0,0,0,0,
525    0,0,0,0,0,0,0,0,
526    128,0,0,0,0,0,0,0,
527    0,1,1,0,1,1,0,0,
528    0,0,0,0,0,0,0,0,
529    0,0,0,0,0,0,0,0,
530    1,0,0,0,128,0,0,0,
531    128,128,128,128,0,0,128,0,
532    28,28,28,28,28,28,28,28,
533    28,28,0,0,0,0,0,128,
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    18,18,18,128,128,0,128,16,
538    0,26,26,26,26,26,26,18,
539    18,18,18,18,18,18,18,18,
540    18,18,18,18,18,18,18,18,
541    18,18,18,128,128,0,0,0,
542    0,0,0,0,0,1,0,0,
543    0,0,0,0,0,0,0,0,
544    0,0,0,0,0,0,0,0,
545    0,0,0,0,0,0,0,0,
546    1,0,0,0,0,0,0,0,
547    0,0,18,0,0,0,0,0,
548    0,0,20,20,0,18,0,0,
549    0,20,18,0,0,0,0,0,
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    18,18,18,18,18,18,18,18,
555    18,18,18,18,18,18,18,18,
556    18,18,18,18,18,18,18,0,
557    18,18,18,18,18,18,18,18
558    };
559    
 static const int utf8_table3[] = {  
   0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};  
560    
561    
562    
563    #ifndef HAVE_STRERROR
564  /*************************************************  /*************************************************
565  *         Print compiled regex                   *  *     Provide strerror() for non-ANSI libraries  *
566  *************************************************/  *************************************************/
567    
568  /* The code for doing this is held in a separate file that is also included in  /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
569  pcre.c when it is compiled with the debug switch. It defines a function called  in their libraries, but can provide the same facility by this simple
570  print_internals(), which uses a table of opcode lengths defined by the macro  alternative function. */
571  OP_LENGTHS, whose name must be OP_lengths. It also uses a table that translates  
572  Unicode property names to numbers; this is kept in a separate file. */  extern int   sys_nerr;
573    extern char *sys_errlist[];
574  static uschar OP_lengths[] = { OP_LENGTHS };  
575    char *
576  #include "ucp.h"  strerror(int n)
577  #include "ucptypetable.c"  {
578  #include "printint.c"  if (n < 0 || n >= sys_nerr) return "unknown error number";
579    return sys_errlist[n];
580    }
581    #endif /* HAVE_STRERROR */
582    
583    
584    /*************************************************
585    *         JIT memory callback                    *
586    *************************************************/
587    
588    static pcre_jit_stack* jit_callback(void *arg)
589    {
590    return (pcre_jit_stack *)arg;
591    }
592    
593    
594    /*************************************************
595    *        Read or extend an input line            *
596    *************************************************/
597    
598    /* Input lines are read into buffer, but both patterns and data lines can be
599    continued over multiple input lines. In addition, if the buffer fills up, we
600    want to automatically expand it so as to be able to handle extremely large
601    lines that are needed for certain stress tests. When the input buffer is
602    expanded, the other two buffers must also be expanded likewise, and the
603    contents of pbuffer, which are a copy of the input for callouts, must be
604    preserved (for when expansion happens for a data line). This is not the most
605    optimal way of handling this, but hey, this is just a test program!
606    
607    Arguments:
608      f            the file to read
609      start        where in buffer to start (this *must* be within buffer)
610      prompt       for stdin or readline()
611    
612    Returns:       pointer to the start of new data
613                   could be a copy of start, or could be moved
614                   NULL if no data read and EOF reached
615    */
616    
617    static pcre_uint8 *
618    extend_inputline(FILE *f, pcre_uint8 *start, const char *prompt)
619    {
620    pcre_uint8 *here = start;
621    
622    for (;;)
623      {
624      int rlen = (int)(buffer_size - (here - buffer));
625    
626      if (rlen > 1000)
627        {
628        int dlen;
629    
630        /* If libreadline support is required, use readline() to read a line if the
631        input is a terminal. Note that readline() removes the trailing newline, so
632        we must put it back again, to be compatible with fgets(). */
633    
634    #ifdef SUPPORT_LIBREADLINE
635        if (isatty(fileno(f)))
636          {
637          size_t len;
638          char *s = readline(prompt);
639          if (s == NULL) return (here == start)? NULL : start;
640          len = strlen(s);
641          if (len > 0) add_history(s);
642          if (len > rlen - 1) len = rlen - 1;
643          memcpy(here, s, len);
644          here[len] = '\n';
645          here[len+1] = 0;
646          free(s);
647          }
648        else
649    #endif
650    
651        /* Read the next line by normal means, prompting if the file is stdin. */
652    
653          {
654          if (f == stdin) printf("%s", prompt);
655          if (fgets((char *)here, rlen,  f) == NULL)
656            return (here == start)? NULL : start;
657          }
658    
659        dlen = (int)strlen((char *)here);
660        if (dlen > 0 && here[dlen - 1] == '\n') return start;
661        here += dlen;
662        }
663    
664      else
665        {
666        int new_buffer_size = 2*buffer_size;
667        pcre_uint8 *new_buffer = (unsigned char *)malloc(new_buffer_size);
668        pcre_uint8 *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
669        pcre_uint8 *new_pbuffer = (unsigned char *)malloc(new_buffer_size);
670    
671        if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
672          {
673          fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
674          exit(1);
675          }
676    
677        memcpy(new_buffer, buffer, buffer_size);
678        memcpy(new_pbuffer, pbuffer, buffer_size);
679    
680        buffer_size = new_buffer_size;
681    
682        start = new_buffer + (start - buffer);
683        here = new_buffer + (here - buffer);
684    
685        free(buffer);
686        free(dbuffer);
687        free(pbuffer);
688    
689        buffer = new_buffer;
690        dbuffer = new_dbuffer;
691        pbuffer = new_pbuffer;
692        }
693      }
694    
695    return NULL;  /* Control never gets here */
696    }
697    
698    
699    
700    
701    
702    
703    
# Line 122  static uschar OP_lengths[] = { OP_LENGTH Line 707  static uschar OP_lengths[] = { OP_LENGTH
707    
708  /* 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
709  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
710  for unpicking the -o argument, so just keep it simple.  for unpicking arguments, so just keep it simple.
711    
712  Arguments:  Arguments:
713    str           string to be converted    str           string to be converted
# Line 143  return(result); Line 728  return(result);
728    
729    
730    
 /*************************************************  
 *       Convert character value to UTF-8         *  
 *************************************************/  
   
 /* This function takes an integer value in the range 0 - 0x7fffffff  
 and encodes it as a UTF-8 character in 0 to 6 bytes.  
   
 Arguments:  
   cvalue     the character value  
   buffer     pointer to buffer for result - at least 6 bytes long  
   
 Returns:     number of characters placed in the buffer  
              -1 if input character is negative  
              0 if input character is positive but too big (only when  
              int is longer than 32 bits)  
 */  
   
 static int  
 ord2utf8(int cvalue, unsigned char *buffer)  
 {  
 register int i, j;  
 for (i = 0; i < sizeof(utf8_table1)/sizeof(int); i++)  
   if (cvalue <= utf8_table1[i]) break;  
 if (i >= sizeof(utf8_table1)/sizeof(int)) return 0;  
 if (cvalue < 0) return -1;  
   
 buffer += i;  
 for (j = i; j > 0; j--)  
  {  
  *buffer-- = 0x80 | (cvalue & 0x3f);  
  cvalue >>= 6;  
  }  
 *buffer = utf8_table2[i] | cvalue;  
 return i + 1;  
 }  
   
731    
732  /*************************************************  /*************************************************
733  *            Convert UTF-8 string to value       *  *            Convert UTF-8 string to value       *
# Line 188  return i + 1; Line 737  return i + 1;
737  and returns the value of the character.  and returns the value of the character.
738    
739  Argument:  Argument:
740    buffer   a pointer to the byte vector    utf8bytes   a pointer to the byte vector
741    vptr     a pointer to an int to receive the value    vptr        a pointer to an int to receive the value
742    
743  Returns:   >  0 => the number of bytes consumed  Returns:      >  0 => the number of bytes consumed
744             -6 to 0 => malformed UTF-8 character at offset = (-return)                -6 to 0 => malformed UTF-8 character at offset = (-return)
745  */  */
746    
747    #if !defined NOUTF8
748    
749  static int  static int
750  utf82ord(unsigned char *buffer, int *vptr)  utf82ord(unsigned char *utf8bytes, int *vptr)
751  {  {
752  int c = *buffer++;  int c = *utf8bytes++;
753  int d = c;  int d = c;
754  int i, j, s;  int i, j, s;
755    
# Line 218  d = (c & utf8_table3[i]) << s; Line 769  d = (c & utf8_table3[i]) << s;
769    
770  for (j = 0; j < i; j++)  for (j = 0; j < i; j++)
771    {    {
772    c = *buffer++;    c = *utf8bytes++;
773    if ((c & 0xc0) != 0x80) return -(j+1);    if ((c & 0xc0) != 0x80) return -(j+1);
774    s -= 6;    s -= 6;
775    d |= (c & 0x3f) << s;    d |= (c & 0x3f) << s;
# Line 226  for (j = 0; j < i; j++) Line 777  for (j = 0; j < i; j++)
777    
778  /* Check that encoding was the correct unique one */  /* Check that encoding was the correct unique one */
779    
780  for (j = 0; j < sizeof(utf8_table1)/sizeof(int); j++)  for (j = 0; j < utf8_table1_size; j++)
781    if (d <= utf8_table1[j]) break;    if (d <= utf8_table1[j]) break;
782  if (j != i) return -(i+1);  if (j != i) return -(i+1);
783    
# Line 236  if (j != i) return -(i+1); Line 787  if (j != i) return -(i+1);
787  return i+1;  return i+1;
788  }  }
789    
790    #endif
791    
792    
793    
794    /*************************************************
795    *       Convert character value to UTF-8         *
796    *************************************************/
797    
798    /* This function takes an integer value in the range 0 - 0x7fffffff
799    and encodes it as a UTF-8 character in 0 to 6 bytes.
