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

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

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

revision 75 by nigel, Sat Feb 24 21:40:37 2007 UTC revision 427 by ph10, Fri Aug 28 09:55:54 2009 UTC
# Line 4  Line 4 
4    
5  /* This program was hacked up as a tester for PCRE. I really should have  /* This program was hacked up as a tester for PCRE. I really should have
6  written it more tidily in the first place. Will I ever learn? It has grown and  written it more tidily in the first place. Will I ever learn? It has grown and
7  been extended and consequently is now rather untidy in places.  been extended and consequently is now rather, er, *very* untidy in places.
8    
9  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
10  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 44  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51  /* We need the internal info for displaying the results of pcre_study(). Also  #ifdef SUPPORT_LIBREADLINE
52  for getting the opcodes for showing compiled code. */  #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57    #endif
58    
59    
60    /* A number of things vary for Windows builds. Originally, pcretest opened its
61    input and output without "b"; then I was told that "b" was needed in some
62    environments, so it was added for release 5.0 to both the input and output. (It
63    makes no difference on Unix-like systems.) Later I was told that it is wrong
64    for the input on Windows. I've now abstracted the modes into two macros that
65    are set here, to make it easier to fiddle with them, and removed "b" from the
66    input mode under Windows. */
67    
68    #if defined(_WIN32) || defined(WIN32)
69    #include <io.h>                /* For _setmode() */
70    #include <fcntl.h>             /* For _O_BINARY */
71    #define INPUT_MODE   "r"
72    #define OUTPUT_MODE  "wb"
73    
74    #ifndef isatty
75    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
76    #endif                         /* though in some environments they seem to   */
77                                   /* be already defined, hence the #ifndefs.    */
78    #ifndef fileno
79    #define fileno _fileno
80    #endif
81    
82    #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"
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.
122    
123    The definition of the macro PRINTABLE, which determines whether to print an
124    output character as-is or as a hex value when showing compiled patterns, is
125    contained in this file. We uses it here also, in cases when the locale has not
126    been explicitly changed, so as to get consistent output from systems that
127    differ in their output from isprint() even in the "C" locale. */
128    
129    #include "pcre_printint.src"
130    
131    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
132    
 #define PCRE_SPY        /* For Win32 build, import data, not export */  
 #include "internal.h"  
133    
134  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
135  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 58  Makefile. */ Line 139  Makefile. */
139  #include "pcreposix.h"  #include "pcreposix.h"
140  #endif  #endif
141    
142    /* It is also possible, for the benefit of the version currently imported into
143    Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
144    interface to the DFA matcher (NODFA), and without the doublecheck of the old
145    "info" function (define NOINFOCHECK). In fact, we automatically cut out the
146    UTF8 support if PCRE is built without it. */
147    
148    #ifndef SUPPORT_UTF8
149    #ifndef NOUTF8
150    #define NOUTF8
151    #endif
152    #endif
153    
154    
155    /* Other parameters */
156    
157  #ifndef CLOCKS_PER_SEC  #ifndef CLOCKS_PER_SEC
158  #ifdef CLK_TCK  #ifdef CLK_TCK
159  #define CLOCKS_PER_SEC CLK_TCK  #define CLOCKS_PER_SEC CLK_TCK
# Line 66  Makefile. */ Line 162  Makefile. */
162  #endif  #endif
163  #endif  #endif
164    
165  #define LOOPREPEAT 500000  /* This is the default loop count for timing. */
166    
167  #define BUFFER_SIZE 30000  #define LOOPREPEAT 500000
 #define PBUFFER_SIZE BUFFER_SIZE  
 #define DBUFFER_SIZE BUFFER_SIZE  
168    
169    /* Static variables */
170    
171  static FILE *outfile;  static FILE *outfile;
172  static int log_store = 0;  static int log_store = 0;
# Line 79  static int callout_count; Line 174  static int callout_count;
174  static int callout_extra;  static int callout_extra;
175  static int callout_fail_count;  static int callout_fail_count;
176  static int callout_fail_id;  static int callout_fail_id;
177    static int debug_lengths;
178  static int first_callout;  static int first_callout;
179    static int locale_set = 0;
180  static int show_malloc;  static int show_malloc;
181  static int use_utf8;  static int use_utf8;
182  static size_t gotten_store;  static size_t gotten_store;
183    
184    /* The buffers grow automatically if very long input lines are encountered. */
185    
186    static int buffer_size = 50000;
187    static uschar *buffer = NULL;
188    static uschar *dbuffer = NULL;
189  static uschar *pbuffer = NULL;  static uschar *pbuffer = NULL;
190    
191    
 static const int utf8_table1[] = {  
   0x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff};  
192    
193  static const int utf8_table2[] = {  /*************************************************
194    0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc};  *        Read or extend an input line            *
195    *************************************************/
196    
197  static const int utf8_table3[] = {  /* Input lines are read into buffer, but both patterns and data lines can be
198    0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};  continued over multiple input lines. In addition, if the buffer fills up, we
199    want to automatically expand it so as to be able to handle extremely large
200    lines that are needed for certain stress tests. When the input buffer is
201    expanded, the other two buffers must also be expanded likewise, and the
202    contents of pbuffer, which are a copy of the input for callouts, must be
203    preserved (for when expansion happens for a data line). This is not the most
204    optimal way of handling this, but hey, this is just a test program!
205    
206    Arguments:
207      f            the file to read
208      start        where in buffer to start (this *must* be within buffer)
209      prompt       for stdin or readline()
210    
211    Returns:       pointer to the start of new data
212                   could be a copy of start, or could be moved
213                   NULL if no data read and EOF reached
214    */
215    
216    static uschar *
217    extend_inputline(FILE *f, uschar *start, const char *prompt)
218    {
219    uschar *here = start;
220    
221    for (;;)
222      {
223      int rlen = buffer_size - (here - buffer);
224    
225      if (rlen > 1000)
226        {
227        int dlen;
228    
229        /* If libreadline support is required, use readline() to read a line if the
230        input is a terminal. Note that readline() removes the trailing newline, so
231        we must put it back again, to be compatible with fgets(). */
232    
233    #ifdef SUPPORT_LIBREADLINE
234        if (isatty(fileno(f)))
235          {
236          size_t len;
237          char *s = readline(prompt);
238          if (s == NULL) return (here == start)? NULL : start;
239          len = strlen(s);
240          if (len > 0) add_history(s);
241          if (len > rlen - 1) len = rlen - 1;
242          memcpy(here, s, len);
243          here[len] = '\n';
244          here[len+1] = 0;
245          free(s);
246          }
247        else
248    #endif
249    
250        /* Read the next line by normal means, prompting if the file is stdin. */
251    
252          {
253          if (f == stdin) printf(prompt);
254          if (fgets((char *)here, rlen,  f) == NULL)
255            return (here == start)? NULL : start;
256          }
257    
258        dlen = (int)strlen((char *)here);
259        if (dlen > 0 && here[dlen - 1] == '\n') return start;
260        here += dlen;
261        }
262    
263      else
264        {
265        int new_buffer_size = 2*buffer_size;
266        uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);
267        uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
268        uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);
269    
270        if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
271          {
272          fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
273          exit(1);
274          }
275    
276        memcpy(new_buffer, buffer, buffer_size);
277        memcpy(new_pbuffer, pbuffer, buffer_size);
278    
279        buffer_size = new_buffer_size;
280    
281        start = new_buffer + (start - buffer);
282        here = new_buffer + (here - buffer);
283    
284        free(buffer);
285        free(dbuffer);
286        free(pbuffer);
287    
288        buffer = new_buffer;
289        dbuffer = new_dbuffer;
290        pbuffer = new_pbuffer;
291        }
292      }
293    
294    return NULL;  /* Control never gets here */
295    }
296    
297    
298    
 /*************************************************  
 *         Print compiled regex                   *  
 *************************************************/  
299    
 /* The code for doing this is held in a separate file that is also included in  
 pcre.c when it is compiled with the debug switch. It defines a function called  
 print_internals(), which uses a table of opcode lengths defined by the macro  
 OP_LENGTHS, whose name must be OP_lengths. It also uses a table that translates  
 Unicode property names to numbers; this is kept in a separate file. */  
   
 static uschar OP_lengths[] = { OP_LENGTHS };  
   
