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

Diff of /code/trunk/pcretest.c

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

revision 255 by ph10, Wed Sep 19 08:50:04 2007 UTC revision 836 by ph10, Wed Dec 28 17:16:11 2011 UTC
# Line 4  Line 4 
4    
5  /* This program was hacked up as a tester for PCRE. I really should have  /* This program was hacked up as a tester for PCRE. I really should have
6  written it more tidily in the first place. Will I ever learn? It has grown and  written it more tidily in the first place. Will I ever learn? It has grown and
7  been extended and consequently is now rather, er, *very* untidy in places.  been extended and consequently is now rather, er, *very* untidy in places. The
8    addition of 16-bit support has made it even worse. :-(
9    
10  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
11  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 35  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
37  */  */
38    
39    /* This program now supports the testing of both the 8-bit and 16-bit PCRE
40    libraries in a single program. This is different from the modules such as
41    pcre_compile.c in the library itself, which are compiled separately for each
42    mode. If both modes are enabled, for example, pcre_compile.c is compiled twice
43    (the second time with COMPILE_PCRE16 defined). By contrast, pcretest.c is
44    compiled only once. Therefore, it must not make use of any of the macros from
45    pcre_internal.h that depend on COMPILE_PCRE8 or COMPILE_PCRE16. It does,
46    however, make use of SUPPORT_PCRE8 and SUPPORT_PCRE16 to ensure that it calls
47    only supported library functions. */
48    
49    
50  #ifdef HAVE_CONFIG_H  #ifdef HAVE_CONFIG_H
51  #include "config.h"  #include "config.h"
# Line 48  POSSIBILITY OF SUCH DAMAGE. Line 59  POSSIBILITY OF SUCH DAMAGE.
59  #include <locale.h>  #include <locale.h>
60  #include <errno.h>  #include <errno.h>
61    
62    #ifdef SUPPORT_LIBREADLINE
63    #ifdef HAVE_UNISTD_H
64    #include <unistd.h>
65    #endif
66    #include <readline/readline.h>
67    #include <readline/history.h>
68    #endif
69    
70    
71  /* A number of things vary for Windows builds. Originally, pcretest opened its  /* A number of things vary for Windows builds. Originally, pcretest opened its
72  input and output without "b"; then I was told that "b" was needed in some  input and output without "b"; then I was told that "b" was needed in some
# Line 63  input mode under Windows. */ Line 82  input mode under Windows. */
82  #define INPUT_MODE   "r"  #define INPUT_MODE   "r"
83  #define OUTPUT_MODE  "wb"  #define OUTPUT_MODE  "wb"
84    
85    #ifndef isatty
86    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
87    #endif                         /* though in some environments they seem to   */
88                                   /* be already defined, hence the #ifndefs.    */
89    #ifndef fileno
90    #define fileno _fileno
91    #endif
92    
93    /* A user sent this fix for Borland Builder 5 under Windows. */
94    
95    #ifdef __BORLANDC__
96    #define _setmode(handle, mode) setmode(handle, mode)
97    #endif
98    
99    /* Not Windows */
100    
101  #else  #else
102  #include <sys/time.h>          /* These two includes are needed */  #include <sys/time.h>          /* These two includes are needed */
103  #include <sys/resource.h>      /* for setrlimit(). */  #include <sys/resource.h>      /* for setrlimit(). */
# Line 81  here before pcre_internal.h so that the Line 116  here before pcre_internal.h so that the
116  appropriately for an application, not for building PCRE. */  appropriately for an application, not for building PCRE. */
117    
118  #include "pcre.h"  #include "pcre.h"
119    
120    #if defined SUPPORT_PCRE16 && !defined SUPPORT_PCRE8
121    /* Configure internal macros to 16 bit mode. */
122    #define COMPILE_PCRE16
123    #endif
124    
125  #include "pcre_internal.h"  #include "pcre_internal.h"
126    
127  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* The pcre_printint() function, which prints the internal form of a compiled
128  two copies, we include the source file here, changing the names of the external  regex, is held in a separate file so that (a) it can be compiled in either
129  symbols to prevent clashes. */  8-bit or 16-bit mode, and (b) it can be #included directly in pcre_compile.c
130    when that is compiled in debug mode. */
 #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  
131    
132  #include "pcre_tables.c"  #ifdef SUPPORT_PCRE8
133    void pcre_printint(pcre *external_re, FILE *f, BOOL print_lengths);
134    #endif
135    #ifdef SUPPORT_PCRE16
136    void pcre16_printint(pcre *external_re, FILE *f, BOOL print_lengths);
137    #endif
138    
139    /* We need access to some of the data tables that PCRE uses. So as not to have
140    to keep two copies, we include the source file here, changing the names of the
141    external symbols to prevent clashes. */
142    
143    #define PCRE_INCLUDED
144    #undef PRIV
145    #define PRIV(name) name
146    
147  /* We also need the pcre_printint() function for printing out compiled  #include "pcre_tables.c"
 patterns. This function is in a separate file so that it can be included in  
 pcre_compile.c when that module is compiled with debugging enabled.  
148    
149  The definition of the macro PRINTABLE, which determines whether to print an  /* The definition of the macro PRINTABLE, which determines whether to print an
150  output character as-is or as a hex value when showing compiled patterns, is  output character as-is or as a hex value when showing compiled patterns, is
151  contained in this file. We uses it here also, in cases when the locale has not  the same as in the printint.src file. We uses it here in cases when the locale
152  been explicitly changed, so as to get consistent output from systems that  has not been explicitly changed, so as to get consistent output from systems
153  differ in their output from isprint() even in the "C" locale. */  that differ in their output from isprint() even in the "C" locale. */
154    
155  #include "pcre_printint.src"  #ifdef EBCDIC
156    #define PRINTABLE(c) ((c) >= 64 && (c) < 255)
157    #else
158    #define PRINTABLE(c) ((c) >= 32 && (c) < 127)
159    #endif
160    
161  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  #define PRINTOK(c) (locale_set? isprint(c) : PRINTABLE(c))
162    
163    /* Posix support is disabled in 16 bit only mode. */
164    #if defined SUPPORT_PCRE16 && !defined SUPPORT_PCRE8 && !defined NOPOSIX
165    #define NOPOSIX
166    #endif
167    
168  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
169  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 122  Makefile. */ Line 173  Makefile. */
173  #include "pcreposix.h"  #include "pcreposix.h"
174  #endif  #endif
175    
176  /* It is also possible, for the benefit of the version currently imported into  /* It is also possible, originally for the benefit of a version that was
177  Exim, to build pcretest without support for UTF8 (define NOUTF8), without the  imported into Exim, to build pcretest without support for UTF8 or UTF16 (define
178  interface to the DFA matcher (NODFA), and without the doublecheck of the old  NOUTF), without the interface to the DFA matcher (NODFA). In fact, we
179  "info" function (define NOINFOCHECK). In fact, we automatically cut out the  automatically cut out the UTF support if PCRE is built without it. */
180  UTF8 support if PCRE is built without it. */  
181    #ifndef SUPPORT_UTF
182  #ifndef SUPPORT_UTF8  #ifndef NOUTF
183  #ifndef NOUTF8  #define NOUTF
184  #define NOUTF8  #endif
185  #endif  #endif
186    
187    /* To make the code a bit tidier for 8-bit and 16-bit support, we define macros
188    for all the pcre[16]_xxx functions (except pcre16_fullinfo, which is called
189    only from one place and is handled differently). I couldn't dream up any way of
190    using a single macro to do this in a generic way, because of the many different
191    argument requirements. We know that at least one of SUPPORT_PCRE8 and
192    SUPPORT_PCRE16 must be set. First define macros for each individual mode; then
193    use these in the definitions of generic macros.
194    
195    **** Special note about the PCHARSxxx macros: the address of the string to be
196    printed is always given as two arguments: a base address followed by an offset.
197    The base address is cast to the correct data size for 8 or 16 bit data; the
198    offset is in units of this size. If the string were given as base+offset in one
199    argument, the casting might be incorrectly applied. */
200    
201    #ifdef SUPPORT_PCRE8
202    
203    #define PCHARS8(lv, p, offset, len, f) \
204      lv = pchars((pcre_uint8 *)(p) + offset, len, f)
205    
206    #define PCHARSV8(p, offset, len, f) \
207      (void)pchars((pcre_uint8 *)(p) + offset, len, f)
208    
209    #define READ_CAPTURE_NAME8(p, cn8, cn16, re) \
210      p = read_capture_name8(p, cn8, re)
211    
212    #define SET_PCRE_CALLOUT8(callout) \
213      pcre_callout = callout
214    
215    #define STRLEN8(p) ((int)strlen((char *)p))
216    
217    
218    #define PCRE_COMPILE8(re, pat, options, error, erroffset, tables) \
219      re = pcre_compile((char *)pat, options, error, erroffset, tables)
220    
221    #define PCRE_COPY_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
222        namesptr, cbuffer, size) \
223      rc = pcre_copy_named_substring(re, (char *)bptr, offsets, count, \
224        (char *)namesptr, cbuffer, size)
225    
226    #define PCRE_COPY_SUBSTRING8(rc, bptr, offsets, count, i, cbuffer, size) \
227      rc = pcre_copy_substring((char *)bptr, offsets, count, i, cbuffer, size)
228    
229    #define PCRE_DFA_EXEC8(count, re, extra, bptr, len, start_offset, options, \
230        offsets, size_offsets, workspace, size_workspace) \
231      count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset, options, \
232        offsets, size_offsets, workspace, size_workspace)
233    
234    #define PCRE_EXEC8(count, re, extra, bptr, len, start_offset, options, \
235        offsets, size_offsets) \
236      count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options, \
237        offsets, size_offsets)
238    
239    #define PCRE_FREE_STUDY8(extra) \
240      pcre_free_study(extra)
241    
242    #define PCRE_FREE_SUBSTRING8(substring) \
243      pcre_free_substring(substring)
244    
245    #define PCRE_FREE_SUBSTRING_LIST8(listptr) \
246      pcre_free_substring_list(listptr)
247    
248    #define PCRE_GET_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
249        getnamesptr, subsptr) \
250      rc = pcre_get_named_substring(re, (char *)bptr, offsets, count, \
251        (char *)getnamesptr, subsptr)
252    
253    #define PCRE_GET_STRINGNUMBER8(n, rc, ptr) \
254      n = pcre_get_stringnumber(re, (char *)ptr)
255    
256    #define PCRE_GET_SUBSTRING8(rc, bptr, offsets, count, i, subsptr) \
257      rc = pcre_get_substring((char *)bptr, offsets, count, i, subsptr)
258    
259    #define PCRE_GET_SUBSTRING_LIST8(rc, bptr, offsets, count, listptr) \
260      rc = pcre_get_substring_list((const char *)bptr, offsets, count, listptr)
261    
262    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER8(re, extra, tables) \
263      pcre_pattern_to_host_byte_order(re, extra, tables)
264    
265    #define PCRE_PRINTINT8(re, outfile, debug_lengths) \
266      pcre_printint(re, outfile, debug_lengths)
267    
268    #define PCRE_STUDY8(extra, re, options, error) \
269      extra = pcre_study(re, options, error)
270    
271    #endif /* SUPPORT_PCRE8 */
272    
273    /* -----------------------------------------------------------*/
274    
275    #ifdef SUPPORT_PCRE16
276    
277    #define PCHARS16(lv, p, offset, len, f) \
278      lv = pchars16((PCRE_SPTR16)(p) + offset, len, f)
279    
280    #define PCHARSV16(p, offset, len, f) \
281      (void)pchars16((PCRE_SPTR16)(p) + offset, len, f)
282    
283    #define READ_CAPTURE_NAME16(p, cn8, cn16, re) \
284      p = read_capture_name16(p, cn16, re)
285    
286    #define STRLEN16(p) ((int)strlen16((PCRE_SPTR16)p))
287    
288    #define SET_PCRE_CALLOUT16(callout) \
289      pcre16_callout = callout
290    
291    
292    #define PCRE_COMPILE16(re, pat, options, error, erroffset, tables) \
293      re = pcre16_compile((PCRE_SPTR16)pat, options, error, erroffset, tables)
294    
295    #define PCRE_COPY_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
296        namesptr, cbuffer, size) \
297      rc = pcre16_copy_named_substring(re, (PCRE_SPTR16)bptr, offsets, count, \
298        (PCRE_SPTR16)namesptr, (PCRE_SCHAR16 *)cbuffer, size/2)
299    
300    #define PCRE_COPY_SUBSTRING16(rc, bptr, offsets, count, i, cbuffer, size) \
301      rc = pcre16_copy_substring((PCRE_SPTR16)bptr, offsets, count, i, \
302        (PCRE_SCHAR16 *)cbuffer, size/2)
303    
304    #define PCRE_DFA_EXEC16(count, re, extra, bptr, len, start_offset, options, \
305        offsets, size_offsets, workspace, size_workspace) \
306      count = pcre16_dfa_exec(re, extra, (PCRE_SPTR16)bptr, len, start_offset, \
307        options, offsets, size_offsets, workspace, size_workspace)
308    
309    #define PCRE_EXEC16(count, re, extra, bptr, len, start_offset, options, \
310        offsets, size_offsets) \
311      count = pcre16_exec(re, extra, (PCRE_SPTR16)bptr, len, start_offset, \
312        options, offsets, size_offsets)
313    
314    #define PCRE_FREE_STUDY16(extra) \
315      pcre16_free_study(extra)
316    
317    #define PCRE_FREE_SUBSTRING16(substring) \
318      pcre16_free_substring((PCRE_SPTR16)substring)
319    
320    #define PCRE_FREE_SUBSTRING_LIST16(listptr) \
321      pcre16_free_substring_list((PCRE_SPTR16 *)listptr)
322    
323    #define PCRE_GET_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
324        getnamesptr, subsptr) \
325      rc = pcre16_get_named_substring(re, (PCRE_SPTR16)bptr, offsets, count, \
326        (PCRE_SPTR16)getnamesptr, (PCRE_SPTR16 *)(void*)subsptr)
327    
328    #define PCRE_GET_STRINGNUMBER16(n, rc, ptr) \
329      n = pcre16_get_stringnumber(re, (PCRE_SPTR16)ptr)
330    
331    #define PCRE_GET_SUBSTRING16(rc, bptr, offsets, count, i, subsptr) \
332      rc = pcre16_get_substring((PCRE_SPTR16)bptr, offsets, count, i, \
333        (PCRE_SPTR16 *)(void*)subsptr)
334    
335    #define PCRE_GET_SUBSTRING_LIST16(rc, bptr, offsets, count, listptr) \
336      rc = pcre16_get_substring_list((PCRE_SPTR16)bptr, offsets, count, \
337        (PCRE_SPTR16 **)(void*)listptr)
338    
339    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER16(re, extra, tables) \
340      pcre16_pattern_to_host_byte_order(re, extra, tables)
341    
342    #define PCRE_PRINTINT16(re, outfile, debug_lengths) \
343      pcre16_printint(re, outfile, debug_lengths)
344    
345    #define PCRE_STUDY16(extra, re, options, error) \
346      extra = pcre16_study(re, options, error)
347    
348    #endif /* SUPPORT_PCRE16 */
349    
350    
351    /* ----- Both modes are supported; a runtime test is needed, except for
352    pcre_config(), and the JIT stack functions, when it doesn't matter which
353    version is called. ----- */
354    
355    #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
356    
357    #define CHAR_SIZE (use_pcre16? 2:1)
358    
359    #define PCHARS(lv, p, offset, len, f) \
360      if (use_pcre16) \
361        PCHARS16(lv, p, offset, len, f); \
362      else \
363        PCHARS8(lv, p, offset, len, f)
364    
365    #define PCHARSV(p, offset, len, f) \
366      if (use_pcre16) \
367        PCHARSV16(p, offset, len, f); \
368      else \
369        PCHARSV8(p, offset, len, f)
370    
371    #define READ_CAPTURE_NAME(p, cn8, cn16, re) \
372      if (use_pcre16) \
373        READ_CAPTURE_NAME16(p, cn8, cn16, re); \
374      else \
375        READ_CAPTURE_NAME8(p, cn8, cn16, re)
376    
377    #define SET_PCRE_CALLOUT(callout) \
378      if (use_pcre16) \
379        SET_PCRE_CALLOUT16(callout); \
380      else \
381        SET_PCRE_CALLOUT8(callout)
382    
383    #define STRLEN(p) (use_pcre16? STRLEN16(p) : STRLEN8(p))
384    
385    #define PCRE_ASSIGN_JIT_STACK pcre_assign_jit_stack
386    
387    #define PCRE_COMPILE(re, pat, options, error, erroffset, tables) \
388      if (use_pcre16) \
389        PCRE_COMPILE16(re, pat, options, error, erroffset, tables); \
390      else \
391        PCRE_COMPILE8(re, pat, options, error, erroffset, tables)
392    
393    #define PCRE_CONFIG pcre_config
394    
395    #define PCRE_COPY_NAMED_SUBSTRING(rc, re, bptr, offsets, count, \
396        namesptr, cbuffer, size) \
397      if (use_pcre16) \
398        PCRE_COPY_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
399          namesptr, cbuffer, size); \
400      else \
401        PCRE_COPY_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
402          namesptr, cbuffer, size)
403    
404    #define PCRE_COPY_SUBSTRING(rc, bptr, offsets, count, i, cbuffer, size) \
405      if (use_pcre16) \
406        PCRE_COPY_SUBSTRING16(rc, bptr, offsets, count, i, cbuffer, size); \
407      else \
408        PCRE_COPY_SUBSTRING8(rc, bptr, offsets, count, i, cbuffer, size)
409    
410    #define PCRE_DFA_EXEC(count, re, extra, bptr, len, start_offset, options, \
411        offsets, size_offsets, workspace, size_workspace) \
412      if (use_pcre16) \
413        PCRE_DFA_EXEC16(count, re, extra, bptr, len, start_offset, options, \
414          offsets, size_offsets, workspace, size_workspace); \
415      else \
416        PCRE_DFA_EXEC8(count, re, extra, bptr, len, start_offset, options, \
417          offsets, size_offsets, workspace, size_workspace)
418    
419    #define PCRE_EXEC(count, re, extra, bptr, len, start_offset, options, \
420        offsets, size_offsets) \
421      if (use_pcre16) \
422        PCRE_EXEC16(count, re, extra, bptr, len, start_offset, options, \
423          offsets, size_offsets); \
424      else \
425        PCRE_EXEC8(count, re, extra, bptr, len, start_offset, options, \
426          offsets, size_offsets)
427    
428    #define PCRE_FREE_STUDY(extra) \
429      if (use_pcre16) \
430        PCRE_FREE_STUDY16(extra); \
431      else \
432        PCRE_FREE_STUDY8(extra)
433    
434    #define PCRE_FREE_SUBSTRING(substring) \
435      if (use_pcre16) \
436        PCRE_FREE_SUBSTRING16(substring); \
437      else \
438        PCRE_FREE_SUBSTRING8(substring)
439    
440    #define PCRE_FREE_SUBSTRING_LIST(listptr) \
441      if (use_pcre16) \
442        PCRE_FREE_SUBSTRING_LIST16(listptr); \
443      else \
444        PCRE_FREE_SUBSTRING_LIST8(listptr)
445    
446    #define PCRE_GET_NAMED_SUBSTRING(rc, re, bptr, offsets, count, \
447        getnamesptr, subsptr) \
448      if (use_pcre16) \
449        PCRE_GET_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
450          getnamesptr, subsptr); \
451      else \
452        PCRE_GET_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
453          getnamesptr, subsptr)
454    
455    #define PCRE_GET_STRINGNUMBER(n, rc, ptr) \
456      if (use_pcre16) \
457        PCRE_GET_STRINGNUMBER16(n, rc, ptr); \
458      else \
459        PCRE_GET_STRINGNUMBER8(n, rc, ptr)
460    
461    #define PCRE_GET_SUBSTRING(rc, bptr, use_offsets, count, i, subsptr) \
462      if (use_pcre16) \
463        PCRE_GET_SUBSTRING16(rc, bptr, use_offsets, count, i, subsptr); \
464      else \
465        PCRE_GET_SUBSTRING8(rc, bptr, use_offsets, count, i, subsptr)
466    
467    #define PCRE_GET_SUBSTRING_LIST(rc, bptr, offsets, count, listptr) \
468      if (use_pcre16) \
469        PCRE_GET_SUBSTRING_LIST16(rc, bptr, offsets, count, listptr); \
470      else \
471        PCRE_GET_SUBSTRING_LIST8(rc, bptr, offsets, count, listptr)
472    
473    #define PCRE_JIT_STACK_ALLOC pcre_jit_stack_alloc
474    #define PCRE_JIT_STACK_FREE pcre_jit_stack_free
475    
476    #define PCRE_MAKETABLES \
477      (use_pcre16? pcre16_maketables() : pcre_maketables())
478    
479    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER(re, extra, tables) \
480      if (use_pcre16) \
481        PCRE_PATTERN_TO_HOST_BYTE_ORDER16(re, extra, tables); \
482      else \
483        PCRE_PATTERN_TO_HOST_BYTE_ORDER8(re, extra, tables)
484    
485    #define PCRE_PRINTINT(re, outfile, debug_lengths) \
486      if (use_pcre16) \
487        PCRE_PRINTINT16(re, outfile, debug_lengths); \
488      else \
489        PCRE_PRINTINT8(re, outfile, debug_lengths)
490    
491    #define PCRE_STUDY(extra, re, options, error) \
492      if (use_pcre16) \
493        PCRE_STUDY16(extra, re, options, error); \
494      else \
495        PCRE_STUDY8(extra, re, options, error)
496    
497    /* ----- Only 8-bit mode is supported ----- */
498    
499    #elif defined SUPPORT_PCRE8
500    #define CHAR_SIZE                 1
501    #define PCHARS                    PCHARS8
502    #define PCHARSV                   PCHARSV8
503    #define READ_CAPTURE_NAME         READ_CAPTURE_NAME8
504    #define SET_PCRE_CALLOUT          SET_PCRE_CALLOUT8
505    #define STRLEN                    STRLEN8
506    #define PCRE_ASSIGN_JIT_STACK     pcre_assign_jit_stack
507    #define PCRE_COMPILE              PCRE_COMPILE8
508    #define PCRE_CONFIG               pcre_config
509    #define PCRE_COPY_NAMED_SUBSTRING PCRE_COPY_NAMED_SUBSTRING8
510    #define PCRE_COPY_SUBSTRING       PCRE_COPY_SUBSTRING8
511    #define PCRE_DFA_EXEC             PCRE_DFA_EXEC8
512    #define PCRE_EXEC                 PCRE_EXEC8
513    #define PCRE_FREE_STUDY           PCRE_FREE_STUDY8
514    #define PCRE_FREE_SUBSTRING       PCRE_FREE_SUBSTRING8
515    #define PCRE_FREE_SUBSTRING_LIST  PCRE_FREE_SUBSTRING_LIST8
516    #define PCRE_GET_NAMED_SUBSTRING  PCRE_GET_NAMED_SUBSTRING8
517    #define PCRE_GET_STRINGNUMBER     PCRE_GET_STRINGNUMBER8
518    #define PCRE_GET_SUBSTRING        PCRE_GET_SUBSTRING8
519    #define PCRE_GET_SUBSTRING_LIST   PCRE_GET_SUBSTRING_LIST8
520    #define PCRE_JIT_STACK_ALLOC      pcre_jit_stack_alloc
521    #define PCRE_JIT_STACK_FREE       pcre_jit_stack_free
522    #define PCRE_MAKETABLES           pcre_maketables()
523    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER PCRE_PATTERN_TO_HOST_BYTE_ORDER8
524    #define PCRE_PRINTINT             PCRE_PRINTINT8
525    #define PCRE_STUDY                PCRE_STUDY8
526    
527    /* ----- Only 16-bit mode is supported ----- */
528    
529    #else
530    #define CHAR_SIZE                 2
531    #define PCHARS                    PCHARS16
532    #define PCHARSV                   PCHARSV16
533    #define READ_CAPTURE_NAME         READ_CAPTURE_NAME16
534    #define SET_PCRE_CALLOUT          SET_PCRE_CALLOUT16
535    #define STRLEN                    STRLEN16
536    #define PCRE_ASSIGN_JIT_STACK     pcre16_assign_jit_stack
537    #define PCRE_COMPILE              PCRE_COMPILE16
538    #define PCRE_CONFIG               pcre16_config
539    #define PCRE_COPY_NAMED_SUBSTRING PCRE_COPY_NAMED_SUBSTRING16
540    #define PCRE_COPY_SUBSTRING       PCRE_COPY_SUBSTRING16
541    #define PCRE_DFA_EXEC             PCRE_DFA_EXEC16
542    #define PCRE_EXEC                 PCRE_EXEC16
543    #define PCRE_FREE_STUDY           PCRE_FREE_STUDY16
544    #define PCRE_FREE_SUBSTRING       PCRE_FREE_SUBSTRING16
545    #define PCRE_FREE_SUBSTRING_LIST  PCRE_FREE_SUBSTRING_LIST16
546    #define PCRE_GET_NAMED_SUBSTRING  PCRE_GET_NAMED_SUBSTRING16
547    #define PCRE_GET_STRINGNUMBER     PCRE_GET_STRINGNUMBER16
548    #define PCRE_GET_SUBSTRING        PCRE_GET_SUBSTRING16
549    #define PCRE_GET_SUBSTRING_LIST   PCRE_GET_SUBSTRING_LIST16
550    #define PCRE_JIT_STACK_ALLOC      pcre16_jit_stack_alloc
551    #define PCRE_JIT_STACK_FREE       pcre16_jit_stack_free
552    #define PCRE_MAKETABLES           pcre16_maketables()
553    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER PCRE_PATTERN_TO_HOST_BYTE_ORDER16
554    #define PCRE_PRINTINT             PCRE_PRINTINT16
555    #define PCRE_STUDY                PCRE_STUDY16
556  #endif  #endif
557    
558    /* ----- End of mode-specific function call macros ----- */
559    
560    
561  /* Other parameters */  /* Other parameters */
562    
# Line 161  static int debug_lengths; Line 584  static int debug_lengths;
584  static int first_callout;  static int first_callout;
585  static int locale_set = 0;  static int locale_set = 0;
586  static int show_malloc;  static int show_malloc;
587  static int use_utf8;  static int use_utf;
588  static size_t gotten_store;  static size_t gotten_store;
589    static size_t first_gotten_store = 0;
590    static const unsigned char *last_callout_mark = NULL;
591    
592  /* The buffers grow automatically if very long input lines are encountered. */  /* The buffers grow automatically if very long input lines are encountered. */
593    
594  static int buffer_size = 50000;  static int buffer_size = 50000;
595  static uschar *buffer = NULL;  static pcre_uint8 *buffer = NULL;
596  static uschar *dbuffer = NULL;  static pcre_uint8 *dbuffer = NULL;
597  static uschar *pbuffer = NULL;  static pcre_uint8 *pbuffer = NULL;
598    
599    /* Another buffer is needed translation to 16-bit character strings. It will
600    obtained and extended as required. */
601    
602    #ifdef SUPPORT_PCRE16
603    static int buffer16_size = 0;
604    static pcre_uint16 *buffer16 = NULL;
605    
606    #ifdef SUPPORT_PCRE8
607    
608    /* We need the table of operator lengths that is used for 16-bit compiling, in
609    order to swap bytes in a pattern for saving/reloading testing. Luckily, the
610    data is defined as a macro. However, we must ensure that LINK_SIZE is adjusted
611    appropriately for the 16-bit world. Just as a safety check, make sure that
612    COMPILE_PCRE16 is *not* set. */
613    
614  /*************************************************  #ifdef COMPILE_PCRE16
615  *        Read or extend an input line            *  #error COMPILE_PCRE16 must not be set when compiling pcretest.c
616  *************************************************/  #endif
   