800    
801    Arguments:
802      cvalue     the character value
803      utf8bytes  pointer to buffer for result - at least 6 bytes long
804    
805    Returns:     number of characters placed in the buffer
806    */
807    
808    #if !defined NOUTF8
809    
810    static int
811    ord2utf8(int cvalue, pcre_uint8 *utf8bytes)
812    {
813    register int i, j;
814    for (i = 0; i < utf8_table1_size; i++)
815      if (cvalue <= utf8_table1[i]) break;
816    utf8bytes += i;
817    for (j = i; j > 0; j--)
818     {
819     *utf8bytes-- = 0x80 | (cvalue & 0x3f);
820     cvalue >>= 6;
821     }
822    *utf8bytes = utf8_table2[i] | cvalue;
823    return i + 1;
824    }
825    
826    #endif
827    
828    
829    
830  /*************************************************  /*************************************************
# Line 248  chars without printing. */ Line 837  chars without printing. */
837    
838  static int pchars(unsigned char *p, int length, FILE *f)  static int pchars(unsigned char *p, int length, FILE *f)
839  {  {
840  int c;  int c = 0;
841  int yield = 0;  int yield = 0;
842    
843  while (length-- > 0)  while (length-- > 0)
844    {    {
845    #if !defined NOUTF8
846    if (use_utf8)    if (use_utf8)
847      {      {
848      int rc = utf82ord(p, &c);      int rc = utf82ord(p, &c);
# Line 261  while (length-- > 0) Line 851  while (length-- > 0)
851        {        {
852        length -= rc - 1;        length -= rc - 1;
853        p += rc;        p += rc;
854        if (c < 256 && isprint(c))        if (PRINTHEX(c))
855          {          {
856          if (f != NULL) fprintf(f, "%c", c);          if (f != NULL) fprintf(f, "%c", c);
857          yield++;          yield++;
858          }          }
859        else        else
860          {          {
861          int n;          int n = 4;
862          if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);          if (f != NULL) fprintf(f, "\\x{%02x}", c);
863          yield += n;          yield += (n <= 0x000000ff)? 2 :
864                     (n <= 0x00000fff)? 3 :
865                     (n <= 0x0000ffff)? 4 :
866                     (n <= 0x000fffff)? 5 : 6;
867          }          }
868        continue;        continue;
869        }        }
870      }      }
871    #endif
872    
873     /* Not UTF-8, or malformed UTF-8  */     /* Not UTF-8, or malformed UTF-8  */
874    
875    if (isprint(c = *(p++)))    c = *p++;
876      if (PRINTHEX(c))
877      {      {
878      if (f != NULL) fprintf(f, "%c", c);      if (f != NULL) fprintf(f, "%c", c);
879      yield++;      yield++;
# Line 376  fprintf(outfile, "%.*s", (cb->next_item_ Line 971  fprintf(outfile, "%.*s", (cb->next_item_
971  fprintf(outfile, "\n");  fprintf(outfile, "\n");
972  first_callout = 0;  first_callout = 0;
973    
974    if (cb->mark != last_callout_mark)
975      {
976      fprintf(outfile, "Latest Mark: %s\n",
977        (cb->mark == NULL)? "<unset>" : (char *)(cb->mark));
978      last_callout_mark = cb->mark;
979      }
980    
981  if (cb->callout_data != NULL)  if (cb->callout_data != NULL)
982    {    {
983    int callout_data = *((int *)(cb->callout_data));    int callout_data = *((int *)(cb->callout_data));
# Line 395  return (cb->callout_number != callout_fa Line 997  return (cb->callout_number != callout_fa
997  *            Local malloc functions              *  *            Local malloc functions              *
998  *************************************************/  *************************************************/
999    
1000  /* Alternative malloc function, to test functionality and show the size of the  /* Alternative malloc function, to test functionality and save the size of a
1001  compiled re. */  compiled re. The show_malloc variable is set only during matching. */
1002    
1003  static void *new_malloc(size_t size)  static void *new_malloc(size_t size)
1004  {  {
1005  void *block = malloc(size);  void *block = malloc(size);
1006  gotten_store = size;  gotten_store = size;
1007  if (show_malloc)  if (show_malloc)
1008    fprintf(outfile, "malloc       %3d %p\n", size, block);    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);
1009  return block;  return block;
1010  }  }
1011    
# Line 414  if (show_malloc) Line 1016  if (show_malloc)
1016  free(block);  free(block);
1017  }  }
1018    
   
1019  /* For recursion malloc/free, to test stacking calls */  /* For recursion malloc/free, to test stacking calls */
1020    
1021  static void *stack_malloc(size_t size)  static void *stack_malloc(size_t size)
1022  {  {
1023  void *block = malloc(size);  void *block = malloc(size);
1024  if (show_malloc)  if (show_malloc)
1025    fprintf(outfile, "stack_malloc %3d %p\n", size, block);    fprintf(outfile, "stack_malloc %3d %p\n", (int)size, block);
1026  return block;  return block;
1027  }  }
1028    
# Line 452  if ((rc = pcre_fullinfo(re, study, optio Line 1053  if ((rc = pcre_fullinfo(re, study, optio
1053  *         Byte flipping function                 *  *         Byte flipping function                 *
1054  *************************************************/  *************************************************/
1055    
1056  static long int  static unsigned long int
1057  byteflip(long int value, int n)  byteflip(unsigned long int value, int n)
1058  {  {
1059  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
1060  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
# Line 466  return ((value & 0x000000ff) << 24) | Line 1067  return ((value & 0x000000ff) << 24) |
1067    
1068    
1069  /*************************************************  /*************************************************
1070    *        Check match or recursion limit          *
1071    *************************************************/
1072    
1073    static int
1074    check_match_limit(pcre *re, pcre_extra *extra, pcre_uint8 *bptr, int len,
1075      int start_offset, int options, int *use_offsets, int use_size_offsets,
1076      int flag, unsigned long int *limit, int errnumber, const char *msg)
1077    {
1078    int count;
1079    int min = 0;
1080    int mid = 64;
1081    int max = -1;
1082    
1083    extra->flags |= flag;
1084    
1085    for (;;)
1086      {
1087      *limit = mid;
1088    
1089      count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,
1090        use_offsets, use_size_offsets);
1091    
1092      if (count == errnumber)
1093        {
1094        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
1095        min = mid;
1096        mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
1097        }
1098    
1099      else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
1100                             count == PCRE_ERROR_PARTIAL)
1101        {
1102        if (mid == min + 1)
1103          {
1104          fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
1105          break;
1106          }
1107        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
1108        max = mid;
1109        mid = (min + mid)/2;
1110        }
1111      else break;    /* Some other error */
1112      }
1113    
1114    extra->flags &= ~flag;
1115    return count;
1116    }
1117    
1118    
1119    
1120    /*************************************************
1121    *         Case-independent strncmp() function    *
1122    *************************************************/
1123    
1124    /*
1125    Arguments:
1126      s         first string
1127      t         second string
1128      n         number of characters to compare
1129    
1130    Returns:    < 0, = 0, or > 0, according to the comparison
1131    */
1132    
1133    static int
1134    strncmpic(pcre_uint8 *s, pcre_uint8 *t, int n)
1135    {
1136    while (n--)
1137      {
1138      int c = tolower(*s++) - tolower(*t++);
1139      if (c) return c;
1140      }
1141    return 0;
1142    }
1143    
1144    
1145    
1146    /*************************************************
1147    *         Check newline indicator                *
1148    *************************************************/
1149    
1150    /* This is used both at compile and run-time to check for <xxx> escapes. Print
1151    a message and return 0 if there is no match.
