/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 520 by ph10, Sat May 22 18:54:05 2010 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.10 03 May-2010
5    ------------------------
6    
7    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
8        THEN.
9    
10    2.  (*ACCEPT) was not working when inside an atomic group.
11    
12    3.  Inside a character class, \B is treated as a literal by default, but
13        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
14        causes the error). The code is unchanged, but I tidied the documentation.
15    
16    4.  Inside a character class, PCRE always treated \R and \X as literals,
17        whereas Perl faults them if its -w option is set. I have changed PCRE so
18        that it faults them when PCRE_EXTRA is set.
19    
20    5.  Added support for \N, which always matches any character other than
21        newline. (It is the same as "." when PCRE_DOTALL is not set.)
22    
23    6.  When compiling pcregrep with newer versions of gcc which may have
24        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
25        declared with attribute warn_unused_result" were given. Just casting the
26        result to (void) does not stop the warnings; a more elaborate fudge is
27        needed. I've used a macro to implement this.
28    
29    7.  Minor change to pcretest.c to avoid a compiler warning.
30    
31    8.  Added four artifical Unicode properties to help with an option to make
32        \s etc use properties (see next item). The new properties are: Xan
33        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
34    
35    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
36        use Unicode properties. (*UCP) at the start of a pattern can be used to set
37        this option. Modified pcretest to add /W to test this facility. Added
38        REG_UCP to make it available via the POSIX interface.
39    
40    10. Added --line-buffered to pcregrep.
41    
42    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
43        studied, and the match started with a letter with a code point greater than
44        127 whose first byte was different to the first byte of the other case of
45        the letter, the other case of this starting letter was not recognized.
46    
47    
48    Version 8.02 19-Mar-2010
49    ------------------------
50    
51    1.  The Unicode data tables have been updated to Unicode 5.2.0.
52    
53    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
54        configured.
55    
56    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
57        original author of that file, following a query about its status.
58    
59    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
60        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
61    
62    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
63        quantifier applied to a forward-referencing subroutine call, could compile
64        incorrect code or give the error "internal error: previously-checked
65        referenced subpattern not found".
66    
67    6.  Both MS Visual Studio and Symbian OS have problems with initializing
68        variables to point to external functions. For these systems, therefore,
69        pcre_malloc etc. are now initialized to local functions that call the
70        relevant global functions.
71    
72    7.  There were two entries missing in the vectors called coptable and poptable
73        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
74        I've fixed the data, and added a kludgy way of testing at compile time that
75        the lengths are correct (equal to the number of opcodes).
76    
77    8.  Following on from 7, I added a similar kludge to check the length of the
78        eint vector in pcreposix.c.
79    
80    9.  Error texts for pcre_compile() are held as one long string to avoid too
81        much relocation at load time. To find a text, the string is searched,
82        counting zeros. There was no check for running off the end of the string,
83        which could happen if a new error number was added without updating the
84        string.
85    
86    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
87    
88    11. \K was not working if it appeared in an atomic group or in a group that
89        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
90        \K is "not well defined" if used in an assertion. PCRE now accepts it if
91        the assertion is positive, but not if it is negative.
92    
93    12. Change 11 fortuitously reduced the size of the stack frame used in the
94        "match()" function of pcre_exec.c by one pointer. Forthcoming
95        implementation of support for (*MARK) will need an extra pointer on the
96        stack; I have reserved it now, so that the stack frame size does not
97        decrease.
98    
99    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
100        item in branch that calls a recursion is a subroutine call - as in the
101        second branch in the above example - was incorrectly given the compile-
102        time error "recursive call could loop indefinitely" because pcre_compile()
103        was not correctly checking the subroutine for matching a non-empty string.
104    
105    14. The checks for overrunning compiling workspace could trigger after an
106        overrun had occurred. This is a "should never occur" error, but it can be
107        triggered by pathological patterns such as hundreds of nested parentheses.
108        The checks now trigger 100 bytes before the end of the workspace.
109    
110    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
111    
112    
113    Version 8.01 19-Jan-2010
114    ------------------------
115    
116    1.  If a pattern contained a conditional subpattern with only one branch (in
117        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
118        computed the wrong minimum data length (which is of course zero for such
119        subpatterns). This could cause incorrect "no match" results.
120    
121    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
122        the pattern is reset in the first branch, pcre_compile() failed with
123        "internal error: code overflow at offset...". This happened only when
124        the reset was to the original external option setting. (An optimization
125        abstracts leading options settings into an external setting, which was the
126        cause of this.)
127    
128    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
129        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
130        assertion pattern did not match (meaning that the assertion was true), it
131        was incorrectly treated as false if the SKIP had been reached during the
132        matching. This also applied to assertions used as conditions.
133    
134    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
135        assertion subpattern, including such a pattern used as a condition,
136        unpredictable results occurred, instead of the error return
137        PCRE_ERROR_DFA_UITEM.
138    
139    5.  The C++ GlobalReplace function was not working like Perl for the special
140        situation when an empty string is matched. It now does the fancy magic
141        stuff that is necessary.
142    
143    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
144        removed. (These were left over from very, very early versions of PCRE.)
145    
146    7.  Some cosmetic changes to the code to make life easier when compiling it
147        as part of something else:
148    
149        (a) Change DEBUG to PCRE_DEBUG.
150    
151        (b) In pcre_compile(), rename the member of the "branch_chain" structure
152            called "current" as "current_branch", to prevent a collision with the
153            Linux macro when compiled as a kernel module.
154    
155        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
156            prevent a collision with the Linux macro when compiled as a kernel
157            module.
158    
159    8.  In pcre_compile() there are some checks for integer overflows that used to
160        cast potentially large values to (double). This has been changed to that
161        when building, a check for int64_t is made, and if it is found, it is used
162        instead, thus avoiding the use of floating point arithmetic. (There is no
163        other use of FP in PCRE.) If int64_t is not found, the fallback is to
164        double.
165    
166    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
167        2005 (difference between two addresses compared to an unsigned value).
168    
169    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
170        custom one, because of the following reported problem in Windows:
171    
172          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
173              under Win32.
174          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
175              therefore missing the function definition.
176          - The compiler thus generates a "C" signature for the test function.
177          - The linker fails to find the "C" function.
178          - PCRE fails to configure if asked to do so against libbz2.
179    
180    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
181        messages were output:
182    
183          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
184          rerunning libtoolize, to keep the correct libtool macros in-tree.
185          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
186    
187        I have done both of these things.
188    
189    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
190        most of the time, it *can* run out if it is given a pattern that contains a
191        runaway infinite recursion. I updated the discussion in the pcrestack man
192        page.
193    
194    13. Now that we have gone to the x.xx style of version numbers, the minor
195        version may start with zero. Using 08 or 09 is a bad idea because users
196        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
197        interpreted as invalid octal numbers. I've updated the previous comment in
198        configure.ac, and also added a check that gives an error if 08 or 09 are
199        used.
200    
201    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
202        causing partial matching to fail when the end of the subject matched \W
203        in a UTF-8 pattern where \W was quantified with a minimum of 3.
204    
205    15. There were some discrepancies between the declarations in pcre_internal.h
206        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
207        their definitions. The declarations used "const uschar *" and the
208        definitions used USPTR. Even though USPTR is normally defined as "const
209        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
210        reported that: "This difference in casting confuses some C++ compilers, for
211        example, SunCC recognizes above declarations as different functions and
212        generates broken code for hbpcre." I have changed the declarations to use
213        USPTR.
214    
215    16. GNU libtool is named differently on some systems. The autogen.sh script now
216        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
217        (FreeBSD).
218    
219    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
220        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
221        comment: "Figure out how to create a longlong from a string: strtoll and
222        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
223        instance, but it only takes 2 args instead of 3!"
224    
225    18. A subtle bug concerned with back references has been fixed by a change of
226        specification, with a corresponding code fix. A pattern such as
227        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
228        refers, was giving matches when it shouldn't. For example, xa=xaaa would
229        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
230        same bug. Such groups have to be quantified to be useful, or contained
231        inside another quantified group. (If there's no repetition, the reference
232        can never match.) The problem arises because, having left the group and
233        moved on to the rest of the pattern, a later failure that backtracks into
234        the group uses the captured value from the final iteration of the group
235        rather than the correct earlier one. I have fixed this in PCRE by forcing
236        any group that contains a reference to itself to be an atomic group; that
237        is, there cannot be any backtracking into it once it has completed. This is
238        similar to recursive and subroutine calls.
239    
240    
241    Version 8.00 19-Oct-09
242    ----------------------
243    
244    1.  The table for translating pcre_compile() error codes into POSIX error codes
245        was out-of-date, and there was no check on the pcre_compile() error code
246        being within the table. This could lead to an OK return being given in
247        error.
248    
249    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
250        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
251        in a Windows environment.
252    
253    3.  The pcregrep --count option prints the count for each file even when it is
254        zero, as does GNU grep. However, pcregrep was also printing all files when
255        --files-with-matches was added. Now, when both options are given, it prints
256        counts only for those files that have at least one match. (GNU grep just
257        prints the file name in this circumstance, but including the count seems
258        more useful - otherwise, why use --count?) Also ensured that the
259        combination -clh just lists non-zero counts, with no names.
