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

Diff of /code/trunk/ChangeLog

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

revision 9 by nigel, Sat Feb 24 21:38:13 2007 UTC revision 545 by ph10, Wed Jun 16 10:51:15 2010 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.10 16-Jun-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        (#976).
47    
48    12. If a pattern that was studied started with a repeated Unicode property
49        test, for example, \p{Nd}+, there was the theoretical possibility of
50        setting up an incorrect bitmap of starting bytes, but fortunately it could
51        not have actually happened in practice until change 8 above was made (it
52        added property types that matched character-matching opcodes).
53    
54    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
55        possible starting bytes for non-anchored patterns.
56    
57    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
58        \R, and also a number of cases that involve Unicode properties, both
59        explicit and implicit when PCRE_UCP is set.
60    
61    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
62        input, it could crash or give wrong results if characters with values
63        greater than 0xc0 were present in the subject string. (Detail: it assumed
64        UTF-8 input when processing these items.)
65    
66    16. Added a lot of (int) casts to avoid compiler warnings in systems where
67        size_t is 64-bit (#991).
68    
69    17. Added a check for running out of memory when PCRE is compiled with
70        --disable-stack-for-recursion (#990).
71    
72    18. If the last data line in a file for pcretest does not have a newline on
73        the end, a newline was missing in the output.
74    
75    19. The default pcre_chartables.c file recognizes only ASCII characters (values
76        less than 128) in its various bitmaps. However, there is a facility for
77        generating tables according to the current locale when PCRE is compiled. It
78        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
79        characters, are recognized by isspace() and therefore were getting set in
80        these tables, and indeed these tables seem to approximate to ISO 8859. This
81        caused a problem in UTF-8 mode when pcre_study() was used to create a list
82        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
83        which of course cannot start UTF-8 characters. I have changed the code so
84        that only real ASCII characters (less than 128) and the correct starting
85        bytes for UTF-8 encodings are set for characters greater than 127 when in
86        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
87        altogether.)
88    
89    20. Added the /T option to pcretest so as to be able to run tests with non-
90        standard character tables, thus making it possible to include the tests
91        used for 19 above in the standard set of tests.
92    
93    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
94        reference to a subpattern the other side of a comment that contains an
95        opening parenthesis caused either an internal compiling error, or a
96        reference to the wrong subpattern.
97    
98    
99    Version 8.02 19-Mar-2010
100    ------------------------
101    
102    1.  The Unicode data tables have been updated to Unicode 5.2.0.
103    
104    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
105        configured.
106    
107    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
108        original author of that file, following a query about its status.
109    
110    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
111        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
112    
113    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
114        quantifier applied to a forward-referencing subroutine call, could compile
115        incorrect code or give the error "internal error: previously-checked
116        referenced subpattern not found".
117    
118    6.  Both MS Visual Studio and Symbian OS have problems with initializing
119        variables to point to external functions. For these systems, therefore,
120        pcre_malloc etc. are now initialized to local functions that call the
121        relevant global functions.
122    
123    7.  There were two entries missing in the vectors called coptable and poptable
124        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
125        I've fixed the data, and added a kludgy way of testing at compile time that
126        the lengths are correct (equal to the number of opcodes).
127    
128    8.  Following on from 7, I added a similar kludge to check the length of the
129        eint vector in pcreposix.c.
130    
131    9.  Error texts for pcre_compile() are held as one long string to avoid too
132        much relocation at load time. To find a text, the string is searched,
133        counting zeros. There was no check for running off the end of the string,
134        which could happen if a new error number was added without updating the
135        string.
136    
137    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
138    
139    11. \K was not working if it appeared in an atomic group or in a group that
140        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
141        \K is "not well defined" if used in an assertion. PCRE now accepts it if
142        the assertion is positive, but not if it is negative.
143    
144    12. Change 11 fortuitously reduced the size of the stack frame used in the
145        "match()" function of pcre_exec.c by one pointer. Forthcoming
146        implementation of support for (*MARK) will need an extra pointer on the
147        stack; I have reserved it now, so that the stack frame size does not
148        decrease.
149    
150    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
151        item in branch that calls a recursion is a subroutine call - as in the
152        second branch in the above example - was incorrectly given the compile-
153        time error "recursive call could loop indefinitely" because pcre_compile()
154        was not correctly checking the subroutine for matching a non-empty string.
155    
156    14. The checks for overrunning compiling workspace could trigger after an
157        overrun had occurred. This is a "should never occur" error, but it can be
158        triggered by pathological patterns such as hundreds of nested parentheses.
159        The checks now trigger 100 bytes before the end of the workspace.
160    
161    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
162    
163    
164    Version 8.01 19-Jan-2010
165    ------------------------
166    
167    1.  If a pattern contained a conditional subpattern with only one branch (in
168        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
169        computed the wrong minimum data length (which is of course zero for such
170        subpatterns). This could cause incorrect "no match" results.
