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

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

Parent Directory | Revision Log | View Patch Patch

-code/trunk/pcretest.c
revision 55 by nigel,
Sat Feb 24 21:39:46 2007 UTC
+code/branches/pcre16/pcretest.c
revision 810 by ph10,
Mon Dec 19 13:34:10 2011 UTC
 Line 2
  *             PCRE testing program               *
  *************************************************/
+ /* This program was hacked up as a tester for PCRE. I really should have
+ written it more tidily in the first place. Will I ever learn? It has grown and
+ been extended and consequently is now rather, er, *very* untidy in places. The
+ addition of 16-bit support has made it even worse. :-(
+ -----------------------------------------------------------------------------
+ Redistribution and use in source and binary forms, with or without
+ modification, are permitted provided that the following conditions are met:
+     * Redistributions of source code must retain the above copyright notice,
+       this list of conditions and the following disclaimer.
+     * Redistributions in binary form must reproduce the above copyright
+       notice, this list of conditions and the following disclaimer in the
+       documentation and/or other materials provided with the distribution.
+     * Neither the name of the University of Cambridge nor the names of its
+       contributors may be used to endorse or promote products derived from
+       this software without specific prior written permission.
+ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ POSSIBILITY OF SUCH DAMAGE.
+ -----------------------------------------------------------------------------
+ */
+ #ifdef HAVE_CONFIG_H
+ #include "config.h"
+ #endif
  #include <ctype.h>
  #include <stdio.h>
  #include <string.h>
  #include <stdlib.h>
  #include <time.h>
  #include <locale.h>
+ #include <errno.h>
+ #ifdef SUPPORT_LIBREADLINE
+ #ifdef HAVE_UNISTD_H
+ #include <unistd.h>
+ #endif
+ #include <readline/readline.h>
+ #include <readline/history.h>
+ #endif
+ /* A number of things vary for Windows builds. Originally, pcretest opened its
+ input and output without "b"; then I was told that "b" was needed in some
+ environments, so it was added for release 5.0 to both the input and output. (It
+ makes no difference on Unix-like systems.) Later I was told that it is wrong
+ for the input on Windows. I've now abstracted the modes into two macros that
+ are set here, to make it easier to fiddle with them, and removed "b" from the
+ input mode under Windows. */
+ #if defined(_WIN32) || defined(WIN32)
+ #include <io.h>                /* For _setmode() */
+ #include <fcntl.h>             /* For _O_BINARY */
+ #define INPUT_MODE   "r"
+ #define OUTPUT_MODE  "wb"
+ #ifndef isatty
+ #define isatty _isatty         /* This is what Windows calls them, I'm told, */
+ #endif                         /* though in some environments they seem to   */
+                                /* be already defined, hence the #ifndefs.    */
+ #ifndef fileno
+ #define fileno _fileno
+ #endif
- /* Use the internal info for displaying the results of pcre_study(). */
+ /* A user sent this fix for Borland Builder 5 under Windows. */
+ #ifdef __BORLANDC__
+ #define _setmode(handle, mode) setmode(handle, mode)
+ #endif
+ /* Not Windows */
+ #else
+ #include <sys/time.h>          /* These two includes are needed */
+ #include <sys/resource.h>      /* for setrlimit(). */
+ #define INPUT_MODE   "rb"
+ #define OUTPUT_MODE  "wb"
+ #endif
+ /* We have to include pcre_internal.h because we need the internal info for
+ displaying the results of pcre_study() and we also need to know about the
+ internal macros, structures, and other internal data values; pcretest has
+ "inside information" compared to a program that strictly follows the PCRE API.
+ Although pcre_internal.h does itself include pcre.h, we explicitly include it
+ here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
+ appropriately for an application, not for building PCRE. */
+ #include "pcre.h"
+ #include "pcre_internal.h"
+ /* The pcre_printint() function, which prints the internal form of a compiled
+ regex, is held in a separate file so that (a) it can be compiled in either
+-bit or 16-bit mode, and (b) it can be #included directly in pcre_compile.c
+ when that is compiled in debug mode. */
+ #ifdef SUPPORT_PCRE8
+ void pcre_printint(pcre *external_re, FILE *f, BOOL print_lengths);
+ #endif
+ #ifdef SUPPORT_PCRE16
+ void pcre16_printint(pcre *external_re, FILE *f, BOOL print_lengths);
+ #endif
- #include "internal.h"
+ /* We need access to some of the data tables that PCRE uses. So as not to have
+ to keep two copies, we include the source file here, changing the names of the
+ external symbols to prevent clashes. */
+ #define _pcre_ucp_gentype      ucp_gentype
+ #define _pcre_ucp_typerange    ucp_typerange
+ #define _pcre_utf8_table1      utf8_table1
+ #define _pcre_utf8_table1_size utf8_table1_size
+ #define _pcre_utf8_table2      utf8_table2
+ #define _pcre_utf8_table3      utf8_table3
+ #define _pcre_utf8_table4      utf8_table4
+ #define _pcre_utt              utt
+ #define _pcre_utt_size         utt_size
+ #define _pcre_utt_names        utt_names
+ #define _pcre_OP_lengths       OP_lengths
+ #include "pcre_tables.c"
+ /* The definition of the macro PRINTABLE, which determines whether to print an
+ output character as-is or as a hex value when showing compiled patterns, is
+ the same as in the printint.src file. We uses it here in cases when the locale
+ has not been explicitly changed, so as to get consistent output from systems
+ that differ in their output from isprint() even in the "C" locale. */
+ #ifdef EBCDIC
+ #define PRINTABLE(c) ((c) >= 64 && (c) < 255)
+ #else
+ #define PRINTABLE(c) ((c) >= 32 && (c) < 127)
+ #endif
+ #define PRINTOK(c) (locale_set? isprint(c) : PRINTABLE(c))
  /* It is possible to compile this test program without including support for
  testing the POSIX interface, though this is not available via the standard
-Line 21 
 Makefile. */
+Line 160 
 Makefile. */
  #include "pcreposix.h"
  #endif
+ /* It is also possible, originally for the benefit of a version that was
+ imported into Exim, to build pcretest without support for UTF8 (define NOUTF8),
+ without the interface to the DFA matcher (NODFA), and without the doublecheck
+ of the old "info" function (define NOINFOCHECK). In fact, we automatically cut
+ out the UTF8 support if PCRE is built without it. */
+ #ifndef SUPPORT_UTF8
+ #ifndef NOUTF8
+ #define NOUTF8
+ #endif
+ #endif
+ /* To make the code a bit tidier for 8-bit and 16-bit support, we define macros
+ for all the pcre[16]_xxx functions (except pcre16_fullinfo, which is called
+ only from one place and is handled differently). I couldn't dream up any way of
+ using a single macro to do this in a generic way, because of the many different
+ argument requirements. We know that at least one of SUPPORT_PCRE8 and
+ SUPPORT_PCRE16 must be set. First define macros for each individual mode; then
+ use these in the definitions of generic macros. */
+ #ifdef SUPPORT_PCRE8
+ #define PCHARS8(lv, p, len, f) \
+   lv = pchars((pcre_uint8 *)p, len, f)
+ #define PCHARSV8(p, len, f) \
+   (void)pchars((pcre_uint8 *)p, len, f)
+ #define PCRE_COMPILE8(re, pat, options, error, erroffset, tables) \
+   re = pcre_compile((char *)pat, options, error, erroffset, tables)
+ #define PCRE_EXEC8(count, re, extra, bptr, len, start_offset, options, \
+     offsets, size_offsets) \
+   count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options, \
+     offsets, size_offsets)
+ #define PCRE_STUDY8(extra, re, options, error) \
+   extra = pcre_study(re, options, error)
+ #define PCRE_FREE_STUDY8(extra) \
+   pcre_free_study(extra)
+ #endif /* SUPPORT_PCRE8 */
+ #ifdef SUPPORT_PCRE16
+ #define PCHARS16(lv, p, len, f) \
+   lv = pchars16((PCRE_SPTR16)p, len, f)
+ #define PCHARSV16(p, len, f) \
+   (void)pchars16((PCRE_SPTR16)p, len, f)
+ #define PCRE_COMPILE16(re, pat, options, error, erroffset, tables) \
+   re = pcre16_compile((PCRE_SPTR16)pat, options, error, erroffset, tables)
+ #define PCRE_EXEC16(count, re, extra, bptr, len, start_offset, options, \
+     offsets, size_offsets) \
+   count = pcre16_exec(re, extra, (PCRE_SPTR16)bptr, len, start_offset, \
+     options, offsets, size_offsets)
+ #define PCRE_FREE_STUDY16(extra) \
+   pcre16_free_study(extra)
+ #define PCRE_STUDY16(extra, re, options, error) \
+   extra = pcre16_study(re, options, error)
+ #endif /* SUPPORT_PCRE16 */
+ /* ----- Both modes are supported; a runtime test is needed ----- */
+ #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
+ #define PCHARS(lv, p, len, f) \
+   if (use_pcre16) \
+     PCHARS16(lv, p, len, f); \
+   else \
+     PCHARS8(lv, p, len, f)
+ #define PCHARSV(p, len, f) \
+   if (use_pcre16) \
+     PCHARSV16(p, len, f); \
+   else \
+     PCHARSV8(p, len, f)
+ #define PCRE_COMPILE(re, pat, options, error, erroffset, tables) \
+   if (use_pcre16) \
+     PCRE_COMPILE16(re, pat, options, error, erroffset, tables); \
+   else \
+     PCRE_COMPILE8(re, pat, options, error, erroffset, tables)
+ #define PCRE_EXEC(count, re, extra, bptr, len, start_offset, options, \
+     offsets, size_offsets) \
+   if (use_pcre16) \
+     PCRE_EXEC16(count, re, extra, bptr, len, start_offset, options, \
+       offsets, size_offsets); \
+   else \
+     PCRE_EXEC8(count, re, extra, bptr, len, start_offset, options, \
+       offsets, size_offsets)
+ #define PCRE_FREE_STUDY(extra) \
+   if (use_pcre16) \
+     PCRE_FREE_STUDY16(extra); \
+   else \
+     PCRE_FREE_STUDY8(extra)
+ #define PCRE_STUDY(extra, re, options, error) \
+   if (use_pcre16) \
+     PCRE_STUDY16(extra, re, options, error); \
+   else \
+     PCRE_STUDY8(extra, re, options, error)
+ /* ----- Only 8-bit mode is supported ----- */
+ #elif defined SUPPORT_PCRE8
+ #define PCHARS           PCHARS8
+ #define PCHARSV          PCHARSV8
+ #define PCRE_COMPILE     PCRE_COMPILE8
+ #define PCRE_EXEC        PCRE_EXEC8
+ #define PCRE_FREE_STUDY  PCRE_FREE_STUDY8
+ #define PCRE_STUDY       PCRE_STUDY8
+ /* ----- Only 16-bit mode is supported ----- */
+ #else
+ #define PCHARS           PCHARS16
+ #define PCHARSV          PCHARSV16
+ #define PCRE_COMPILE     PCRE_COMPILE16
+ #define PCRE_EXEC        PCRE_EXEC16
+ #define PCRE_FREE_STUDY  PCRE_FREE_STUDY16
+ #define PCRE_STUDY       PCRE_STUDY16
+ #endif
+ /* ----- End of mode-specific function call macros ----- */
+ /* Other parameters */
  #ifndef CLOCKS_PER_SEC
  #ifdef CLK_TCK
  #define CLOCKS_PER_SEC CLK_TCK
-Line 29 
 Makefile. */
+Line 305 
 Makefile. */
  #endif
  #endif
- #define LOOPREPEAT 20000
+ /* This is the default loop count for timing. */
+ #define LOOPREPEAT 500000
+ /* Static variables */
  static FILE *outfile;
  static int log_store = 0;
+ static int callout_count;
+ static int callout_extra;
+ static int callout_fail_count;
+ static int callout_fail_id;
+ static int debug_lengths;
+ static int first_callout;
+ static int locale_set = 0;
+ static int show_malloc;
+ static int use_utf;
  static size_t gotten_store;
+ static size_t first_gotten_store = 0;
+ static const unsigned char *last_callout_mark = NULL;
+ /* The buffers grow automatically if very long input lines are encountered. */
+ static int buffer_size = 50000;
+ static pcre_uint8 *buffer = NULL;
+ static pcre_uint8 *dbuffer = NULL;
+ static pcre_uint8 *pbuffer = NULL;
+ #ifdef SUPPORT_PCRE16
+ static int buffer16_size = 0;
+ static pcre_uint16 *buffer16 = NULL;
+ #endif
- static int utf8_table1[] = {
+ /* If we have 8-bit support, default use_pcre16 to false; if there is also
-x0000007f, 0x000007ff, 0x0000ffff, 0x001fffff, 0x03ffffff, 0x7fffffff};
+-bit support, it can be changed by an option. If there is no 8-bit support,
+ there must be 16-bit support, so default it to 1. */
- static int utf8_table2[] = {
+ #ifdef SUPPORT_PCRE8
-, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc};
+ static int use_pcre16 = 0;
+ #else
+ static int use_pcre16 = 1;
+ #endif
- static int utf8_table3[] = {
+ /* Textual explanations for runtime error codes */
-xff, 0x1f, 0x0f, 0x07, 0x03, 0x01};
+ static const char *errtexts[] = {
+   NULL,  /* 0 is no error */
+   NULL,  /* NOMATCH is handled specially */
+   "NULL argument passed",
+   "bad option value",
+   "magic number missing",
+   "unknown opcode - pattern overwritten?",
+   "no more memory",
+   NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
+   "match limit exceeded",
+   "callout error code",
+   NULL,  /* BADUTF8 is handled specially */
+   "bad UTF-8 offset",
+   NULL,  /* PARTIAL is handled specially */
+   "not used - internal error",
+   "internal error - pattern overwritten?",
+   "bad count value",
+   "item unsupported for DFA matching",
+   "backreference condition or recursion test not supported for DFA matching",
+   "match limit not supported for DFA matching",
+   "workspace size exceeded in DFA matching",
+   "too much recursion for DFA matching",
+   "recursion limit exceeded",
+   "not used - internal error",
+   "invalid combination of newline options",
+   "bad offset value",
+   NULL,  /* SHORTUTF8 is handled specially */
+   "nested recursion at the same subject position",
+   "JIT stack limit reached",
+   "pattern compiled in wrong mode (8-bit/16-bit error)"
+ };
  /*************************************************
- *       Convert character value to UTF-8         *
+ *         Alternate character tables             *
  *************************************************/
- /* This function takes an integer value in the range 0 - 0x7fffffff
+ /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
- and encodes it as a UTF-8 character in 0 to 6 bytes.
