/[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 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 507 by ph10, Wed Mar 10 16:08:01 2010 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 44  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51  #define PCRE_SPY        /* For Win32 build, import data, not export */  #ifdef SUPPORT_LIBREADLINE
52    #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57    #endif
58    
59    
60    /* A number of things vary for Windows builds. Originally, pcretest opened its
61    input and output without "b"; then I was told that "b" was needed in some
62    environments, so it was added for release 5.0 to both the input and output. (It
63    makes no difference on Unix-like systems.) Later I was told that it is wrong
64    for the input on Windows. I've now abstracted the modes into two macros that
65    are set here, to make it easier to fiddle with them, and removed "b" from the
66    input mode under Windows. */
67    
68  /* We need the internal info for displaying the results of pcre_study() and  #if defined(_WIN32) || defined(WIN32)
69  other internal data; pcretest also uses some of the fixed tables, and generally  #include <io.h>                /* For _setmode() */
70  has "inside information" compared to a program that strictly follows the PCRE  #include <fcntl.h>             /* For _O_BINARY */
71  API. */  #define INPUT_MODE   "r"
72    #define OUTPUT_MODE  "wb"
73    
74    #ifndef isatty
75    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
76    #endif                         /* though in some environments they seem to   */
77                                   /* be already defined, hence the #ifndefs.    */
78    #ifndef fileno
79    #define fileno _fileno
80    #endif
81    
82    #else
83    #include <sys/time.h>          /* These two includes are needed */
84    #include <sys/resource.h>      /* for setrlimit(). */
85    #define INPUT_MODE   "rb"
86    #define OUTPUT_MODE  "wb"
87    #endif
88    
89    
90    /* We have to include pcre_internal.h because we need the internal info for
91    displaying the results of pcre_study() and we also need to know about the
92    internal macros, structures, and other internal data values; pcretest has
93    "inside information" compared to a program that strictly follows the PCRE API.
94    
95    Although pcre_internal.h does itself include pcre.h, we explicitly include it
96    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
97    appropriately for an application, not for building PCRE. */
98    
99    #include "pcre.h"
100  #include "pcre_internal.h"  #include "pcre_internal.h"
101    
102    /* We need access to some of the data tables that PCRE uses. So as not to have
103    to keep two copies, we include the source file here, changing the names of the
104    external symbols to prevent clashes. */
105    
106    #define _pcre_ucp_gentype      ucp_gentype
107    #define _pcre_utf8_table1      utf8_table1
108    #define _pcre_utf8_table1_size utf8_table1_size
109    #define _pcre_utf8_table2      utf8_table2
110    #define _pcre_utf8_table3      utf8_table3
111    #define _pcre_utf8_table4      utf8_table4
112    #define _pcre_utt              utt
113    #define _pcre_utt_size         utt_size
114    #define _pcre_utt_names        utt_names
115    #define _pcre_OP_lengths       OP_lengths
116    
117    #include "pcre_tables.c"
118    
119    /* We also need the pcre_printint() function for printing out compiled
120    patterns. This function is in a separate file so that it can be included in
121    pcre_compile.c when that module is compiled with debugging enabled. It needs to
122    know which case is being compiled. */
123    
124    #define COMPILING_PCRETEST
125    #include "pcre_printint.src"
126    
127    /* The definition of the macro PRINTABLE, which determines whether to print an
128    output character as-is or as a hex value when showing compiled patterns, is
129    contained in the printint.src file. We uses it here also, in cases when the
130    locale has not been explicitly changed, so as to get consistent output from
131    systems that differ in their output from isprint() even in the "C" locale. */
132    
133    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
134    
135  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
136  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 62  Makefile. */ Line 140  Makefile. */
140  #include "pcreposix.h"  #include "pcreposix.h"
141  #endif  #endif
142    
143    /* It is also possible, for the benefit of the version currently imported into
144    Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
145    interface to the DFA matcher (NODFA), and without the doublecheck of the old
146    "info" function (define NOINFOCHECK). In fact, we automatically cut out the
147    UTF8 support if PCRE is built without it. */
148    
149    #ifndef SUPPORT_UTF8
150    #ifndef NOUTF8
151    #define NOUTF8
152    #endif
153    #endif
154    
155    
156    /* Other parameters */
157    
158  #ifndef CLOCKS_PER_SEC  #ifndef CLOCKS_PER_SEC
159  #ifdef CLK_TCK  #ifdef CLK_TCK
160  #define CLOCKS_PER_SEC CLK_TCK  #define CLOCKS_PER_SEC CLK_TCK
# Line 70  Makefile. */ Line 163  Makefile. */
163  #endif  #endif
164  #endif  #endif
165    
166  #define LOOPREPEAT 500000  /* This is the default loop count for timing. */
167    
168  #define BUFFER_SIZE 30000  #define LOOPREPEAT 500000
 #define PBUFFER_SIZE BUFFER_SIZE  
 #define DBUFFER_SIZE BUFFER_SIZE  
169    
170    /* Static variables */
171    
172  static FILE *outfile;  static FILE *outfile;
173  static int log_store = 0;  static int log_store = 0;
# Line 83  static int callout_count; Line 175  static int callout_count;
175  static int callout_extra;  static int callout_extra;
176  static int callout_fail_count;  static int callout_fail_count;
177  static int callout_fail_id;  static int callout_fail_id;
178    static int debug_lengths;
179  static int first_callout;  static int first_callout;
180    static int locale_set = 0;
181  static int show_malloc;  static int show_malloc;
182  static int use_utf8;  static int use_utf8;
183  static size_t gotten_store;  static size_t gotten_store;
184    
185    /* The buffers grow automatically if very long input lines are encountered. */
186    
187    static int buffer_size = 50000;
188    static uschar *buffer = NULL;
189    static uschar *dbuffer = NULL;
190  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
191    
192    
193    
194  /*************************************************  /*************************************************
195    *        Read or extend an input line            *
196    *************************************************/
197    
198    /* Input lines are read into buffer, but both patterns and data lines can be
199    continued over multiple input lines. In addition, if the buffer fills up, we
200    want to automatically expand it so as to be able to handle extremely large
201    lines that are needed for certain stress tests. When the input buffer is
202    expanded, the other two buffers must also be expanded likewise, and the
203    contents of pbuffer, which are a copy of the input for callouts, must be
204    preserved (for when expansion happens for a data line). This is not the most
205    optimal way of handling this, but hey, this is just a test program!
