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

Diff of /code/trunk/pcretest.c

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

revision 91 by nigel, Sat Feb 24 21:41:34 2007 UTC revision 455 by ph10, Sat Sep 26 19:12:32 2009 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 44  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51  #ifndef _WIN32  #ifdef SUPPORT_LIBREADLINE
52  #include <sys/resource.h>  #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  #endif
88    
 #define PCRE_SPY        /* For Win32 build, import data, not export */  
89    
90  /* We include pcre_internal.h because we need the internal info for displaying  /* We have to include pcre_internal.h because we need the internal info for
91  the results of pcre_study() and we also need to know about the internal  displaying the results of pcre_study() and we also need to know about the
92  macros, structures, and other internal data values; pcretest has "inside  internal macros, structures, and other internal data values; pcretest has
93  information" compared to a program that strictly follows the PCRE API. */  "inside information" compared to a program that strictly follows the PCRE API.
94    
95    Although pcre_internal.h does itself include pcre.h, we explicitly include it
96    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
97    appropriately for an application, not for building PCRE. */
98    
99    #include "pcre.h"
100  #include "pcre_internal.h"  #include "pcre_internal.h"
101    
102  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* We need access to some of the data tables that PCRE uses. So as not to have
103  two copies, we include the source file here, changing the names of the external  to keep two copies, we include the source file here, changing the names of the
104  symbols to prevent clashes. */  external symbols to prevent clashes. */
105    
106    #define _pcre_ucp_gentype      ucp_gentype
107  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
108  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
109  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 68  symbols to prevent clashes. */ Line 111  symbols to prevent clashes. */
111  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
112  #define _pcre_utt              utt  #define _pcre_utt              utt
113  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
114    #define _pcre_utt_names        utt_names
115  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
116    
117  #include "pcre_tables.c"  #include "pcre_tables.c"
118    
119  /* We also need the pcre_printint() function for printing out compiled  /* 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  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. */  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"  #include "pcre_printint.src"
130    
131    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
132    
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 87  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 imported into Exim, to  /* It is also possible, for the benefit of the version currently imported into
143  build pcretest without support for UTF8 (define NOUTF8), without the interface  Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
144  to the DFA matcher (NODFA), and without the doublecheck of the old "info"  interface to the DFA matcher (NODFA), and without the doublecheck of the old
145  function (define NOINFOCHECK). */  "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 */  /* Other parameters */
# Line 103  function (define NOINFOCHECK). */ Line 162  function (define NOINFOCHECK). */
162  #endif  #endif
163  #endif  #endif
164    
165    /* This is the default loop count for timing. */
166    
167  #define LOOPREPEAT 500000  #define LOOPREPEAT 500000
168    
169  /* Static variables */  /* Static variables */
# Line 113  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;
# Line 143  optimal way of handling this, but hey, t Line 206  optimal way of handling this, but hey, t
206  Arguments:  Arguments:
207    f            the file to read    f            the file to read
208    start        where in buffer to start (this *must* be within buffer)    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  Returns:       pointer to the start of new data
212                 could be a copy of start, or could be moved                 could be a copy of start, or could be moved
# Line 150  Returns: pointer to the start of n Line 214  Returns: pointer to the start of n
214  */  */
215    
216  static uschar *  static uschar *
217  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
218  {  {
219  uschar *here = start;  uschar *here = start;
220    
221  for (;;)  for (;;)
222    {    {
223    int rlen = buffer_size - (here - buffer);    int rlen = buffer_size - (here - buffer);
224    
225    if (rlen > 1000)    if (rlen > 1000)
226      {      {
227      int dlen;      int dlen;
228      if (fgets((char *)here, rlen,  f) == NULL)  
229        return (here == start)? NULL : start;      /* 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);      dlen = (int)strlen((char *)here);
259      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
260      here += dlen;      here += dlen;
# Line 213  return NULL; /* Control never gets here Line 306  return NULL; /* Control never gets here
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 311  Arguments: Line 404  Arguments:
404  Returns:     number of characters placed in the buffer  Returns:     number of characters placed in the buffer
405  */  */
406    
407    #if !defined NOUTF8
408    
409  static int  static int
410  ord2utf8(int cvalue, uschar *utf8bytes)  ord2utf8(int cvalue, uschar *utf8bytes)
411  {  {
# Line 327  for (j = i; j > 0; j--) Line 422  for (j = i; j > 0; j--)
422  return i + 1;  return i + 1;
423  }  }
424    
425    #endif
426    
427    
428    
429  /*************************************************  /*************************************************
# Line 353  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        }        }
# Line 371  while (length-- > 0) Line 471  while (length-- > 0)
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 610  return count; Line 711  return count;
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                *  *         Check newline indicator                *
741  *************************************************/  *************************************************/
742    
743  /* This is used both at compile and run-time to check for <xxx> escapes, where  /* This is used both at compile and run-time to check for <xxx> escapes, where
744  xxx is LF, CR, or CRLF. Print a message and return 0 if there is no match.  xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is
745    no match.