+ using the default tables of the library. However, the T option can be used to
+ select alternate sets of tables, for different kinds of testing. Note also that
+ the L (locale) option also adjusts the tables. */
+ /* This is the set of tables distributed as default with PCRE. It recognizes
+ only ASCII characters. */
+ static const pcre_uint8 tables0[] = {
+ /* This table is a lower casing table. */
+,  1,  2,  3,  4,  5,  6,  7,
+,  9, 10, 11, 12, 13, 14, 15,
+, 17, 18, 19, 20, 21, 22, 23,
+, 25, 26, 27, 28, 29, 30, 31,
+, 33, 34, 35, 36, 37, 38, 39,
+, 41, 42, 43, 44, 45, 46, 47,
+, 49, 50, 51, 52, 53, 54, 55,
+, 57, 58, 59, 60, 61, 62, 63,
+, 97, 98, 99,100,101,102,103,
+,105,106,107,108,109,110,111,
+,113,114,115,116,117,118,119,
+,121,122, 91, 92, 93, 94, 95,
+, 97, 98, 99,100,101,102,103,
+,105,106,107,108,109,110,111,
+,113,114,115,116,117,118,119,
+,121,122,123,124,125,126,127,
+,129,130,131,132,133,134,135,
+,137,138,139,140,141,142,143,
+,145,146,147,148,149,150,151,
+,153,154,155,156,157,158,159,
+,161,162,163,164,165,166,167,
+,169,170,171,172,173,174,175,
+,177,178,179,180,181,182,183,
+,185,186,187,188,189,190,191,
+,193,194,195,196,197,198,199,
+,201,202,203,204,205,206,207,
+,209,210,211,212,213,214,215,
+,217,218,219,220,221,222,223,
+,225,226,227,228,229,230,231,
+,233,234,235,236,237,238,239,
+,241,242,243,244,245,246,247,
+,249,250,251,252,253,254,255,
+ /* This table is a case flipping table. */
+,  1,  2,  3,  4,  5,  6,  7,
+,  9, 10, 11, 12, 13, 14, 15,
+, 17, 18, 19, 20, 21, 22, 23,
+, 25, 26, 27, 28, 29, 30, 31,
+, 33, 34, 35, 36, 37, 38, 39,
+, 41, 42, 43, 44, 45, 46, 47,
+, 49, 50, 51, 52, 53, 54, 55,
+, 57, 58, 59, 60, 61, 62, 63,
+, 97, 98, 99,100,101,102,103,
+,105,106,107,108,109,110,111,
+,113,114,115,116,117,118,119,
+,121,122, 91, 92, 93, 94, 95,
+, 65, 66, 67, 68, 69, 70, 71,
+, 73, 74, 75, 76, 77, 78, 79,
+, 81, 82, 83, 84, 85, 86, 87,
+, 89, 90,123,124,125,126,127,
+,129,130,131,132,133,134,135,
+,137,138,139,140,141,142,143,
+,145,146,147,148,149,150,151,
+,153,154,155,156,157,158,159,
+,161,162,163,164,165,166,167,
+,169,170,171,172,173,174,175,
+,177,178,179,180,181,182,183,
+,185,186,187,188,189,190,191,
+,193,194,195,196,197,198,199,
+,201,202,203,204,205,206,207,
+,209,210,211,212,213,214,215,
+,217,218,219,220,221,222,223,
+,225,226,227,228,229,230,231,
+,233,234,235,236,237,238,239,
+,241,242,243,244,245,246,247,
+,249,250,251,252,253,254,255,
+ /* This table contains bit maps for various character classes. Each map is 32
+ bytes long and the bits run from the least significant end of each byte. The
+ classes that have their own maps are: space, xdigit, digit, upper, lower, word,
+ graph, print, punct, and cntrl. Other classes are built from combinations. */
+x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
+x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
+xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
+xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
+xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
+x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+ /* This table identifies various classes of character by individual bits:
+x01   white space character
+x02   letter
+x04   decimal digit
+x08   hexadecimal digit
+x10   alphanumeric or '_'
+x80   regular expression metacharacter or binary zero
+ */
- Arguments:
+x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
-   cvalue     the character value
+x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
-   buffer     pointer to buffer for result - at least 6 bytes long
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
+x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
+x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
+x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
+x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
+x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
+x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
+x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
+x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
+x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
+x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
+x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
+x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
+x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
+ /* This is a set of tables that came orginally from a Windows user. It seems to
+ be at least an approximation of ISO 8859. In particular, there are characters
+ greater than 128 that are marked as spaces, letters, etc. */
+ static const pcre_uint8 tables1[] = {
+,1,2,3,4,5,6,7,
+,9,10,11,12,13,14,15,
+,17,18,19,20,21,22,23,
+,25,26,27,28,29,30,31,
+,33,34,35,36,37,38,39,
+,41,42,43,44,45,46,47,
+,49,50,51,52,53,54,55,
+,57,58,59,60,61,62,63,
+,97,98,99,100,101,102,103,
+,105,106,107,108,109,110,111,
+,113,114,115,116,117,118,119,
+,121,122,91,92,93,94,95,
+,97,98,99,100,101,102,103,
+,105,106,107,108,109,110,111,
+,113,114,115,116,117,118,119,
+,121,122,123,124,125,126,127,
+,129,130,131,132,133,134,135,
+,137,138,139,140,141,142,143,
+,145,146,147,148,149,150,151,
+,153,154,155,156,157,158,159,
+,161,162,163,164,165,166,167,
+,169,170,171,172,173,174,175,
+,177,178,179,180,181,182,183,
+,185,186,187,188,189,190,191,
+,225,226,227,228,229,230,231,
+,233,234,235,236,237,238,239,
+,241,242,243,244,245,246,215,
+,249,250,251,252,253,254,223,
+,225,226,227,228,229,230,231,
+,233,234,235,236,237,238,239,
+,241,242,243,244,245,246,247,
+,249,250,251,252,253,254,255,
+,1,2,3,4,5,6,7,
+,9,10,11,12,13,14,15,
+,17,18,19,20,21,22,23,
+,25,26,27,28,29,30,31,
+,33,34,35,36,37,38,39,
+,41,42,43,44,45,46,47,
+,49,50,51,52,53,54,55,
+,57,58,59,60,61,62,63,
+,97,98,99,100,101,102,103,
+,105,106,107,108,109,110,111,
+,113,114,115,116,117,118,119,
+,121,122,91,92,93,94,95,
+,65,66,67,68,69,70,71,
+,73,74,75,76,77,78,79,
+,81,82,83,84,85,86,87,
+,89,90,123,124,125,126,127,
+,129,130,131,132,133,134,135,
+,137,138,139,140,141,142,143,
+,145,146,147,148,149,150,151,
+,153,154,155,156,157,158,159,
+,161,162,163,164,165,166,167,
+,169,170,171,172,173,174,175,
+,177,178,179,180,181,182,183,
+,185,186,187,188,189,190,191,
+,225,226,227,228,229,230,231,
+,233,234,235,236,237,238,239,
+,241,242,243,244,245,246,215,
+,249,250,251,252,253,254,223,
+,193,194,195,196,197,198,199,
+,201,202,203,204,205,206,207,
+,209,210,211,212,213,214,247,
+,217,218,219,220,221,222,255,
+,62,0,0,1,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,1,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,255,3,
+,0,0,0,126,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,255,3,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,12,2,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,0,0,
+,255,255,7,0,0,0,0,
+,0,0,0,0,0,0,0,
+,255,127,127,0,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,254,255,255,7,
+,0,0,0,0,4,32,4,
+,0,0,128,255,255,127,255,
+,0,0,0,0,0,255,3,
+,255,255,135,254,255,255,7,
+,0,0,0,0,4,44,6,
+,255,127,255,255,255,127,255,
+,0,0,0,254,255,255,255,
+,255,255,255,255,255,255,127,
+,0,0,0,254,255,255,255,
+,255,255,255,255,255,255,255,
+,2,0,0,255,255,255,255,
+,255,255,255,255,255,255,127,
+,0,0,0,255,255,255,255,
+,255,255,255,255,255,255,255,
+,0,0,0,254,255,0,252,
+,0,0,248,1,0,0,120,
+,0,0,0,254,255,255,255,
+,0,128,0,0,0,128,0,
+,255,255,255,0,0,0,0,
+,0,0,0,0,0,0,128,
+,255,255,255,0,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,0,0,
+,1,1,0,1,1,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,128,0,0,0,
+,128,128,128,0,0,128,0,
+,28,28,28,28,28,28,28,
+,28,0,0,0,0,0,128,
+,26,26,26,26,26,26,18,
+,18,18,18,18,18,18,18,
+,18,18,18,18,18,18,18,
+,18,18,128,128,0,128,16,
+,26,26,26,26,26,26,18,
+,18,18,18,18,18,18,18,
+,18,18,18,18,18,18,18,
+,18,18,128,128,0,0,0,
+,0,0,0,0,1,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,0,0,
+,0,0,0,0,0,0,0,
+,0,18,0,0,0,0,0,
+,0,20,20,0,18,0,0,
+,20,18,0,0,0,0,0,
+,18,18,18,18,18,18,18,
+,18,18,18,18,18,18,18,
+,18,18,18,18,18,18,0,
+,18,18,18,18,18,18,18,
+,18,18,18,18,18,18,18,
+,18,18,18,18,18,18,18,
+,18,18,18,18,18,18,0,
+,18,18,18,18,18,18,18
+ };
- Returns:     number of characters placed in the buffer
-              -1 if input character is negative
-if input character is positive but too big (only when
-              int is longer than 32 bits)
- */
- static int
- ord2utf8(int cvalue, unsigned char *buffer)
+ #ifndef HAVE_STRERROR
+ /*************************************************
+ *     Provide strerror() for non-ANSI libraries  *
+ *************************************************/
+ /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
+ in their libraries, but can provide the same facility by this simple
+ alternative function. */
+ extern int   sys_nerr;
+ extern char *sys_errlist[];
+ char *
+ strerror(int n)
  {
- register int i, j;
+ if (n < 0 || n >= sys_nerr) return "unknown error number";
- for (i = 0; i < sizeof(utf8_table1)/sizeof(int); i++)
+ return sys_errlist[n];
-   if (cvalue <= utf8_table1[i]) break;
+ }
- if (i >= sizeof(utf8_table1)/sizeof(int)) return 0;
+ #endif /* HAVE_STRERROR */
- if (cvalue < 0) return -1;
- *buffer++ = utf8_table2[i] | (cvalue & utf8_table3[i]);
- cvalue >>= 6 - i;
+ /*************************************************
- for (j = 0; j < i; j++)
+ *         JIT memory callback                    *
-   {
+ *************************************************/
-   *buffer++ = 0x80 | (cvalue & 0x3f);
-   cvalue >>= 6;
+ static pcre_jit_stack* jit_callback(void *arg)
-   }
+ {
- return i + 1;
+ return (pcre_jit_stack *)arg;
  }
-Line 92 
 return i + 1;
+Line 749 
 return i + 1;
  and returns the value of the character.
  Argument:
-   buffer   a pointer to the byte vector
+   utf8bytes   a pointer to the byte vector
-   vptr     a pointer to an int to receive the value
+   vptr        a pointer to an int to receive the value
- Returns:   >  0 => the number of bytes consumed
+ Returns:      >  0 => the number of bytes consumed
-            -6 to 0 => malformed UTF-8 character at offset = (-return)
+               -6 to 0 => malformed UTF-8 character at offset = (-return)
  */
- int
+ #if !defined NOUTF8
- utf82ord(unsigned char *buffer, int *vptr)
+ static int
+ utf82ord(pcre_uint8 *utf8bytes, int *vptr)
  {
- int c = *buffer++;
+ int c = *utf8bytes++;
  int d = c;
  int i, j, s;
-Line 117 
 if (i == 0 || i == 6) return 0;        /
+Line 776 
 if (i == 0 || i == 6) return 0;        /
  /* i now has a value in the range 1-5 */
- d = c & utf8_table3[i];
+ s = 6*i;
- s = 6 - i;
+ d = (c & utf8_table3[i]) << s;
  for (j = 0; j < i; j++)
    {
-   c = *buffer++;
+   c = *utf8bytes++;
    if ((c & 0xc0) != 0x80) return -(j+1);
+   s -= 6;
    d |= (c & 0x3f) << s;
-   s += 6;
    }
  /* Check that encoding was the correct unique one */
- for (j = 0; j < sizeof(utf8_table1)/sizeof(int); j++)
+ for (j = 0; j < utf8_table1_size; j++)
    if (d <= utf8_table1[j]) break;
  if (j != i) return -(i+1);
-Line 140 
 if (j != i) return -(i+1);
+Line 799 
 if (j != i) return -(i+1);
  return i+1;
  }
+ #endif
+ /*************************************************
+ *       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.