260    
261    4.  The long form of the pcregrep -F option was incorrectly implemented as
262        --fixed_strings instead of --fixed-strings. This is an incompatible change,
263        but it seems right to fix it, and I didn't think it was worth preserving
264        the old behaviour.
265    
266    5.  The command line items --regex=pattern and --regexp=pattern were not
267        recognized by pcregrep, which required --regex pattern or --regexp pattern
268        (with a space rather than an '='). The man page documented the '=' forms,
269        which are compatible with GNU grep; these now work.
270    
271    6.  No libpcreposix.pc file was created for pkg-config; there was just
272        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
273    
274    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
275        when UCP support is not needed, by modifying the Python script that
276        generates it from Unicode data files. This should not matter if the module
277        is correctly used as a library, but I received one complaint about 50K of
278        unwanted data. My guess is that the person linked everything into his
279        program rather than using a library. Anyway, it does no harm.
280    
281    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
282        was a minimum greater than 1 for a wide character in a possessive
283        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
284        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
285        character. Chaos in the form of incorrect output or a compiling loop could
286        result.
287    
288    9.  The restrictions on what a pattern can contain when partial matching is
289        requested for pcre_exec() have been removed. All patterns can now be
290        partially matched by this function. In addition, if there are at least two
291        slots in the offset vector, the offset of the earliest inspected character
292        for the match and the offset of the end of the subject are set in them when
293        PCRE_ERROR_PARTIAL is returned.
294    
295    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
296        synonymous with PCRE_PARTIAL, for backwards compatibility, and
297        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
298        and may be more useful for multi-segment matching.
299    
300    11. Partial matching with pcre_exec() is now more intuitive. A partial match
301        used to be given if ever the end of the subject was reached; now it is
302        given only if matching could not proceed because another character was
303        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
304        string "Z", which now yields "no match" instead of "partial match". In the
305        case of pcre_dfa_exec(), "no match" is given if every matching path for the
306        final character ended with (*FAIL).
307    
308    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
309        if the pattern had a "must contain" character that was already found in the
310        earlier partial match, unless partial matching was again requested. For
311        example, with the pattern /dog.(body)?/, the "must contain" character is
312        "g". If the first part-match was for the string "dog", restarting with
313        "sbody" failed. This bug has been fixed.
314    
315    13. The string returned by pcre_dfa_exec() after a partial match has been
316        changed so that it starts at the first inspected character rather than the
317        first character of the match. This makes a difference only if the pattern
318        starts with a lookbehind assertion or \b or \B (\K is not supported by
319        pcre_dfa_exec()). It's an incompatible change, but it makes the two
320        matching functions compatible, and I think it's the right thing to do.
321    
322    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
323        so that the demonstration program is easily available in environments where
324        PCRE has not been installed from source.
325    
326    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
327        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
328        library.
329    
330    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
331        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
332        is not the first non-POSIX option to be added. Clearly some people find
333        these options useful.
334    
335    17. If a caller to the POSIX matching function regexec() passes a non-zero
336        value for nmatch with a NULL value for pmatch, the value of
337        nmatch is forced to zero.
338    
339    18. RunGrepTest did not have a test for the availability of the -u option of
340        the diff command, as RunTest does. It now checks in the same way as
341        RunTest, and also checks for the -b option.
342    
343    19. If an odd number of negated classes containing just a single character
344        interposed, within parentheses, between a forward reference to a named
345        subpattern and the definition of the subpattern, compilation crashed with
346        an internal error, complaining that it could not find the referenced
347        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
348        [The bug was that it was starting one character too far in when skipping
349        over the character class, thus treating the ] as data rather than
350        terminating the class. This meant it could skip too much.]
351    
352    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
353        /g option in pcretest when the pattern contains \K, which makes it possible
354        to have an empty string match not at the start, even when the pattern is
355        anchored. Updated pcretest and pcredemo to use this option.
356    
357    21. If the maximum number of capturing subpatterns in a recursion was greater
358        than the maximum at the outer level, the higher number was returned, but
359        with unset values at the outer level. The correct (outer level) value is
360        now given.
361    
362    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
363        PCRE did not set those parentheses (unlike Perl). I have now found a way to
364        make it do so. The string so far is captured, making this feature
365        compatible with Perl.
366    
367    23. The tests have been re-organized, adding tests 11 and 12, to make it
368        possible to check the Perl 5.10 features against Perl 5.10.
369    
370    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
371        pattern matches a fixed length string. PCRE did not allow this; now it
372        does. Neither allows recursion.
373    
374    25. I finally figured out how to implement a request to provide the minimum
375        length of subject string that was needed in order to match a given pattern.
376        (It was back references and recursion that I had previously got hung up
377        on.) This code has now been added to pcre_study(); it finds a lower bound
378        to the length of subject needed. It is not necessarily the greatest lower
379        bound, but using it to avoid searching strings that are too short does give
380        some useful speed-ups. The value is available to calling programs via
381        pcre_fullinfo().
382    
383    26. While implementing 25, I discovered to my embarrassment that pcretest had
384        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
385        study optimizations had never been tested with that matching function.
386        Oops. What is worse, even when it was passed study data, there was a bug in
387        pcre_dfa_exec() that meant it never actually used it. Double oops. There
388        were also very few tests of studied patterns with pcre_dfa_exec().
389    
390    27. If (?| is used to create subpatterns with duplicate numbers, they are now
391        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
392        on the other side of the coin, they are no longer allowed to have different
393        names, because these cannot be distinguished in PCRE, and this has caused
394        confusion. (This is a difference from Perl.)
395    
396    28. When duplicate subpattern names are present (necessarily with different
397        numbers, as required by 27 above), and a test is made by name in a
398        conditional pattern, either for a subpattern having been matched, or for
399        recursion in such a pattern, all the associated numbered subpatterns are
400        tested, and the overall condition is true if the condition is true for any
401        one of them. This is the way Perl works, and is also more like the way
402        testing by number works.
403    
404    
405    Version 7.9 11-Apr-09
406    ---------------------
407    
408    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
409        (pcretest), all targets were linked against these libraries. This included
410        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
411        libraries. This caused unwanted dependencies to be created. This problem
412        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
413        pcretest is linked with readline.
414    
415    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
416        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
417        moved outside it again, because FALSE and TRUE are already defined in AIX,
418        but BOOL is not.
419    
420    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
421        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
422    
423    4.  The pcregrep documentation said spaces were inserted as well as colons (or
424        hyphens) following file names and line numbers when outputting matching
425        lines. This is not true; no spaces are inserted. I have also clarified the
426        wording for the --colour (or --color) option.
427    
428    5.  In pcregrep, when --colour was used with -o, the list of matching strings
429        was not coloured; this is different to GNU grep, so I have changed it to be
430        the same.
431    
432    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
433        each matching line was coloured. Now it goes on to look for further matches
434        of any of the test patterns, which is the same behaviour as GNU grep.
435    
436    7.  A pattern that could match an empty string could cause pcregrep to loop; it
437        doesn't make sense to accept an empty string match in pcregrep, so I have
438        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
439        seems to be how GNU grep behaves.
440    
441    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
442        start or after a newline", because the conditional assertion was not being
443        correctly handled. The rule now is that both the assertion and what follows
444        in the first alternative must satisfy the test.
445    
446    9.  If auto-callout was enabled in a pattern with a conditional group whose
447        condition was an assertion, PCRE could crash during matching, both with
448        pcre_exec() and pcre_dfa_exec().
449    
450    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
451        used for matching.
452    
453    11. Unicode property support in character classes was not working for
454        characters (bytes) greater than 127 when not in UTF-8 mode.
455    
456    12. Added the -M command line option to pcretest.
457    
458    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
459    
460    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
461    
462    16. Added comments and documentation about mis-use of no_arg in the C++
463        wrapper.
464    
465    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
466        from Martin Jerabek that uses macro names for all relevant character and
467        string constants.
468    
469    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
470        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
471        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
472        these, but not everybody uses configure.
473    
474    19. A conditional group that had only one branch was not being correctly
475        recognized as an item that could match an empty string. This meant that an
476        enclosing group might also not be so recognized, causing infinite looping
477        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
478        with the subject "ab", where knowledge that the repeated group can match
479        nothing is needed in order to break the loop.
480    
481    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
482        exec(), but without supplying a callout function, matching went wrong.
483    
484    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
485        leak if the size of the offset vector was greater than 30. When the vector
486        is smaller, the saved offsets during recursion go onto a local stack
487        vector, but for larger vectors malloc() is used. It was failing to free
488        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
489        error, in fact).
490    
491    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
492        heapframe that is used only when UTF-8 support is enabled. This caused no
493        problem, but was untidy.
494    
495    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
496        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
497        included within another project.
498    
499    24. Steven Van Ingelgem's patches to add more options to the CMake support,
500        slightly modified by me:
501    
502          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
503              not building pcregrep.
504    
505          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
506              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
507    
508    25. Forward references, both numeric and by name, in patterns that made use of
509        duplicate group numbers, could behave incorrectly or give incorrect errors,
510        because when scanning forward to find the reference group, PCRE was not
511        taking into account the duplicate group numbers. A pattern such as
512        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
513    
514    26. Changed a few more instances of "const unsigned char *" to USPTR, making
515        the feature of a custom pointer more persuasive (as requested by a user).