171    
172    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
173        the pattern is reset in the first branch, pcre_compile() failed with
174        "internal error: code overflow at offset...". This happened only when
175        the reset was to the original external option setting. (An optimization
176        abstracts leading options settings into an external setting, which was the
177        cause of this.)
178    
179    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
180        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
181        assertion pattern did not match (meaning that the assertion was true), it
182        was incorrectly treated as false if the SKIP had been reached during the
183        matching. This also applied to assertions used as conditions.
184    
185    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
186        assertion subpattern, including such a pattern used as a condition,
187        unpredictable results occurred, instead of the error return
188        PCRE_ERROR_DFA_UITEM.
189    
190    5.  The C++ GlobalReplace function was not working like Perl for the special
191        situation when an empty string is matched. It now does the fancy magic
192        stuff that is necessary.
193    
194    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
195        removed. (These were left over from very, very early versions of PCRE.)
196    
197    7.  Some cosmetic changes to the code to make life easier when compiling it
198        as part of something else:
199    
200        (a) Change DEBUG to PCRE_DEBUG.
201    
202        (b) In pcre_compile(), rename the member of the "branch_chain" structure
203            called "current" as "current_branch", to prevent a collision with the
204            Linux macro when compiled as a kernel module.
205    
206        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
207            prevent a collision with the Linux macro when compiled as a kernel
208            module.
209    
210    8.  In pcre_compile() there are some checks for integer overflows that used to
211        cast potentially large values to (double). This has been changed to that
212        when building, a check for int64_t is made, and if it is found, it is used
213        instead, thus avoiding the use of floating point arithmetic. (There is no
214        other use of FP in PCRE.) If int64_t is not found, the fallback is to
215        double.
216    
217    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
218        2005 (difference between two addresses compared to an unsigned value).
219    
220    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
221        custom one, because of the following reported problem in Windows:
222    
223          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
224              under Win32.
225          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
226              therefore missing the function definition.
227          - The compiler thus generates a "C" signature for the test function.
228          - The linker fails to find the "C" function.
229          - PCRE fails to configure if asked to do so against libbz2.
230    
231    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
232        messages were output:
233    
234          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
235          rerunning libtoolize, to keep the correct libtool macros in-tree.
236          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
237    
238        I have done both of these things.
239    
240    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
241        most of the time, it *can* run out if it is given a pattern that contains a
242        runaway infinite recursion. I updated the discussion in the pcrestack man
243        page.
244    
245    13. Now that we have gone to the x.xx style of version numbers, the minor
246        version may start with zero. Using 08 or 09 is a bad idea because users
247        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
248        interpreted as invalid octal numbers. I've updated the previous comment in
249        configure.ac, and also added a check that gives an error if 08 or 09 are
250        used.
251    
252    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
253        causing partial matching to fail when the end of the subject matched \W
254        in a UTF-8 pattern where \W was quantified with a minimum of 3.
255    
256    15. There were some discrepancies between the declarations in pcre_internal.h
257        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
258        their definitions. The declarations used "const uschar *" and the
259        definitions used USPTR. Even though USPTR is normally defined as "const
260        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
261        reported that: "This difference in casting confuses some C++ compilers, for
262        example, SunCC recognizes above declarations as different functions and
263        generates broken code for hbpcre." I have changed the declarations to use
264        USPTR.
265    
266    16. GNU libtool is named differently on some systems. The autogen.sh script now
267        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
268        (FreeBSD).
269    
270    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
271        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
272        comment: "Figure out how to create a longlong from a string: strtoll and
273        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
274        instance, but it only takes 2 args instead of 3!"
275    
276    18. A subtle bug concerned with back references has been fixed by a change of
277        specification, with a corresponding code fix. A pattern such as
278        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
279        refers, was giving matches when it shouldn't. For example, xa=xaaa would
280        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
281        same bug. Such groups have to be quantified to be useful, or contained
282        inside another quantified group. (If there's no repetition, the reference
283        can never match.) The problem arises because, having left the group and
284        moved on to the rest of the pattern, a later failure that backtracks into
285        the group uses the captured value from the final iteration of the group
286        rather than the correct earlier one. I have fixed this in PCRE by forcing
287        any group that contains a reference to itself to be an atomic group; that
288        is, there cannot be any backtracking into it once it has completed. This is
289        similar to recursive and subroutine calls.
290    
291    
292    Version 8.00 19-Oct-09
293    ----------------------
294    
295    1.  The table for translating pcre_compile() error codes into POSIX error codes
296        was out-of-date, and there was no check on the pcre_compile() error code
297        being within the table. This could lead to an OK return being given in
298        error.
299    
300    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
301        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
302        in a Windows environment.
303    
304    3.  The pcregrep --count option prints the count for each file even when it is
305        zero, as does GNU grep. However, pcregrep was also printing all files when
306        --files-with-matches was added. Now, when both options are given, it prints
307        counts only for those files that have at least one match. (GNU grep just
308        prints the file name in this circumstance, but including the count seems
309        more useful - otherwise, why use --count?) Also ensured that the
310        combination -clh just lists non-zero counts, with no names.