- /* Debugging function to print the internal form of the regex. This is the same
+ Arguments:
- code as contained in pcre.c under the DEBUG macro. */
+   cvalue     the character value
+   utf8bytes  pointer to buffer for result - at least 6 bytes long
- static const char *OP_names[] = {
+ Returns:     number of characters placed in the buffer
-   "End", "\\A", "\\B", "\\b", "\\D", "\\d",
+ */
-   "\\S", "\\s", "\\W", "\\w", "\\Z", "\\z",
-   "Opt", "^", "$", "Any", "chars", "not",
-   "*", "*?", "+", "+?", "?", "??", "{", "{", "{",
-   "*", "*?", "+", "+?", "?", "??", "{", "{", "{",
-   "*", "*?", "+", "+?", "?", "??", "{", "{", "{",
-   "*", "*?", "+", "+?", "?", "??", "{", "{",
-   "class", "Ref", "Recurse",
-   "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",
-   "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cref",
-   "Brazero", "Braminzero", "Branumber", "Bra"
- };
+ #if !defined NOUTF8
- static void print_internals(pcre *re)
+ static int
+ ord2utf8(int cvalue, pcre_uint8 *utf8bytes)
  {
- unsigned char *code = ((real_pcre *)re)->code;
+ register int i, j;
+ for (i = 0; i < utf8_table1_size; i++)
+   if (cvalue <= utf8_table1[i]) break;
+ utf8bytes += i;
+ for (j = i; j > 0; j--)
+  {
+  *utf8bytes-- = 0x80 | (cvalue & 0x3f);
+  cvalue >>= 6;
+  }
+ *utf8bytes = utf8_table2[i] | cvalue;
+ return i + 1;
+ }
+ #endif
+ #ifdef SUPPORT_PCRE16
+ /*************************************************
+ *         Convert a string to 16-bit             *
+ *************************************************/
+ /* In non-UTF mode, the space needed for a 16-bit string is exactly double the
+-bit size. For a UTF-8 string, the size needed for UTF-16 is no more than
+ double, because up to 0xffff uses no more than 3 bytes in UTF-8 but possibly 4
+ in UTF-16. Higher values use 4 bytes in UTF-8 and up to 4 bytes in UTF-16. The
+ result is always left in buffer16.
- fprintf(outfile, "------------------------------------------------------------------\n");
+ Arguments:
+   p          points to a byte string
+   utf        true if UTF-8 (to be converted to UTF-16)
+   len        number of bytes in the string (excluding trailing zero)
+ Returns:     number of 16-bit data items used (excluding trailing zero)
+              OR -1 if a UTF-8 string is malformed
+ */
- for(;;)
+ static int
+ to16(pcre_uint8 *p, int utf, int len)
+ {
+ pcre_uint16 *pp;
+ if (buffer16_size < 2*len + 2)
+   {
+   if (buffer16 != NULL) free(buffer16);
+   buffer16_size = 2*len + 2;
+   buffer16 = (pcre_uint16 *)malloc(buffer16_size);
+   if (buffer16 == NULL)
+     {
+     fprintf(stderr, "pcretest: malloc(%d) failed for buffer16\n", buffer16_size);
+     exit(1);
+     }
+   }
+ pp = buffer16;
+ if (!utf)
+   {
+   while (len-- > 0) *pp++ = *p++;
+   }
+ else
    {
    int c;
-   int charlength;
+   while (len > 0)
+     {
+     int chlen = utf82ord(p, &c);
+     if (chlen <= 0) return -1;
+     p += chlen;
+     len -= chlen;
+     if (c < 0x10000) *pp++ = c; else
+       {
+       c -= 0x10000;
+       *pp++ = 0xD800 | (c >> 10);
+       *pp++ = 0xDC00 | (c & 0x3ff);
+       }
+     }
+   }
-   fprintf(outfile, "%3d ", (int)(code - ((real_pcre *)re)->code));
+ *pp = 0;
+ return pp - buffer16;
+ }
+ #endif
+ /*************************************************
+ *        Read or extend an input line            *
+ *************************************************/
+ /* Input lines are read into buffer, but both patterns and data lines can be
+ continued over multiple input lines. In addition, if the buffer fills up, we
+ want to automatically expand it so as to be able to handle extremely large
+ lines that are needed for certain stress tests. When the input buffer is
+ expanded, the other two buffers must also be expanded likewise, and the
+ contents of pbuffer, which are a copy of the input for callouts, must be
+ preserved (for when expansion happens for a data line). This is not the most
+ optimal way of handling this, but hey, this is just a test program!
+ Arguments:
+   f            the file to read
+   start        where in buffer to start (this *must* be within buffer)
+   prompt       for stdin or readline()
+ Returns:       pointer to the start of new data
+                could be a copy of start, or could be moved
+                NULL if no data read and EOF reached
+ */
-   if (*code >= OP_BRA)
+ static pcre_uint8 *
+ extend_inputline(FILE *f, pcre_uint8 *start, const char *prompt)
+ {
+ pcre_uint8 *here = start;
+ for (;;)
+   {
+   int rlen = (int)(buffer_size - (here - buffer));
+   if (rlen > 1000)
      {
-     if (*code - OP_BRA > EXTRACT_BASIC_MAX)
+     int dlen;
-       fprintf(outfile, "%3d Bra extra", (code[1] << 8) + code[2]);
+     /* If libreadline support is required, use readline() to read a line if the
+     input is a terminal. Note that readline() removes the trailing newline, so
+     we must put it back again, to be compatible with fgets(). */
+ #ifdef SUPPORT_LIBREADLINE
+     if (isatty(fileno(f)))
+       {
+       size_t len;
+       char *s = readline(prompt);
+       if (s == NULL) return (here == start)? NULL : start;
+       len = strlen(s);
+       if (len > 0) add_history(s);
+       if (len > rlen - 1) len = rlen - 1;
+       memcpy(here, s, len);
+       here[len] = '\n';
+       here[len+1] = 0;
+       free(s);
+       }
      else
-       fprintf(outfile, "%3d Bra %d", (code[1] << 8) + code[2], *code - OP_BRA);
+ #endif
-     code += 2;
+     /* Read the next line by normal means, prompting if the file is stdin. */
+       {
+       if (f == stdin) printf("%s", prompt);
+       if (fgets((char *)here, rlen,  f) == NULL)
+         return (here == start)? NULL : start;
+       }
+     dlen = (int)strlen((char *)here);
+     if (dlen > 0 && here[dlen - 1] == '\n') return start;
+     here += dlen;
      }
-   else switch(*code)
+   else
      {
-     case OP_END:
+     int new_buffer_size = 2*buffer_size;
-     fprintf(outfile, "    %s\n", OP_names[*code]);
+     pcre_uint8 *new_buffer = (pcre_uint8 *)malloc(new_buffer_size);
-     fprintf(outfile, "------------------------------------------------------------------\n");
+     pcre_uint8 *new_dbuffer = (pcre_uint8 *)malloc(new_buffer_size);
-     return;
+     pcre_uint8 *new_pbuffer = (pcre_uint8 *)malloc(new_buffer_size);
-     case OP_OPT:
-     fprintf(outfile, " %.2x %s", code[1], OP_names[*code]);
-     code++;
-     break;
-     case OP_CHARS:
-     charlength = *(++code);
-     fprintf(outfile, "%3d ", charlength);
-     while (charlength-- > 0)
-       if (isprint(c = *(++code))) fprintf(outfile, "%c", c);
-         else fprintf(outfile, "\\x%02x", c);
-     break;
-     case OP_KETRMAX:
-     case OP_KETRMIN:
-     case OP_ALT:
-     case OP_KET:
-     case OP_ASSERT:
-     case OP_ASSERT_NOT:
-     case OP_ASSERTBACK:
-     case OP_ASSERTBACK_NOT:
-     case OP_ONCE:
-     case OP_COND:
-     case OP_BRANUMBER:
-     case OP_REVERSE:
-     case OP_CREF:
-     fprintf(outfile, "%3d %s", (code[1] << 8) + code[2], OP_names[*code]);
-     code += 2;
-     break;
-     case OP_STAR:
-     case OP_MINSTAR:
-     case OP_PLUS:
-     case OP_MINPLUS:
-     case OP_QUERY:
-     case OP_MINQUERY:
-     case OP_TYPESTAR:
-     case OP_TYPEMINSTAR:
-     case OP_TYPEPLUS:
-     case OP_TYPEMINPLUS:
-     case OP_TYPEQUERY:
-     case OP_TYPEMINQUERY:
-     if (*code >= OP_TYPESTAR)
-       fprintf(outfile, "    %s", OP_names[code[1]]);
-     else if (isprint(c = code[1])) fprintf(outfile, "    %c", c);
-       else fprintf(outfile, "    \\x%02x", c);
-     fprintf(outfile, "%s", OP_names[*code++]);
-     break;
-     case OP_EXACT:
-     case OP_UPTO:
-     case OP_MINUPTO:
-     if (isprint(c = code[3])) fprintf(outfile, "    %c{", c);
-       else fprintf(outfile, "    \\x%02x{", c);
-     if (*code != OP_EXACT) fprintf(outfile, ",");
-     fprintf(outfile, "%d}", (code[1] << 8) + code[2]);
-     if (*code == OP_MINUPTO) fprintf(outfile, "?");
-     code += 3;
-     break;
-     case OP_TYPEEXACT:
-     case OP_TYPEUPTO:
-     case OP_TYPEMINUPTO:
-     fprintf(outfile, "    %s{", OP_names[code[3]]);
-     if (*code != OP_TYPEEXACT) fprintf(outfile, "0,");
-     fprintf(outfile, "%d}", (code[1] << 8) + code[2]);
-     if (*code == OP_TYPEMINUPTO) fprintf(outfile, "?");
-     code += 3;
-     break;
-     case OP_NOT:
-     if (isprint(c = *(++code))) fprintf(outfile, "    [^%c]", c);
-       else fprintf(outfile, "    [^\\x%02x]", c);
-     break;
-     case OP_NOTSTAR:
-     case OP_NOTMINSTAR:
-     case OP_NOTPLUS:
-     case OP_NOTMINPLUS:
-     case OP_NOTQUERY:
-     case OP_NOTMINQUERY:
-     if (isprint(c = code[1])) fprintf(outfile, "    [^%c]", c);
-       else fprintf(outfile, "    [^\\x%02x]", c);
-     fprintf(outfile, "%s", OP_names[*code++]);
-     break;
-     case OP_NOTEXACT:
-     case OP_NOTUPTO:
-     case OP_NOTMINUPTO:
-     if (isprint(c = code[3])) fprintf(outfile, "    [^%c]{", c);
-       else fprintf(outfile, "    [^\\x%02x]{", c);
-     if (*code != OP_NOTEXACT) fprintf(outfile, ",");
-     fprintf(outfile, "%d}", (code[1] << 8) + code[2]);
-     if (*code == OP_NOTMINUPTO) fprintf(outfile, "?");
-     code += 3;
-     break;
-     case OP_REF:
-     fprintf(outfile, "    \\%d", (code[1] << 8) | code[2]);
-     code += 3;
-     goto CLASS_REF_REPEAT;
-     case OP_CLASS:
-       {
-       int i, min, max;
-       code++;
-       fprintf(outfile, "    [");
-       for (i = 0; i < 256; i++)
-         {
-         if ((code[i/8] & (1 << (i&7))) != 0)
-           {
-           int j;
-           for (j = i+1; j < 256; j++)
-             if ((code[j/8] & (1 << (j&7))) == 0) break;
-           if (i == '-' || i == ']') fprintf(outfile, "\\");
-           if (isprint(i)) fprintf(outfile, "%c", i); else fprintf(outfile, "\\x%02x", i);
-           if (--j > i)
-             {
-             fprintf(outfile, "-");
-             if (j == '-' || j == ']') fprintf(outfile, "\\");
-             if (isprint(j)) fprintf(outfile, "%c", j); else fprintf(outfile, "\\x%02x", j);
-             }
-           i = j;
-           }
-         }
-       fprintf(outfile, "]");
-       code += 32;
-       CLASS_REF_REPEAT:
+     if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
+       {
+       fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
+       exit(1);
+       }
-       switch(*code)
+     memcpy(new_buffer, buffer, buffer_size);
-         {
+     memcpy(new_pbuffer, pbuffer, buffer_size);
-         case OP_CRSTAR:
-         case OP_CRMINSTAR:
-         case OP_CRPLUS:
-         case OP_CRMINPLUS:
-         case OP_CRQUERY:
-         case OP_CRMINQUERY:
-         fprintf(outfile, "%s", OP_names[*code]);
-         break;
-         case OP_CRRANGE:
+     buffer_size = new_buffer_size;
-         case OP_CRMINRANGE:
-         min = (code[1] << 8) + code[2];
-         max = (code[3] << 8) + code[4];
-         if (max == 0) fprintf(outfile, "{%d,}", min);
-         else fprintf(outfile, "{%d,%d}", min, max);
-         if (*code == OP_CRMINRANGE) fprintf(outfile, "?");
-         code += 4;
-         break;
-         default:
+     start = new_buffer + (start - buffer);
-         code--;
+     here = new_buffer + (here - buffer);
-         }
-       }
-     break;
-     /* Anything else is just a one-node item */
+     free(buffer);
+     free(dbuffer);
+     free(pbuffer);
-     default:
+     buffer = new_buffer;
-     fprintf(outfile, "    %s", OP_names[*code]);
+     dbuffer = new_dbuffer;
-     break;
+     pbuffer = new_pbuffer;
      }
+   }
+ return NULL;  /* Control never gets here */
+ }
-   code++;
+ /*************************************************
-   fprintf(outfile, "\n");
+ *          Read number from string               *
+ *************************************************/
+ /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess
+ around with conditional compilation, just do the job by hand. It is only used
+ for unpicking arguments, so just keep it simple.
+ Arguments:
+   str           string to be converted
+   endptr        where to put the end pointer
+ Returns:        the unsigned long
+ */
+ static int
+ get_value(pcre_uint8 *str, pcre_uint8 **endptr)
+ {
+ int result = 0;
+ while(*str != 0 && isspace(*str)) str++;
+ while (isdigit(*str)) result = result * 10 + (int)(*str++ - '0');
+ *endptr = str;
+ return(result);
+ }
+ /*************************************************
+ *             Print one character                *
+ *************************************************/
+ /* Print a single character either literally, or as a hex escape. */
+ static int pchar(int c, FILE *f)
+ {
+ if (PRINTOK(c))
+   {
+   if (f != NULL) fprintf(f, "%c", c);
+   return 1;
    }
+ if (c < 0x100)
+   {
+   if (use_utf)
+     {
+     if (f != NULL) fprintf(f, "\\x{%02x}", c);
+     return 6;
+     }
+   else
+     {
+     if (f != NULL) fprintf(f, "\\x%02x", c);
+     return 4;
+     }
+   }
+ if (f != NULL) fprintf(f, "\\x{%02x}", c);
+ return (c <= 0x000000ff)? 6 :
+        (c <= 0x00000fff)? 7 :
+        (c <= 0x0000ffff)? 8 :
+        (c <= 0x000fffff)? 9 : 10;
  }
- /* Character string printing function. A "normal" and a UTF-8 version. */
+ #ifdef SUPPORT_PCRE8
+ /*************************************************
+ *         Print 8-bit character string           *
+ *************************************************/
+ /* Must handle UTF-8 strings in utf8 mode. Yields number of characters printed.
