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

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

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

revision 75 by nigel, Sat Feb 24 21:40:37 2007 UTC revision 255 by ph10, Wed Sep 19 08:50:04 2007 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    
 /* We need the internal info for displaying the results of pcre_study(). Also  
 for getting the opcodes for showing compiled code. */  
51    
52  #define PCRE_SPY        /* For Win32 build, import data, not export */  /* A number of things vary for Windows builds. Originally, pcretest opened its
53  #include "internal.h"  input and output without "b"; then I was told that "b" was needed in some
54    environments, so it was added for release 5.0 to both the input and output. (It
55    makes no difference on Unix-like systems.) Later I was told that it is wrong
56    for the input on Windows. I've now abstracted the modes into two macros that
57    are set here, to make it easier to fiddle with them, and removed "b" from the
58    input mode under Windows. */
59    
60    #if defined(_WIN32) || defined(WIN32)
61    #include <io.h>                /* For _setmode() */
62    #include <fcntl.h>             /* For _O_BINARY */
63    #define INPUT_MODE   "r"
64    #define OUTPUT_MODE  "wb"
65    
66    #else
67    #include <sys/time.h>          /* These two includes are needed */
68    #include <sys/resource.h>      /* for setrlimit(). */
69    #define INPUT_MODE   "rb"
70    #define OUTPUT_MODE  "wb"
71    #endif
72    
73    
74    /* We have to include pcre_internal.h because we need the internal info for
75    displaying the results of pcre_study() and we also need to know about the
76    internal macros, structures, and other internal data values; pcretest has
77    "inside information" compared to a program that strictly follows the PCRE API.
78    
79    Although pcre_internal.h does itself include pcre.h, we explicitly include it
80    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
81    appropriately for an application, not for building PCRE. */
82    
83    #include "pcre.h"
84    #include "pcre_internal.h"
85    
86    /* We need access to the data tables that PCRE uses. So as not to have to keep
87    two copies, we include the source file here, changing the names of the external
88    symbols to prevent clashes. */
89    
90    #define _pcre_utf8_table1      utf8_table1
91    #define _pcre_utf8_table1_size utf8_table1_size
92    #define _pcre_utf8_table2      utf8_table2
93    #define _pcre_utf8_table3      utf8_table3
94    #define _pcre_utf8_table4      utf8_table4
95    #define _pcre_utt              utt
96    #define _pcre_utt_size         utt_size
97    #define _pcre_utt_names        utt_names
98    #define _pcre_OP_lengths       OP_lengths
99    
100    #include "pcre_tables.c"
101    
102    /* We also need the pcre_printint() function for printing out compiled
103    patterns. This function is in a separate file so that it can be included in
104    pcre_compile.c when that module is compiled with debugging enabled.
105    
106    The definition of the macro PRINTABLE, which determines whether to print an
107    output character as-is or as a hex value when showing compiled patterns, is
108    contained in this file. We uses it here also, in cases when the locale has not
109    been explicitly changed, so as to get consistent output from systems that
110    differ in their output from isprint() even in the "C" locale. */
111    
112    #include "pcre_printint.src"
113    
114    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
115    
116    
117  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
118  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 122  Makefile. */
122  #include "pcreposix.h"  #include "pcreposix.h"
123  #endif  #endif
124    
125    /* It is also possible, for the benefit of the version currently imported into
126    Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
127    interface to the DFA matcher (NODFA), and without the doublecheck of the old
128    "info" function (define NOINFOCHECK). In fact, we automatically cut out the
129    UTF8 support if PCRE is built without it. */
130    
131    #ifndef SUPPORT_UTF8
132    #ifndef NOUTF8
133    #define NOUTF8
134    #endif
135    #endif
136    
137    
138    /* Other parameters */
139    
140  #ifndef CLOCKS_PER_SEC  #ifndef CLOCKS_PER_SEC
141  #ifdef CLK_TCK  #ifdef CLK_TCK
142  #define CLOCKS_PER_SEC CLK_TCK  #define CLOCKS_PER_SEC CLK_TCK
# Line 66  Makefile. */ Line 145  Makefile. */
145  #endif  #endif
146  #endif  #endif
147    
148  #define LOOPREPEAT 500000  /* This is the default loop count for timing. */
149    
150  #define BUFFER_SIZE 30000  #define LOOPREPEAT 500000
 #define PBUFFER_SIZE BUFFER_SIZE  
 #define DBUFFER_SIZE BUFFER_SIZE  
151    
152    /* Static variables */
153    
154  static FILE *outfile;  static FILE *outfile;
155  static int log_store = 0;  static int log_store = 0;
# Line 79  static int callout_count; Line 157  static int callout_count;
157  static int callout_extra;  static int callout_extra;
158  static int callout_fail_count;  static int callout_fail_count;
159  static int callout_fail_id;  static int callout_fail_id;
160    static int debug_lengths;
161  static int first_callout;  static int first_callout;
162    static int locale_set = 0;
163  static int show_malloc;  static int show_malloc;
164  static int use_utf8;  static int use_utf8;
165  static size_t gotten_store;  static size_t gotten_store;
166    
167    /* The buffers grow automatically if very long input lines are encountered. */
168    
169    static int buffer_size = 50000;
170    static uschar *buffer = NULL;
171    static uschar *dbuffer = NULL;
172  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
173    
174    
 static const int utf8_table1[] = {  
   0x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff};  
175    
176  static const int utf8_table2[] = {  /*************************************************
177    0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc};  *        Read or extend an input line            *
178    *************************************************/
179    
180  static const int utf8_table3[] = {  /* Input lines are read into buffer, but both patterns and data lines can be
181    0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};  continued over multiple input lines. In addition, if the buffer fills up, we
182    want to automatically expand it so as to be able to handle extremely large
183    lines that are needed for certain stress tests. When the input buffer is
184    expanded, the other two buffers must also be expanded likewise, and the
185    contents of pbuffer, which are a copy of the input for callouts, must be
186    preserved (for when expansion happens for a data line). This is not the most
187    optimal way of handling this, but hey, this is just a test program!
