/[pcre]/code/trunk/pcretest.c
ViewVC logotype

Diff of /code/trunk/pcretest.c

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

revision 75 by nigel, Sat Feb 24 21:40:37 2007 UTC revision 674 by ph10, Fri Aug 26 18:17:51 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  #define PCRE_SPY        /* For Win32 build, import data, not export */  
98  #include "internal.h"  /* 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_utf8_char_sizes  utf8_char_sizes
122    #define _pcre_utt              utt
123    #define _pcre_utt_size         utt_size
124    #define _pcre_utt_names        utt_names
125    #define _pcre_OP_lengths       OP_lengths
126    
127    #include "pcre_tables.c"
128    
129    /* We also need the pcre_printint() function for printing out compiled
130    patterns. This function is in a separate file so that it can be included in
131    pcre_compile.c when that module is compiled with debugging enabled. It needs to
132    know which case is being compiled. */
133    
134    #define COMPILING_PCRETEST
135    #include "pcre_printint.src"
136    
137    /* The definition of the macro PRINTABLE, which determines whether to print an
138    output character as-is or as a hex value when showing compiled patterns, is
139    contained in the printint.src file. We uses it here also, in cases when the
140    locale has not been explicitly changed, so as to get consistent output from
141    systems that differ in their output from isprint() even in the "C" locale. */
142    
143    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
144    
145  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
146  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 150  Makefile. */
150  #include "pcreposix.h"  #include "pcreposix.h"
151  #endif  #endif
152    
153    /* It is also possible, for the benefit of the version currently imported into
154    Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
155    interface to the DFA matcher (NODFA), and without the doublecheck of the old
156    "info" function (define NOINFOCHECK). In fact, we automatically cut out the
157    UTF8 support if PCRE is built without it. */
158    
159    #ifndef SUPPORT_UTF8
160    #ifndef NOUTF8
161    #define NOUTF8
162    #endif
163    #endif
164    
165    
166    /* Other parameters */
167    
168  #ifndef CLOCKS_PER_SEC  #ifndef CLOCKS_PER_SEC
169  #ifdef CLK_TCK  #ifdef CLK_TCK
170  #define CLOCKS_PER_SEC CLK_TCK  #define CLOCKS_PER_SEC CLK_TCK
# Line 66  Makefile. */ Line 173  Makefile. */
173  #endif  #endif
174  #endif  #endif
175    
176  #define LOOPREPEAT 500000  /* This is the default loop count for timing. */
177    
178  #define BUFFER_SIZE 30000  #define LOOPREPEAT 500000
 #define PBUFFER_SIZE BUFFER_SIZE  
 #define DBUFFER_SIZE BUFFER_SIZE  
179    
180    /* Static variables */
181    
182  static FILE *outfile;  static FILE *outfile;
183  static int log_store = 0;  static int log_store = 0;
# Line 79  static int callout_count; Line 185  static int callout_count;
185  static int callout_extra;  static int callout_extra;
186  static int callout_fail_count;  static int callout_fail_count;
187  static int callout_fail_id;  static int callout_fail_id;
188    static int debug_lengths;
189  static int first_callout;  static int first_callout;
190    static int locale_set = 0;
191  static int show_malloc;  static int show_malloc;
192  static int use_utf8;  static int use_utf8;
193  static size_t gotten_store;  static size_t gotten_store;
194    static const unsigned char *last_callout_mark = NULL;
195    
196    /* The buffers grow automatically if very long input lines are encountered. */
197    
198    static int buffer_size = 50000;
199    static uschar *buffer = NULL;
200    static uschar *dbuffer = NULL;
201  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
202    
203    /* Textual explanations for runtime error codes */
204    
205    static const char *errtexts[] = {
206      NULL,  /* 0 is no error */
207      NULL,  /* NOMATCH is handled specially */
208      "NULL argument passed",
209      "bad option value",
210      "magic number missing",
211      "unknown opcode - pattern overwritten?",
212      "no more memory",
213      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
214      "match limit exceeded",
215      "callout error code",
216      NULL,  /* BADUTF8 is handled specially */
217      "bad UTF-8 offset",
218      NULL,  /* PARTIAL is handled specially */
219      "not used - internal error",
220      "internal error - pattern overwritten?",
221      "bad count value",
222      "item unsupported for DFA matching",
223      "backreference condition or recursion test not supported for DFA matching",
224      "match limit not supported for DFA matching",
225      "workspace size exceeded in DFA matching",
226      "too much recursion for DFA matching",
227      "recursion limit exceeded",
228      "not used - internal error",
229      "invalid combination of newline options",
230      "bad offset value",
231      NULL,  /* SHORTUTF8 is handled specially */
232      "nested recursion at the same subject position"
233    };
234    
235    
236    /*************************************************
237    *         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      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
385      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    
560    
 static const int utf8_table1[] = {  
   0x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff};  
561    
 static const int utf8_table2[] = {  
   0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc};  
562    
563  static const int utf8_table3[] = {  #ifndef HAVE_STRERROR
564    0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};  /*************************************************
565    *     Provide strerror() for non-ANSI libraries  *
566    *************************************************/
567    
568    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
569    in their libraries, but can provide the same facility by this simple
570    alternative function. */
571    
572    extern int   sys_nerr;
573    extern char *sys_errlist[];
574    
575    char *
576    strerror(int n)
577    {
578    if (n < 0 || n >= sys_nerr) return "unknown error number";
579    return sys_errlist[n];
580    }
581    #endif /* HAVE_STRERROR */
582    
583    
584  /*************************************************  /*************************************************
585  *         Print compiled regex                   *  *         JIT memory callback                    *
586  *************************************************/  *************************************************/
587    
588  /* The code for doing this is held in a separate file that is also included in  static pcre_jit_stack* jit_callback(void *arg)
589  pcre.c when it is compiled with the debug switch. It defines a function called  {
590  print_internals(), which uses a table of opcode lengths defined by the macro  return (pcre_jit_stack *)arg;
591  OP_LENGTHS, whose name must be OP_lengths. It also uses a table that translates  }
592  Unicode property names to numbers; this is kept in a separate file. */  
593    
594  static uschar OP_lengths[] = { OP_LENGTHS };  /*************************************************
595    *        Read or extend an input line            *
596  #include "ucp.h"  *************************************************/
597  #include "ucptypetable.c"  
598  #include "printint.c"  /* 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 uschar *
618    extend_inputline(FILE *f, uschar *start, const char *prompt)
619    {
620    uschar *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        uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);
668        uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
669        uschar *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, uschar *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 449  if ((rc = pcre_fullinfo(re, study, optio Line 1050  if ((rc = pcre_fullinfo(re, study, optio
1050    
1051    
1052  /*************************************************  /*************************************************
1053    *      Check for supported JIT architecture      *
1054    *************************************************/
1055    
1056    /* If it won't JIT-compile a very simple regex, return FALSE. */
1057    
1058    static int check_jit_arch(void)
1059    {
1060    const char *error;
1061    int erroffset, rc;
1062    pcre *re = pcre_compile("abc", 0, &error, &erroffset, NULL);
1063    pcre_extra *extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
1064    rc = extra != NULL && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 &&
1065      extra->executable_jit != NULL;
1066    pcre_free_study(extra);
1067    free(re);
1068    return rc;
1069    }
1070    
1071    
1072    /*************************************************
1073  *         Byte flipping function                 *  *         Byte flipping function                 *
1074  *************************************************/  *************************************************/
1075    
1076  static long int  static unsigned long int
1077  byteflip(long int value, int n)  byteflip(unsigned long int value, int n)
1078  {  {
1079  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
1080  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
# Line 466  return ((value & 0x000000ff) << 24) | Line 1087  return ((value & 0x000000ff) << 24) |
1087    
1088    
1089  /*************************************************  /*************************************************
1090    *        Check match or recursion limit          *
1091    *************************************************/
1092    
1093    static int
1094    check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,
1095      int start_offset, int options, int *use_offsets, int use_size_offsets,
1096      int flag, unsigned long int *limit, int errnumber, const char *msg)
1097    {
1098    int count;
1099    int min = 0;
1100    int mid = 64;
1101    int max = -1;
1102    
1103    extra->flags |= flag;
1104    
1105    for (;;)
1106      {
1107      *limit = mid;
1108    
1109      count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,
1110        use_offsets, use_size_offsets);
1111    
1112      if (count == errnumber)
1113        {
1114        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
1115        min = mid;
1116        mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
1117        }
1118    
1119      else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
1120                             count == PCRE_ERROR_PARTIAL)
1121        {
1122        if (mid == min + 1)
1123          {
1124          fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
1125          break;
1126          }
1127        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
1128        max = mid;
1129        mid = (min + mid)/2;
1130        }
1131      else break;    /* Some other error */
1132      }
1133    
1134    extra->flags &= ~flag;
1135    return count;
1136    }
1137    
1138    
1139    
1140    /*************************************************
1141    *         Case-independent strncmp() function    *
1142    *************************************************/
1143    
1144    /*
1145    Arguments:
1146      s         first string
1147      t         second string
1148      n         number of characters to compare
1149    
1150    Returns:    < 0, = 0, or > 0, according to the comparison
1151    */
1152    
1153    static int
1154    strncmpic(uschar *s, uschar *t, int n)
1155    {
1156    while (n--)
1157      {
1158      int c = tolower(*s++) - tolower(*t++);
1159      if (c) return c;
1160      }
1161    return 0;
1162    }
1163    
1164    
1165    
1166    /*************************************************
1167    *         Check newline indicator                *
1168    *************************************************/
1169    
1170    /* This is used both at compile and run-time to check for <xxx> escapes. Print
1171    a message and return 0 if there is no match.