311    
312    4.  The long form of the pcregrep -F option was incorrectly implemented as
313        --fixed_strings instead of --fixed-strings. This is an incompatible change,
314        but it seems right to fix it, and I didn't think it was worth preserving
315        the old behaviour.
316    
317    5.  The command line items --regex=pattern and --regexp=pattern were not
318        recognized by pcregrep, which required --regex pattern or --regexp pattern
319        (with a space rather than an '='). The man page documented the '=' forms,
320        which are compatible with GNU grep; these now work.
321    
322    6.  No libpcreposix.pc file was created for pkg-config; there was just
323        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
324    
325    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
326        when UCP support is not needed, by modifying the Python script that
327        generates it from Unicode data files. This should not matter if the module
328        is correctly used as a library, but I received one complaint about 50K of
329        unwanted data. My guess is that the person linked everything into his
330        program rather than using a library. Anyway, it does no harm.
331    
332    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
333        was a minimum greater than 1 for a wide character in a possessive
334        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
335        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
336        character. Chaos in the form of incorrect output or a compiling loop could
337        result.
338    
339    9.  The restrictions on what a pattern can contain when partial matching is
340        requested for pcre_exec() have been removed. All patterns can now be
341        partially matched by this function. In addition, if there are at least two
342        slots in the offset vector, the offset of the earliest inspected character
343        for the match and the offset of the end of the subject are set in them when
344        PCRE_ERROR_PARTIAL is returned.
345    
346    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
347        synonymous with PCRE_PARTIAL, for backwards compatibility, and
348        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
349        and may be more useful for multi-segment matching.
350    
351    11. Partial matching with pcre_exec() is now more intuitive. A partial match
352        used to be given if ever the end of the subject was reached; now it is
353        given only if matching could not proceed because another character was
354        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
355        string "Z", which now yields "no match" instead of "partial match". In the
356        case of pcre_dfa_exec(), "no match" is given if every matching path for the
357        final character ended with (*FAIL).
358    
359    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
360        if the pattern had a "must contain" character that was already found in the
361        earlier partial match, unless partial matching was again requested. For
362        example, with the pattern /dog.(body)?/, the "must contain" character is
363        "g". If the first part-match was for the string "dog", restarting with
364        "sbody" failed. This bug has been fixed.
365    
366    13. The string returned by pcre_dfa_exec() after a partial match has been
367        changed so that it starts at the first inspected character rather than the
368        first character of the match. This makes a difference only if the pattern
369        starts with a lookbehind assertion or \b or \B (\K is not supported by
370        pcre_dfa_exec()). It's an incompatible change, but it makes the two
371        matching functions compatible, and I think it's the right thing to do.
372    
373    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
374        so that the demonstration program is easily available in environments where
375        PCRE has not been installed from source.
376    
377    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
378        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
379        library.
380    
381    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
382        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
383        is not the first non-POSIX option to be added. Clearly some people find
384        these options useful.
385    
386    17. If a caller to the POSIX matching function regexec() passes a non-zero
387        value for nmatch with a NULL value for pmatch, the value of
388        nmatch is forced to zero.
389    
390    18. RunGrepTest did not have a test for the availability of the -u option of
391        the diff command, as RunTest does. It now checks in the same way as
392        RunTest, and also checks for the -b option.
393    
394    19. If an odd number of negated classes containing just a single character
395        interposed, within parentheses, between a forward reference to a named
396        subpattern and the definition of the subpattern, compilation crashed with
397        an internal error, complaining that it could not find the referenced
398        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
399        [The bug was that it was starting one character too far in when skipping
400        over the character class, thus treating the ] as data rather than
401        terminating the class. This meant it could skip too much.]
402    
403    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
404        /g option in pcretest when the pattern contains \K, which makes it possible
405        to have an empty string match not at the start, even when the pattern is
406        anchored. Updated pcretest and pcredemo to use this option.
407    
408    21. If the maximum number of capturing subpatterns in a recursion was greater
409        than the maximum at the outer level, the higher number was returned, but
410        with unset values at the outer level. The correct (outer level) value is
411        now given.
412    
413    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
414        PCRE did not set those parentheses (unlike Perl). I have now found a way to
415        make it do so. The string so far is captured, making this feature
416        compatible with Perl.
417    
418    23. The tests have been re-organized, adding tests 11 and 12, to make it
419        possible to check the Perl 5.10 features against Perl 5.10.
420    
421    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
422        pattern matches a fixed length string. PCRE did not allow this; now it
423        does. Neither allows recursion.
424    
425    25. I finally figured out how to implement a request to provide the minimum
426        length of subject string that was needed in order to match a given pattern.
427        (It was back references and recursion that I had previously got hung up
428        on.) This code has now been added to pcre_study(); it finds a lower bound
429        to the length of subject needed. It is not necessarily the greatest lower
430        bound, but using it to avoid searching strings that are too short does give
431        some useful speed-ups. The value is available to calling programs via
432        pcre_fullinfo().