1152    
1153    Arguments:
1154      p           points after the leading '<'
1155      f           file for error message
1156    
1157    Returns:      appropriate PCRE_NEWLINE_xxx flags, or 0
1158    */
1159    
1160    static int
1161    check_newline(pcre_uint8 *p, FILE *f)
1162    {
1163    if (strncmpic(p, (pcre_uint8 *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1164    if (strncmpic(p, (pcre_uint8 *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1165    if (strncmpic(p, (pcre_uint8 *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1166    if (strncmpic(p, (pcre_uint8 *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1167    if (strncmpic(p, (pcre_uint8 *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1168    if (strncmpic(p, (pcre_uint8 *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1169    if (strncmpic(p, (pcre_uint8 *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1170    fprintf(f, "Unknown newline type at: <%s\n", p);
1171    return 0;
1172    }
1173    
1174    
1175    
1176    /*************************************************
1177    *             Usage function                     *
1178    *************************************************/
1179    
1180    static void
1181    usage(void)
1182    {
1183    printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1184    printf("Input and output default to stdin and stdout.\n");
1185    #ifdef SUPPORT_LIBREADLINE
1186    printf("If input is a terminal, readline() is used to read from it.\n");
1187    #else
1188    printf("This version of pcretest is not linked with readline().\n");
1189    #endif
1190    printf("\nOptions:\n");
1191    printf("  -b       show compiled code (bytecode)\n");
1192    printf("  -C       show PCRE compile-time options and exit\n");
1193    printf("  -d       debug: show compiled code and information (-b and -i)\n");
1194    #if !defined NODFA
1195    printf("  -dfa     force DFA matching for all subjects\n");
1196    #endif
1197    printf("  -help    show usage information\n");
1198    printf("  -i       show information about compiled patterns\n"
1199           "  -M       find MATCH_LIMIT minimum for each subject\n"
1200           "  -m       output memory used information\n"
1201           "  -o <n>   set size of offsets vector to <n>\n");
1202    #if !defined NOPOSIX
1203    printf("  -p       use POSIX interface\n");
1204    #endif
1205    printf("  -q       quiet: do not output PCRE version number at start\n");
1206    printf("  -S <n>   set stack size to <n> megabytes\n");
1207    printf("  -s       force each pattern to be studied at basic level\n"
1208           "  -s+      force each pattern to be studied, using JIT if available\n"
1209           "  -t       time compilation and execution\n");
1210    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1211    printf("  -tm      time execution (matching) only\n");
1212    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
1213    }
1214    
1215    
1216    
1217    /*************************************************
1218  *                Main Program                    *  *                Main Program                    *
1219  *************************************************/  *************************************************/
1220    
# Line 478  int main(int argc, char **argv) Line 1227  int main(int argc, char **argv)
1227  FILE *infile = stdin;  FILE *infile = stdin;
1228  int options = 0;  int options = 0;
1229  int study_options = 0;  int study_options = 0;
1230    int default_find_match_limit = FALSE;
1231  int op = 1;  int op = 1;
1232  int timeit = 0;  int timeit = 0;
1233    int timeitm = 0;
1234  int showinfo = 0;  int showinfo = 0;
1235  int showstore = 0;  int showstore = 0;
1236    int force_study = -1;
1237    int force_study_options = 0;
1238    int quiet = 0;
1239  int size_offsets = 45;  int size_offsets = 45;
1240  int size_offsets_max;  int size_offsets_max;
1241  int *offsets;  int *offsets = NULL;
1242  #if !defined NOPOSIX  #if !defined NOPOSIX
1243  int posix = 0;  int posix = 0;
1244  #endif  #endif
1245  int debug = 0;  int debug = 0;
1246  int done = 0;  int done = 0;
1247    int all_use_dfa = 0;
1248    int yield = 0;
1249    int stack_size;
1250    
1251    pcre_jit_stack *jit_stack = NULL;
1252    
1253    
1254  unsigned char *buffer;  /* These vectors store, end-to-end, a list of captured substring names. Assume
1255  unsigned char *dbuffer;  that 1024 is plenty long enough for the few names we'll be testing. */
1256    
1257    pcre_uchar copynames[1024];
1258    pcre_uchar getnames[1024];
1259    
1260    pcre_uchar *copynamesptr;
1261    pcre_uchar *getnamesptr;
1262    
1263  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* Get buffers from malloc() so that Electric Fence will check their misuse
1264  when I am debugging. */  when I am debugging. They grow automatically when very long lines are read. */
1265    
1266  buffer = (unsigned char *)malloc(BUFFER_SIZE);  buffer = (pcre_uint8 *)malloc(buffer_size);
1267  dbuffer = (unsigned char *)malloc(DBUFFER_SIZE);  dbuffer = (pcre_uint8 *)malloc(buffer_size);
1268  pbuffer = (unsigned char *)malloc(PBUFFER_SIZE);  pbuffer = (pcre_uint8 *)malloc(buffer_size);
   
 /* The outfile variable is static so that new_malloc can use it. The _setmode()  
 stuff is some magic that I don't understand, but which apparently does good  
 things in Windows. It's related to line terminations.  */  
1269    
1270  #if defined(_WIN32) || defined(WIN32)  /* The outfile variable is static so that new_malloc can use it. */
 _setmode( _fileno( stdout ), 0x8000 );  
 #endif  /* defined(_WIN32) || defined(WIN32) */  
1271    
1272  outfile = stdout;  outfile = stdout;
1273    
1274    /* The following  _setmode() stuff is some Windows magic that tells its runtime
1275    library to translate CRLF into a single LF character. At least, that's what
1276    I've been told: never having used Windows I take this all on trust. Originally
1277    it set 0x8000, but then I was advised that _O_BINARY was better. */
1278    
1279    #if defined(_WIN32) || defined(WIN32)
1280    _setmode( _fileno( stdout ), _O_BINARY );
1281    #endif
1282    
1283  /* Scan options */  /* Scan options */
1284    
1285  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
1286    {    {
1287    unsigned char *endptr;    unsigned char *endptr;
1288    
1289    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1290      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 0;
1291    else if (strcmp(argv[op], "-t") == 0) timeit = 1;    else if (strcmp(argv[op], "-s+") == 0)
1292        {
1293        force_study = 1;
1294        force_study_options = PCRE_STUDY_JIT_COMPILE;
1295        }
1296      else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1297      else if (strcmp(argv[op], "-b") == 0) debug = 1;
1298    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1299    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1300      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1301    #if !defined NODFA
1302      else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1303    #endif
1304    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
1305        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),
1306          *endptr == 0))          *endptr == 0))
# Line 529  while (argc > 1 && argv[op][0] == '-') Line 1308  while (argc > 1 && argv[op][0] == '-')
1308      op++;      op++;
1309      argc--;      argc--;
1310      }      }
1311      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
1312        {
1313        int both = argv[op][2] == 0;
1314        int temp;
1315        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
1316                         *endptr == 0))
1317          {
1318          timeitm = temp;
1319          op++;
1320          argc--;
1321          }
1322        else timeitm = LOOPREPEAT;
1323        if (both) timeit = timeitm;
1324        }
1325      else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
1326          ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1327            *endptr == 0))
1328        {
1329    #if defined(_WIN32) || defined(WIN32) || defined(__minix)
1330        printf("PCRE: -S not supported on this OS\n");
1331        exit(1);
1332    #else
1333        int rc;
1334        struct rlimit rlim;
1335        getrlimit(RLIMIT_STACK, &rlim);
1336        rlim.rlim_cur = stack_size * 1024 * 1024;
1337        rc = setrlimit(RLIMIT_STACK, &rlim);
1338        if (rc != 0)
1339          {
1340        printf("PCRE: setrlimit() failed with error %d\n", rc);
1341        exit(1);
1342          }
1343        op++;
1344        argc--;
1345    #endif
1346        }
1347  #if !defined NOPOSIX  #if !defined NOPOSIX
1348    else if (strcmp(argv[op], "-p") == 0) posix = 1;    else if (strcmp(argv[op], "-p") == 0) posix = 1;
1349  #endif  #endif
1350    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1351      {      {
1352      int rc;      int rc;
1353        unsigned long int lrc;
1354      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1355      printf("Compiled with\n");      printf("Compiled with\n");
1356      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
1357      printf("  %sUTF-8 support\n", rc? "" : "No ");      printf("  %sUTF-8 support\n", rc? "" : "No ");
1358      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1359      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1360        (void)pcre_config(PCRE_CONFIG_JIT, &rc);
1361        if (rc)
1362          printf("  Just-in-time compiler support\n");
1363        else
1364          printf("  No just-in-time compiler support\n");
1365      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1366      printf("  Newline character is %s\n", (rc == '\r')? "CR" : "LF");      /* Note that these values are always the ASCII values, even
1367        in EBCDIC environments. CR is 13 and NL is 10. */
1368        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1369          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1370          (rc == -2)? "ANYCRLF" :
1371          (rc == -1)? "ANY" : "???");
1372        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1373        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1374                                         "all Unicode newlines");
1375      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1376      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1377      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1378      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1379      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1380      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1381        (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1382        printf("  Default recursion depth limit = %ld\n", lrc);
1383      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1384      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1385      exit(0);      goto EXIT;
1386        }
1387      else if (strcmp(argv[op], "-help") == 0 ||
1388               strcmp(argv[op], "--help") == 0)
1389        {
1390        usage();
1391        goto EXIT;
1392      }      }
1393    else    else
1394      {      {
1395      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
1396      printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");      usage();
1397      printf("  -C     show PCRE compile-time options and exit\n");      yield = 1;
1398      printf("  -d     debug: show compiled code; implies -i\n"      goto EXIT;
            "  -i     show information about compiled pattern\n"  
            "  -m     output memory used information\n"  
            "  -o <n> set size of offsets vector to <n>\n");  
 #if !defined NOPOSIX  
     printf("  -p     use POSIX interface\n");  
 #endif  
     printf("  -s     output store (memory) used information\n"  
            "  -t     time compilation and execution\n");  
     return 1;  
1399      }      }
1400    op++;    op++;
1401    argc--;    argc--;
# Line 580  offsets = (int *)malloc(size_offsets_max Line 1408  offsets = (int *)malloc(size_offsets_max
1408  if (offsets == NULL)  if (offsets == NULL)
1409    {    {
1410    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1411      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1412    return 1;    yield = 1;
1413      goto EXIT;
1414    }    }
1415    
1416  /* Sort out the input and output files */  /* Sort out the input and output files */
1417    
1418  if (argc > 1)  if (argc > 1)
1419    {    {
1420    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
1421    if (infile == NULL)    if (infile == NULL)