1172    
1173    Arguments:
1174      p           points after the leading '<'
1175      f           file for error message
1176    
1177    Returns:      appropriate PCRE_NEWLINE_xxx flags, or 0
1178    */
1179    
1180    static int
1181    check_newline(uschar *p, FILE *f)
1182    {
1183    if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
1184    if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
1185    if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
1186    if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
1187    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
1188    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
1189    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
1190    fprintf(f, "Unknown newline type at: <%s\n", p);
1191    return 0;
1192    }
1193    
1194    
1195    
1196    /*************************************************
1197    *             Usage function                     *
1198    *************************************************/
1199    
1200    static void
1201    usage(void)
1202    {
1203    printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
1204    printf("Input and output default to stdin and stdout.\n");
1205    #ifdef SUPPORT_LIBREADLINE
1206    printf("If input is a terminal, readline() is used to read from it.\n");
1207    #else
1208    printf("This version of pcretest is not linked with readline().\n");
1209    #endif
1210    printf("\nOptions:\n");
1211    printf("  -b       show compiled code (bytecode)\n");
1212    printf("  -C       show PCRE compile-time options and exit\n");
1213    printf("  -d       debug: show compiled code and information (-b and -i)\n");
1214    #if !defined NODFA
1215    printf("  -dfa     force DFA matching for all subjects\n");
1216    #endif
1217    printf("  -help    show usage information\n");
1218    printf("  -i       show information about compiled patterns\n"
1219           "  -M       find MATCH_LIMIT minimum for each subject\n"
1220           "  -m       output memory used information\n"
1221           "  -o <n>   set size of offsets vector to <n>\n");
1222    #if !defined NOPOSIX
1223    printf("  -p       use POSIX interface\n");
1224    #endif
1225    printf("  -q       quiet: do not output PCRE version number at start\n");
1226    printf("  -S <n>   set stack size to <n> megabytes\n");
1227    printf("  -s       force each pattern to be studied at basic level\n"
1228           "  -s+      force each pattern to be studied, using JIT if available\n"
1229           "  -t       time compilation and execution\n");
1230    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
1231    printf("  -tm      time execution (matching) only\n");
1232    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
1233    }
1234    
1235    
1236    
1237    /*************************************************
1238  *                Main Program                    *  *                Main Program                    *
1239  *************************************************/  *************************************************/
1240    
# Line 478  int main(int argc, char **argv) Line 1247  int main(int argc, char **argv)
1247  FILE *infile = stdin;  FILE *infile = stdin;
1248  int options = 0;  int options = 0;
1249  int study_options = 0;  int study_options = 0;
1250    int default_find_match_limit = FALSE;
1251  int op = 1;  int op = 1;
1252  int timeit = 0;  int timeit = 0;
1253    int timeitm = 0;
1254  int showinfo = 0;  int showinfo = 0;
1255  int showstore = 0;  int showstore = 0;
1256    int force_study = -1;
1257    int force_study_options = 0;
1258    int quiet = 0;
1259  int size_offsets = 45;  int size_offsets = 45;
1260  int size_offsets_max;  int size_offsets_max;
1261  int *offsets;  int *offsets = NULL;
1262  #if !defined NOPOSIX  #if !defined NOPOSIX
1263  int posix = 0;  int posix = 0;
1264  #endif  #endif
1265  int debug = 0;  int debug = 0;
1266  int done = 0;  int done = 0;
1267    int all_use_dfa = 0;
1268    int yield = 0;
1269    int stack_size;
1270    
1271    pcre_jit_stack *jit_stack = NULL;
1272    
1273    
1274  unsigned char *buffer;  /* These vectors store, end-to-end, a list of captured substring names. Assume
1275  unsigned char *dbuffer;  that 1024 is plenty long enough for the few names we'll be testing. */
1276    
1277    uschar copynames[1024];
1278    uschar getnames[1024];
1279    
1280    uschar *copynamesptr;
1281    uschar *getnamesptr;
1282    
1283  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* Get buffers from malloc() so that Electric Fence will check their misuse
1284  when I am debugging. */  when I am debugging. They grow automatically when very long lines are read. */
1285    
1286  buffer = (unsigned char *)malloc(BUFFER_SIZE);  buffer = (unsigned char *)malloc(buffer_size);
1287  dbuffer = (unsigned char *)malloc(DBUFFER_SIZE);  dbuffer = (unsigned char *)malloc(buffer_size);
1288  pbuffer = (unsigned char *)malloc(PBUFFER_SIZE);  pbuffer = (unsigned char *)malloc(buffer_size);
   
 /* The outfile variable is static so that new_malloc can use it. The _setmode()  
 stuff is some magic that I don't understand, but which apparently does good  
 things in Windows. It's related to line terminations.  */  
1289    
1290  #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) */  
1291    
1292  outfile = stdout;  outfile = stdout;
1293    
1294    /* The following  _setmode() stuff is some Windows magic that tells its runtime
1295    library to translate CRLF into a single LF character. At least, that's what
1296    I've been told: never having used Windows I take this all on trust. Originally
1297    it set 0x8000, but then I was advised that _O_BINARY was better. */
1298    
1299    #if defined(_WIN32) || defined(WIN32)
1300    _setmode( _fileno( stdout ), _O_BINARY );
1301    #endif
1302    
1303  /* Scan options */  /* Scan options */
1304    
1305  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
1306    {    {
1307    unsigned char *endptr;    unsigned char *endptr;
1308    
1309    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
1310      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 0;
1311    else if (strcmp(argv[op], "-t") == 0) timeit = 1;    else if (strcmp(argv[op], "-s+") == 0)
1312        {
1313        force_study = 1;
1314        force_study_options = PCRE_STUDY_JIT_COMPILE;
1315        }
1316      else if (strcmp(argv[op], "-q") == 0) quiet = 1;
1317      else if (strcmp(argv[op], "-b") == 0) debug = 1;
1318    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
1319    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
1320      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
1321    #if !defined NODFA
1322      else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
1323    #endif
1324    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
1325        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),
1326          *endptr == 0))          *endptr == 0))
# Line 529  while (argc > 1 && argv[op][0] == '-') Line 1328  while (argc > 1 && argv[op][0] == '-')
1328      op++;      op++;
1329      argc--;      argc--;
1330      }      }
1331      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
1332        {
1333        int both = argv[op][2] == 0;
1334        int temp;
1335        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
1336                         *endptr == 0))
1337          {
1338          timeitm = temp;
1339          op++;
1340          argc--;
1341          }
1342        else timeitm = LOOPREPEAT;
1343        if (both) timeit = timeitm;
1344        }
1345      else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
1346          ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
1347            *endptr == 0))
1348        {
1349    #if defined(_WIN32) || defined(WIN32) || defined(__minix)
1350        printf("PCRE: -S not supported on this OS\n");
1351        exit(1);
1352    #else
1353        int rc;
1354        struct rlimit rlim;
1355        getrlimit(RLIMIT_STACK, &rlim);
1356        rlim.rlim_cur = stack_size * 1024 * 1024;
1357        rc = setrlimit(RLIMIT_STACK, &rlim);
1358        if (rc != 0)
1359          {
1360        printf("PCRE: setrlimit() failed with error %d\n", rc);
1361        exit(1);
1362          }
1363        op++;
1364        argc--;
1365    #endif
1366        }
1367  #if !defined NOPOSIX  #if !defined NOPOSIX
1368    else if (strcmp(argv[op], "-p") == 0) posix = 1;    else if (strcmp(argv[op], "-p") == 0) posix = 1;
1369  #endif  #endif
1370    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
1371      {      {
1372      int rc;      int rc;
1373        unsigned long int lrc;
1374      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
1375      printf("Compiled with\n");      printf("Compiled with\n");
1376      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
1377      printf("  %sUTF-8 support\n", rc? "" : "No ");      printf("  %sUTF-8 support\n", rc? "" : "No ");
1378      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
1379      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
1380        (void)pcre_config(PCRE_CONFIG_JIT, &rc);
1381        if (rc)
1382          printf("  Just-in-time compiler support%s\n", check_jit_arch()?