746    
747  Arguments:  Arguments:
748    p           points after the leading '<'    p           points after the leading '<'
# Line 626  Returns: appropriate PCRE_NEWLINE_x Line 754  Returns: appropriate PCRE_NEWLINE_x
754  static int  static int
755  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
756  {  {
757  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
758  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
759  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  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);  fprintf(f, "Unknown newline type at: <%s\n", p);
765  return 0;  return 0;
766  }  }
# Line 636  return 0; Line 768  return 0;
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 648  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;  int quiet = 0;
# Line 681  buffer = (unsigned char *)malloc(buffer_ Line 855  buffer = (unsigned char *)malloc(buffer_
855  dbuffer = (unsigned char *)malloc(buffer_size);  dbuffer = (unsigned char *)malloc(buffer_size);
856  pbuffer = (unsigned char *)malloc(buffer_size);  pbuffer = (unsigned char *)malloc(buffer_size);
857    
858  /* The outfile variable is static so that new_malloc can use it. The _setmode()  /* The outfile variable is static so that new_malloc can use it. */
 stuff is some magic that I don't understand, but which apparently does good  
 things in Windows. It's related to line terminations.  */  
   
 #if defined(_WIN32) || defined(WIN32)  
 _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 699  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;
   else if (strcmp(argv[op], "-t") == 0) timeit = 1;  
879    else if (strcmp(argv[op], "-q") == 0) quiet = 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  #if !defined NODFA
885    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
886  #endif  #endif
# Line 713  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 &&    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
909        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
910          *endptr == 0))          *endptr == 0))
911      {      {
912  #ifdef _WIN32  #if defined(_WIN32) || defined(WIN32)
913      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
914      exit(1);      exit(1);
915  #else  #else
# Line 741  while (argc > 1 && argv[op][0] == '-') Line 933  while (argc > 1 && argv[op][0] == '-')
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 748  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 sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
945        (rc == '\n')? "LF" : "CRLF");      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, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
960      printf("  Default recursion depth limit = %d\n", rc);      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 [options] [<input> [<output>]]\n");      usage();
     printf("  -C     show PCRE compile-time options and exit\n");  
     printf("  -d     debug: show compiled code; implies -i\n");  
 #if !defined NODFA  
     printf("  -dfa   force DFA matching for all subjects\n");  
 #endif  
     printf("  -i     show information about compiled pattern\n"  
            "  -m     output memory used information\n"  
            "  -o <n> set size of offsets vector to <n>\n");  
 #if !defined NOPOSIX  
     printf("  -p     use POSIX interface\n");  
 #endif  
     printf("  -S <n> set stack size to <n> megabytes\n");  
     printf("  -s     output store (memory) used information\n"  
            "  -t     time compilation and execution\n");  
975      yield = 1;      yield = 1;
976      goto EXIT;      goto EXIT;
977      }      }
# Line 794  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    yield = 1;    yield = 1;
991    goto EXIT;    goto EXIT;
992    }    }
# Line 803  if (offsets == NULL) Line 995  if (offsets == NULL)
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]);
# Line 814  if (argc > 1) Line 1006  if (argc > 1)
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]);
# Line 859  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 (extend_inputline(infile, buffer) == 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 964  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 979  while (!done) Line 1172  while (!done)
1172        pp++;        pp++;
1173        }        }
1174      if (*pp != 0) break;      if (*pp != 0) break;
1175      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1176        {        {
1177        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1178        done = 1;        done = 1;
# Line 989  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 1020  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;
# Line 1037  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        /* The '\r' test here is so that it works on Windows */        /* The '\r' test here is so that it works on Windows. */
1245        while (*ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;        /* 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 1063  while (!done) Line 1264  while (!done)
1264    
1265        case '<':        case '<':
1266          {          {
1267          int x = check_newline(pp, outfile);          if (strncmp((char *)pp, "JS>", 3) == 0)
1268          if (x == 0) goto SKIP_DATA;            {
1269          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1270          while (*pp++ != '>');            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;        break;
1281    
# Line 1096  while (!done) Line 1305  while (!done)
1305      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1306      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1307      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1308        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1309    
1310      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1311    
# Line 1116  while (!done) Line 1326  while (!done)
1326  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1327    
1328      {      {
1329      if (timeit)      unsigned long int get_options;
1330    
1331        if (timeit > 0)
1332        {        {
1333        register int i;        register int i;
1334        clock_t time_taken;        clock_t time_taken;
1335        clock_t start_time = clock();        clock_t start_time = clock();
1336        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1337          {          {
1338          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1339          if (re != NULL) free(re);          if (re != NULL) free(re);