 #include "ucp.h"  
 #include "ucptypetable.c"  
 #include "printint.c"  
300    
301    
302    
# Line 122  static uschar OP_lengths[] = { OP_LENGTH Line 306  static uschar OP_lengths[] = { OP_LENGTH
306    
307  /* 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
308  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
309  for unpicking the -o argument, so just keep it simple.  for unpicking arguments, so just keep it simple.
310    
311  Arguments:  Arguments:
312    str           string to be converted    str           string to be converted
# Line 143  return(result); Line 327  return(result);
327    
328    
329    
 /*************************************************  
 *       Convert character value to UTF-8         *  
 *************************************************/  
   
 /* This function takes an integer value in the range 0 - 0x7fffffff  
 and encodes it as a UTF-8 character in 0 to 6 bytes.  
   
 Arguments:  
   cvalue     the character value  
   buffer     pointer to buffer for result - at least 6 bytes long  
   
 Returns:     number of characters placed in the buffer  
              -1 if input character is negative  
              0 if input character is positive but too big (only when  
              int is longer than 32 bits)  
 */  
   
 static int  
 ord2utf8(int cvalue, unsigned char *buffer)  
 {  
 register int i, j;  
 for (i = 0; i < sizeof(utf8_table1)/sizeof(int); i++)  
   if (cvalue <= utf8_table1[i]) break;  
 if (i >= sizeof(utf8_table1)/sizeof(int)) return 0;  
 if (cvalue < 0) return -1;  
   
 buffer += i;  
 for (j = i; j > 0; j--)  
  {  
  *buffer-- = 0x80 | (cvalue & 0x3f);  
  cvalue >>= 6;  
  }  
 *buffer = utf8_table2[i] | cvalue;  
 return i + 1;  
 }  
   