1383            "" : " (but this architecture is unsupported)");
1384        else
1385          printf("  No just-in-time compiler support\n");
1386      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
1387      printf("  Newline character is %s\n", (rc == '\r')? "CR" : "LF");      /* Note that these values are always the ASCII values, even
1388        in EBCDIC environments. CR is 13 and NL is 10. */
1389        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
1390          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
1391          (rc == -2)? "ANYCRLF" :
1392          (rc == -1)? "ANY" : "???");
1393        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
1394        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
1395                                         "all Unicode newlines");
1396      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
1397      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
1398      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
1399      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
1400      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
1401      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
1402        (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
1403        printf("  Default recursion depth limit = %ld\n", lrc);
1404      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
1405      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
1406      exit(0);      goto EXIT;
1407        }
1408      else if (strcmp(argv[op], "-help") == 0 ||
1409               strcmp(argv[op], "--help") == 0)
1410        {
1411        usage();
1412        goto EXIT;
1413      }      }
1414    else    else
1415      {      {
1416      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
1417      printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");      usage();
1418      printf("  -C     show PCRE compile-time options and exit\n");      yield = 1;
1419      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;  
1420      }      }
1421    op++;    op++;
1422    argc--;    argc--;
# Line 580  offsets = (int *)malloc(size_offsets_max Line 1429  offsets = (int *)malloc(size_offsets_max
1429  if (offsets == NULL)  if (offsets == NULL)
1430    {    {
1431    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
1432      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
1433    return 1;    yield = 1;
1434      goto EXIT;
1435    }    }
1436    
1437  /* Sort out the input and output files */  /* Sort out the input and output files */
1438    
1439  if (argc > 1)  if (argc > 1)
1440    {    {
1441    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
1442    if (infile == NULL)    if (infile == NULL)
1443      {      {
1444      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
1445      return 1;      yield = 1;
1446        goto EXIT;
1447      }      }
1448    }    }
1449    
1450  if (argc > 2)  if (argc > 2)
1451    {    {
1452    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
1453    if (outfile == NULL)    if (outfile == NULL)
1454      {      {
1455      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
1456      return 1;      yield = 1;
1457        goto EXIT;
1458      }      }
1459    }    }
1460    
# Line 613  pcre_free = new_free; Line 1465  pcre_free = new_free;
1465  pcre_stack_malloc = stack_malloc;  pcre_stack_malloc = stack_malloc;
1466  pcre_stack_free = stack_free;  pcre_stack_free = stack_free;
1467    
1468  /* Heading line, then prompt for first regex if stdin */  /* Heading line unless quiet, then prompt for first regex if stdin */
1469    
1470  fprintf(outfile, "PCRE version %s\n\n", pcre_version());  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
1471    
1472  /* Main loop */  /* Main loop */
1473    
# Line 630  while (!done) Line 1482  while (!done)
1482  #endif  #endif
1483    
1484    const char *error;    const char *error;
1485      unsigned char *markptr;
1486    unsigned char *p, *pp, *ppp;    unsigned char *p, *pp, *ppp;
1487    unsigned char *to_file = NULL;    unsigned char *to_file = NULL;
1488    const unsigned char *tables = NULL;    const unsigned char *tables = NULL;
1489    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
1490    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1491      int do_allcaps = 0;
1492      int do_mark = 0;
1493    int do_study = 0;    int do_study = 0;
1494      int no_force_study = 0;
1495    int do_debug = debug;    int do_debug = debug;
1496    int do_G = 0;    int do_G = 0;
1497    int do_g = 0;    int do_g = 0;
1498    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1499    int do_showrest = 0;    int do_showrest = 0;
1500      int do_showcaprest = 0;
1501    int do_flip = 0;    int do_flip = 0;
1502    int erroroffset, len, delimiter;    int erroroffset, len, delimiter, poffset;
1503    
1504    use_utf8 = 0;    use_utf8 = 0;
1505      debug_lengths = 1;
1506    
1507    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) break;  
1508    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1509    fflush(outfile);    fflush(outfile);
1510    
# Line 659  while (!done) Line 1516  while (!done)
1516    
1517    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
1518      {      {
1519      unsigned long int magic;      unsigned long int magic, get_options;
1520      uschar sbuf[8];      uschar sbuf[8];
1521      FILE *f;      FILE *f;
1522    
# Line 702  while (!done) Line 1559  while (!done)
1559          }          }
1560        }        }
1561    
1562      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
1563        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
1564    
1565      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
1566    
1567      new_info(re, NULL, PCRE_INFO_OPTIONS, &options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1568      use_utf8 = (options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1569    
1570      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
1571    
1572      if (true_study_size != 0)      if (true_study_size != 0)
1573        {        {
# Line 726  while (!done) Line 1583  while (!done)
1583          {          {
1584          FAIL_READ:          FAIL_READ:
1585          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
1586          if (extra != NULL) new_free(extra);          if (extra != NULL) pcre_free_study(extra);
1587          if (re != NULL) new_free(re);          if (re != NULL) new_free(re);
1588          fclose(f);          fclose(f);
1589          continue;          continue;
# Line 747  while (!done) Line 1604  while (!done)
1604    
1605    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1606      {      {
1607      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1608      goto SKIP_DATA;      goto SKIP_DATA;
1609      }      }
1610    
1611    pp = p;    pp = p;
1612      poffset = (int)(p - buffer);
1613    
1614    for(;;)    for(;;)
1615      {      {
# Line 762  while (!done) Line 1620  while (!done)
1620        pp++;        pp++;
1621        }        }
1622      if (*pp != 0) break;      if (*pp != 0) break;
1623        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)  
1624        {        {
1625        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1626        done = 1;        done = 1;
# Line 780  while (!done) Line 1629  while (!done)
1629      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1630      }      }
1631    
1632      /* The buffer may have moved while being extended; reset the start of data
1633      pointer to the correct relative point in the buffer. */
1634    
1635      p = buffer + poffset;
1636    
1637    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1638    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1639    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 1649  while (!done)
1649    /* Look for options after final delimiter */    /* Look for options after final delimiter */
1650    
1651    options = 0;    options = 0;
   study_options = 0;  
1652    log_store = showstore;  /* default from command line */    log_store = showstore;  /* default from command line */
1653    
1654    while (*pp != 0)    while (*pp != 0)
1655      {      {
1656      switch (*pp++)      switch (*pp++)
1657        {        {
1658          case 'f': options |= PCRE_FIRSTLINE; break;
1659        case 'g': do_g = 1; break;        case 'g': do_g = 1; break;
1660        case 'i': options |= PCRE_CASELESS; break;        case 'i': options |= PCRE_CASELESS; break;
1661        case 'm': options |= PCRE_MULTILINE; break;        case 'm': options |= PCRE_MULTILINE; break;
1662        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
1663        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
1664    
1665        case '+': do_showrest = 1; break;        case '+':
1666          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
1667          break;
1668    
1669          case '=': do_allcaps = 1; break;
1670        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1671          case 'B': do_debug = 1; break;
1672        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1673        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1674        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
1675        case 'F': do_flip = 1; break;        case 'F': do_flip = 1; break;
1676        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1677        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1678          case 'J': options |= PCRE_DUPNAMES; break;
1679          case 'K': do_mark = 1; break;
1680        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1681        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1682    
# Line 823  while (!done) Line 1684  while (!done)
1684        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
1685  #endif  #endif
1686    
1687        case 'S': do_study = 1; break;        case 'S':
1688          if (do_study == 0)
1689            {
1690            do_study = 1;
1691            if (*pp == '+')
1692              {
1693              study_options |= PCRE_STUDY_JIT_COMPILE;
1694              pp++;
1695              }
1696            }
1697          else
1698            {
1699            do_study = 0;
1700            no_force_study = 1;
1701            }
1702          break;
1703    
1704        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1705          case 'W': options |= PCRE_UCP; break;
1706        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1707          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
1708          case 'Z': debug_lengths = 0; break;
1709        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1710        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1711    
1712          case 'T':
1713          switch (*pp++)
1714            {
1715            case '0': tables = tables0; break;
1716            case '1': tables = tables1; break;
1717    
1718            case '\r':
1719            case '\n':
1720            case ' ':
1721            case 0:
1722            fprintf(outfile, "** Missing table number after /T\n");
1723            goto SKIP_DATA;
1724    
1725            default:
1726            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
1727            goto SKIP_DATA;
1728            }
1729          break;
1730    
1731        case 'L':        case 'L':
1732        ppp = pp;        ppp = pp;
1733        while (*ppp != '\n' && *ppp != ' ') ppp++;        /* The '\r' test here is so that it works on Windows. */
1734          /* The '0' test is just in case this is an unterminated line. */
1735          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1736        *ppp = 0;        *ppp = 0;
1737        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1738          {          {
1739          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1740          goto SKIP_DATA;          goto SKIP_DATA;
1741          }          }
1742          locale_set = 1;
1743        tables = pcre_maketables();        tables = pcre_maketables();
1744        pp = ppp;        pp = ppp;
1745        break;        break;
# Line 849  while (!done) Line 1751  while (!done)
1751        *pp = 0;        *pp = 0;
1752        break;        break;
1753    
1754        case '\n': case ' ': break;        case '<':
1755            {
1756            if (strncmpic(pp, (uschar *)"JS>", 3) == 0)
1757              {
1758              options |= PCRE_JAVASCRIPT_COMPAT;
1759              pp += 3;
1760              }
1761            else
1762              {
1763              int x = check_newline(pp, outfile);
1764              if (x == 0) goto SKIP_DATA;
1765              options |= x;
1766              while (*pp++ != '>');
1767              }
1768            }
1769          break;
1770    
1771          case '\r':                      /* So that it works in Windows */
1772          case '\n':
1773          case ' ':
1774          break;
1775    
1776        default:        default:
1777        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);
# Line 869  while (!done) Line 1791  while (!done)
1791    
1792      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
1793      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
1794        if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1795        if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1796        if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1797        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
1798        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1799    
1800      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1801    
1802      /* 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 1804  while (!done)
1804    
1805      if (rc != 0)      if (rc != 0)
1806        {        {
1807        (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);        (void)regerror(rc, &preg, (char *)buffer, buffer_size);
1808        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
1809        goto SKIP_DATA;        goto SKIP_DATA;
1810        }        }
# Line 888  while (!done) Line 1816  while (!done)
1816  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1817    
1818      {      {
1819      if (timeit)      unsigned long int get_options;
1820    
1821        if (timeit > 0)
1822        {        {
1823        register int i;        register int i;
1824        clock_t time_taken;        clock_t time_taken;
1825        clock_t start_time = clock();        clock_t start_time = clock();
1826        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1827          {          {
1828          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1829          if (re != NULL) free(re);          if (re != NULL) free(re);
1830          }          }
1831        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1832        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1833          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1834            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1835        }        }
1836    
# Line 917  while (!done) Line 1847  while (!done)
1847          {          {
1848          for (;;)          for (;;)
1849            {            {
1850            if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1851              {              {
1852              done = 1;              done = 1;
1853              goto CONTINUE;              goto CONTINUE;
# Line 931  while (!done) Line 1861  while (!done)
1861        goto CONTINUE;        goto CONTINUE;
1862        }        }
1863    
1864      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1865      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
1866      returns only limited data. Check that it agrees with the newer one. */      lines. */
1867    
1868        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1869        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1870    
1871        /* Print information if required. There are now two info-returning
1872        functions. The old one has a limited interface and returns only limited
1873        data. Check that it agrees with the newer one. */
1874    
1875      if (log_store)      if (log_store)
1876        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 947  while (!done) Line 1884  while (!done)
1884      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
1885      regex_gotten_store = gotten_store;      regex_gotten_store = gotten_store;
1886    
1887      /* 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
1888      help with the matching. */      help with the matching, unless the pattern has the SS option, which
1889        suppresses the effect of /S (used for a few test patterns where studying is
1890        never sensible). */
1891    
1892      if (do_study)      if (do_study || (force_study >= 0 && !no_force_study))
1893        {        {
1894        if (timeit)        if (timeit > 0)
1895          {          {
1896          register int i;          register int i;
1897          clock_t time_taken;          clock_t time_taken;
1898          clock_t start_time = clock();          clock_t start_time = clock();
1899          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1900            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options | force_study_options, &error);
1901          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1902          if (extra != NULL) free(extra);          if (extra != NULL) pcre_free_study(extra);
1903          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1904            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1905              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1906          }          }
1907        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options | force_study_options, &error);
1908        if (error != NULL)        if (error != NULL)
1909          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
1910        else if (extra != NULL)        else if (extra != NULL)
1911          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
1912        }        }
1913    
1914        /* If /K was present, we set up for handling MARK data. */
1915    
1916        if (do_mark)
1917          {
1918          if (extra == NULL)
1919            {
1920            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1921            extra->flags = 0;
1922            }
1923          extra->mark = &markptr;
1924          extra->flags |= PCRE_EXTRA_MARK;
1925          }
1926    
1927      /* 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
1928      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
1929      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 1932  while (!done)
1932      if (do_flip)      if (do_flip)
1933        {        {
1934        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1935        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1936            byteflip(rre->magic_number, sizeof(rre->magic_number));
1937        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1938        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1939        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1940        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1941        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1942        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1943        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1944          rre->first_byte =
1945            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1946          rre->req_byte =
1947            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1948          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1949          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1950        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1951          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1952        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1953            sizeof(rre->name_count));
1954    
1955        if (extra != NULL)        if (extra != NULL)
1956          {          {
1957          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1958          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1959          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1960            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1961          }          }
1962        }        }
1963    
# Line 1005  while (!done) Line 1965  while (!done)
1965    
1966      SHOW_INFO:      SHOW_INFO:
1967    
1968        if (do_debug)
1969          {
1970          fprintf(outfile, "------------------------------------------------------------------\n");
1971          pcre_printint(re, outfile, debug_lengths);
1972          }
1973    
1974        /* We already have the options in get_options (see above) */
1975    
1976      if (do_showinfo)      if (do_showinfo)
1977        {        {
1978        unsigned long int get_options, all_options;        unsigned long int all_options;
1979    #if !defined NOINFOCHECK
1980        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1981        int count, backrefmax, first_char, need_char;  #endif
1982          int count, backrefmax, first_char, need_char, okpartial, jchanged,
1983            hascrorlf;
1984        int nameentrysize, namecount;        int nameentrysize, namecount;
1985        const uschar *nametable;        const uschar *nametable;
1986    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         print_internals(re, outfile);  
         }  
   
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1987        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1988        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1989        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1028  while (!done) Line 1992  while (!done)
1992        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1993        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1994        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1995          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1996          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1997          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1998    