188    
189    Arguments:
190      f            the file to read
191      start        where in buffer to start (this *must* be within buffer)
192    
193    Returns:       pointer to the start of new data
194                   could be a copy of start, or could be moved
195                   NULL if no data read and EOF reached
196    */
197    
198    static uschar *
199    extend_inputline(FILE *f, uschar *start)
200    {
201    uschar *here = start;
202    
203    for (;;)
204      {
205      int rlen = buffer_size - (here - buffer);
206    
207      if (rlen > 1000)
208        {
209        int dlen;
210        if (fgets((char *)here, rlen,  f) == NULL)
211          return (here == start)? NULL : start;
212        dlen = (int)strlen((char *)here);
213        if (dlen > 0 && here[dlen - 1] == '\n') return start;
214        here += dlen;
215        }
216    
217      else
218        {
219        int new_buffer_size = 2*buffer_size;
220        uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);
221        uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
222        uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);
223    
224        if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
225          {
226          fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
227          exit(1);
228          }
229    
230        memcpy(new_buffer, buffer, buffer_size);
231        memcpy(new_pbuffer, pbuffer, buffer_size);
232    
233        buffer_size = new_buffer_size;
234    
235        start = new_buffer + (start - buffer);
236        here = new_buffer + (here - buffer);
237    
238        free(buffer);
239        free(dbuffer);
240        free(pbuffer);
241    
242        buffer = new_buffer;
243        dbuffer = new_dbuffer;
244        pbuffer = new_pbuffer;
245        }
246      }
247    
248    return NULL;  /* Control never gets here */
249    }
250    
251    
252    
 /*************************************************  
 *         Print compiled regex                   *  
 *************************************************/  
253    
 /* The code for doing this is held in a separate file that is also included in  
 pcre.c when it is compiled with the debug switch. It defines a function called  
 print_internals(), which uses a table of opcode lengths defined by the macro  
 OP_LENGTHS, whose name must be OP_lengths. It also uses a table that translates  
 Unicode property names to numbers; this is kept in a separate file. */  
   
 static uschar OP_lengths[] = { OP_LENGTHS };  
   
 #include "ucp.h"  
 #include "ucptypetable.c"  
 #include "printint.c"  
254    
255    
256    
# Line 122  static uschar OP_lengths[] = { OP_LENGTH Line 260  static uschar OP_lengths[] = { OP_LENGTH
260    
261  /* 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
262  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
263  for unpicking the -o argument, so just keep it simple.  for unpicking arguments, so just keep it simple.
264    
265  Arguments:  Arguments:
266    str           string to be converted    str           string to be converted
# Line 143  return(result); Line 281  return(result);
281    
282    
283    
 /*************************************************  
 *       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;  
 }  
   
284    
285  /*************************************************  /*************************************************
286  *            Convert UTF-8 string to value       *  *            Convert UTF-8 string to value       *
# Line 188  return i + 1; Line 290  return i + 1;
290  and returns the value of the character.  and returns the value of the character.
291    
292  Argument:  Argument:
293    buffer   a pointer to the byte vector    utf8bytes   a pointer to the byte vector
294    vptr     a pointer to an int to receive the value    vptr        a pointer to an int to receive the value
295    
296  Returns:   >  0 => the number of bytes consumed  Returns:      >  0 => the number of bytes consumed
297             -6 to 0 => malformed UTF-8 character at offset = (-return)                -6 to 0 => malformed UTF-8 character at offset = (-return)
298  */  */
299    
300    #if !defined NOUTF8
301    
302  static int  static int
303  utf82ord(unsigned char *buffer, int *vptr)  utf82ord(unsigned char *utf8bytes, int *vptr)
304  {  {
305  int c = *buffer++;  int c = *utf8bytes++;
306  int d = c;  int d = c;
307  int i, j, s;  int i, j, s;
308    
# Line 218  d = (c & utf8_table3[i]) << s; Line 322  d = (c & utf8_table3[i]) << s;
322    
323  for (j = 0; j < i; j++)  for (j = 0; j < i; j++)
324    {    {
325    c = *buffer++;    c = *utf8bytes++;
326    if ((c & 0xc0) != 0x80) return -(j+1);    if ((c & 0xc0) != 0x80) return -(j+1);
327    s -= 6;    s -= 6;
328    d |= (c & 0x3f) << s;    d |= (c & 0x3f) << s;
# Line 226  for (j = 0; j < i; j++) Line 330  for (j = 0; j < i; j++)
330    
331  /* Check that encoding was the correct unique one */  /* Check that encoding was the correct unique one */
332    
333  for (j = 0; j < sizeof(utf8_table1)/sizeof(int); j++)  for (j = 0; j < utf8_table1_size; j++)
334    if (d <= utf8_table1[j]) break;    if (d <= utf8_table1[j]) break;
335  if (j != i) return -(i+1);  if (j != i) return -(i+1);
336    
# Line 236  if (j != i) return -(i+1); Line 340  if (j != i) return -(i+1);
340  return i+1;  return i+1;
341  }  }
342    
343    #endif
344    
345    
346    
347    /*************************************************
348    *       Convert character value to UTF-8         *
349    *************************************************/
350    
351    /* This function takes an integer value in the range 0 - 0x7fffffff
352    and encodes it as a UTF-8 character in 0 to 6 bytes.
353    
354    Arguments:
355      cvalue     the character value
356      utf8bytes  pointer to buffer for result - at least 6 bytes long
357    
358    Returns:     number of characters placed in the buffer
359    */
360    
361    #if !defined NOUTF8
362    
363    static int
364    ord2utf8(int cvalue, uschar *utf8bytes)
365    {
366    register int i, j;
367    for (i = 0; i < utf8_table1_size; i++)
368      if (cvalue <= utf8_table1[i]) break;
369    utf8bytes += i;
370    for (j = i; j > 0; j--)
371     {
372     *utf8bytes-- = 0x80 | (cvalue & 0x3f);
373     cvalue >>= 6;
374     }
375    *utf8bytes = utf8_table2[i] | cvalue;
376    return i + 1;
377    }
378    
379    #endif
380    
381    
382    
383  /*************************************************  /*************************************************
# Line 248  chars without printing. */ Line 390  chars without printing. */
390    
391  static int pchars(unsigned char *p, int length, FILE *f)  static int pchars(unsigned char *p, int length, FILE *f)
392  {  {
393  int c;  int c = 0;
394  int yield = 0;  int yield = 0;
395    
396  while (length-- > 0)  while (length-- > 0)
397    {    {
398    #if !defined NOUTF8
399    if (use_utf8)    if (use_utf8)
400      {      {
401      int rc = utf82ord(p, &c);      int rc = utf82ord(p, &c);
# Line 261  while (length-- > 0) Line 404  while (length-- > 0)
404        {        {
405        length -= rc - 1;        length -= rc - 1;
406        p += rc;        p += rc;
407        if (c < 256 && isprint(c))        if (PRINTHEX(c))
408          {          {
409          if (f != NULL) fprintf(f, "%c", c);          if (f != NULL) fprintf(f, "%c", c);
410          yield++;          yield++;
411          }          }
412        else        else
413          {          {
414          int n;          int n = 4;
415          if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);          if (f != NULL) fprintf(f, "\\x{%02x}", c);
416          yield += n;          yield += (n <= 0x000000ff)? 2 :
417                     (n <= 0x00000fff)? 3 :
418                     (n <= 0x0000ffff)? 4 :
419                     (n <= 0x000fffff)? 5 : 6;
420          }          }
421        continue;        continue;
422        }        }
423      }      }
424    #endif
425    
426     /* Not UTF-8, or malformed UTF-8  */     /* Not UTF-8, or malformed UTF-8  */
427    
428    if (isprint(c = *(p++)))    c = *p++;
429      if (PRINTHEX(c))
430      {      {
431      if (f != NULL) fprintf(f, "%c", c);      if (f != NULL) fprintf(f, "%c", c);
432      yield++;      yield++;
# Line 403  static void *new_malloc(size_t size) Line 551  static void *new_malloc(size_t size)
551  void *block = malloc(size);  void *block = malloc(size);
552  gotten_store = size;  gotten_store = size;
553  if (show_malloc)  if (show_malloc)
554    fprintf(outfile, "malloc       %3d %p\n", size, block);    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);
555  return block;  return block;
556  }  }
557    
# Line 421  static void *stack_malloc(size_t size) Line 569  static void *stack_malloc(size_t size)
569  {  {
570  void *block = malloc(size);  void *block = malloc(size);
571  if (show_malloc)  if (show_malloc)
572    fprintf(outfile, "stack_malloc %3d %p\n", size, block);    fprintf(outfile, "stack_malloc %3d %p\n", (int)size, block);
573  return block;  return block;
574  }  }
575    
# Line 452  if ((rc = pcre_fullinfo(re, study, optio Line 600  if ((rc = pcre_fullinfo(re, study, optio
600  *         Byte flipping function                 *  *         Byte flipping function                 *
601  *************************************************/  *************************************************/