+ If handed a NULL file, just counts chars without printing. */
- static void pchars(unsigned char *p, int length, int utf8)
+ static int pchars(pcre_uint8 *p, int length, FILE *f)
  {
- int c;
+ int c = 0;
+ int yield = 0;
  while (length-- > 0)
    {
-   if (utf8)
+ #if !defined NOUTF8
+   if (use_utf)
      {
      int rc = utf82ord(p, &c);
-     if (rc > 0)
+     if (rc > 0 && rc <= length + 1)   /* Mustn't run over the end */
        {
        length -= rc - 1;
        p += rc;
-       if (c < 256 && isprint(c)) fprintf(outfile, "%c", c);
+       yield += pchar(c, f);
-         else fprintf(outfile, "\\x{%02x}", c);
+       continue;
-       continue;
        }
      }
+ #endif
+   c = *p++;
+   yield += pchar(c, f);
+   }
+ return yield;
+ }
+ #endif
+ #ifdef SUPPORT_PCRE16
+ /*************************************************
+ *           Print 16-bit character string        *
+ *************************************************/
+ /* Must handle UTF-16 strings in utf mode. Yields number of characters printed.
+ If handed a NULL file, just counts chars without printing. */
+ static int pchars16(PCRE_SPTR16 p, int length, FILE *f)
+ {
+ int yield = 0;
+ while (length-- > 0)
+   {
+   int c = *p++ & 0xffff;
+ #if !defined NOUTF8
+   if (use_utf && c >= 0xD800 && c < 0xDC00 && length > 0)
+     {
+     int d = *p & 0xffff;
+     if (d >= 0xDC00 && d < 0xDFFF)
+       {
+       c = ((c & 0x3ff) << 10) + (d & 0x3ff) + 0x10000;
+       length--;
+       p++;
+       }
+     }
+ #endif
+   yield += pchar(c, f);
+   }
+ return yield;
+ }
+ #endif
-    /* Not UTF-8, or malformed UTF-8  */
-   if (isprint(c = *(p++))) fprintf(outfile, "%c", c);
-     else fprintf(outfile, "\\x%02x", c);
+ /*************************************************
+ *              Callout function                  *
+ *************************************************/
+ /* Called from PCRE as a result of the (?C) item. We print out where we are in
+ the match. Yield zero unless more callouts than the fail count, or the callout
+ data is not zero. */
+ static int callout(pcre_callout_block *cb)
+ {
+ FILE *f = (first_callout | callout_extra)? outfile : NULL;
+ int i, pre_start, post_start, subject_length;
+ if (callout_extra)
+   {
+   fprintf(f, "Callout %d: last capture = %d\n",
+     cb->callout_number, cb->capture_last);
+   for (i = 0; i < cb->capture_top * 2; i += 2)
+     {
+     if (cb->offset_vector[i] < 0)
+       fprintf(f, "%2d: <unset>\n", i/2);
+     else
+       {
+       fprintf(f, "%2d: ", i/2);
+       PCHARSV(cb->subject + cb->offset_vector[i],
+         cb->offset_vector[i+1] - cb->offset_vector[i], f);
+       fprintf(f, "\n");
+       }
+     }
+   }
+ /* Re-print the subject in canonical form, the first time or if giving full
+ datails. On subsequent calls in the same match, we use pchars just to find the
+ printed lengths of the substrings. */
+ if (f != NULL) fprintf(f, "--->");
+ PCHARS(pre_start, cb->subject, cb->start_match, f);
+ PCHARS(post_start, cb->subject + cb->start_match,
+   cb->current_position - cb->start_match, f);
+ PCHARS(subject_length, cb->subject, cb->subject_length, NULL);
+ PCHARSV(cb->subject + cb->current_position,
+   cb->subject_length - cb->current_position, f);
+ if (f != NULL) fprintf(f, "\n");
+ /* Always print appropriate indicators, with callout number if not already
+ shown. For automatic callouts, show the pattern offset. */
+ if (cb->callout_number == 255)
+   {
+   fprintf(outfile, "%+3d ", cb->pattern_position);
+   if (cb->pattern_position > 99) fprintf(outfile, "\n    ");
+   }
+ else
+   {
+   if (callout_extra) fprintf(outfile, "    ");
+     else fprintf(outfile, "%3d ", cb->callout_number);
+   }
+ for (i = 0; i < pre_start; i++) fprintf(outfile, " ");
+ fprintf(outfile, "^");
+ if (post_start > 0)
+   {
+   for (i = 0; i < post_start - 1; i++) fprintf(outfile, " ");
+   fprintf(outfile, "^");
+   }
+ for (i = 0; i < subject_length - pre_start - post_start + 4; i++)
+   fprintf(outfile, " ");
+ fprintf(outfile, "%.*s", (cb->next_item_length == 0)? 1 : cb->next_item_length,
+   pbuffer + cb->pattern_position);
+ fprintf(outfile, "\n");
+ first_callout = 0;
+ if (cb->mark != last_callout_mark)
+   {
+   fprintf(outfile, "Latest Mark: %s\n",
+     (cb->mark == NULL)? "<unset>" : (char *)(cb->mark));
+   last_callout_mark = cb->mark;
+   }
+ if (cb->callout_data != NULL)
+   {
+   int callout_data = *((int *)(cb->callout_data));
+   if (callout_data != 0)
+     {
+     fprintf(outfile, "Callout data = %d\n", callout_data);
+     return callout_data;
+     }
+   }
+ return (cb->callout_number != callout_fail_id)? 0 :
+        (++callout_count >= callout_fail_count)? 1 : 0;
+ }
+ /*************************************************
+ *            Local malloc functions              *
+ *************************************************/
+ /* Alternative malloc function, to test functionality and save the size of a
+ compiled re, which is the first store request that pcre_compile() makes. The
+ show_malloc variable is set only during matching. */
+ static void *new_malloc(size_t size)
+ {
+ void *block = malloc(size);
+ gotten_store = size;
+ if (first_gotten_store == 0) first_gotten_store = size;
+ if (show_malloc)
+   fprintf(outfile, "malloc       %3d %p\n", (int)size, block);
+ return block;
+ }
+ static void new_free(void *block)
+ {
+ if (show_malloc)
+   fprintf(outfile, "free             %p\n", block);
+ free(block);
+ }
+ /* For recursion malloc/free, to test stacking calls */
+ static void *stack_malloc(size_t size)
+ {
+ void *block = malloc(size);
+ if (show_malloc)
+   fprintf(outfile, "stack_malloc %3d %p\n", (int)size, block);
+ return block;
+ }
+ static void stack_free(void *block)
+ {
+ if (show_malloc)
+   fprintf(outfile, "stack_free       %p\n", block);
+ free(block);
+ }
+ /*************************************************
+ *          Call pcre_fullinfo()                  *
+ *************************************************/
+ /* Get one piece of information from the pcre_fullinfo() function. When only
+ one of 8-bit or 16-bit is supported, use_pcre16 should always have the correct
+ value, but the code is defensive. */
+ static void new_info(pcre *re, pcre_extra *study, int option, void *ptr)
+ {
+ int rc;
+ if (use_pcre16)
+ #ifdef SUPPORT_PCRE16
+   rc = pcre16_fullinfo(re, study, option, ptr);
+ #else
+   rc = PCRE_ERROR_BADMODE;
+ #endif
+ else
+ #ifdef SUPPORT_PCRE8
+   rc = pcre_fullinfo(re, study, option, ptr);
+ #else
+   rc = PCRE_ERROR_BADMODE;
+ #endif
+ if (rc < 0) fprintf(outfile, "Error %d from pcre%s_fullinfo(%d)\n", rc,
+   use_pcre16? "16" : "", option);
+ }
+ /*************************************************
+ *         Byte flipping function                 *
+ *************************************************/
+ static unsigned long int
+ byteflip(unsigned long int value, int n)
+ {
+ if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
+ return ((value & 0x000000ff) << 24) |
+        ((value & 0x0000ff00) <<  8) |
+        ((value & 0x00ff0000) >>  8) |
+        ((value & 0xff000000) >> 24);
+ }
+ /*************************************************
+ *        Check match or recursion limit          *
+ *************************************************/
+ static int
+ check_match_limit(pcre *re, pcre_extra *extra, pcre_uint8 *bptr, int len,
+   int start_offset, int options, int *use_offsets, int use_size_offsets,
+   int flag, unsigned long int *limit, int errnumber, const char *msg)
+ {
+ int count;
+ int min = 0;
+ int mid = 64;
+ int max = -1;
+ extra->flags |= flag;
+ for (;;)
+   {
+   *limit = mid;
+   PCRE_EXEC(count, re, extra, bptr, len, start_offset, options,
+     use_offsets, use_size_offsets);
+   if (count == errnumber)
+     {
+     /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
+     min = mid;
+     mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
+     }
+   else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
+                          count == PCRE_ERROR_PARTIAL)
+     {
+     if (mid == min + 1)
+       {
+       fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
+       break;
+       }
+     /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
+     max = mid;
+     mid = (min + mid)/2;
+     }
+   else break;    /* Some other error */
+   }
+ extra->flags &= ~flag;
+ return count;
+ }
+ /*************************************************
+ *         Case-independent strncmp() function    *
+ *************************************************/
+ /*
+ Arguments:
+   s         first string
+   t         second string
+   n         number of characters to compare
+ Returns:    < 0, = 0, or > 0, according to the comparison
+ */
+ static int
+ strncmpic(pcre_uint8 *s, pcre_uint8 *t, int n)
+ {
+ while (n--)
+   {
+   int c = tolower(*s++) - tolower(*t++);
+   if (c) return c;
    }
+ return 0;
  }
- /* Alternative malloc function, to test functionality and show the size of the
+ /*************************************************
- compiled re. */
+ *         Check newline indicator                *
+ *************************************************/
+ /* This is used both at compile and run-time to check for <xxx> escapes. Print
+ a message and return 0 if there is no match.