433    
434    26. While implementing 25, I discovered to my embarrassment that pcretest had
435        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
436        study optimizations had never been tested with that matching function.
437        Oops. What is worse, even when it was passed study data, there was a bug in
438        pcre_dfa_exec() that meant it never actually used it. Double oops. There
439        were also very few tests of studied patterns with pcre_dfa_exec().
440    
441    27. If (?| is used to create subpatterns with duplicate numbers, they are now
442        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
443        on the other side of the coin, they are no longer allowed to have different
444        names, because these cannot be distinguished in PCRE, and this has caused
445        confusion. (This is a difference from Perl.)
446    
447    28. When duplicate subpattern names are present (necessarily with different
448        numbers, as required by 27 above), and a test is made by name in a
449        conditional pattern, either for a subpattern having been matched, or for
450        recursion in such a pattern, all the associated numbered subpatterns are
451        tested, and the overall condition is true if the condition is true for any
452        one of them. This is the way Perl works, and is also more like the way
453        testing by number works.
454    
455    
456    Version 7.9 11-Apr-09
457    ---------------------
458    
459    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
460        (pcretest), all targets were linked against these libraries. This included
461        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
462        libraries. This caused unwanted dependencies to be created. This problem
463        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
464        pcretest is linked with readline.
465    
466    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
467        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
468        moved outside it again, because FALSE and TRUE are already defined in AIX,
469        but BOOL is not.
470    
471    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
472        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
473    
474    4.  The pcregrep documentation said spaces were inserted as well as colons (or
475        hyphens) following file names and line numbers when outputting matching
476        lines. This is not true; no spaces are inserted. I have also clarified the
477        wording for the --colour (or --color) option.
478    
479    5.  In pcregrep, when --colour was used with -o, the list of matching strings
480        was not coloured; this is different to GNU grep, so I have changed it to be
481        the same.
482    
483    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
484        each matching line was coloured. Now it goes on to look for further matches
485        of any of the test patterns, which is the same behaviour as GNU grep.
486    
487    7.  A pattern that could match an empty string could cause pcregrep to loop; it
488        doesn't make sense to accept an empty string match in pcregrep, so I have
489        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
490        seems to be how GNU grep behaves.
491    
492    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
493        start or after a newline", because the conditional assertion was not being
494        correctly handled. The rule now is that both the assertion and what follows
495        in the first alternative must satisfy the test.
496    
497    9.  If auto-callout was enabled in a pattern with a conditional group whose
498        condition was an assertion, PCRE could crash during matching, both with
499        pcre_exec() and pcre_dfa_exec().
500    
501    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
502        used for matching.
503    
504    11. Unicode property support in character classes was not working for
505        characters (bytes) greater than 127 when not in UTF-8 mode.
506    
507    12. Added the -M command line option to pcretest.
508    
509    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
510    
511    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
512    
513    16. Added comments and documentation about mis-use of no_arg in the C++
514        wrapper.
515    
516    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
517        from Martin Jerabek that uses macro names for all relevant character and
518        string constants.
519    
520    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
521        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
522        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
523        these, but not everybody uses configure.
524    
525    19. A conditional group that had only one branch was not being correctly
526        recognized as an item that could match an empty string. This meant that an
527        enclosing group might also not be so recognized, causing infinite looping
528        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
529        with the subject "ab", where knowledge that the repeated group can match
530        nothing is needed in order to break the loop.
531    
532    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
533        exec(), but without supplying a callout function, matching went wrong.
534    
535    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
536        leak if the size of the offset vector was greater than 30. When the vector
537        is smaller, the saved offsets during recursion go onto a local stack
538        vector, but for larger vectors malloc() is used. It was failing to free
539        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
540        error, in fact).
541    
542    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
543        heapframe that is used only when UTF-8 support is enabled. This caused no
544        problem, but was untidy.
545    
546    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
547        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
548        included within another project.
549    
550    24. Steven Van Ingelgem's patches to add more options to the CMake support,
551        slightly modified by me:
552    
553          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
554              not building pcregrep.
555    
556          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
557              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
558    
559    25. Forward references, both numeric and by name, in patterns that made use of
560        duplicate group numbers, could behave incorrectly or give incorrect errors,
561        because when scanning forward to find the reference group, PCRE was not
562        taking into account the duplicate group numbers. A pattern such as
563        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
564    
565    26. Changed a few more instances of "const unsigned char *" to USPTR, making
566        the feature of a custom pointer more persuasive (as requested by a user).
567    
568    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
569        pcretest.c, inside #ifndefs, because it seems they are sometimes already
570        pre-defined.
571    
572    28. Added support for (*UTF8) at the start of a pattern.
573    
574    29. Arrange for flags added by the "release type" setting in CMake to be shown
575        in the configuration summary.