 /* 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!  
617    
618  Arguments:  #if LINK_SIZE == 2
619    f            the file to read  #undef LINK_SIZE
620    start        where in buffer to start (this *must* be within buffer)  #define LINK_SIZE 1
621    #elif LINK_SIZE == 3 || LINK_SIZE == 4
622    #undef LINK_SIZE
623    #define LINK_SIZE 2
624    #else
625    #error LINK_SIZE must be either 2, 3, or 4
626    #endif
627    
628  Returns:       pointer to the start of new data  #endif /* SUPPORT_PCRE8 */
                could be a copy of start, or could be moved  
                NULL if no data read and EOF reached  
 */  
629    
630  static uschar *  static const pcre_uint16 OP_lengths16[] = { OP_LENGTHS };
631  extend_inputline(FILE *f, uschar *start)  #endif  /* SUPPORT_PCRE16 */
 {  
 uschar *here = start;  
632    
633  for (;;)  /* If we have 8-bit support, default use_pcre16 to false; if there is also
634    {  16-bit support, it can be changed by an option. If there is no 8-bit support,
635    int rlen = buffer_size - (here - buffer);  there must be 16-bit support, so default it to 1. */
636    
637    if (rlen > 1000)  #ifdef SUPPORT_PCRE8
638      {  static int use_pcre16 = 0;
639      int dlen;  #else
640      if (fgets((char *)here, rlen,  f) == NULL)  static int use_pcre16 = 1;
641        return (here == start)? NULL : start;  #endif
     dlen = (int)strlen((char *)here);  
     if (dlen > 0 && here[dlen - 1] == '\n') return start;  
     here += dlen;  
     }  
642    
643    else  /* Textual explanations for runtime error codes */
     {  
     int new_buffer_size = 2*buffer_size;  
     uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);  
     uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);  
     uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);  
644    
645      if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)  static const char *errtexts[] = {
646        {    NULL,  /* 0 is no error */
647        fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);    NULL,  /* NOMATCH is handled specially */
648        exit(1);    "NULL argument passed",
649        }    "bad option value",
650      "magic number missing",
651      "unknown opcode - pattern overwritten?",
652      "no more memory",
653      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
654      "match limit exceeded",
655      "callout error code",
656      NULL,  /* BADUTF8/16 is handled specially */
657      NULL,  /* BADUTF8/16 offset is handled specially */
658      NULL,  /* PARTIAL is handled specially */
659      "not used - internal error",
660      "internal error - pattern overwritten?",
661      "bad count value",
662      "item unsupported for DFA matching",
663      "backreference condition or recursion test not supported for DFA matching",
664      "match limit not supported for DFA matching",
665      "workspace size exceeded in DFA matching",
666      "too much recursion for DFA matching",
667      "recursion limit exceeded",
668      "not used - internal error",
669      "invalid combination of newline options",
670      "bad offset value",
671      NULL,  /* SHORTUTF8/16 is handled specially */
672      "nested recursion at the same subject position",
673      "JIT stack limit reached",
674      "pattern compiled in wrong mode: 8-bit/16-bit error"
675    };
676    
     memcpy(new_buffer, buffer, buffer_size);  
     memcpy(new_pbuffer, pbuffer, buffer_size);  
677    
678      buffer_size = new_buffer_size;  /*************************************************
679    *         Alternate character tables             *
680    *************************************************/
681    
682      start = new_buffer + (start - buffer);  /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
683      here = new_buffer + (here - buffer);  using the default tables of the library. However, the T option can be used to
684    select alternate sets of tables, for different kinds of testing. Note also that
685    the L (locale) option also adjusts the tables. */
686    
687    /* This is the set of tables distributed as default with PCRE. It recognizes
688    only ASCII characters. */
689    
690    static const pcre_uint8 tables0[] = {
691    
692    /* This table is a lower casing table. */
693    
694        0,  1,  2,  3,  4,  5,  6,  7,
695        8,  9, 10, 11, 12, 13, 14, 15,
696       16, 17, 18, 19, 20, 21, 22, 23,
697       24, 25, 26, 27, 28, 29, 30, 31,
698       32, 33, 34, 35, 36, 37, 38, 39,
699       40, 41, 42, 43, 44, 45, 46, 47,
700       48, 49, 50, 51, 52, 53, 54, 55,
701       56, 57, 58, 59, 60, 61, 62, 63,
702       64, 97, 98, 99,100,101,102,103,
703      104,105,106,107,108,109,110,111,
704      112,113,114,115,116,117,118,119,
705      120,121,122, 91, 92, 93, 94, 95,
706       96, 97, 98, 99,100,101,102,103,
707      104,105,106,107,108,109,110,111,
708      112,113,114,115,116,117,118,119,
709      120,121,122,123,124,125,126,127,
710      128,129,130,131,132,133,134,135,
711      136,137,138,139,140,141,142,143,
712      144,145,146,147,148,149,150,151,
713      152,153,154,155,156,157,158,159,
714      160,161,162,163,164,165,166,167,
715      168,169,170,171,172,173,174,175,
716      176,177,178,179,180,181,182,183,
717      184,185,186,187,188,189,190,191,
718      192,193,194,195,196,197,198,199,
719      200,201,202,203,204,205,206,207,
720      208,209,210,211,212,213,214,215,
721      216,217,218,219,220,221,222,223,
722      224,225,226,227,228,229,230,231,
723      232,233,234,235,236,237,238,239,
724      240,241,242,243,244,245,246,247,
725      248,249,250,251,252,253,254,255,
726    
727    /* This table is a case flipping table. */
728    
729        0,  1,  2,  3,  4,  5,  6,  7,
730        8,  9, 10, 11, 12, 13, 14, 15,
731       16, 17, 18, 19, 20, 21, 22, 23,
732       24, 25, 26, 27, 28, 29, 30, 31,
733       32, 33, 34, 35, 36, 37, 38, 39,
734       40, 41, 42, 43, 44, 45, 46, 47,
735       48, 49, 50, 51, 52, 53, 54, 55,
736       56, 57, 58, 59, 60, 61, 62, 63,
737       64, 97, 98, 99,100,101,102,103,
738      104,105,106,107,108,109,110,111,
739      112,113,114,115,116,117,118,119,
740      120,121,122, 91, 92, 93, 94, 95,
741       96, 65, 66, 67, 68, 69, 70, 71,
742       72, 73, 74, 75, 76, 77, 78, 79,
743       80, 81, 82, 83, 84, 85, 86, 87,
744       88, 89, 90,123,124,125,126,127,
745      128,129,130,131,132,133,134,135,
746      136,137,138,139,140,141,142,143,
747      144,145,146,147,148,149,150,151,
748      152,153,154,155,156,157,158,159,
749      160,161,162,163,164,165,166,167,
750      168,169,170,171,172,173,174,175,
751      176,177,178,179,180,181,182,183,
752      184,185,186,187,188,189,190,191,
753      192,193,194,195,196,197,198,199,
754      200,201,202,203,204,205,206,207,
755      208,209,210,211,212,213,214,215,
756      216,217,218,219,220,221,222,223,
757      224,225,226,227,228,229,230,231,
758      232,233,234,235,236,237,238,239,
759      240,241,242,243,244,245,246,247,
760      248,249,250,251,252,253,254,255,
761    
762    /* This table contains bit maps for various character classes. Each map is 32
763    bytes long and the bits run from the least significant end of each byte. The
764    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
765    graph, print, punct, and cntrl. Other classes are built from combinations. */
766    
767      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
768      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
769      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
770      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
771    
772      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
773      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
774      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
775      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
776    
777      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
778      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
779      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
780      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
781    
782      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
783      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
784      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
785      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
786    
787      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
788      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
789      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
790      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
791    
792      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
793      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
794      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
795      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
796    
797      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
798      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
799      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
800      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
801    
802      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
803      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
804      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
805      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
806    
807      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
808      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
809      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
810      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
811    
812      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
813      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
814      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
815      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
816    
817    /* This table identifies various classes of character by individual bits:
818      0x01   white space character
819      0x02   letter
820      0x04   decimal digit
821      0x08   hexadecimal digit
822      0x10   alphanumeric or '_'
823      0x80   regular expression metacharacter or binary zero
824    */
825    
826      free(buffer);    0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
827      free(dbuffer);    0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
828      free(pbuffer);    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
829      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
830      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
831      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
832      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
833      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
834      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
835      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
836      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
837      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
838      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
839      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
840      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
841      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
842      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
843      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
844      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
845      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
846      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
847      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
848      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
849      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
850      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
851      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
852      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
853      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
854      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
855      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
856      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
857      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
858    
859    /* This is a set of tables that came orginally from a Windows user. It seems to
860    be at least an approximation of ISO 8859. In particular, there are characters
861    greater than 128 that are marked as spaces, letters, etc. */
862    
863    static const pcre_uint8 tables1[] = {
864    0,1,2,3,4,5,6,7,
865    8,9,10,11,12,13,14,15,
866    16,17,18,19,20,21,22,23,
867    24,25,26,27,28,29,30,31,
868    32,33,34,35,36,37,38,39,
869    40,41,42,43,44,45,46,47,
870    48,49,50,51,52,53,54,55,
871    56,57,58,59,60,61,62,63,
872    64,97,98,99,100,101,102,103,
873    104,105,106,107,108,109,110,111,
874    112,113,114,115,116,117,118,119,
875    120,121,122,91,92,93,94,95,
876    96,97,98,99,100,101,102,103,
877    104,105,106,107,108,109,110,111,
878    112,113,114,115,116,117,118,119,
879    120,121,122,123,124,125,126,127,
880    128,129,130,131,132,133,134,135,
881    136,137,138,139,140,141,142,143,
882    144,145,146,147,148,149,150,151,
883    152,153,154,155,156,157,158,159,
884    160,161,162,163,164,165,166,167,
885    168,169,170,171,172,173,174,175,
886    176,177,178,179,180,181,182,183,
887    184,185,186,187,188,189,190,191,
888    224,225,226,227,228,229,230,231,
889    232,233,234,235,236,237,238,239,
890    240,241,242,243,244,245,246,215,
891    248,249,250,251,252,253,254,223,
892    224,225,226,227,228,229,230,231,
893    232,233,234,235,236,237,238,239,
894    240,241,242,243,244,245,246,247,
895    248,249,250,251,252,253,254,255,
896    0,1,2,3,4,5,6,7,
897    8,9,10,11,12,13,14,15,
898    16,17,18,19,20,21,22,23,
899    24,25,26,27,28,29,30,31,
900    32,33,34,35,36,37,38,39,
901    40,41,42,43,44,45,46,47,
902    48,49,50,51,52,53,54,55,
903    56,57,58,59,60,61,62,63,
904    64,97,98,99,100,101,102,103,
905    104,105,106,107,108,109,110,111,
906    112,113,114,115,116,117,118,119,
907    120,121,122,91,92,93,94,95,
908    96,65,66,67,68,69,70,71,
909    72,73,74,75,76,77,78,79,
910    80,81,82,83,84,85,86,87,
911    88,89,90,123,124,125,126,127,
912    128,129,130,131,132,133,134,135,
913    136,137,138,139,140,141,142,143,
914    144,145,146,147,148,149,150,151,
915    152,153,154,155,156,157,158,159,
916    160,161,162,163,164,165,166,167,
917    168,169,170,171,172,173,174,175,
918    176,177,178,179,180,181,182,183,
919    184,185,186,187,188,189,190,191,
920    224,225,226,227,228,229,230,231,
921    232,233,234,235,236,237,238,239,
922    240,241,242,243,244,245,246,215,
923    248,249,250,251,252,253,254,223,
924    192,193,194,195,196,197,198,199,
925    200,201,202,203,204,205,206,207,
926    208,209,210,211,212,213,214,247,
927    216,217,218,219,220,221,222,255,
928    0,62,0,0,1,0,0,0,
929    0,0,0,0,0,0,0,0,
930    32,0,0,0,1,0,0,0,
931    0,0,0,0,0,0,0,0,
932    0,0,0,0,0,0,255,3,
933    126,0,0,0,126,0,0,0,
934    0,0,0,0,0,0,0,0,
935    0,0,0,0,0,0,0,0,
936    0,0,0,0,0,0,255,3,
937    0,0,0,0,0,0,0,0,
938    0,0,0,0,0,0,12,2,
939    0,0,0,0,0,0,0,0,
940    0,0,0,0,0,0,0,0,
941    254,255,255,7,0,0,0,0,
942    0,0,0,0,0,0,0,0,
943    255,255,127,127,0,0,0,0,
944    0,0,0,0,0,0,0,0,
945    0,0,0,0,254,255,255,7,
946    0,0,0,0,0,4,32,4,
947    0,0,0,128,255,255,127,255,
948    0,0,0,0,0,0,255,3,
949    254,255,255,135,254,255,255,7,
950    0,0,0,0,0,4,44,6,
951    255,255,127,255,255,255,127,255,
952    0,0,0,0,254,255,255,255,
953    255,255,255,255,255,255,255,127,
954    0,0,0,0,254,255,255,255,
955    255,255,255,255,255,255,255,255,
956    0,2,0,0,255,255,255,255,
957    255,255,255,255,255,255,255,127,
958    0,0,0,0,255,255,255,255,
959    255,255,255,255,255,255,255,255,
960    0,0,0,0,254,255,0,252,
961    1,0,0,248,1,0,0,120,
962    0,0,0,0,254,255,255,255,
963    0,0,128,0,0,0,128,0,
964    255,255,255,255,0,0,0,0,
965    0,0,0,0,0,0,0,128,
966    255,255,255,255,0,0,0,0,
967    0,0,0,0,0,0,0,0,
968    128,0,0,0,0,0,0,0,
969    0,1,1,0,1,1,0,0,
970    0,0,0,0,0,0,0,0,
971    0,0,0,0,0,0,0,0,
972    1,0,0,0,128,0,0,0,
973    128,128,128,128,0,0,128,0,
974    28,28,28,28,28,28,28,28,
975    28,28,0,0,0,0,0,128,
976    0,26,26,26,26,26,26,18,
977    18,18,18,18,18,18,18,18,
978    18,18,18,18,18,18,18,18,
979    18,18,18,128,128,0,128,16,
980    0,26,26,26,26,26,26,18,
981    18,18,18,18,18,18,18,18,
982    18,18,18,18,18,18,18,18,
983    18,18,18,128,128,0,0,0,
984    0,0,0,0,0,1,0,0,
985    0,0,0,0,0,0,0,0,
986    0,0,0,0,0,0,0,0,
987    0,0,0,0,0,0,0,0,
988    1,0,0,0,0,0,0,0,
989    0,0,18,0,0,0,0,0,
990    0,0,20,20,0,18,0,0,
991    0,20,18,0,0,0,0,0,
992    18,18,18,18,18,18,18,18,
993    18,18,18,18,18,18,18,18,
994    18,18,18,18,18,18,18,0,
995    18,18,18,18,18,18,18,18,
996    18,18,18,18,18,18,18,18,
997    18,18,18,18,18,18,18,18,
998    18,18,18,18,18,18,18,0,
999    18,18,18,18,18,18,18,18
1000    };
1001    
     buffer = new_buffer;  
     dbuffer = new_dbuffer;  
     pbuffer = new_pbuffer;  
     }  
   }  
1002    
 return NULL;  /* Control never gets here */  
 }  
1003    
1004    
1005    #ifndef HAVE_STRERROR
1006    /*************************************************
1007    *     Provide strerror() for non-ANSI libraries  *
1008    *************************************************/
1009    
1010    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
1011    in their libraries, but can provide the same facility by this simple
1012    alternative function. */
1013    
1014    extern int   sys_nerr;
1015    extern char *sys_errlist[];
1016    
1017    char *
1018    strerror(int n)
1019    {
1020    if (n < 0 || n >= sys_nerr) return "unknown error number";
1021    return sys_errlist[n];
1022    }
1023    #endif /* HAVE_STRERROR */
1024    
1025    
1026  /*************************************************  /*************************************************
1027  *          Read number from string               *  *         JIT memory callback                    *
1028  *************************************************/  *************************************************/
1029    
1030  /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess  static pcre_jit_stack* jit_callback(void *arg)
 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(unsigned char *str, unsigned char **endptr)  
1031  {  {
1032  int result = 0;  return (pcre_jit_stack *)arg;
 while(*str != 0 && isspace(*str)) str++;  
 while (isdigit(*str)) result = result * 10 + (int)(*str++ - '0');  
 *endptr = str;  
 return(result);  
1033  }  }
1034    
1035    
1036    #if !defined NOUTF || defined SUPPORT_PCRE16
   