+ Arguments:
+   p           points after the leading '<'
+   f           file for error message
- static void *new_malloc(size_t size)
+ Returns:      appropriate PCRE_NEWLINE_xxx flags, or 0
+ */
+ static int
+ check_newline(pcre_uint8 *p, FILE *f)
  {
- gotten_store = size;
+ if (strncmpic(p, (pcre_uint8 *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
- if (log_store)
+ if (strncmpic(p, (pcre_uint8 *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
-   fprintf(outfile, "Memory allocation (code space): %d\n",
+ if (strncmpic(p, (pcre_uint8 *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
-     (int)((int)size - offsetof(real_pcre, code[0])));
+ if (strncmpic(p, (pcre_uint8 *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
- return malloc(size);
+ if (strncmpic(p, (pcre_uint8 *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
+ if (strncmpic(p, (pcre_uint8 *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
+ if (strncmpic(p, (pcre_uint8 *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
+ fprintf(f, "Unknown newline type at: <%s\n", p);
+ return 0;
  }
+ /*************************************************
+ *             Usage function                     *
+ *************************************************/
- /* Get one piece of information from the pcre_fullinfo() function */
+ static void
+ usage(void)
- static void new_info(pcre *re, pcre_extra *study, int option, void *ptr)
  {
- int rc;
+ printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
- if ((rc = pcre_fullinfo(re, study, option, ptr)) < 0)
+ printf("Input and output default to stdin and stdout.\n");
-   fprintf(outfile, "Error %d from pcre_fullinfo(%d)\n", rc, option);
+ #ifdef SUPPORT_LIBREADLINE
+ printf("If input is a terminal, readline() is used to read from it.\n");
+ #else
+ printf("This version of pcretest is not linked with readline().\n");
+ #endif
+ printf("\nOptions:\n");
+ #ifdef SUPPORT_PCRE16
+ printf("  -16      use 16-bit interface\n");
+ #endif
+ printf("  -b       show compiled code (bytecode)\n");
+ printf("  -C       show PCRE compile-time options and exit\n");
+ printf("  -d       debug: show compiled code and information (-b and -i)\n");
+ #if !defined NODFA
+ printf("  -dfa     force DFA matching for all subjects\n");
+ #endif
+ printf("  -help    show usage information\n");
+ printf("  -i       show information about compiled patterns\n"
+        "  -M       find MATCH_LIMIT minimum for each subject\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("  -q       quiet: do not output PCRE version number at start\n");
+ printf("  -S <n>   set stack size to <n> megabytes\n");
+ printf("  -s       force each pattern to be studied at basic level\n"
+        "  -s+      force each pattern to be studied, using JIT if available\n"
+        "  -t       time compilation and execution\n");
+ printf("  -t <n>   time compilation and execution, repeating <n> times\n");
+ printf("  -tm      time execution (matching) only\n");
+ printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
  }
+ /*************************************************
+ *                Main Program                    *
+ *************************************************/
  /* Read lines from named file or stdin and write to named file or stdout; lines
  consist of a regular expression, in delimiters and optionally followed by
-Line 429 
 int main(int argc, char **argv)
+Line 1509 
 int main(int argc, char **argv)
  FILE *infile = stdin;
  int options = 0;
  int study_options = 0;
+ int default_find_match_limit = FALSE;
  int op = 1;
  int timeit = 0;
+ int timeitm = 0;
  int showinfo = 0;
  int showstore = 0;
+ int force_study = -1;
+ int force_study_options = 0;
+ int quiet = 0;
  int size_offsets = 45;
  int size_offsets_max;
- int *offsets;
+ int *offsets = NULL;
  #if !defined NOPOSIX
  int posix = 0;
  #endif
  int debug = 0;
  int done = 0;
- unsigned char buffer[30000];
+ int all_use_dfa = 0;
- unsigned char dbuffer[1024];
+ int yield = 0;
+ int stack_size;
+ pcre_jit_stack *jit_stack = NULL;
+ /* These vectors store, end-to-end, a list of captured substring names. Assume
+ that 1024 is plenty long enough for the few names we'll be testing. */
+ pcre_uchar copynames[1024];
+ pcre_uchar getnames[1024];
+ pcre_uchar *copynamesptr;
+ pcre_uchar *getnamesptr;
+ /* Get buffers from malloc() so that valgrind will check their misuse when
+ debugging. They grow automatically when very long lines are read. The 16-bit
+ buffer (buffer16) is obtained only if needed. */
+ buffer = (pcre_uint8 *)malloc(buffer_size);
+ dbuffer = (pcre_uint8 *)malloc(buffer_size);
+ pbuffer = (pcre_uint8 *)malloc(buffer_size);
- /* Static so that new_malloc can use it. */
+ /* The outfile variable is static so that new_malloc can use it. */
  outfile = stdout;
+ /* The following  _setmode() stuff is some Windows magic that tells its runtime
+ library to translate CRLF into a single LF character. At least, that's what
+ I've been told: never having used Windows I take this all on trust. Originally
+ it set 0x8000, but then I was advised that _O_BINARY was better. */
+ #if defined(_WIN32) || defined(WIN32)
+ _setmode( _fileno( stdout ), _O_BINARY );
+ #endif
  /* Scan options */
  while (argc > 1 && argv[op][0] == '-')
    {
-   char *endptr;
+   pcre_uint8 *endptr;
+   if (strcmp(argv[op], "-m") == 0) showstore = 1;
+   else if (strcmp(argv[op], "-s") == 0) force_study = 0;
+   else if (strcmp(argv[op], "-s+") == 0)
+     {
+     force_study = 1;
+     force_study_options = PCRE_STUDY_JIT_COMPILE;
+     }
+ #ifdef SUPPORT_PCRE16
+   else if (strcmp(argv[op], "-16") == 0) use_pcre16 = 1;
+ #endif
-   if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
+   else if (strcmp(argv[op], "-q") == 0) quiet = 1;
-     showstore = 1;
+   else if (strcmp(argv[op], "-b") == 0) debug = 1;
-   else if (strcmp(argv[op], "-t") == 0) timeit = 1;
    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
+   else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
+ #if !defined NODFA
+   else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
+ #endif
    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
-       ((size_offsets = strtoul(argv[op+1], &endptr, 10)), *endptr == 0))
+       ((size_offsets = get_value((pcre_uint8 *)argv[op+1], &endptr)),
+         *endptr == 0))
+     {
+     op++;
+     argc--;
+     }
+   else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
+     {
+     int both = argv[op][2] == 0;
+     int temp;
+     if (argc > 2 && (temp = get_value((pcre_uint8 *)argv[op+1], &endptr),
+                      *endptr == 0))
+       {
+       timeitm = temp;
+       op++;
+       argc--;
+       }
+     else timeitm = LOOPREPEAT;
+     if (both) timeit = timeitm;
+     }
+   else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
+       ((stack_size = get_value((pcre_uint8 *)argv[op+1], &endptr)),
+         *endptr == 0))
      {
+ #if defined(_WIN32) || defined(WIN32) || defined(__minix)
+     printf("PCRE: -S not supported on this OS\n");
+     exit(1);
+ #else
+     int rc;
+     struct rlimit rlim;
+     getrlimit(RLIMIT_STACK, &rlim);
+     rlim.rlim_cur = stack_size * 1024 * 1024;
+     rc = setrlimit(RLIMIT_STACK, &rlim);
+     if (rc != 0)
+       {
+     printf("PCRE: setrlimit() failed with error %d\n", rc);
+     exit(1);
+       }
      op++;
      argc--;
+ #endif
      }
  #if !defined NOPOSIX
    else if (strcmp(argv[op], "-p") == 0) posix = 1;
  #endif
+   else if (strcmp(argv[op], "-C") == 0)
+     {
+     int rc;
+     unsigned long int lrc;
+     printf("PCRE version %s\n", pcre_version());
+     printf("Compiled with\n");
+ /* At least one of SUPPORT_PCRE8 and SUPPORT_PCRE16 will be set. If both
+ are set, either both UTFs are supported or both are not supported. */
+ #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
+     printf("  8-bit and 16-bit support\n");
+     (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
+     if (rc)
+       printf("  UTF-8 and UTF-16 support\n");
+     else
+       printf("  No UTF-8 or UTF-16 support\n");
+ #elif defined SUPPORT_PCRE8
+     printf("  8-bit support only\n");
+     (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
+     printf("  %sUTF-8 support\n", rc? "" : "No ");
+ #else
+     printf("  16-bit support only\n");
+     (void)pcre16_config(PCRE_CONFIG_UTF16, &rc);
+     printf("  %sUTF-16 support\n", rc? "" : "No ");
+ #endif
+     (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
+     printf("  %sUnicode properties support\n", rc? "" : "No ");
+     (void)pcre_config(PCRE_CONFIG_JIT, &rc);
+     if (rc)
+       printf("  Just-in-time compiler support\n");
+     else
+       printf("  No just-in-time compiler support\n");
+     (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
+     /* Note that these values are always the ASCII values, even
+     in EBCDIC environments. CR is 13 and NL is 10. */
+     printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
+       (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
+       (rc == -2)? "ANYCRLF" :
+       (rc == -1)? "ANY" : "???");
+     (void)pcre_config(PCRE_CONFIG_BSR, &rc);
+     printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
+                                      "all Unicode newlines");
+     (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
+     printf("  Internal link size = %d\n", rc);
+     (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
+     printf("  POSIX malloc threshold = %d\n", rc);
+     (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
+     printf("  Default match limit = %ld\n", lrc);
+     (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
+     printf("  Default recursion depth limit = %ld\n", lrc);
+     (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
+     printf("  Match recursion uses %s\n", rc? "stack" : "heap");
+     goto EXIT;
+     }
+   else if (strcmp(argv[op], "-help") == 0 ||
+            strcmp(argv[op], "--help") == 0)
+     {
+     usage();
+     goto EXIT;
+     }
    else
      {
      printf("** Unknown or malformed option %s\n", argv[op]);
-     printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");
+     usage();
-     printf("  -d     debug: show compiled code; implies -i\n"
+     yield = 1;
-            "  -i     show information about compiled pattern\n"
+     goto EXIT;
-            "  -o <n> set size of offsets vector to <n>\n");
- #if !defined NOPOSIX
-     printf("  -p     use POSIX interface\n");
- #endif
-     printf("  -s     output store information\n"
-            "  -t     time compilation and execution\n");
-     return 1;
      }
    op++;
    argc--;
-Line 489 
 while (argc > 1 && argv[op][0] == '-')
+Line 1709 
 while (argc > 1 && argv[op][0] == '-')
  /* Get the store for the offsets vector, and remember what it was */
  size_offsets_max = size_offsets;
- offsets = malloc(size_offsets_max * sizeof(int));
+ offsets = (int *)malloc(size_offsets_max * sizeof(int));
  if (offsets == NULL)
    {
    printf("** Failed to get %d bytes of memory for offsets vector\n",
-     size_offsets_max * sizeof(int));
+     (int)(size_offsets_max * sizeof(int)));
-   return 1;
+   yield = 1;
+   goto EXIT;
    }
  /* Sort out the input and output files */
  if (argc > 1)
    {
-   infile = fopen(argv[op], "r");
+   infile = fopen(argv[op], INPUT_MODE);
    if (infile == NULL)
      {
      printf("** Failed to open %s\n", argv[op]);
-     return 1;
+     yield = 1;
+     goto EXIT;
      }
    }
  if (argc > 2)
    {
-   outfile = fopen(argv[op+1], "w");
+   outfile = fopen(argv[op+1], OUTPUT_MODE);
    if (outfile == NULL)
      {
      printf("** Failed to open %s\n", argv[op+1]);
-     return 1;
+     yield = 1;
+     goto EXIT;
      }
    }
  /* Set alternative malloc function */
+ #ifdef SUPPORT_PCRE8
  pcre_malloc = new_malloc;
+ pcre_free = new_free;
+ pcre_stack_malloc = stack_malloc;
+ pcre_stack_free = stack_free;
+ #endif
+ #ifdef SUPPORT_PCRE16
+ pcre16_malloc = new_malloc;
+ pcre16_free = new_free;
+ pcre16_stack_malloc = stack_malloc;
+ pcre16_stack_free = stack_free;
+ #endif
- /* Heading line, then prompt for first regex if stdin */
+ /* Heading line unless quiet, then prompt for first regex if stdin */
- fprintf(outfile, "PCRE version %s\n\n", pcre_version());
+ if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
  /* Main loop */
-Line 540 
 while (!done)
+Line 1775 
 while (!done)
  #endif
    const char *error;
-   unsigned char *p, *pp, *ppp;
+   pcre_uint8 *markptr;
-   const unsigned char *tables = NULL;
+   pcre_uint8 *p, *pp, *ppp;
+   pcre_uint8 *to_file = NULL;
+   const pcre_uint8 *tables = NULL;
+   unsigned long int true_size, true_study_size = 0;
+   size_t size, regex_gotten_store;
+   int do_allcaps = 0;
+   int do_mark = 0;
    int do_study = 0;
+   int no_force_study = 0;
    int do_debug = debug;
    int do_G = 0;
    int do_g = 0;
    int do_showinfo = showinfo;
    int do_showrest = 0;
-   int utf8 = 0;
+   int do_showcaprest = 0;
-   int erroroffset, len, delimiter;
+   int do_flip = 0;
+   int erroroffset, len, delimiter, poffset;
+   use_utf = 0;
+   debug_lengths = 1;
-   if (infile == stdin) printf("  re> ");
+   if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
-   if (fgets((char *)buffer, sizeof(buffer), infile) == NULL) break;
    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
+   fflush(outfile);
    p = buffer;
    while (isspace(*p)) p++;
    if (*p == 0) continue;
-   /* Get the delimiter and seek the end of the pattern; if is isn't
+   /* See if the pattern is to be loaded pre-compiled from a file. */
-   complete, read more. */
+   if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
+     {
+     unsigned long int magic, get_options;
+     pcre_uint8 sbuf[8];
+     FILE *f;
+     p++;
+     pp = p + (int)strlen((char *)p);
+     while (isspace(pp[-1])) pp--;
+     *pp = 0;
+     f = fopen((char *)p, "rb");
+     if (f == NULL)
+       {
+       fprintf(outfile, "Failed to open %s: %s\n", p, strerror(errno));
+       continue;
+       }
+     if (fread(sbuf, 1, 8, f) != 8) goto FAIL_READ;
+     true_size =
+       (sbuf[0] << 24) | (sbuf[1] << 16) | (sbuf[2] << 8) | sbuf[3];
+     true_study_size =
+       (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7];
+     re = (real_pcre *)new_malloc(true_size);
+     regex_gotten_store = first_gotten_store;
+     if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ;
+     magic = ((real_pcre *)re)->magic_number;
+     if (magic != MAGIC_NUMBER)
+       {
+       if (byteflip(magic, sizeof(magic)) == MAGIC_NUMBER)
+         {
+         do_flip = 1;
+         }
+       else
+         {
+         fprintf(outfile, "Data in %s is not a compiled PCRE regex\n", p);
+         fclose(f);
+         continue;
+         }
+       }
+     fprintf(outfile, "Compiled pattern%s loaded from %s\n",
+       do_flip? " (byte-inverted)" : "", p);
+     /* Need to know if UTF-8 for printing data strings */
+     new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
+     use_utf = (get_options & PCRE_UTF8) != 0;
+     /* Now see if there is any following study data. */
+     if (true_study_size != 0)
+       {
+       pcre_study_data *psd;
+       extra = (pcre_extra *)new_malloc(sizeof(pcre_extra) + true_study_size);
+       extra->flags = PCRE_EXTRA_STUDY_DATA;
+       psd = (pcre_study_data *)(((char *)extra) + sizeof(pcre_extra));
+       extra->study_data = psd;
+       if (fread(psd, 1, true_study_size, f) != true_study_size)
+         {
+         FAIL_READ:
+         fprintf(outfile, "Failed to read data from %s\n", p);
+         if (extra != NULL)
+           {
+           PCRE_FREE_STUDY(extra);
+           }
+         if (re != NULL) new_free(re);
+         fclose(f);
+         continue;
+         }
+       fprintf(outfile, "Study data loaded from %s\n", p);
+       do_study = 1;     /* To get the data output if requested */
+       }
+     else fprintf(outfile, "No study data\n");
+     fclose(f);
+     goto SHOW_INFO;
+     }
+   /* In-line pattern (the usual case). Get the delimiter and seek the end of
+   the pattern; if is isn't complete, read more. */
    delimiter = *p++;
    if (isalnum(delimiter) || delimiter == '\\')
      {
-     fprintf(outfile, "** Delimiter must not be alphameric or \\\n");
+     fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
      goto SKIP_DATA;
      }
    pp = p;
+   poffset = (int)(p - buffer);
    for(;;)
      {
-Line 581 
 while (!done)
+Line 1916 
 while (!done)
        pp++;
        }
      if (*pp != 0) break;
+     if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
-     len = sizeof(buffer) - (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)
        {
        fprintf(outfile, "** Unexpected EOF\n");
        done = 1;
-Line 599 
 while (!done)
+Line 1925 
 while (!done)
      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
      }
+   /* The buffer may have moved while being extended; reset the start of data
+   pointer to the correct relative point in the buffer. */
+   p = buffer + poffset;
    /* If the first character after the delimiter is backslash, make
    the pattern end with backslash. This is purely to provide a way
    of testing for the error message when a pattern ends with backslash. */
    if (pp[1] == '\\') *pp++ = '\\';
-   /* Terminate the pattern at the delimiter */
+   /* Terminate the pattern at the delimiter, and save a copy of the pattern
+   for callouts. */
    *pp++ = 0;
+   strcpy((char *)pbuffer, (char *)p);
    /* Look for options after final delimiter */
-Line 619 
 while (!done)
+Line 1952 
 while (!done)
      {
      switch (*pp++)
        {
+       case 'f': options |= PCRE_FIRSTLINE; break;
        case 'g': do_g = 1; break;
        case 'i': options |= PCRE_CASELESS; break;
        case 'm': options |= PCRE_MULTILINE; break;
        case 's': options |= PCRE_DOTALL; break;