1999    #if !defined NOINFOCHECK
2000        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
2001        if (count < 0) fprintf(outfile,        if (count < 0) fprintf(outfile,
2002          "Error %d from pcre_info()\n", count);          "Error %d from pcre_info()\n", count);
# Line 1046  while (!done) Line 2014  while (!done)
2014            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
2015              get_options, old_options);              get_options, old_options);
2016          }          }
2017    #endif
2018    
2019        if (size != regex_gotten_store) fprintf(outfile,        if (size != regex_gotten_store) fprintf(outfile,
2020          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
2021          size, regex_gotten_store);          (int)size, (int)regex_gotten_store);
2022    
2023        fprintf(outfile, "Capturing subpattern count = %d\n", count);        fprintf(outfile, "Capturing subpattern count = %d\n", count);
2024        if (backrefmax > 0)        if (backrefmax > 0)
# Line 1067  while (!done) Line 2036  while (!done)
2036            }            }
2037          }          }
2038    
2039        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
2040        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
2041    
2042        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
2043        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");  
2044    
2045        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
2046          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",
2047            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
2048            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
2049            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
2050            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
2051              ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
2052            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
2053              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
2054              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
2055            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
2056            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
2057            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
2058              ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
2059            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
2060            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");            ((get_options & PCRE_UCP) != 0)? " ucp" : "",
2061              ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
2062              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
2063              ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
2064    
2065          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
2066    
2067          switch (get_options & PCRE_NEWLINE_BITS)
2068            {
2069            case PCRE_NEWLINE_CR:
2070            fprintf(outfile, "Forced newline sequence: CR\n");
2071            break;
2072    
2073            case PCRE_NEWLINE_LF:
2074            fprintf(outfile, "Forced newline sequence: LF\n");
2075            break;
2076    
2077        if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)          case PCRE_NEWLINE_CRLF:
2078          fprintf(outfile, "Case state changes\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
2079            break;
2080    
2081            case PCRE_NEWLINE_ANYCRLF:
2082            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
2083            break;
2084    
2085            case PCRE_NEWLINE_ANY:
2086            fprintf(outfile, "Forced newline sequence: ANY\n");
2087            break;
2088    
2089            default:
2090            break;
2091            }
2092    
2093        if (first_char == -1)        if (first_char == -1)
2094          {          {
2095          fprintf(outfile, "First char at start or follows \\n\n");          fprintf(outfile, "First char at start or follows newline\n");
2096          }          }
2097        else if (first_char < 0)        else if (first_char < 0)
2098          {          {
# Line 1108  while (!done) Line 2103  while (!done)
2103          int ch = first_char & 255;          int ch = first_char & 255;
2104          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
2105            "" : " (caseless)";            "" : " (caseless)";
2106          if (isprint(ch))          if (PRINTHEX(ch))
2107            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
2108          else          else
2109            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1123  while (!done) Line 2118  while (!done)
2118          int ch = need_char & 255;          int ch = need_char & 255;
2119          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
2120            "" : " (caseless)";            "" : " (caseless)";
2121          if (isprint(ch))          if (PRINTHEX(ch))
2122            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
2123          else          else
2124            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1132  while (!done) Line 2127  while (!done)
2127        /* 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
2128        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
2129        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
2130        flipped.) */        flipped.) If study was forced by an external -s, don't show this
2131          information unless -i or -d was also present. This means that, except
2132          when auto-callouts are involved, the output from runs with and without
2133          -s should be identical. */
2134    
2135        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
2136          {          {
2137          if (extra == NULL)          if (extra == NULL)
2138            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
2139          else          else
2140            {            {
2141            uschar *start_bits = NULL;            uschar *start_bits = NULL;
2142            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
2143    
2144              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
2145              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
2146    
2147              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
2148            if (start_bits == NULL)            if (start_bits == NULL)
2149              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
2150            else            else
2151              {              {
2152              int i;              int i;
# Line 1159  while (!done) Line 2161  while (!done)
2161                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
2162                    c = 2;                    c = 2;
2163                    }                    }
2164                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
2165                    {                    {
2166                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
2167                    c += 2;                    c += 2;
# Line 1174  while (!done) Line 2176  while (!done)
2176              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2177              }              }
2178            }            }
2179    
2180            /* Show this only if the JIT was set by /S, not by -s. */
2181    
2182            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
2183              {
2184              int jit;
2185              new_info(re, extra, PCRE_INFO_JIT, &jit);
2186              if (jit)
2187                fprintf(outfile, "JIT study was successful\n");
2188              else
2189    #ifdef SUPPORT_JIT
2190                fprintf(outfile, "JIT study was not successful\n");
2191    #else
2192                fprintf(outfile, "JIT support is not available in this version of PCRE\n");
2193    #endif
2194              }
2195          }          }
2196        }        }
2197    
# Line 1191  while (!done) Line 2209  while (!done)
2209        else        else
2210          {          {
2211          uschar sbuf[8];          uschar sbuf[8];
2212          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
2213          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
2214          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
2215          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
2216    
2217          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
2218          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
2219          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
2220          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
2221    
2222          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
2223              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1208  while (!done) Line 2226  while (!done)
2226            }            }
2227          else          else
2228            {            {
2229            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
2230    
2231              /* If there is study data, write it. */
2232    
2233            if (extra != NULL)            if (extra != NULL)
2234              {              {
2235              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 2243  while (!done)
2243            }            }
2244          fclose(f);          fclose(f);
2245          }          }
2246    
2247          new_free(re);
2248          if (extra != NULL) pcre_free_study(extra);
2249          if (locale_set)
2250            {
2251            new_free((void *)tables);
2252            setlocale(LC_CTYPE, "C");
2253            locale_set = 0;
2254            }
2255        continue;  /* With next regex */        continue;  /* With next regex */
2256        }        }
2257      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1230  while (!done) Line 2260  while (!done)
2260    
2261    for (;;)    for (;;)
2262      {      {
2263      unsigned char *q;      uschar *q;
2264      unsigned char *bptr = dbuffer;      uschar *bptr;
2265      int *use_offsets = offsets;      int *use_offsets = offsets;
2266      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
2267      int callout_data = 0;      int callout_data = 0;
2268      int callout_data_set = 0;      int callout_data_set = 0;
2269      int count, c;      int count, c;
2270      int copystrings = 0;      int copystrings = 0;
2271      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
2272      int getstrings = 0;      int getstrings = 0;
2273      int getlist = 0;      int getlist = 0;
2274      int gmatched = 0;      int gmatched = 0;
2275      int start_offset = 0;      int start_offset = 0;
2276        int start_offset_sign = 1;
2277      int g_notempty = 0;      int g_notempty = 0;
2278        int use_dfa = 0;
2279    
2280      options = 0;      options = 0;
2281    
2282        *copynames = 0;
2283        *getnames = 0;
2284    