206    
207    Arguments:
208      f            the file to read
209      start        where in buffer to start (this *must* be within buffer)
210      prompt       for stdin or readline()
211    
212    Returns:       pointer to the start of new data
213                   could be a copy of start, or could be moved
214                   NULL if no data read and EOF reached
215    */
216    
217    static uschar *
218    extend_inputline(FILE *f, uschar *start, const char *prompt)
219    {
220    uschar *here = start;
221    
222    for (;;)
223      {
224      int rlen = buffer_size - (here - buffer);
225    
226      if (rlen > 1000)
227        {
228        int dlen;
229    
230        /* If libreadline support is required, use readline() to read a line if the
231        input is a terminal. Note that readline() removes the trailing newline, so
232        we must put it back again, to be compatible with fgets(). */
233    
234    #ifdef SUPPORT_LIBREADLINE
235        if (isatty(fileno(f)))
236          {
237          size_t len;
238          char *s = readline(prompt);
239          if (s == NULL) return (here == start)? NULL : start;
240          len = strlen(s);
241          if (len > 0) add_history(s);
242          if (len > rlen - 1) len = rlen - 1;
243          memcpy(here, s, len);
244          here[len] = '\n';
245          here[len+1] = 0;
246          free(s);
247          }
248        else
249    #endif
250    
251        /* Read the next line by normal means, prompting if the file is stdin. */
252    
253          {
254          if (f == stdin) printf(prompt);
255          if (fgets((char *)here, rlen,  f) == NULL)
256            return (here == start)? NULL : start;
257          }
258    
259        dlen = (int)strlen((char *)here);
260        if (dlen > 0 && here[dlen - 1] == '\n') return start;
261        here += dlen;
262        }
263    
264      else
265        {
266        int new_buffer_size = 2*buffer_size;
267        uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);
268        uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
269        uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);
270    
271        if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
272          {
273          fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
274          exit(1);
275          }
276    
277        memcpy(new_buffer, buffer, buffer_size);
278        memcpy(new_pbuffer, pbuffer, buffer_size);
279    
280        buffer_size = new_buffer_size;
281    
282        start = new_buffer + (start - buffer);
283        here = new_buffer + (here - buffer);
284    
285        free(buffer);
286        free(dbuffer);
287        free(pbuffer);
288    
289        buffer = new_buffer;
290        dbuffer = new_dbuffer;
291        pbuffer = new_pbuffer;
292        }
293      }
294    
295    return NULL;  /* Control never gets here */
296    }
297    
298    
299    
300    
301    
302    
303    
304    /*************************************************
305  *          Read number from string               *  *          Read number from string               *
306  *************************************************/  *************************************************/
307    
308  /* 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
309  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
310  for unpicking the -o argument, so just keep it simple.  for unpicking arguments, so just keep it simple.
311    
312  Arguments:  Arguments:
313    str           string to be converted    str           string to be converted
# Line 128  return(result); Line 337  return(result);
337  and returns the value of the character.  and returns the value of the character.
338    
339  Argument:  Argument:
340    buffer   a pointer to the byte vector    utf8bytes   a pointer to the byte vector
341    vptr     a pointer to an int to receive the value    vptr        a pointer to an int to receive the value
342    
343  Returns:   >  0 => the number of bytes consumed  Returns:      >  0 => the number of bytes consumed
344             -6 to 0 => malformed UTF-8 character at offset = (-return)                -6 to 0 => malformed UTF-8 character at offset = (-return)
345  */  */
346    
347    #if !defined NOUTF8
348    
349  static int  static int
350  utf82ord(unsigned char *buffer, int *vptr)  utf82ord(unsigned char *utf8bytes, int *vptr)
351  {  {
352  int c = *buffer++;  int c = *utf8bytes++;
353  int d = c;  int d = c;
354  int i, j, s;  int i, j, s;
355    
# Line 154  if (i == 0 || i == 6) return 0;        / Line 365  if (i == 0 || i == 6) return 0;        /
365  /* i now has a value in the range 1-5 */  /* i now has a value in the range 1-5 */
366    
367  s = 6*i;  s = 6*i;
368  d = (c & _pcre_utf8_table3[i]) << s;  d = (c & utf8_table3[i]) << s;
369    
370  for (j = 0; j < i; j++)  for (j = 0; j < i; j++)
371    {    {
372    c = *buffer++;    c = *utf8bytes++;
373    if ((c & 0xc0) != 0x80) return -(j+1);    if ((c & 0xc0) != 0x80) return -(j+1);
374    s -= 6;    s -= 6;
375    d |= (c & 0x3f) << s;    d |= (c & 0x3f) << s;
# Line 166  for (j = 0; j < i; j++) Line 377  for (j = 0; j < i; j++)
377    
378  /* Check that encoding was the correct unique one */  /* Check that encoding was the correct unique one */
379    
380  for (j = 0; j < _pcre_utf8_table1_size; j++)  for (j = 0; j < utf8_table1_size; j++)
381    if (d <= _pcre_utf8_table1[j]) break;    if (d <= utf8_table1[j]) break;
382  if (j != i) return -(i+1);  if (j != i) return -(i+1);
383    
384  /* Valid value */  /* Valid value */
# Line 176  if (j != i) return -(i+1); Line 387  if (j != i) return -(i+1);
387  return i+1;  return i+1;
388  }  }
389    
390    #endif
391    
392    
393    
394    /*************************************************
395    *       Convert character value to UTF-8         *
396    *************************************************/
397    
398    /* This function takes an integer value in the range 0 - 0x7fffffff
399    and encodes it as a UTF-8 character in 0 to 6 bytes.