1340          }          }
1341        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1342        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1343          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1344            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1345        }        }
1346    
# Line 1145  while (!done) Line 1357  while (!done)
1357          {          {
1358          for (;;)          for (;;)
1359            {            {
1360            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1361              {              {
1362              done = 1;              done = 1;
1363              goto CONTINUE;              goto CONTINUE;
# Line 1159  while (!done) Line 1371  while (!done)
1371        goto CONTINUE;        goto CONTINUE;
1372        }        }
1373    
1374      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1375      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
1376      returns only limited data. Check that it agrees with the newer one. */      lines. */
1377    
1378        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1379        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1380    
1381        /* Print information if required. There are now two info-returning
1382        functions. The old one has a limited interface and returns only limited
1383        data. Check that it agrees with the newer one. */
1384    
1385      if (log_store)      if (log_store)
1386        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1180  while (!done) Line 1399  while (!done)
1399    
1400      if (do_study)      if (do_study)
1401        {        {
1402        if (timeit)        if (timeit > 0)
1403          {          {
1404          register int i;          register int i;
1405          clock_t time_taken;          clock_t time_taken;
1406          clock_t start_time = clock();          clock_t start_time = clock();
1407          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1408            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options, &error);
1409          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1410          if (extra != NULL) free(extra);          if (extra != NULL) free(extra);
1411          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1412            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1413              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1414          }          }
1415        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options, &error);
# Line 1208  while (!done) Line 1427  while (!done)
1427      if (do_flip)      if (do_flip)
1428        {        {
1429        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1430        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1431            byteflip(rre->magic_number, sizeof(rre->magic_number));
1432        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1433        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1434        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1435        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1436        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1437        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1438        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1439          rre->first_byte =
1440            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1441          rre->req_byte =
1442            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1443          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1444          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1445        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1446          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1447        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1448            sizeof(rre->name_count));
1449    
1450        if (extra != NULL)        if (extra != NULL)
1451          {          {
1452          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1453          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          rsd->size = byteflip(rsd->size, sizeof(rsd->size));
1454          rsd->options = byteflip(rsd->options, sizeof(rsd->options));          rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
1455            rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
1456          }          }
1457        }        }
1458    
# Line 1233  while (!done) Line 1460  while (!done)
1460    
1461      SHOW_INFO:      SHOW_INFO:
1462    
1463        if (do_debug)
1464          {
1465          fprintf(outfile, "------------------------------------------------------------------\n");
1466          pcre_printint(re, outfile, debug_lengths);
1467          }
1468    
1469        /* We already have the options in get_options (see above) */
1470    
1471      if (do_showinfo)      if (do_showinfo)
1472        {        {
1473        unsigned long int get_options, all_options;        unsigned long int all_options;
1474  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1475        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1476  #endif  #endif
1477        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1478            hascrorlf;
1479        int nameentrysize, namecount;        int nameentrysize, namecount;
1480        const uschar *nametable;        const uschar *nametable;
1481    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         pcre_printint(re, outfile);  
         }  
   
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1482        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1483        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1484        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1258  while (!done) Line 1487  while (!done)
1487        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1488        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1489        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1490          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1491          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1492          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1493    
1494  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1495        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1299  while (!done) Line 1531  while (!done)
1531            }            }
1532          }          }
1533    