2285        copynamesptr = copynames;
2286        getnamesptr = getnames;
2287    
2288      pcre_callout = callout;      pcre_callout = callout;
2289      first_callout = 1;      first_callout = 1;
2290        last_callout_mark = NULL;
2291      callout_extra = 0;      callout_extra = 0;
2292      callout_count = 0;      callout_count = 0;
2293      callout_fail_count = 999999;      callout_fail_count = 999999;
2294      callout_fail_id = -1;      callout_fail_id = -1;
2295      show_malloc = 0;      show_malloc = 0;
2296    
2297      if (infile == stdin) printf("data> ");      if (extra != NULL) extra->flags &=
2298      if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
2299    
2300        len = 0;
2301        for (;;)
2302        {        {
2303        done = 1;        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
2304        goto CONTINUE;          {
2305            if (len > 0)    /* Reached EOF without hitting a newline */
2306              {
2307              fprintf(outfile, "\n");
2308              break;
2309              }
2310            done = 1;
2311            goto CONTINUE;
2312            }
2313          if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
2314          len = (int)strlen((char *)buffer);
2315          if (buffer[len-1] == '\n') break;
2316        }        }
     if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);  
2317    
     len = (int)strlen((char *)buffer);  
2318      while (len > 0 && isspace(buffer[len-1])) len--;      while (len > 0 && isspace(buffer[len-1])) len--;
2319      buffer[len] = 0;      buffer[len] = 0;
2320      if (len == 0) break;      if (len == 0) break;
# Line 1271  while (!done) Line 2322  while (!done)
2322      p = buffer;      p = buffer;
2323      while (isspace(*p)) p++;      while (isspace(*p)) p++;
2324    
2325      q = dbuffer;      bptr = q = dbuffer;
2326      while ((c = *p++) != 0)      while ((c = *p++) != 0)
2327        {        {
2328        int i = 0;        int i = 0;
# Line 1293  while (!done) Line 2344  while (!done)
2344          c -= '0';          c -= '0';
2345          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
2346            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
2347    
2348    #if !defined NOUTF8
2349            if (use_utf8 && c > 255)
2350              {
2351              unsigned char buff8[8];
2352              int ii, utn;
2353              utn = ord2utf8(c, buff8);
2354              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2355              c = buff8[ii];   /* Last byte */
2356              }
2357    #endif
2358          break;          break;
2359    
2360          case 'x':          case 'x':
2361    
2362          /* Handle \x{..} specially - new Perl thing for utf8 */          /* Handle \x{..} specially - new Perl thing for utf8 */
2363    
2364    #if !defined NOUTF8
2365          if (*p == '{')          if (*p == '{')
2366            {            {
2367            unsigned char *pt = p;            unsigned char *pt = p;
# Line 1309  while (!done) Line 2372  while (!done)
2372              {              {
2373              unsigned char buff8[8];              unsigned char buff8[8];
2374              int ii, utn;              int ii, utn;
2375              utn = ord2utf8(c, buff8);              if (use_utf8)
2376              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
2377              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
2378                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
2379                  c = buff8[ii];   /* Last byte */
2380                  }
2381                else
2382                 {
2383                 if (c > 255)
2384                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
2385                     "UTF-8 mode is not enabled.\n"
2386                     "** Truncation will probably give the wrong result.\n", c);
2387                 }
2388              p = pt + 1;              p = pt + 1;
2389              break;              break;
2390              }              }
2391            /* Not correct form; fall through */            /* Not correct form; fall through */
2392            }            }
2393    #endif
2394    
2395          /* Ordinary \x */          /* Ordinary \x */
2396    
# Line 1333  while (!done) Line 2407  while (!done)
2407          continue;          continue;
2408    
2409          case '>':          case '>':
2410            if (*p == '-')
2411              {
2412              start_offset_sign = -1;
2413              p++;
2414              }
2415          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
2416            start_offset *= start_offset_sign;
2417          continue;          continue;
2418    
2419          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1352  while (!done) Line 2432  while (!done)
2432            }            }
2433          else if (isalnum(*p))          else if (isalnum(*p))
2434            {            {
2435            uschar name[256];            uschar *npp = copynamesptr;
           uschar *npp = name;  
2436            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2437              *npp++ = 0;
2438            *npp = 0;            *npp = 0;
2439            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)copynamesptr);
2440            if (n < 0)            if (n < 0)
2441              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
2442            else copystrings |= 1 << n;            copynamesptr = npp;
2443            }            }
2444          else if (*p == '+')          else if (*p == '+')
2445            {            {
# Line 1397  while (!done) Line 2477  while (!done)
2477            }            }
2478          continue;          continue;
2479    
2480    #if !defined NODFA
2481            case 'D':
2482    #if !defined NOPOSIX
2483            if (posix || do_posix)
2484              printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");
2485            else
2486    #endif
2487              use_dfa = 1;
2488            continue;
2489    #endif
2490    
2491    #if !defined NODFA
2492            case 'F':
2493            options |= PCRE_DFA_SHORTEST;
2494            continue;
2495    #endif
2496    
2497          case 'G':          case 'G':
2498          if (isdigit(*p))          if (isdigit(*p))
2499            {            {
# Line 1405  while (!done) Line 2502  while (!done)
2502            }            }
2503          else if (isalnum(*p))          else if (isalnum(*p))
2504            {            {
2505            uschar name[256];            uschar *npp = getnamesptr;
           uschar *npp = name;  
2506            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
2507              *npp++ = 0;
2508            *npp = 0;            *npp = 0;
2509            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)getnamesptr);
2510            if (n < 0)            if (n < 0)
2511              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
2512            else getstrings |= 1 << n;            getnamesptr = npp;
2513            }            }
2514          continue;          continue;
2515    
2516            case 'J':
2517            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2518            if (extra != NULL
2519                && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
2520                && extra->executable_jit != NULL)
2521              {
2522              if (jit_stack != NULL) pcre_jit_stack_free(jit_stack);
2523              jit_stack = pcre_jit_stack_alloc(1, n * 1024);
2524              pcre_assign_jit_callback(extra, jit_callback, jit_stack);
2525              }
2526            continue;
2527    
2528          case 'L':          case 'L':
2529          getlist = 1;          getlist = 1;
# Line 1425  while (!done) Line 2534  while (!done)
2534          continue;          continue;
2535    
2536          case 'N':          case 'N':
2537          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
2538              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
2539            else
2540              options |= PCRE_NOTEMPTY;
2541          continue;          continue;
2542    
2543          case 'O':          case 'O':
# Line 1438  while (!done) Line 2550  while (!done)
2550            if (offsets == NULL)            if (offsets == NULL)
2551              {              {
2552              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
2553                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
2554              return 1;              yield = 1;
2555                goto EXIT;
2556              }              }
2557            }            }
2558          use_size_offsets = n;          use_size_offsets = n;
# Line 1447  while (!done) Line 2560  while (!done)
2560          continue;          continue;
2561    
2562          case 'P':          case 'P':
2563          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2564              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2565          continue;          continue;
2566    
2567            case 'Q':
2568            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2569            if (extra == NULL)
2570              {
2571              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2572              extra->flags = 0;
2573              }
2574            extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
2575            extra->match_limit_recursion = n;
2576            continue;
2577    
2578            case 'q':
2579            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2580            if (extra == NULL)
2581              {
2582              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2583              extra->flags = 0;
2584              }
2585            extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
2586            extra->match_limit = n;
2587            continue;
2588    
2589    #if !defined NODFA
2590            case 'R':
2591            options |= PCRE_DFA_RESTART;
2592            continue;
2593    #endif
2594    
2595          case 'S':          case 'S':
2596          show_malloc = 1;          show_malloc = 1;
2597          continue;          continue;
2598    
2599            case 'Y':
2600            options |= PCRE_NO_START_OPTIMIZE;
2601            continue;
2602    
2603          case 'Z':          case 'Z':
2604          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2605          continue;          continue;
# Line 1461  while (!done) Line 2607  while (!done)
2607          case '?':          case '?':
2608          options |= PCRE_NO_UTF8_CHECK;          options |= PCRE_NO_UTF8_CHECK;
2609          continue;          continue;
2610    
2611            case '<':
2612              {
2613              int x = check_newline(p, outfile);
2614              if (x == 0) goto NEXT_DATA;
2615              options |= x;
2616              while (*p++ != '>');
2617              }
2618            continue;
2619          }          }
2620        *q++ = c;        *q++ = c;
2621        }        }
2622      *q = 0;      *q = 0;