1422      {      {
1423      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
1424      return 1;      yield = 1;
1425        goto EXIT;
1426      }      }
1427    }    }
1428    
1429  if (argc > 2)  if (argc > 2)
1430    {    {
1431    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
1432    if (outfile == NULL)    if (outfile == NULL)
1433      {      {
1434      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
1435      return 1;      yield = 1;
1436        goto EXIT;
1437      }      }
1438    }    }
1439    
# Line 613  pcre_free = new_free; Line 1444  pcre_free = new_free;
1444  pcre_stack_malloc = stack_malloc;  pcre_stack_malloc = stack_malloc;
1445  pcre_stack_free = stack_free;  pcre_stack_free = stack_free;
1446    
1447  /* Heading line, then prompt for first regex if stdin */  /* Heading line unless quiet, then prompt for first regex if stdin */
1448    
1449  fprintf(outfile, "PCRE version %s\n\n", pcre_version());  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
1450    
1451  /* Main loop */  /* Main loop */
1452    
# Line 630  while (!done) Line 1461  while (!done)
1461  #endif  #endif
1462    
1463    const char *error;    const char *error;
1464      unsigned char *markptr;
1465    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1466    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1467    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1468    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1469    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1470      int do_allcaps = 0;
1471      int do_mark = 0;
1472    int do_study = 0;    int do_study = 0;
1473      int no_force_study = 0;
1474    int do_debug = debug;    int do_debug = debug;
1475    int do_G = 0;    int do_G = 0;
1476    int do_g = 0;    int do_g = 0;
1477    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1478    int do_showrest = 0;    int do_showrest = 0;
1479      int do_showcaprest = 0;
1480    int do_flip = 0;    int do_flip = 0;
1481    int erroroffset, len, delimiter;    int erroroffset, len, delimiter, poffset;
1482    
1483    use_utf8 = 0;    use_utf8 = 0;
1484      debug_lengths = 1;
1485    
1486    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) break;  
1487    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1488    fflush(outfile);    fflush(outfile);
1489    
# Line 659  while (!done) Line 1495  while (!done)
1495    
1496    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
1497      {      {
1498      unsigned long int magic;      unsigned long int magic, get_options;
1499      uschar sbuf[8];      pcre_uint8 sbuf[8];
1500      FILE *f;      FILE *f;
1501    
1502      p++;      p++;
# Line 702  while (!done) Line 1538  while (!done)
1538          }          }
1539        }        }
1540    
1541      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1542        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1543    
1544      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
1545    
1546      new_info(re, NULL, PCRE_INFO_OPTIONS, &options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1547      use_utf8 = (options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1548    
1549      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1550    
1551      if (true_study_size != 0)      if (true_study_size != 0)
1552        {        {
# Line 726  while (!done) Line 1562  while (!done)
1562          {          {
1563          FAIL_READ:          FAIL_READ:
1564          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
1565          if (extra != NULL) new_free(extra);          if (extra != NULL) pcre_free_study(extra);
1566          if (re != NULL) new_free(re);          if (re != NULL) new_free(re);
1567          fclose(f);          fclose(f);
1568          continue;          continue;
# Line 747  while (!done) Line 1583  while (!done)
1583    
1584    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1585      {      {
1586      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1587      goto SKIP_DATA;      goto SKIP_DATA;
1588      }      }
1589    
1590    pp = p;    pp = p;
1591      poffset = (int)(p - buffer);
1592    
1593    for(;;)    for(;;)
1594      {      {
# Line 762  while (!done) Line 1599  while (!done)
1599        pp++;        pp++;
1600        }        }
1601      if (*pp != 0) break;      if (*pp != 0) break;
1602        if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     len = BUFFER_SIZE - (pp - buffer);  
     if (len < 256)  
       {  
       fprintf(outfile, "** Expression too long - missing delimiter?\n");  
       goto SKIP_DATA;  
       }  
   
     if (infile == stdin) printf("    > ");  
     if (fgets((char *)pp, len, infile) == NULL)  
1603        {        {
1604        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1605        done = 1;        done = 1;
# Line 780  while (!done) Line 1608  while (!done)
1608      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1609      }      }
1610    
1611      /* The buffer may have moved while being extended; reset the start of data
1612      pointer to the correct relative point in the buffer. */
1613    
1614      p = buffer + poffset;
1615    
1616    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1617    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1618    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 795  while (!done) Line 1628  while (!done)
1628    /* Look for options after final delimiter */    /* Look for options after final delimiter */
1629    
1630    options = 0;    options = 0;
   study_options = 0;  
1631    log_store = showstore;  /* default from command line */    log_store = showstore;  /* default from command line */
1632    
1633    while (*pp != 0)    while (*pp != 0)
1634      {      {
1635      switch (*pp++)      switch (*pp++)
1636        {        {
1637          case 'f': options |= PCRE_FIRSTLINE; break;
1638        case 'g': do_g = 1; break;        case 'g': do_g = 1; break;
1639        case 'i': options |= PCRE_CASELESS; break;        case 'i': options |= PCRE_CASELESS; break;
1640        case 'm': options |= PCRE_MULTILINE; break;        case 'm': options |= PCRE_MULTILINE; break;
1641        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1642        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1643    
1644        case '+': do_showrest = 1; break;        case '+':
1645          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1646          break;
1647    
1648          case '=': do_allcaps = 1; break;
1649        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1650          case 'B': do_debug = 1; break;
1651        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1652        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1653        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
1654        case 'F': do_flip = 1; break;        case 'F': do_flip = 1; break;
1655        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1656        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1657          case 'J': options |= PCRE_DUPNAMES; break;
1658          case 'K': do_mark = 1; break;
1659        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1660        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1661    
# Line 823  while (!done) Line 1663  while (!done)
1663        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1664  #endif  #endif
1665    
1666        case 'S': do_study = 1; break;        case 'S':
1667          if (do_study == 0)
1668            {
1669            do_study = 1;
1670            if (*pp == '+')
1671              {
1672              study_options |= PCRE_STUDY_JIT_COMPILE;
1673              pp++;
1674              }
1675            }
1676          else
1677            {
1678            do_study = 0;
1679            no_force_study = 1;
1680            }
1681          break;
1682    
1683        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1684          case 'W': options |= PCRE_UCP; break;
1685        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1686          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1687          case 'Z': debug_lengths = 0; break;
1688        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1689        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1690    
1691          case 'T':
1692          switch (*pp++)
1693            {
1694            case '0': tables = tables0; break;
1695            case '1': tables = tables1; break;
1696    
1697            case '\r':
1698            case '\n':
1699            case ' ':
1700            case 0:
1701            fprintf(outfile, "** Missing table number after /T\n");
1702            goto SKIP_DATA;
1703    
1704            default:
1705            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1706            goto SKIP_DATA;
1707            }
1708          break;
1709    
1710        case 'L':        case 'L':
1711        ppp = pp;        ppp = pp;
1712        while (*ppp != '\n' && *ppp != ' ') ppp++;        /* The '\r' test here is so that it works on Windows. */
1713          /* The '0' test is just in case this is an unterminated line. */
1714          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1715        *ppp = 0;        *ppp = 0;
1716        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1717          {          {
1718          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1719          goto SKIP_DATA;          goto SKIP_DATA;
1720          }          }
1721          locale_set = 1;
1722        tables = pcre_maketables();        tables = pcre_maketables();
1723        pp = ppp;        pp = ppp;
1724        break;        break;
# Line 849  while (!done) Line 1730  while (!done)
1730        *pp = 0;        *pp = 0;
1731        break;        break;
1732    
1733        case '\n': case ' ': break;        case '<':
1734            {
1735            if (strncmpic(pp, (pcre_uint8 *)"JS>", 3) == 0)
1736              {
1737              options |= PCRE_JAVASCRIPT_COMPAT;
1738              pp += 3;
1739              }
1740            else
1741              {
1742              int x = check_newline(pp, outfile);
1743              if (x == 0) goto SKIP_DATA;
1744              options |= x;
1745              while (*pp++ != '>');
1746              }
1747            }
1748          break;
1749    
1750          case '\r':                      /* So that it works in Windows */
1751          case '\n':
1752          case ' ':
1753          break;
1754    
1755        default:        default:
1756        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);
# Line 869  while (!done) Line 1770  while (!done)
1770    
1771      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
1772      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
1773        if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1774        if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1775        if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1776        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1777        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1778    
1779      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1780    
1781      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
# Line 876  while (!done) Line 1783  while (!done)
1783    
1784      if (rc != 0)      if (rc != 0)
1785        {        {
1786        (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);        (void)regerror(rc, &preg, (char *)buffer, buffer_size);
1787        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
1788        goto SKIP_DATA;        goto SKIP_DATA;
1789        }        }
# Line 888  while (!done) Line 1795  while (!done)
1795  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1796    
1797      {      {
1798      if (timeit)      unsigned long int get_options;
1799    
1800        if (timeit > 0)
1801        {        {
1802        register int i;        register int i;
1803        clock_t time_taken;        clock_t time_taken;
1804        clock_t start_time = clock();        clock_t start_time = clock();
1805        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1806          {          {
1807          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1808          if (re != NULL) free(re);          if (re != NULL) free(re);
1809          }          }
1810        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1811        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1812          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1813            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1814        }        }