330    
331  /*************************************************  /*************************************************
332  *            Convert UTF-8 string to value       *  *            Convert UTF-8 string to value       *
# Line 188  return i + 1; Line 336  return i + 1;
336  and returns the value of the character.  and returns the value of the character.
337    
338  Argument:  Argument:
339    buffer   a pointer to the byte vector    utf8bytes   a pointer to the byte vector
340    vptr     a pointer to an int to receive the value    vptr        a pointer to an int to receive the value
341    
342  Returns:   >  0 => the number of bytes consumed  Returns:      >  0 => the number of bytes consumed
343             -6 to 0 => malformed UTF-8 character at offset = (-return)                -6 to 0 => malformed UTF-8 character at offset = (-return)
344  */  */
345    
346    #if !defined NOUTF8
347    
348  static int  static int
349  utf82ord(unsigned char *buffer, int *vptr)  utf82ord(unsigned char *utf8bytes, int *vptr)
350  {  {
351  int c = *buffer++;  int c = *utf8bytes++;
352  int d = c;  int d = c;
353  int i, j, s;  int i, j, s;
354    
# Line 218  d = (c & utf8_table3[i]) << s; Line 368  d = (c & utf8_table3[i]) << s;
368    
369  for (j = 0; j < i; j++)  for (j = 0; j < i; j++)
370    {    {
371    c = *buffer++;    c = *utf8bytes++;
372    if ((c & 0xc0) != 0x80) return -(j+1);    if ((c & 0xc0) != 0x80) return -(j+1);
373    s -= 6;    s -= 6;
374    d |= (c & 0x3f) << s;    d |= (c & 0x3f) << s;
# Line 226  for (j = 0; j < i; j++) Line 376  for (j = 0; j < i; j++)
376    
377  /* Check that encoding was the correct unique one */  /* Check that encoding was the correct unique one */
378    
379  for (j = 0; j < sizeof(utf8_table1)/sizeof(int); j++)  for (j = 0; j < utf8_table1_size; j++)
380    if (d <= utf8_table1[j]) break;    if (d <= utf8_table1[j]) break;
381  if (j != i) return -(i+1);  if (j != i) return -(i+1);
382    
# Line 236  if (j != i) return -(i+1); Line 386  if (j != i) return -(i+1);
386  return i+1;  return i+1;
387  }  }
388    
389    #endif
390    
391    
392    
393    /*************************************************
394    *       Convert character value to UTF-8         *
395    *************************************************/
396    
397    /* This function takes an integer value in the range 0 - 0x7fffffff
398    and encodes it as a UTF-8 character in 0 to 6 bytes.
399    
400    Arguments:
401      cvalue     the character value
402      utf8bytes  pointer to buffer for result - at least 6 bytes long
403    
404    Returns:     number of characters placed in the buffer
405    */
406    
407    #if !defined NOUTF8
408    
409    static int
410    ord2utf8(int cvalue, uschar *utf8bytes)
411    {
412    register int i, j;
413    for (i = 0; i < utf8_table1_size; i++)
414      if (cvalue <= utf8_table1[i]) break;
415    utf8bytes += i;
416    for (j = i; j > 0; j--)
417     {
418     *utf8bytes-- = 0x80 | (cvalue & 0x3f);
419     cvalue >>= 6;
420     }
421    *utf8bytes = utf8_table2[i] | cvalue;
422    return i + 1;
423    }
424    
425    #endif
426    
427    
428    
429  /*************************************************  /*************************************************
# Line 248  chars without printing. */ Line 436  chars without printing. */
436    
437  static int pchars(unsigned char *p, int length, FILE *f)  static int pchars(unsigned char *p, int length, FILE *f)
438  {  {
439  int c;  int c = 0;
440  int yield = 0;  int yield = 0;
441    
442  while (length-- > 0)  while (length-- > 0)
443    {    {
444    #if !defined NOUTF8
445    if (use_utf8)    if (use_utf8)
446      {      {
447      int rc = utf82ord(p, &c);      int rc = utf82ord(p, &c);
# Line 261  while (length-- > 0) Line 450  while (length-- > 0)
450        {        {
451        length -= rc - 1;        length -= rc - 1;
452        p += rc;        p += rc;
453        if (c < 256 && isprint(c))        if (PRINTHEX(c))
454          {          {
455          if (f != NULL) fprintf(f, "%c", c);          if (f != NULL) fprintf(f, "%c", c);
456          yield++;          yield++;
457          }          }
458        else        else
459          {          {
460          int n;          int n = 4;
461          if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);          if (f != NULL) fprintf(f, "\\x{%02x}", c);
462          yield += n;          yield += (n <= 0x000000ff)? 2 :
463                     (n <= 0x00000fff)? 3 :
464                     (n <= 0x0000ffff)? 4 :
465                     (n <= 0x000fffff)? 5 : 6;
466          }          }
467        continue;        continue;
468        }        }
469      }      }
470    #endif
471    
472     /* Not UTF-8, or malformed UTF-8  */     /* Not UTF-8, or malformed UTF-8  */
473    
474    if (isprint(c = *(p++)))    c = *p++;
475      if (PRINTHEX(c))
476      {      {
477      if (f != NULL) fprintf(f, "%c", c);      if (f != NULL) fprintf(f, "%c", c);
478      yield++;      yield++;
# Line 403  static void *new_malloc(size_t size) Line 597  static void *new_malloc(size_t size)
597  void *block = malloc(size);  void *block = malloc(size);
598  gotten_store = size;  gotten_store = size;
599  if (show_malloc)  if (show_malloc)
600    fprintf(outfile, "malloc       %3d %p\n", size, block);    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);
601  return block;  return block;
602  }  }
603    
# Line 421  static void *stack_malloc(size_t size) Line 615  static void *stack_malloc(size_t size)
615  {  {
616  void *block = malloc(size);  void *block = malloc(size);
617  if (show_malloc)  if (show_malloc)
618    fprintf(outfile, "stack_malloc %3d %p\n", size, block);    fprintf(outfile, "stack_malloc %3d %p\n", (int)size, block);
619  return block;  return block;
620  }  }
621    
# Line 452  if ((rc = pcre_fullinfo(re, study, optio Line 646  if ((rc = pcre_fullinfo(re, study, optio
646  *         Byte flipping function                 *  *         Byte flipping function                 *
647  *************************************************/  *************************************************/
648    
649  static long int  static unsigned long int
650  byteflip(long int value, int n)  byteflip(unsigned long int value, int n)
651  {  {
652  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
653  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
# Line 466  return ((value & 0x000000ff) << 24) | Line 660  return ((value & 0x000000ff) << 24) |
660    
661    
662  /*************************************************  /*************************************************
663    *        Check match or recursion limit          *
664    *************************************************/
665    
666    static int
667    check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,
668      int start_offset, int options, int *use_offsets, int use_size_offsets,
669      int flag, unsigned long int *limit, int errnumber, const char *msg)
670    {
671    int count;
672    int min = 0;
673    int mid = 64;
674    int max = -1;
675    
676    extra->flags |= flag;
677    
678    for (;;)
679      {
680      *limit = mid;
681    
682      count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,
683        use_offsets, use_size_offsets);
684    
685      if (count == errnumber)
686        {
687        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
688        min = mid;
689        mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
690        }
691    
692      else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
693                             count == PCRE_ERROR_PARTIAL)
694        {
695        if (mid == min + 1)
696          {
697          fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
698          break;
699          }
700        /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
701        max = mid;
702        mid = (min + mid)/2;
703        }
704      else break;    /* Some other error */
705      }
706    
707    extra->flags &= ~flag;
708    return count;
709    }
710    
711    
712    
713    /*************************************************
714    *         Case-independent strncmp() function    *
715    *************************************************/
716    
717    /*
718    Arguments:
719      s         first string
720      t         second string
721      n         number of characters to compare
722    
723    Returns:    < 0, = 0, or > 0, according to the comparison
724    */
725    
726    static int
727    strncmpic(uschar *s, uschar *t, int n)
728    {
729    while (n--)
730      {
731      int c = tolower(*s++) - tolower(*t++);
732      if (c) return c;
733      }
734    return 0;
735    }
736    
737    
738    
739    /*************************************************
740    *         Check newline indicator                *
741    *************************************************/
742    
743    /* This is used both at compile and run-time to check for <xxx> escapes, where
744    xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is
745    no match.
746    
747    Arguments:
748      p           points after the leading '<'
749      f           file for error message
750    
751    Returns:      appropriate PCRE_NEWLINE_xxx flags, or 0
752    */
753    
754    static int
755    check_newline(uschar *p, FILE *f)
756    {
757    if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
758    if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
759    if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
760    if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
761    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
762    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
763    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
764    fprintf(f, "Unknown newline type at: <%s\n", p);
765    return 0;
766    }
767    
768    
769    
770    /*************************************************
771    *             Usage function                     *
772    *************************************************/
773    
774    static void
775    usage(void)
776    {
777    printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
778    printf("Input and output default to stdin and stdout.\n");
779    #ifdef SUPPORT_LIBREADLINE
780    printf("If input is a terminal, readline() is used to read from it.\n");
781    #else
782    printf("This version of pcretest is not linked with readline().\n");
783    #endif
784    printf("\nOptions:\n");
785    printf("  -b       show compiled code (bytecode)\n");
786    printf("  -C       show PCRE compile-time options and exit\n");
787    printf("  -d       debug: show compiled code and information (-b and -i)\n");
788    #if !defined NODFA
789    printf("  -dfa     force DFA matching for all subjects\n");
790    #endif
791    printf("  -help    show usage information\n");
792    printf("  -i       show information about compiled patterns\n"
793           "  -M       find MATCH_LIMIT minimum for each subject\n"
794           "  -m       output memory used information\n"
795           "  -o <n>   set size of offsets vector to <n>\n");
796    #if !defined NOPOSIX
797    printf("  -p       use POSIX interface\n");
798    #endif
799    printf("  -q       quiet: do not output PCRE version number at start\n");
800    printf("  -S <n>   set stack size to <n> megabytes\n");
801    printf("  -s       output store (memory) used information\n"