        case 'x': options |= PCRE_EXTENDED; break;
-       case '+': do_showrest = 1; break;
+       case '+':
+       if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
+       break;
+       case '=': do_allcaps = 1; break;
        case 'A': options |= PCRE_ANCHORED; break;
+       case 'B': do_debug = 1; break;
+       case 'C': options |= PCRE_AUTO_CALLOUT; break;
        case 'D': do_debug = do_showinfo = 1; break;
        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
+       case 'F': do_flip = 1; break;
        case 'G': do_G = 1; break;
        case 'I': do_showinfo = 1; break;
+       case 'J': options |= PCRE_DUPNAMES; break;
+       case 'K': do_mark = 1; break;
        case 'M': log_store = 1; break;
+       case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
  #if !defined NOPOSIX
        case 'P': do_posix = 1; break;
  #endif
-       case 'S': do_study = 1; break;
+       case 'S':
+       if (do_study == 0)
+         {
+         do_study = 1;
+         if (*pp == '+')
+           {
+           study_options |= PCRE_STUDY_JIT_COMPILE;
+           pp++;
+           }
+         }
+       else
+         {
+         do_study = 0;
+         no_force_study = 1;
+         }
+       break;
        case 'U': options |= PCRE_UNGREEDY; break;
+       case 'W': options |= PCRE_UCP; break;
        case 'X': options |= PCRE_EXTRA; break;
-       case '8': options |= PCRE_UTF8; utf8 = 1; break;
+       case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
+       case 'Z': debug_lengths = 0; break;
+       case '8': options |= PCRE_UTF8; use_utf = 1; break;
+       case '?': options |= PCRE_NO_UTF8_CHECK; break;
+       case 'T':
+       switch (*pp++)
+         {
+         case '0': tables = tables0; break;
+         case '1': tables = tables1; break;
+         case '\r':
+         case '\n':
+         case ' ':
+         case 0:
+         fprintf(outfile, "** Missing table number after /T\n");
+         goto SKIP_DATA;
+         default:
+         fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
+         goto SKIP_DATA;
+         }
+       break;
        case 'L':
        ppp = pp;
-       while (*ppp != '\n' && *ppp != ' ') ppp++;
+       /* The '\r' test here is so that it works on Windows. */
+       /* The '0' test is just in case this is an unterminated line. */
+       while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
        *ppp = 0;
        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
          {
          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
          goto SKIP_DATA;
          }
+       locale_set = 1;
        tables = pcre_maketables();
        pp = ppp;
        break;
-       case '\n': case ' ': break;
+       case '>':
+       to_file = pp;
+       while (*pp != 0) pp++;
+       while (isspace(pp[-1])) pp--;
+       *pp = 0;
+       break;
+       case '<':
+         {
+         if (strncmpic(pp, (pcre_uint8 *)"JS>", 3) == 0)
+           {
+           options |= PCRE_JAVASCRIPT_COMPAT;
+           pp += 3;
+           }
+         else
+           {
+           int x = check_newline(pp, outfile);
+           if (x == 0) goto SKIP_DATA;
+           options |= x;
+           while (*pp++ != '>');
+           }
+         }
+       break;
+       case '\r':                      /* So that it works in Windows */
+       case '\n':
+       case ' ':
+       break;
        default:
        fprintf(outfile, "** Unknown option '%c'\n", pp[-1]);
        goto SKIP_DATA;
-Line 664 
 while (!done)
+Line 2078 
 while (!done)
    /* Handle compiling via the POSIX interface, which doesn't support the
    timing, showing, or debugging options, nor the ability to pass over
-   local character tables. */
+   local character tables. Neither does it have 16-bit support. */
  #if !defined NOPOSIX
    if (posix || do_posix)
      {
      int rc;
      int cflags = 0;
      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
+     if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
+     if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
+     if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
+     if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
+     if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
+     first_gotten_store = 0;
      rc = regcomp(&preg, (char *)p, cflags);
      /* Compilation failed; go back for another re, skipping to blank line
-Line 680 
 while (!done)
+Line 2102 
 while (!done)
      if (rc != 0)
        {
-       (void)regerror(rc, &preg, (char *)buffer, sizeof(buffer));
+       (void)regerror(rc, &preg, (char *)buffer, buffer_size);
        fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
        goto SKIP_DATA;
        }
-Line 692 
 while (!done)
+Line 2114 
 while (!done)
  #endif  /* !defined NOPOSIX */
      {
-     if (timeit)
+     unsigned long int get_options;
+     /* In 16-bit mode, convert the input. */
+ #ifdef SUPPORT_PCRE16
+     if (use_pcre16)
+       {
+       if (to16(p, options & PCRE_UTF8, (int)strlen((char *)p)) < 0)
+         {
+         fprintf(outfile, "**Failed: invalid UTF-8 string cannot be "
+           "converted to UTF-16\n");
+         goto SKIP_DATA;
+         }
+       p = (pcre_uint8 *)buffer16;
+       }
+ #endif
+     /* Compile many times when timing */
+     if (timeit > 0)
        {
        register int i;
        clock_t time_taken;
        clock_t start_time = clock();
-       for (i = 0; i < LOOPREPEAT; i++)
+       for (i = 0; i < timeit; i++)
          {
-         re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
+         PCRE_COMPILE(re, p, options, &error, &erroroffset, tables);
          if (re != NULL) free(re);
          }
        time_taken = clock() - start_time;
-       fprintf(outfile, "Compile time %.3f milliseconds\n",
+       fprintf(outfile, "Compile time %.4f milliseconds\n",
-         ((double)time_taken * 1000.0) /
+         (((double)time_taken * 1000.0) / (double)timeit) /
-         ((double)LOOPREPEAT * (double)CLOCKS_PER_SEC));
+           (double)CLOCKS_PER_SEC);
        }
-     re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
+     first_gotten_store = 0;
+     PCRE_COMPILE(re, p, options, &error, &erroroffset, tables);
      /* Compilation failed; go back for another re, skipping to blank line
      if non-interactive. */
-Line 721 
 while (!done)
+Line 2163 
 while (!done)
          {
          for (;;)
            {
-           if (fgets((char *)buffer, sizeof(buffer), infile) == NULL)
+           if (extend_inputline(infile, buffer, NULL) == NULL)
              {
              done = 1;
              goto CONTINUE;
-Line 735 
 while (!done)
+Line 2177 
 while (!done)
        goto CONTINUE;
        }
-     /* Compilation succeeded; print data if required. There are now two
+     /* Compilation succeeded. It is now possible to set the UTF-8 option from
-     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
-     returns only limited data. Check that it agrees with the newer one. */
+     lines. */
+     new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
+     if ((get_options & PCRE_UTF8) != 0) use_utf = 1;
+     /* Extract the size for possible writing before possibly flipping it,
+     and remember the store that was got. */
+     true_size = ((real_pcre *)re)->size;
+     regex_gotten_store = first_gotten_store;
+     /* Output code size information if requested */
+     if (log_store)
+       fprintf(outfile, "Memory allocation (code space): %d\n",
+         (int)(first_gotten_store -
+               sizeof(real_pcre) -
+               ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size));
+     /* If -s or /S was present, study the regex to generate additional info to
+     help with the matching, unless the pattern has the SS option, which
+     suppresses the effect of /S (used for a few test patterns where studying is
+     never sensible). */
+     if (do_study || (force_study >= 0 && !no_force_study))
+       {
+       if (timeit > 0)
+         {
+         register int i;
+         clock_t time_taken;
+         clock_t start_time = clock();
+         for (i = 0; i < timeit; i++)
+           {
+           PCRE_STUDY(extra, re, study_options | force_study_options, &error);
+           }
+         time_taken = clock() - start_time;
+         if (extra != NULL)
+           {
+           PCRE_FREE_STUDY(extra);
+           }
+         fprintf(outfile, "  Study time %.4f milliseconds\n",
+           (((double)time_taken * 1000.0) / (double)timeit) /
+             (double)CLOCKS_PER_SEC);
+         }
+       PCRE_STUDY(extra, re, study_options | force_study_options, &error);
+       if (error != NULL)
+         fprintf(outfile, "Failed to study: %s\n", error);
+       else if (extra != NULL)
+         {
+         true_study_size = ((pcre_study_data *)(extra->study_data))->size;
+         if (log_store)
+           {
+           size_t jitsize;
+           new_info(re, extra, PCRE_INFO_JITSIZE, &jitsize);
+           if (jitsize != 0)
+             fprintf(outfile, "Memory allocation (JIT code): %d\n", jitsize);
+           }
+         }
+       }
+     /* If /K was present, we set up for handling MARK data. */
+     if (do_mark)
+       {
+       if (extra == NULL)
+         {
+         extra = (pcre_extra *)malloc(sizeof(pcre_extra));
+         extra->flags = 0;
+         }
+       extra->mark = &markptr;
+       extra->flags |= PCRE_EXTRA_MARK;
+       }
+     /* If the 'F' option was present, we flip the bytes of all the integer
+     fields in the regex data block and the study block. This is to make it
+     possible to test PCRE's handling of byte-flipped patterns, e.g. those
+     compiled on a different architecture. */
+     if (do_flip)
+       {
+       real_pcre *rre = (real_pcre *)re;
+       rre->magic_number =
+         byteflip(rre->magic_number, sizeof(rre->magic_number));
+       rre->size = byteflip(rre->size, sizeof(rre->size));
+       rre->options = byteflip(rre->options, sizeof(rre->options));
+       rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
+       rre->top_bracket =
+         (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
+       rre->top_backref =
+         (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
+       rre->first_char =
+         (pcre_uint16)byteflip(rre->first_char, sizeof(rre->first_char));
+       rre->req_char =
+         (pcre_uint16)byteflip(rre->req_char, sizeof(rre->req_char));
+       rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
+         sizeof(rre->name_table_offset));
+       rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
+         sizeof(rre->name_entry_size));
+       rre->name_count = (pcre_uint16)byteflip(rre->name_count,
+         sizeof(rre->name_count));
+       if (extra != NULL)
+         {
+         pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
+         rsd->size = byteflip(rsd->size, sizeof(rsd->size));
+         rsd->flags = byteflip(rsd->flags, sizeof(rsd->flags));
+         rsd->minlength = byteflip(rsd->minlength, sizeof(rsd->minlength));
+         }
+       }
+     /* Extract and display information from the compiled data if required. */
+     SHOW_INFO:
+     if (do_debug)
+       {
+       fprintf(outfile, "------------------------------------------------------------------\n");
+ #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
+       if (use_pcre16)
+         pcre16_printint(re, outfile, debug_lengths);
+       else
+         pcre_printint(re, outfile, debug_lengths);
+ #elif defined SUPPORT_PCRE8
+       pcre_printint(re, outfile, debug_lengths);
+ #else
+       pcre16_printint(re, outfile, debug_lengths);
+ #endif
+       }
+     /* We already have the options in get_options (see above) */
      if (do_showinfo)
        {
-       unsigned long int get_options;
+       unsigned long int all_options;
+ #if !defined NOINFOCHECK
        int old_first_char, old_options, old_count;
-       int count, backrefmax, first_char, need_char;
+ #endif
-       size_t size;
+       int count, backrefmax, first_char, need_char, okpartial, jchanged,
+         hascrorlf;
-       if (do_debug) print_internals(re);
+       int nameentrysize, namecount;
+       const pcre_uchar *nametable;
-       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
        new_info(re, NULL, PCRE_INFO_SIZE, &size);
        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
-       new_info(re, NULL, PCRE_INFO_FIRSTCHAR, &first_char);
+       new_info(re, NULL, PCRE_INFO_FIRSTBYTE, &first_char);
        new_info(re, NULL, PCRE_INFO_LASTLITERAL, &need_char);
+       new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
-       old_count = pcre_info(re, &old_options, &old_first_char);
+       new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
-       if (count < 0) fprintf(outfile,
+       new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
-         "Error %d from pcre_info()\n", count);
+       new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
-       else
+       new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
-         {
+       new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
-         if (old_count != count) fprintf(outfile,
-           "Count disagreement: pcre_fullinfo=%d pcre_info=%d\n", count,
+       /* The old, obsolete function pcre_info() works only in 8-bit mode. Check
-             old_count);
+       that it gives the same results as the new function. */
-         if (old_first_char != first_char) fprintf(outfile,
+ #if !defined NOINFOCHECK
-           "First char disagreement: pcre_fullinfo=%d pcre_info=%d\n",
+       if (!use_pcre16)
-             first_char, old_first_char);
+         {
+         old_count = pcre_info(re, &old_options, &old_first_char);
-         if (old_options != (int)get_options) fprintf(outfile,
+         if (count < 0) fprintf(outfile,
-           "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
+           "Error %d from pcre_info()\n", count);
-             get_options, old_options);
+         else
+           {
+           if (old_count != count) fprintf(outfile,
+             "Count disagreement: pcre_fullinfo=%d pcre_info=%d\n", count,
+               old_count);
+           if (old_first_char != first_char) fprintf(outfile,
+             "First char disagreement: pcre_fullinfo=%d pcre_info=%d\n",
+               first_char, old_first_char);
+           if (old_options != (int)get_options) fprintf(outfile,
+             "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
+               get_options, old_options);
+           }
          }
+ #endif
-       if (size != gotten_store) fprintf(outfile,
+       if (size != regex_gotten_store) fprintf(outfile,
          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
-         size, gotten_store);
+         (int)size, (int)regex_gotten_store);
        fprintf(outfile, "Capturing subpattern count = %d\n", count);
        if (backrefmax > 0)
          fprintf(outfile, "Max back reference = %d\n", backrefmax);
+       if (namecount > 0)
+         {
+         fprintf(outfile, "Named capturing subpatterns:\n");
+         while (namecount-- > 0)
+           {
+           fprintf(outfile, "  %s %*s%3d\n", nametable + 2,
+             nameentrysize - 3 - (int)strlen((char *)nametable + 2), "",
+             GET2(nametable, 0));
+           nametable += nameentrysize;
+           }
+         }
+       if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
+       if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
+       all_options = ((real_pcre *)re)->options;
+       if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
        if (get_options == 0) fprintf(outfile, "No options\n");
-         else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s\n",
+         else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
+           ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
+           ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
+           ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
-           ((get_options & PCRE_UTF8) != 0)? " utf8" : "");
+           ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
+           ((get_options & PCRE_UTF8) != 0)? " utf" : "",
+           ((get_options & PCRE_UCP) != 0)? " ucp" : "",
+           ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf_check" : "",
+           ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
+           ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
+       if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
+       switch (get_options & PCRE_NEWLINE_BITS)
+         {
+         case PCRE_NEWLINE_CR:
+         fprintf(outfile, "Forced newline sequence: CR\n");
+         break;
+         case PCRE_NEWLINE_LF:
+         fprintf(outfile, "Forced newline sequence: LF\n");
+         break;
+         case PCRE_NEWLINE_CRLF:
+         fprintf(outfile, "Forced newline sequence: CRLF\n");
+         break;
+         case PCRE_NEWLINE_ANYCRLF:
+         fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
+         break;
+         case PCRE_NEWLINE_ANY:
+         fprintf(outfile, "Forced newline sequence: ANY\n");
+         break;
-       if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)
+         default:
-         fprintf(outfile, "Case state changes\n");
+         break;
+         }
        if (first_char == -1)
          {
-         fprintf(outfile, "First char at start or follows \\n\n");
+         fprintf(outfile, "First char at start or follows newline\n");
          }
        else if (first_char < 0)
          {
-Line 805 
 while (!done)
+Line 2443 
 while (!done)
          }
        else
          {
-         if (isprint(first_char))
+         const char *caseless =
-           fprintf(outfile, "First char = \'%c\'\n", first_char);
+           ((((real_pcre *)re)->flags & PCRE_FCH_CASELESS) == 0)?