602    
603  static long int  static unsigned long int
604  byteflip(long int value, int n)  byteflip(unsigned long int value, int n)
605  {  {
606  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
607  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
# Line 466  return ((value & 0x000000ff) << 24) | Line 614  return ((value & 0x000000ff) << 24) |
614    
615    
616  /*************************************************  /*************************************************
617    *        Check match or recursion limit          *
618    *************************************************/
619    
620    static int
621    check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,
622      int start_offset, int options, int *use_offsets, int use_size_offsets,
623      int flag, unsigned long int *limit, int errnumber, const char *msg)
624    {
625    int count;
626    int min = 0;
627    int mid = 64;
628    int max = -1;
629    
630    extra->flags |= flag;
631    
632    for (;;)
633      {
634      *limit = mid;
635    
636      count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,
637        use_offsets, use_size_offsets);
638    
639      if (count == errnumber)
640        {
641        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
642        min = mid;
643        mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
644        }
645    
646      else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
647                             count == PCRE_ERROR_PARTIAL)
648        {
649        if (mid == min + 1)
650          {
651          fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
652          break;
653          }
654        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
655        max = mid;
656        mid = (min + mid)/2;
657        }
658      else break;    /* Some other error */
659      }
660    
661    extra->flags &= ~flag;
662    return count;
663    }
664    
665    
666    
667    /*************************************************
668    *         Case-independent strncmp() function    *
669    *************************************************/
670    
671    /*
672    Arguments:
673      s         first string
674      t         second string
675      n         number of characters to compare
676    
677    Returns:    < 0, = 0, or > 0, according to the comparison
678    */
679    
680    static int
681    strncmpic(uschar *s, uschar *t, int n)
682    {
683    while (n--)
684      {
685      int c = tolower(*s++) - tolower(*t++);
686      if (c) return c;
687      }
688    return 0;
689    }
690    
691    
692    
693    /*************************************************
694    *         Check newline indicator                *
695    *************************************************/
696    
697    /* This is used both at compile and run-time to check for <xxx> escapes, where
698    xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is
699    no match.
700    
701    Arguments:
702      p           points after the leading '<'
703      f           file for error message
704    
705    Returns:      appropriate PCRE_NEWLINE_xxx flags, or 0
706    */
707    
708    static int
709    check_newline(uschar *p, FILE *f)
710    {
711    if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
712    if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
713    if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
714    if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
715    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
716    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
717    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
718    fprintf(f, "Unknown newline type at: <%s\n", p);
719    return 0;
720    }
721    
722    
723    
724    /*************************************************
725    *             Usage function                     *
726    *************************************************/
727    
728    static void
729    usage(void)
730    {
731    printf("Usage:     pcretest [options] [<input> [<output>]]\n");
732    printf("  -b       show compiled code (bytecode)\n");
733    printf("  -C       show PCRE compile-time options and exit\n");
734    printf("  -d       debug: show compiled code and information (-b and -i)\n");
735    #if !defined NODFA
736    printf("  -dfa     force DFA matching for all subjects\n");
737    #endif
738    printf("  -help    show usage information\n");
739    printf("  -i       show information about compiled patterns\n"
740           "  -m       output memory used information\n"
741           "  -o <n>   set size of offsets vector to <n>\n");
742    #if !defined NOPOSIX
743    printf("  -p       use POSIX interface\n");
744    #endif
745    printf("  -q       quiet: do not output PCRE version number at start\n");
746    printf("  -S <n>   set stack size to <n> megabytes\n");
747    printf("  -s       output store (memory) used information\n"
748           "  -t       time compilation and execution\n");
749    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
750    printf("  -tm      time execution (matching) only\n");
751    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
752    }
753    
754    
755    
756    /*************************************************
757  *                Main Program                    *  *                Main Program                    *
758  *************************************************/  *************************************************/
759    
# Line 480  int options = 0; Line 768  int options = 0;
768  int study_options = 0;  int study_options = 0;
769  int op = 1;  int op = 1;
770  int timeit = 0;  int timeit = 0;
771    int timeitm = 0;
772  int showinfo = 0;  int showinfo = 0;
773  int showstore = 0;  int showstore = 0;
774    int quiet = 0;
775  int size_offsets = 45;  int size_offsets = 45;
776  int size_offsets_max;  int size_offsets_max;
777  int *offsets;  int *offsets = NULL;
778  #if !defined NOPOSIX  #if !defined NOPOSIX
779  int posix = 0;  int posix = 0;
780  #endif  #endif
781  int debug = 0;  int debug = 0;
782  int done = 0;  int done = 0;
783    int all_use_dfa = 0;
784    int yield = 0;
785    int stack_size;
786    
787    /* These vectors store, end-to-end, a list of captured substring names. Assume
788    that 1024 is plenty long enough for the few names we'll be testing. */
789    
790    uschar copynames[1024];
791    uschar getnames[1024];
792    
793  unsigned char *buffer;  uschar *copynamesptr;
794  unsigned char *dbuffer;  uschar *getnamesptr;
795    
796  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* Get buffers from malloc() so that Electric Fence will check their misuse
797  when I am debugging. */  when I am debugging. They grow automatically when very long lines are read. */
798    
799  buffer = (unsigned char *)malloc(BUFFER_SIZE);  buffer = (unsigned char *)malloc(buffer_size);
800  dbuffer = (unsigned char *)malloc(DBUFFER_SIZE);  dbuffer = (unsigned char *)malloc(buffer_size);
801  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.  */  
802    
803  #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) */  
804    
805  outfile = stdout;  outfile = stdout;
806    
807    /* The following  _setmode() stuff is some Windows magic that tells its runtime
808    library to translate CRLF into a single LF character. At least, that's what
809    I've been told: never having used Windows I take this all on trust. Originally
810    it set 0x8000, but then I was advised that _O_BINARY was better. */
811    
812    #if defined(_WIN32) || defined(WIN32)
813    _setmode( _fileno( stdout ), _O_BINARY );
814    #endif
815    
816  /* Scan options */  /* Scan options */
817    
818  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
# Line 519  while (argc > 1 && argv[op][0] == '-') Line 821  while (argc > 1 && argv[op][0] == '-')
821    
822    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
823      showstore = 1;      showstore = 1;
824    else if (strcmp(argv[op], "-t") == 0) timeit = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
825      else if (strcmp(argv[op], "-b") == 0) debug = 1;
826    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
827    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
828    #if !defined NODFA
829      else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
830    #endif
831    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
832        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),
833          *endptr == 0))          *endptr == 0))
# Line 529  while (argc > 1 && argv[op][0] == '-') Line 835  while (argc > 1 && argv[op][0] == '-')
835      op++;      op++;
836      argc--;      argc--;
837      }      }
838      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
839        {
840        int both = argv[op][2] == 0;
841        int temp;
842        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
843                         *endptr == 0))
844          {
845          timeitm = temp;
846          op++;
847          argc--;
848          }
849        else timeitm = LOOPREPEAT;
850        if (both) timeit = timeitm;
851        }
852      else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
853          ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
854            *endptr == 0))
855        {
856    #if defined(_WIN32) || defined(WIN32)