400    
401    Arguments:
402      cvalue     the character value
403      utf8bytes  pointer to buffer for result - at least 6 bytes long
404    
405    Returns:     number of characters placed in the buffer
406    */
407    
408    #if !defined NOUTF8
409    
410    static int
411    ord2utf8(int cvalue, uschar *utf8bytes)
412    {
413    register int i, j;
414    for (i = 0; i < utf8_table1_size; i++)
415      if (cvalue <= utf8_table1[i]) break;
416    utf8bytes += i;
417    for (j = i; j > 0; j--)
418     {
419     *utf8bytes-- = 0x80 | (cvalue & 0x3f);
420     cvalue >>= 6;
421     }
422    *utf8bytes = utf8_table2[i] | cvalue;
423    return i + 1;
424    }
425    
426    #endif
427    
428    
429    
430  /*************************************************  /*************************************************
# Line 188  chars without printing. */ Line 437  chars without printing. */
437    
438  static int pchars(unsigned char *p, int length, FILE *f)  static int pchars(unsigned char *p, int length, FILE *f)
439  {  {
440  int c;  int c = 0;
441  int yield = 0;  int yield = 0;
442    
443  while (length-- > 0)  while (length-- > 0)
444    {    {
445    #if !defined NOUTF8
446    if (use_utf8)    if (use_utf8)
447      {      {
448      int rc = utf82ord(p, &c);      int rc = utf82ord(p, &c);
# Line 201  while (length-- > 0) Line 451  while (length-- > 0)
451        {        {
452        length -= rc - 1;        length -= rc - 1;
453        p += rc;        p += rc;
454        if (c < 256 && isprint(c))        if (PRINTHEX(c))
455          {          {
456          if (f != NULL) fprintf(f, "%c", c);          if (f != NULL) fprintf(f, "%c", c);
457          yield++;          yield++;
458          }          }
459        else        else
460          {          {
461          int n;          int n = 4;
462          if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);          if (f != NULL) fprintf(f, "\\x{%02x}", c);
463          yield += n;          yield += (n <= 0x000000ff)? 2 :
464                     (n <= 0x00000fff)? 3 :
465                     (n <= 0x0000ffff)? 4 :
466                     (n <= 0x000fffff)? 5 : 6;
467          }          }
468        continue;        continue;
469        }        }
470      }      }
471    #endif
472    
473     /* Not UTF-8, or malformed UTF-8  */     /* Not UTF-8, or malformed UTF-8  */
474    
475    if (isprint(c = *(p++)))    c = *p++;
476      if (PRINTHEX(c))
477      {      {
478      if (f != NULL) fprintf(f, "%c", c);      if (f != NULL) fprintf(f, "%c", c);
479      yield++;      yield++;
# Line 392  if ((rc = pcre_fullinfo(re, study, optio Line 647  if ((rc = pcre_fullinfo(re, study, optio
647  *         Byte flipping function                 *  *         Byte flipping function                 *
648  *************************************************/  *************************************************/
649    
650  static long int  static unsigned long int
651  byteflip(long int value, int n)  byteflip(unsigned long int value, int n)
652  {  {
653  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
654  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
# Line 406  return ((value & 0x000000ff) << 24) | Line 661  return ((value & 0x000000ff) << 24) |
661    
662    
663  /*************************************************  /*************************************************
664    *        Check match or recursion limit          *
665    *************************************************/
666    
667    static int
668    check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,
669      int start_offset, int options, int *use_offsets, int use_size_offsets,
670      int flag, unsigned long int *limit, int errnumber, const char *msg)
671    {
672    int count;
673    int min = 0;
674    int mid = 64;
675    int max = -1;
676    
677    extra->flags |= flag;
678    
679    for (;;)
680      {
681      *limit = mid;
682    
683      count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,
684        use_offsets, use_size_offsets);
685    
686      if (count == errnumber)
687        {
688        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
689        min = mid;
690        mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
691        }
692    
693      else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
694                             count == PCRE_ERROR_PARTIAL)
695        {
696        if (mid == min + 1)
697          {
698          fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
699          break;
700          }
701        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
702        max = mid;
703        mid = (min + mid)/2;
704        }
705      else break;    /* Some other error */
706      }
707    
708    extra->flags &= ~flag;
709    return count;
710    }
711    
712    
713    
714    /*************************************************
715    *         Case-independent strncmp() function    *
716    *************************************************/
717    
718    /*
719    Arguments:
720      s         first string
721      t         second string
722      n         number of characters to compare
723    
724    Returns:    < 0, = 0, or > 0, according to the comparison
725    */
726    
727    static int
728    strncmpic(uschar *s, uschar *t, int n)
729    {
730    while (n--)
731      {
732      int c = tolower(*s++) - tolower(*t++);
733      if (c) return c;
734      }
735    return 0;
736    }
737    
738    
739    
740    /*************************************************
741    *         Check newline indicator                *
742    *************************************************/
743    
744    /* This is used both at compile and run-time to check for <xxx> escapes, where
745    xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is
746    no match.
747    
748    Arguments:
749      p           points after the leading '<'
750      f           file for error message
751    
752    Returns:      appropriate PCRE_NEWLINE_xxx flags, or 0
753    */
754    
755    static int
756    check_newline(uschar *p, FILE *f)
757    {
758    if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
759    if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
760    if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
761    if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
762    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
763    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
764    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
765    fprintf(f, "Unknown newline type at: <%s\n", p);
766    return 0;
767    }
768    
769    
770    
771    /*************************************************
772    *             Usage function                     *
773    *************************************************/
774    
775    static void
776    usage(void)
777    {
778    printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
779    printf("Input and output default to stdin and stdout.\n");
780    #ifdef SUPPORT_LIBREADLINE
781    printf("If input is a terminal, readline() is used to read from it.\n");
782    #else
783    printf("This version of pcretest is not linked with readline().\n");
784    #endif
785    printf("\nOptions:\n");
786    printf("  -b       show compiled code (bytecode)\n");
787    printf("  -C       show PCRE compile-time options and exit\n");
788    printf("  -d       debug: show compiled code and information (-b and -i)\n");
789    #if !defined NODFA
790    printf("  -dfa     force DFA matching for all subjects\n");
791    #endif
792    printf("  -help    show usage information\n");
793    printf("  -i       show information about compiled patterns\n"
794           "  -M       find MATCH_LIMIT minimum for each subject\n"
795           "  -m       output memory used information\n"
796           "  -o <n>   set size of offsets vector to <n>\n");
797    #if !defined NOPOSIX
798    printf("  -p       use POSIX interface\n");
799    #endif
800    printf("  -q       quiet: do not output PCRE version number at start\n");
801    printf("  -S <n>   set stack size to <n> megabytes\n");
802    printf("  -s       output store (memory) used information\n"
803           "  -t       time compilation and execution\n");
804    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
805    printf("  -tm      time execution (matching) only\n");
806    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
807    }
808    
809    
810    
811    /*************************************************
812  *                Main Program                    *  *                Main Program                    *
813  *************************************************/  *************************************************/
814    
# Line 418  int main(int argc, char **argv) Line 821  int main(int argc, char **argv)
821  FILE *infile = stdin;  FILE *infile = stdin;
822  int options = 0;  int options = 0;
823  int study_options = 0;  int study_options = 0;
824    int default_find_match_limit = FALSE;
825  int op = 1;  int op = 1;
826  int timeit = 0;  int timeit = 0;
827    int timeitm = 0;
828  int showinfo = 0;  int showinfo = 0;
829  int showstore = 0;  int showstore = 0;
830    int quiet = 0;
831  int size_offsets = 45;  int size_offsets = 45;
832  int size_offsets_max;  int size_offsets_max;
833  int *offsets = NULL;  int *offsets = NULL;
# Line 432  int debug = 0; Line 838  int debug = 0;
838  int done = 0;  int done = 0;
839  int all_use_dfa = 0;  int all_use_dfa = 0;
840  int yield = 0;  int yield = 0;
841    int stack_size;
842    
843  unsigned char *buffer;  /* These vectors store, end-to-end, a list of captured substring names. Assume
844  unsigned char *dbuffer;  that 1024 is plenty long enough for the few names we'll be testing. */
845    
846    uschar copynames[1024];
847    uschar getnames[1024];
848    
849    uschar *copynamesptr;
850    uschar *getnamesptr;
851    
852  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* Get buffers from malloc() so that Electric Fence will check their misuse
853  when I am debugging. */  when I am debugging. They grow automatically when very long lines are read. */
854    
855  buffer = (unsigned char *)malloc(BUFFER_SIZE);  buffer = (unsigned char *)malloc(buffer_size);
856  dbuffer = (unsigned char *)malloc(DBUFFER_SIZE);  dbuffer = (unsigned char *)malloc(buffer_size);
857  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.  */  
858    
859  #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) */  
860    
861  outfile = stdout;  outfile = stdout;
862    
863    /* The following  _setmode() stuff is some Windows magic that tells its runtime