1534        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1535        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1536    
1537        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1538        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");  
1539    
1540        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1541          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s\n",          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
1542            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1543            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1544            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1545            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1546            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1547            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1548              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1549              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1550            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1551            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1552            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
# Line 1327  while (!done) Line 1555  while (!done)
1555            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1556            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1557    
1558        switch (get_options & PCRE_NEWLINE_CRLF)        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1559    
1560          switch (get_options & PCRE_NEWLINE_BITS)
1561          {          {
1562          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
1563          fprintf(outfile, "Forced newline sequence: CR\n");          fprintf(outfile, "Forced newline sequence: CR\n");
# Line 1341  while (!done) Line 1571  while (!done)
1571          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
1572          break;          break;
1573    
1574            case PCRE_NEWLINE_ANYCRLF:
1575            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1576            break;
1577    
1578            case PCRE_NEWLINE_ANY:
1579            fprintf(outfile, "Forced newline sequence: ANY\n");
1580            break;
1581    
1582          default:          default:
1583          break;          break;
1584          }          }
# Line 1358  while (!done) Line 1596  while (!done)
1596          int ch = first_char & 255;          int ch = first_char & 255;
1597          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
1598            "" : " (caseless)";            "" : " (caseless)";
1599          if (isprint(ch))          if (PRINTHEX(ch))
1600            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
1601          else          else
1602            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1373  while (!done) Line 1611  while (!done)
1611          int ch = need_char & 255;          int ch = need_char & 255;
1612          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
1613            "" : " (caseless)";            "" : " (caseless)";
1614          if (isprint(ch))          if (PRINTHEX(ch))
1615            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
1616          else          else
1617            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1391  while (!done) Line 1629  while (!done)
1629          else          else
1630            {            {
1631            uschar *start_bits = NULL;            uschar *start_bits = NULL;
1632              int minlength;
1633    
1634              new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
1635              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
1636    
1637            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
   
1638            if (start_bits == NULL)            if (start_bits == NULL)
1639              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "No set of starting bytes\n");
1640            else            else
1641              {              {
1642              int i;              int i;
# Line 1409  while (!done) Line 1651  while (!done)
1651                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
1652                    c = 2;                    c = 2;
1653                    }                    }
1654                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
1655                    {                    {
1656                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
1657                    c += 2;                    c += 2;
# Line 1441  while (!done) Line 1683  while (!done)
1683        else        else
1684          {          {
1685          uschar sbuf[8];          uschar sbuf[8];
1686          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1687          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1688          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1689          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1690    
1691          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1692          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1693          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1694          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1695    
1696          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1697              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1468  while (!done) Line 1710  while (!done)
1710                  strerror(errno));                  strerror(errno));
1711                }                }
1712              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
1713    
1714              }              }
1715            }            }
1716          fclose(f);          fclose(f);
# Line 1485  while (!done) Line 1728  while (!done)
1728    for (;;)    for (;;)
1729      {      {
1730      uschar *q;      uschar *q;
1731      uschar *bptr = dbuffer;      uschar *bptr;
1732      int *use_offsets = offsets;      int *use_offsets = offsets;
1733      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1734      int callout_data = 0;      int callout_data = 0;
1735      int callout_data_set = 0;      int callout_data_set = 0;
1736      int count, c;      int count, c;
1737      int copystrings = 0;      int copystrings = 0;
1738      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
1739      int getstrings = 0;      int getstrings = 0;
1740      int getlist = 0;      int getlist = 0;
1741      int gmatched = 0;      int gmatched = 0;