857        printf("PCRE: -S not supported on this OS\n");
858        exit(1);
859    #else
860        int rc;
861        struct rlimit rlim;
862        getrlimit(RLIMIT_STACK, &rlim);
863        rlim.rlim_cur = stack_size * 1024 * 1024;
864        rc = setrlimit(RLIMIT_STACK, &rlim);
865        if (rc != 0)
866          {
867        printf("PCRE: setrlimit() failed with error %d\n", rc);
868        exit(1);
869          }
870        op++;
871        argc--;
872    #endif
873        }
874  #if !defined NOPOSIX  #if !defined NOPOSIX
875    else if (strcmp(argv[op], "-p") == 0) posix = 1;    else if (strcmp(argv[op], "-p") == 0) posix = 1;
876  #endif  #endif
# Line 542  while (argc > 1 && argv[op][0] == '-') Line 884  while (argc > 1 && argv[op][0] == '-')
884      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
885      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
886      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
887      printf("  Newline character is %s\n", (rc == '\r')? "CR" : "LF");      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :
888          (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :
889          (rc == -2)? "ANYCRLF" :
890          (rc == -1)? "ANY" : "???");
891        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
892        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
893                                         "all Unicode newlines");
894      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
895      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
896      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
897      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
898      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);
899      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %d\n", rc);
900        (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);
901        printf("  Default recursion depth limit = %d\n", rc);
902      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
903      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
904      exit(0);      goto EXIT;
905        }
906      else if (strcmp(argv[op], "-help") == 0 ||
907               strcmp(argv[op], "--help") == 0)
908        {
909        usage();
910        goto EXIT;
911      }      }
912    else    else
913      {      {
914      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
915      printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");      usage();
916      printf("  -C     show PCRE compile-time options and exit\n");      yield = 1;
917      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;  
918      }      }
919    op++;    op++;
920    argc--;    argc--;
# Line 580  offsets = (int *)malloc(size_offsets_max Line 927  offsets = (int *)malloc(size_offsets_max
927  if (offsets == NULL)  if (offsets == NULL)
928    {    {
929    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
930      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
931    return 1;    yield = 1;
932      goto EXIT;
933    }    }
934    
935  /* Sort out the input and output files */  /* Sort out the input and output files */
936    
937  if (argc > 1)  if (argc > 1)
938    {    {
939    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
940    if (infile == NULL)    if (infile == NULL)
941      {      {
942      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
943      return 1;      yield = 1;
944        goto EXIT;
945      }      }
946    }    }
947    
948  if (argc > 2)  if (argc > 2)
949    {    {
950    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
951    if (outfile == NULL)    if (outfile == NULL)
952      {      {
953      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
954      return 1;      yield = 1;
955        goto EXIT;
956      }      }
957    }    }
958    
# Line 613  pcre_free = new_free; Line 963  pcre_free = new_free;
963  pcre_stack_malloc = stack_malloc;  pcre_stack_malloc = stack_malloc;
964  pcre_stack_free = stack_free;  pcre_stack_free = stack_free;
965    
966  /* Heading line, then prompt for first regex if stdin */  /* Heading line unless quiet, then prompt for first regex if stdin */
967    
968  fprintf(outfile, "PCRE version %s\n\n", pcre_version());  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
969    
970  /* Main loop */  /* Main loop */
971    
# Line 642  while (!done) Line 992  while (!done)
992    int do_showinfo = showinfo;    int do_showinfo = showinfo;
993    int do_showrest = 0;    int do_showrest = 0;
994    int do_flip = 0;    int do_flip = 0;
995    int erroroffset, len, delimiter;    int erroroffset, len, delimiter, poffset;
996    
997    use_utf8 = 0;    use_utf8 = 0;
998      debug_lengths = 1;
999    
1000    if (infile == stdin) printf("  re> ");    if (infile == stdin) printf("  re> ");
1001    if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) break;    if (extend_inputline(infile, buffer) == NULL) break;
1002    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1003    fflush(outfile);    fflush(outfile);
1004    
# Line 659  while (!done) Line 1010  while (!done)
1010    
1011    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
1012      {      {
1013      unsigned long int magic;      unsigned long int magic, get_options;
1014      uschar sbuf[8];      uschar sbuf[8];
1015      FILE *f;      FILE *f;
1016    
# Line 707  while (!done) Line 1058  while (!done)
1058    
1059      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
1060    
1061      new_info(re, NULL, PCRE_INFO_OPTIONS, &options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1062      use_utf8 = (options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1063    
1064      /* Now see if there is any following study data */      /* Now see if there is any following study data */
1065    
# Line 752  while (!done) Line 1103  while (!done)
1103      }      }
1104    
1105    pp = p;    pp = p;
1106      poffset = p - buffer;
1107    
1108    for(;;)    for(;;)
1109      {      {
# Line 762  while (!done) Line 1114  while (!done)
1114        pp++;        pp++;
1115        }        }
1116      if (*pp != 0) break;      if (*pp != 0) break;
   
     len = BUFFER_SIZE - (pp - buffer);  
     if (len < 256)  
       {  
       fprintf(outfile, "** Expression too long - missing delimiter?\n");  
       goto SKIP_DATA;  
       }  
   
1117      if (infile == stdin) printf("    > ");      if (infile == stdin) printf("    > ");
1118      if (fgets((char *)pp, len, infile) == NULL)      if ((pp = extend_inputline(infile, pp)) == NULL)
1119        {        {
1120        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1121        done = 1;        done = 1;
# Line 780  while (!done) Line 1124  while (!done)
1124      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1125      }      }
1126    
1127      /* The buffer may have moved while being extended; reset the start of data
1128      pointer to the correct relative point in the buffer. */
1129    
1130      p = buffer + poffset;
1131    
1132    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1133    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1134    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 802  while (!done) Line 1151  while (!done)
1151      {      {
1152      switch (*pp++)      switch (*pp++)
1153        {        {
1154          case 'f': options |= PCRE_FIRSTLINE; break;
1155        case 'g': do_g = 1; break;        case 'g': do_g = 1; break;
1156        case 'i': options |= PCRE_CASELESS; break;        case 'i': options |= PCRE_CASELESS; break;
1157        case 'm': options |= PCRE_MULTILINE; break;        case 'm': options |= PCRE_MULTILINE; break;
# Line 810  while (!done) Line 1160  while (!done)
1160    
1161        case '+': do_showrest = 1; break;        case '+': do_showrest = 1; break;
1162        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1163          case 'B': do_debug = 1; break;
1164        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1165        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1166        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
1167        case 'F': do_flip = 1; break;        case 'F': do_flip = 1; break;
1168        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1169        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1170          case 'J': options |= PCRE_DUPNAMES; break;
1171        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1172        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1173    
# Line 826  while (!done) Line 1178  while (!done)
1178        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1179        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1180        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1181          case 'Z': debug_lengths = 0; break;
1182        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1183        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1184    
1185        case 'L':        case 'L':
1186        ppp = pp;        ppp = pp;
1187        while (*ppp != '\n' && *ppp != ' ') ppp++;        /* The '\r' test here is so that it works on Windows. */
1188          /* The '0' test is just in case this is an unterminated line. */
1189          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1190        *ppp = 0;        *ppp = 0;
1191        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1192          {          {