1037  /*************************************************  /*************************************************
1038  *            Convert UTF-8 string to value       *  *            Convert UTF-8 string to value       *
1039  *************************************************/  *************************************************/
# Line 297  Returns:      >  0 => the number of byte Line 1049  Returns:      >  0 => the number of byte
1049                -6 to 0 => malformed UTF-8 character at offset = (-return)                -6 to 0 => malformed UTF-8 character at offset = (-return)
1050  */  */
1051    
 #if !defined NOUTF8  
   
1052  static int  static int
1053  utf82ord(unsigned char *utf8bytes, int *vptr)  utf82ord(pcre_uint8 *utf8bytes, int *vptr)
1054  {  {
1055  int c = *utf8bytes++;  int c = *utf8bytes++;
1056  int d = c;  int d = c;
# Line 339  if (j != i) return -(i+1); Line 1089  if (j != i) return -(i+1);
1089  *vptr = d;  *vptr = d;
1090  return i+1;  return i+1;
1091  }  }
1092    #endif /* NOUTF || SUPPORT_PCRE16 */
 #endif  
1093    
1094    
1095    
1096    #if !defined NOUTF || defined SUPPORT_PCRE16
1097  /*************************************************  /*************************************************
1098  *       Convert character value to UTF-8         *  *       Convert character value to UTF-8         *
1099  *************************************************/  *************************************************/
# Line 358  Arguments: Line 1108  Arguments:
1108  Returns:     number of characters placed in the buffer  Returns:     number of characters placed in the buffer
1109  */  */
1110    
 #if !defined NOUTF8  
   
1111  static int  static int
1112  ord2utf8(int cvalue, uschar *utf8bytes)  ord2utf8(int cvalue, pcre_uint8 *utf8bytes)
1113  {  {
1114  register int i, j;  register int i, j;
1115  for (i = 0; i < utf8_table1_size; i++)  for (i = 0; i < utf8_table1_size; i++)
# Line 375  for (j = i; j > 0; j--) Line 1123  for (j = i; j > 0; j--)
1123  *utf8bytes = utf8_table2[i] | cvalue;  *utf8bytes = utf8_table2[i] | cvalue;
1124  return i + 1;  return i + 1;
1125  }  }
1126    #endif /* NOUTF || SUPPORT_PCRE16 */
1127    
1128    
1129    
1130    #ifdef SUPPORT_PCRE16
1131    /*************************************************
1132    *         Convert a string to 16-bit             *
1133    *************************************************/
1134    
1135    /* In non-UTF mode, the space needed for a 16-bit string is exactly double the
1136    8-bit size. For a UTF-8 string, the size needed for UTF-16 is no more than
1137    double, because up to 0xffff uses no more than 3 bytes in UTF-8 but possibly 4
1138    in UTF-16. Higher values use 4 bytes in UTF-8 and up to 4 bytes in UTF-16. The
1139    result is always left in buffer16.
1140    
1141    Note that this function does not object to surrogate values. This is
1142    deliberate; it makes it possible to construct UTF-16 strings that are invalid,
1143    for the purpose of testing that they are correctly faulted.
1144    
1145    Patterns to be converted are either plain ASCII or UTF-8; data lines are always
1146    in UTF-8 so that values greater than 255 can be handled.
1147    
1148    Arguments:
1149      data       TRUE if converting a data line; FALSE for a regex
1150      p          points to a byte string
1151      utf        true if UTF-8 (to be converted to UTF-16)
1152      len        number of bytes in the string (excluding trailing zero)
1153    
1154    Returns:     number of 16-bit data items used (excluding trailing zero)
1155                 OR -1 if a UTF-8 string is malformed
1156                 OR -2 if a value > 0x10ffff is encountered
1157                 OR -3 if a value > 0xffff is encountered when not in UTF mode
1158    */
1159    
1160    static int
1161    to16(int data, pcre_uint8 *p, int utf, int len)
1162    {
1163    pcre_uint16 *pp;
1164    
1165    if (buffer16_size < 2*len + 2)
1166      {
1167      if (buffer16 != NULL) free(buffer16);
1168      buffer16_size = 2*len + 2;
1169      buffer16 = (pcre_uint16 *)malloc(buffer16_size);
1170      if (buffer16 == NULL)
1171        {
1172        fprintf(stderr, "pcretest: malloc(%d) failed for buffer16\n", buffer16_size);
1173        exit(1);
1174        }
1175      }
1176    
1177    pp = buffer16;
1178    
1179    if (!utf && !data)
1180      {
1181      while (len-- > 0) *pp++ = *p++;
1182      }
1183    
1184    else
1185      {
1186      int c = 0;
1187      while (len > 0)
1188        {
1189        int chlen = utf82ord(p, &c);
1190        if (chlen <= 0) return -1;
1191        if (c > 0x10ffff) return -2;
1192        p += chlen;
1193        len -= chlen;
1194        if (c < 0x10000) *pp++ = c; else
1195          {
1196          if (!utf) return -3;
1197          c -= 0x10000;
1198          *pp++ = 0xD800 | (c >> 10);
1199          *pp++ = 0xDC00 | (c & 0x3ff);
1200          }
1201        }
1202      }
1203    
1204    *pp = 0;
1205    return pp - buffer16;
1206    }
1207  #endif  #endif
1208    
1209    
1210    /*************************************************
1211    *        Read or extend an input line            *
1212    *************************************************/
1213    
1214    /* Input lines are read into buffer, but both patterns and data lines can be
1215    continued over multiple input lines. In addition, if the buffer fills up, we
1216    want to automatically expand it so as to be able to handle extremely large
1217    lines that are needed for certain stress tests. When the input buffer is
1218    expanded, the other two buffers must also be expanded likewise, and the
1219    contents of pbuffer, which are a copy of the input for callouts, must be
1220    preserved (for when expansion happens for a data line). This is not the most
1221    optimal way of handling this, but hey, this is just a test program!
1222    
1223    Arguments:
1224      f            the file to read
1225      start        where in buffer to start (this *must* be within buffer)
1226      prompt       for stdin or readline()
1227    
1228    Returns:       pointer to the start of new data
1229                   could be a copy of start, or could be moved
1230                   NULL if no data read and EOF reached
1231    */
1232    
1233    static pcre_uint8 *
1234    extend_inputline(FILE *f, pcre_uint8 *start, const char *prompt)
1235    {
1236    pcre_uint8 *here = start;
1237    
1238    for (;;)
1239      {
1240      int rlen = (int)(buffer_size - (here - buffer));
1241    
1242      if (rlen > 1000)
1243        {
1244        int dlen;
1245    
1246        /* If libreadline support is required, use readline() to read a line if the
1247        input is a terminal. Note that readline() removes the trailing newline, so
1248        we must put it back again, to be compatible with fgets(). */
1249    
1250    #ifdef SUPPORT_LIBREADLINE
1251        if (isatty(fileno(f)))
1252          {
1253          size_t len;
1254          char *s = readline(prompt);
1255          if (s == NULL) return (here == start)? NULL : start;
1256          len = strlen(s);
1257          if (len > 0) add_history(s);
1258          if (len > rlen - 1) len = rlen - 1;
1259          memcpy(here, s, len);
1260          here[len] = '\n';
1261          here[len+1] = 0;
1262          free(s);
1263          }
1264        else
1265    #endif
1266    
1267        /* Read the next line by normal means, prompting if the file is stdin. */
1268    
1269          {
1270          if (f == stdin) printf("%s", prompt);
1271          if (fgets((char *)here, rlen,  f) == NULL)
1272            return (here == start)? NULL : start;
1273          }
1274    
1275        dlen = (int)strlen((char *)here);
1276        if (dlen > 0 && here[dlen - 1] == '\n') return start;
1277        here += dlen;
1278        }
1279    
1280      else
1281        {
1282        int new_buffer_size = 2*buffer_size;
1283        pcre_uint8 *new_buffer = (pcre_uint8 *)malloc(new_buffer_size);
1284        pcre_uint8 *new_dbuffer = (pcre_uint8 *)malloc(new_buffer_size);
1285        pcre_uint8 *new_pbuffer = (pcre_uint8 *)malloc(new_buffer_size);
1286    
1287        if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
1288          {
1289          fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
1290          exit(1);
1291          }
1292    
1293        memcpy(new_buffer, buffer, buffer_size);
1294        memcpy(new_pbuffer, pbuffer, buffer_size);
1295    
1296        buffer_size = new_buffer_size;
1297    
1298        start = new_buffer + (start - buffer);
1299        here = new_buffer + (here - buffer);
1300    
1301        free(buffer);
1302        free(dbuffer);
1303        free(pbuffer);
1304    
1305        buffer = new_buffer;
1306        dbuffer = new_dbuffer;
1307        pbuffer = new_pbuffer;
1308        }
1309      }
1310    
1311    return NULL;  /* Control never gets here */
1312    }
1313    
1314    
1315    
1316  /*************************************************  /*************************************************
1317  *             Print character string             *  *          Read number from string               *
1318  *************************************************/  *************************************************/
1319    
1320  /* Character string printing function. Must handle UTF-8 strings in utf8  /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess
1321  mode. Yields number of characters printed. If handed a NULL file, just counts  around with conditional compilation, just do the job by hand. It is only used
1322  chars without printing. */  for unpicking arguments, so just keep it simple.
1323    
1324  static int pchars(unsigned char *p, int length, FILE *f)  Arguments:
1325      str           string to be converted
1326      endptr        where to put the end pointer
1327    
1328    Returns:        the unsigned long
1329    */
1330    
1331    static int
1332    get_value(pcre_uint8 *str, pcre_uint8 **endptr)
1333    {
1334    int result = 0;
1335    while(*str != 0 && isspace(*str)) str++;
1336    while (isdigit(*str)) result = result * 10 + (int)(*str++ - '0');
1337    *endptr = str;
1338    return(result);
1339    }
1340    
1341    
1342    
1343    /*************************************************
1344    *             Print one character                *
1345    *************************************************/
1346    
1347    /* Print a single character either literally, or as a hex escape. */
1348    
1349    static int pchar(int c, FILE *f)
1350    {
1351    if (PRINTOK(c))
1352      {
1353      if (f != NULL) fprintf(f, "%c", c);
1354      return 1;
1355      }
1356    
1357    if (c < 0x100)
1358      {
1359      if (use_utf)
1360        {
1361        if (f != NULL) fprintf(f, "\\x{%02x}", c);
1362        return 6;
1363        }
1364      else
1365        {
1366        if (f != NULL) fprintf(f, "\\x%02x", c);
1367        return 4;
1368        }
1369      }
1370    
1371    if (f != NULL) fprintf(f, "\\x{%02x}", c);
1372    return (c <= 0x000000ff)? 6 :
1373           (c <= 0x00000fff)? 7 :
1374           (c <= 0x0000ffff)? 8 :
1375           (c <= 0x000fffff)? 9 : 10;
1376    }
1377    
1378    
1379    
1380    #ifdef SUPPORT_PCRE8
1381    /*************************************************
1382    *         Print 8-bit character string           *
1383    *************************************************/
1384    
1385    /* Must handle UTF-8 strings in utf8 mode. Yields number of characters printed.
1386    If handed a NULL file, just counts chars without printing. */
1387    
1388    static int pchars(pcre_uint8 *p, int length, FILE *f)
1389  {  {
1390  int c = 0;  int c = 0;
1391  int yield = 0;  int yield = 0;
1392    
1393    if (length < 0)
1394      length = strlen((char *)p);
1395    
1396  while (length-- > 0)  while (length-- > 0)
1397    {    {
1398  #if !defined NOUTF8  #if !defined NOUTF
1399    if (use_utf8)    if (use_utf)
1400      {      {
1401      int rc = utf82ord(p, &c);      int rc = utf82ord(p, &c);
   