516    
517    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
518        pcretest.c, inside #ifndefs, because it seems they are sometimes already
519        pre-defined.
520    
521    28. Added support for (*UTF8) at the start of a pattern.
522    
523    29. Arrange for flags added by the "release type" setting in CMake to be shown
524        in the configuration summary.
525    
526    
527    Version 7.8 05-Sep-08
528    ---------------------
529    
530    1.  Replaced UCP searching code with optimized version as implemented for Ad
531        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
532        stage table and inline lookup instead of a function, giving speed ups of 2
533        to 5 times on some simple patterns that I tested. Permission was given to
534        distribute the MultiStage2.py script that generates the tables (it's not in
535        the tarball, but is in the Subversion repository).
536    
537    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
538        scripts.
539    
540    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
541        a group with a zero qualifier. The result of the study could be incorrect,
542        or the function might crash, depending on the pattern.
543    
544    4.  Caseless matching was not working for non-ASCII characters in back
545        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
546        It now works when Unicode Property Support is available.
547    
548    5.  In pcretest, an escape such as \x{de} in the data was always generating
549        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
550        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
551        truncation.
552    
553    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
554    
555    7.  Added two (int) casts to pcregrep when printing the difference of two
556        pointers, in case they are 64-bit values.
557    
558    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
559        test 2 if it fails.
560    
561    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
562        and a #define of that name to empty if it is not externally set. This is to
563        allow users of MSVC to set it if necessary.
564    
565    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
566        the convenience functions in the pcre_get.c source file.
567    
568    11. An option change at the start of a pattern that had top-level alternatives
569        could cause overwriting and/or a crash. This command provoked a crash in
570        some environments:
571    
572          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
573    
574        This potential security problem was recorded as CVE-2008-2371.
575    
576    12. For a pattern where the match had to start at the beginning or immediately
577        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
578        pcre_dfa_exec() could read past the end of the passed subject if there was
579        no match. To help with detecting such bugs (e.g. with valgrind), I modified
580        pcretest so that it places the subject at the end of its malloc-ed buffer.
581    
582    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
583        exec() might read past the end of the data buffer in UTF-8 mode.
584    
585    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
586        the data contained the byte 0x85 as part of a UTF-8 character within its
587        first line. This applied both to normal and DFA matching.
588    
589    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
590        /^[^d]*?$/8 failed to match "abc".
591    
592    16. Added a missing copyright notice to pcrecpp_internal.h.
593    
594    17. Make it more clear in the documentation that values returned from
595        pcre_exec() in ovector are byte offsets, not character counts.
596    
597    18. Tidied a few places to stop certain compilers from issuing warnings.
598    
599    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
600        supplied by Stefan Weber. I made a further small update for 7.8 because
601        there is a change of source arrangements: the pcre_searchfuncs.c module is
602        replaced by pcre_ucd.c.
603    
604    
605    Version 7.7 07-May-08
606    ---------------------
607    
608    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
609        a string to a long long, pretend we don't even have a long long." This is
610        done by checking for the strtoq, strtoll, and _strtoi64 functions.
611    
612    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
613        pre-7.6 versions, which defined a global no_arg variable instead of putting
614        it in the RE class. (See also #8 below.)
615    
616    3.  Remove a line of dead code, identified by coverity and reported by Nuno
617        Lopes.
618    
619    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
620    
621        (1) The include/exclude patterns were being applied to the whole pathnames
622            of files, instead of just to the final components.
623    
624        (2) If there was more than one level of directory, the subdirectories were
625            skipped unless they satisfied the include/exclude conditions. This is
626            inconsistent with GNU grep (and could even be seen as contrary to the
627            pcregrep specification - which I improved to make it absolutely clear).
628            The action now is always to scan all levels of directory, and just
629            apply the include/exclude patterns to regular files.
630    
631    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
632        --exclude_dir in the tests to avoid scanning .svn directories.
633    
634    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
635        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
636        doesn't support NULs in patterns.
637    
638    7.  Added some missing "const"s to declarations of static tables in
639        pcre_compile.c and pcre_dfa_exec.c.
640    
641    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
642        caused by fix #2  above. (Subsequently also a second patch to fix the
643        first patch. And a third patch - this was a messy problem.)
644    
645    9.  Applied Craig's patch to remove the use of push_back().
646    
647    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
648        matching function regexec().
649    
650    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
651        which, however, unlike Perl's \g{...}, are subroutine calls, not back
652        references. PCRE supports relative numbers with this syntax (I don't think
653        Oniguruma does).
654    
655    12. Previously, a group with a zero repeat such as (...){0} was completely
656        omitted from the compiled regex. However, this means that if the group
657        was called as a subroutine from elsewhere in the pattern, things went wrong
658        (an internal error was given). Such groups are now left in the compiled
659        pattern, with a new opcode that causes them to be skipped at execution
660        time.
661    
662    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
663        to the way PCRE behaves:
664    
665        (a) A lone ] character is dis-allowed (Perl treats it as data).
666    
667        (b) A back reference to an unmatched subpattern matches an empty string
668            (Perl fails the current match path).
669    
670        (c) A data ] in a character class must be notated as \] because if the
671            first data character in a class is ], it defines an empty class. (In
672            Perl it is not possible to have an empty class.) The empty class []
673            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
674            The negative empty class [^] matches any one character, independently
675            of the DOTALL setting.
676    
677    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
678        non-existent subpattern following a character class starting with ']' and
679        containing () gave an internal compiling error instead of "reference to
680        non-existent subpattern". Fortunately, when the pattern did exist, the
681        compiled code was correct. (When scanning forwards to check for the
682        existencd of the subpattern, it was treating the data ']' as terminating
683        the class, so got the count wrong. When actually compiling, the reference
684        was subsequently set up correctly.)
685    
686    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
687        it was being rejected as not supported by pcre_dfa_exec(), even though
688        other assertions are supported. I have made pcre_dfa_exec() support
689        (*FAIL).
690    
691    16. The implementation of 13c above involved the invention of a new opcode,
692        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
693        cannot be changed at match time, I realized I could make a small
694        improvement to matching performance by compiling OP_ALLANY instead of
695        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
696        on the OP_ANY path.
697    
698    17. Compiling pcretest on Windows with readline support failed without the
699        following two fixes: (1) Make the unistd.h include conditional on
700        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
701    
702    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
703        ncurses library to be included for pcretest when ReadLine support is
704        requested, but also to allow for it to be overridden. This patch came from
705        Daniel Bergström.
706    
707    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
708        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
709        any errors with the current Unicode tables. Thanks to Peter Kankowski for
710        spotting this.
711    
712    
713    Version 7.6 28-Jan-08
714    ---------------------
715    
716    1.  A character class containing a very large number of characters with
717        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
718        overflow.
719    
720    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
721        HAVE_LONG_LONG is not defined.
722    
723    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
724        bring them up to date and include new features. This patch includes:
725    
726        - Fixed PH's badly added libz and libbz2 support.
727        - Fixed a problem with static linking.
728        - Added pcredemo. [But later removed - see 7 below.]
729        - Fixed dftables problem and added an option.
730        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
731            HAVE_LONG_LONG.
732        - Added readline support for pcretest.
733        - Added an listing of the option settings after cmake has run.
734    
735    4.  A user submitted a patch to Makefile that makes it easy to create
736        "pcre.dll" under mingw when using Configure/Make. I added stuff to
737        Makefile.am that cause it to include this special target, without
738        affecting anything else. Note that the same mingw target plus all
739        the other distribution libraries and programs are now supported
740        when configuring with CMake (see 6 below) instead of with
741        Configure/Make.
742    
743    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
744        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
745        exported in the Windows port". It has not yet been confirmed that the patch
746        solves the problem, but it does no harm.
747    
748    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
749        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
750        with CMake, and also correct the comment about stack recursion.
751    
752    7.  Remove the automatic building of pcredemo from the ./configure system and
753        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
754        of a program that users should build themselves after PCRE is installed, so
755        building it automatically is not really right. What is more, it gave
756        trouble in some build environments.
757    
758    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
759    
760    
761    Version 7.5 10-Jan-08
762    ---------------------
763    
764    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
765        values in parens when parsing an RE using the C++ wrapper."
766    
767    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
768        Characters greater than 255 were excluded from the class instead of being
769        included.
770    
771    3.  The same bug as (2) above applied to negated POSIX classes such as
772        [:^space:].
773    
774    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
775        defined or documented. It seems to have been a typo for PCRE_STATIC, so
776        I have changed it.
777    
778    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
779        first named subpattern) and a construct such as (?&a) would reference the
780        first named subpattern whose name started with "a" (in other words, the
781        length check was missing). Both these problems are fixed. "Subpattern name
782        expected" is now given for (?&) (a zero-length name), and this patch also
783        makes it give the same error for \k'' (previously it complained that that
784        was a reference to a non-existent subpattern).
785    
786    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
787        this is right because (?- can be followed by option settings as well as by
788        digits. I have, however, made the messages clearer.
789    
790    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
791        than the number used in the conditional) now cause a compile-time error.