2623      len = q - dbuffer;      len = (int)(q - dbuffer);
2624    
2625        /* Move the data to the end of the buffer so that a read over the end of
2626        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2627        we are using the POSIX interface, we must include the terminating zero. */
2628    
2629    #if !defined NOPOSIX
2630        if (posix || do_posix)
2631          {
2632          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2633          bptr += buffer_size - len - 1;
2634          }
2635        else
2636    #endif
2637          {
2638          memmove(bptr + buffer_size - len, bptr, len);
2639          bptr += buffer_size - len;
2640          }
2641    
2642        if ((all_use_dfa || use_dfa) && find_match_limit)
2643          {
2644          printf("**Match limit not relevant for DFA matching: ignored\n");
2645          find_match_limit = 0;
2646          }
2647    
2648      /* Handle matching via the POSIX interface, which does not      /* Handle matching via the POSIX interface, which does not
2649      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 2658  while (!done)
2658          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2659        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2660        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2661          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2662    
2663        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2664    
2665        if (rc != 0)        if (rc != 0)
2666          {          {
2667          (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);          (void)regerror(rc, &preg, (char *)buffer, buffer_size);
2668          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
2669          }          }
2670          else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
2671                  != 0)
2672            {
2673            fprintf(outfile, "Matched with REG_NOSUB\n");
2674            }
2675        else        else
2676          {          {
2677          size_t i;          size_t i;
# Line 1499  while (!done) Line 2683  while (!done)
2683              (void)pchars(dbuffer + pmatch[i].rm_so,              (void)pchars(dbuffer + pmatch[i].rm_so,
2684                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
2685              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2686              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
2687                {                {
2688                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
2689                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
2690                  outfile);                  outfile);
2691                fprintf(outfile, "\n");                fprintf(outfile, "\n");
# Line 1519  while (!done) Line 2703  while (!done)
2703    
2704      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2705        {        {
2706        if (timeit)        markptr = NULL;
2707    
2708          if (timeitm > 0)
2709          {          {
2710          register int i;          register int i;
2711          clock_t time_taken;          clock_t time_taken;
2712          clock_t start_time = clock();          clock_t start_time = clock();
2713          for (i = 0; i < LOOPREPEAT; i++)  
2714    #if !defined NODFA
2715            if (all_use_dfa || use_dfa)
2716              {
2717              int workspace[1000];
2718              for (i = 0; i < timeitm; i++)
2719                count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2720                  options | g_notempty, use_offsets, use_size_offsets, workspace,
2721                  sizeof(workspace)/sizeof(int));
2722              }
2723            else
2724    #endif
2725    
2726            for (i = 0; i < timeitm; i++)
2727            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2728              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2729    
2730          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2731          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2732            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2733              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2734          }          }
2735    
2736        /* 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
2737        varying limits in order to find the minimum value. */        varying limits in order to find the minimum value for the match limit and
2738          for the recursion limit. The match limits are relevant only to the normal
2739          running of pcre_exec(), so disable the JIT optimization. This makes it
2740          possible to run the same set of tests with and without JIT externally
2741          requested. */
2742    
2743        if (find_match_limit)        if (find_match_limit)
2744          {          {
         int min = 0;  
         int mid = 64;  
         int max = -1;  
   
2745          if (extra == NULL)          if (extra == NULL)
2746            {            {
2747            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2748            extra->flags = 0;            extra->flags = 0;
2749            }            }
2750          extra->flags |= PCRE_EXTRA_MATCH_LIMIT;          else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
2751    
2752          for (;;)          (void)check_match_limit(re, extra, bptr, len, start_offset,
2753            {            options|g_notempty, use_offsets, use_size_offsets,
2754            extra->match_limit = mid;            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
2755            count = pcre_exec(re, extra, (char *)bptr, len, start_offset,            PCRE_ERROR_MATCHLIMIT, "match()");
2756              options | g_notempty, use_offsets, use_size_offsets);  
2757            if (count == PCRE_ERROR_MATCHLIMIT)          count = check_match_limit(re, extra, bptr, len, start_offset,
2758              {            options|g_notempty, use_offsets, use_size_offsets,
2759              /* fprintf(outfile, "Testing match limit = %d\n", mid); */            PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
2760              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;  
2761          }          }
2762    
2763        /* 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 2779  while (!done)
2779        /* 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
2780        value of match_limit. */        value of match_limit. */
2781    
2782        else  #if !defined NODFA
2783          else if (all_use_dfa || use_dfa)
2784          {          {
2785          count = pcre_exec(re, extra, (char *)bptr, len,          int workspace[1000];
2786            start_offset, options | g_notempty, use_offsets, use_size_offsets);          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2787              options | g_notempty, use_offsets, use_size_offsets, workspace,
2788              sizeof(workspace)/sizeof(int));
2789            if (count == 0)
2790              {
2791              fprintf(outfile, "Matched, but too many subsidiary matches\n");
2792              count = use_size_offsets/2;
2793              }
2794          }          }
2795    #endif
2796    
2797        if (count == 0)        else
2798          {          {
2799          fprintf(outfile, "Matched, but too many substrings\n");          count = pcre_exec(re, extra, (char *)bptr, len,
2800          count = use_size_offsets/3;            start_offset, options | g_notempty, use_offsets, use_size_offsets);
2801            if (count == 0)
2802              {
2803              fprintf(outfile, "Matched, but too many substrings\n");
2804              count = use_size_offsets/3;
2805              }
2806          }          }
2807    
2808        /* Matched */        /* Matched */
2809    
2810        if (count >= 0)        if (count >= 0)
2811          {          {
2812          int i;          int i, maxcount;
2813    
2814    #if !defined NODFA
2815            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2816    #endif
2817              maxcount = use_size_offsets/3;
2818    
2819            /* This is a check against a lunatic return value. */
2820    
2821            if (count > maxcount)
2822              {
2823              fprintf(outfile,
2824                "** PCRE error: returned count %d is too big for offset size %d\n",
2825                count, use_size_offsets);
2826              count = use_size_offsets/3;
2827              if (do_g || do_G)
2828                {
2829                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2830                do_g = do_G = FALSE;        /* Break g/G loop */
2831                }
2832              }
2833    
2834            /* do_allcaps requests showing of all captures in the pattern, to check
2835            unset ones at the end. */
2836    
2837            if (do_allcaps)
2838              {
2839              new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
2840              count++;   /* Allow for full match */
2841              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
2842              }
2843    
2844            /* Output the captured substrings */
2845    
2846          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2847            {            {
2848            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
2849                {
2850                if (use_offsets[i] != -1)
2851                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2852                    use_offsets[i], i);
2853                if (use_offsets[i+1] != -1)
2854                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
2855                    use_offsets[i+1], i+1);
2856              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
2857                }
2858            else            else
2859              {              {
2860              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
2861              (void)pchars(bptr + use_offsets[i],              (void)pchars(bptr + use_offsets[i],
2862                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
2863              fprintf(outfile, "\n");              fprintf(outfile, "\n");
2864              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
2865                {                {
2866                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
2867                  {                (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],
2868                  fprintf(outfile, " 0+ ");                  outfile);
2869                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
2870                }                }
2871              }              }
2872            }            }
2873    
2874            if (markptr != NULL) fprintf(outfile, "MK: %s\n", markptr);
2875    
2876          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2877            {            {
2878            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
2879              {              {
2880              char copybuffer[16];              char copybuffer[256];