1402      if (rc > 0 && rc <= length + 1)   /* Mustn't run over the end */      if (rc > 0 && rc <= length + 1)   /* Mustn't run over the end */
1403        {        {
1404        length -= rc - 1;        length -= rc - 1;
1405        p += rc;        p += rc;
1406        if (PRINTHEX(c))        yield += pchar(c, f);
         {  
         if (f != NULL) fprintf(f, "%c", c);  
         yield++;  
         }  
       else  
         {  
         int n = 4;  
         if (f != NULL) fprintf(f, "\\x{%02x}", c);  
         yield += (n <= 0x000000ff)? 2 :  
                  (n <= 0x00000fff)? 3 :  
                  (n <= 0x0000ffff)? 4 :  
                  (n <= 0x000fffff)? 5 : 6;  
         }  
1407        continue;        continue;
1408        }        }
1409      }      }
1410  #endif  #endif
1411      c = *p++;
1412      yield += pchar(c, f);
1413      }
1414    
1415     /* Not UTF-8, or malformed UTF-8  */  return yield;
1416    }
1417    #endif
1418    
1419    c = *p++;  
1420    if (PRINTHEX(c))  
1421      {  #ifdef SUPPORT_PCRE16
1422      if (f != NULL) fprintf(f, "%c", c);  /*************************************************
1423      yield++;  *    Find length of 0-terminated 16-bit string   *
1424      }  *************************************************/
1425    else  
1426    static int strlen16(PCRE_SPTR16 p)
1427    {
1428    int len = 0;
1429    while (*p++ != 0) len++;
1430    return len;
1431    }
1432    #endif  /* SUPPORT_PCRE16 */
1433    
1434    
1435    #ifdef SUPPORT_PCRE16
1436    /*************************************************
1437    *           Print 16-bit character string        *
1438    *************************************************/
1439    
1440    /* Must handle UTF-16 strings in utf mode. Yields number of characters printed.
1441    If handed a NULL file, just counts chars without printing. */
1442    
1443    static int pchars16(PCRE_SPTR16 p, int length, FILE *f)
1444    {
1445    int yield = 0;
1446    
1447    if (length < 0)
1448      length = strlen16(p);
1449    
1450    while (length-- > 0)
1451      {
1452      int c = *p++ & 0xffff;
1453    #if !defined NOUTF
1454      if (use_utf && c >= 0xD800 && c < 0xDC00 && length > 0)
1455      {      {
1456      if (f != NULL) fprintf(f, "\\x%02x", c);      int d = *p & 0xffff;
1457      yield += 4;      if (d >= 0xDC00 && d < 0xDFFF)
1458          {
1459          c = ((c & 0x3ff) << 10) + (d & 0x3ff) + 0x10000;
1460          length--;
1461          p++;
1462          }
1463      }      }
1464    #endif
1465      yield += pchar(c, f);
1466    }    }
1467    
1468  return yield;  return yield;
1469  }  }
1470    #endif  /* SUPPORT_PCRE16 */
1471    
1472    
1473    
1474    #ifdef SUPPORT_PCRE8
1475    /*************************************************
1476    *     Read a capture name (8-bit) and check it   *
1477    *************************************************/
1478    
1479    static pcre_uint8 *
1480    read_capture_name8(pcre_uint8 *p, pcre_uint8 **pp, pcre *re)
1481    {
1482    pcre_uint8 *npp = *pp;
1483    while (isalnum(*p)) *npp++ = *p++;
1484    *npp++ = 0;
1485    *npp = 0;
1486    if (pcre_get_stringnumber(re, (char *)(*pp)) < 0)
1487      {
1488      fprintf(outfile, "no parentheses with name \"");
1489      PCHARSV(*pp, 0, -1, outfile);
1490      fprintf(outfile, "\"\n");
1491      }
1492    
1493    *pp = npp;
1494    return p;
1495    }
1496    #endif  /* SUPPORT_PCRE8 */
1497    
1498    
1499    
1500    #ifdef SUPPORT_PCRE16
1501    /*************************************************
1502    *     Read a capture name (16-bit) and check it  *
1503    *************************************************/
1504    
1505    /* Note that the text being read is 8-bit. */
1506    
1507    static pcre_uint8 *
1508    read_capture_name16(pcre_uint8 *p, pcre_uint16 **pp, pcre *re)
1509    {
1510    pcre_uint16 *npp = *pp;
1511    while (isalnum(*p)) *npp++ = *p++;
1512    *npp++ = 0;
1513    *npp = 0;
1514    if (pcre16_get_stringnumber(re, (PCRE_SPTR16)(*pp)) < 0)
1515      {
1516      fprintf(outfile, "no parentheses with name \"");
1517      PCHARSV(*pp, 0, -1, outfile);
1518      fprintf(outfile, "\"\n");
1519      }
1520    *pp = npp;
1521    return p;
1522    }
1523    #endif  /* SUPPORT_PCRE16 */
1524    
1525    
1526    
# Line 468  if (callout_extra) Line 1549  if (callout_extra)
1549      else      else
1550        {        {
1551        fprintf(f, "%2d: ", i/2);        fprintf(f, "%2d: ", i/2);
1552        (void)pchars((unsigned char *)cb->subject + cb->offset_vector[i],        PCHARSV(cb->subject, cb->offset_vector[i],
1553          cb->offset_vector[i+1] - cb->offset_vector[i], f);          cb->offset_vector[i+1] - cb->offset_vector[i], f);
1554        fprintf(f, "\n");        fprintf(f, "\n");
1555        }        }
# Line 481  printed lengths of the substrings. */ Line 1562  printed lengths of the substrings. */
1562    
1563  if (f != NULL) fprintf(f, "--->");  if (f != NULL) fprintf(f, "--->");
1564    
1565  pre_start = pchars((unsigned char *)cb->subject, cb->start_match, f);  PCHARS(pre_start, cb->subject, 0, cb->start_match, f);
1566  post_start = pchars((unsigned char *)(cb->subject + cb->start_match),  PCHARS(post_start, cb->subject, cb->start_match,
1567    cb->current_position - cb->start_match, f);    cb->current_position - cb->start_match, f);
1568    
1569  subject_length = pchars((unsigned char *)cb->subject, cb->subject_length, NULL);  PCHARS(subject_length, cb->subject, 0, cb->subject_length, NULL);
1570    
1571  (void)pchars((unsigned char *)(cb->subject + cb->current_position),  PCHARSV(cb->subject, cb->current_position,
1572    cb->subject_length - cb->current_position, f);    cb->subject_length - cb->current_position, f);
1573    
1574  if (f != NULL) fprintf(f, "\n");  if (f != NULL) fprintf(f, "\n");
# Line 524  fprintf(outfile, "%.*s", (cb->next_item_ Line 1605  fprintf(outfile, "%.*s", (cb->next_item_
1605  fprintf(outfile, "\n");  fprintf(outfile, "\n");
1606  first_callout = 0;  first_callout = 0;
1607    
1608    if (cb->mark != last_callout_mark)
1609      {
1610      if (cb->mark == NULL)
1611        fprintf(outfile, "Latest Mark: <unset>\n");
1612      else
1613        {
1614        fprintf(outfile, "Latest Mark: ");
1615        PCHARSV(cb->mark, 0, -1, outfile);
1616        putc('\n', outfile);
1617        }
1618      last_callout_mark = cb->mark;
1619      }
1620    
1621  if (cb->callout_data != NULL)  if (cb->callout_data != NULL)
1622    {    {
1623    int callout_data = *((int *)(cb->callout_data));    int callout_data = *((int *)(cb->callout_data));
# Line 543  return (cb->callout_number != callout_fa Line 1637  return (cb->callout_number != callout_fa
1637  *            Local malloc functions              *  *            Local malloc functions              *
1638  *************************************************/  *************************************************/
1639    
1640  /* Alternative malloc function, to test functionality and show the size of the  /* Alternative malloc function, to test functionality and save the size of a
1641  compiled re. */  compiled re, which is the first store request that pcre_compile() makes. The
1642    show_malloc variable is set only during matching. */
1643    
1644  static void *new_malloc(size_t size)  static void *new_malloc(size_t size)
1645  {  {
1646  void *block = malloc(size);  void *block = malloc(size);
1647  gotten_store = size;  gotten_store = size;
1648    if (first_gotten_store == 0) first_gotten_store = size;
1649  if (show_malloc)  if (show_malloc)
1650    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);
1651  return block;  return block;
# Line 562  if (show_malloc) Line 1658  if (show_malloc)
1658  free(block);  free(block);
1659  }  }
1660    
   