576    
577    
578    Version 7.8 05-Sep-08
579    ---------------------
580    
581    1.  Replaced UCP searching code with optimized version as implemented for Ad
582        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
583        stage table and inline lookup instead of a function, giving speed ups of 2
584        to 5 times on some simple patterns that I tested. Permission was given to
585        distribute the MultiStage2.py script that generates the tables (it's not in
586        the tarball, but is in the Subversion repository).
587    
588    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
589        scripts.
590    
591    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
592        a group with a zero qualifier. The result of the study could be incorrect,
593        or the function might crash, depending on the pattern.
594    
595    4.  Caseless matching was not working for non-ASCII characters in back
596        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
597        It now works when Unicode Property Support is available.
598    
599    5.  In pcretest, an escape such as \x{de} in the data was always generating
600        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
601        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
602        truncation.
603    
604    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
605    
606    7.  Added two (int) casts to pcregrep when printing the difference of two
607        pointers, in case they are 64-bit values.
608    
609    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
610        test 2 if it fails.
611    
612    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
613        and a #define of that name to empty if it is not externally set. This is to
614        allow users of MSVC to set it if necessary.
615    
616    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
617        the convenience functions in the pcre_get.c source file.
618    
619    11. An option change at the start of a pattern that had top-level alternatives
620        could cause overwriting and/or a crash. This command provoked a crash in
621        some environments:
622    
623          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
624    
625        This potential security problem was recorded as CVE-2008-2371.
626    
627    12. For a pattern where the match had to start at the beginning or immediately
628        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
629        pcre_dfa_exec() could read past the end of the passed subject if there was
630        no match. To help with detecting such bugs (e.g. with valgrind), I modified
631        pcretest so that it places the subject at the end of its malloc-ed buffer.
632    
633    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
634        exec() might read past the end of the data buffer in UTF-8 mode.
635    
636    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
637        the data contained the byte 0x85 as part of a UTF-8 character within its
638        first line. This applied both to normal and DFA matching.
639    
640    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
641        /^[^d]*?$/8 failed to match "abc".
642    
643    16. Added a missing copyright notice to pcrecpp_internal.h.
644    
645    17. Make it more clear in the documentation that values returned from
646        pcre_exec() in ovector are byte offsets, not character counts.
647    
648    18. Tidied a few places to stop certain compilers from issuing warnings.
649    
650    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
651        supplied by Stefan Weber. I made a further small update for 7.8 because
652        there is a change of source arrangements: the pcre_searchfuncs.c module is
653        replaced by pcre_ucd.c.
654    
655    
656    Version 7.7 07-May-08
657    ---------------------
658    
659    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
660        a string to a long long, pretend we don't even have a long long." This is
661        done by checking for the strtoq, strtoll, and _strtoi64 functions.
662    
663    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
664        pre-7.6 versions, which defined a global no_arg variable instead of putting
665        it in the RE class. (See also #8 below.)
666    
667    3.  Remove a line of dead code, identified by coverity and reported by Nuno
668        Lopes.
669    
670    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
671    
672        (1) The include/exclude patterns were being applied to the whole pathnames
673            of files, instead of just to the final components.
674    
675        (2) If there was more than one level of directory, the subdirectories were
676            skipped unless they satisfied the include/exclude conditions. This is
677            inconsistent with GNU grep (and could even be seen as contrary to the
678            pcregrep specification - which I improved to make it absolutely clear).
679            The action now is always to scan all levels of directory, and just
680            apply the include/exclude patterns to regular files.
681    
682    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
683        --exclude_dir in the tests to avoid scanning .svn directories.
684    
685    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
686        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
687        doesn't support NULs in patterns.
688    
689    7.  Added some missing "const"s to declarations of static tables in
690        pcre_compile.c and pcre_dfa_exec.c.
691    
692    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
693        caused by fix #2  above. (Subsequently also a second patch to fix the
694        first patch. And a third patch - this was a messy problem.)
695    
696    9.  Applied Craig's patch to remove the use of push_back().
697    
698    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
699        matching function regexec().
700    
701    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
702        which, however, unlike Perl's \g{...}, are subroutine calls, not back
703        references. PCRE supports relative numbers with this syntax (I don't think
704        Oniguruma does).
705    
706    12. Previously, a group with a zero repeat such as (...){0} was completely
707        omitted from the compiled regex. However, this means that if the group
708        was called as a subroutine from elsewhere in the pattern, things went wrong
709        (an internal error was given). Such groups are now left in the compiled
710        pattern, with a new opcode that causes them to be skipped at execution
711        time.
712    
713    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
714        to the way PCRE behaves:
715    
716        (a) A lone ] character is dis-allowed (Perl treats it as data).
717    
718        (b) A back reference to an unmatched subpattern matches an empty string
719            (Perl fails the current match path).
720    
721        (c) A data ] in a character class must be notated as \] because if the
722            first data character in a class is ], it defines an empty class. (In
723            Perl it is not possible to have an empty class.) The empty class []
724            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
725            The negative empty class [^] matches any one character, independently
726            of the DOTALL setting.