1815    
# Line 917  while (!done) Line 1826  while (!done)
1826          {          {
1827          for (;;)          for (;;)
1828            {            {
1829            if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1830              {              {
1831              done = 1;              done = 1;
1832              goto CONTINUE;              goto CONTINUE;
# Line 931  while (!done) Line 1840  while (!done)
1840        goto CONTINUE;        goto CONTINUE;
1841        }        }
1842    
1843      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1844      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
1845      returns only limited data. Check that it agrees with the newer one. */      lines. */
1846    
1847        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1848        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1849    
1850        /* Print information if required. There are now two info-returning
1851        functions. The old one has a limited interface and returns only limited
1852        data. Check that it agrees with the newer one. */
1853    
1854      if (log_store)      if (log_store)
1855        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 947  while (!done) Line 1863  while (!done)
1863      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1864      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1865    
1866      /* If /S was present, study the regexp to generate additional info to      /* If -s or /S was present, study the regex to generate additional info to
1867      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1868        suppresses the effect of /S (used for a few test patterns where studying is
1869        never sensible). */
1870    
1871      if (do_study)      if (do_study || (force_study >= 0 && !no_force_study))
1872        {        {
1873        if (timeit)        if (timeit > 0)
1874          {          {
1875          register int i;          register int i;
1876          clock_t time_taken;          clock_t time_taken;
1877          clock_t start_time = clock();          clock_t start_time = clock();
1878          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1879            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options | force_study_options, &error);
1880          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1881          if (extra != NULL) free(extra);          if (extra != NULL) pcre_free_study(extra);
1882          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1883            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1884              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1885          }          }
1886        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options | force_study_options, &error);
1887        if (error != NULL)        if (error != NULL)
1888          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
1889        else if (extra != NULL)        else if (extra != NULL)
1890          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1891        }        }
1892    
1893        /* If /K was present, we set up for handling MARK data. */
1894    
1895        if (do_mark)
1896          {
1897          if (extra == NULL)
1898            {
1899            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1900            extra->flags = 0;
1901            }
1902          extra->mark = &markptr;
1903          extra->flags |= PCRE_EXTRA_MARK;
1904          }
1905    
1906      /* 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
1907      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
1908      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 980  while (!done) Line 1911  while (!done)
1911      if (do_flip)      if (do_flip)
1912        {        {
1913        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1914        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1915            byteflip(rre->magic_number, sizeof(rre->magic_number));
1916        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1917        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1918        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1919        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1920        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1921        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1922        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1923          rre->first_char =
1924            (pcre_uint16)byteflip(rre->first_char, sizeof(rre->first_char));
1925          rre->req_char =
1926            (pcre_uint16)byteflip(rre->req_char, sizeof(rre->req_char));
1927          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1928          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1929        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1930          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1931        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1932            sizeof(rre->name_count));
1933    
1934        if (extra != NULL)        if (extra != NULL)
1935          {          {
1936          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1937          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1938          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1939            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1940          }          }
1941        }        }
1942    
# Line 1005  while (!done) Line 1944  while (!done)
1944    
1945      SHOW_INFO:      SHOW_INFO:
1946    
1947        if (do_debug)
1948          {
1949          fprintf(outfile, "------------------------------------------------------------------\n");
1950          pcre_printint(re, outfile, debug_lengths);
1951          }
1952    
1953        /* We already have the options in get_options (see above) */
1954    
1955      if (do_showinfo)      if (do_showinfo)
1956        {        {
1957        unsigned long int get_options, all_options;        unsigned long int all_options;
1958    #if !defined NOINFOCHECK
1959        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1960        int count, backrefmax, first_char, need_char;  #endif
1961          int count, backrefmax, first_char, need_char, okpartial, jchanged,
1962            hascrorlf;
1963        int nameentrysize, namecount;        int nameentrysize, namecount;
1964        const uschar *nametable;        const pcre_uchar *nametable;
1965    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         print_internals(re, outfile);  
         }  
   
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1966        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1967        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1968        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1028  while (!done) Line 1971  while (!done)
1971        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1972        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1973        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1974          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1975          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1976          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1977    
1978    #if !defined NOINFOCHECK
1979        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
1980        if (count < 0) fprintf(outfile,        if (count < 0) fprintf(outfile,
1981          "Error %d from pcre_info()\n", count);          "Error %d from pcre_info()\n", count);
# Line 1046  while (!done) Line 1993  while (!done)
1993            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
1994              get_options, old_options);              get_options, old_options);
1995          }          }
1996    #endif
1997    
1998        if (size != regex_gotten_store) fprintf(outfile,        if (size != regex_gotten_store) fprintf(outfile,
1999          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
2000          size, regex_gotten_store);          (int)size, (int)regex_gotten_store);
2001    
2002        fprintf(outfile, "Capturing subpattern count = %d\n", count);        fprintf(outfile, "Capturing subpattern count = %d\n", count);
2003        if (backrefmax > 0)        if (backrefmax > 0)
# Line 1067  while (!done) Line 2015  while (!done)
2015            }            }
2016          }          }
2017    
2018        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
2019        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
2020    
2021        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
2022        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");  
2023    
2024        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
2025          else fprintf(outfile, "Options:%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",
2026            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
2027            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
2028            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
2029            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
2030              ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
2031            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
2032              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
2033              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
2034            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
2035            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
2036            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
2037              ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
2038            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
2039            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");            ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2040              ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2041              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2042              ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2043    
2044        if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
2045          fprintf(outfile, "Case state changes\n");  
2046          switch (get_options & PCRE_NEWLINE_BITS)
2047            {
2048            case PCRE_NEWLINE_CR:
2049            fprintf(outfile, "Forced newline sequence: CR\n");
2050            break;
2051    
2052            case PCRE_NEWLINE_LF:
2053            fprintf(outfile, "Forced newline sequence: LF\n");
2054            break;
2055    
2056            case PCRE_NEWLINE_CRLF:
2057            fprintf(outfile, "Forced newline sequence: CRLF\n");
2058            break;
2059    
2060            case PCRE_NEWLINE_ANYCRLF:
2061            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
2062            break;
2063    
2064            case PCRE_NEWLINE_ANY:
2065            fprintf(outfile, "Forced newline sequence: ANY\n");
2066            break;
2067    
2068            default:
2069            break;
2070            }
2071    
2072        if (first_char == -1)        if (first_char == -1)
2073          {          {
2074          fprintf(outfile, "First char at start or follows \\n\n");          fprintf(outfile, "First char at start or follows newline\n");
2075          }          }
2076        else if (first_char < 0)        else if (first_char < 0)
2077          {          {
# Line 1105  while (!done) Line 2079  while (!done)
2079          }          }
2080        else        else
2081          {          {
2082          int ch = first_char & 255;          const char *caseless =
2083          const char *caseless = ((first_char & REQ_CASELESS) == 0)?            ((((real_pcre *)re)->flags & PCRE_FCH_CASELESS) == 0)?
2084            "" : " (caseless)";            "" : " (caseless)";
2085          if (isprint(ch))  
2086            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);          if (PRINTHEX(first_char))
2087              fprintf(outfile, "First char = \'%c\'%s\n", first_char, caseless);
2088          else          else
2089            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", first_char, caseless);
2090          }          }
2091    
2092        if (need_char < 0)        if (need_char < 0)
# Line 1120  while (!done) Line 2095  while (!done)
2095          }          }
2096        else        else
2097          {          {
2098          int ch = need_char & 255;          const char *caseless =
2099          const char *caseless = ((need_char & REQ_CASELESS) == 0)?            ((((real_pcre *)re)->flags & PCRE_RCH_CASELESS) == 0)?