1661  /* For recursion malloc/free, to test stacking calls */  /* For recursion malloc/free, to test stacking calls */
1662    
1663  static void *stack_malloc(size_t size)  static void *stack_malloc(size_t size)
# Line 585  free(block); Line 1680  free(block);
1680  *          Call pcre_fullinfo()                  *  *          Call pcre_fullinfo()                  *
1681  *************************************************/  *************************************************/
1682    
1683  /* Get one piece of information from the pcre_fullinfo() function */  /* Get one piece of information from the pcre_fullinfo() function. When only
1684    one of 8-bit or 16-bit is supported, use_pcre16 should always have the correct
1685    value, but the code is defensive.
1686    
1687    Arguments:
1688      re        compiled regex
1689      study     study data
1690      option    PCRE_INFO_xxx option
1691      ptr       where to put the data
1692    
1693    Returns:    0 when OK, < 0 on error
1694    */
1695    
1696  static void new_info(pcre *re, pcre_extra *study, int option, void *ptr)  static int
1697    new_info(pcre *re, pcre_extra *study, int option, void *ptr)
1698  {  {
1699  int rc;  int rc;
1700  if ((rc = pcre_fullinfo(re, study, option, ptr)) < 0)  
1701    fprintf(outfile, "Error %d from pcre_fullinfo(%d)\n", rc, option);  if (use_pcre16)
1702    #ifdef SUPPORT_PCRE16
1703      rc = pcre16_fullinfo(re, study, option, ptr);
1704    #else
1705      rc = PCRE_ERROR_BADMODE;
1706    #endif
1707    else
1708    #ifdef SUPPORT_PCRE8
1709      rc = pcre_fullinfo(re, study, option, ptr);
1710    #else
1711      rc = PCRE_ERROR_BADMODE;
1712    #endif
1713    
1714    if (rc < 0)
1715      {
1716      fprintf(outfile, "Error %d from pcre%s_fullinfo(%d)\n", rc,
1717        use_pcre16? "16" : "", option);
1718      if (rc == PCRE_ERROR_BADMODE)
1719        fprintf(outfile, "Running in %s-bit mode but pattern was compiled in "
1720          "%s-bit mode\n", use_pcre16? "16":"8", use_pcre16? "8":"16");
1721      }
1722    
1723    return rc;
1724  }  }
1725    
1726    
1727    
1728  /*************************************************  /*************************************************
1729  *         Byte flipping function                 *  *             Swap byte functions                *
1730  *************************************************/  *************************************************/
1731    
1732  static unsigned long int  /* The following functions swap the bytes of a pcre_uint16 and pcre_uint32
1733  byteflip(unsigned long int value, int n)  value, respectively.
1734    
1735    Arguments:
1736      value        any number
1737    
1738    Returns:       the byte swapped value
1739    */
1740    
1741    static pcre_uint32
1742    swap_uint32(pcre_uint32 value)
1743  {  {
 if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  
1744  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
1745         ((value & 0x0000ff00) <<  8) |         ((value & 0x0000ff00) <<  8) |
1746         ((value & 0x00ff0000) >>  8) |         ((value & 0x00ff0000) >>  8) |
1747         ((value & 0xff000000) >> 24);         (value >> 24);
1748    }
1749    
1750    static pcre_uint16
1751    swap_uint16(pcre_uint16 value)
1752    {
1753    return (value >> 8) | (value << 8);
1754  }  }
1755    
1756    
1757    
1758    /*************************************************
1759    *        Flip bytes in a compiled pattern        *
1760    *************************************************/
1761    
1762    /* This function is called if the 'F' option was present on a pattern that is
1763    to be written to a file. We flip the bytes of all the integer fields in the
1764    regex data block and the study block. In 16-bit mode this also flips relevant
1765    bytes in the pattern itself. This is to make it possible to test PCRE's
1766    ability to reload byte-flipped patterns, e.g. those compiled on a different
1767    architecture. */
1768    
1769    static void
1770    regexflip(pcre *ere, pcre_extra *extra)
1771    {
1772    real_pcre *re = (real_pcre *)ere;
1773    #ifdef SUPPORT_PCRE16
1774    int op;
1775    pcre_uint16 *ptr = (pcre_uint16 *)re + re->name_table_offset;
1776    int length = re->name_count * re->name_entry_size;
1777    #ifdef SUPPORT_UTF
1778    BOOL utf = (re->options & PCRE_UTF16) != 0;
1779    BOOL utf16_char = FALSE;
1780    #endif /* SUPPORT_UTF */
1781    #endif /* SUPPORT_PCRE16 */
1782    
1783    /* Always flip the bytes in the main data block and study blocks. */
1784    
1785    re->magic_number = REVERSED_MAGIC_NUMBER;
1786    re->size = swap_uint32(re->size);
1787    re->options = swap_uint32(re->options);
1788    re->flags = swap_uint16(re->flags);
1789    re->top_bracket = swap_uint16(re->top_bracket);
1790    re->top_backref = swap_uint16(re->top_backref);
1791    re->first_char = swap_uint16(re->first_char);
1792    re->req_char = swap_uint16(re->req_char);
1793    re->name_table_offset = swap_uint16(re->name_table_offset);
1794    re->name_entry_size = swap_uint16(re->name_entry_size);
1795    re->name_count = swap_uint16(re->name_count);
1796    
1797    if (extra != NULL)
1798      {
1799      pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1800      rsd->size = swap_uint32(rsd->size);
1801      rsd->flags = swap_uint32(rsd->flags);
1802      rsd->minlength = swap_uint32(rsd->minlength);
1803      }
1804    
1805    /* In 8-bit mode, that is all we need to do. In 16-bit mode we must swap bytes
1806    in the name table, if present, and then in the pattern itself. */
1807    
1808    #ifdef SUPPORT_PCRE16
1809    if (!use_pcre16) return;
1810    
1811    while(TRUE)
1812      {
1813      /* Swap previous characters. */
1814      while (length-- > 0)
1815        {
1816        *ptr = swap_uint16(*ptr);
1817        ptr++;
1818        }
1819    #ifdef SUPPORT_UTF
1820      if (utf16_char)
1821        {
1822        if ((ptr[-1] & 0xfc00) == 0xd800)
1823          {
1824          /* We know that there is only one extra character in UTF-16. */
1825          *ptr = swap_uint16(*ptr);
1826          ptr++;
1827          }
1828        }
1829      utf16_char = FALSE;
1830    #endif /* SUPPORT_UTF */
1831    
1832      /* Get next opcode. */
1833    
1834      length = 0;
1835      op = *ptr;
1836      *ptr++ = swap_uint16(op);
1837    
1838      switch (op)
1839        {
1840        case OP_END:
1841        return;
1842    
1843    #ifdef SUPPORT_UTF
1844        case OP_CHAR:
1845        case OP_CHARI:
1846        case OP_NOT:
1847        case OP_NOTI:
1848        case OP_STAR:
1849        case OP_MINSTAR:
1850        case OP_PLUS:
1851        case OP_MINPLUS:
1852        case OP_QUERY:
1853        case OP_MINQUERY:
1854        case OP_UPTO:
1855        case OP_MINUPTO:
1856        case OP_EXACT:
1857        case OP_POSSTAR:
1858        case OP_POSPLUS:
1859        case OP_POSQUERY:
1860        case OP_POSUPTO:
1861        case OP_STARI:
1862        case OP_MINSTARI:
1863        case OP_PLUSI:
1864        case OP_MINPLUSI:
1865        case OP_QUERYI:
1866        case OP_MINQUERYI:
1867        case OP_UPTOI:
1868        case OP_MINUPTOI:
1869        case OP_EXACTI:
1870        case OP_POSSTARI:
1871        case OP_POSPLUSI:
1872        case OP_POSQUERYI:
1873        case OP_POSUPTOI:
1874        case OP_NOTSTAR:
1875        case OP_NOTMINSTAR:
1876        case OP_NOTPLUS:
1877        case OP_NOTMINPLUS:
1878        case OP_NOTQUERY:
1879        case OP_NOTMINQUERY:
1880        case OP_NOTUPTO:
1881        case OP_NOTMINUPTO:
1882        case OP_NOTEXACT:
1883        case OP_NOTPOSSTAR:
1884        case OP_NOTPOSPLUS:
1885        case OP_NOTPOSQUERY:
1886        case OP_NOTPOSUPTO:
1887        case OP_NOTSTARI:
1888        case OP_NOTMINSTARI:
1889        case OP_NOTPLUSI:
1890        case OP_NOTMINPLUSI:
1891        case OP_NOTQUERYI:
1892        case OP_NOTMINQUERYI:
1893        case OP_NOTUPTOI:
1894        case OP_NOTMINUPTOI:
1895        case OP_NOTEXACTI:
1896        case OP_NOTPOSSTARI:
1897        case OP_NOTPOSPLUSI:
1898        case OP_NOTPOSQUERYI:
1899        case OP_NOTPOSUPTOI:
1900        if (utf) utf16_char = TRUE;
1901    #endif
1902        /* Fall through. */
1903    
1904        default:
1905        length = OP_lengths16[op] - 1;
1906        break;
1907    
1908        case OP_CLASS:
1909        case OP_NCLASS:
1910        /* Skip the character bit map. */
1911        ptr += 32/sizeof(pcre_uint16);
1912        length = 0;
1913        break;
1914    
1915        case OP_XCLASS:
1916        /* Reverse the size of the XCLASS instance. */
1917        ptr++;
1918        *ptr = swap_uint16(*ptr);
1919        if (LINK_SIZE > 1)
1920          {
1921          /* LINK_SIZE can be 1 or 2 in 16 bit mode. */
1922          ptr++;
1923          *ptr = swap_uint16(*ptr);
1924          }
1925        ptr++;
1926    
1927        if (LINK_SIZE > 1)
1928          length = ((ptr[-LINK_SIZE] << 16) | ptr[-LINK_SIZE + 1]) -
1929            (1 + LINK_SIZE + 1);
1930        else
1931          length = ptr[-LINK_SIZE] - (1 + LINK_SIZE + 1);
1932    
1933        op = *ptr;
1934        *ptr = swap_uint16(op);
1935        if ((op & XCL_MAP) != 0)
1936          {
1937          /* Skip the character bit map. */
1938          ptr += 32/sizeof(pcre_uint16);
1939          length -= 32/sizeof(pcre_uint16);
1940          }
1941        break;
1942        }
1943      }
1944    /* Control should never reach here in 16 bit mode. */
1945    #endif /* SUPPORT_PCRE16 */
1946    }
1947    
1948    
1949    
1950  /*************************************************  /*************************************************
1951  *        Check match or recursion limit          *  *        Check match or recursion limit          *
1952  *************************************************/  *************************************************/
1953    
1954  static int  static int
1955  check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,  check_match_limit(pcre *re, pcre_extra *extra, pcre_uint8 *bptr, int len,
1956    int start_offset, int options, int *use_offsets, int use_size_offsets,    int start_offset, int options, int *use_offsets, int use_size_offsets,
1957    int flag, unsigned long int *limit, int errnumber, const char *msg)    int flag, unsigned long int *limit, int errnumber, const char *msg)
1958  {  {
# Line 633  for (;;) Line 1967  for (;;)
1967    {    {
1968    *limit = mid;    *limit = mid;
1969    
1970    count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,    PCRE_EXEC(count, re, extra, bptr, len, start_offset, options,
1971      use_offsets, use_size_offsets);      use_offsets, use_size_offsets);
1972    
1973    if (count == errnumber)    if (count == errnumber)
# Line 678  Returns:    < 0, = 0, or > 0, according Line 2012  Returns:    < 0, = 0, or > 0, according
2012  */  */
2013    
2014  static int  static int
2015  strncmpic(uschar *s, uschar *t, int n)  strncmpic(pcre_uint8 *s, pcre_uint8 *t, int n)
2016  {  {
2017  while (n--)  while (n--)
2018    {    {
# Line 694  return 0; Line 2028  return 0;
2028  *         Check newline indicator                *  *         Check newline indicator                *
2029  *************************************************/  *************************************************/
2030    
2031  /* This is used both at compile and run-time to check for <xxx> escapes, where  /* This is used both at compile and run-time to check for <xxx> escapes. Print
2032  xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is  a message and return 0 if there is no match.
 no match.  
2033    
2034  Arguments:  Arguments:
2035    p           points after the leading '<'    p           points after the leading '<'
# Line 706  Returns:      appropriate PCRE_NEWLINE_x Line 2039  Returns:      appropriate PCRE_NEWLINE_x
2039  */  */
2040    
2041  static int  static int
2042  check_newline(uschar *p, FILE *f)  check_newline(pcre_uint8 *p, FILE *f)
2043  {  {
2044  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (pcre_uint8 *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
2045  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (pcre_uint8 *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
2046  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (pcre_uint8 *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
2047  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;  if (strncmpic(p, (pcre_uint8 *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
2048  if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (pcre_uint8 *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
2049  if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;  if (strncmpic(p, (pcre_uint8 *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
2050  if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;  if (strncmpic(p, (pcre_uint8 *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
2051  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
2052  return 0;  return 0;
2053  }  }
# Line 728  return 0; Line 2061  return 0;
2061  static void  static void
2062  usage(void)  usage(void)
2063  {  {
2064  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
2065    printf("Input and output default to stdin and stdout.\n");
2066    #ifdef SUPPORT_LIBREADLINE
2067    printf("If input is a terminal, readline() is used to read from it.\n");
2068    #else
2069    printf("This version of pcretest is not linked with readline().\n");
2070    #endif
2071    printf("\nOptions:\n");
2072    #ifdef SUPPORT_PCRE16
2073    printf("  -16      use 16-bit interface\n");
2074    #endif
2075  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
2076  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
2077    printf("  -C arg   show a specific compile-time option\n");
2078    printf("           and exit with its value. The arg can be:\n");
2079    printf("     linksize     internal link size [2, 3, 4]\n");
2080    printf("     pcre8        8 bit library support enabled [0, 1]\n");
2081    printf("     pcre16       16 bit library support enabled [0, 1]\n");
2082    printf("     utf          Unicode Transformation Format supported [0, 1]\n");
2083    printf("     ucp          Unicode Properties supported [0, 1]\n");
2084    printf("     jit          Just-in-time compiler supported [0, 1]\n");
2085  printf("  -d       debug: show compiled code and information (-b and -i)\n");  printf("  -d       debug: show compiled code and information (-b and -i)\n");
2086  #if !defined NODFA  #if !defined NODFA
2087  printf("  -dfa     force DFA matching for all subjects\n");  printf("  -dfa     force DFA matching for all subjects\n");
2088  #endif  #endif
2089  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
2090  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
2091           "  -M       find MATCH_LIMIT minimum for each subject\n"
2092         "  -m       output memory used information\n"         "  -m       output memory used information\n"
2093         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
2094  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 744  printf("  -p       use POSIX interface\n Line 2096  printf("  -p       use POSIX interface\n
2096  #endif  #endif
2097  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
2098  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
2099  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied at basic level\n"
2100           "  -s+      force each pattern to be studied, using JIT if available\n"
2101         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
2102  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