2881              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,
2882                i, copybuffer, sizeof(copybuffer));                i, copybuffer, sizeof(copybuffer));
2883              if (rc < 0)              if (rc < 0)
# Line 1651  while (!done) Line 2887  while (!done)
2887              }              }
2888            }            }
2889    
2890            for (copynamesptr = copynames;
2891                 *copynamesptr != 0;
2892                 copynamesptr += (int)strlen((char*)copynamesptr) + 1)
2893              {
2894              char copybuffer[256];
2895              int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,
2896                count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));
2897              if (rc < 0)
2898                fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);
2899              else
2900                fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);
2901              }
2902    
2903          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2904            {            {
2905            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
# Line 1663  while (!done) Line 2912  while (!done)
2912              else              else
2913                {                {
2914                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
               /* free((void *)substring); */  
2915                pcre_free_substring(substring);                pcre_free_substring(substring);
2916                }                }
2917              }              }
2918            }            }
2919    
2920            for (getnamesptr = getnames;
2921                 *getnamesptr != 0;
2922                 getnamesptr += (int)strlen((char*)getnamesptr) + 1)
2923              {
2924              const char *substring;
2925              int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,
2926                count, (char *)getnamesptr, &substring);
2927              if (rc < 0)
2928                fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);
2929              else
2930                {
2931                fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);
2932                pcre_free_substring(substring);
2933                }
2934              }
2935    
2936          if (getlist)          if (getlist)
2937            {            {
2938            const char **stringlist;            const char **stringlist;
# Line 1682  while (!done) Line 2946  while (!done)
2946                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);
2947              if (stringlist[i] != NULL)              if (stringlist[i] != NULL)
2948                fprintf(outfile, "string list not terminated by NULL\n");                fprintf(outfile, "string list not terminated by NULL\n");
             /* free((void *)stringlist); */  
2949              pcre_free_substring_list(stringlist);              pcre_free_substring_list(stringlist);
2950              }              }
2951            }            }
# Line 1692  while (!done) Line 2955  while (!done)
2955    
2956        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2957          {          {
2958          fprintf(outfile, "Partial match\n");          if (markptr == NULL) fprintf(outfile, "Partial match");
2959              else fprintf(outfile, "Partial match, mark=%s", markptr);
2960            if (use_size_offsets > 1)
2961              {
2962              fprintf(outfile, ": ");
2963              pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2964                outfile);
2965              }
2966            fprintf(outfile, "\n");
2967          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2968          }          }
2969    
2970        /* 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
2971        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
2972        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
2973        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2974        offset values to achieve this. We won't be at the end of the string -  
2975        that was checked before setting g_notempty. */        Complication arises in the case when the newline convention is "any",
2976          "crlf", or "anycrlf". If the previous match was at the end of a line
2977          terminated by CRLF, an advance of one character just passes the \r,
2978          whereas we should prefer the longer newline sequence, as does the code in
2979          pcre_exec(). Fudge the offset value to achieve this. We check for a
2980          newline setting in the pattern; if none was set, use pcre_config() to
2981          find the default.
2982    
2983          Otherwise, in the case of UTF-8 matching, the advance must be one
2984          character, not one byte. */
2985    
2986        else        else
2987          {          {
2988          if (g_notempty != 0)          if (g_notempty != 0)
2989            {            {
2990            int onechar = 1;            int onechar = 1;
2991              unsigned int obits = ((real_pcre *)re)->options;
2992            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2993            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2994                {
2995                int d;
2996                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2997                /* Note that these values are always the ASCII ones, even in
2998                EBCDIC environments. CR = 13, NL = 10. */
2999                obits = (d == 13)? PCRE_NEWLINE_CR :
3000                        (d == 10)? PCRE_NEWLINE_LF :
3001                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
3002                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
3003                        (d == -1)? PCRE_NEWLINE_ANY : 0;
3004                }
3005              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
3006                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_CRLF ||
3007                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
3008                  &&
3009                  start_offset < len - 1 &&
3010                  bptr[start_offset] == '\r' &&
3011                  bptr[start_offset+1] == '\n')
3012                onechar++;
3013              else if (use_utf8)
3014              {              {
3015              while (start_offset + onechar < len)              while (start_offset + onechar < len)
3016                {                {
3017                int tb = bptr[start_offset+onechar];                if ((bptr[start_offset+onechar] & 0xc0) != 0x80) break;
3018                if (tb <= 127) break;                onechar++;
               tb &= 0xc0;  
               if (tb != 0 && tb != 0xc0) onechar++;  
3019                }                }
3020              }              }
3021            use_offsets[1] = start_offset + onechar;            use_offsets[1] = start_offset + onechar;
3022            }            }
3023          else          else
3024            {            {
3025            if (count == PCRE_ERROR_NOMATCH)            switch(count)
3026              {              {
3027              if (gmatched == 0) fprintf(outfile, "No match\n");              case PCRE_ERROR_NOMATCH:
3028                if (gmatched == 0)
3029                  {
3030                  if (markptr == NULL) fprintf(outfile, "No match\n");
3031                    else fprintf(outfile, "No match, mark = %s\n", markptr);
3032                  }
3033                break;
3034    
3035                case PCRE_ERROR_BADUTF8:
3036                case PCRE_ERROR_SHORTUTF8:
3037                fprintf(outfile, "Error %d (%s UTF-8 string)", count,
3038                  (count == PCRE_ERROR_BADUTF8)? "bad" : "short");
3039                if (use_size_offsets >= 2)
3040                  fprintf(outfile, " offset=%d reason=%d", use_offsets[0],
3041                    use_offsets[1]);
3042                fprintf(outfile, "\n");
3043                break;
3044    
3045                default:
3046                if (count < 0 && (-count) < sizeof(errtexts)/sizeof(const char *))
3047                  fprintf(outfile, "Error %d (%s)\n", count, errtexts[-count]);
3048                else
3049                  fprintf(outfile, "Error %d (Unexpected value)\n", count);
3050                break;
3051              }              }
3052            else fprintf(outfile, "Error %d\n", count);  
3053            break;  /* Out of the /g loop */            break;  /* Out of the /g loop */
3054            }            }
3055          }          }
# Line 1737  while (!done) Line 3059  while (!done)
3059        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
3060    
3061        /* 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
3062        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
3063        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
3064        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
3065        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
3066        character. */        character. */
3067    
3068        g_notempty = 0;        g_notempty = 0;
3069    
3070        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
3071          {          {
3072          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
3073          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
3074          }          }
3075    
3076        /* 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 3085  while (!done)
3085          len -= use_offsets[1];          len -= use_offsets[1];
3086          }          }
3087        }  /* End of loop for /g and /G */        }  /* End of loop for /g and /G */
3088    
3089        NEXT_DATA: continue;
3090      }    /* End of loop for data lines */      }    /* End of loop for data lines */
3091    
3092    CONTINUE:    CONTINUE:
# Line 1770  while (!done) Line 3095  while (!done)
3095    if (posix || do_posix) regfree(&preg);    if (posix || do_posix) regfree(&preg);
3096  #endif  #endif
3097    
3098    if (re != NULL) free(re);    if (re != NULL) new_free(re);
3099    if (extra != NULL) free(extra);    if (extra != NULL) pcre_free_study(extra);
3100    if (tables != NULL)    if (locale_set)
3101      {      {
3102      free((void *)tables);      new_free((void *)tables);
3103      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
3104        locale_set = 0;
3105      }      }
3106      if (jit_stack != NULL)
3107        {
3108        pcre_jit_stack_free(jit_stack);
3109        jit_stack = NULL;
3110        }
3111    }    }
3112    
3113  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");
3114  return 0;  
3115    EXIT:
3116    
3117    if (infile != NULL && infile != stdin) fclose(infile);
3118    if (outfile != NULL && outfile != stdout) fclose(outfile);
3119    
3120    free(buffer);
3121    free(dbuffer);
3122    free(pbuffer);
3123    free(offsets);
3124    
3125    return yield;
3126  }  }
3127    
3128  /* End */  /* End of pcretest.c */

Legend:
Removed from v.75  
changed lines
  Added in v.674

  ViewVC Help
Powered by ViewVC 1.1.5