802           "  -t       time compilation and execution\n");
803    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
804    printf("  -tm      time execution (matching) only\n");
805    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
806    }
807    
808    
809    
810    /*************************************************
811  *                Main Program                    *  *                Main Program                    *
812  *************************************************/  *************************************************/
813    
# Line 478  int main(int argc, char **argv) Line 820  int main(int argc, char **argv)
820  FILE *infile = stdin;  FILE *infile = stdin;
821  int options = 0;  int options = 0;
822  int study_options = 0;  int study_options = 0;
823    int default_find_match_limit = FALSE;
824  int op = 1;  int op = 1;
825  int timeit = 0;  int timeit = 0;
826    int timeitm = 0;
827  int showinfo = 0;  int showinfo = 0;
828  int showstore = 0;  int showstore = 0;
829    int quiet = 0;
830  int size_offsets = 45;  int size_offsets = 45;
831  int size_offsets_max;  int size_offsets_max;
832  int *offsets;  int *offsets = NULL;
833  #if !defined NOPOSIX  #if !defined NOPOSIX
834  int posix = 0;  int posix = 0;
835  #endif  #endif
836  int debug = 0;  int debug = 0;
837  int done = 0;  int done = 0;
838    int all_use_dfa = 0;
839    int yield = 0;
840    int stack_size;
841    
842    /* These vectors store, end-to-end, a list of captured substring names. Assume
843    that 1024 is plenty long enough for the few names we'll be testing. */
844    
845    uschar copynames[1024];
846    uschar getnames[1024];
847    
848  unsigned char *buffer;  uschar *copynamesptr;
849  unsigned char *dbuffer;  uschar *getnamesptr;
850    
851  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* Get buffers from malloc() so that Electric Fence will check their misuse
852  when I am debugging. */  when I am debugging. They grow automatically when very long lines are read. */
853    
854  buffer = (unsigned char *)malloc(BUFFER_SIZE);  buffer = (unsigned char *)malloc(buffer_size);
855  dbuffer = (unsigned char *)malloc(DBUFFER_SIZE);  dbuffer = (unsigned char *)malloc(buffer_size);
856  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.  */  
857    
858  #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) */  
859    
860  outfile = stdout;  outfile = stdout;
861    
862    /* The following  _setmode() stuff is some Windows magic that tells its runtime
863    library to translate CRLF into a single LF character. At least, that's what
864    I've been told: never having used Windows I take this all on trust. Originally
865    it set 0x8000, but then I was advised that _O_BINARY was better. */
866    
867    #if defined(_WIN32) || defined(WIN32)
868    _setmode( _fileno( stdout ), _O_BINARY );
869    #endif
870    
871  /* Scan options */  /* Scan options */
872    
873  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
# Line 519  while (argc > 1 && argv[op][0] == '-') Line 876  while (argc > 1 && argv[op][0] == '-')
876    
877    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
878      showstore = 1;      showstore = 1;
879    else if (strcmp(argv[op], "-t") == 0) timeit = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
880      else if (strcmp(argv[op], "-b") == 0) debug = 1;
881    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
882    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
883      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
884    #if !defined NODFA
885      else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
886    #endif
887    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
888        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),
889          *endptr == 0))          *endptr == 0))
# Line 529  while (argc > 1 && argv[op][0] == '-') Line 891  while (argc > 1 && argv[op][0] == '-')
891      op++;      op++;
892      argc--;      argc--;
893      }      }
894      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
895        {
896        int both = argv[op][2] == 0;
897        int temp;
898        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
899                         *endptr == 0))
900          {
901          timeitm = temp;
902          op++;
903          argc--;
904          }
905        else timeitm = LOOPREPEAT;
906        if (both) timeit = timeitm;
907        }
908      else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
909          ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
910            *endptr == 0))
911        {
912    #if defined(_WIN32) || defined(WIN32)
913        printf("PCRE: -S not supported on this OS\n");
914        exit(1);
915    #else
916        int rc;
917        struct rlimit rlim;
918        getrlimit(RLIMIT_STACK, &rlim);
919        rlim.rlim_cur = stack_size * 1024 * 1024;
920        rc = setrlimit(RLIMIT_STACK, &rlim);
921        if (rc != 0)
922          {
923        printf("PCRE: setrlimit() failed with error %d\n", rc);
924        exit(1);
925          }
926        op++;
927        argc--;
928    #endif
929        }
930  #if !defined NOPOSIX  #if !defined NOPOSIX
931    else if (strcmp(argv[op], "-p") == 0) posix = 1;    else if (strcmp(argv[op], "-p") == 0) posix = 1;
932  #endif  #endif
933    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
934      {      {
935      int rc;      int rc;
936        unsigned long int lrc;
937      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
938      printf("Compiled with\n");      printf("Compiled with\n");
939      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 542  while (argc > 1 && argv[op][0] == '-') Line 941  while (argc > 1 && argv[op][0] == '-')
941      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
942      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
943      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
944      printf("  Newline character is %s\n", (rc == '\r')? "CR" : "LF");      /* Note that these values are always the ASCII values, even
945        in EBCDIC environments. CR is 13 and NL is 10. */
946        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
947          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
948          (rc == -2)? "ANYCRLF" :
949          (rc == -1)? "ANY" : "???");
950        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
951        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
952                                         "all Unicode newlines");
953      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
954      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
955      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
956      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
957      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
958      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
959        (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
960        printf("  Default recursion depth limit = %ld\n", lrc);
961      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
962      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
963      exit(0);      goto EXIT;
964        }
965      else if (strcmp(argv[op], "-help") == 0 ||
966               strcmp(argv[op], "--help") == 0)
967        {
968        usage();
969        goto EXIT;
970      }      }
971    else    else
972      {      {
973      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
974      printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");      usage();
975      printf("  -C     show PCRE compile-time options and exit\n");      yield = 1;
976      printf("  -d     debug: show compiled code; implies -i\n"      goto EXIT;
            "  -i     show information about compiled pattern\n"  
            "  -m     output memory used information\n"  
            "  -o <n> set size of offsets vector to <n>\n");  
 #if !defined NOPOSIX  
     printf("  -p     use POSIX interface\n");  
 #endif  
     printf("  -s     output store (memory) used information\n"  
            "  -t     time compilation and execution\n");  
     return 1;  
977      }      }
978    op++;    op++;
979    argc--;    argc--;
# Line 580  offsets = (int *)malloc(size_offsets_max Line 986  offsets = (int *)malloc(size_offsets_max
986  if (offsets == NULL)  if (offsets == NULL)
987    {    {
988    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
989      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
990    return 1;    yield = 1;
991      goto EXIT;
992    }    }
993    
994  /* Sort out the input and output files */  /* Sort out the input and output files */
995    
996  if (argc > 1)  if (argc > 1)
997    {    {
998    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
999    if (infile == NULL)    if (infile == NULL)
1000      {      {
1001      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
1002      return 1;      yield = 1;
1003        goto EXIT;
1004      }      }
1005    }    }
1006    
1007  if (argc > 2)  if (argc > 2)
1008    {    {
1009    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
1010    if (outfile == NULL)    if (outfile == NULL)
1011      {      {
1012      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
1013      return 1;      yield = 1;
1014        goto EXIT;
1015      }      }
1016    }    }
1017    
# Line 613  pcre_free = new_free; Line 1022  pcre_free = new_free;
1022  pcre_stack_malloc = stack_malloc;  pcre_stack_malloc = stack_malloc;
1023  pcre_stack_free = stack_free;  pcre_stack_free = stack_free;
1024    
1025  /* Heading line, then prompt for first regex if stdin */  /* Heading line unless quiet, then prompt for first regex if stdin */
1026    
1027  fprintf(outfile, "PCRE version %s\n\n", pcre_version());  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
1028    
1029  /* Main loop */  /* Main loop */
1030    
# Line 642  while (!done) Line 1051  while (!done)
1051    int do_showinfo = showinfo;    int do_showinfo = showinfo;
1052    int do_showrest = 0;    int do_showrest = 0;
1053    int do_flip = 0;    int do_flip = 0;
1054    int erroroffset, len, delimiter;    int erroroffset, len, delimiter, poffset;
1055    
1056    use_utf8 = 0;    use_utf8 = 0;
1057      debug_lengths = 1;
1058    
1059    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) break;  
1060    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1061    fflush(outfile);    fflush(outfile);
1062    
# Line 659  while (!done) Line 1068  while (!done)
1068    
1069    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
1070      {      {
1071      unsigned long int magic;      unsigned long int magic, get_options;
1072      uschar sbuf[8];      uschar sbuf[8];
1073      FILE *f;      FILE *f;
1074    
# Line 707  while (!done) Line 1116  while (!done)
1116    
1117      /* Need to know if UTF-8 for printing data strings */      /* Need to know if UTF-8 for printing data strings */
1118    
1119      new_info(re, NULL, PCRE_INFO_OPTIONS, &options);      new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1120      use_utf8 = (options & PCRE_UTF8) != 0;      use_utf8 = (get_options & PCRE_UTF8) != 0;
1121    
1122      /* Now see if there is any following study data */      /* Now see if there is any following study data */
1123    
# Line 747  while (!done) Line 1156  while (!done)
1156    
1157    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1158      {      {