2103  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 764  options, followed by a set of test data, Line 2117  options, followed by a set of test data,
2117  int main(int argc, char **argv)  int main(int argc, char **argv)
2118  {  {
2119  FILE *infile = stdin;  FILE *infile = stdin;
2120    const char *version;
2121  int options = 0;  int options = 0;
2122  int study_options = 0;  int study_options = 0;
2123    int default_find_match_limit = FALSE;
2124  int op = 1;  int op = 1;
2125  int timeit = 0;  int timeit = 0;
2126  int timeitm = 0;  int timeitm = 0;
2127  int showinfo = 0;  int showinfo = 0;
2128  int showstore = 0;  int showstore = 0;
2129    int force_study = -1;
2130    int force_study_options = 0;
2131  int quiet = 0;  int quiet = 0;
2132  int size_offsets = 45;  int size_offsets = 45;
2133  int size_offsets_max;  int size_offsets_max;
# Line 784  int all_use_dfa = 0; Line 2141  int all_use_dfa = 0;
2141  int yield = 0;  int yield = 0;
2142  int stack_size;  int stack_size;
2143    
2144  /* These vectors store, end-to-end, a list of captured substring names. Assume  pcre_jit_stack *jit_stack = NULL;
 that 1024 is plenty long enough for the few names we'll be testing. */  
   
 uschar copynames[1024];  
 uschar getnames[1024];  
   
 uschar *copynamesptr;  
 uschar *getnamesptr;  
2145    
2146  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* These vectors store, end-to-end, a list of zero-terminated captured
2147  when I am debugging. They grow automatically when very long lines are read. */  substring names, each list itself being terminated by an empty name. Assume
2148    that 1024 is plenty long enough for the few names we'll be testing. It is
2149  buffer = (unsigned char *)malloc(buffer_size);  easiest to keep separate 8-bit and 16-bit versions, using the 16-bit version
2150  dbuffer = (unsigned char *)malloc(buffer_size);  for the actual memory, to ensure alignment. By defining these variables always
2151  pbuffer = (unsigned char *)malloc(buffer_size);  (whether or not 8-bit or 16-bit is supported), we avoid too much mess with
2152    #ifdefs in the code. */
2153    
2154    pcre_uint16 copynames[1024];
2155    pcre_uint16 getnames[1024];
2156    
2157    pcre_uint16 *cn16ptr;
2158    pcre_uint16 *gn16ptr;
2159    
2160    pcre_uint8 *copynames8 = (pcre_uint8 *)copynames;
2161    pcre_uint8 *getnames8 = (pcre_uint8 *)getnames;
2162    pcre_uint8 *cn8ptr;
2163    pcre_uint8 *gn8ptr;
2164    
2165    /* Get buffers from malloc() so that valgrind will check their misuse when
2166    debugging. They grow automatically when very long lines are read. The 16-bit
2167    buffer (buffer16) is obtained only if needed. */
2168    
2169    buffer = (pcre_uint8 *)malloc(buffer_size);
2170    dbuffer = (pcre_uint8 *)malloc(buffer_size);
2171    pbuffer = (pcre_uint8 *)malloc(buffer_size);
2172    
2173  /* The outfile variable is static so that new_malloc can use it. */  /* The outfile variable is static so that new_malloc can use it. */
2174    
# Line 813  it set 0x8000, but then I was advised th Line 2183  it set 0x8000, but then I was advised th
2183  _setmode( _fileno( stdout ), _O_BINARY );  _setmode( _fileno( stdout ), _O_BINARY );
2184  #endif  #endif
2185    
2186    /* Get the version number: both pcre_version() and pcre16_version() give the
2187    same answer. We just need to ensure that we call one that is available. */
2188    
2189    #ifdef SUPPORT_PCRE8
2190    version = pcre_version();
2191    #else
2192    version = pcre16_version();
2193    #endif
2194    
2195  /* Scan options */  /* Scan options */
2196    
2197  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
2198    {    {
2199    unsigned char *endptr;    pcre_uint8 *endptr;
2200    
2201    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
2202      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 0;
2203      else if (strcmp(argv[op], "-s+") == 0)
2204        {
2205        force_study = 1;
2206        force_study_options = PCRE_STUDY_JIT_COMPILE;
2207        }
2208      else if (strcmp(argv[op], "-16") == 0)
2209        {
2210    #ifdef SUPPORT_PCRE16
2211        use_pcre16 = 1;
2212    #else
2213        printf("** This version of PCRE was built without 16-bit support\n");
2214        exit(1);
2215    #endif
2216        }
2217    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
2218    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
2219    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
2220    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
2221      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
2222  #if !defined NODFA  #if !defined NODFA
2223    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
2224  #endif  #endif
2225    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
2226        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),        ((size_offsets = get_value((pcre_uint8 *)argv[op+1], &endptr)),
2227          *endptr == 0))          *endptr == 0))
2228      {      {
2229      op++;      op++;
# Line 839  while (argc > 1 && argv[op][0] == '-') Line 2233  while (argc > 1 && argv[op][0] == '-')
2233      {      {
2234      int both = argv[op][2] == 0;      int both = argv[op][2] == 0;
2235      int temp;      int temp;
2236      if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),      if (argc > 2 && (temp = get_value((pcre_uint8 *)argv[op+1], &endptr),
2237                       *endptr == 0))                       *endptr == 0))
2238        {        {
2239        timeitm = temp;        timeitm = temp;
# Line 850  while (argc > 1 && argv[op][0] == '-') Line 2244  while (argc > 1 && argv[op][0] == '-')
2244      if (both) timeit = timeitm;      if (both) timeit = timeitm;
2245      }      }
2246    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
2247        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((pcre_uint8 *)argv[op+1], &endptr)),
2248          *endptr == 0))          *endptr == 0))
2249      {      {
2250  #if defined(_WIN32) || defined(WIN32)  #if defined(_WIN32) || defined(WIN32) || defined(__minix)
2251      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
2252      exit(1);      exit(1);
2253  #else  #else
# Line 877  while (argc > 1 && argv[op][0] == '-') Line 2271  while (argc > 1 && argv[op][0] == '-')
2271    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
2272      {      {
2273      int rc;      int rc;
2274      printf("PCRE version %s\n", pcre_version());      unsigned long int lrc;
2275    
2276        if (argc > 2)
2277          {
2278          if (strcmp(argv[op + 1], "linksize") == 0)
2279            {
2280            (void)PCRE_CONFIG(PCRE_CONFIG_LINK_SIZE, &rc);
2281            printf("%d\n", rc);
2282            yield = rc;
2283            goto EXIT;
2284            }
2285          if (strcmp(argv[op + 1], "pcre8") == 0)
2286            {
2287    #ifdef SUPPORT_PCRE8
2288            printf("1\n");
2289            yield = 1;
2290    #else
2291            printf("0\n");
2292            yield = 0;
2293    #endif
2294            goto EXIT;
2295            }
2296          if (strcmp(argv[op + 1], "pcre16") == 0)
2297            {
2298    #ifdef SUPPORT_PCRE16
2299            printf("1\n");
2300            yield = 1;
2301    #else
2302            printf("0\n");
2303            yield = 0;
2304    #endif
2305            goto EXIT;
2306            }
2307          if (strcmp(argv[op + 1], "utf") == 0)
2308            {
2309    #ifdef SUPPORT_PCRE8
2310            (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
2311            printf("%d\n", rc);
2312            yield = rc;
2313    #else
2314            (void)pcre16_config(PCRE_CONFIG_UTF16, &rc);
2315            printf("%d\n", rc);
2316            yield = rc;
2317    #endif
2318            goto EXIT;
2319            }
2320          if (strcmp(argv[op + 1], "ucp") == 0)
2321            {
2322            (void)PCRE_CONFIG(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
2323            printf("%d\n", rc);
2324            yield = rc;
2325            goto EXIT;
2326            }
2327          if (strcmp(argv[op + 1], "jit") == 0)
2328            {
2329            (void)PCRE_CONFIG(PCRE_CONFIG_JIT, &rc);
2330            printf("%d\n", rc);
2331            yield = rc;
2332            goto EXIT;
2333            }
2334          printf("Unknown option: %s\n", argv[op + 1]);
2335          goto EXIT;
2336          }
2337    
2338        printf("PCRE version %s\n", version);
2339      printf("Compiled with\n");      printf("Compiled with\n");
2340    
2341    /* At least one of SUPPORT_PCRE8 and SUPPORT_PCRE16 will be set. If both
2342    are set, either both UTFs are supported or both are not supported. */
2343    
2344    #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
2345        printf("  8-bit and 16-bit support\n");
2346        (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
2347        if (rc)
2348          printf("  UTF-8 and UTF-16 support\n");
2349        else
2350          printf("  No UTF-8 or UTF-16 support\n");
2351    #elif defined SUPPORT_PCRE8
2352        printf("  8-bit support only\n");
2353      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
2354      printf("  %sUTF-8 support\n", rc? "" : "No ");      printf("  %sUTF-8 support\n", rc? "" : "No ");
2355      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);  #else
2356        printf("  16-bit support only\n");
2357        (void)pcre16_config(PCRE_CONFIG_UTF16, &rc);
2358        printf("  %sUTF-16 support\n", rc? "" : "No ");
2359    #endif
2360    
2361        (void)PCRE_CONFIG(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
2362      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
2363      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_JIT, &rc);
2364      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      if (rc)
2365        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :        printf("  Just-in-time compiler support\n");
2366        else
2367          printf("  No just-in-time compiler support\n");
2368        (void)PCRE_CONFIG(PCRE_CONFIG_NEWLINE, &rc);
2369        /* Note that these values are always the ASCII values, even
2370        in EBCDIC environments. CR is 13 and NL is 10. */
2371        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
2372          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
2373        (rc == -2)? "ANYCRLF" :        (rc == -2)? "ANYCRLF" :
2374        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
2375      (void)pcre_config(PCRE_CONFIG_BSR, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_BSR, &rc);
2376      printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :      printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
2377                                       "all Unicode newlines");                                       "all Unicode newlines");
2378      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_LINK_SIZE, &rc);
2379      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
2380      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
2381      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
2382      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_MATCH_LIMIT, &lrc);
2383      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
2384      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
2385      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
2386      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_STACKRECURSE, &rc);
2387      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
2388      goto EXIT;      goto EXIT;
2389      }      }
# Line 958  if (argc > 2) Line 2442  if (argc > 2)
2442    
2443  /* Set alternative malloc function */  /* Set alternative malloc function */
2444    
2445    #ifdef SUPPORT_PCRE8
2446  pcre_malloc = new_malloc;  pcre_malloc = new_malloc;
2447  pcre_free = new_free;  pcre_free = new_free;
2448  pcre_stack_malloc = stack_malloc;  pcre_stack_malloc = stack_malloc;
2449  pcre_stack_free = stack_free;  pcre_stack_free = stack_free;
2450    #endif
2451    
2452    #ifdef SUPPORT_PCRE16
2453    pcre16_malloc = new_malloc;
2454    pcre16_free = new_free;
2455    pcre16_stack_malloc = stack_malloc;
2456    pcre16_stack_free = stack_free;
2457    #endif
2458    
2459  /* Heading line unless quiet, then prompt for first regex if stdin */  /* Heading line unless quiet, then prompt for first regex if stdin */
2460    
2461  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", version);
2462    
2463  /* Main loop */  /* Main loop */
2464    
# Line 980  while (!done) Line 2473  while (!done)
2473  #endif  #endif
2474    
2475    const char *error;    const char *error;
2476    unsigned char *p, *pp, *ppp;    pcre_uint8 *markptr;
2477    unsigned char *to_file = NULL;    pcre_uint8 *p, *pp, *ppp;
2478    const unsigned char *tables = NULL;    pcre_uint8 *to_file = NULL;
2479      const pcre_uint8 *tables = NULL;
2480    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
2481    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
2482      int do_allcaps = 0;
2483      int do_mark = 0;
2484    int do_study = 0;    int do_study = 0;
2485      int no_force_study = 0;
2486    int do_debug = debug;    int do_debug = debug;
2487    int do_G = 0;    int do_G = 0;
2488    int do_g = 0;    int do_g = 0;
2489    int do_showinfo = showinfo;    int do_showinfo = showinfo;
2490    int do_showrest = 0;    int do_showrest = 0;
2491      int do_showcaprest = 0;
2492    int do_flip = 0;    int do_flip = 0;
2493    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
2494    
2495    use_utf8 = 0;    use_utf = 0;
2496    debug_lengths = 1;    debug_lengths = 1;
2497    
2498    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
2499    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
2500    fflush(outfile);    fflush(outfile);
2501    
# Line 1011  while (!done) Line 2508  while (!done)
2508    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
2509      {      {
2510      unsigned long int magic, get_options;      unsigned long int magic, get_options;
2511      uschar sbuf[8];      pcre_uint8 sbuf[8];
2512      FILE *f;      FILE *f;
2513    
2514      p++;      p++;
# Line 1034  while (!done) Line 2531  while (!done)
2531        (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7];        (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7];
2532    
2533      re = (real_pcre *)new_malloc(true_size);      re = (real_pcre *)new_malloc(true_size);
2534      regex_gotten_store = gotten_store;      regex_gotten_store = first_gotten_store;
2535    
2536      if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ;      if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ;
2537    
2538      magic = ((real_pcre *)re)->magic_number;      magic = ((real_pcre *)re)->magic_number;
2539      if (magic != MAGIC_NUMBER)      if (magic != MAGIC_NUMBER)
2540        {        {
2541        if (byteflip(magic, sizeof(magic)) == MAGIC_NUMBER)        if (swap_uint32(magic) == MAGIC_NUMBER)
2542          {          {
2543          do_flip = 1;          do_flip = 1;
2544          }          }
# Line 1053  while (!done) Line 2550  while (!done)
2550          }          }
2551        }        }
2552    
2553      fprintf(outfile, "Compiled regex%s loaded from %s\n",      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
2554        do_flip? " (byte-inverted)" : "", p);        do_flip? " (byte-inverted)" : "", p);
2555    
2556      /* Need to know if UTF-8 for printing data strings */      /* Now see if there is any following study data. */
   