2100            "" : " (caseless)";            "" : " (caseless)";
2101          if (isprint(ch))  
2102            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);          if (PRINTHEX(need_char))
2103              fprintf(outfile, "Need char = \'%c\'%s\n", need_char, caseless);
2104          else          else
2105            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", need_char, caseless);
2106          }          }
2107    
2108        /* Don't output study size; at present it is in any case a fixed        /* Don't output study size; at present it is in any case a fixed
2109        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2110        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
2111        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2112          information unless -i or -d was also present. This means that, except
2113          when auto-callouts are involved, the output from runs with and without
2114          -s should be identical. */
2115    
2116        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
2117          {          {
2118          if (extra == NULL)          if (extra == NULL)
2119            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2120          else          else
2121            {            {
2122            uschar *start_bits = NULL;            pcre_uint8 *start_bits = NULL;
2123            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2124    
2125              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2126              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2127    
2128              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2129            if (start_bits == NULL)            if (start_bits == NULL)
2130              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2131            else            else
2132              {              {
2133              int i;              int i;
# Line 1159  while (!done) Line 2142  while (!done)
2142                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
2143                    c = 2;                    c = 2;
2144                    }                    }
2145                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
2146                    {                    {
2147                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
2148                    c += 2;                    c += 2;
# Line 1174  while (!done) Line 2157  while (!done)
2157              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2158              }              }
2159            }            }
2160    
2161            /* Show this only if the JIT was set by /S, not by -s. */
2162    
2163            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
2164              {
2165              int jit;
2166              new_info(re, extra, PCRE_INFO_JIT, &jit);
2167              if (jit)
2168                fprintf(outfile, "JIT study was successful\n");
2169              else
2170    #ifdef SUPPORT_JIT
2171                fprintf(outfile, "JIT study was not successful\n");
2172    #else
2173                fprintf(outfile, "JIT support is not available in this version of PCRE\n");
2174    #endif
2175              }
2176          }          }
2177        }        }
2178    
# Line 1190  while (!done) Line 2189  while (!done)
2189          }          }
2190        else        else
2191          {          {
2192          uschar sbuf[8];          pcre_uint8 sbuf[8];
2193          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (pcre_uint8)((true_size >> 24) & 255);
2194          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (pcre_uint8)((true_size >> 16) & 255);
2195          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (pcre_uint8)((true_size >>  8) & 255);
2196          sbuf[3] = (true_size)  & 255;          sbuf[3] = (pcre_uint8)((true_size) & 255);
2197    
2198          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (pcre_uint8)((true_study_size >> 24) & 255);
2199          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (pcre_uint8)((true_study_size >> 16) & 255);
2200          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (pcre_uint8)((true_study_size >>  8) & 255);
2201          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (pcre_uint8)((true_study_size) & 255);
2202    
2203          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2204              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1208  while (!done) Line 2207  while (!done)
2207            }            }
2208          else          else
2209            {            {
2210            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2211    
2212              /* If there is study data, write it. */
2213    
2214            if (extra != NULL)            if (extra != NULL)
2215              {              {
2216              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1222  while (!done) Line 2224  while (!done)
2224            }            }
2225          fclose(f);          fclose(f);
2226          }          }
2227    
2228          new_free(re);
2229          if (extra != NULL) pcre_free_study(extra);
2230          if (locale_set)
2231            {
2232            new_free((void *)tables);
2233            setlocale(LC_CTYPE, "C");
2234            locale_set = 0;
2235            }
2236        continue;  /* With next regex */        continue;  /* With next regex */
2237        }        }
2238      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1230  while (!done) Line 2241  while (!done)
2241    
2242    for (;;)    for (;;)
2243      {      {
2244      unsigned char *q;      pcre_uint8 *q;
2245      unsigned char *bptr = dbuffer;      pcre_uint8 *bptr;
2246      int *use_offsets = offsets;      int *use_offsets = offsets;
2247      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2248      int callout_data = 0;      int callout_data = 0;
2249      int callout_data_set = 0;      int callout_data_set = 0;
2250      int count, c;      int count, c;
2251      int copystrings = 0;      int copystrings = 0;
2252      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2253      int getstrings = 0;      int getstrings = 0;
2254      int getlist = 0;      int getlist = 0;
2255      int gmatched = 0;      int gmatched = 0;
2256      int start_offset = 0;      int start_offset = 0;
2257        int start_offset_sign = 1;
2258      int g_notempty = 0;      int g_notempty = 0;
2259        int use_dfa = 0;
2260    
2261      options = 0;      options = 0;
2262    
2263        *copynames = 0;
2264        *getnames = 0;
2265    
2266        copynamesptr = copynames;
2267        getnamesptr = getnames;
2268    
2269      pcre_callout = callout;      pcre_callout = callout;
2270      first_callout = 1;      first_callout = 1;
2271        last_callout_mark = NULL;
2272      callout_extra = 0;      callout_extra = 0;
2273      callout_count = 0;      callout_count = 0;
2274      callout_fail_count = 999999;      callout_fail_count = 999999;
2275      callout_fail_id = -1;      callout_fail_id = -1;
2276      show_malloc = 0;      show_malloc = 0;
2277    
2278      if (infile == stdin) printf("data> ");      if (extra != NULL) extra->flags &=
2279      if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
2280    
2281        len = 0;
2282        for (;;)
2283        {        {
2284        done = 1;        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
2285        goto CONTINUE;          {
2286            if (len > 0)    /* Reached EOF without hitting a newline */
2287              {
2288              fprintf(outfile, "\n");
2289              break;
2290              }
2291            done = 1;
2292            goto CONTINUE;
2293            }
2294          if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
2295          len = (int)strlen((char *)buffer);
2296          if (buffer[len-1] == '\n') break;
2297        }        }
     if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);  
2298    
     len = (int)strlen((char *)buffer);  
2299      while (len > 0 && isspace(buffer[len-1])) len--;      while (len > 0 && isspace(buffer[len-1])) len--;
2300      buffer[len] = 0;      buffer[len] = 0;
2301      if (len == 0) break;      if (len == 0) break;
# Line 1271  while (!done) Line 2303  while (!done)
2303      p = buffer;      p = buffer;
2304      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2305    
2306      q = dbuffer;      bptr = q = dbuffer;
2307      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2308        {        {
2309        int i = 0;        int i = 0;
# Line 1293  while (!done) Line 2325  while (!done)
2325          c -= '0';          c -= '0';
2326          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
2327            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
2328    
2329    #if !defined NOUTF8
2330            if (use_utf8 && c > 255)
2331              {
2332              unsigned char buff8[8];
2333              int ii, utn;
2334              utn = ord2utf8(c, buff8);
2335              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2336              c = buff8[ii];   /* Last byte */
2337              }
2338    #endif
2339          break;          break;
2340    
2341          case 'x':          case 'x':
2342    
2343          /* Handle \x{..} specially - new Perl thing for utf8 */          /* Handle \x{..} specially - new Perl thing for utf8 */
2344    
2345    #if !defined NOUTF8
2346          if (*p == '{')          if (*p == '{')
2347            {            {
2348            unsigned char *pt = p;            unsigned char *pt = p;
2349            c = 0;            c = 0;
2350            while (isxdigit(*(++pt)))  
2351              c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W');            /* We used to have "while (isxdigit(*(++pt)))" here, but it fails
2352              when isxdigit() is a macro that refers to its argument more than
2353              once. This is banned by the C Standard, but apparently happens in at
2354              least one MacOS environment. */
2355    
2356              for (pt++; isxdigit(*pt); pt++)
2357                c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'a' - 10);
2358            if (*pt == '}')            if (*pt == '}')
2359              {              {
2360              unsigned char buff8[8];              unsigned char buff8[8];
2361              int ii, utn;              int ii, utn;
2362              utn = ord2utf8(c, buff8);              if (use_utf8)
2363              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2364              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2365                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2366                  c = buff8[ii];   /* Last byte */
2367                  }
2368                else
2369                 {
2370                 if (c > 255)
2371                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2372                     "UTF-8 mode is not enabled.\n"
2373                     "** Truncation will probably give the wrong result.\n", c);
2374                 }
2375              p = pt + 1;              p = pt + 1;
2376              break;              break;
2377              }              }
2378            /* Not correct form; fall through */            /* Not correct form; fall through */
2379            }            }
2380    #endif
2381    
2382          /* Ordinary \x */          /* Ordinary \x */
2383    
2384          c = 0;          c = 0;
2385          while (i++ < 2 && isxdigit(*p))          while (i++ < 2 && isxdigit(*p))
2386            {            {
2387            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W');            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'a' - 10);
2388            p++;            p++;
2389            }            }
2390          break;          break;
# Line 1333  while (!done) Line 2394  while (!done)
2394          continue;          continue;
2395    
2396          case '>':          case '>':
2397            if (*p == '-')
2398              {
2399              start_offset_sign = -1;
2400              p++;
2401              }
2402          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2403            start_offset *= start_offset_sign;
2404          continue;          continue;
2405    
2406          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1352  while (!done) Line 2419  while (!done)
2419            }            }
2420          else if (isalnum(*p))          else if (isalnum(*p))
2421            {            {
2422            uschar name[256];            pcre_uchar *npp = copynamesptr;
           uschar *npp = name;  
2423            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2424              *npp++ = 0;
2425            *npp = 0;            *npp = 0;
2426            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)copynamesptr);
2427            if (n < 0)            if (n < 0)
2428              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
2429            else copystrings |= 1 << n;            copynamesptr = npp;
2430            }            }
2431          else if (*p == '+')          else if (*p == '+')
2432            {            {
# Line 1397  while (!done) Line 2464  while (!done)
2464            }            }
2465          continue;          continue;
2466    
2467    #if !defined NODFA
2468            case 'D':
2469    #if !defined NOPOSIX
2470            if (posix || do_posix)
2471              printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");
2472            else
2473    #endif
2474              use_dfa = 1;
2475            continue;
2476    #endif
2477    
2478    #if !defined NODFA
2479            case 'F':
2480            options |= PCRE_DFA_SHORTEST;
2481            continue;