727    
728    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
729        non-existent subpattern following a character class starting with ']' and
730        containing () gave an internal compiling error instead of "reference to
731        non-existent subpattern". Fortunately, when the pattern did exist, the
732        compiled code was correct. (When scanning forwards to check for the
733        existencd of the subpattern, it was treating the data ']' as terminating
734        the class, so got the count wrong. When actually compiling, the reference
735        was subsequently set up correctly.)
736    
737    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
738        it was being rejected as not supported by pcre_dfa_exec(), even though
739        other assertions are supported. I have made pcre_dfa_exec() support
740        (*FAIL).
741    
742    16. The implementation of 13c above involved the invention of a new opcode,
743        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
744        cannot be changed at match time, I realized I could make a small
745        improvement to matching performance by compiling OP_ALLANY instead of
746        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
747        on the OP_ANY path.
748    
749    17. Compiling pcretest on Windows with readline support failed without the
750        following two fixes: (1) Make the unistd.h include conditional on
751        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
752    
753    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
754        ncurses library to be included for pcretest when ReadLine support is
755        requested, but also to allow for it to be overridden. This patch came from
756        Daniel Bergström.
757    
758    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
759        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
760        any errors with the current Unicode tables. Thanks to Peter Kankowski for
761        spotting this.
762    
763    
764    Version 7.6 28-Jan-08
765    ---------------------
766    
767    1.  A character class containing a very large number of characters with
768        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
769        overflow.
770    
771    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
772        HAVE_LONG_LONG is not defined.
773    
774    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
775        bring them up to date and include new features. This patch includes:
776    
777        - Fixed PH's badly added libz and libbz2 support.
778        - Fixed a problem with static linking.
779        - Added pcredemo. [But later removed - see 7 below.]
780        - Fixed dftables problem and added an option.
781        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
782            HAVE_LONG_LONG.
783        - Added readline support for pcretest.
784        - Added an listing of the option settings after cmake has run.
785    
786    4.  A user submitted a patch to Makefile that makes it easy to create
787        "pcre.dll" under mingw when using Configure/Make. I added stuff to
788        Makefile.am that cause it to include this special target, without
789        affecting anything else. Note that the same mingw target plus all
790        the other distribution libraries and programs are now supported
791        when configuring with CMake (see 6 below) instead of with
792        Configure/Make.
793    
794    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
795        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
796        exported in the Windows port". It has not yet been confirmed that the patch
797        solves the problem, but it does no harm.
798    
799    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
800        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
801        with CMake, and also correct the comment about stack recursion.
802    
803    7.  Remove the automatic building of pcredemo from the ./configure system and
804        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
805        of a program that users should build themselves after PCRE is installed, so
806        building it automatically is not really right. What is more, it gave
807        trouble in some build environments.
808    
809    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
810    
811    
812    Version 7.5 10-Jan-08
813    ---------------------
814    
815    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
816        values in parens when parsing an RE using the C++ wrapper."
817    
818    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
819        Characters greater than 255 were excluded from the class instead of being
820        included.
821    
822    3.  The same bug as (2) above applied to negated POSIX classes such as
823        [:^space:].
824    
825    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
826        defined or documented. It seems to have been a typo for PCRE_STATIC, so
827        I have changed it.
828    
829    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
830        first named subpattern) and a construct such as (?&a) would reference the
831        first named subpattern whose name started with "a" (in other words, the
832        length check was missing). Both these problems are fixed. "Subpattern name
833        expected" is now given for (?&) (a zero-length name), and this patch also
834        makes it give the same error for \k'' (previously it complained that that
835        was a reference to a non-existent subpattern).
836    
837    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
838        this is right because (?- can be followed by option settings as well as by
839        digits. I have, however, made the messages clearer.
840    
841    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
842        than the number used in the conditional) now cause a compile-time error.
843        This is actually not compatible with Perl, which accepts such patterns, but
844        treats the conditional as always being FALSE (as PCRE used to), but it
845        seems to me that giving a diagnostic is better.
846    
847    8.  Change "alphameric" to the more common word "alphanumeric" in comments
848        and messages.
849    
850    9.  Fix two occurrences of "backslash" in comments that should have been
851        "backspace".
852    
853    10. Remove two redundant lines of code that can never be obeyed (their function
854        was moved elsewhere).
855    
856    11. The program that makes PCRE's Unicode character property table had a bug
857        which caused it to generate incorrect table entries for sequences of
858        characters that have the same character type, but are in different scripts.