1159      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1160      goto SKIP_DATA;      goto SKIP_DATA;
1161      }      }
1162    
1163    pp = p;    pp = p;
1164      poffset = p - buffer;
1165    
1166    for(;;)    for(;;)
1167      {      {
# Line 762  while (!done) Line 1172  while (!done)
1172        pp++;        pp++;
1173        }        }
1174      if (*pp != 0) break;      if (*pp != 0) break;
1175        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)  
1176        {        {
1177        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1178        done = 1;        done = 1;
# Line 780  while (!done) Line 1181  while (!done)
1181      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1182      }      }
1183    
1184      /* The buffer may have moved while being extended; reset the start of data
1185      pointer to the correct relative point in the buffer. */
1186    
1187      p = buffer + poffset;
1188    
1189    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1190    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1191    of testing for the error message when a pattern ends with backslash. */    of testing for the error message when a pattern ends with backslash. */
# Line 802  while (!done) Line 1208  while (!done)
1208      {      {
1209      switch (*pp++)      switch (*pp++)
1210        {        {
1211          case 'f': options |= PCRE_FIRSTLINE; break;
1212        case 'g': do_g = 1; break;        case 'g': do_g = 1; break;
1213        case 'i': options |= PCRE_CASELESS; break;        case 'i': options |= PCRE_CASELESS; break;
1214        case 'm': options |= PCRE_MULTILINE; break;        case 'm': options |= PCRE_MULTILINE; break;
# Line 810  while (!done) Line 1217  while (!done)
1217    
1218        case '+': do_showrest = 1; break;        case '+': do_showrest = 1; break;
1219        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1220          case 'B': do_debug = 1; break;
1221        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1222        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1223        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
1224        case 'F': do_flip = 1; break;        case 'F': do_flip = 1; break;
1225        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
1226        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
1227          case 'J': options |= PCRE_DUPNAMES; break;
1228        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
1229        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
1230    
# Line 826  while (!done) Line 1235  while (!done)
1235        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1236        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1237        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1238          case 'Z': debug_lengths = 0; break;
1239        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1240        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1241    
1242        case 'L':        case 'L':
1243        ppp = pp;        ppp = pp;
1244        while (*ppp != '\n' && *ppp != ' ') ppp++;        /* The '\r' test here is so that it works on Windows. */
1245          /* The '0' test is just in case this is an unterminated line. */
1246          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1247        *ppp = 0;        *ppp = 0;
1248        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1249          {          {
1250          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1251          goto SKIP_DATA;          goto SKIP_DATA;
1252          }          }
1253          locale_set = 1;
1254        tables = pcre_maketables();        tables = pcre_maketables();
1255        pp = ppp;        pp = ppp;
1256        break;        break;
# Line 849  while (!done) Line 1262  while (!done)
1262        *pp = 0;        *pp = 0;
1263        break;        break;
1264    
1265        case '\n': case ' ': break;        case '<':
1266            {
1267            if (strncmp((char *)pp, "JS>", 3) == 0)
1268              {
1269              options |= PCRE_JAVASCRIPT_COMPAT;
1270              pp += 3;
1271              }
1272            else
1273              {
1274              int x = check_newline(pp, outfile);
1275              if (x == 0) goto SKIP_DATA;
1276              options |= x;
1277              while (*pp++ != '>');
1278              }
1279            }
1280          break;
1281    
1282          case '\r':                      /* So that it works in Windows */
1283          case '\n':
1284          case ' ':
1285          break;
1286    
1287        default:        default:
1288        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);
# Line 869  while (!done) Line 1302  while (!done)
1302    
1303      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
1304      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
1305        if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1306        if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1307        if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1308    
1309      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1310    
1311      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
# Line 876  while (!done) Line 1313  while (!done)
1313    
1314      if (rc != 0)      if (rc != 0)
1315        {        {
1316        (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);        (void)regerror(rc, &preg, (char *)buffer, buffer_size);
1317        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
1318        goto SKIP_DATA;        goto SKIP_DATA;
1319        }        }
# Line 888  while (!done) Line 1325  while (!done)
1325  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1326    
1327      {      {
1328      if (timeit)      unsigned long int get_options;
1329    
1330        if (timeit > 0)
1331        {        {
1332        register int i;        register int i;
1333        clock_t time_taken;        clock_t time_taken;
1334        clock_t start_time = clock();        clock_t start_time = clock();
1335        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1336          {          {
1337          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1338          if (re != NULL) free(re);          if (re != NULL) free(re);
1339          }          }
1340        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1341        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1342          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1343            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1344        }        }
1345    
# Line 917  while (!done) Line 1356  while (!done)
1356          {          {
1357          for (;;)          for (;;)
1358            {            {
1359            if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1360              {              {
1361              done = 1;              done = 1;
1362              goto CONTINUE;              goto CONTINUE;
# Line 931  while (!done) Line 1370  while (!done)
1370        goto CONTINUE;        goto CONTINUE;
1371        }        }
1372    
1373      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1374      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
1375      returns only limited data. Check that it agrees with the newer one. */      lines. */
1376    
1377        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1378        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1379    
1380        /* Print information if required. There are now two info-returning
1381        functions. The old one has a limited interface and returns only limited
1382        data. Check that it agrees with the newer one. */
1383    
1384      if (log_store)      if (log_store)
1385        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 952  while (!done) Line 1398  while (!done)
1398    
1399      if (do_study)      if (do_study)
1400        {        {
1401        if (timeit)        if (timeit > 0)
1402          {          {
1403          register int i;          register int i;
1404          clock_t time_taken;          clock_t time_taken;
1405          clock_t start_time = clock();          clock_t start_time = clock();
1406          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1407            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options, &error);
1408          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1409          if (extra != NULL) free(extra);          if (extra != NULL) free(extra);
1410          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1411            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1412              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1413          }          }
1414        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options, &error);
# Line 980  while (!done) Line 1426  while (!done)
1426      if (do_flip)      if (do_flip)
1427        {        {
1428        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1429        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1430            byteflip(rre->magic_number, sizeof(rre->magic_number));
1431        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1432        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1433        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1434        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1435        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1436        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1437        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1438          rre->first_byte =
1439            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1440          rre->req_byte =
1441            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1442          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1443          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1444        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1445          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1446        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1447            sizeof(rre->name_count));
1448    
1449        if (extra != NULL)        if (extra != NULL)
1450          {          {
# Line 1005  while (!done) Line 1458  while (!done)
1458    
1459      SHOW_INFO:      SHOW_INFO:
1460    
1461        if (do_debug)
1462          {
1463          fprintf(outfile, "------------------------------------------------------------------\n");
1464          pcre_printint(re, outfile, debug_lengths);
1465          }
1466    
1467        /* We already have the options in get_options (see above) */
1468    
1469      if (do_showinfo)      if (do_showinfo)
1470        {        {
1471        unsigned long int get_options, all_options;        unsigned long int all_options;
1472    #if !defined NOINFOCHECK
1473        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1474        int count, backrefmax, first_char, need_char;  #endif
1475          int count, backrefmax, first_char, need_char, okpartial, jchanged,
1476            hascrorlf;
1477        int nameentrysize, namecount;        int nameentrysize, namecount;
1478        const uschar *nametable;        const uschar *nametable;
1479    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         print_internals(re, outfile);  
         }  
   