1193          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1194          goto SKIP_DATA;          goto SKIP_DATA;
1195          }          }
1196          locale_set = 1;
1197        tables = pcre_maketables();        tables = pcre_maketables();
1198        pp = ppp;        pp = ppp;
1199        break;        break;
# Line 849  while (!done) Line 1205  while (!done)
1205        *pp = 0;        *pp = 0;
1206        break;        break;
1207    
1208        case '\n': case ' ': break;        case '<':
1209            {
1210            int x = check_newline(pp, outfile);
1211            if (x == 0) goto SKIP_DATA;
1212            options |= x;
1213            while (*pp++ != '>');
1214            }
1215          break;
1216    
1217          case '\r':                      /* So that it works in Windows */
1218          case '\n':
1219          case ' ':
1220          break;
1221    
1222        default:        default:
1223        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);
# Line 869  while (!done) Line 1237  while (!done)
1237    
1238      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
1239      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
1240        if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1241        if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1242        if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1243    
1244      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1245    
1246      /* 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 1248  while (!done)
1248    
1249      if (rc != 0)      if (rc != 0)
1250        {        {
1251        (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);        (void)regerror(rc, &preg, (char *)buffer, buffer_size);
1252        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
1253        goto SKIP_DATA;        goto SKIP_DATA;
1254        }        }
# Line 888  while (!done) Line 1260  while (!done)
1260  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1261    
1262      {      {
1263      if (timeit)      if (timeit > 0)
1264        {        {
1265        register int i;        register int i;
1266        clock_t time_taken;        clock_t time_taken;
1267        clock_t start_time = clock();        clock_t start_time = clock();
1268        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1269          {          {
1270          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1271          if (re != NULL) free(re);          if (re != NULL) free(re);
1272          }          }
1273        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1274        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1275          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1276            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1277        }        }
1278    
# Line 917  while (!done) Line 1289  while (!done)
1289          {          {
1290          for (;;)          for (;;)
1291            {            {
1292            if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)            if (extend_inputline(infile, buffer) == NULL)
1293              {              {
1294              done = 1;              done = 1;
1295              goto CONTINUE;              goto CONTINUE;
# Line 952  while (!done) Line 1324  while (!done)
1324    
1325      if (do_study)      if (do_study)
1326        {        {
1327        if (timeit)        if (timeit > 0)
1328          {          {
1329          register int i;          register int i;
1330          clock_t time_taken;          clock_t time_taken;
1331          clock_t start_time = clock();          clock_t start_time = clock();
1332          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1333            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options, &error);
1334          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1335          if (extra != NULL) free(extra);          if (extra != NULL) free(extra);
1336          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1337            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1338              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1339          }          }
1340        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options, &error);
# Line 980  while (!done) Line 1352  while (!done)
1352      if (do_flip)      if (do_flip)
1353        {        {
1354        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1355        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1356            byteflip(rre->magic_number, sizeof(rre->magic_number));
1357        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1358        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1359        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1360        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1361        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1362        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1363        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1364          rre->first_byte =
1365            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1366          rre->req_byte =
1367            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1368          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1369          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1370        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1371          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1372        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1373            sizeof(rre->name_count));
1374    
1375        if (extra != NULL)        if (extra != NULL)
1376          {          {
# Line 1005  while (!done) Line 1384  while (!done)
1384    
1385      SHOW_INFO:      SHOW_INFO:
1386    
1387        if (do_debug)
1388          {
1389          fprintf(outfile, "------------------------------------------------------------------\n");
1390          pcre_printint(re, outfile, debug_lengths);
1391          }
1392    
1393      if (do_showinfo)      if (do_showinfo)
1394        {        {
1395        unsigned long int get_options, all_options;        unsigned long int get_options, all_options;
1396    #if !defined NOINFOCHECK
1397        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1398        int count, backrefmax, first_char, need_char;  #endif
1399          int count, backrefmax, first_char, need_char, okpartial, jchanged,
1400            hascrorlf;
1401        int nameentrysize, namecount;        int nameentrysize, namecount;
1402        const uschar *nametable;        const uschar *nametable;
1403    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         print_internals(re, outfile);  
         }  
   
1404        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1405        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1406        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
# Line 1028  while (!done) Line 1410  while (!done)
1410        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1411        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1412        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1413          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1414          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1415          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1416    
1417    #if !defined NOINFOCHECK
1418        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
1419        if (count < 0) fprintf(outfile,        if (count < 0) fprintf(outfile,
1420          "Error %d from pcre_info()\n", count);          "Error %d from pcre_info()\n", count);
# Line 1046  while (!done) Line 1432  while (!done)
1432            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
1433              get_options, old_options);              get_options, old_options);
1434          }          }
1435    #endif
1436    
1437        if (size != regex_gotten_store) fprintf(outfile,        if (size != regex_gotten_store) fprintf(outfile,
1438          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
1439          size, regex_gotten_store);          (int)size, (int)regex_gotten_store);
1440    
1441        fprintf(outfile, "Capturing subpattern count = %d\n", count);        fprintf(outfile, "Capturing subpattern count = %d\n", count);
1442        if (backrefmax > 0)        if (backrefmax > 0)
# Line 1067  while (!done) Line 1454  while (!done)
1454            }            }
1455          }          }
1456    
1457        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1458        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1459    
1460        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1461        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");  
1462    
1463        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1464          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\n",