+           "" : " (caseless)";
+         if (PRINTOK(first_char))
+           fprintf(outfile, "First char = \'%c\'%s\n", first_char, caseless);
+         else
+           {
+           fprintf(outfile, "First char = ");
+           pchar(first_char, outfile);
+           fprintf(outfile, "%s\n", caseless);
+           }
+         }
+       if (need_char < 0)
+         {
+         fprintf(outfile, "No need char\n");
+         }
+       else
+         {
+         const char *caseless =
+           ((((real_pcre *)re)->flags & PCRE_RCH_CASELESS) == 0)?
+           "" : " (caseless)";
+         if (PRINTOK(need_char))
+           fprintf(outfile, "Need char = \'%c\'%s\n", need_char, caseless);
          else
-           fprintf(outfile, "First char = %d\n", first_char);
+           fprintf(outfile, "Need char = %d%s\n", need_char, caseless);
          }
-       if (need_char < 0)
+       /* Don't output study size; at present it is in any case a fixed
-         {
+       value, but it varies, depending on the computer architecture, and
-         fprintf(outfile, "No need char\n");
+       so messes up the test suite. (And with the /F option, it might be
-         }
+       flipped.) If study was forced by an external -s, don't show this
-       else
+       information unless -i or -d was also present. This means that, except
-         {
+       when auto-callouts are involved, the output from runs with and without
-         if (isprint(need_char))
+       -s should be identical. */
-           fprintf(outfile, "Need char = \'%c\'\n", need_char);
-         else
+       if (do_study || (force_study >= 0 && showinfo && !no_force_study))
-           fprintf(outfile, "Need char = %d\n", need_char);
+         {
+         if (extra == NULL)
+           fprintf(outfile, "Study returned NULL\n");
+         else
+           {
+           pcre_uint8 *start_bits = NULL;
+           int minlength;
+           new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength);
+           fprintf(outfile, "Subject length lower bound = %d\n", minlength);
+           new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
+           if (start_bits == NULL)
+             fprintf(outfile, "No set of starting bytes\n");
+           else
+             {
+             int i;
+             int c = 24;
+             fprintf(outfile, "Starting byte set: ");
+             for (i = 0; i < 256; i++)
+               {
+               if ((start_bits[i/8] & (1<<(i&7))) != 0)
+                 {
+                 if (c > 75)
+                   {
+                   fprintf(outfile, "\n  ");
+                   c = 2;
+                   }
+                 if (PRINTOK(i) && i != ' ')
+                   {
+                   fprintf(outfile, "%c ", i);
+                   c += 2;
+                   }
+                 else
+                   {
+                   fprintf(outfile, "\\x%02x ", i);
+                   c += 5;
+                   }
+                 }
+               }
+             fprintf(outfile, "\n");
+             }
+           }
+         /* Show this only if the JIT was set by /S, not by -s. */
+         if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
+           {
+           int jit;
+           new_info(re, extra, PCRE_INFO_JIT, &jit);
+           if (jit)
+             fprintf(outfile, "JIT study was successful\n");
+           else
+ #ifdef SUPPORT_JIT
+             fprintf(outfile, "JIT study was not successful\n");
+ #else
+             fprintf(outfile, "JIT support is not available in this version of PCRE\n");
+ #endif
+           }
          }
        }
-     /* If /S was present, study the regexp to generate additional info to
+     /* If the '>' option was present, we write out the regex to a file, and
-     help with the matching. */
+     that is all. The first 8 bytes of the file are the regex length and then
+     the study length, in big-endian order. */
-     if (do_study)
+     if (to_file != NULL)
        {
-       if (timeit)
+       FILE *f = fopen((char *)to_file, "wb");
+       if (f == NULL)
          {
-         register int i;
+         fprintf(outfile, "Unable to open %s: %s\n", to_file, strerror(errno));
-         clock_t time_taken;
-         clock_t start_time = clock();
-         for (i = 0; i < LOOPREPEAT; i++)
-           extra = pcre_study(re, study_options, &error);
-         time_taken = clock() - start_time;
-         if (extra != NULL) free(extra);
-         fprintf(outfile, "  Study time %.3f milliseconds\n",
-           ((double)time_taken * 1000.0)/
-           ((double)LOOPREPEAT * (double)CLOCKS_PER_SEC));
          }
+       else
-       extra = pcre_study(re, study_options, &error);
-       if (error != NULL)
-         fprintf(outfile, "Failed to study: %s\n", error);
-       else if (extra == NULL)
-         fprintf(outfile, "Study returned NULL\n");
-       else if (do_showinfo)
          {
-         uschar *start_bits = NULL;
+         pcre_uint8 sbuf[8];
-         new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);
+         sbuf[0] = (pcre_uint8)((true_size >> 24) & 255);
-         if (start_bits == NULL)
+         sbuf[1] = (pcre_uint8)((true_size >> 16) & 255);
-           fprintf(outfile, "No starting character set\n");
+         sbuf[2] = (pcre_uint8)((true_size >>  8) & 255);
+         sbuf[3] = (pcre_uint8)((true_size) & 255);
+         sbuf[4] = (pcre_uint8)((true_study_size >> 24) & 255);
+         sbuf[5] = (pcre_uint8)((true_study_size >> 16) & 255);
+         sbuf[6] = (pcre_uint8)((true_study_size >>  8) & 255);
+         sbuf[7] = (pcre_uint8)((true_study_size) & 255);
+         if (fwrite(sbuf, 1, 8, f) < 8 ||
+             fwrite(re, 1, true_size, f) < true_size)
+           {
+           fprintf(outfile, "Write error on %s: %s\n", to_file, strerror(errno));
+           }
          else
            {
-           int i;
+           fprintf(outfile, "Compiled pattern written to %s\n", to_file);
-           int c = 24;
-           fprintf(outfile, "Starting character set: ");
+           /* If there is study data, write it. */
-           for (i = 0; i < 256; i++)
+           if (extra != NULL)
              {
-             if ((start_bits[i/8] & (1<<(i%8))) != 0)
+             if (fwrite(extra->study_data, 1, true_study_size, f) <
+                 true_study_size)
                {
-               if (c > 75)
+               fprintf(outfile, "Write error on %s: %s\n", to_file,
-                 {
+                 strerror(errno));
-                 fprintf(outfile, "\n  ");
-                 c = 2;
-                 }
-               if (isprint(i) && i != ' ')
-                 {
-                 fprintf(outfile, "%c ", i);
-                 c += 2;
-                 }
-               else
-                 {
-                 fprintf(outfile, "\\x%02x ", i);
-                 c += 5;
-                 }
                }
+             else fprintf(outfile, "Study data written to %s\n", to_file);
              }
-           fprintf(outfile, "\n");
            }
+         fclose(f);
+         }
+       new_free(re);
+       if (extra != NULL)
+         {
+         PCRE_FREE_STUDY(extra);
+         }
+       if (locale_set)
+         {
+         new_free((void *)tables);
+         setlocale(LC_CTYPE, "C");
+         locale_set = 0;
          }
+       continue;  /* With next regex */
        }
-     }
+     }        /* End of non-POSIX compile */
    /* Read data lines and test them */
    for (;;)
      {
-     unsigned char *q;
+     pcre_uint8 *q;
-     unsigned char *bptr = dbuffer;
+     pcre_uint8 *bptr;
+     int *use_offsets = offsets;
      int use_size_offsets = size_offsets;
+     int callout_data = 0;
+     int callout_data_set = 0;
      int count, c;
      int copystrings = 0;
+     int find_match_limit = default_find_match_limit;
      int getstrings = 0;
      int getlist = 0;
      int gmatched = 0;
      int start_offset = 0;
+     int start_offset_sign = 1;
      int g_notempty = 0;
+     int use_dfa = 0;
      options = 0;
-     if (infile == stdin) printf("data> ");
+     *copynames = 0;
-     if (fgets((char *)buffer, sizeof(buffer), infile) == NULL)
+     *getnames = 0;
+     copynamesptr = copynames;
+     getnamesptr = getnames;
+     pcre_callout = callout;
+     first_callout = 1;
+     last_callout_mark = NULL;
+     callout_extra = 0;
+     callout_count = 0;
+     callout_fail_count = 999999;
+     callout_fail_id = -1;
+     show_malloc = 0;
+     if (extra != NULL) extra->flags &=
+       ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
+     len = 0;
+     for (;;)
        {
-       done = 1;
+       if (extend_inputline(infile, buffer + len, "data> ") == NULL)
-       goto CONTINUE;
+         {
+         if (len > 0)    /* Reached EOF without hitting a newline */
+           {
+           fprintf(outfile, "\n");
+           break;
+           }
+         done = 1;
+         goto CONTINUE;
+         }
+       if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
+       len = (int)strlen((char *)buffer);
+       if (buffer[len-1] == '\n') break;
        }
-     if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
-     len = (int)strlen((char *)buffer);
      while (len > 0 && isspace(buffer[len-1])) len--;
      buffer[len] = 0;
      if (len == 0) break;
-Line 920 
 while (!done)
+Line 2674 
 while (!done)
      p = buffer;
      while (isspace(*p)) p++;
-     q = dbuffer;
+     bptr = q = dbuffer;
      while ((c = *p++) != 0)
        {
        int i = 0;
        int n = 0;
        if (c == '\\') switch ((c = *p++))
          {
          case 'a': c =    7; break;
-Line 941 
 while (!done)
+Line 2696 
 while (!done)
          c -= '0';
          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
            c = c * 8 + *p++ - '0';
+ #if !defined NOUTF8
+         if (use_utf && c > 255)
+           {
+           pcre_uint8 buff8[8];
+           int ii, utn;
+           utn = ord2utf8(c, buff8);
+           for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
+           c = buff8[ii];   /* Last byte */
+           }
+ #endif
          break;
          case 'x':
          /* Handle \x{..} specially - new Perl thing for utf8 */
+ #if !defined NOUTF8
          if (*p == '{')
            {
-           unsigned char *pt = p;
+           pcre_uint8 *pt = p;
            c = 0;
-           while (isxdigit(*(++pt)))
-             c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W');
+           /* We used to have "while (isxdigit(*(++pt)))" here, but it fails
+           when isxdigit() is a macro that refers to its argument more than
+           once. This is banned by the C Standard, but apparently happens in at
+           least one MacOS environment. */
+           for (pt++; isxdigit(*pt); pt++)
+             c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'a' - 10);
            if (*pt == '}')
              {
-             unsigned char buffer[8];
+             pcre_uint8 buff8[8];
              int ii, utn;
-             utn = ord2utf8(c, buffer);
+             if (use_utf)
-             for (ii = 0; ii < utn - 1; ii++) *q++ = buffer[ii];
+               {
-             c = buffer[ii];   /* Last byte */
+               utn = ord2utf8(c, buff8);
+               for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
+               c = buff8[ii];   /* Last byte */
+               }
+             else
+              {
+              if (c > 255)
+                fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
+                  "UTF-8 mode is not enabled.\n"
+                  "** Truncation will probably give the wrong result.\n", c);
+              }
              p = pt + 1;
              break;
              }
            /* Not correct form; fall through */
            }
+ #endif
          /* Ordinary \x */
          c = 0;
          while (i++ < 2 && isxdigit(*p))
            {
-           c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W');
+           c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'a' - 10);
            p++;
            }
          break;
-         case 0:   /* Allows for an empty line */
+         case 0:   /* \ followed by EOF allows for an empty line */
          p--;
          continue;
+         case '>':
+         if (*p == '-')
+           {
+           start_offset_sign = -1;
+           p++;
+           }
+         while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
+         start_offset *= start_offset_sign;
+         continue;
          case 'A':  /* Option setting */
          options |= PCRE_ANCHORED;
          continue;
-Line 989 
 while (!done)
+Line 2783 
 while (!done)
          continue;
          case 'C':
-         while(isdigit(*p)) n = n * 10 + *p++ - '0';
+         if (isdigit(*p))    /* Set copy string */
-         copystrings |= 1 << n;
+           {
+           while(isdigit(*p)) n = n * 10 + *p++ - '0';
+           copystrings |= 1 << n;
+           }
+         else if (isalnum(*p))
+           {
+           pcre_uchar *npp = copynamesptr;
+           while (isalnum(*p)) *npp++ = *p++;
+           *npp++ = 0;
+           *npp = 0;
+           n = pcre_get_stringnumber(re, (char *)copynamesptr);
+           if (n < 0)
+             fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
+           copynamesptr = npp;
+           }
+         else if (*p == '+')
+           {
+           callout_extra = 1;
+           p++;
+           }
+         else if (*p == '-')
+           {
+           pcre_callout = NULL;
+           p++;
+           }
+         else if (*p == '!')