864    library to translate CRLF into a single LF character. At least, that's what
865    I've been told: never having used Windows I take this all on trust. Originally
866    it set 0x8000, but then I was advised that _O_BINARY was better. */
867    
868    #if defined(_WIN32) || defined(WIN32)
869    _setmode( _fileno( stdout ), _O_BINARY );
870    #endif
871    
872  /* Scan options */  /* Scan options */
873    
874  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
# Line 461  while (argc > 1 && argv[op][0] == '-') Line 877  while (argc > 1 && argv[op][0] == '-')
877    
878    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
879      showstore = 1;      showstore = 1;
880    else if (strcmp(argv[op], "-t") == 0) timeit = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
881      else if (strcmp(argv[op], "-b") == 0) debug = 1;
882    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
883    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
884      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
885    #if !defined NODFA
886    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
887    #endif
888    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
889        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),
890          *endptr == 0))          *endptr == 0))
# Line 472  while (argc > 1 && argv[op][0] == '-') Line 892  while (argc > 1 && argv[op][0] == '-')
892      op++;      op++;
893      argc--;      argc--;
894      }      }
895      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
896        {
897        int both = argv[op][2] == 0;
898        int temp;
899        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
900                         *endptr == 0))
901          {
902          timeitm = temp;
903          op++;
904          argc--;
905          }
906        else timeitm = LOOPREPEAT;
907        if (both) timeit = timeitm;
908        }
909      else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
910          ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
911            *endptr == 0))
912        {
913    #if defined(_WIN32) || defined(WIN32)
914        printf("PCRE: -S not supported on this OS\n");
915        exit(1);
916    #else
917        int rc;
918        struct rlimit rlim;
919        getrlimit(RLIMIT_STACK, &rlim);
920        rlim.rlim_cur = stack_size * 1024 * 1024;
921        rc = setrlimit(RLIMIT_STACK, &rlim);
922        if (rc != 0)
923          {
924        printf("PCRE: setrlimit() failed with error %d\n", rc);
925        exit(1);
926          }
927        op++;
928        argc--;
929    #endif
930        }
931  #if !defined NOPOSIX  #if !defined NOPOSIX
932    else if (strcmp(argv[op], "-p") == 0) posix = 1;    else if (strcmp(argv[op], "-p") == 0) posix = 1;
933  #endif  #endif
934    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
935      {      {
936      int rc;      int rc;
937        unsigned long int lrc;
938      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
939      printf("Compiled with\n");      printf("Compiled with\n");
940      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 485  while (argc > 1 && argv[op][0] == '-') Line 942  while (argc > 1 && argv[op][0] == '-')
942      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
943      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
944      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
945      printf("  Newline character is %s\n", (rc == '\r')? "CR" : "LF");      /* Note that these values are always the ASCII values, even
946        in EBCDIC environments. CR is 13 and NL is 10. */
947        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
948          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
949          (rc == -2)? "ANYCRLF" :
950          (rc == -1)? "ANY" : "???");
951        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
952        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
953                                         "all Unicode newlines");
954      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
955      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
956      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
957      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
958      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
959      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
960        (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
961        printf("  Default recursion depth limit = %ld\n", lrc);
962      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
963      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
964      exit(0);      goto EXIT;
965        }
966      else if (strcmp(argv[op], "-help") == 0 ||
967               strcmp(argv[op], "--help") == 0)
968        {
969        usage();
970        goto EXIT;
971      }      }
972    else    else
973      {      {
974      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
975      printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");      usage();
     printf("  -C     show PCRE compile-time options and exit\n");  
     printf("  -d     debug: show compiled code; implies -i\n");  
     printf("  -dfa   force DFA matching for all subjects\n");  
     printf("  -i     show information about compiled pattern\n"  
            "  -m     output memory used information\n"  
            "  -o <n> set size of offsets vector to <n>\n");  
 #if !defined NOPOSIX  
     printf("  -p     use POSIX interface\n");  
 #endif  
     printf("  -s     output store (memory) used information\n"  
            "  -t     time compilation and execution\n");  
976      yield = 1;      yield = 1;
977      goto EXIT;      goto EXIT;
978      }      }
# Line 525  offsets = (int *)malloc(size_offsets_max Line 987  offsets = (int *)malloc(size_offsets_max
987  if (offsets == NULL)  if (offsets == NULL)
988    {    {
989    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
990      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
991    yield = 1;    yield = 1;
992    goto EXIT;    goto EXIT;
993    }    }
# Line 534  if (offsets == NULL) Line 996  if (offsets == NULL)
996    
997  if (argc > 1)  if (argc > 1)
998    {    {
999    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
1000    if (infile == NULL)    if (infile == NULL)
1001      {      {
1002      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
# Line 545  if (argc > 1) Line 1007  if (argc > 1)
1007    
1008  if (argc > 2)  if (argc > 2)
1009    {    {
1010    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
1011    if (outfile == NULL)    if (outfile == NULL)
1012      {      {
1013      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
# Line 561  pcre_free = new_free; Line 1023  pcre_free = new_free;
1023  pcre_stack_malloc = stack_malloc;  pcre_stack_malloc = stack_malloc;
1024  pcre_stack_free = stack_free;  pcre_stack_free = stack_free;
1025    
1026  /* Heading line, then prompt for first regex if stdin */  /* Heading line unless quiet, then prompt for first regex if stdin */
1027    
1028  fprintf(outfile, "PCRE version %s\n\n", pcre_version());  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
1029    
1030  /* Main loop */  /* Main loop */
1031    
# Line 590  while (!done) Line 1052  while (!done)
1052    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1053    int do_showrest = 0;    int do_showrest = 0;
1054    int do_flip = 0;    int do_flip = 0;
1055    int erroroffset, len, delimiter;    int erroroffset, len, delimiter, poffset;
1056    
1057    use_utf8 = 0;    use_utf8 = 0;
1058      debug_lengths = 1;
1059    
1060    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) break;  
1061    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1062    fflush(outfile);    fflush(outfile);
1063    
# Line 607  while (!done) Line 1069  while (!done)
1069    
1070    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
1071      {      {
1072      unsigned long int magic;      unsigned long int magic, get_options;
1073      uschar sbuf[8];      uschar sbuf[8];
1074      FILE *f;      FILE *f;
1075    
# Line 655  while (!done) Line 1117  while (!done)
1117    
1118      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
1119    
1120      new_info(re, NULL, PCRE_INFO_OPTIONS, &options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1121      use_utf8 = (options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1122    
1123      /* Now see if there is any following study data */      /* Now see if there is any following study data */
1124    
# Line 695  while (!done) Line 1157  while (!done)
1157    
1158    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1159      {      {
1160      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1161      goto SKIP_DATA;      goto SKIP_DATA;
1162      }      }
1163    
1164    pp = p;    pp = p;
1165      poffset = p - buffer;
1166    
1167    for(;;)    for(;;)
1168      {      {
# Line 710  while (!done) Line 1173  while (!done)
1173        pp++;        pp++;
1174        }        }
1175      if (*pp != 0) break;      if (*pp != 0) break;
1176        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)  
1177        {        {
1178        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1179        done = 1;        done = 1;
# Line 728  while (!done) Line 1182  while (!done)
1182      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1183      }      }
1184    
1185      /* The buffer may have moved while being extended; reset the start of data
1186      pointer to the correct relative point in the buffer. */
1187    
1188      p = buffer + poffset;
1189    
1190    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1191    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1192    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 759  while (!done) Line 1218  while (!done)
1218    
1219        case '+': do_showrest = 1; break;        case '+': do_showrest = 1; break;
1220        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1221          case 'B': do_debug = 1; break;
1222        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1223        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1224        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
1225        case 'F': do_flip = 1; break;        case 'F': do_flip = 1; break;
1226        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1227        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1228          case 'J': options |= PCRE_DUPNAMES; break;
1229        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1230        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1231    