1465            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1466            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1467            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1468            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1469              ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1470            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1471              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1472              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1473            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1474            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1475            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1476              ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1477            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1478            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1479              ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1480    
1481          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1482    
1483          switch (get_options & PCRE_NEWLINE_BITS)
1484            {
1485            case PCRE_NEWLINE_CR:
1486            fprintf(outfile, "Forced newline sequence: CR\n");
1487            break;
1488    
1489            case PCRE_NEWLINE_LF:
1490            fprintf(outfile, "Forced newline sequence: LF\n");
1491            break;
1492    
1493            case PCRE_NEWLINE_CRLF:
1494            fprintf(outfile, "Forced newline sequence: CRLF\n");
1495            break;
1496    
1497        if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)          case PCRE_NEWLINE_ANYCRLF:
1498          fprintf(outfile, "Case state changes\n");          fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1499            break;
1500    
1501            case PCRE_NEWLINE_ANY:
1502            fprintf(outfile, "Forced newline sequence: ANY\n");
1503            break;
1504    
1505            default:
1506            break;
1507            }
1508    
1509        if (first_char == -1)        if (first_char == -1)
1510          {          {
1511          fprintf(outfile, "First char at start or follows \\n\n");          fprintf(outfile, "First char at start or follows newline\n");
1512          }          }
1513        else if (first_char < 0)        else if (first_char < 0)
1514          {          {
# Line 1108  while (!done) Line 1519  while (!done)
1519          int ch = first_char & 255;          int ch = first_char & 255;
1520          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
1521            "" : " (caseless)";            "" : " (caseless)";
1522          if (isprint(ch))          if (PRINTHEX(ch))
1523            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
1524          else          else
1525            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1123  while (!done) Line 1534  while (!done)
1534          int ch = need_char & 255;          int ch = need_char & 255;
1535          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
1536            "" : " (caseless)";            "" : " (caseless)";
1537          if (isprint(ch))          if (PRINTHEX(ch))
1538            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
1539          else          else
1540            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1159  while (!done) Line 1570  while (!done)
1570                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
1571                    c = 2;                    c = 2;
1572                    }                    }
1573                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
1574                    {                    {
1575                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
1576                    c += 2;                    c += 2;
# Line 1191  while (!done) Line 1602  while (!done)
1602        else        else
1603          {          {
1604          uschar sbuf[8];          uschar sbuf[8];
1605          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1606          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1607          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1608          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1609    
1610          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1611          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1612          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1613          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1614    
1615          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1616              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1218  while (!done) Line 1629  while (!done)
1629                  strerror(errno));                  strerror(errno));
1630                }                }
1631              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
1632    
1633              }              }
1634            }            }
1635          fclose(f);          fclose(f);
1636          }          }
1637    
1638          new_free(re);
1639          if (extra != NULL) new_free(extra);
1640          if (tables != NULL) new_free((void *)tables);
1641        continue;  /* With next regex */        continue;  /* With next regex */
1642        }        }
1643      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1230  while (!done) Line 1646  while (!done)
1646    
1647    for (;;)    for (;;)
1648      {      {
1649      unsigned char *q;      uschar *q;
1650      unsigned char *bptr = dbuffer;      uschar *bptr;
1651      int *use_offsets = offsets;      int *use_offsets = offsets;
1652      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1653      int callout_data = 0;      int callout_data = 0;
# Line 1244  while (!done) Line 1660  while (!done)
1660      int gmatched = 0;      int gmatched = 0;
1661      int start_offset = 0;      int start_offset = 0;
1662      int g_notempty = 0;      int g_notempty = 0;
1663        int use_dfa = 0;
1664    
1665      options = 0;      options = 0;
1666    
1667        *copynames = 0;
1668        *getnames = 0;
1669    
1670        copynamesptr = copynames;
1671        getnamesptr = getnames;
1672    
1673      pcre_callout = callout;      pcre_callout = callout;
1674      first_callout = 1;      first_callout = 1;
1675      callout_extra = 0;      callout_extra = 0;
# Line 1255  while (!done) Line 1678  while (!done)
1678      callout_fail_id = -1;      callout_fail_id = -1;
1679      show_malloc = 0;      show_malloc = 0;
1680    
1681      if (infile == stdin) printf("data> ");      if (extra != NULL) extra->flags &=
1682      if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
1683    
1684        len = 0;
1685        for (;;)
1686        {        {
1687        done = 1;        if (infile == stdin) printf("data> ");
1688        goto CONTINUE;        if (extend_inputline(infile, buffer + len) == NULL)
1689            {
1690            if (len > 0) break;
1691            done = 1;
1692            goto CONTINUE;
1693            }
1694          if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1695          len = (int)strlen((char *)buffer);
1696          if (buffer[len-1] == '\n') break;
1697        }        }
     if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);  
1698    
     len = (int)strlen((char *)buffer);  
1699      while (len > 0 && isspace(buffer[len-1])) len--;      while (len > 0 && isspace(buffer[len-1])) len--;
1700      buffer[len] = 0;      buffer[len] = 0;
1701      if (len == 0) break;      if (len == 0) break;
# Line 1271  while (!done) Line 1703  while (!done)
1703      p = buffer;      p = buffer;
1704      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1705    
1706      q = dbuffer;      bptr = q = dbuffer;
1707      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1708        {        {
1709        int i = 0;        int i = 0;
# Line 1293  while (!done) Line 1725  while (!done)
1725          c -= '0';          c -= '0';
1726          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
1727            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
1728    
1729    #if !defined NOUTF8
1730            if (use_utf8 && c > 255)
1731              {
1732              unsigned char buff8[8];
1733              int ii, utn;
1734              utn = ord2utf8(c, buff8);
1735              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1736              c = buff8[ii];   /* Last byte */
1737              }
1738    #endif
1739          break;          break;
1740    
1741          case 'x':          case 'x':
1742    
1743          /* Handle \x{..} specially - new Perl thing for utf8 */          /* Handle \x{..} specially - new Perl thing for utf8 */
1744    
1745    #if !defined NOUTF8
1746          if (*p == '{')          if (*p == '{')
1747            {            {
1748            unsigned char *pt = p;            unsigned char *pt = p;
# Line 1317  while (!done) Line 1761  while (!done)
1761              }              }
1762            /* Not correct form; fall through */            /* Not correct form; fall through */
1763            }            }
1764    #endif
1765    
1766          /* Ordinary \x */          /* Ordinary \x */
1767    
# Line 1352  while (!done) Line 1797  while (!done)
1797            }            }
1798          else if (isalnum(*p))          else if (isalnum(*p))
1799            {            {
1800            uschar name[256];            uschar *npp = copynamesptr;
           uschar *npp = name;  
1801            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
1802              *npp++ = 0;
1803            *npp = 0;            *npp = 0;
1804            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)copynamesptr);