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1480        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1481        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1482        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1028  while (!done) Line 1485  while (!done)
1485        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1486        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1487        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1488          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1489          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1490          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1491    
1492    #if !defined NOINFOCHECK
1493        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
1494        if (count < 0) fprintf(outfile,        if (count < 0) fprintf(outfile,
1495          "Error %d from pcre_info()\n", count);          "Error %d from pcre_info()\n", count);
# Line 1046  while (!done) Line 1507  while (!done)
1507            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
1508              get_options, old_options);              get_options, old_options);
1509          }          }
1510    #endif
1511    
1512        if (size != regex_gotten_store) fprintf(outfile,        if (size != regex_gotten_store) fprintf(outfile,
1513          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
1514          size, regex_gotten_store);          (int)size, (int)regex_gotten_store);
1515    
1516        fprintf(outfile, "Capturing subpattern count = %d\n", count);        fprintf(outfile, "Capturing subpattern count = %d\n", count);
1517        if (backrefmax > 0)        if (backrefmax > 0)
# Line 1067  while (!done) Line 1529  while (!done)
1529            }            }
1530          }          }
1531    
1532        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1533        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1534    
1535        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1536        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");  
1537    
1538        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1539          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s\n",          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
1540            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1541            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1542            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1543            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1544              ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1545            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1546              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1547              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1548            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1549            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1550            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
1551              ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
1552            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
1553            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1554              ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1555    
1556        if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1557          fprintf(outfile, "Case state changes\n");  
1558          switch (get_options & PCRE_NEWLINE_BITS)
1559            {
1560            case PCRE_NEWLINE_CR:
1561            fprintf(outfile, "Forced newline sequence: CR\n");
1562            break;
1563    
1564            case PCRE_NEWLINE_LF:
1565            fprintf(outfile, "Forced newline sequence: LF\n");
1566            break;
1567    
1568            case PCRE_NEWLINE_CRLF:
1569            fprintf(outfile, "Forced newline sequence: CRLF\n");
1570            break;
1571    
1572            case PCRE_NEWLINE_ANYCRLF:
1573            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1574            break;
1575    
1576            case PCRE_NEWLINE_ANY:
1577            fprintf(outfile, "Forced newline sequence: ANY\n");
1578            break;
1579    
1580            default:
1581            break;
1582            }
1583    
1584        if (first_char == -1)        if (first_char == -1)
1585          {          {
1586          fprintf(outfile, "First char at start or follows \\n\n");          fprintf(outfile, "First char at start or follows newline\n");
1587          }          }
1588        else if (first_char < 0)        else if (first_char < 0)
1589          {          {
# Line 1108  while (!done) Line 1594  while (!done)
1594          int ch = first_char & 255;          int ch = first_char & 255;
1595          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
1596            "" : " (caseless)";            "" : " (caseless)";
1597          if (isprint(ch))          if (PRINTHEX(ch))
1598            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
1599          else          else
1600            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1123  while (!done) Line 1609  while (!done)
1609          int ch = need_char & 255;          int ch = need_char & 255;
1610          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
1611            "" : " (caseless)";            "" : " (caseless)";
1612          if (isprint(ch))          if (PRINTHEX(ch))
1613            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
1614          else          else
1615            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1159  while (!done) Line 1645  while (!done)
1645                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
1646                    c = 2;                    c = 2;
1647                    }                    }
1648                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
1649                    {                    {
1650                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
1651                    c += 2;                    c += 2;
# Line 1191  while (!done) Line 1677  while (!done)
1677        else        else
1678          {          {
1679          uschar sbuf[8];          uschar sbuf[8];
1680          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1681          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1682          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1683          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1684    
1685          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1686          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1687          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1688          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1689    
1690          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1691              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1218  while (!done) Line 1704  while (!done)
1704                  strerror(errno));                  strerror(errno));
1705                }                }
1706              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
1707    
1708              }              }
1709            }            }
1710          fclose(f);          fclose(f);
1711          }          }
1712    
1713          new_free(re);
1714          if (extra != NULL) new_free(extra);
1715          if (tables != NULL) new_free((void *)tables);
1716        continue;  /* With next regex */        continue;  /* With next regex */
1717        }        }
1718      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1230  while (!done) Line 1721  while (!done)
1721    
1722    for (;;)    for (;;)
1723      {      {
1724      unsigned char *q;      uschar *q;
1725      unsigned char *bptr = dbuffer;      uschar *bptr;
1726      int *use_offsets = offsets;      int *use_offsets = offsets;
1727      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1728      int callout_data = 0;      int callout_data = 0;
1729      int callout_data_set = 0;      int callout_data_set = 0;
1730      int count, c;      int count, c;
1731      int copystrings = 0;      int copystrings = 0;
1732      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
1733      int getstrings = 0;      int getstrings = 0;
1734      int getlist = 0;      int getlist = 0;
1735      int gmatched = 0;      int gmatched = 0;
1736      int start_offset = 0;      int start_offset = 0;
1737      int g_notempty = 0;      int g_notempty = 0;
1738        int use_dfa = 0;
1739    
1740      options = 0;      options = 0;
1741    
1742        *copynames = 0;
1743        *getnames = 0;
1744    
1745        copynamesptr = copynames;
1746        getnamesptr = getnames;
1747    
1748      pcre_callout = callout;      pcre_callout = callout;
1749      first_callout = 1;      first_callout = 1;
1750      callout_extra = 0;      callout_extra = 0;
# Line 1255  while (!done) Line 1753  while (!done)
1753      callout_fail_id = -1;      callout_fail_id = -1;
1754      show_malloc = 0;      show_malloc = 0;
1755    
1756      if (infile == stdin) printf("data> ");      if (extra != NULL) extra->flags &=
1757      if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
1758    
1759        len = 0;
1760        for (;;)
1761        {        {
1762        done = 1;        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
1763        goto CONTINUE;          {
1764            if (len > 0) break;
1765            done = 1;
1766            goto CONTINUE;
1767            }
1768          if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1769          len = (int)strlen((char *)buffer);
1770          if (buffer[len-1] == '\n') break;
1771        }        }
     if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);  
1772    
     len = (int)strlen((char *)buffer);  
1773      while (len > 0 && isspace(buffer[len-1])) len--;      while (len > 0 && isspace(buffer[len-1])) len--;
1774      buffer[len] = 0;      buffer[len] = 0;
1775      if (len == 0) break;      if (len == 0) break;
# Line 1271  while (!done) Line 1777  while (!done)
1777      p = buffer;      p = buffer;
1778      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1779    
1780      q = dbuffer;      bptr = q = dbuffer;
1781      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1782        {        {
1783        int i = 0;        int i = 0;
# Line 1293  while (!done) Line 1799  while (!done)
1799          c -= '0';          c -= '0';
1800          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
1801            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
1802    
1803    #if !defined NOUTF8
1804            if (use_utf8 && c > 255)
1805              {
1806              unsigned char buff8[8];
1807              int ii, utn;
1808              utn = ord2utf8(c, buff8);
1809              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1810              c = buff8[ii];   /* Last byte */
1811              }
1812    #endif
1813          break;          break;
1814    
1815          case 'x':          case 'x':
1816    
1817          /* Handle \x{..} specially - new Perl thing for utf8 */          /* Handle \x{..} specially - new Perl thing for utf8 */
1818    
1819    #if !defined NOUTF8
1820          if (*p == '{')          if (*p == '{')
1821            {            {
1822            unsigned char *pt = p;            unsigned char *pt = p;
# Line 1309  while (!done) Line 1827  while (!done)
1827              {              {
1828              unsigned char buff8[8];              unsigned char buff8[8];
1829              int ii, utn;              int ii, utn;
1830              utn = ord2utf8(c, buff8);              if (use_utf8)
1831              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
1832              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
1833                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1834                  c = buff8[ii];   /* Last byte */
1835                  }
1836                else
1837                 {
1838                 if (c > 255)
1839                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
1840                     "UTF-8 mode is not enabled.\n"
1841                     "** Truncation will probably give the wrong result.\n", c);
1842                 }
1843              p = pt + 1;              p = pt + 1;
1844              break;              break;
1845              }              }
1846            /* Not correct form; fall through */            /* Not correct form; fall through */
1847            }            }
1848    #endif
1849    
1850          /* Ordinary \x */          /* Ordinary \x */
1851    
# Line 1352  while (!done) Line 1881  while (!done)
1881            }            }
1882          else if (isalnum(*p))          else if (isalnum(*p))
1883            {            {
1884            uschar name[256];            uschar *npp = copynamesptr;
           uschar *npp = name;  
1885            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
1886              *npp++ = 0;
1887            *npp = 0;            *npp = 0;
1888            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)copynamesptr);
1889            if (n < 0)            if (n < 0)
1890              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
1891            else copystrings |= 1 << n;            copynamesptr = npp;
1892            }            }
1893          else if (*p == '+')          else if (*p == '+')
1894            {            {
# Line 1397  while (!done) Line 1926  while (!done)
1926            }            }
1927          continue;          continue;
1928    
1929    #if !defined NODFA
1930            case 'D':
1931    #if !defined NOPOSIX
1932            if (posix || do_posix)
1933              printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");
1934            else
1935    #endif
1936              use_dfa = 1;
1937            continue;
1938    
1939            case 'F':
1940            options |= PCRE_DFA_SHORTEST;
1941            continue;
1942    #endif
1943    
1944          case 'G':          case 'G':
1945          if (isdigit(*p))          if (isdigit(*p))
1946            {            {
# Line 1405  while (!done) Line 1949  while (!done)
1949            }            }
1950          else if (isalnum(*p))          else if (isalnum(*p))
1951            {            {
1952            uschar name[256];            uschar *npp = getnamesptr;
           uschar *npp = name;  
1953            while (isalnum(*p)) *npp++ = *p++;            while (isalnum(*p)) *npp++ = *p++;
1954              *npp++ = 0;
1955            *npp = 0;            *npp = 0;
1956            n = pcre_get_stringnumber(re, (char *)name);            n = pcre_get_stringnumber(re, (char *)getnamesptr);
1957            if (n < 0)            if (n < 0)
1958              fprintf(outfile, "no parentheses with name \"%s\"\n", name);              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
1959            else getstrings |= 1 << n;            getnamesptr = npp;
1960            }            }
1961          continue;          continue;
1962    
# Line 1438  while (!done) Line 1982  while (!done)
1982            if (offsets == NULL)            if (offsets == NULL)
1983              {              {
1984              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1985                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1986              return 1;              yield = 1;
1987                goto EXIT;
1988              }              }
1989            }            }
1990          use_size_offsets = n;          use_size_offsets = n;
# Line 1447  while (!done) Line 1992  while (!done)
1992          continue;          continue;
1993    
1994          case 'P':          case 'P':
1995          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
1996              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
1997          continue;          continue;
1998    
1999            case 'Q':
2000            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2001            if (extra == NULL)
2002              {
2003              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2004              extra->flags = 0;
2005              }
2006            extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
2007            extra->match_limit_recursion = n;
2008            continue;
2009    
2010            case 'q':
2011            while(isdigit(*p)) n = n * 10 + *p++ - '0';
2012            if (extra == NULL)
2013              {
2014              extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2015              extra->flags = 0;
2016              }
2017            extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
2018            extra->match_limit = n;
2019            continue;
2020    
2021    #if !defined NODFA
2022            case 'R':
2023            options |= PCRE_DFA_RESTART;
2024            continue;
2025    #endif
2026    
2027          case 'S':          case 'S':
2028          show_malloc = 1;          show_malloc = 1;
2029          continue;          continue;
2030    
2031            case 'Y':
2032            options |= PCRE_NO_START_OPTIMIZE;
2033            continue;
2034    
2035          case 'Z':          case 'Z':
2036          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2037          continue;          continue;
# Line 1461  while (!done) Line 2039  while (!done)
2039          case '?':          case '?':
2040          options |= PCRE_NO_UTF8_CHECK;          options |= PCRE_NO_UTF8_CHECK;
2041          continue;          continue;
2042    
2043            case '<':
2044              {
2045              int x = check_newline(p, outfile);
2046              if (x == 0) goto NEXT_DATA;
2047              options |= x;
2048              while (*p++ != '>');
2049              }
2050            continue;
2051          }          }
2052        *q++ = c;        *q++ = c;
2053        }        }
2054      *q = 0;      *q = 0;
2055      len = q - dbuffer;      len = q - dbuffer;
2056    
2057        /* Move the data to the end of the buffer so that a read over the end of
2058        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2059        we are using the POSIX interface, we must include the terminating zero. */
2060    
2061    #if !defined NOPOSIX
2062        if (posix || do_posix)
2063          {
2064          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2065          bptr += buffer_size - len - 1;
2066          }
2067        else
2068    #endif
2069          {
2070          memmove(bptr + buffer_size - len, bptr, len);
2071          bptr += buffer_size - len;
2072          }
2073    
2074        if ((all_use_dfa || use_dfa) && find_match_limit)
2075          {
2076          printf("**Match limit not relevant for DFA matching: ignored\n");
2077          find_match_limit = 0;
2078          }
2079    
2080      /* Handle matching via the POSIX interface, which does not      /* Handle matching via the POSIX interface, which does not
2081      support timing or playing with the match limit or callout data. */      support timing or playing with the match limit or callout data. */
2082    
# Line 1480  while (!done) Line 2090  while (!done)
2090          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2091        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2092        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2093          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2094    
2095        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2096    
2097        if (rc != 0)        if (rc != 0)
2098          {          {
2099          (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);          (void)regerror(rc, &preg, (char *)buffer, buffer_size);
2100          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
2101          }          }
2102          else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
2103                  != 0)
2104            {
2105            fprintf(outfile, "Matched with REG_NOSUB\n");
2106            }
2107        else        else
2108          {          {
2109          size_t i;          size_t i;
# Line 1519  while (!done) Line 2135  while (!done)
2135    
2136      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2137        {        {
2138        if (timeit)        if (timeitm > 0)
2139          {          {
2140          register int i;          register int i;
2141          clock_t time_taken;          clock_t time_taken;
2142          clock_t start_time = clock();          clock_t start_time = clock();
2143          for (i = 0; i < LOOPREPEAT; i++)  
2144    #if !defined NODFA
2145            if (all_use_dfa || use_dfa)
2146              {
2147              int workspace[1000];
2148              for (i = 0; i < timeitm; i++)
2149                count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,
2150                  options | g_notempty, use_offsets, use_size_offsets, workspace,
2151                  sizeof(workspace)/sizeof(int));
2152              }
2153            else
2154    #endif
2155    
2156            for (i = 0; i < timeitm; i++)
2157            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2158              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2159    
2160          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2161          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2162            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2163              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2164          }          }
2165    
2166        /* 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
2167        varying limits in order to find the minimum value. */        varying limits in order to find the minimum value for the match limit and
2168          for the recursion limit. */
2169    
2170        if (find_match_limit)        if (find_match_limit)
2171          {          {
         int min = 0;  
         int mid = 64;  
         int max = -1;  
   