# Line 775  while (!done) Line 1236  while (!done)
1236        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1237        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1238        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1239          case 'Z': debug_lengths = 0; break;
1240        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1241        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1242    
1243        case 'L':        case 'L':
1244        ppp = pp;        ppp = pp;
1245        /* The '\r' test here is so that it works on Windows */        /* The '\r' test here is so that it works on Windows. */
1246        while (*ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;        /* The '0' test is just in case this is an unterminated line. */
1247          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1248        *ppp = 0;        *ppp = 0;
1249        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1250          {          {
1251          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1252          goto SKIP_DATA;          goto SKIP_DATA;
1253          }          }
1254          locale_set = 1;
1255        tables = pcre_maketables();        tables = pcre_maketables();
1256        pp = ppp;        pp = ppp;
1257        break;        break;
# Line 799  while (!done) Line 1263  while (!done)
1263        *pp = 0;        *pp = 0;
1264        break;        break;
1265    
1266          case '<':
1267            {
1268            if (strncmp((char *)pp, "JS>", 3) == 0)
1269              {
1270              options |= PCRE_JAVASCRIPT_COMPAT;
1271              pp += 3;
1272              }
1273            else
1274              {
1275              int x = check_newline(pp, outfile);
1276              if (x == 0) goto SKIP_DATA;
1277              options |= x;
1278              while (*pp++ != '>');
1279              }
1280            }
1281          break;
1282    
1283        case '\r':                      /* So that it works in Windows */        case '\r':                      /* So that it works in Windows */
1284        case '\n':        case '\n':
1285        case ' ':        case ' ':
# Line 823  while (!done) Line 1304  while (!done)
1304      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
1305      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
1306      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1307        if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1308        if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1309        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1310    
1311      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1312    
1313      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
# Line 830  while (!done) Line 1315  while (!done)
1315    
1316      if (rc != 0)      if (rc != 0)
1317        {        {
1318        (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);        (void)regerror(rc, &preg, (char *)buffer, buffer_size);
1319        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
1320        goto SKIP_DATA;        goto SKIP_DATA;
1321        }        }
# Line 842  while (!done) Line 1327  while (!done)
1327  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1328    
1329      {      {
1330      if (timeit)      unsigned long int get_options;
1331    
1332        if (timeit > 0)
1333        {        {
1334        register int i;        register int i;
1335        clock_t time_taken;        clock_t time_taken;
1336        clock_t start_time = clock();        clock_t start_time = clock();
1337        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1338          {          {
1339          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1340          if (re != NULL) free(re);          if (re != NULL) free(re);
1341          }          }
1342        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1343        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1344          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1345            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1346        }        }
1347    
# Line 871  while (!done) Line 1358  while (!done)
1358          {          {
1359          for (;;)          for (;;)
1360            {            {
1361            if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1362              {              {
1363              done = 1;              done = 1;
1364              goto CONTINUE;              goto CONTINUE;
# Line 885  while (!done) Line 1372  while (!done)
1372        goto CONTINUE;        goto CONTINUE;
1373        }        }
1374    
1375      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1376      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
1377      returns only limited data. Check that it agrees with the newer one. */      lines. */
1378    
1379        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1380        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1381    
1382        /* Print information if required. There are now two info-returning
1383        functions. The old one has a limited interface and returns only limited
1384        data. Check that it agrees with the newer one. */
1385    
1386      if (log_store)      if (log_store)
1387        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 906  while (!done) Line 1400  while (!done)
1400    
1401      if (do_study)      if (do_study)
1402        {        {
1403        if (timeit)        if (timeit > 0)
1404          {          {
1405          register int i;          register int i;
1406          clock_t time_taken;          clock_t time_taken;
1407          clock_t start_time = clock();          clock_t start_time = clock();
1408          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1409            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options, &error);
1410          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1411          if (extra != NULL) free(extra);          if (extra != NULL) free(extra);
1412          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1413            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1414              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1415          }          }
1416        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options, &error);
# Line 934  while (!done) Line 1428  while (!done)
1428      if (do_flip)      if (do_flip)
1429        {        {
1430        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1431        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1432            byteflip(rre->magic_number, sizeof(rre->magic_number));
1433        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1434        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1435        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1436        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1437        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1438        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1439        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1440          rre->first_byte =
1441            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1442          rre->req_byte =
1443            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1444          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1445          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1446        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1447          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1448        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1449            sizeof(rre->name_count));
1450    
1451        if (extra != NULL)        if (extra != NULL)
1452          {          {
1453          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1454          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1455          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1456            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1457          }          }
1458        }        }
1459    
# Line 959  while (!done) Line 1461  while (!done)
1461    
1462      SHOW_INFO:      SHOW_INFO:
1463    
1464        if (do_debug)
1465          {
1466          fprintf(outfile, "------------------------------------------------------------------\n");
1467          pcre_printint(re, outfile, debug_lengths);
1468          }
1469    
1470        /* We already have the options in get_options (see above) */
1471    
1472      if (do_showinfo)      if (do_showinfo)
1473        {        {
1474        unsigned long int get_options, all_options;        unsigned long int all_options;
1475    #if !defined NOINFOCHECK
1476        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1477        int count, backrefmax, first_char, need_char;  #endif
1478          int count, backrefmax, first_char, need_char, okpartial, jchanged,
1479            hascrorlf;
1480        int nameentrysize, namecount;        int nameentrysize, namecount;
1481        const uschar *nametable;        const uschar *nametable;
1482    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         _pcre_printint(re, outfile);  
         }  
   
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1483        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1484        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1485        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 982  while (!done) Line 1488  while (!done)
1488        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1489        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1490        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1491          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1492          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1493          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1494    
1495    #if !defined NOINFOCHECK
1496        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
1497        if (count < 0) fprintf(outfile,        if (count < 0) fprintf(outfile,
1498          "Error %d from pcre_info()\n", count);          "Error %d from pcre_info()\n", count);
# Line 1000  while (!done) Line 1510  while (!done)
1510            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
1511              get_options, old_options);              get_options, old_options);
1512          }          }
1513    #endif
1514    
1515        if (size != regex_gotten_store) fprintf(outfile,        if (size != regex_gotten_store) fprintf(outfile,