2482    #endif
2483    
2484          case 'G':          case 'G':
2485          if (isdigit(*p))          if (isdigit(*p))
2486            {            {
# Line 1405  while (!done) Line 2489  while (!done)
2489            }            }
2490          else if (isalnum(*p))          else if (isalnum(*p))
2491            {            {
2492            uschar name[256];            pcre_uchar *npp = getnamesptr;
           uschar *npp = name;  
2493            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2494              *npp++ = 0;
2495            *npp = 0;            *npp = 0;
2496            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)getnamesptr);
2497            if (n < 0)            if (n < 0)
2498              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
2499            else getstrings |= 1 << n;            getnamesptr = npp;
2500              }
2501            continue;
2502    
2503            case 'J':
2504            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2505            if (extra != NULL
2506                && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
2507                && extra->executable_jit != NULL)
2508              {
2509              if (jit_stack != NULL) pcre_jit_stack_free(jit_stack);
2510              jit_stack = pcre_jit_stack_alloc(1, n * 1024);
2511              pcre_assign_jit_stack(extra, jit_callback, jit_stack);
2512            }            }
2513          continue;          continue;
2514    
# Line 1425  while (!done) Line 2521  while (!done)
2521          continue;          continue;
2522    
2523          case 'N':          case 'N':
2524          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2525              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2526            else
2527              options |= PCRE_NOTEMPTY;
2528          continue;          continue;
2529    
2530          case 'O':          case 'O':
# Line 1438  while (!done) Line 2537  while (!done)
2537            if (offsets == NULL)            if (offsets == NULL)
2538              {              {
2539              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2540                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2541              return 1;              yield = 1;
2542                goto EXIT;
2543              }              }
2544            }            }
2545          use_size_offsets = n;          use_size_offsets = n;
# Line 1447  while (!done) Line 2547  while (!done)
2547          continue;          continue;
2548    
2549          case 'P':          case 'P':
2550          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2551              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2552          continue;          continue;
2553    
2554            case 'Q':
2555            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2556            if (extra == NULL)
2557              {
2558              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2559              extra->flags = 0;
2560              }
2561            extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
2562            extra->match_limit_recursion = n;
2563            continue;
2564    
2565            case 'q':
2566            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2567            if (extra == NULL)
2568              {
2569              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2570              extra->flags = 0;
2571              }
2572            extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
2573            extra->match_limit = n;
2574            continue;
2575    
2576    #if !defined NODFA
2577            case 'R':
2578            options |= PCRE_DFA_RESTART;
2579            continue;
2580    #endif
2581    
2582          case 'S':          case 'S':
2583          show_malloc = 1;          show_malloc = 1;
2584          continue;          continue;
2585    
2586            case 'Y':
2587            options |= PCRE_NO_START_OPTIMIZE;
2588            continue;
2589    
2590          case 'Z':          case 'Z':
2591          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2592          continue;          continue;
# Line 1461  while (!done) Line 2594  while (!done)
2594          case '?':          case '?':
2595          options |= PCRE_NO_UTF8_CHECK;          options |= PCRE_NO_UTF8_CHECK;
2596          continue;          continue;
2597    
2598            case '<':
2599              {
2600              int x = check_newline(p, outfile);
2601              if (x == 0) goto NEXT_DATA;
2602              options |= x;
2603              while (*p++ != '>');
2604              }
2605            continue;
2606          }          }
2607        *q++ = c;        *q++ = c;
2608        }        }
2609      *q = 0;      *q = 0;
2610      len = q - dbuffer;      len = (int)(q - dbuffer);
2611    
2612        /* Move the data to the end of the buffer so that a read over the end of
2613        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2614        we are using the POSIX interface, we must include the terminating zero. */
2615    
2616    #if !defined NOPOSIX
2617        if (posix || do_posix)
2618          {
2619          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2620          bptr += buffer_size - len - 1;
2621          }
2622        else
2623    #endif
2624          {
2625          memmove(bptr + buffer_size - len, bptr, len);
2626          bptr += buffer_size - len;
2627          }
2628    
2629        if ((all_use_dfa || use_dfa) && find_match_limit)
2630          {
2631          printf("**Match limit not relevant for DFA matching: ignored\n");
2632          find_match_limit = 0;
2633          }
2634    
2635      /* Handle matching via the POSIX interface, which does not      /* Handle matching via the POSIX interface, which does not
2636      support timing or playing with the match limit or callout data. */      support timing or playing with the match limit or callout data. */
# Line 1480  while (!done) Line 2645  while (!done)
2645          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2646        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2647        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2648          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2649    
2650        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2651    
2652        if (rc != 0)        if (rc != 0)
2653          {          {
2654          (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);          (void)regerror(rc, &preg, (char *)buffer, buffer_size);
2655          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
2656          }          }
2657          else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
2658                  != 0)
2659            {
2660            fprintf(outfile, "Matched with REG_NOSUB\n");
2661            }
2662        else        else
2663          {          {
2664          size_t i;          size_t i;
# Line 1499  while (!done) Line 2670  while (!done)
2670              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2671                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2672              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2673              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2674                {                {
2675                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2676                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2677                  outfile);                  outfile);
2678                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 1519  while (!done) Line 2690  while (!done)
2690    
2691      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2692        {        {
2693        if (timeit)        markptr = NULL;
2694    
2695          if (timeitm > 0)
2696          {          {
2697          register int i;          register int i;
2698          clock_t time_taken;          clock_t time_taken;
2699          clock_t start_time = clock();          clock_t start_time = clock();
2700          for (i = 0; i < LOOPREPEAT; i++)  
2701    #if !defined NODFA
2702            if (all_use_dfa || use_dfa)
2703              {
2704              int workspace[1000];
2705              for (i = 0; i < timeitm; i++)
2706                count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2707                  options | g_notempty, use_offsets, use_size_offsets, workspace,
2708                  sizeof(workspace)/sizeof(int));
2709              }
2710            else
2711    #endif
2712    
2713            for (i = 0; i < timeitm; i++)
2714            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2715              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2716    
2717          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2718          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2719            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2720              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2721          }          }
2722    
2723        /* If find_match_limit is set, we want to do repeated matches with        /* If find_match_limit is set, we want to do repeated matches with
2724        varying limits in order to find the minimum value. */        varying limits in order to find the minimum value for the match limit and
2725          for the recursion limit. The match limits are relevant only to the normal
2726          running of pcre_exec(), so disable the JIT optimization. This makes it
2727          possible to run the same set of tests with and without JIT externally
2728          requested. */
2729    
2730        if (find_match_limit)        if (find_match_limit)
2731          {          {
         int min = 0;  
         int mid = 64;  
         int max = -1;  
   
2732          if (extra == NULL)          if (extra == NULL)
2733            {            {
2734            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2735            extra->flags = 0;            extra->flags = 0;
2736            }            }
2737          extra->flags |= PCRE_EXTRA_MATCH_LIMIT;          else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
2738    
2739          for (;;)          (void)check_match_limit(re, extra, bptr, len, start_offset,
2740            {            options|g_notempty, use_offsets, use_size_offsets,
2741            extra->match_limit = mid;            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
2742            count = pcre_exec(re, extra, (char *)bptr, len, start_offset,            PCRE_ERROR_MATCHLIMIT, "match()");
2743              options | g_notempty, use_offsets, use_size_offsets);  
2744            if (count == PCRE_ERROR_MATCHLIMIT)          count = check_match_limit(re, extra, bptr, len, start_offset,
2745              {            options|g_notempty, use_offsets, use_size_offsets,
2746              /* fprintf(outfile, "Testing match limit = %d\n", mid); */            PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
2747              min = mid;            PCRE_ERROR_RECURSIONLIMIT, "match() recursion");
             mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;  
             }  
           else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||  
                                  count == PCRE_ERROR_PARTIAL)  
             {  
             if (mid == min + 1)  
               {  
               fprintf(outfile, "Minimum match limit = %d\n", mid);  
               break;  
               }  
             /* fprintf(outfile, "Testing match limit = %d\n", mid); */  
             max = mid;  
             mid = (min + mid)/2;  
             }  
           else break;    /* Some other error */  
           }  
   
         extra->flags &= ~PCRE_EXTRA_MATCH_LIMIT;  
2748          }          }
2749    
2750        /* If callout_data is set, use the interface with additional data */        /* If callout_data is set, use the interface with additional data */
# Line 1597  while (!done) Line 2766  while (!done)
2766        /* The normal case is just to do the match once, with the default        /* The normal case is just to do the match once, with the default
2767        value of match_limit. */        value of match_limit. */
2768    
2769        else  #if !defined NODFA
2770          else if (all_use_dfa || use_dfa)
2771          {          {
2772          count = pcre_exec(re, extra, (char *)bptr, len,          int workspace[1000];
2773            start_offset, options | g_notempty, use_offsets, use_size_offsets);          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2774              options | g_notempty, use_offsets, use_size_offsets, workspace,
2775              sizeof(workspace)/sizeof(int));
2776            if (count == 0)
2777              {
2778              fprintf(outfile, "Matched, but too many subsidiary matches\n");
2779              count = use_size_offsets/2;
2780              }
2781          }          }
2782    #endif
2783    
2784        if (count == 0)        else
2785          {          {
2786          fprintf(outfile, "Matched, but too many substrings\n");          count = pcre_exec(re, extra, (char *)bptr, len,
2787          count = use_size_offsets/3;            start_offset, options | g_notempty, use_offsets, use_size_offsets);
2788            if (count == 0)
2789              {
2790              fprintf(outfile, "Matched, but too many substrings\n");
2791              count = use_size_offsets/3;
2792              }
2793          }          }
2794    
2795        /* Matched */        /* Matched */
2796    
2797        if (count >= 0)        if (count >= 0)
2798          {          {
2799          int i;          int i, maxcount;
2800    
2801    #if !defined NODFA
2802            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2803    #endif