     new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
     use_utf8 = (get_options & PCRE_UTF8) != 0;  
   
     /* Now see if there is any following study data */  
2557    
2558      if (true_study_size != 0)      if (true_study_size != 0)
2559        {        {
# Line 1077  while (!done) Line 2569  while (!done)
2569          {          {
2570          FAIL_READ:          FAIL_READ:
2571          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
2572          if (extra != NULL) new_free(extra);          if (extra != NULL)
2573              {
2574              PCRE_FREE_STUDY(extra);
2575              }
2576          if (re != NULL) new_free(re);          if (re != NULL) new_free(re);
2577          fclose(f);          fclose(f);
2578          continue;          continue;
# Line 1085  while (!done) Line 2580  while (!done)
2580        fprintf(outfile, "Study data loaded from %s\n", p);        fprintf(outfile, "Study data loaded from %s\n", p);
2581        do_study = 1;     /* To get the data output if requested */        do_study = 1;     /* To get the data output if requested */
2582        }        }
2583      else fprintf(outfile, "No study data\n");      else fprintf(outfile, "No study data\n");
2584    
2585        /* Flip the necessary bytes. */
2586        if (do_flip)
2587          {
2588          PCRE_PATTERN_TO_HOST_BYTE_ORDER(re, extra, NULL);
2589          }
2590    
2591        /* Need to know if UTF-8 for printing data strings. */
2592    
2593        if (new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options) < 0) continue;
2594        use_utf = (get_options & PCRE_UTF8) != 0;
2595    
2596      fclose(f);      fclose(f);
2597      goto SHOW_INFO;      goto SHOW_INFO;
2598      }      }
2599    
2600    /* In-line pattern (the usual case). Get the delimiter and seek the end of    /* In-line pattern (the usual case). Get the delimiter and seek the end of
2601    the pattern; if is isn't complete, read more. */    the pattern; if it isn't complete, read more. */
2602    
2603    delimiter = *p++;    delimiter = *p++;
2604    
2605    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
2606      {      {
2607      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
2608      goto SKIP_DATA;      goto SKIP_DATA;
2609      }      }
2610    
2611    pp = p;    pp = p;
2612    poffset = p - buffer;    poffset = (int)(p - buffer);
2613    
2614    for(;;)    for(;;)
2615      {      {
# Line 1114  while (!done) Line 2620  while (!done)
2620        pp++;        pp++;
2621        }        }
2622      if (*pp != 0) break;      if (*pp != 0) break;
2623      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
2624        {        {
2625        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
2626        done = 1;        done = 1;
# Line 1158  while (!done) Line 2663  while (!done)
2663        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
2664        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
2665    
2666        case '+': do_showrest = 1; break;        case '+':
2667          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
2668          break;
2669    
2670          case '=': do_allcaps = 1; break;
2671        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
2672        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
2673        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1168  while (!done) Line 2677  while (!done)
2677        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
2678        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
2679        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
2680          case 'K': do_mark = 1; break;
2681        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
2682        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
2683    
# Line 1175  while (!done) Line 2685  while (!done)
2685        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
2686  #endif  #endif
2687    
2688        case 'S': do_study = 1; break;        case 'S':
2689          if (do_study == 0)
2690            {
2691            do_study = 1;
2692            if (*pp == '+')
2693              {
2694              study_options |= PCRE_STUDY_JIT_COMPILE;
2695              pp++;
2696              }
2697            }
2698          else
2699            {
2700            do_study = 0;
2701            no_force_study = 1;
2702            }
2703          break;
2704    
2705        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
2706          case 'W': options |= PCRE_UCP; break;
2707        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
2708          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
2709        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
2710        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf = 1; break;
2711        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
2712    
2713          case 'T':
2714          switch (*pp++)
2715            {
2716            case '0': tables = tables0; break;
2717            case '1': tables = tables1; break;
2718    
2719            case '\r':
2720            case '\n':
2721            case ' ':
2722            case 0:
2723            fprintf(outfile, "** Missing table number after /T\n");
2724            goto SKIP_DATA;
2725    
2726            default:
2727            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
2728            goto SKIP_DATA;
2729            }
2730          break;
2731    
2732        case 'L':        case 'L':
2733        ppp = pp;        ppp = pp;
2734        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1194  while (!done) Line 2741  while (!done)
2741          goto SKIP_DATA;          goto SKIP_DATA;
2742          }          }
2743        locale_set = 1;        locale_set = 1;
2744        tables = pcre_maketables();        tables = PCRE_MAKETABLES;
2745        pp = ppp;        pp = ppp;
2746        break;        break;
2747    
# Line 1207  while (!done) Line 2754  while (!done)
2754    
2755        case '<':        case '<':
2756          {          {
2757          int x = check_newline(pp, outfile);          if (strncmpic(pp, (pcre_uint8 *)"JS>", 3) == 0)
2758          if (x == 0) goto SKIP_DATA;            {
2759          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
2760          while (*pp++ != '>');            pp += 3;
2761              }
2762            else
2763              {
2764              int x = check_newline(pp, outfile);
2765              if (x == 0) goto SKIP_DATA;
2766              options |= x;
2767              while (*pp++ != '>');
2768              }
2769          }          }
2770        break;        break;
2771    
# Line 1227  while (!done) Line 2782  while (!done)
2782    
2783    /* Handle compiling via the POSIX interface, which doesn't support the    /* Handle compiling via the POSIX interface, which doesn't support the
2784    timing, showing, or debugging options, nor the ability to pass over    timing, showing, or debugging options, nor the ability to pass over
2785    local character tables. */    local character tables. Neither does it have 16-bit support. */
2786    
2787  #if !defined NOPOSIX  #if !defined NOPOSIX
2788    if (posix || do_posix)    if (posix || do_posix)
# Line 1240  while (!done) Line 2795  while (!done)
2795      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
2796      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
2797      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
2798        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
2799        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
2800    
2801        first_gotten_store = 0;
2802      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
2803    
2804      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
# Line 1260  while (!done) Line 2818  while (!done)
2818  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
2819    
2820      {      {
2821        unsigned long int get_options;
2822    
2823        /* In 16-bit mode, convert the input. */
2824    
2825    #ifdef SUPPORT_PCRE16
2826        if (use_pcre16)
2827          {
2828          switch(to16(FALSE, p, options & PCRE_UTF8, (int)strlen((char *)p)))
2829            {
2830            case -1:
2831            fprintf(outfile, "**Failed: invalid UTF-8 string cannot be "
2832              "converted to UTF-16\n");
2833            goto SKIP_DATA;
2834    
2835            case -2:
2836            fprintf(outfile, "**Failed: character value greater than 0x10ffff "
2837              "cannot be converted to UTF-16\n");
2838            goto SKIP_DATA;
2839    
2840            case -3: /* "Impossible error" when to16 is called arg1 FALSE */
2841            fprintf(outfile, "**Failed: character value greater than 0xffff "
2842              "cannot be converted to 16-bit in non-UTF mode\n");
2843            goto SKIP_DATA;
2844    
2845            default:
2846            break;
2847            }
2848          p = (pcre_uint8 *)buffer16;
2849          }
2850    #endif
2851    
2852        /* Compile many times when timing */
2853    
2854      if (timeit > 0)      if (timeit > 0)
2855        {        {
2856        register int i;        register int i;
# Line 1267  while (!done) Line 2858  while (!done)
2858        clock_t start_time = clock();        clock_t start_time = clock();
2859        for (i = 0; i < timeit; i++)        for (i = 0; i < timeit; i++)
2860          {          {
2861          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          PCRE_COMPILE(re, p, options, &error, &erroroffset, tables);
2862          if (re != NULL) free(re);          if (re != NULL) free(re);
2863          }          }
2864        time_taken = clock() - start_time;        time_taken = clock() - start_time;
# Line 1276  while (!done) Line 2867  while (!done)
2867            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
2868        }        }
2869    
2870      re = pcre_compile((char *)p, options, &error, &erroroffset, tables);      first_gotten_store = 0;
2871        PCRE_COMPILE(re, p, options, &error, &erroroffset, tables);
2872    
2873      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
2874      if non-interactive. */      if non-interactive. */
# Line 1289  while (!done) Line 2881  while (!done)
2881          {          {
2882          for (;;)          for (;;)
2883            {            {
2884            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
2885              {              {
2886              done = 1;              done = 1;
2887              goto CONTINUE;              goto CONTINUE;
# Line 1303  while (!done) Line 2895  while (!done)
2895        goto CONTINUE;        goto CONTINUE;
2896        }        }
2897    
2898      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
2899      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
2900      returns only limited data. Check that it agrees with the newer one. */      lines. */
2901    
2902      if (log_store)      if (new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options) < 0)
2903        fprintf(outfile, "Memory allocation (code space): %d\n",        goto SKIP_DATA;
2904          (int)(gotten_store -      if ((get_options & PCRE_UTF8) != 0) use_utf = 1;
               sizeof(real_pcre) -  
               ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size));  
2905    
2906      /* Extract the size for possible writing before possibly flipping it,      /* Extract the size for possible writing before possibly flipping it,
2907      and remember the store that was got. */      and remember the store that was got. */
2908    
2909      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
2910      regex_gotten_store = gotten_store;      regex_gotten_store = first_gotten_store;
2911    
2912        /* Output code size information if requested */
2913    
2914        if (log_store)
2915          fprintf(outfile, "Memory allocation (code space): %d\n",
2916            (int)(first_gotten_store -
2917                  sizeof(real_pcre) -
2918                  ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size));
2919    
2920      /* If /S was present, study the regexp to generate additional info to      /* If -s or /S was present, study the regex to generate additional info to
2921      help with the matching. */      help with the matching, unless the pattern has the SS option, which
2922        suppresses the effect of /S (used for a few test patterns where studying is
2923        never sensible). */
2924    
2925      if (do_study)      if (do_study || (force_study >= 0 && !no_force_study))
2926        {        {
2927        if (timeit > 0)        if (timeit > 0)
2928          {          {
# Line 1330  while (!done) Line 2930  while (!done)
2930          clock_t time_taken;          clock_t time_taken;
2931          clock_t start_time = clock();          clock_t start_time = clock();
2932          for (i = 0; i < timeit; i++)          for (i = 0; i < timeit; i++)
2933            extra = pcre_study(re, study_options, &error);            {
2934              PCRE_STUDY(extra, re, study_options | force_study_options, &error);
2935              }
2936          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2937          if (extra != NULL) free(extra);          if (extra != NULL)
2938              {
2939              PCRE_FREE_STUDY(extra);
2940              }
2941          fprintf(outfile, "  Study time %.4f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
2942            (((double)time_taken * 1000.0) / (double)timeit) /            (((double)time_taken * 1000.0) / (double)timeit) /
2943              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2944          }          }
2945        extra = pcre_study(re, study_options, &error);        PCRE_STUDY(extra, re, study_options | force_study_options, &error);
2946        if (error != NULL)        if (error != NULL)
2947          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
2948        else if (extra != NULL)        else if (extra != NULL)
2949            {
2950          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
2951            if (log_store)
2952              {
2953              size_t jitsize;
2954              if (new_info(re, extra, PCRE_INFO_JITSIZE, &jitsize) == 0 &&
2955                  jitsize != 0)
2956                fprintf(outfile, "Memory allocation (JIT code): %d\n", (int)jitsize);
2957              }
2958            }
2959        }        }
2960    
2961      /* If the 'F' option was present, we flip the bytes of all the integer      /* If /K was present, we set up for handling MARK data. */
     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. */  
2962    
2963      if (do_flip)      if (do_mark)
2964        {        {
2965        real_pcre *rre = (real_pcre *)re;        if (extra == NULL)
       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_byte =  
         (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));  
       rre->req_byte =  
         (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));  
       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)  
2966          {          {
2967          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          extra = (pcre_extra *)malloc(sizeof(pcre_extra));
2968          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          extra->flags = 0;
         rsd->options = byteflip(rsd->options, sizeof(rsd->options));  
2969          }          }
2970          extra->mark = &markptr;
2971          extra->flags |= PCRE_EXTRA_MARK;
2972        }        }
2973    
2974      /* Extract information from the compiled data if required */      /* Extract and display information from the compiled data if required. */
2975    
2976      SHOW_INFO:      SHOW_INFO:
2977    
2978      if (do_debug)      if (do_debug)
2979        {        {
2980        fprintf(outfile, "------------------------------------------------------------------\n");        fprintf(outfile, "------------------------------------------------------------------\n");
2981        pcre_printint(re, outfile, debug_lengths);        PCRE_PRINTINT(re, outfile, debug_lengths);
2982        }        }
2983    
2984        /* We already have the options in get_options (see above) */
2985    
2986      if (do_showinfo)      if (do_showinfo)
2987        {        {
2988        unsigned long int get_options, all_options;        unsigned long int all_options;
 #if !defined NOINFOCHECK  
       int old_first_char, old_options, old_count;  
 #endif  
2989        int count, backrefmax, first_char, need_char, okpartial, jchanged,        int count, backrefmax, first_char, need_char, okpartial, jchanged,
2990          hascrorlf;          hascrorlf;
2991        int nameentrysize, namecount;        int nameentrysize, namecount;
2992        const uschar *nametable;        const pcre_uint8 *nametable;
2993    
2994        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);        if (new_info(re, NULL, PCRE_INFO_SIZE, &size) +
2995        new_info(re, NULL, PCRE_INFO_SIZE, &size);            new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count) +
2996        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);            new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax) +
2997        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);            new_info(re, NULL, PCRE_INFO_FIRSTBYTE, &first_char) +
2998        new_info(re, NULL, PCRE_INFO_FIRSTBYTE, &first_char);            new_info(re, NULL, PCRE_INFO_LASTLITERAL, &need_char) +
2999        new_info(re, NULL, PCRE_INFO_LASTLITERAL, &need_char);            new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize) +
3000        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);            new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount) +
3001        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);            new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable) +
3002        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);            new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial) +
3003        new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);            new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged) +
3004        new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);            new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf)
3005        new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);            != 0)
3006            goto SKIP_DATA;
 #if !defined NOINFOCHECK  
       old_count = pcre_info(re, &old_options, &old_first_char);  
       if (count < 0) fprintf(outfile,  
         "Error %d from pcre_info()\n", count);  
       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  
3007    
3008        if (size != regex_gotten_store) fprintf(outfile,        if (size != regex_gotten_store) fprintf(outfile,
3009          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
# Line 1447  while (!done) Line 3018  while (!done)
3018          fprintf(outfile, "Named capturing subpatterns:\n");          fprintf(outfile, "Named capturing subpatterns:\n");
3019          while (namecount-- > 0)          while (namecount-- > 0)
3020            {            {
3021            fprintf(outfile, "  %s %*s%3d\n", nametable + 2,  #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
3022              nameentrysize - 3 - (int)strlen((char *)nametable + 2), "",            int imm2_size = use_pcre16 ? 1 : 2;
3023              GET2(nametable, 0));  #else
3024              int imm2_size = IMM2_SIZE;
3025    #endif
3026              int length = (int)STRLEN(nametable + imm2_size);
3027              fprintf(outfile, "  ");
3028              PCHARSV(nametable, imm2_size, length, outfile);
3029              while (length++ < nameentrysize - imm2_size) putc(' ', outfile);
3030    #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
3031              fprintf(outfile, "%3d\n", use_pcre16?
3032                 (int)(((PCRE_SPTR16)nametable)[0])
3033                :((int)nametable[0] << 8) | (int)nametable[1]);
3034              nametable += nameentrysize * (use_pcre16 ? 2 : 1);
3035    #else
3036              fprintf(outfile, "%3d\n", GET2(nametable, 0));
3037    #ifdef SUPPORT_PCRE8
3038            nametable += nameentrysize;            nametable += nameentrysize;
3039    #else
3040              nametable += nameentrysize * 2;
3041    #endif
3042    #endif
3043            }            }
3044          }          }
3045    
# Line 1458  while (!done) Line 3047  while (!done)
3047        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
3048    
3049        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
3050        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = swap_uint32(all_options);
3051    
3052        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
3053          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
3054            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
3055            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
3056            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
# Line 1474  while (!done) Line 3063  while (!done)
3063            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
3064            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
3065            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
3066            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf" : "",
3067            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_UCP) != 0)? " ucp" : "",
3068              ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf_check" : "",
3069              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
3070            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
3071    
3072        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
# Line 1516  while (!done) Line 3107  while (!done)
3107          }          }
3108        else        else
3109          {          {
3110          int ch = first_char & 255;          const char *caseless =
3111          const char *caseless = ((first_char & REQ_CASELESS) == 0)?            ((((real_pcre *)re)->flags & PCRE_FCH_CASELESS) == 0)?
3112            "" : " (caseless)";            "" : " (caseless)";
3113          if (PRINTHEX(ch))  
3114            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);          if (PRINTOK(first_char))
3115              fprintf(outfile, "First char = \'%c\'%s\n", first_char, caseless);
3116          else          else
3117            fprintf(outfile, "First char = %d%s\n", ch, caseless);            {
3118              fprintf(outfile, "First char = ");
3119              pchar(first_char, outfile);
3120              fprintf(outfile, "%s\n", caseless);
3121              }
3122          }          }
3123    
3124        if (need_char < 0)        if (need_char < 0)
# Line 1531  while (!done) Line 3127  while (!done)
3127          }          }
3128        else        else
3129          {          {
3130          int ch = need_char & 255;          const char *caseless =
3131          const char *caseless = ((need_char & REQ_CASELESS) == 0)?            ((((real_pcre *)re)->flags & PCRE_RCH_CASELESS) == 0)?
3132            "" : " (caseless)";            "" : " (caseless)";
3133          if (PRINTHEX(ch))  
3134            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);          if (PRINTOK(need_char))
3135              fprintf(outfile, "Need char = \'%c\'%s\n", need_char, caseless);
3136          else          else
3137            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            {
3138              fprintf(outfile, "Need char = ");
3139              pchar(need_char, outfile);
3140              fprintf(outfile, "%s\n", caseless);
3141              }
3142          }          }
3143    
3144        /* Don't output study size; at present it is in any case a fixed        /* Don't output study size; at present it is in any case a fixed
3145        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
3146        so messes up the test suite. (And with the /F option, it might be        so messes up the test suite. (And with the /F option, it might be
3147        flipped.) */        flipped.) If study was forced by an external -s, don't show this
3148          information unless -i or -d was also present. This means that, except
3149          when auto-callouts are involved, the output from runs with and without
3150          -s should be identical. */
3151    
3152        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
3153          {          {
3154          if (extra == NULL)          if (extra == NULL)
3155            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
3156          else          else
3157            {            {
3158            uschar *start_bits = NULL;            pcre_uint8 *start_bits = NULL;
3159            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
3160    
3161            if (start_bits == NULL)            if (new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength) == 0)
3162              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
3163            else  
3164              if (new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits) == 0)
3165              {              {
3166              int i;              if (start_bits == NULL)
3167              int c = 24;                fprintf(outfile, "No set of starting bytes\n");
3168              fprintf(outfile, "Starting byte set: ");              else
             for (i = 0; i < 256; i++)  
3169                {                {
3170                if ((start_bits[i/8] & (1<<(i&7))) != 0)                int i;
3171                  int c = 24;
3172                  fprintf(outfile, "Starting byte set: ");
3173                  for (i = 0; i < 256; i++)
3174                  {                  {
3175                  if (c > 75)                  if ((start_bits[i/8] & (1<<(i&7))) != 0)
                   {  
                   fprintf(outfile, "\n  ");  
                   c = 2;  
                   }  
                 if (PRINTHEX(i) && i != ' ')  
3176                    {                    {
3177                    fprintf(outfile, "%c ", i);                    if (c > 75)
3178                    c += 2;                      {
3179                    }                      fprintf(outfile, "\n  ");
3180                  else                      c = 2;
3181                    {                      }
3182                    fprintf(outfile, "\\x%02x ", i);                    if (PRINTOK(i) && i != ' ')
3183                    c += 5;                      {
3184                        fprintf(outfile, "%c ", i);
3185                        c += 2;
3186                        }
3187                      else
3188                        {
3189                        fprintf(outfile, "\\x%02x ", i);
3190                        c += 5;
3191                        }
3192                    }                    }
3193                  }                  }
3194                  fprintf(outfile, "\n");
3195                }                }
3196              fprintf(outfile, "\n");              }
3197              }
3198    
3199            /* Show this only if the JIT was set by /S, not by -s. */
3200    
3201            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
3202              {
3203              int jit;
3204              if (new_info(re, extra, PCRE_INFO_JIT, &jit) == 0)
3205                {
3206                if (jit)
3207                  fprintf(outfile, "JIT study was successful\n");
3208                else
3209    #ifdef SUPPORT_JIT
3210                  fprintf(outfile, "JIT study was not successful\n");
3211    #else
3212                  fprintf(outfile, "JIT support is not available in this version of PCRE\n");
3213    #endif
3214              }              }
3215            }            }
3216          }          }
# Line 1601  while (!done) Line 3229  while (!done)
3229          }          }
3230        else        else
3231          {          {
3232          uschar sbuf[8];          pcre_uint8 sbuf[8];
3233          sbuf[0] = (uschar)((true_size >> 24) & 255);  
3234          sbuf[1] = (uschar)((true_size >> 16) & 255);          if (do_flip) regexflip(re, extra);
3235          sbuf[2] = (uschar)((true_size >>  8) & 255);          sbuf[0] = (pcre_uint8)((true_size >> 24) & 255);
3236          sbuf[3] = (uschar)((true_size) & 255);          sbuf[1] = (pcre_uint8)((true_size >> 16) & 255);
3237            sbuf[2] = (pcre_uint8)((true_size >>  8) & 255);
3238          sbuf[4] = (uschar)((true_study_size >> 24) & 255);          sbuf[3] = (pcre_uint8)((true_size) & 255);
3239          sbuf[5] = (uschar)((true_study_size >> 16) & 255);          sbuf[4] = (pcre_uint8)((true_study_size >> 24) & 255);
3240          sbuf[6] = (uschar)((true_study_size >>  8) & 255);          sbuf[5] = (pcre_uint8)((true_study_size >> 16) & 255);
3241          sbuf[7] = (uschar)((true_study_size) & 255);          sbuf[6] = (pcre_uint8)((true_study_size >>  8) & 255);
3242            sbuf[7] = (pcre_uint8)((true_study_size) & 255);
3243    
3244          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
3245              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1619  while (!done) Line 3248  while (!done)
3248            }            }
3249          else          else
3250            {            {
3251            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
3252    
3253              /* If there is study data, write it. */
3254    
3255            if (extra != NULL)            if (extra != NULL)
3256              {              {
3257              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1629  while (!done) Line 3261  while (!done)
3261                  strerror(errno));                  strerror(errno));
3262                }                }
3263              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
3264              }              }
3265            }            }
3266          fclose(f);          fclose(f);
3267          }          }
3268    
3269        new_free(re);        new_free(re);
3270        if (extra != NULL) new_free(extra);        if (extra != NULL)
3271        if (tables != NULL) new_free((void *)tables);          {
3272            PCRE_FREE_STUDY(extra);
3273            }
3274          if (locale_set)
3275            {
3276            new_free((void *)tables);
3277            setlocale(LC_CTYPE, "C");
3278            locale_set = 0;
3279            }
3280        continue;  /* With next regex */        continue;  /* With next regex */
3281        }        }
3282      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1646  while (!done) Line 3285  while (!done)
3285    
3286    for (;;)    for (;;)
3287      {      {
3288      uschar *q;      pcre_uint8 *q;
3289      uschar *bptr;      pcre_uint8 *bptr;
3290      int *use_offsets = offsets;      int *use_offsets = offsets;
3291      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
3292      int callout_data = 0;      int callout_data = 0;
3293      int callout_data_set = 0;      int callout_data_set = 0;
3294      int count, c;      int count, c;
3295      int copystrings = 0;      int copystrings = 0;
3296      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
3297      int getstrings = 0;      int getstrings = 0;
3298      int getlist = 0;      int getlist = 0;
3299      int gmatched = 0;      int gmatched = 0;
3300      int start_offset = 0;      int start_offset = 0;
3301        int start_offset_sign = 1;
3302      int g_notempty = 0;      int g_notempty = 0;
3303      int use_dfa = 0;      int use_dfa = 0;
3304    
     options = 0;  
   
3305      *copynames = 0;      *copynames = 0;
3306      *getnames = 0;      *getnames = 0;
3307    
3308      copynamesptr = copynames;      cn16ptr = copynames;
3309      getnamesptr = getnames;      gn16ptr = getnames;
3310        cn8ptr = copynames8;
3311        gn8ptr = getnames8;
3312    
3313      pcre_callout = callout;      SET_PCRE_CALLOUT(callout);
3314      first_callout = 1;      first_callout = 1;
3315        last_callout_mark = NULL;
3316      callout_extra = 0;      callout_extra = 0;
3317      callout_count = 0;      callout_count = 0;
3318      callout_fail_count = 999999;      callout_fail_count = 999999;
3319      callout_fail_id = -1;      callout_fail_id = -1;
3320      show_malloc = 0;      show_malloc = 0;
3321        options = 0;
3322    
3323      if (extra != NULL) extra->flags &=      if (extra != NULL) extra->flags &=
3324        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
# Line 1684  while (!done) Line 3326  while (!done)
3326      len = 0;      len = 0;
3327      for (;;)      for (;;)
3328        {        {
3329        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
3330          {          {
3331          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
3332              {
3333              fprintf(outfile, "\n");
3334              break;
3335              }
3336          done = 1;          done = 1;
3337          goto CONTINUE;          goto CONTINUE;
3338          }          }
# Line 1708  while (!done) Line 3353  while (!done)
3353        {        {
3354        int i = 0;        int i = 0;
3355        int n = 0;        int n = 0;
3356    
3357        if (c == '\\') switch ((c = *p++))        /* In UTF mode, input can be UTF-8, so just copy all non-backslash bytes.
3358          In non-UTF mode, allow the value of the byte to fall through to later,
3359          where values greater than 127 are turned into UTF-8 when running in
3360          16-bit mode. */
3361    
3362          if (c != '\\')
3363            {
3364            if (use_utf)
3365              {
3366              *q++ = c;
3367              continue;
3368              }
3369            }
3370    
3371          /* Handle backslash escapes */
3372    
3373          else switch ((c = *p++))
3374          {          {
3375          case 'a': c =    7; break;          case 'a': c =    7; break;
3376          case 'b': c = '\b'; break;          case 'b': c = '\b'; break;
# Line 1725  while (!done) Line 3386  while (!done)
3386          c -= '0';          c -= '0';
3387          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
3388            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
   
 #if !defined NOUTF8  
         if (use_utf8 && c > 255)  
           {  
           unsigned char 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  
3389          break;          break;
3390    
3391          case 'x':          case 'x':
   