859        It amalgamated them into a single range, with the script of the first of
860        them. In other words, some characters were in the wrong script. There were
861        thirteen such cases, affecting characters in the following ranges:
862    
863          U+002b0 - U+002c1
864          U+0060c - U+0060d
865          U+0061e - U+00612
866          U+0064b - U+0065e
867          U+0074d - U+0076d
868          U+01800 - U+01805
869          U+01d00 - U+01d77
870          U+01d9b - U+01dbf
871          U+0200b - U+0200f
872          U+030fc - U+030fe
873          U+03260 - U+0327f
874          U+0fb46 - U+0fbb1
875          U+10450 - U+1049d
876    
877    12. The -o option (show only the matching part of a line) for pcregrep was not
878        compatible with GNU grep in that, if there was more than one match in a
879        line, it showed only the first of them. It now behaves in the same way as
880        GNU grep.
881    
882    13. If the -o and -v options were combined for pcregrep, it printed a blank
883        line for every non-matching line. GNU grep prints nothing, and pcregrep now
884        does the same. The return code can be used to tell if there were any
885        non-matching lines.
886    
887    14. Added --file-offsets and --line-offsets to pcregrep.
888    
889    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
890        infinitely looping recursion. The bug was that positive lookaheads were not
891        being skipped when checking for a possible empty match (negative lookaheads
892        and both kinds of lookbehind were skipped).
893    
894    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
895        inclusion of <windows.h> to before rather than after the definition of
896        INVALID_FILE_ATTRIBUTES (patch from David Byron).
897    
898    17. Specifying a possessive quantifier with a specific limit for a Unicode
899        character property caused pcre_compile() to compile bad code, which led at
900        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
901        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
902        caused the error; without that there was no problem.
903    
904    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
905    
906    19. Added --enable-pcretest-libreadline.
907    
908    20. In pcrecpp.cc, the variable 'count' was incremented twice in
909        RE::GlobalReplace(). As a result, the number of replacements returned was
910        double what it should be. I removed one of the increments, but Craig sent a
911        later patch that removed the other one (the right fix) and added unit tests
912        that check the return values (which was not done before).
913    
914    21. Several CMake things:
915    
916        (1) Arranged that, when cmake is used on Unix, the libraries end up with
917            the names libpcre and libpcreposix, not just pcre and pcreposix.
918    
919        (2) The above change means that pcretest and pcregrep are now correctly
920            linked with the newly-built libraries, not previously installed ones.
921    
922        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
923    
924    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
925        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
926        UTF-8 newline character). The key issue is that the pattern starts .*;
927        this means that the match must be either at the beginning, or after a
928        newline. The bug was in the code for advancing after a failed match and
929        checking that the new position followed a newline. It was not taking
930        account of UTF-8 characters correctly.
931    
932    23. PCRE was behaving differently from Perl in the way it recognized POSIX
933        character classes. PCRE was not treating the sequence [:...:] as a
934        character class unless the ... were all letters. Perl, however, seems to
935        allow any characters between [: and :], though of course it rejects as
936        unknown any "names" that contain non-letters, because all the known class
937        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
938        for example, whereas PCRE did not - it did not recognize a POSIX character
939        class. This seemed a bit dangerous, so the code has been changed to be
940        closer to Perl. The behaviour is not identical to Perl, because PCRE will
941        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
942        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
943        Perl does, and where it didn't before.
944    
945    24. Rewrite so as to remove the single use of %n from pcregrep because in some
946        Windows environments %n is disabled by default.
947    
948    
949    Version 7.4 21-Sep-07
950    ---------------------
951    
952    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
953        means that a class such as [\s] counted as "explicit reference to CR or
954        LF". That isn't really right - the whole point of the change was to try to
955        help when there was an actual mention of one of the two characters. So now
956        the change happens only if \r or \n (or a literal CR or LF) character is
957        encountered.
958    
959    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
960        of both had grown to the point where there were only 3 bits left.
961        Fortunately, there was spare space in the data structure, and so I have
962        moved the internal flags into a new 16-bit field to free up more option
963        bits.
964    
965    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
966        but did not set the internal JCHANGED flag - either of these is enough to
967        control the way the "get" function works - but the PCRE_INFO_JCHANGED
968        facility is supposed to tell if (?J) was ever used, so now (?J) at the
969        start sets both bits.
970    
971    4.  Added options (at build time, compile time, exec time) to change \R from
972        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
973    
974    5.  doc/pcresyntax.html was missing from the distribution.
975    
976    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
977        compatibility, even though it is no longer used.
978    
979    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
980        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
981        windows.h file is present (where different names are used). [This was
982        reversed later after testing - see 16 below.]
983    
984    8.  Changed all #include <config.h> to #include "config.h". There were also
985        some further <pcre.h> cases that I changed to "pcre.h".
986    
987    9.  When pcregrep was used with the --colour option, it missed the line ending
988        sequence off the lines that it output.
989    
990    10. It was pointed out to me that arrays of string pointers cause lots of
991        relocations when a shared library is dynamically loaded. A technique of
992        using a single long string with a table of offsets can drastically reduce
993        these. I have refactored PCRE in four places to do this. The result is
994        dramatic:
995    
996          Originally:                          290
997          After changing UCP table:            187
998          After changing error message table:   43
999          After changing table of "verbs"       36
1000          After changing table of Posix names   22
1001    
1002        Thanks to the folks working on Gregex for glib for this insight.
1003    
1004    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1005        unicode-properties was also set.