792        This is actually not compatible with Perl, which accepts such patterns, but
793        treats the conditional as always being FALSE (as PCRE used to), but it
794        seems to me that giving a diagnostic is better.
795    
796    8.  Change "alphameric" to the more common word "alphanumeric" in comments
797        and messages.
798    
799    9.  Fix two occurrences of "backslash" in comments that should have been
800        "backspace".
801    
802    10. Remove two redundant lines of code that can never be obeyed (their function
803        was moved elsewhere).
804    
805    11. The program that makes PCRE's Unicode character property table had a bug
806        which caused it to generate incorrect table entries for sequences of
807        characters that have the same character type, but are in different scripts.
808        It amalgamated them into a single range, with the script of the first of
809        them. In other words, some characters were in the wrong script. There were
810        thirteen such cases, affecting characters in the following ranges:
811    
812          U+002b0 - U+002c1
813          U+0060c - U+0060d
814          U+0061e - U+00612
815          U+0064b - U+0065e
816          U+0074d - U+0076d
817          U+01800 - U+01805
818          U+01d00 - U+01d77
819          U+01d9b - U+01dbf
820          U+0200b - U+0200f
821          U+030fc - U+030fe
822          U+03260 - U+0327f
823          U+0fb46 - U+0fbb1
824          U+10450 - U+1049d
825    
826    12. The -o option (show only the matching part of a line) for pcregrep was not
827        compatible with GNU grep in that, if there was more than one match in a
828        line, it showed only the first of them. It now behaves in the same way as
829        GNU grep.
830    
831    13. If the -o and -v options were combined for pcregrep, it printed a blank
832        line for every non-matching line. GNU grep prints nothing, and pcregrep now
833        does the same. The return code can be used to tell if there were any
834        non-matching lines.
835    
836    14. Added --file-offsets and --line-offsets to pcregrep.
837    
838    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
839        infinitely looping recursion. The bug was that positive lookaheads were not
840        being skipped when checking for a possible empty match (negative lookaheads
841        and both kinds of lookbehind were skipped).
842    
843    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
844        inclusion of <windows.h> to before rather than after the definition of
845        INVALID_FILE_ATTRIBUTES (patch from David Byron).
846    
847    17. Specifying a possessive quantifier with a specific limit for a Unicode
848        character property caused pcre_compile() to compile bad code, which led at
849        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
850        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
851        caused the error; without that there was no problem.
852    
853    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
854    
855    19. Added --enable-pcretest-libreadline.
856    
857    20. In pcrecpp.cc, the variable 'count' was incremented twice in
858        RE::GlobalReplace(). As a result, the number of replacements returned was
859        double what it should be. I removed one of the increments, but Craig sent a
860        later patch that removed the other one (the right fix) and added unit tests
861        that check the return values (which was not done before).
862    
863    21. Several CMake things:
864    
865        (1) Arranged that, when cmake is used on Unix, the libraries end up with
866            the names libpcre and libpcreposix, not just pcre and pcreposix.
867    
868        (2) The above change means that pcretest and pcregrep are now correctly
869            linked with the newly-built libraries, not previously installed ones.
870    
871        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
872    
873    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
874        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
875        UTF-8 newline character). The key issue is that the pattern starts .*;
876        this means that the match must be either at the beginning, or after a
877        newline. The bug was in the code for advancing after a failed match and
878        checking that the new position followed a newline. It was not taking
879        account of UTF-8 characters correctly.
880    
881    23. PCRE was behaving differently from Perl in the way it recognized POSIX
882        character classes. PCRE was not treating the sequence [:...:] as a
883        character class unless the ... were all letters. Perl, however, seems to
884        allow any characters between [: and :], though of course it rejects as
885        unknown any "names" that contain non-letters, because all the known class
886        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
887        for example, whereas PCRE did not - it did not recognize a POSIX character
888        class. This seemed a bit dangerous, so the code has been changed to be
889        closer to Perl. The behaviour is not identical to Perl, because PCRE will
890        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
891        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
892        Perl does, and where it didn't before.
893    
894    24. Rewrite so as to remove the single use of %n from pcregrep because in some
895        Windows environments %n is disabled by default.
896    
897    
898    Version 7.4 21-Sep-07
899    ---------------------
900    
901    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
902        means that a class such as [\s] counted as "explicit reference to CR or
903        LF". That isn't really right - the whole point of the change was to try to
904        help when there was an actual mention of one of the two characters. So now
905        the change happens only if \r or \n (or a literal CR or LF) character is
906        encountered.
907    
908    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
909        of both had grown to the point where there were only 3 bits left.
910        Fortunately, there was spare space in the data structure, and so I have
911        moved the internal flags into a new 16-bit field to free up more option
912        bits.
913    
914    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
915        but did not set the internal JCHANGED flag - either of these is enough to
916        control the way the "get" function works - but the PCRE_INFO_JCHANGED
917        facility is supposed to tell if (?J) was ever used, so now (?J) at the
918        start sets both bits.
919    
920    4.  Added options (at build time, compile time, exec time) to change \R from
921        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
922    
923    5.  doc/pcresyntax.html was missing from the distribution.
924    
925    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
926        compatibility, even though it is no longer used.
927    
928    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
929        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
930        windows.h file is present (where different names are used). [This was
931        reversed later after testing - see 16 below.]
932    
933    8.  Changed all #include <config.h> to #include "config.h". There were also
934        some further <pcre.h> cases that I changed to "pcre.h".
935    
936    9.  When pcregrep was used with the --colour option, it missed the line ending
937        sequence off the lines that it output.
938    
939    10. It was pointed out to me that arrays of string pointers cause lots of
940        relocations when a shared library is dynamically loaded. A technique of
941        using a single long string with a table of offsets can drastically reduce
942        these. I have refactored PCRE in four places to do this. The result is
943        dramatic:
944    
945          Originally:                          290
946          After changing UCP table:            187
947          After changing error message table:   43
948          After changing table of "verbs"       36
949          After changing table of Posix names   22
950    
951        Thanks to the folks working on Gregex for glib for this insight.
952    
953    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
954        unicode-properties was also set.
955    
956    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
957    
958    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
959        checked only for CRLF.
960    
961    14. Added casts to pcretest.c to avoid compiler warnings.
962    
963    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
964    
965    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
966        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
967        entirely. This removes changes made in 7 above.
968    
969    17. The CMake files have been updated, and there is now more information about
970        building with CMake in the NON-UNIX-USE document.
971    
972    
973    Version 7.3 28-Aug-07
974    ---------------------
975    
976     1. In the rejigging of the build system that eventually resulted in 7.1, the
977        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
978        brackets there is not right, since it causes compilers to look for an
979        installed pcre.h, not the version that is in the source that is being
980        compiled (which of course may be different). I have changed it back to:
981    
982          #include "pcre.h"
983    
984        I have a vague recollection that the change was concerned with compiling in
985        different directories, but in the new build system, that is taken care of
986        by the VPATH setting the Makefile.
987    
988     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
989        when the subject happened to end in the byte 0x85 (e.g. if the last
990        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
991        characters but of course it shouldn't be taken as a newline when it is part
992        of another character. The bug was that, for an unlimited repeat of . in
993        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
994        characters when looking for a newline.
995    
996     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
997    
998     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
999        in debug output.
1000    
1001     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1002        long printing in the pcrecpp unittest when running under MinGW.
1003    
1004     6. ESC_K was left out of the EBCDIC table.
1005    
1006     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1007        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1008        limit also applies to "virtual nesting" when a pattern is recursive, and in
1009        this case 1000 isn't so big. I have been able to remove this limit at the
1010        expense of backing off one optimization in certain circumstances. Normally,
1011        when pcre_exec() would call its internal match() function recursively and
1012        immediately return the result unconditionally, it uses a "tail recursion"
1013        feature to save stack. However, when a subpattern that can match an empty
1014        string has an unlimited repetition quantifier, it no longer makes this
1015        optimization. That gives it a stack frame in which to save the data for
1016        checking that an empty string has been matched. Previously this was taken
1017        from the 1000-entry workspace that had been reserved. So now there is no
1018        explicit limit, but more stack is used.
1019    
1020     8. Applied Daniel's patches to solve problems with the import/export magic
1021        syntax that is required for Windows, and which was going wrong for the
1022        pcreposix and pcrecpp parts of the library. These were overlooked when this
1023        problem was solved for the main library.
1024    
1025     9. There were some crude static tests to avoid integer overflow when computing
1026        the size of patterns that contain repeated groups with explicit upper
1027        limits. As the maximum quantifier is 65535, the maximum group length was
1028        set at 30,000 so that the product of these two numbers did not overflow a
1029        32-bit integer. However, it turns out that people want to use groups that
1030        are longer than 30,000 bytes (though not repeat them that many times).
1031        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1032        made it possible to implement the integer overflow checks in a much more
1033        dynamic way, which I have now done. The artificial limitation on group
1034        length has been removed - we now have only the limit on the total length of
1035        the compiled pattern, which depends on the LINK_SIZE setting.
1036    
1037    10. Fixed a bug in the documentation for get/copy named substring when
1038        duplicate names are permitted. If none of the named substrings are set, the
1039        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1040        empty string.