1805            if (n < 0)            if (n < 0)
1806              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
1807            else copystrings |= 1 << n;            copynamesptr = npp;
1808            }            }
1809          else if (*p == '+')          else if (*p == '+')
1810            {            {
# Line 1397  while (!done) Line 1842  while (!done)
1842            }            }
1843          continue;          continue;
1844    
1845    #if !defined NODFA
1846            case 'D':
1847    #if !defined NOPOSIX
1848            if (posix || do_posix)
1849              printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");
1850            else
1851    #endif
1852              use_dfa = 1;
1853            continue;
1854    
1855            case 'F':
1856            options |= PCRE_DFA_SHORTEST;
1857            continue;
1858    #endif
1859    
1860          case 'G':          case 'G':
1861          if (isdigit(*p))          if (isdigit(*p))
1862            {            {
# Line 1405  while (!done) Line 1865  while (!done)
1865            }            }
1866          else if (isalnum(*p))          else if (isalnum(*p))
1867            {            {
1868            uschar name[256];            uschar *npp = getnamesptr;
           uschar *npp = name;  
1869            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
1870              *npp++ = 0;
1871            *npp = 0;            *npp = 0;
1872            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)getnamesptr);
1873            if (n < 0)            if (n < 0)
1874              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
1875            else getstrings |= 1 << n;            getnamesptr = npp;
1876            }            }
1877          continue;          continue;
1878    
# Line 1438  while (!done) Line 1898  while (!done)
1898            if (offsets == NULL)            if (offsets == NULL)
1899              {              {
1900              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1901                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1902              return 1;              yield = 1;
1903                goto EXIT;
1904              }              }
1905            }            }
1906          use_size_offsets = n;          use_size_offsets = n;
# Line 1450  while (!done) Line 1911  while (!done)
1911          options |= PCRE_PARTIAL;          options |= PCRE_PARTIAL;
1912          continue;          continue;
1913    
1914            case 'Q':
1915            while(isdigit(*p)) n = n * 10 + *p++ - '0';
1916            if (extra == NULL)
1917              {
1918              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1919              extra->flags = 0;
1920              }
1921            extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
1922            extra->match_limit_recursion = n;
1923            continue;
1924    
1925            case 'q':
1926            while(isdigit(*p)) n = n * 10 + *p++ - '0';
1927            if (extra == NULL)
1928              {
1929              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
1930              extra->flags = 0;
1931              }
1932            extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
1933            extra->match_limit = n;
1934            continue;
1935    
1936    #if !defined NODFA
1937            case 'R':
1938            options |= PCRE_DFA_RESTART;
1939            continue;
1940    #endif
1941    
1942          case 'S':          case 'S':
1943          show_malloc = 1;          show_malloc = 1;
1944          continue;          continue;
# Line 1461  while (!done) Line 1950  while (!done)
1950          case '?':          case '?':
1951          options |= PCRE_NO_UTF8_CHECK;          options |= PCRE_NO_UTF8_CHECK;
1952          continue;          continue;
1953    
1954            case '<':
1955              {
1956              int x = check_newline(p, outfile);
1957              if (x == 0) goto NEXT_DATA;
1958              options |= x;
1959              while (*p++ != '>');
1960              }
1961            continue;
1962          }          }
1963        *q++ = c;        *q++ = c;
1964        }        }
1965      *q = 0;      *q = 0;
1966      len = q - dbuffer;      len = q - dbuffer;
1967    
1968        if ((all_use_dfa || use_dfa) && find_match_limit)
1969          {
1970          printf("**Match limit not relevant for DFA matching: ignored\n");
1971          find_match_limit = 0;
1972          }
1973    
1974      /* Handle matching via the POSIX interface, which does not      /* Handle matching via the POSIX interface, which does not
1975      support timing or playing with the match limit or callout data. */      support timing or playing with the match limit or callout data. */
1976    
# Line 1485  while (!done) Line 1989  while (!done)
1989    
1990        if (rc != 0)        if (rc != 0)
1991          {          {
1992          (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);          (void)regerror(rc, &preg, (char *)buffer, buffer_size);
1993          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
1994          }          }
1995          else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
1996                  != 0)
1997            {
1998            fprintf(outfile, "Matched with REG_NOSUB\n");
1999            }
2000        else        else
2001          {          {
2002          size_t i;          size_t i;
# Line 1519  while (!done) Line 2028  while (!done)
2028    
2029      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2030        {        {
2031        if (timeit)        if (timeitm > 0)
2032          {          {
2033          register int i;          register int i;
2034          clock_t time_taken;          clock_t time_taken;
2035          clock_t start_time = clock();          clock_t start_time = clock();
2036          for (i = 0; i < LOOPREPEAT; i++)  
2037    #if !defined NODFA
2038            if (all_use_dfa || use_dfa)
2039              {
2040              int workspace[1000];
2041              for (i = 0; i < timeitm; i++)
2042                count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,
2043                  options | g_notempty, use_offsets, use_size_offsets, workspace,
2044                  sizeof(workspace)/sizeof(int));
2045              }
2046            else
2047    #endif
2048    
2049            for (i = 0; i < timeitm; i++)
2050            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2051              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2052    
2053          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2054          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2055            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2056              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2057          }          }
2058    
2059        /* 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
2060        varying limits in order to find the minimum value. */        varying limits in order to find the minimum value for the match limit and
2061          for the recursion limit. */
2062    
2063        if (find_match_limit)        if (find_match_limit)
2064          {          {
         int min = 0;  
         int mid = 64;  
         int max = -1;  
   
2065          if (extra == NULL)          if (extra == NULL)
2066            {            {
2067            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2068            extra->flags = 0;            extra->flags = 0;
2069            }            }
         extra->flags |= PCRE_EXTRA_MATCH_LIMIT;  
   
         for (;;)  
           {  
           extra->match_limit = mid;  
           count = pcre_exec(re, extra, (char *)bptr, len, start_offset,  
             options | g_notempty, use_offsets, use_size_offsets);  
           if (count == PCRE_ERROR_MATCHLIMIT)  
             {  
             /* fprintf(outfile, "Testing match limit = %d\n", mid); */  
             min = mid;  
             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 */  
           }  
2070    
2071          extra->flags &= ~PCRE_EXTRA_MATCH_LIMIT;          (void)check_match_limit(re, extra, bptr, len, start_offset,
2072              options|g_notempty, use_offsets, use_size_offsets,
2073              PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
2074              PCRE_ERROR_MATCHLIMIT, "match()");
2075    
2076            count = check_match_limit(re, extra, bptr, len, start_offset,
2077              options|g_notempty, use_offsets, use_size_offsets,
2078              PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
2079              PCRE_ERROR_RECURSIONLIMIT, "match() recursion");
2080          }          }
2081    
2082        /* 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 2098  while (!done)
2098        /* 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
2099        value of match_limit. */        value of match_limit. */
2100    
2101        else  #if !defined NODFA
2102          else if (all_use_dfa || use_dfa)
2103          {          {
2104          count = pcre_exec(re, extra, (char *)bptr, len,          int workspace[1000];
2105            start_offset, options | g_notempty, use_offsets, use_size_offsets);          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,
2106              options | g_notempty, use_offsets, use_size_offsets, workspace,