1516          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
# Line 1021  while (!done) Line 1532  while (!done)
1532            }            }
1533          }          }
1534    
1535        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1536        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1537    
1538        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1539        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");  
1540    
1541        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1542          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s\n",          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
1543            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1544            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1545            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1546            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1547            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1548            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1549              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1550              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1551            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1552            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1553            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1554              ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1555            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1556            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1557              ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1558    
1559          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1560    
1561        if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)        switch (get_options & PCRE_NEWLINE_BITS)
1562          fprintf(outfile, "Case state changes\n");          {
1563            case PCRE_NEWLINE_CR:
1564            fprintf(outfile, "Forced newline sequence: CR\n");
1565            break;
1566    
1567            case PCRE_NEWLINE_LF:
1568            fprintf(outfile, "Forced newline sequence: LF\n");
1569            break;
1570    
1571            case PCRE_NEWLINE_CRLF:
1572            fprintf(outfile, "Forced newline sequence: CRLF\n");
1573            break;
1574    
1575            case PCRE_NEWLINE_ANYCRLF:
1576            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1577            break;
1578    
1579            case PCRE_NEWLINE_ANY:
1580            fprintf(outfile, "Forced newline sequence: ANY\n");
1581            break;
1582    
1583            default:
1584            break;
1585            }
1586    
1587        if (first_char == -1)        if (first_char == -1)
1588          {          {
1589          fprintf(outfile, "First char at start or follows \\n\n");          fprintf(outfile, "First char at start or follows newline\n");
1590          }          }
1591        else if (first_char < 0)        else if (first_char < 0)
1592          {          {
# Line 1063  while (!done) Line 1597  while (!done)
1597          int ch = first_char & 255;          int ch = first_char & 255;
1598          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
1599            "" : " (caseless)";            "" : " (caseless)";
1600          if (isprint(ch))          if (PRINTHEX(ch))
1601            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
1602          else          else
1603            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1078  while (!done) Line 1612  while (!done)
1612          int ch = need_char & 255;          int ch = need_char & 255;
1613          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
1614            "" : " (caseless)";            "" : " (caseless)";
1615          if (isprint(ch))          if (PRINTHEX(ch))
1616            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
1617          else          else
1618            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1096  while (!done) Line 1630  while (!done)
1630          else          else
1631            {            {
1632            uschar *start_bits = NULL;            uschar *start_bits = NULL;
1633            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
1634    
1635              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
1636              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
1637    
1638              new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
1639            if (start_bits == NULL)            if (start_bits == NULL)
1640              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
1641            else            else
1642              {              {
1643              int i;              int i;
# Line 1114  while (!done) Line 1652  while (!done)
1652                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
1653                    c = 2;                    c = 2;
1654                    }                    }
1655                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
1656                    {                    {
1657                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
1658                    c += 2;                    c += 2;
# Line 1146  while (!done) Line 1684  while (!done)
1684        else        else
1685          {          {
1686          uschar sbuf[8];          uschar sbuf[8];
1687          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1688          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1689          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1690          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1691    
1692          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1693          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1694          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1695          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1696    
1697          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1698              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1173  while (!done) Line 1711  while (!done)
1711                  strerror(errno));                  strerror(errno));
1712                }                }
1713              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
1714    
1715              }              }
1716            }            }
1717          fclose(f);          fclose(f);
# Line 1189  while (!done) Line 1728  while (!done)
1728    
1729    for (;;)    for (;;)
1730      {      {
1731      unsigned char *q;      uschar *q;
1732      unsigned char *bptr = dbuffer;      uschar *bptr;
1733      int *use_offsets = offsets;      int *use_offsets = offsets;
1734      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1735      int callout_data = 0;      int callout_data = 0;
1736      int callout_data_set = 0;      int callout_data_set = 0;
1737      int count, c;      int count, c;
1738      int copystrings = 0;      int copystrings = 0;
1739      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
1740      int getstrings = 0;      int getstrings = 0;
1741      int getlist = 0;      int getlist = 0;
1742      int gmatched = 0;      int gmatched = 0;
# Line 1207  while (!done) Line 1746  while (!done)
1746    
1747      options = 0;      options = 0;
1748    
1749        *copynames = 0;
1750        *getnames = 0;
1751    
1752        copynamesptr = copynames;
1753        getnamesptr = getnames;
1754    
1755      pcre_callout = callout;      pcre_callout = callout;
1756      first_callout = 1;      first_callout = 1;
1757      callout_extra = 0;      callout_extra = 0;
# Line 1215  while (!done) Line 1760  while (!done)
1760      callout_fail_id = -1;      callout_fail_id = -1;
1761      show_malloc = 0;      show_malloc = 0;
1762    
1763      if (infile == stdin) printf("data> ");      if (extra != NULL) extra->flags &=
1764      if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
1765    
1766        len = 0;
1767        for (;;)
1768        {        {
1769        done = 1;        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
1770        goto CONTINUE;          {
1771            if (len > 0) break;
1772            done = 1;
1773            goto CONTINUE;
1774            }
1775          if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1776          len = (int)strlen((char *)buffer);
1777          if (buffer[len-1] == '\n') break;
1778        }        }
     if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);  
1779    
     len = (int)strlen((char *)buffer);  
1780      while (len > 0 && isspace(buffer[len-1])) len--;      while (len > 0 && isspace(buffer[len-1])) len--;
1781      buffer[len] = 0;      buffer[len] = 0;
1782      if (len == 0) break;      if (len == 0) break;
# Line 1231  while (!done) Line 1784  while (!done)
1784      p = buffer;      p = buffer;
1785      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1786    
1787      q = dbuffer;      bptr = q = dbuffer;
1788      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1789        {        {
1790        int i = 0;        int i = 0;
# Line 1253  while (!done) Line 1806  while (!done)
1806          c -= '0';          c -= '0';
1807          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
1808            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
1809    
1810    #if !defined NOUTF8
1811            if (use_utf8 && c > 255)
1812              {
1813              unsigned char buff8[8];
1814              int ii, utn;
1815              utn = ord2utf8(c, buff8);
1816              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1817              c = buff8[ii];   /* Last byte */
1818              }
1819    #endif
1820          break;          break;
1821    
1822          case 'x':          case 'x':
1823    
1824          /* Handle \x{..} specially - new Perl thing for utf8 */          /* Handle \x{..} specially - new Perl thing for utf8 */
1825    
1826    #if !defined NOUTF8
1827          if (*p == '{')          if (*p == '{')
1828            {            {
1829            unsigned char *pt = p;            unsigned char *pt = p;
# Line 1269  while (!done) Line 1834  while (!done)
1834              {              {
1835              unsigned char buff8[8];              unsigned char buff8[8];
1836              int ii, utn;              int ii, utn;
1837              utn = _pcre_ord2utf8(c, buff8);              if (use_utf8)
1838              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
1839              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