# Line 1522  while (!done) Line 1765  while (!done)
1765      len = 0;      len = 0;
1766      for (;;)      for (;;)
1767        {        {
1768        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
1769          {          {
1770          if (len > 0) break;          if (len > 0) break;
1771          done = 1;          done = 1;
# Line 1541  while (!done) Line 1783  while (!done)
1783      p = buffer;      p = buffer;
1784      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1785    
1786      q = dbuffer;      bptr = q = dbuffer;
1787      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1788        {        {
1789        int i = 0;        int i = 0;
# Line 1591  while (!done) Line 1833  while (!done)
1833              {              {
1834              unsigned char buff8[8];              unsigned char buff8[8];
1835              int ii, utn;              int ii, utn;
1836              utn = ord2utf8(c, buff8);              if (use_utf8)
1837              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
1838              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
1839                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1840                  c = buff8[ii];   /* Last byte */
1841                  }
1842                else
1843                 {
1844                 if (c > 255)
1845                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
1846                     "UTF-8 mode is not enabled.\n"
1847                     "** Truncation will probably give the wrong result.\n", c);
1848                 }
1849              p = pt + 1;              p = pt + 1;
1850              break;              break;
1851              }              }
# Line 1723  while (!done) Line 1975  while (!done)
1975          continue;          continue;
1976    
1977          case 'N':          case 'N':
1978          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
1979              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
1980            else
1981              options |= PCRE_NOTEMPTY;
1982          continue;          continue;
1983    
1984          case 'O':          case 'O':
# Line 1736  while (!done) Line 1991  while (!done)
1991            if (offsets == NULL)            if (offsets == NULL)
1992              {              {
1993              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1994                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1995              yield = 1;              yield = 1;
1996              goto EXIT;              goto EXIT;
1997              }              }
# Line 1746  while (!done) Line 2001  while (!done)
2001          continue;          continue;
2002    
2003          case 'P':          case 'P':
2004          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2005              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2006          continue;          continue;
2007    
2008          case 'Q':          case 'Q':
# Line 1781  while (!done) Line 2037  while (!done)
2037          show_malloc = 1;          show_malloc = 1;
2038          continue;          continue;
2039    
2040            case 'Y':
2041            options |= PCRE_NO_START_OPTIMIZE;
2042            continue;
2043    
2044          case 'Z':          case 'Z':
2045          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2046          continue;          continue;
# Line 1803  while (!done) Line 2063  while (!done)
2063      *q = 0;      *q = 0;
2064      len = q - dbuffer;      len = q - dbuffer;
2065    
2066        /* Move the data to the end of the buffer so that a read over the end of
2067        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2068        we are using the POSIX interface, we must include the terminating zero. */
2069    
2070    #if !defined NOPOSIX
2071        if (posix || do_posix)
2072          {
2073          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2074          bptr += buffer_size - len - 1;
2075          }
2076        else
2077    #endif
2078          {
2079          memmove(bptr + buffer_size - len, bptr, len);
2080          bptr += buffer_size - len;
2081          }
2082    
2083      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2084        {        {
2085        printf("**Match limit not relevant for DFA matching: ignored\n");        printf("**Match limit not relevant for DFA matching: ignored\n");
# Line 1822  while (!done) Line 2099  while (!done)
2099          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2100        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2101        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2102          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2103    
2104        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2105    
# Line 1866  while (!done) Line 2144  while (!done)
2144    
2145      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2146        {        {
2147        if (timeit)        if (timeitm > 0)
2148          {          {
2149          register int i;          register int i;
2150          clock_t time_taken;          clock_t time_taken;
# Line 1876  while (!done) Line 2154  while (!done)
2154          if (all_use_dfa || use_dfa)          if (all_use_dfa || use_dfa)
2155            {            {
2156            int workspace[1000];            int workspace[1000];
2157            for (i = 0; i < LOOPREPEAT; i++)            for (i = 0; i < timeitm; i++)
2158              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2159                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2160                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
2161            }            }
2162          else          else
2163  #endif  #endif
2164    
2165          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeitm; i++)
2166            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2167              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2168    
2169          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2170          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2171            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2172              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2173          }          }
2174    
# Line 1940  while (!done) Line 2218  while (!done)
2218        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
2219          {          {
2220          int workspace[1000];          int workspace[1000];