2107              sizeof(workspace)/sizeof(int));
2108            if (count == 0)
2109              {
2110              fprintf(outfile, "Matched, but too many subsidiary matches\n");
2111              count = use_size_offsets/2;
2112              }
2113          }          }
2114    #endif
2115    
2116        if (count == 0)        else
2117          {          {
2118          fprintf(outfile, "Matched, but too many substrings\n");          count = pcre_exec(re, extra, (char *)bptr, len,
2119          count = use_size_offsets/3;            start_offset, options | g_notempty, use_offsets, use_size_offsets);
2120            if (count == 0)
2121              {
2122              fprintf(outfile, "Matched, but too many substrings\n");
2123              count = use_size_offsets/3;
2124              }
2125          }          }
2126    
2127        /* Matched */        /* Matched */
2128    
2129        if (count >= 0)        if (count >= 0)
2130          {          {
2131          int i;          int i, maxcount;
2132    
2133    #if !defined NODFA
2134            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2135    #endif
2136              maxcount = use_size_offsets/3;
2137    
2138            /* This is a check against a lunatic return value. */
2139    
2140            if (count > maxcount)
2141              {
2142              fprintf(outfile,
2143                "** PCRE error: returned count %d is too big for offset size %d\n",
2144                count, use_size_offsets);
2145              count = use_size_offsets/3;
2146              if (do_g || do_G)
2147                {
2148                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2149                do_g = do_G = FALSE;        /* Break g/G loop */
2150                }
2151              }
2152    
2153          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2154            {            {
2155            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
# Line 1641  while (!done) Line 2177  while (!done)
2177            {            {
2178            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
2179              {              {
2180              char copybuffer[16];              char copybuffer[256];
2181              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,
2182                i, copybuffer, sizeof(copybuffer));                i, copybuffer, sizeof(copybuffer));
2183              if (rc < 0)              if (rc < 0)
# Line 1651  while (!done) Line 2187  while (!done)
2187              }              }
2188            }            }
2189    
2190            for (copynamesptr = copynames;
2191                 *copynamesptr != 0;
2192                 copynamesptr += (int)strlen((char*)copynamesptr) + 1)
2193              {
2194              char copybuffer[256];
2195              int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,
2196                count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));
2197              if (rc < 0)
2198                fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);
2199              else
2200                fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);
2201              }
2202    
2203          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2204            {            {
2205            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
# Line 1663  while (!done) Line 2212  while (!done)
2212              else              else
2213                {                {
2214                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
               /* free((void *)substring); */  
2215                pcre_free_substring(substring);                pcre_free_substring(substring);
2216                }                }
2217              }              }
2218            }            }
2219    
2220            for (getnamesptr = getnames;
2221                 *getnamesptr != 0;
2222                 getnamesptr += (int)strlen((char*)getnamesptr) + 1)
2223              {
2224              const char *substring;
2225              int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,
2226                count, (char *)getnamesptr, &substring);
2227              if (rc < 0)
2228                fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);
2229              else
2230                {
2231                fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);
2232                pcre_free_substring(substring);
2233                }
2234              }
2235    
2236          if (getlist)          if (getlist)
2237            {            {
2238            const char **stringlist;            const char **stringlist;
# Line 1692  while (!done) Line 2256  while (!done)
2256    
2257        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2258          {          {
2259          fprintf(outfile, "Partial match\n");          fprintf(outfile, "Partial match");
2260    #if !defined NODFA
2261            if ((all_use_dfa || use_dfa) && use_size_offsets > 2)
2262              fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],
2263                bptr + use_offsets[0]);
2264    #endif
2265            fprintf(outfile, "\n");
2266          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2267          }          }
2268    
2269        /* 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
2270        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
2271        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
2272        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2273        offset values to achieve this. We won't be at the end of the string -  
2274        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2275          "anycrlf". If the previous match was at the end of a line terminated by
2276          CRLF, an advance of one character just passes the \r, whereas we should
2277          prefer the longer newline sequence, as does the code in pcre_exec().
2278          Fudge the offset value to achieve this.
2279    
2280          Otherwise, in the case of UTF-8 matching, the advance must be one
2281          character, not one byte. */
2282    
2283        else        else
2284          {          {
2285          if (g_notempty != 0)          if (g_notempty != 0)
2286            {            {
2287            int onechar = 1;            int onechar = 1;
2288              unsigned int obits = ((real_pcre *)re)->options;
2289            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2290            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2291                {
2292                int d;
2293                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2294                obits = (d == '\r')? PCRE_NEWLINE_CR :
2295                        (d == '\n')? PCRE_NEWLINE_LF :
2296                        (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :
2297                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2298                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2299                }
2300              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2301                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2302                  &&
2303                  start_offset < len - 1 &&
2304                  bptr[start_offset] == '\r' &&
2305                  bptr[start_offset+1] == '\n')
2306                onechar++;
2307              else if (use_utf8)
2308              {              {
2309              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2310                {                {
# Line 1744  while (!done) Line 2339  while (!done)
2339        character. */        character. */
2340    
2341        g_notempty = 0;        g_notempty = 0;
2342    
2343        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2344          {          {
2345          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
# Line 1762  while (!done) Line 2358  while (!done)
2358          len -= use_offsets[1];          len -= use_offsets[1];
2359          }          }
2360        }  /* End of loop for /g and /G */        }  /* End of loop for /g and /G */
2361    
2362        NEXT_DATA: continue;
2363      }    /* End of loop for data lines */      }    /* End of loop for data lines */
2364    
2365    CONTINUE:    CONTINUE:
# Line 1770  while (!done) Line 2368  while (!done)
2368    if (posix || do_posix) regfree(&preg);    if (posix || do_posix) regfree(&preg);
2369  #endif  #endif
2370    
2371    if (re != NULL) free(re);    if (re != NULL) new_free(re);
2372    if (extra != NULL) free(extra);    if (extra != NULL) new_free(extra);
2373    if (tables != NULL)    if (tables != NULL)
2374      {      {
2375      free((void *)tables);      new_free((void *)tables);
2376      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
2377        locale_set = 0;
2378      }      }
2379    }    }
2380    
2381  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");
2382  return 0;  
2383    EXIT:
2384    
2385    if (infile != NULL && infile != stdin) fclose(infile);
2386    if (outfile != NULL && outfile != stdout) fclose(outfile);
2387    
2388    free(buffer);
2389    free(dbuffer);
2390    free(pbuffer);
2391    free(offsets);
2392    
2393    return yield;
2394  }  }
2395    
2396  /* End */  /* End of pcretest.c */

Legend:
Removed from v.75  
changed lines
  Added in v.255

  ViewVC Help
Powered by ViewVC 1.1.5