1840                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1841                  c = buff8[ii];   /* Last byte */
1842                  }
1843                else
1844                 {
1845                 if (c > 255)
1846                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
1847                     "UTF-8 mode is not enabled.\n"
1848                     "** Truncation will probably give the wrong result.\n", c);
1849                 }
1850              p = pt + 1;              p = pt + 1;
1851              break;              break;
1852              }              }
1853            /* Not correct form; fall through */            /* Not correct form; fall through */
1854            }            }
1855    #endif
1856    
1857          /* Ordinary \x */          /* Ordinary \x */
1858    
# Line 1312  while (!done) Line 1888  while (!done)
1888            }            }
1889          else if (isalnum(*p))          else if (isalnum(*p))
1890            {            {
1891            uschar name[256];            uschar *npp = copynamesptr;
           uschar *npp = name;  
1892            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
1893              *npp++ = 0;
1894            *npp = 0;            *npp = 0;
1895            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)copynamesptr);
1896            if (n < 0)            if (n < 0)
1897              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
1898            else copystrings |= 1 << n;            copynamesptr = npp;
1899            }            }
1900          else if (*p == '+')          else if (*p == '+')
1901            {            {
# Line 1357  while (!done) Line 1933  while (!done)
1933            }            }
1934          continue;          continue;
1935    
1936    #if !defined NODFA
1937          case 'D':          case 'D':
1938    #if !defined NOPOSIX
1939          if (posix || do_posix)          if (posix || do_posix)
1940            printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");            printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");
1941          else          else
1942    #endif
1943            use_dfa = 1;            use_dfa = 1;
1944          continue;          continue;
1945    
1946          case 'F':          case 'F':
1947          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
1948          continue;          continue;
1949    #endif
1950    
1951          case 'G':          case 'G':
1952          if (isdigit(*p))          if (isdigit(*p))
# Line 1376  while (!done) Line 1956  while (!done)
1956            }            }
1957          else if (isalnum(*p))          else if (isalnum(*p))
1958            {            {
1959            uschar name[256];            uschar *npp = getnamesptr;
           uschar *npp = name;  
1960            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
1961              *npp++ = 0;
1962            *npp = 0;            *npp = 0;
1963            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)getnamesptr);
1964            if (n < 0)            if (n < 0)
1965              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
1966            else getstrings |= 1 << n;            getnamesptr = npp;
1967            }            }
1968          continue;          continue;
1969    
# Line 1396  while (!done) Line 1976  while (!done)
1976          continue;          continue;
1977    
1978          case 'N':          case 'N':
1979          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
1980              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
1981            else
1982              options |= PCRE_NOTEMPTY;
1983          continue;          continue;
1984    
1985          case 'O':          case 'O':
# Line 1409  while (!done) Line 1992  while (!done)
1992            if (offsets == NULL)            if (offsets == NULL)
1993              {              {
1994              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1995                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1996              yield = 1;              yield = 1;
1997              goto EXIT;              goto EXIT;
1998              }              }
# Line 1419  while (!done) Line 2002  while (!done)
2002          continue;          continue;
2003    
2004          case 'P':          case 'P':
2005          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2006              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2007            continue;
2008    
2009            case 'Q':
2010            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2011            if (extra == NULL)
2012              {
2013              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2014              extra->flags = 0;
2015              }
2016            extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
2017            extra->match_limit_recursion = n;
2018            continue;
2019    
2020            case 'q':
2021            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2022            if (extra == NULL)
2023              {
2024              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2025              extra->flags = 0;
2026              }
2027            extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
2028            extra->match_limit = n;
2029          continue;          continue;
2030    
2031    #if !defined NODFA
2032          case 'R':          case 'R':
2033          options |= PCRE_DFA_RESTART;          options |= PCRE_DFA_RESTART;
2034          continue;          continue;
2035    #endif
2036    
2037          case 'S':          case 'S':
2038          show_malloc = 1;          show_malloc = 1;
2039          continue;          continue;
2040    
2041            case 'Y':
2042            options |= PCRE_NO_START_OPTIMIZE;
2043            continue;
2044    
2045          case 'Z':          case 'Z':
2046          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2047          continue;          continue;
# Line 1437  while (!done) Line 2049  while (!done)
2049          case '?':          case '?':
2050          options |= PCRE_NO_UTF8_CHECK;          options |= PCRE_NO_UTF8_CHECK;
2051          continue;          continue;
2052    
2053            case '<':
2054              {
2055              int x = check_newline(p, outfile);
2056              if (x == 0) goto NEXT_DATA;
2057              options |= x;
2058              while (*p++ != '>');
2059              }
2060            continue;
2061          }          }
2062        *q++ = c;        *q++ = c;
2063        }        }
2064      *q = 0;      *q = 0;
2065      len = q - dbuffer;      len = q - dbuffer;
2066    
2067        /* Move the data to the end of the buffer so that a read over the end of
2068        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2069        we are using the POSIX interface, we must include the terminating zero. */
2070    
2071    #if !defined NOPOSIX
2072        if (posix || do_posix)
2073          {
2074          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2075          bptr += buffer_size - len - 1;
2076          }
2077        else
2078    #endif
2079          {
2080          memmove(bptr + buffer_size - len, bptr, len);
2081          bptr += buffer_size - len;
2082          }
2083    
2084      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2085        {        {
2086        printf("**Match limit not relevant for DFA matching: ignored\n");        printf("**Match limit not relevant for DFA matching: ignored\n");
# Line 1462  while (!done) Line 2100  while (!done)
2100          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2101        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2102        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2103          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2104    
2105        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2106    
2107        if (rc != 0)        if (rc != 0)
2108          {          {
2109          (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);          (void)regerror(rc, &preg, (char *)buffer, buffer_size);
2110          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
2111          }          }
2112          else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
2113                  != 0)
2114            {
2115            fprintf(outfile, "Matched with REG_NOSUB\n");
2116            }
2117        else        else
2118          {          {
2119          size_t i;          size_t i;
# Line 1501  while (!done) Line 2145  while (!done)
2145    
2146      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2147        {        {
2148        if (timeit)        if (timeitm > 0)
2149          {          {
2150          register int i;          register int i;
2151          clock_t time_taken;          clock_t time_taken;
2152          clock_t start_time = clock();          clock_t start_time = clock();
2153    
2154    #if !defined NODFA
2155          if (all_use_dfa || use_dfa)          if (all_use_dfa || use_dfa)
2156            {            {
2157            int workspace[1000];            int workspace[1000];
2158            for (i = 0; i < LOOPREPEAT; i++)            for (i = 0; i < timeitm; i++)
2159              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2160                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2161                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2162            }            }
2163          else          else
2164    #endif
2165    
2166          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeitm; i++)
2167            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2168              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2169    
2170          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2171          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2172            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2173              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2174          }          }
2175    
2176        /* 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
2177        varying limits in order to find the minimum value. */        varying limits in order to find the minimum value for the match limit and
2178          for the recursion limit. */
2179    
2180        if (find_match_limit)        if (find_match_limit)
2181          {          {
         int min = 0;  
         int mid = 64;  
         int max = -1;  
   
2182          if (extra == NULL)          if (extra == NULL)
2183            {            {