2172          if (extra == NULL)          if (extra == NULL)
2173            {            {
2174            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2175            extra->flags = 0;            extra->flags = 0;
2176            }            }
         extra->flags |= PCRE_EXTRA_MATCH_LIMIT;  
2177    
2178          for (;;)          (void)check_match_limit(re, extra, bptr, len, start_offset,
2179            {            options|g_notempty, use_offsets, use_size_offsets,
2180            extra->match_limit = mid;            PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
2181            count = pcre_exec(re, extra, (char *)bptr, len, start_offset,            PCRE_ERROR_MATCHLIMIT, "match()");
2182              options | g_notempty, use_offsets, use_size_offsets);  
2183            if (count == PCRE_ERROR_MATCHLIMIT)          count = check_match_limit(re, extra, bptr, len, start_offset,
2184              {            options|g_notempty, use_offsets, use_size_offsets,
2185              /* fprintf(outfile, "Testing match limit = %d\n", mid); */            PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
2186              min = mid;            PCRE_ERROR_RECURSIONLIMIT, "match() recursion");
             mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;  
             }  
           else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||  
                                  count == PCRE_ERROR_PARTIAL)  
             {  
             if (mid == min + 1)  
               {  
               fprintf(outfile, "Minimum match limit = %d\n", mid);  
               break;  
               }  
             /* fprintf(outfile, "Testing match limit = %d\n", mid); */  
             max = mid;  
             mid = (min + mid)/2;  
             }  
           else break;    /* Some other error */  
           }  
   