+           {
+           callout_fail_id = 0;
+           p++;
+           while(isdigit(*p))
+             callout_fail_id = callout_fail_id * 10 + *p++ - '0';
+           callout_fail_count = 0;
+           if (*p == '!')
+             {
+             p++;
+             while(isdigit(*p))
+               callout_fail_count = callout_fail_count * 10 + *p++ - '0';
+             }
+           }
+         else if (*p == '*')
+           {
+           int sign = 1;
+           callout_data = 0;
+           if (*(++p) == '-') { sign = -1; p++; }
+           while(isdigit(*p))
+             callout_data = callout_data * 10 + *p++ - '0';
+           callout_data *= sign;
+           callout_data_set = 1;
+           }
+         continue;
+ #if !defined NODFA
+         case 'D':
+ #if !defined NOPOSIX
+         if (posix || do_posix)
+           printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");
+         else
+ #endif
+           use_dfa = 1;
          continue;
+ #endif
+ #if !defined NODFA
+         case 'F':
+         options |= PCRE_DFA_SHORTEST;
+         continue;
+ #endif
          case 'G':
+         if (isdigit(*p))
+           {
+           while(isdigit(*p)) n = n * 10 + *p++ - '0';
+           getstrings |= 1 << n;
+           }
+         else if (isalnum(*p))
+           {
+           pcre_uchar *npp = getnamesptr;
+           while (isalnum(*p)) *npp++ = *p++;
+           *npp++ = 0;
+           *npp = 0;
+           n = pcre_get_stringnumber(re, (char *)getnamesptr);
+           if (n < 0)
+             fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
+           getnamesptr = npp;
+           }
+         continue;
+         case 'J':
          while(isdigit(*p)) n = n * 10 + *p++ - '0';
-         getstrings |= 1 << n;
+         if (extra != NULL
+             && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
+             && extra->executable_jit != NULL)
+           {
+           if (jit_stack != NULL) pcre_jit_stack_free(jit_stack);
+           jit_stack = pcre_jit_stack_alloc(1, n * 1024);
+           pcre_assign_jit_stack(extra, jit_callback, jit_stack);
+           }
          continue;
          case 'L':
          getlist = 1;
          continue;
+         case 'M':
+         find_match_limit = 1;
+         continue;
          case 'N':
-         options |= PCRE_NOTEMPTY;
+         if ((options & PCRE_NOTEMPTY) != 0)
+           options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
+         else
+           options |= PCRE_NOTEMPTY;
          continue;
          case 'O':
          while(isdigit(*p)) n = n * 10 + *p++ - '0';
          if (n > size_offsets_max)
            {
- if (offsets != NULL)
-           free(offsets);
            size_offsets_max = n;
-           offsets = malloc(size_offsets_max * sizeof(int));
+           free(offsets);
+           use_offsets = offsets = (int *)malloc(size_offsets_max * sizeof(int));
            if (offsets == NULL)
              {
              printf("** Failed to get %d bytes of memory for offsets vector\n",
-               size_offsets_max * sizeof(int));
+               (int)(size_offsets_max * sizeof(int)));
-             return 1;
+             yield = 1;
+             goto EXIT;
              }
            }
          use_size_offsets = n;
+         if (n == 0) use_offsets = NULL;   /* Ensures it can't write to it */
+         continue;
- if (n == 0)
+         case 'P':
-   {
+         options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
-   free(offsets);
+           PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
-   offsets = NULL;
+         continue;
-   size_offsets_max = 0;
-   }
+         case 'Q':
+         while(isdigit(*p)) n = n * 10 + *p++ - '0';
+         if (extra == NULL)
+           {
+           extra = (pcre_extra *)malloc(sizeof(pcre_extra));
+           extra->flags = 0;
+           }
+         extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
+         extra->match_limit_recursion = n;
+         continue;
+         case 'q':
+         while(isdigit(*p)) n = n * 10 + *p++ - '0';
+         if (extra == NULL)
+           {
+           extra = (pcre_extra *)malloc(sizeof(pcre_extra));
+           extra->flags = 0;
+           }
+         extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
+         extra->match_limit = n;
+         continue;
+ #if !defined NODFA
+         case 'R':
+         options |= PCRE_DFA_RESTART;
+         continue;
+ #endif
+         case 'S':
+         show_malloc = 1;
+         continue;
+         case 'Y':
+         options |= PCRE_NO_START_OPTIMIZE;
          continue;
          case 'Z':
          options |= PCRE_NOTEOL;
          continue;
+         case '?':
+         options |= PCRE_NO_UTF8_CHECK;
+         continue;
+         case '<':
+           {
+           int x = check_newline(p, outfile);
+           if (x == 0) goto NEXT_DATA;
+           options |= x;
+           while (*p++ != '>');
+           }
+         continue;
          }
        *q++ = c;
        }
      *q = 0;
-     len = q - dbuffer;
+     len = (int)(q - dbuffer);
+     /* Move the data to the end of the buffer so that a read over the end of
+     the buffer will be seen by valgrind, even if it doesn't cause a crash. If
+     we are using the POSIX interface, we must include the terminating zero. */
+ #if !defined NOPOSIX
+     if (posix || do_posix)
+       {
+       memmove(bptr + buffer_size - len - 1, bptr, len + 1);
+       bptr += buffer_size - len - 1;
+       }
+     else
+ #endif
+       {
+       memmove(bptr + buffer_size - len, bptr, len);
+       bptr += buffer_size - len;
+       }
+     if ((all_use_dfa || use_dfa) && find_match_limit)
+       {
+       printf("**Match limit not relevant for DFA matching: ignored\n");
+       find_match_limit = 0;
+       }
      /* Handle matching via the POSIX interface, which does not
-     support timing. */
+     support timing or playing with the match limit or callout data. */
  #if !defined NOPOSIX
      if (posix || do_posix)
        {
        int rc;
        int eflags = 0;
-       regmatch_t *pmatch = malloc(sizeof(regmatch_t) * use_size_offsets);
+       regmatch_t *pmatch = NULL;
+       if (use_size_offsets > 0)
+         pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
+       if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
        if (rc != 0)
          {
-         (void)regerror(rc, &preg, (char *)buffer, sizeof(buffer));
+         (void)regerror(rc, &preg, (char *)buffer, buffer_size);
          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
          }
+       else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
+               != 0)
+         {
+         fprintf(outfile, "Matched with REG_NOSUB\n");
+         }
        else
          {
          size_t i;
-         for (i = 0; i < use_size_offsets; i++)
+         for (i = 0; i < (size_t)use_size_offsets; i++)
            {
            if (pmatch[i].rm_so >= 0)
              {
              fprintf(outfile, "%2d: ", (int)i);
-             pchars(dbuffer + pmatch[i].rm_so,
+             PCHARSV(dbuffer + pmatch[i].rm_so,
-               pmatch[i].rm_eo - pmatch[i].rm_so, utf8);
+               pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
              fprintf(outfile, "\n");
-             if (i == 0 && do_showrest)
+             if (do_showcaprest || (i == 0 && do_showrest))
                {
-               fprintf(outfile, " 0+ ");
+               fprintf(outfile, "%2d+ ", (int)i);
-               pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo, utf8);
+               PCHARSV(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,
+                 outfile);
                fprintf(outfile, "\n");
                }
              }
            }
          }
        free(pmatch);
+       goto NEXT_DATA;
        }
+ #endif  /* !defined NOPOSIX */
      /* Handle matching via the native interface - repeats for /g and /G */
-     else
+ #ifdef SUPPORT_PCRE16
- #endif  /* !defined NOPOSIX */
+     if (use_pcre16)
+       {
+       len = to16(bptr, (((real_pcre *)re)->options) & PCRE_UTF8, len);
+       if (len < 0)
+         {
+         fprintf(outfile, "**Failed: invalid UTF-8 string cannot be "
+           "converted to UTF-16\n");
+         goto NEXT_DATA;
+         }
+       bptr = (pcre_uint8 *)buffer16;
+       }
+ #endif
      for (;; gmatched++)    /* Loop for /g or /G */
        {
-       if (timeit)
+       markptr = NULL;
+       if (timeitm > 0)
          {
          register int i;
          clock_t time_taken;
          clock_t start_time = clock();
-         for (i = 0; i < LOOPREPEAT; i++)
-           count = pcre_exec(re, extra, (char *)bptr, len,
+ #if !defined NODFA
-             start_offset, options | g_notempty, offsets, use_size_offsets);
+         if (all_use_dfa || use_dfa)
+           {
+           int workspace[1000];
+           for (i = 0; i < timeitm; i++)
+             count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
+               options | g_notempty, use_offsets, use_size_offsets, workspace,
+               sizeof(workspace)/sizeof(int));
+           }
+         else
+ #endif
+         for (i = 0; i < timeitm; i++)
+           {
+           PCRE_EXEC(count, re, extra, bptr, len,
+             start_offset, options | g_notempty, use_offsets, use_size_offsets);
+           }
          time_taken = clock() - start_time;
-         fprintf(outfile, "Execute time %.3f milliseconds\n",
+         fprintf(outfile, "Execute time %.4f milliseconds\n",
-           ((double)time_taken * 1000.0)/
+           (((double)time_taken * 1000.0) / (double)timeitm) /
-           ((double)LOOPREPEAT * (double)CLOCKS_PER_SEC));
+             (double)CLOCKS_PER_SEC);
+         }
+       /* If find_match_limit is set, we want to do repeated matches with
+       varying limits in order to find the minimum value for the match limit and
+       for the recursion limit. The match limits are relevant only to the normal
+       running of pcre_exec(), so disable the JIT optimization. This makes it
+       possible to run the same set of tests with and without JIT externally
+       requested. */
+       if (find_match_limit)
+         {
+         if (extra == NULL)
+           {
+           extra = (pcre_extra *)malloc(sizeof(pcre_extra));
+           extra->flags = 0;
+           }
+         else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
+         (void)check_match_limit(re, extra, bptr, len, start_offset,
+           options|g_notempty, use_offsets, use_size_offsets,
+           PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
+           PCRE_ERROR_MATCHLIMIT, "match()");
+         count = check_match_limit(re, extra, bptr, len, start_offset,
+           options|g_notempty, use_offsets, use_size_offsets,
+           PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
+           PCRE_ERROR_RECURSIONLIMIT, "match() recursion");
          }
-       count = pcre_exec(re, extra, (char *)bptr, len,
+       /* If callout_data is set, use the interface with additional data */
-         start_offset, options | g_notempty, offsets, use_size_offsets);
+       else if (callout_data_set)
+         {
+         if (extra == NULL)
+           {
+           extra = (pcre_extra *)malloc(sizeof(pcre_extra));
+           extra->flags = 0;
+           }
+         extra->flags |= PCRE_EXTRA_CALLOUT_DATA;
+         extra->callout_data = &callout_data;
+         PCRE_EXEC(count, re, extra, bptr, len, start_offset,
+           options | g_notempty, use_offsets, use_size_offsets);
+         extra->flags &= ~PCRE_EXTRA_CALLOUT_DATA;
+         }
+       /* The normal case is just to do the match once, with the default
+       value of match_limit. */
+ #if !defined NODFA
+       else if (all_use_dfa || use_dfa)
+         {
+         int workspace[1000];
+         count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset,
+           options | g_notempty, use_offsets, use_size_offsets, workspace,
+           sizeof(workspace)/sizeof(int));
+         if (count == 0)
+           {
+           fprintf(outfile, "Matched, but too many subsidiary matches\n");
+           count = use_size_offsets/2;
+           }
+         }
+ #endif
-       if (count == 0)
+       else
          {
-         fprintf(outfile, "Matched, but too many substrings\n");
+         PCRE_EXEC(count, re, extra, bptr, len, start_offset,
-         count = use_size_offsets/3;
+           options | g_notempty, use_offsets, use_size_offsets);
+         if (count == 0)
+           {
+           fprintf(outfile, "Matched, but too many substrings\n");
+           count = use_size_offsets/3;
+           }
          }
        /* Matched */
        if (count >= 0)
          {
-         int i;
+         int i, maxcount;
+ #if !defined NODFA
+         if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
+ #endif
+           maxcount = use_size_offsets/3;
+         /* This is a check against a lunatic return value. */
+         if (count > maxcount)
+           {
+           fprintf(outfile,
+             "** PCRE error: returned count %d is too big for offset size %d\n",
+             count, use_size_offsets);
+           count = use_size_offsets/3;
+           if (do_g || do_G)
+             {
+             fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
+             do_g = do_G = FALSE;        /* Break g/G loop */
+             }
+           }
+         /* do_allcaps requests showing of all captures in the pattern, to check
+         unset ones at the end. */
+         if (do_allcaps)
+           {
+           new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
+           count++;   /* Allow for full match */
+           if (count * 2 > use_size_offsets) count = use_size_offsets/2;
+           }
+         /* Output the captured substrings */
          for (i = 0; i < count * 2; i += 2)
            {
-           if (offsets[i] < 0)
+           if (use_offsets[i] < 0)
+             {
+             if (use_offsets[i] != -1)
+               fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
+                 use_offsets[i], i);
+             if (use_offsets[i+1] != -1)
+               fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
+                 use_offsets[i+1], i+1);
              fprintf(outfile, "%2d: <unset>\n", i/2);
+             }
            else
              {
              fprintf(outfile, "%2d: ", i/2);
-             pchars(bptr + offsets[i], offsets[i+1] - offsets[i], utf8);
+             PCHARSV(bptr + use_offsets[i],
+               use_offsets[i+1] - use_offsets[i], outfile);
              fprintf(outfile, "\n");
-             if (i == 0)
+             if (do_showcaprest || (i == 0 && do_showrest))
                {
-               if (do_showrest)
+               fprintf(outfile, "%2d+ ", i/2);
-                 {
+               PCHARSV(bptr + use_offsets[i+1], len - use_offsets[i+1],