2804              maxcount = use_size_offsets/3;
2805    
2806            /* This is a check against a lunatic return value. */
2807    
2808            if (count > maxcount)
2809              {
2810              fprintf(outfile,
2811                "** PCRE error: returned count %d is too big for offset size %d\n",
2812                count, use_size_offsets);
2813              count = use_size_offsets/3;
2814              if (do_g || do_G)
2815                {
2816                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2817                do_g = do_G = FALSE;        /* Break g/G loop */
2818                }
2819              }
2820    
2821            /* do_allcaps requests showing of all captures in the pattern, to check
2822            unset ones at the end. */
2823    
2824            if (do_allcaps)
2825              {
2826              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2827              count++;   /* Allow for full match */
2828              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2829              }
2830    
2831            /* Output the captured substrings */
2832    
2833          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2834            {            {
2835            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2836                {
2837                if (use_offsets[i] != -1)
2838                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2839                    use_offsets[i], i);
2840                if (use_offsets[i+1] != -1)
2841                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2842                    use_offsets[i+1], i+1);
2843              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2844                }
2845            else            else
2846              {              {
2847              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2848              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2849                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2850              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2851              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2852                {                {
2853                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2854                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2855                  fprintf(outfile, " 0+ ");                  outfile);
2856                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2857                }                }
2858              }              }
2859            }            }
2860    
2861            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2862    
2863          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2864            {            {
2865            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
2866              {              {
2867              char copybuffer[16];              char copybuffer[256];
2868              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,
2869                i, copybuffer, sizeof(copybuffer));                i, copybuffer, sizeof(copybuffer));
2870              if (rc < 0)              if (rc < 0)
# Line 1651  while (!done) Line 2874  while (!done)
2874              }              }
2875            }            }
2876    
2877            for (copynamesptr = copynames;
2878                 *copynamesptr != 0;
2879                 copynamesptr += (int)strlen((char*)copynamesptr) + 1)
2880              {
2881              char copybuffer[256];
2882              int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,
2883                count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));
2884              if (rc < 0)
2885                fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);
2886              else
2887                fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);
2888              }
2889    
2890          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2891            {            {
2892            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
# Line 1663  while (!done) Line 2899  while (!done)
2899              else              else
2900                {                {
2901                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
               /* free((void *)substring); */  
2902                pcre_free_substring(substring);                pcre_free_substring(substring);
2903                }                }
2904              }              }
2905            }            }
2906    
2907            for (getnamesptr = getnames;
2908                 *getnamesptr != 0;
2909                 getnamesptr += (int)strlen((char*)getnamesptr) + 1)
2910              {
2911              const char *substring;
2912              int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,
2913                count, (char *)getnamesptr, &substring);
2914              if (rc < 0)
2915                fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);
2916              else
2917                {
2918                fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);
2919                pcre_free_substring(substring);
2920                }
2921              }
2922    
2923          if (getlist)          if (getlist)
2924            {            {
2925            const char **stringlist;            const char **stringlist;
# Line 1682  while (!done) Line 2933  while (!done)
2933                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);
2934              if (stringlist[i] != NULL)              if (stringlist[i] != NULL)
2935                fprintf(outfile, "string list not terminated by NULL\n");                fprintf(outfile, "string list not terminated by NULL\n");
             /* free((void *)stringlist); */  
2936              pcre_free_substring_list(stringlist);              pcre_free_substring_list(stringlist);
2937              }              }
2938            }            }
# Line 1692  while (!done) Line 2942  while (!done)
2942    
2943        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2944          {          {
2945          fprintf(outfile, "Partial match\n");          if (markptr == NULL) fprintf(outfile, "Partial match");
2946              else fprintf(outfile, "Partial match, mark=%s", markptr);
2947            if (use_size_offsets > 1)
2948              {
2949              fprintf(outfile, ": ");
2950              pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2951                outfile);
2952              }
2953            fprintf(outfile, "\n");
2954          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2955          }          }
2956    
2957        /* 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
2958        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
2959        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
2960        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2961        offset values to achieve this. We won't be at the end of the string -  
2962        that was checked before setting g_notempty. */        Complication arises in the case when the newline convention is "any",
2963          "crlf", or "anycrlf". If the previous match was at the end of a line
2964          terminated by CRLF, an advance of one character just passes the \r,
2965          whereas we should prefer the longer newline sequence, as does the code in
2966          pcre_exec(). Fudge the offset value to achieve this. We check for a
2967          newline setting in the pattern; if none was set, use pcre_config() to
2968          find the default.
2969    
2970          Otherwise, in the case of UTF-8 matching, the advance must be one
2971          character, not one byte. */
2972    
2973        else        else
2974          {          {
2975          if (g_notempty != 0)          if (g_notempty != 0)
2976            {            {
2977            int onechar = 1;            int onechar = 1;
2978              unsigned int obits = ((real_pcre *)re)->options;
2979            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2980            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2981                {
2982                int d;
2983                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2984                /* Note that these values are always the ASCII ones, even in
2985                EBCDIC environments. CR = 13, NL = 10. */
2986                obits = (d == 13)? PCRE_NEWLINE_CR :
2987                        (d == 10)? PCRE_NEWLINE_LF :
2988                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2989                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2990                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2991                }
2992              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2993                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
2994                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2995                  &&
2996                  start_offset < len - 1 &&
2997                  bptr[start_offset] == '\r' &&
2998                  bptr[start_offset+1] == '\n')
2999                onechar++;
3000              else if (use_utf8)
3001              {              {
3002              while (start_offset + onechar < len)              while (start_offset + onechar < len)
3003                {                {
3004                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
3005                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
3006                }                }
3007              }              }
3008            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
3009            }            }
3010          else          else
3011            {            {
3012            if (count == PCRE_ERROR_NOMATCH)            switch(count)
3013              {              {
3014              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
3015                if (gmatched == 0)
3016                  {
3017                  if (markptr == NULL) fprintf(outfile, "No match\n");
3018                    else fprintf(outfile, "No match, mark = %s\n", markptr);
3019                  }
3020                break;
3021    
3022                case PCRE_ERROR_BADUTF8:
3023                case PCRE_ERROR_SHORTUTF8:
3024                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
3025                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
3026                if (use_size_offsets >= 2)
3027                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
3028                    use_offsets[1]);
3029                fprintf(outfile, "\n");
3030                break;
3031    
3032                default:
3033                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
3034                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
3035                else
3036                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
3037                break;
3038              }              }
3039            else fprintf(outfile, "Error %d\n", count);  
3040            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
3041            }            }
3042          }          }
# Line 1737  while (!done) Line 3046  while (!done)
3046        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
3047    
3048        /* 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
3049        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
3050        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
3051        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
3052        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
3053        character. */        character. */
3054    
3055        g_notempty = 0;        g_notempty = 0;
3056    
3057        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
3058          {          {
3059          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
3060          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
3061          }          }
3062    
3063        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 1762  while (!done) Line 3072  while (!done)
3072          len -= use_offsets[1];          len -= use_offsets[1];
3073          }          }
3074        }  /* End of loop for /g and /G */        }  /* End of loop for /g and /G */
3075    
3076        NEXT_DATA: continue;
3077      }    /* End of loop for data lines */      }    /* End of loop for data lines */
3078    
3079    CONTINUE:    CONTINUE:
# Line 1770  while (!done) Line 3082  while (!done)
3082    if (posix || do_posix) regfree(&preg);    if (posix || do_posix) regfree(&preg);
3083  #endif  #endif
3084    
3085    if (re != NULL) free(re);    if (re != NULL) new_free(re);
3086    if (extra != NULL) free(extra);    if (extra != NULL) pcre_free_study(extra);
3087    if (tables != NULL)    if (locale_set)
3088      {      {
3089      free((void *)tables);      new_free((void *)tables);
3090      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
3091        locale_set = 0;
3092        }
3093      if (jit_stack != NULL)
3094        {
3095        pcre_jit_stack_free(jit_stack);
3096        jit_stack = NULL;
3097      }      }
3098    }    }
3099    
3100  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");
3101  return 0;  
3102    EXIT:
3103    
3104    if (infile != NULL && infile != stdin) fclose(infile);
3105    if (outfile != NULL && outfile != stdout) fclose(outfile);
3106    
3107    free(buffer);
3108    free(dbuffer);
3109    free(pbuffer);
3110    free(offsets);
3111    
3112    return yield;
3113  }  }
3114    
3115  /* End */  /* End of pcretest.c */

Legend:
Removed from v.75  
changed lines
  Added in v.774

  ViewVC Help
Powered by ViewVC 1.1.5