2184            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2185            extra->flags = 0;            extra->flags = 0;
2186            }            }
         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 */  
           }  
2187    
2188          extra->flags &= ~PCRE_EXTRA_MATCH_LIMIT;          (void)check_match_limit(re, extra, bptr, len, start_offset,
2189              options|g_notempty, use_offsets, use_size_offsets,
2190              PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
2191              PCRE_ERROR_MATCHLIMIT, "match()");
2192    
2193            count = check_match_limit(re, extra, bptr, len, start_offset,
2194              options|g_notempty, use_offsets, use_size_offsets,
2195              PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
2196              PCRE_ERROR_RECURSIONLIMIT, "match() recursion");
2197          }          }
2198    
2199        /* If callout_data is set, use the interface with additional data */        /* If callout_data is set, use the interface with additional data */
# Line 1591  while (!done) Line 2215  while (!done)
2215        /* 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
2216        value of match_limit. */        value of match_limit. */
2217    
2218    #if !defined NODFA
2219        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2220          {          {
2221          int workspace[1000];          int workspace[1000];
2222          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2223            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2224            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2225          if (count == 0)          if (count == 0)
# Line 1603  while (!done) Line 2228  while (!done)
2228            count = use_size_offsets/2;            count = use_size_offsets/2;
2229            }            }
2230          }          }
2231    #endif
2232    
2233        else        else
2234          {          {
# Line 1619  while (!done) Line 2245  while (!done)
2245    
2246        if (count >= 0)        if (count >= 0)
2247          {          {
2248          int i;          int i, maxcount;
2249    
2250    #if !defined NODFA
2251            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2252    #endif
2253              maxcount = use_size_offsets/3;
2254    
2255            /* This is a check against a lunatic return value. */
2256    
2257            if (count > maxcount)
2258              {
2259              fprintf(outfile,
2260                "** PCRE error: returned count %d is too big for offset size %d\n",
2261                count, use_size_offsets);
2262              count = use_size_offsets/3;
2263              if (do_g || do_G)
2264                {
2265                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2266                do_g = do_G = FALSE;        /* Break g/G loop */
2267                }
2268              }
2269    
2270          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2271            {            {
2272            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
# Line 1647  while (!done) Line 2294  while (!done)
2294            {            {
2295            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
2296              {              {
2297              char copybuffer[16];              char copybuffer[256];
2298              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,
2299                i, copybuffer, sizeof(copybuffer));                i, copybuffer, sizeof(copybuffer));
2300              if (rc < 0)              if (rc < 0)
# Line 1657  while (!done) Line 2304  while (!done)
2304              }              }
2305            }            }
2306    
2307            for (copynamesptr = copynames;
2308                 *copynamesptr != 0;
2309                 copynamesptr += (int)strlen((char*)copynamesptr) + 1)
2310              {
2311              char copybuffer[256];
2312              int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,
2313                count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));
2314              if (rc < 0)
2315                fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);
2316              else
2317                fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);
2318              }
2319    
2320          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2321            {            {
2322            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
# Line 1669  while (!done) Line 2329  while (!done)
2329              else              else
2330                {                {
2331                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
               /* free((void *)substring); */  
2332                pcre_free_substring(substring);                pcre_free_substring(substring);
2333                }                }
2334              }              }
2335            }            }
2336    
2337            for (getnamesptr = getnames;
2338                 *getnamesptr != 0;
2339                 getnamesptr += (int)strlen((char*)getnamesptr) + 1)
2340              {
2341              const char *substring;
2342              int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,
2343                count, (char *)getnamesptr, &substring);
2344              if (rc < 0)
2345                fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);
2346              else
2347                {
2348                fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);
2349                pcre_free_substring(substring);
2350                }
2351              }
2352    
2353          if (getlist)          if (getlist)
2354            {            {
2355            const char **stringlist;            const char **stringlist;
# Line 1699  while (!done) Line 2374  while (!done)
2374        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2375          {          {
2376          fprintf(outfile, "Partial match");          fprintf(outfile, "Partial match");
2377          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)          if (use_size_offsets > 1)
2378            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            {
2379              bptr + use_offsets[0]);            fprintf(outfile, ": ");
2380              pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2381                outfile);
2382              }
2383          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2384          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2385          }          }
2386    
2387        /* 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
2388        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
2389        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
2390        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2391        offset values to achieve this. We won't be at the end of the string -  
2392        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2393          "anycrlf". If the previous match was at the end of a line terminated by
2394          CRLF, an advance of one character just passes the \r, whereas we should
2395          prefer the longer newline sequence, as does the code in pcre_exec().
2396          Fudge the offset value to achieve this.
2397    
2398          Otherwise, in the case of UTF-8 matching, the advance must be one
2399          character, not one byte. */
2400    
2401        else        else
2402          {          {
2403          if (g_notempty != 0)          if (g_notempty != 0)
2404            {            {
2405            int onechar = 1;            int onechar = 1;
2406              unsigned int obits = ((real_pcre *)re)->options;
2407            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2408            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2409                {
2410                int d;
2411                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2412                /* Note that these values are always the ASCII ones, even in
2413                EBCDIC environments. CR = 13, NL = 10. */
2414                obits = (d == 13)? PCRE_NEWLINE_CR :
2415                        (d == 10)? PCRE_NEWLINE_LF :
2416                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2417                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2418                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2419                }
2420              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2421                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2422                  &&
2423                  start_offset < len - 1 &&
2424                  bptr[start_offset] == '\r' &&
2425                  bptr[start_offset+1] == '\n')
2426                onechar++;
2427              else if (use_utf8)
2428              {              {
2429              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2430                {                {
# Line 1747  while (!done) Line 2452  while (!done)
2452        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2453    
2454        /* 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
2455        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
2456        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
2457        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
2458        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
2459        character. */        character. */
2460    
2461        g_notempty = 0;        g_notempty = 0;
2462    
2463        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2464          {          {
2465          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2466          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2467          }          }
2468    
2469        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 1772  while (!done) Line 2478  while (!done)
2478          len -= use_offsets[1];          len -= use_offsets[1];
2479          }          }
2480        }  /* End of loop for /g and /G */        }  /* End of loop for /g and /G */
2481    
2482        NEXT_DATA: continue;
2483      }    /* End of loop for data lines */      }    /* End of loop for data lines */
2484    
2485    CONTINUE:    CONTINUE:
# Line 1786  while (!done) Line 2494  while (!done)
2494      {      {
2495      new_free((void *)tables);      new_free((void *)tables);
2496      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
2497        locale_set = 0;
2498      }      }
2499    }    }
2500    

Legend:
Removed from v.77  
changed lines
  Added in v.507

  ViewVC Help
Powered by ViewVC 1.1.5