2221          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
2222            options | g_notempty, use_offsets, use_size_offsets, workspace,            options | g_notempty, use_offsets, use_size_offsets, workspace,
2223            sizeof(workspace)/sizeof(int));            sizeof(workspace)/sizeof(int));
2224          if (count == 0)          if (count == 0)
# Line 1966  while (!done) Line 2244  while (!done)
2244    
2245        if (count >= 0)        if (count >= 0)
2246          {          {
2247          int i;          int i, maxcount;
2248    
2249    #if !defined NODFA
2250            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2251    #endif
2252              maxcount = use_size_offsets/3;
2253    
2254            /* This is a check against a lunatic return value. */
2255    
2256            if (count > maxcount)
2257              {
2258              fprintf(outfile,
2259                "** PCRE error: returned count %d is too big for offset size %d\n",
2260                count, use_size_offsets);
2261              count = use_size_offsets/3;
2262              if (do_g || do_G)
2263                {
2264                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2265                do_g = do_G = FALSE;        /* Break g/G loop */
2266                }
2267              }
2268    
2269          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2270            {            {
2271            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
# Line 2074  while (!done) Line 2373  while (!done)
2373        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2374          {          {
2375          fprintf(outfile, "Partial match");          fprintf(outfile, "Partial match");
2376  #if !defined NODFA          if (use_size_offsets > 1)
2377          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)            {
2378            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            fprintf(outfile, ": ");
2379              bptr + use_offsets[0]);            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2380  #endif              outfile);
2381              }
2382          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2383          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2384          }          }
2385    
2386        /* 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
2387        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
2388        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
2389        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2390        offset values to achieve this. We won't be at the end of the string -  
2391        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2392          "anycrlf". If the previous match was at the end of a line terminated by
2393          CRLF, an advance of one character just passes the \r, whereas we should
2394          prefer the longer newline sequence, as does the code in pcre_exec().
2395          Fudge the offset value to achieve this.
2396    
2397          Otherwise, in the case of UTF-8 matching, the advance must be one
2398          character, not one byte. */
2399    
2400        else        else
2401          {          {
2402          if (g_notempty != 0)          if (g_notempty != 0)
2403            {            {
2404            int onechar = 1;            int onechar = 1;
2405              unsigned int obits = ((real_pcre *)re)->options;
2406            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2407            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2408                {
2409                int d;
2410                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2411                /* Note that these values are always the ASCII ones, even in
2412                EBCDIC environments. CR = 13, NL = 10. */
2413                obits = (d == 13)? PCRE_NEWLINE_CR :
2414                        (d == 10)? PCRE_NEWLINE_LF :
2415                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2416                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2417                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2418                }
2419              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2420                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2421                  &&
2422                  start_offset < len - 1 &&
2423                  bptr[start_offset] == '\r' &&
2424                  bptr[start_offset+1] == '\n')
2425                onechar++;
2426              else if (use_utf8)
2427              {              {
2428              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2429                {                {
# Line 2124  while (!done) Line 2451  while (!done)
2451        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2452    
2453        /* If we have matched an empty string, first check to see if we are at        /* If we have matched an empty string, first check to see if we are at
2454        the end of the subject. If so, the /g loop is over. Otherwise, mimic        the end of the subject. If so, the /g loop is over. Otherwise, mimic what
2455        what Perl's /g options does. This turns out to be rather cunning. First        Perl's /g options does. This turns out to be rather cunning. First we set
2456        we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the        PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED and try the match again at the
2457        same point. If this fails (picked up above) we advance to the next        same point. If this fails (picked up above) we advance to the next
2458        character. */        character. */
2459    
2460        g_notempty = 0;        g_notempty = 0;
2461    
2462        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2463          {          {
2464          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2465          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2466          }          }
2467    
2468        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
# Line 2165  while (!done) Line 2493  while (!done)
2493      {      {
2494      new_free((void *)tables);      new_free((void *)tables);
2495      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
2496        locale_set = 0;
2497      }      }
2498    }    }
2499    

Legend:
Removed from v.91  
changed lines
  Added in v.455

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12