1041    
1042    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1043        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1044        because the ] is interpreted as the first data character and the
1045        terminating ] is not found. PCRE has been made compatible with Perl in this
1046        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1047        cause memory overwriting.
1048    
1049    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1050        string has been matched (to stop an infinite loop). It was not recognizing
1051        a conditional subpattern that could match an empty string if that
1052        subpattern was within another subpattern. For example, it looped when
1053        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1054        condition was not nested. This bug has been fixed.
1055    
1056    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1057        past the start of the subject in the presence of bytes with the top bit
1058        set, for example "\x8aBCD".
1059    
1060    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1061        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1062    
1063    14. Optimized (?!) to (*FAIL).
1064    
1065    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1066        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1067        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1068        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1069        does: it's just the validity check that is more restrictive.
1070    
1071    16. Inserted checks for integer overflows during escape sequence (backslash)
1072        processing, and also fixed erroneous offset values for syntax errors during
1073        backslash processing.
1074    
1075    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1076        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1077    
1078    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1079        caused an overrun.
1080    
1081    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1082        something other than just ASCII characters) inside a group that had an
1083        unlimited repeat caused a loop at compile time (while checking to see
1084        whether the group could match an empty string).
1085    
1086    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1087        [\P{Any}] did so. (Error in the code for printing property names.)
1088    
1089    21. An orphan \E inside a character class could cause a crash.
1090    
1091    22. A repeated capturing bracket such as (A)? could cause a wild memory
1092        reference during compilation.
1093    
1094    23. There are several functions in pcre_compile() that scan along a compiled
1095        expression for various reasons (e.g. to see if it's fixed length for look
1096        behind). There were bugs in these functions when a repeated \p or \P was
1097        present in the pattern. These operators have additional parameters compared
1098        with \d, etc, and these were not being taken into account when moving along
1099        the compiled data. Specifically:
1100    
1101        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1102            length.
1103    
1104        (b) An item such as \pL+ within a repeated group could cause crashes or
1105            loops.
1106    
1107        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1108            "reference to non-existent subpattern" error.
1109    
1110        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1111    
1112    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1113        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1114    
1115    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1116    
1117    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1118        character were causing crashes (broken optimization).
1119    
1120    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1121        \p or \P) caused a compile-time loop.
1122    
1123    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1124        break. For example, the unstudied pattern [\r\n]A does not match the string
1125        "\r\nA" because change 7.0/46 below moves the current point on by two
1126        characters after failing to match at the start. However, the pattern \nA
1127        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1128        the same is true. There doesn't seem any very clean way out of this, but
1129        what I have chosen to do makes the common cases work: PCRE now takes note
1130        of whether there can be an explicit match for \r or \n anywhere in the
1131        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1132        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1133        pattern has explicit CR or LF references.
1134    
1135    29. Added (*CR) etc for changing newline setting at start of pattern.
1136    
1137    
1138    Version 7.2 19-Jun-07
1139    ---------------------
1140    
1141     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1142        which is apparently normally available under Windows.
1143    
1144     2. Re-jig the pcregrep tests with different newline settings in an attempt
1145        to make them independent of the local environment's newline setting.
1146    
1147     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1148    
1149     4. Some of the "internals" tests were previously cut out when the link size
1150        was not 2, because the output contained actual offsets. The recent new
1151        "Z" feature of pcretest means that these can be cut out, making the tests
1152        usable with all link sizes.
1153    
1154     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1155        stack recursion. This gives a massive performance boost under BSD, but just
1156        a small improvement under Linux. However, it saves one field in the frame
1157        in all cases.
1158    
1159     6. Added more features from the forthcoming Perl 5.10:
1160    
1161        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1162            recursion call. It refers to the nth most recently opened parentheses.
1163    
1164        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1165            to be opened parentheses.
1166    
1167        (c) Conditions that refer to capturing parentheses can be specified
1168            relatively, for example, (?(-2)... or (?(+3)...
1169    
1170        (d) \K resets the start of the current match so that everything before
1171            is not part of it.
1172    
1173        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1174    
1175        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1176            reference syntax.
1177    
1178        (g) (?| introduces a group in which the numbering of parentheses in each
1179            alternative starts with the same number.
1180    
1181        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1182    
1183     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1184        PCRE_INFO_JCHANGED.
1185    
1186     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1187        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1188        for detecting groups that can match an empty string.
1189    
1190     9. A pattern with a very large number of alternatives (more than several
1191        hundred) was running out of internal workspace during the pre-compile
1192        phase, where pcre_compile() figures out how much memory will be needed. A
1193        bit of new cunning has reduced the workspace needed for groups with
1194        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1195        workspace instead of running out of the 4096 that are available.
1196    
1197    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1198    
1199    11. Applied patch from Google to remove an optimization that didn't quite work.
1200        The report of the bug said:
1201    
1202          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1203          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1204          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1205    
1206    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1207        it matched the wrong number of bytes.
1208    
1209    
1210    Version 7.1 24-Apr-07
1211    ---------------------
1212    
1213     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1214        that is more "standard", making use of automake and other Autotools. There
1215        is some re-arrangement of the files and adjustment of comments consequent
1216        on this.
1217    
1218     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1219        for recursive directory scanning broke on some systems because the files
1220        are not scanned in any specific order and on different systems the order
1221        was different. A call to "sort" has been inserted into RunGrepTest for the
1222        approprate test as a short-term fix. In the longer term there may be an
1223        alternative.
1224    
1225     3. I had an email from Eric Raymond about problems translating some of PCRE's
1226        man pages to HTML (despite the fact that I distribute HTML pages, some
1227        people do their own conversions for various reasons). The problems
1228        concerned the use of low-level troff macros .br and .in. I have therefore
1229        removed all such uses from the man pages (some were redundant, some could
1230        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1231        HTML has been updated to handle .nf/.fi and to complain if it encounters
1232        .br or .in.
1233    
1234     4. Updated comments in configure.ac that get placed in config.h.in and also
1235        arranged for config.h to be included in the distribution, with the name
1236        config.h.generic, for the benefit of those who have to compile without
1237        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1238    
1239     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1240        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1241        makevp.bat for the current PCRE, using the additional files
1242        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1243    
1244     6. A Windows user reported a minor discrepancy with test 2, which turned out
1245        to be caused by a trailing space on an input line that had got lost in his
1246        copy. The trailing space was an accident, so I've just removed it.
1247    
1248     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1249        that is needed.
1250    
1251     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1252        as "const" (a) because they are and (b) because it helps the PHP
1253        maintainers who have recently made a script to detect big data structures
1254        in the php code that should be moved to the .rodata section. I remembered
1255        to update Builducptable as well, so it won't revert if ucptable.h is ever
1256        re-created.
1257    
1258     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1259        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1260        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1261        support is not required. This saves 1.5-2K of code, which is important in
1262        some applications.
1263    
1264        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1265        so as not to refer to the tables, even though these functions will never be
1266        called when UTF-8 support is disabled. Otherwise there are problems with a
1267        shared library.
1268    
1269    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1270    
1271        (a) It was defining its arguments as char * instead of void *.
1272    
1273        (b) It was assuming that all moves were upwards in memory; this was true
1274            a long time ago when I wrote it, but is no longer the case.
1275    
1276        The emulated memove() is provided for those environments that have neither
1277        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1278        is clearly not the case, as these two bugs were recently reported.
1279    
1280    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1281        and Detrail to create the HTML documentation, the .txt form of the man
1282        pages, and it removes trailing spaces from listed files. It also creates
1283        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1284        case, it wraps all the #defines with #ifndefs. This script should be run
1285        before "make dist".
1286    
1287    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1288        with Unicode property support.
1289    
1290        (a) For a maximizing quantifier, if the two different cases of the
1291            character were of different lengths in their UTF-8 codings (there are
1292            some cases like this - I found 11), and the matching function had to
1293            back up over a mixture of the two cases, it incorrectly assumed they
1294            were both the same length.
1295    
1296        (b) When PCRE was configured to use the heap rather than the stack for
1297            recursion during matching, it was not correctly preserving the data for
1298            the other case of a UTF-8 character when checking ahead for a match
1299            while processing a minimizing repeat. If the check also involved
1300            matching a wide character, but failed, corruption could cause an
1301            erroneous result when trying to check for a repeat of the original
1302            character.
1303    
1304    13. Some tidying changes to the testing mechanism:
1305    
1306        (a) The RunTest script now detects the internal link size and whether there
1307            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1308            values substituted by "configure". (The RunGrepTest script already did
1309            this for UTF-8.) The configure.ac script no longer substitutes the
1310            relevant variables.
1311    
1312        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1313            with length and offset values. This means that the output is different
1314            for different internal link sizes. Test 2 is skipped for link sizes
1315            other than 2 because of this, bypassing the problem. Unfortunately,
1316            there was also a test in test 3 (the locale tests) that used /B and
1317            failed for link sizes other than 2. Rather than cut the whole test out,
1318            I have added a new /Z option to pcretest that replaces the length and
1319            offset values with spaces. This is now used to make test 3 independent
1320            of link size. (Test 2 will be tidied up later.)
1321    
1322    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1323        segmentation fault instead of returning the appropriate error message.