         /* Handle \x{..} specially - new Perl thing for utf8 */  
   
 #if !defined NOUTF8  
3392          if (*p == '{')          if (*p == '{')
3393            {            {
3394            unsigned char *pt = p;            pcre_uint8 *pt = p;
3395            c = 0;            c = 0;
3396            while (isxdigit(*(++pt)))  
3397              c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W');            /* We used to have "while (isxdigit(*(++pt)))" here, but it fails
3398              when isxdigit() is a macro that refers to its argument more than
3399              once. This is banned by the C Standard, but apparently happens in at
3400              least one MacOS environment. */
3401    
3402              for (pt++; isxdigit(*pt); pt++)
3403                c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'a' - 10);
3404            if (*pt == '}')            if (*pt == '}')
3405              {              {
             unsigned char buff8[8];  
             int ii, utn;  
             utn = ord2utf8(c, buff8);  
             for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];  
             c = buff8[ii];   /* Last byte */  
3406              p = pt + 1;              p = pt + 1;
3407              break;              break;
3408              }              }
3409            /* Not correct form; fall through */            /* Not correct form for \x{...}; fall through */
3410            }            }
 #endif  
3411    
3412          /* Ordinary \x */          /* \x without {} always defines just one byte in 8-bit mode. This
3413            allows UTF-8 characters to be constructed byte by byte, and also allows
3414            invalid UTF-8 sequences to be made. Just copy the byte in UTF mode.
3415            Otherwise, pass it down to later code so that it can be turned into
3416            UTF-8 when running in 16-bit mode. */
3417    
3418          c = 0;          c = 0;
3419          while (i++ < 2 && isxdigit(*p))          while (i++ < 2 && isxdigit(*p))
3420            {            {
3421            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W');            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'a' - 10);
3422            p++;            p++;
3423            }            }
3424            if (use_utf)
3425              {
3426              *q++ = c;
3427              continue;
3428              }
3429          break;          break;
3430    
3431          case 0:   /* \ followed by EOF allows for an empty line */          case 0:   /* \ followed by EOF allows for an empty line */
# Line 1778  while (!done) Line 3433  while (!done)
3433          continue;          continue;
3434    
3435          case '>':          case '>':
3436            if (*p == '-')
3437              {
3438              start_offset_sign = -1;
3439              p++;
3440              }
3441          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
3442            start_offset *= start_offset_sign;
3443          continue;          continue;
3444    
3445          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1797  while (!done) Line 3458  while (!done)
3458            }            }
3459          else if (isalnum(*p))          else if (isalnum(*p))
3460            {            {
3461            uschar *npp = copynamesptr;            READ_CAPTURE_NAME(p, &cn8ptr, &cn16ptr, re);
           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;  
3462            }            }
3463          else if (*p == '+')          else if (*p == '+')
3464            {            {
# Line 1813  while (!done) Line 3467  while (!done)
3467            }            }
3468          else if (*p == '-')          else if (*p == '-')
3469            {            {
3470            pcre_callout = NULL;            SET_PCRE_CALLOUT(NULL);
3471            p++;            p++;
3472            }            }
3473          else if (*p == '!')          else if (*p == '!')
# Line 1851  while (!done) Line 3505  while (!done)
3505  #endif  #endif
3506            use_dfa = 1;            use_dfa = 1;
3507          continue;          continue;
3508    #endif
3509    
3510    #if !defined NODFA
3511          case 'F':          case 'F':
3512          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
3513          continue;          continue;
# Line 1865  while (!done) Line 3521  while (!done)
3521            }            }
3522          else if (isalnum(*p))          else if (isalnum(*p))
3523            {            {
3524            uschar *npp = getnamesptr;            READ_CAPTURE_NAME(p, &gn8ptr, &gn16ptr, re);
3525            while (isalnum(*p)) *npp++ = *p++;            }
3526            *npp++ = 0;          continue;
3527            *npp = 0;  
3528            n = pcre_get_stringnumber(re, (char *)getnamesptr);          case 'J':
3529            if (n < 0)          while(isdigit(*p)) n = n * 10 + *p++ - '0';
3530              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);          if (extra != NULL
3531            getnamesptr = npp;              && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
3532                && extra->executable_jit != NULL)
3533              {
3534              if (jit_stack != NULL) PCRE_JIT_STACK_FREE(jit_stack);
3535              jit_stack = PCRE_JIT_STACK_ALLOC(1, n * 1024);
3536              PCRE_ASSIGN_JIT_STACK(extra, jit_callback, jit_stack);
3537            }            }
3538          continue;          continue;
3539    
# Line 1885  while (!done) Line 3546  while (!done)
3546          continue;          continue;
3547    
3548          case 'N':          case 'N':
3549          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
3550              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
3551            else
3552              options |= PCRE_NOTEMPTY;
3553          continue;          continue;
3554    
3555          case 'O':          case 'O':
# Line 1908  while (!done) Line 3572  while (!done)
3572          continue;          continue;
3573    
3574          case 'P':          case 'P':
3575          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
3576              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
3577          continue;          continue;
3578    
3579          case 'Q':          case 'Q':
# Line 1943  while (!done) Line 3608  while (!done)
3608          show_malloc = 1;          show_malloc = 1;
3609          continue;          continue;
3610    
3611            case 'Y':
3612            options |= PCRE_NO_START_OPTIMIZE;
3613            continue;
3614    
3615          case 'Z':          case 'Z':
3616          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
3617          continue;          continue;
# Line 1960  while (!done) Line 3629  while (!done)
3629            }            }
3630          continue;          continue;
3631          }          }
3632        *q++ = c;  
3633          /* We now have a character value in c that may be greater than 255. In
3634          16-bit mode, we always convert characters to UTF-8 so that values greater
3635          than 255 can be passed to non-UTF 16-bit strings. In 8-bit mode we
3636          convert to UTF-8 if we are in UTF mode. Values greater than 127 in UTF
3637          mode must have come from \x{...} or octal constructs because values from
3638          \x.. get this far only in non-UTF mode. */
3639    
3640          if (use_pcre16 || use_utf)
3641            {
3642            pcre_uint8 buff8[8];
3643            int ii, utn;
3644            utn = ord2utf8(c, buff8);
3645            for (ii = 0; ii < utn; ii++) *q++ = buff8[ii];
3646            }
3647          else
3648            {
3649            if (c > 255)
3650              {
3651              fprintf(outfile, "** Character \\x{%x} is greater than 255 "
3652                "and UTF-8 mode is not enabled.\n", c);
3653              fprintf(outfile, "** Truncation will probably give the wrong "
3654                "result.\n");
3655              }
3656            *q++ = c;
3657            }
3658        }        }
3659    
3660        /* Reached end of subject string */
3661    
3662      *q = 0;      *q = 0;
3663      len = q - dbuffer;      len = (int)(q - dbuffer);
3664    
3665        /* Move the data to the end of the buffer so that a read over the end of
3666        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
3667        we are using the POSIX interface, we must include the terminating zero. */
3668    
3669    #if !defined NOPOSIX
3670        if (posix || do_posix)
3671          {
3672          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
3673          bptr += buffer_size - len - 1;
3674          }
3675        else
3676    #endif
3677          {
3678          memmove(bptr + buffer_size - len, bptr, len);
3679          bptr += buffer_size - len;
3680          }
3681    
3682      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
3683        {        {
# Line 1984  while (!done) Line 3698  while (!done)
3698          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
3699        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
3700        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
3701          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
3702    
3703        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
3704    
# Line 2005  while (!done) Line 3720  while (!done)
3720            if (pmatch[i].rm_so >= 0)            if (pmatch[i].rm_so >= 0)
3721              {              {
3722              fprintf(outfile, "%2d: ", (int)i);              fprintf(outfile, "%2d: ", (int)i);
3723              (void)pchars(dbuffer + pmatch[i].rm_so,              PCHARSV(dbuffer, pmatch[i].rm_so,
3724                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
3725              fprintf(outfile, "\n");              fprintf(outfile, "\n");
3726              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
3727                {                {
3728                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
3729                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                PCHARSV(dbuffer, pmatch[i].rm_eo, len - pmatch[i].rm_eo,
3730                  outfile);                  outfile);
3731                fprintf(outfile, "\n");                fprintf(outfile, "\n");
3732                }                }
# Line 2019  while (!done) Line 3734  while (!done)
3734            }            }
3735          }          }
3736        free(pmatch);        free(pmatch);
3737          goto NEXT_DATA;
3738        }        }
3739    
3740    #endif  /* !defined NOPOSIX */
3741    
3742      /* Handle matching via the native interface - repeats for /g and /G */      /* Handle matching via the native interface - repeats for /g and /G */
3743    
3744      else  #ifdef SUPPORT_PCRE16
3745  #endif  /* !defined NOPOSIX */      if (use_pcre16)
3746          {
3747          len = to16(TRUE, bptr, (((real_pcre *)re)->options) & PCRE_UTF8, len);
3748          switch(len)
3749            {
3750            case -1:
3751            fprintf(outfile, "**Failed: invalid UTF-8 string cannot be "
3752              "converted to UTF-16\n");
3753            goto NEXT_DATA;
3754    
3755            case -2:
3756            fprintf(outfile, "**Failed: character value greater than 0x10ffff "
3757              "cannot be converted to UTF-16\n");
3758            goto NEXT_DATA;
3759    
3760            case -3:
3761            fprintf(outfile, "**Failed: character value greater than 0xffff "
3762              "cannot be converted to 16-bit in non-UTF mode\n");
3763            goto NEXT_DATA;
3764    
3765            default:
3766            break;
3767            }
3768          bptr = (pcre_uint8 *)buffer16;
3769          }
3770    #endif
3771    
3772      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
3773        {        {
3774          markptr = NULL;
3775    
3776        if (timeitm > 0)        if (timeitm > 0)
3777          {          {
3778          register int i;          register int i;
# Line 2039  while (!done) Line 3784  while (!done)
3784            {            {
3785            int workspace[1000];            int workspace[1000];
3786            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
3787              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              {
3788                options | g_notempty, use_offsets, use_size_offsets, workspace,              PCRE_DFA_EXEC(count, re, extra, bptr, len, start_offset,
3789                sizeof(workspace)/sizeof(int));                (options | g_notempty), use_offsets, use_size_offsets, workspace,
3790                  (sizeof(workspace)/sizeof(int)));
3791                }
3792            }            }
3793          else          else
3794  #endif  #endif
3795    
3796          for (i = 0; i < timeitm; i++)          for (i = 0; i < timeitm; i++)
3797            count = pcre_exec(re, extra, (char *)bptr, len,            {
3798              start_offset, options | g_notempty, use_offsets, use_size_offsets);            PCRE_EXEC(count, re, extra, bptr, len, start_offset,
3799                (options | g_notempty), use_offsets, use_size_offsets);
3800              }
3801          time_taken = clock() - start_time;          time_taken = clock() - start_time;
3802          fprintf(outfile, "Execute time %.4f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
3803            (((double)time_taken * 1000.0) / (double)timeitm) /            (((double)time_taken * 1000.0) / (double)timeitm) /
# Line 2058  while (!done) Line 3806  while (!done)
3806    
3807        /* If find_match_limit is set, we want to do repeated matches with        /* If find_match_limit is set, we want to do repeated matches with
3808        varying limits in order to find the minimum value for the match limit and        varying limits in order to find the minimum value for the match limit and
3809        for the recursion limit. */        for the recursion limit. The match limits are relevant only to the normal
3810          running of pcre_exec(), so disable the JIT optimization. This makes it
3811          possible to run the same set of tests with and without JIT externally
3812          requested. */
3813    
3814        if (find_match_limit)        if (find_match_limit)
3815          {          {
# Line 2067  while (!done) Line 3818  while (!done)
3818            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
3819            extra->flags = 0;            extra->flags = 0;
3820            }            }
3821            else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
3822    
3823          (void)check_match_limit(re, extra, bptr, len, start_offset,          (void)check_match_limit(re, extra, bptr, len, start_offset,
3824            options|g_notempty, use_offsets, use_size_offsets,            options|g_notempty, use_offsets, use_size_offsets,
# Line 2090  while (!done) Line 3842  while (!done)
3842            }            }
3843          extra->flags |= PCRE_EXTRA_CALLOUT_DATA;          extra->flags |= PCRE_EXTRA_CALLOUT_DATA;
3844          extra->callout_data = &callout_data;          extra->callout_data = &callout_data;
3845          count = pcre_exec(re, extra, (char *)bptr, len, start_offset,          PCRE_EXEC(count, re, extra, bptr, len, start_offset,
3846            options | g_notempty, use_offsets, use_size_offsets);            options | g_notempty, use_offsets, use_size_offsets);
3847          extra->flags &= ~PCRE_EXTRA_CALLOUT_DATA;          extra->flags &= ~PCRE_EXTRA_CALLOUT_DATA;
3848          }          }
# Line 2102  while (!done) Line 3854  while (!done)
3854        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
3855          {          {
3856          int workspace[1000];          int workspace[1000];
3857          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          PCRE_DFA_EXEC(count, re, extra, bptr, len, start_offset,
3858            options | g_notempty, use_offsets, use_size_offsets, workspace,            (options | g_notempty), use_offsets, use_size_offsets, workspace,
3859            sizeof(workspace)/sizeof(int));            (sizeof(workspace)/sizeof(int)));
3860          if (count == 0)          if (count == 0)
3861            {            {
3862            fprintf(outfile, "Matched, but too many subsidiary matches\n");            fprintf(outfile, "Matched, but too many subsidiary matches\n");
# Line 2115  while (!done) Line 3867  while (!done)
3867    
3868        else        else
3869          {          {
3870          count = pcre_exec(re, extra, (char *)bptr, len,          PCRE_EXEC(count, re, extra, bptr, len, start_offset,
3871            start_offset, options | g_notempty, use_offsets, use_size_offsets);            options | g_notempty, use_offsets, use_size_offsets);
3872          if (count == 0)          if (count == 0)
3873            {            {
3874            fprintf(outfile, "Matched, but too many substrings\n");            fprintf(outfile, "Matched, but too many substrings\n");
# Line 2129  while (!done) Line 3881  while (!done)
3881        if (count >= 0)        if (count >= 0)
3882          {          {
3883          int i, maxcount;          int i, maxcount;
3884            void *cnptr, *gnptr;
3885    
3886  #if !defined NODFA  #if !defined NODFA
3887          if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else          if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
# Line 2150  while (!done) Line 3903  while (!done)
3903              }              }
3904            }            }
3905    
3906            /* do_allcaps requests showing of all captures in the pattern, to check
3907            unset ones at the end. */
3908    
3909            if (do_allcaps)
3910              {
3911              if (new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count) < 0)
3912                goto SKIP_DATA;
3913              count++;   /* Allow for full match */
3914              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
3915              }
3916    
3917            /* Output the captured substrings */
3918    
3919          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
3920            {            {
3921            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
3922                {
3923                if (use_offsets[i] != -1)
3924                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
3925                    use_offsets[i], i);
3926                if (use_offsets[i+1] != -1)
3927                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
3928                    use_offsets[i+1], i+1);
3929              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
3930                }
3931            else            else
3932              {              {
3933              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
3934              (void)pchars(bptr + use_offsets[i],              PCHARSV(bptr, use_offsets[i],
3935                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
3936              fprintf(outfile, "\n");              fprintf(outfile, "\n");
3937              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
3938                {                {
3939                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
3940                  {                PCHARSV(bptr, use_offsets[i+1], len - use_offsets[i+1],
3941                  fprintf(outfile, " 0+ ");                  outfile);
3942                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
3943                }                }
3944              }              }
3945            }            }
3946    
3947            if (markptr != NULL)
3948              {
3949              fprintf(outfile, "MK: ");
3950              PCHARSV(markptr, 0, -1, outfile);
3951              fprintf(outfile, "\n");
3952              }
3953    
3954          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
3955            {            {
3956            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
3957              {              {
3958                int rc;
3959              char copybuffer[256];              char copybuffer[256];
3960              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              PCRE_COPY_SUBSTRING(rc, bptr, use_offsets, count, i,
3961                i, copybuffer, sizeof(copybuffer));                copybuffer, sizeof(copybuffer));
3962              if (rc < 0)              if (rc < 0)
3963                fprintf(outfile, "copy substring %d failed %d\n", i, rc);                fprintf(outfile, "copy substring %d failed %d\n", i, rc);
3964              else              else
3965                fprintf(outfile, "%2dC %s (%d)\n", i, copybuffer, rc);                {
3966                  fprintf(outfile, "%2dC ", i);
3967                  PCHARSV(copybuffer, 0, rc, outfile);
3968                  fprintf(outfile, " (%d)\n", rc);
3969                  }
3970              }              }
3971            }            }
3972    
3973          for (copynamesptr = copynames;          cnptr = copynames;
3974               *copynamesptr != 0;          for (;;)
              copynamesptr += (int)strlen((char*)copynamesptr) + 1)  
3975            {            {
3976              int rc;
3977            char copybuffer[256];            char copybuffer[256];
3978            int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,  
3979              count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));            if (use_pcre16)
3980                {
3981                if (*(pcre_uint16 *)cnptr == 0) break;
3982                }
3983              else
3984                {
3985                if (*(pcre_uint8 *)cnptr == 0) break;
3986                }
3987    
3988              PCRE_COPY_NAMED_SUBSTRING(rc, re, bptr, use_offsets, count,
3989                cnptr, copybuffer, sizeof(copybuffer));
3990    
3991            if (rc < 0)            if (rc < 0)
3992              fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);              {
3993                fprintf(outfile, "copy substring ");
3994                PCHARSV(cnptr, 0, -1, outfile);
3995                fprintf(outfile, " failed %d\n", rc);
3996                }
3997            else            else
3998              fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);              {
3999                fprintf(outfile, "  C ");
4000                PCHARSV(copybuffer, 0, rc, outfile);
4001                fprintf(outfile, " (%d) ", rc);
4002                PCHARSV(cnptr, 0, -1, outfile);
4003                putc('\n', outfile);
4004                }
4005    
4006              cnptr = (char *)cnptr + (STRLEN(cnptr) + 1) * CHAR_SIZE;
4007            }            }
4008    
4009          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
4010            {            {
4011            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
4012              {              {
4013                int rc;
4014              const char *substring;              const char *substring;
4015              int rc = pcre_get_substring((char *)bptr, use_offsets, count,              PCRE_GET_SUBSTRING(rc, bptr, use_offsets, count, i, &substring);
               i, &substring);  
4016              if (rc < 0)              if (rc < 0)
4017                fprintf(outfile, "get substring %d failed %d\n", i, rc);                fprintf(outfile, "get substring %d failed %d\n", i, rc);
4018              else              else
4019                {                {
4020                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG ", i);
4021                pcre_free_substring(substring);                PCHARSV(substring, 0, rc, outfile);
4022                  fprintf(outfile, " (%d)\n", rc);
4023                  PCRE_FREE_SUBSTRING(substring);
4024                }                }
4025              }              }
4026            }            }
4027    
4028          for (getnamesptr = getnames;          gnptr = getnames;
4029               *getnamesptr != 0;          for (;;)
              getnamesptr += (int)strlen((char*)getnamesptr) + 1)  
4030            {            {
4031              int rc;
4032            const char *substring;            const char *substring;
4033            int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,  
4034              count, (char *)getnamesptr, &substring);            if (use_pcre16)
4035                {
4036                if (*(pcre_uint16 *)gnptr == 0) break;
4037                }
4038              else
4039                {
4040                if (*(pcre_uint8 *)gnptr == 0) break;
4041                }
4042    
4043              PCRE_GET_NAMED_SUBSTRING(rc, re, bptr, use_offsets, count,
4044                gnptr, &substring);
4045            if (rc < 0)            if (rc < 0)
4046              fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);              {
4047                fprintf(outfile, "get substring ");
4048                PCHARSV(gnptr, 0, -1, outfile);
4049                fprintf(outfile, " failed %d\n", rc);
4050                }
4051            else            else
4052              {              {
4053              fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);              fprintf(outfile, "  G ");
4054              pcre_free_substring(substring);              PCHARSV(substring, 0, rc, outfile);
4055                fprintf(outfile, " (%d) ", rc);