         extra->flags &= ~PCRE_EXTRA_MATCH_LIMIT;  
2187          }          }
2188    
2189        /* If callout_data is set, use the interface with additional data */        /* If callout_data is set, use the interface with additional data */
# Line 1597  while (!done) Line 2205  while (!done)
2205        /* 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
2206        value of match_limit. */        value of match_limit. */
2207    
2208        else  #if !defined NODFA
2209          else if (all_use_dfa || use_dfa)
2210          {          {
2211          count = pcre_exec(re, extra, (char *)bptr, len,          int workspace[1000];
2212            start_offset, options | g_notempty, use_offsets, use_size_offsets);          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,
2213              options | g_notempty, use_offsets, use_size_offsets, workspace,
2214              sizeof(workspace)/sizeof(int));
2215            if (count == 0)
2216              {
2217              fprintf(outfile, "Matched, but too many subsidiary matches\n");
2218              count = use_size_offsets/2;
2219              }
2220          }          }
2221    #endif
2222    
2223        if (count == 0)        else
2224          {          {
2225          fprintf(outfile, "Matched, but too many substrings\n");          count = pcre_exec(re, extra, (char *)bptr, len,
2226          count = use_size_offsets/3;            start_offset, options | g_notempty, use_offsets, use_size_offsets);
2227            if (count == 0)
2228              {
2229              fprintf(outfile, "Matched, but too many substrings\n");
2230              count = use_size_offsets/3;
2231              }
2232          }          }
2233    
2234        /* Matched */        /* Matched */
2235    
2236        if (count >= 0)        if (count >= 0)
2237          {          {
2238          int i;          int i, maxcount;
2239    
2240    #if !defined NODFA
2241            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2242    #endif
2243              maxcount = use_size_offsets/3;
2244    
2245            /* This is a check against a lunatic return value. */
2246    
2247            if (count > maxcount)
2248              {
2249              fprintf(outfile,
2250                "** PCRE error: returned count %d is too big for offset size %d\n",
2251                count, use_size_offsets);
2252              count = use_size_offsets/3;
2253              if (do_g || do_G)
2254                {
2255                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2256                do_g = do_G = FALSE;        /* Break g/G loop */
2257                }
2258              }
2259    
2260          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2261            {            {
2262            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
# Line 1641  while (!done) Line 2284  while (!done)
2284            {            {
2285            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
2286              {              {
2287              char copybuffer[16];              char copybuffer[256];
2288              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,
2289                i, copybuffer, sizeof(copybuffer));                i, copybuffer, sizeof(copybuffer));
2290              if (rc < 0)              if (rc < 0)
# Line 1651  while (!done) Line 2294  while (!done)
2294              }              }
2295            }            }
2296    
2297            for (copynamesptr = copynames;
2298                 *copynamesptr != 0;
2299                 copynamesptr += (int)strlen((char*)copynamesptr) + 1)
2300              {
2301              char copybuffer[256];
2302              int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,
2303                count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));
2304              if (rc < 0)
2305                fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);
2306              else
2307                fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);
2308              }
2309    
2310          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
2311            {            {
2312            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
# Line 1663  while (!done) Line 2319  while (!done)
2319              else              else
2320                {                {
2321                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
               /* free((void *)substring); */  
2322                pcre_free_substring(substring);                pcre_free_substring(substring);
2323                }                }
2324              }              }
2325            }            }
2326    
2327            for (getnamesptr = getnames;
2328                 *getnamesptr != 0;
2329                 getnamesptr += (int)strlen((char*)getnamesptr) + 1)
2330              {
2331              const char *substring;
2332              int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,
2333                count, (char *)getnamesptr, &substring);
2334              if (rc < 0)
2335                fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);
2336              else
2337                {
2338                fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);
2339                pcre_free_substring(substring);
2340                }
2341              }
2342    
2343          if (getlist)          if (getlist)
2344            {            {
2345            const char **stringlist;            const char **stringlist;
# Line 1692  while (!done) Line 2363  while (!done)
2363    
2364        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2365          {          {
2366          fprintf(outfile, "Partial match\n");          fprintf(outfile, "Partial match");
2367            if (use_size_offsets > 1)
2368              {
2369              fprintf(outfile, ": ");
2370              pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2371                outfile);
2372              }
2373            fprintf(outfile, "\n");
2374          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2375          }          }
2376    
2377        /* 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
2378        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
2379        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
2380        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2381        offset values to achieve this. We won't be at the end of the string -  
2382        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2383          "anycrlf". If the previous match was at the end of a line terminated by
2384          CRLF, an advance of one character just passes the \r, whereas we should
2385          prefer the longer newline sequence, as does the code in pcre_exec().
2386          Fudge the offset value to achieve this.
2387    
2388          Otherwise, in the case of UTF-8 matching, the advance must be one
2389          character, not one byte. */
2390    
2391        else        else
2392          {          {
2393          if (g_notempty != 0)          if (g_notempty != 0)
2394            {            {
2395            int onechar = 1;            int onechar = 1;
2396              unsigned int obits = ((real_pcre *)re)->options;
2397            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2398            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2399                {
2400                int d;
2401                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2402                /* Note that these values are always the ASCII ones, even in
2403                EBCDIC environments. CR = 13, NL = 10. */
2404                obits = (d == 13)? PCRE_NEWLINE_CR :
2405                        (d == 10)? PCRE_NEWLINE_LF :
2406                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2407                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2408                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2409                }
2410              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2411                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2412                  &&
2413                  start_offset < len - 1 &&
2414                  bptr[start_offset] == '\r' &&
2415                  bptr[start_offset+1] == '\n')
2416                onechar++;
2417              else if (use_utf8)
2418              {              {
2419              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2420                {                {
# Line 1744  while (!done) Line 2449  while (!done)
2449        character. */        character. */
2450    
2451        g_notempty = 0;        g_notempty = 0;
2452    
2453        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2454          {          {
2455          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
# Line 1762  while (!done) Line 2468  while (!done)
2468          len -= use_offsets[1];          len -= use_offsets[1];
2469          }          }
2470        }  /* End of loop for /g and /G */        }  /* End of loop for /g and /G */
2471    
2472        NEXT_DATA: continue;
2473      }    /* End of loop for data lines */      }    /* End of loop for data lines */
2474    
2475    CONTINUE:    CONTINUE:
# Line 1770  while (!done) Line 2478  while (!done)
2478    if (posix || do_posix) regfree(&preg);    if (posix || do_posix) regfree(&preg);
2479  #endif  #endif
2480    
2481    if (re != NULL) free(re);    if (re != NULL) new_free(re);
2482    if (extra != NULL) free(extra);    if (extra != NULL) new_free(extra);
2483    if (tables != NULL)    if (tables != NULL)
2484      {      {
2485      free((void *)tables);      new_free((void *)tables);
2486      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
2487        locale_set = 0;
2488      }      }
2489    }    }
2490    
2491  if (infile == stdin) fprintf(outfile, "\n");  if (infile == stdin) fprintf(outfile, "\n");
2492  return 0;  
2493    EXIT:
2494    
2495    if (infile != NULL && infile != stdin) fclose(infile);
2496    if (outfile != NULL && outfile != stdout) fclose(outfile);
2497    
2498    free(buffer);
2499    free(dbuffer);
2500    free(pbuffer);
2501    free(offsets);
2502    
2503    return yield;
2504  }  }
2505    
2506  /* End */  /* End of pcretest.c */

Legend:
Removed from v.75  
changed lines
  Added in v.427

  ViewVC Help
Powered by ViewVC 1.1.5