1324    
1325    15. In multiline mode when the newline sequence was set to "any", the pattern
1326        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1327        This doesn't seem right; it now treats the CRLF combination as the line
1328        ending, and so does not match in that case. It's only a pattern such as ^$
1329        that would hit this one: something like ^ABC$ would have failed after \r
1330        and then tried again after \r\n.
1331    
1332    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1333        in an attempt to make files that differ only in their line terminators
1334        compare equal. This works on Linux.
1335    
1336    17. Under certain error circumstances pcregrep might try to free random memory
1337        as it exited. This is now fixed, thanks to valgrind.
1338    
1339    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1340        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1341        was because its rules for how to advance for /g after matching an empty
1342        string at the end of a line did not allow for this case. They now check for
1343        it specially.
1344    
1345    20. pcretest is supposed to handle patterns and data of any length, by
1346        extending its buffers when necessary. It was getting this wrong when the
1347        buffer for a data line had to be extended.
1348    
1349    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1350        CRLF as a newline sequence.
1351    
1352    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1353        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1354        I have nevertheless tidied it up.
1355    
1356    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1357    
1358    24. Added a man page for pcre-config.
1359    
1360    
1361    Version 7.0 19-Dec-06
1362    ---------------------
1363    
1364     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1365        moving to gcc 4.1.1.
1366    
1367     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1368        sys/time.h, which is documented as needed for this function. It doesn't
1369        seem to matter on Linux, but it showed up on some releases of OS X.
1370    
1371     3. It seems that there are systems where bytes whose values are greater than
1372        127 match isprint() in the "C" locale. The "C" locale should be the
1373        default when a C program starts up. In most systems, only ASCII printing
1374        characters match isprint(). This difference caused the output from pcretest
1375        to vary, making some of the tests fail. I have changed pcretest so that:
1376    
1377        (a) When it is outputting text in the compiled version of a pattern, bytes
1378            other than 32-126 are always shown as hex escapes.
1379    
1380        (b) When it is outputting text that is a matched part of a subject string,
1381            it does the same, unless a different locale has been set for the match
1382            (using the /L modifier). In this case, it uses isprint() to decide.
1383    
1384     4. Fixed a major bug that caused incorrect computation of the amount of memory
1385        required for a compiled pattern when options that changed within the
1386        pattern affected the logic of the preliminary scan that determines the
1387        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1388        that the computed length was too small. The symptoms of this bug were
1389        either the PCRE error "internal error: code overflow" from pcre_compile(),
1390        or a glibc crash with a message such as "pcretest: free(): invalid next
1391        size (fast)". Examples of patterns that provoked this bug (shown in
1392        pcretest format) are:
1393    
1394          /(?-x: )/x
1395          /(?x)(?-x: \s*#\s*)/
1396          /((?i)[\x{c0}])/8
1397          /(?i:[\x{c0}])/8
1398    
1399        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1400        is now done differently.
1401    
1402     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1403        wrapper classes; (b) implement a new function in the C++ scanner that is
1404        more efficient than the old way of doing things because it avoids levels of
1405        recursion in the regex matching; (c) add a paragraph to the documentation
1406        for the FullMatch() function.
1407    
1408     6. The escape sequence \n was being treated as whatever was defined as
1409        "newline". Not only was this contrary to the documentation, which states
1410        that \n is character 10 (hex 0A), but it also went horribly wrong when
1411        "newline" was defined as CRLF. This has been fixed.
1412    
1413     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1414        was being set to -1 for the "end of line" case (supposedly a value that no
1415        character can have). Though this value is never used (the check for end of
1416        line is "zero bytes in current character"), it caused compiler complaints.
1417        I've changed it to 0xffffffff.
1418    
1419     8. In pcre_version.c, the version string was being built by a sequence of
1420        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1421        string (as it is for production releases) called a macro with an empty
1422        argument. The C standard says the result of this is undefined. The gcc
1423        compiler treats it as an empty string (which was what was wanted) but it is
1424        reported that Visual C gives an error. The source has been hacked around to
1425        avoid this problem.
1426    
1427     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1428        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1429        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1430        of them did).
1431    
1432    10. Originally, pcretest opened its input and output without "b"; then I was
1433        told that "b" was needed in some environments, so it was added for release
1434        5.0 to both the input and output. (It makes no difference on Unix-like
1435        systems.) Later I was told that it is wrong for the input on Windows. I've
1436        now abstracted the modes into two macros, to make it easier to fiddle with
1437        them, and removed "b" from the input mode under Windows.
1438    
1439    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1440    
1441    12. Added -help and --help to pcretest as an official way of being reminded
1442        of the options.
1443    
1444    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1445        and pcrecpp.cc because they annoy compilers at high warning levels.
1446    
1447    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1448    
1449    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1450        scripts are not C programs :-) and which was not noticed because it works
1451        on Linux.
1452    
1453    16. pcretest is supposed to handle any length of pattern and data line (as one
1454        line or as a continued sequence of lines) by extending its input buffer if
1455        necessary. This feature was broken for very long pattern lines, leading to
1456        a string of junk being passed to pcre_compile() if the pattern was longer
1457        than about 50K.
1458    
1459    17. I have done a major re-factoring of the way pcre_compile() computes the
1460        amount of memory needed for a compiled pattern. Previously, there was code
1461        that made a preliminary scan of the pattern in order to do this. That was
1462        OK when PCRE was new, but as the facilities have expanded, it has become
1463        harder and harder to keep it in step with the real compile phase, and there
1464        have been a number of bugs (see for example, 4 above). I have now found a
1465        cunning way of running the real compile function in a "fake" mode that
1466        enables it to compute how much memory it would need, while actually only
1467        ever using a few hundred bytes of working memory and without too many
1468        tests of the mode. This should make future maintenance and development
1469        easier. A side effect of this work is that the limit of 200 on the nesting
1470        depth of parentheses has been removed (though this was never a serious
1471        limitation, I suspect). However, there is a downside: pcre_compile() now
1472        runs more slowly than before (30% or more, depending on the pattern). I
1473        hope this isn't a big issue. There is no effect on runtime performance.
1474    
1475    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1476        newline (only possible for the last line of a file) and it was a
1477        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1478    
1479    19. Added additional timing features to pcretest. (1) The -tm option now times
1480        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1481        separate command line item, by a number that specifies the number of
1482        repeats to use when timing. The default is 50000; this gives better
1483        precision, but takes uncomfortably long for very large patterns.
1484    
1485    20. Extended pcre_study() to be more clever in cases where a branch of a
1486        subpattern has no definite first character. For example, (a*|b*)[cd] would
1487        previously give no result from pcre_study(). Now it recognizes that the
1488        first character must be a, b, c, or d.
1489    
1490    21. There was an incorrect error "recursive call could loop indefinitely" if
1491        a subpattern (or the entire pattern) that was being tested for matching an
1492        empty string contained only one non-empty item after a nested subpattern.
1493        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1494        incorrectly, because the \d was being skipped in the check.
1495    
1496    22. The pcretest program now has a new pattern option /B and a command line
1497        option -b, which is equivalent to adding /B to every pattern. This causes
1498        it to show the compiled bytecode, without the additional information that
1499        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1500        is the same as /B/I).
1501    
1502    23. A new optimization is now able automatically to treat some sequences such
1503        as a*b as a*+b. More specifically, if something simple (such as a character
1504        or a simple class like \d) has an unlimited quantifier, and is followed by
1505        something that cannot possibly match the quantified thing, the quantifier
1506        is automatically "possessified".
1507    
1508    24. A recursive reference to a subpattern whose number was greater than 39
1509        went wrong under certain circumstances in UTF-8 mode. This bug could also
1510        have affected the operation of pcre_study().
1511    
1512    25. Realized that a little bit of performance could be had by replacing
1513        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1514    
1515    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1516    
1517    27. Possessive quantifiers such as a++ were previously implemented by turning
1518        them into atomic groups such as ($>a+). Now they have their own opcodes,
1519        which improves performance. This includes the automatically created ones
1520        from 23 above.
1521    
1522    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1523        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1524        the first matched character to be a colon. This applied both to named and
1525        numbered groups.
1526    
1527    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1528    
1529    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1530        building PCRE on Windows easier, so I have included it in the distribution.
1531    
1532    31. There is now a check in pcretest against a ridiculously large number being
1533        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1534        loop, the loop is abandoned.
1535    
1536    32. Forward references to subpatterns in conditions such as (?(2)...) where
1537        subpattern 2 is defined later cause pcre_compile() to search forwards in
1538        the pattern for the relevant set of parentheses. This search went wrong
1539        when there were unescaped parentheses in a character class, parentheses
1540        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1541    
1542    33. "Subroutine" calls and backreferences were previously restricted to
1543        referencing subpatterns earlier in the regex. This restriction has now
1544        been removed.
1545    
1546    34. Added a number of extra features that are going to be in Perl 5.10. On the
1547        whole, these are just syntactic alternatives for features that PCRE had
1548        previously implemented using the Python syntax or my own invention. The
1549        other formats are all retained for compatibility.
1550    
1551        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1552            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1553            also .NET compatible.
1554    
1555        (b) A recursion or subroutine call to a named group can now be defined as
1556            (?&name) as well as (?P>name).
1557    
1558        (c) A backreference to a named group can now be defined as \k<name> or
1559            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1560            5.10, are also .NET compatible.
1561    
1562        (d) A conditional reference to a named group can now use the syntax
1563            (?(<name>) or (?('name') as well as (?(name).
1564    
1565        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1566            groups (named and numbered) that are never evaluated inline, but can be
1567            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1568            is always false. There may be only one alternative in such a group.
1569    
1570        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1571            as the simple (?(R). The condition is true only if the most recent
1572            recursion is that of the given number or name. It does not search out
1573            through the entire recursion stack.
1574    
1575        (g) The escape \gN or \g{N} has been added, where N is a positive or
1576            negative number, specifying an absolute or relative reference.
1577    
1578    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1579        some "unreachable code" warnings.
1580    
1581    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1582        things, this adds five new scripts.
1583    
1584    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1585        There were also incompatibilities regarding the handling of \Q..\E inside
1586        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1587        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1588    
1589    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1590        matches an empty string, and forcibly breaks the loop. There were bugs in
1591        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1592        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1593        separate and independent bugs (that affected different cases) have been
1594        fixed.
1595    
1596    39. Refactored the code to abolish the use of different opcodes for small
1597        capturing bracket numbers. This is a tidy that I avoided doing when I
1598        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1599        The new approach is not only tidier, it makes it possible to reduce the
1600        memory needed to fix the previous bug (38).
1601    
1602    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1603        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1604        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1605        mode.
1606    
1607    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1608        report.
1609    
1610    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1611        copy construction and assignment in the C++ wrapper.
1612    
1613    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1614        couple of bugs that could have given wrong results in the "--newline=crlf"
1615        case.
1616    
1617    44. Added a number of casts and did some reorganization of signed/unsigned int
1618        variables following suggestions from Dair Grant. Also renamed the variable
1619        "this" as "item" because it is a C++ keyword.
1620    
1621    45. Arranged for dftables to add
1622    
1623          #include "pcre_internal.h"
1624    
1625        to pcre_chartables.c because without it, gcc 4.x may remove the array
1626        definition from the final binary if PCRE is built into a static library and
1627        dead code stripping is activated.
1628    
1629    46. For an unanchored pattern, if a match attempt fails at the start of a
1630        newline sequence, and the newline setting is CRLF or ANY, and the next two
1631        characters are CRLF, advance by two characters instead of one.
1632    
1633    
1634    Version 6.7 04-Jul-06
1635    ---------------------
1636    
1637     1. In order to handle tests when input lines are enormously long, pcretest has
1638        been re-factored so that it automatically extends its buffers when
1639        necessary. The code is crude, but this _is_ just a test program. The
1640        default size has been increased from 32K to 50K.
1641    
1642     2. The code in pcre_study() was using the value of the re argument before
1643        testing it for NULL. (Of course, in any sensible call of the function, it
1644        won't be NULL.)
1645    
1646     3. The memmove() emulation function in pcre_internal.h, which is used on
1647        systems that lack both memmove() and bcopy() - that is, hardly ever -
1648        was missing a "static" storage class specifier.
1649    
1650     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1651        containing an extended class (one that cannot be represented by a bitmap
1652        because it contains high-valued characters or Unicode property items, e.g.
1653        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1654        pattern, but PCRE should not loop if you do not (it no longer does).
1655        [Detail: two cases were found: (a) a repeated subpattern containing an
1656        extended class; (b) a recursive reference to a subpattern that followed a
1657        previous extended class. It wasn't skipping over the extended class
1658        correctly when UTF-8 mode was not set.]
1659    
1660     5. A negated single-character class was not being recognized as fixed-length
1661        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1662        compile error "lookbehind assertion is not fixed length".
1663    
1664     6. The RunPerlTest auxiliary script was showing an unexpected difference
1665        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1666        write a Perl script that can interpret lines of an input file either as
1667        byte characters or as UTF-8, which is what "perltest" was being required to
1668        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1669        can't do is switch easily at run time between having the "use utf8;" pragma
1670        or not. In the end, I fudged it by using the RunPerlTest script to insert
1671        "use utf8;" explicitly for the UTF-8 tests.
1672    
1673     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1674        the end of the subject string, contrary to the documentation and to what
1675        Perl does. This was true of both matching functions. Now it matches only at
1676        the start of the subject and immediately after *internal* newlines.
1677    
1678     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1679        a pointer to an int instead of a pointer to an unsigned long int. This
1680        caused problems on 64-bit systems.
1681    
1682     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1683        instance of the 'standard' template library not being so standard".
1684    
1685    10. There was no check on the number of named subpatterns nor the maximum
1686        length of a subpattern name. The product of these values is used to compute
1687        the size of the memory block for a compiled pattern. By supplying a very
1688        long subpattern name and a large number of named subpatterns, the size
1689        computation could be caused to overflow. This is now prevented by limiting
1690        the length of names to 32 characters, and the number of named subpatterns
1691        to 10,000.
1692    
1693    11. Subpatterns that are repeated with specific counts have to be replicated in
1694        the compiled pattern. The size of memory for this was computed from the
1695        length of the subpattern and the repeat count. The latter is limited to
1696        65535, but there was no limit on the former, meaning that integer overflow
1697        could in principle occur. The compiled length of a repeated subpattern is
1698        now limited to 30,000 bytes in order to prevent this.
1699    
1700    12. Added the optional facility to have named substrings with the same name.
1701    
1702    13. Added the ability to use a named substring as a condition, using the
1703        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1704        are numbers (not recommended). Forward references are permitted.
1705    
1706    14. Added forward references in named backreferences (if you see what I mean).
1707    
1708    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1709        pattern could run off the end of the subject. For example, the pattern
1710        "(?s)(.{1,5})"8 did this with the subject "ab".
1711    
1712    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1713        PCRE_CASELESS was set when matching characters that were quantified with ?
1714        or *.
1715    
1716    17. A character class other than a single negated character that had a minimum
1717        but no maximum quantifier - for example [ab]{6,} - was not handled
1718        correctly by pce_dfa_exec(). It would match only one character.
1719    
1720    18. A valid (though odd) pattern that looked like a POSIX character
1721        class but used an invalid character after [ (for example [[,abc,]]) caused
1722        pcre_compile() to give the error "Failed: internal error: code overflow" or
1723        in some cases to crash with a glibc free() error. This could even happen if
1724        the pattern terminated after [[ but there just happened to be a sequence of
1725        letters, a binary zero, and a closing ] in the memory that followed.
1726    
1727    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1728        over the years. Originally (before any Unicode support), just the bottom 8
1729        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1730        output from "man perlunicode" includes this:
1731    
1732          The regular expression compiler produces polymorphic opcodes.  That
1733          is, the pattern adapts to the data and automatically switches to
1734          the Unicode character scheme when presented with Unicode data--or
1735          instead uses a traditional byte scheme when presented with byte
1736          data.
1737    
1738        Sadly, a wide octal escape does not cause a switch, and in a string with
1739        no other multibyte characters, these octal escapes are treated as before.
1740        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1741        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1742        Unicode string.
1743    
1744        I have not perpetrated such confusion in PCRE. Up till now, it took just
1745        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1746        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1747        translate to the appropriate multibyte character.
1748    
1749    29. Applied some refactoring to reduce the number of warnings from Microsoft
1750        and Borland compilers. This has included removing the fudge introduced
1751        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1752        a warning about an unused variable.
1753    
1754    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1755        characters since release 4.0, because Perl (from release 5.004) does not.
1756        [Or at least, is documented not to: some releases seem to be in conflict
1757        with the documentation.] However, when a pattern was studied with
1758        pcre_study() and all its branches started with \s, PCRE still included VT
1759        as a possible starting character. Of course, this did no harm; it just
1760        caused an unnecessary match attempt.
1761    
1762    22. Removed a now-redundant internal flag bit that recorded the fact that case
1763        dependency changed within the pattern. This was once needed for "required
1764        byte" processing, but is no longer used. This recovers a now-scarce options
1765        bit. Also moved the least significant internal flag bit to the most-
1766        significant bit of the word, which was not previously used (hangover from
1767        the days when it was an int rather than a uint) to free up another bit for
1768        the future.
1769    
1770    23. Added support for CRLF line endings as well as CR and LF. As well as the
1771        default being selectable at build time, it can now be changed at runtime
1772        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1773        specify that it is scanning data with non-default line endings.
1774    
1775    24. Changed the definition of CXXLINK to make it agree with the definition of
1776        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1777    
1778    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1779        recursions. This makes a big different to stack usage for some patterns.
1780    
1781    26. If a subpattern containing a named recursion or subroutine reference such
1782        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1783        the space required for the compiled pattern went wrong and gave too small a
1784        value. Depending on the environment, this could lead to "Failed: internal
1785        error: code overflow at offset 49" or "glibc detected double free or
1786        corruption" errors.
1787    
1788    27. Applied patches from Google (a) to support the new newline modes and (b) to
1789        advance over multibyte UTF-8 characters in GlobalReplace.
1790    
1791    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1792        difference for some implementation of PCRE in some Windows version.
1793    
1794    29. Added some extra testing facilities to pcretest:
1795    
1796        \q<number>   in a data line sets the "match limit" value
1797        \Q<number>   in a data line sets the "match recursion limt" value
1798        -S <number>  sets the stack size, where <number> is in megabytes
1799    
1800        The -S option isn't available for Windows.
1801    
1802    
1803    Version 6.6 06-Feb-06
1804    ---------------------
1805    
1806     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1807        in pcreposix.h. I have copied the definition from pcre.h.
1808    
1809     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1810        because pcre.h is no longer a built file.
1811    
1812     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1813        not normally included in the compiled code.
1814    
1815    
1816    Version 6.5 01-Feb-06
1817    ---------------------
1818    
1819     1. When using the partial match feature with pcre_dfa_exec(), it was not
1820        anchoring the second and subsequent partial matches at the new starting
1821        point. This could lead to incorrect results. For example, with the pattern
1822        /1234/, partially matching against "123" and then "a4" gave a match.
1823    
1824     2. Changes to pcregrep:
1825    
1826        (a) All non-match returns from pcre_exec() were being treated as failures
1827            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1828            error message is output. Some extra information is given for the
1829            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1830            probably the only errors that are likely to be caused by users (by
1831            specifying a regex that has nested indefinite repeats, for instance).
1832            If there are more than 20 of these errors, pcregrep is abandoned.
1833    
1834        (b) A binary zero was treated as data while matching, but terminated the
1835            output line if it was written out. This has been fixed: binary zeroes
1836            are now no different to any other data bytes.
1837    
1838        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1839            used to set a locale for matching. The --locale=xxxx long option has
1840            been added (no short equivalent) to specify a locale explicitly on the
1841            pcregrep command, overriding the environment variables.
1842    
1843        (d) When -B was used with -n, some line numbers in the output were one less
1844            than they should have been.
1845    
1846        (e) Added the -o (--only-matching) option.
1847    
1848        (f) If -A or -C was used with -c (count only), some lines of context were
1849            accidentally printed for the final match.
1850    
1851        (g) Added the -H (--with-filename) option.
1852    
1853        (h) The combination of options -rh failed to suppress file names for files
1854            that were found from directory arguments.
1855    
1856        (i) Added the -D (--devices) and -d (--directories) options.
1857    
1858        (j) Added the -F (--fixed-strings) option.
1859    
1860        (k) Allow "-" to be used as a file name for -f as well as for a data file.
1861    
1862        (l) Added the --colo(u)r option.
1863    
1864        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1865            is not present by default.
1866    
1867     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1868        items such as (?R) or (?1), when the recursion could match a number of
1869        alternatives. If it matched one of the alternatives, but subsequently,
1870        outside the recursion, there was a failure, the code tried to back up into
1871        the recursion. However, because of the way PCRE is implemented, this is not
1872        possible, and the result was an incorrect result from the match.
1873    
1874        In order to prevent this happening, the specification of recursion has
1875        been changed so that all such subpatterns are automatically treated as
1876        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1877    
1878     4. I had overlooked the fact that, in some locales, there are characters for
1879        which isalpha() is true but neither isupper() nor islower() are true. In
1880        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1881        and ordfeminine) are like this. This affected the treatment of \w and \W
1882        when they appeared in character classes, but not when they appeared outside
1883        a character class. The bit map for "word" characters is now created
1884        separately from the results of isalnum() instead of just taking it from the
1885        upper, lower, and digit maps. (Plus the underscore character, of course.)
1886    
1887     5. The above bug also affected the handling of POSIX character classes such as
1888        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1889        permanent tables. Instead, the bit maps for such a class were previously
1890        created as the appropriate unions of the upper, lower, and digit bitmaps.
1891        Now they are created by subtraction from the [[:word:]] class, which has
1892        its own bitmap.
1893    
1894     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1895        It is created by subtracting the vertical space characters (\x09, \x0a,
1896        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1897        subtraction was done in the overall bitmap for a character class, meaning
1898        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1899        be recognized. This bug has been fixed.
1900    
1901     7. Patches from the folks at Google:
1902    
1903          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1904          real life, but is still worth protecting against".
1905    
1906          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1907          regular expressions".
1908    
1909          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1910          have it.
1911    
1912          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1913          "configure" and the latter not, in order to fix a problem somebody had
1914          with compiling the Arg class on HP-UX.
1915    
1916          (e) Improve the error-handling of the C++ wrapper a little bit.
1917    
1918          (f) New tests for checking recursion limiting.
1919    
1920     8. The pcre_memmove() function, which is used only if the environment does not
1921        have a standard memmove() function (and is therefore rarely compiled),
1922        contained two bugs: (a) use of int instead of size_t, and (b) it was not
1923        returning a result (though PCRE never actually uses the result).
1924    
1925     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1926        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1927        returned instead of calling malloc() with an overflowing number that would
1928        most likely cause subsequent chaos.
1929    
1930    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1931    
1932    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1933        with this option is matched, the nmatch and pmatch options of regexec() are
1934        ignored.
1935    
1936    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1937        provided in case anyone wants to the the POSIX interface with UTF-8
1938        strings.
1939    
1940    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1941        C++ linking (needed for some HP-UX environments).
1942    
1943    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1944        (unused parameter) and in the pcre_printint() function (omitted "default"
1945        switch label when the default is to do nothing).
1946    
1947    15. Added some code to make it possible, when PCRE is compiled as a C++
1948        library, to replace subject pointers for pcre_exec() with a smart pointer
1949        class, thus making it possible to process discontinuous strings.
1950    
1951    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1952        much the same function. They were added by different people who were trying
1953        to make PCRE easy to compile on non-Unix systems. It has been suggested
1954        that PCRE_EXPORT be abolished now that there is more automatic apparatus
1955        for compiling on Windows systems. I have therefore replaced it with
1956        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1957        defaults to "extern" for C or "extern C" for C++, which works fine on
1958        Unix-like systems. It is now possible to override the value of PCRE_DATA_
1959        SCOPE with something explicit in config.h. In addition:
1960    
1961        (a) pcreposix.h still had just "extern" instead of either of these macros;
1962            I have replaced it with PCRE_DATA_SCOPE.
1963    
1964        (b) Functions such as _pcre_xclass(), which are internal to the library,
1965            but external in the C sense, all had PCRE_EXPORT in their definitions.
1966            This is apparently wrong for the Windows case, so I have removed it.
1967            (It makes no difference on Unix-like systems.)
1968    
1969    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1970        of recursive calls to match(). This is different to MATCH_LIMIT because
1971        that limits the total number of calls to match(), not all of which increase
1972        the depth of recursion. Limiting the recursion depth limits the amount of
1973        stack (or heap if NO_RECURSE is set) that is used. The default can be set
1974        when PCRE is compiled, and changed at run time. A patch from Google adds
1975        this functionality to the C++ interface.
1976    
1977    18. Changes to the handling of Unicode character properties:
1978    
1979        (a) Updated the table to Unicode 4.1.0.
1980    
1981        (b) Recognize characters that are not in the table as "Cn" (undefined).
1982    
1983        (c) I revised the way the table is implemented to a much improved format
1984            which includes recognition of ranges. It now supports the ranges that
1985            are defined in UnicodeData.txt, and it also amalgamates other
1986            characters into ranges. This has reduced the number of entries in the
1987            table from around 16,000 to around 3,000, thus reducing its size
1988            considerably. I realized I did not need to use a tree structure after
1989            all - a binary chop search is just as efficient. Having reduced the
1990            number of entries, I extended their size from 6 bytes to 8 bytes to
1991            allow for more data.
1992    
1993        (d) Added support for Unicode script names via properties such as \p{Han}.
1994    
1995    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1996        matching that character.
1997    
1998    20. When matching a repeated Unicode property with a minimum greater than zero,
1999        (for example \pL{2,}), PCRE could look past the end of the subject if it
2000        reached it while seeking the minimum number of characters. This could
2001        happen only if some of the characters were more than one byte long, because
2002        there is a check for at least the minimum number of bytes.
2003    
2004    21. Refactored the implementation of \p and \P so as to be more general, to
2005        allow for more different types of property in future. This has changed the
2006        compiled form incompatibly. Anybody with saved compiled patterns that use
2007        \p or \P will have to recompile them.
2008    
2009    22. Added "Any" and "L&" to the supported property types.
2010    
2011    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2012        but give a compile time error if the value is greater than 0xff.
2013    
2014    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2015        accidentally not being installed or uninstalled.
2016    
2017    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2018        made were to insert the current release number. This seemed silly, because
2019        it made things harder for people building PCRE on systems that don't run
2020        "configure". I have turned pcre.h into a distributed file, no longer built
2021        by "configure", with the version identification directly included. There is
2022        no longer a pcre.h.in file.
2023    
2024        However, this change necessitated a change to the pcre-config script as
2025        well. It is built from pcre-config.in, and one of the substitutions was the
2026        release number. I have updated configure.ac so that ./configure now finds
2027        the release number by grepping pcre.h.
2028    
2029    26. Added the ability to run the tests under valgrind.
2030    
2031    
2032  Version 6.4 05-Sep-05  Version 6.4 05-Sep-05
2033  ---------------------  ---------------------
2034    

Legend:
Removed from v.85  
changed lines
  Added in v.520

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12