1006    
1007    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1008    
1009    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1010        checked only for CRLF.
1011    
1012    14. Added casts to pcretest.c to avoid compiler warnings.
1013    
1014    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1015    
1016    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1017        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1018        entirely. This removes changes made in 7 above.
1019    
1020    17. The CMake files have been updated, and there is now more information about
1021        building with CMake in the NON-UNIX-USE document.
1022    
1023    
1024    Version 7.3 28-Aug-07
1025    ---------------------
1026    
1027     1. In the rejigging of the build system that eventually resulted in 7.1, the
1028        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1029        brackets there is not right, since it causes compilers to look for an
1030        installed pcre.h, not the version that is in the source that is being
1031        compiled (which of course may be different). I have changed it back to:
1032    
1033          #include "pcre.h"
1034    
1035        I have a vague recollection that the change was concerned with compiling in
1036        different directories, but in the new build system, that is taken care of
1037        by the VPATH setting the Makefile.
1038    
1039     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1040        when the subject happened to end in the byte 0x85 (e.g. if the last
1041        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1042        characters but of course it shouldn't be taken as a newline when it is part
1043        of another character. The bug was that, for an unlimited repeat of . in
1044        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1045        characters when looking for a newline.
1046    
1047     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1048    
1049     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1050        in debug output.
1051    
1052     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1053        long printing in the pcrecpp unittest when running under MinGW.
1054    
1055     6. ESC_K was left out of the EBCDIC table.
1056    
1057     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1058        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1059        limit also applies to "virtual nesting" when a pattern is recursive, and in
1060        this case 1000 isn't so big. I have been able to remove this limit at the
1061        expense of backing off one optimization in certain circumstances. Normally,
1062        when pcre_exec() would call its internal match() function recursively and
1063        immediately return the result unconditionally, it uses a "tail recursion"
1064        feature to save stack. However, when a subpattern that can match an empty
1065        string has an unlimited repetition quantifier, it no longer makes this
1066        optimization. That gives it a stack frame in which to save the data for
1067        checking that an empty string has been matched. Previously this was taken
1068        from the 1000-entry workspace that had been reserved. So now there is no
1069        explicit limit, but more stack is used.
1070    
1071     8. Applied Daniel's patches to solve problems with the import/export magic
1072        syntax that is required for Windows, and which was going wrong for the
1073        pcreposix and pcrecpp parts of the library. These were overlooked when this
1074        problem was solved for the main library.
1075    
1076     9. There were some crude static tests to avoid integer overflow when computing
1077        the size of patterns that contain repeated groups with explicit upper
1078        limits. As the maximum quantifier is 65535, the maximum group length was
1079        set at 30,000 so that the product of these two numbers did not overflow a
1080        32-bit integer. However, it turns out that people want to use groups that
1081        are longer than 30,000 bytes (though not repeat them that many times).
1082        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1083        made it possible to implement the integer overflow checks in a much more
1084        dynamic way, which I have now done. The artificial limitation on group
1085        length has been removed - we now have only the limit on the total length of
1086        the compiled pattern, which depends on the LINK_SIZE setting.
1087    
1088    10. Fixed a bug in the documentation for get/copy named substring when
1089        duplicate names are permitted. If none of the named substrings are set, the
1090        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1091        empty string.
1092    
1093    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1094        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1095        because the ] is interpreted as the first data character and the
1096        terminating ] is not found. PCRE has been made compatible with Perl in this
1097        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1098        cause memory overwriting.
1099    
1100    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1101        string has been matched (to stop an infinite loop). It was not recognizing
1102        a conditional subpattern that could match an empty string if that
1103        subpattern was within another subpattern. For example, it looped when
1104        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1105        condition was not nested. This bug has been fixed.
1106    
1107    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1108        past the start of the subject in the presence of bytes with the top bit
1109        set, for example "\x8aBCD".
1110    
1111    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1112        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1113    
1114    14. Optimized (?!) to (*FAIL).
1115    
1116    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1117        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1118        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1119        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1120        does: it's just the validity check that is more restrictive.
1121    
1122    16. Inserted checks for integer overflows during escape sequence (backslash)
1123        processing, and also fixed erroneous offset values for syntax errors during
1124        backslash processing.
1125    
1126    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1127        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1128    
1129    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1130        caused an overrun.
1131    
1132    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1133        something other than just ASCII characters) inside a group that had an
1134        unlimited repeat caused a loop at compile time (while checking to see
1135        whether the group could match an empty string).
1136    
1137    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1138        [\P{Any}] did so. (Error in the code for printing property names.)
1139    
1140    21. An orphan \E inside a character class could cause a crash.
1141    
1142    22. A repeated capturing bracket such as (A)? could cause a wild memory
1143        reference during compilation.
1144    
1145    23. There are several functions in pcre_compile() that scan along a compiled
1146        expression for various reasons (e.g. to see if it's fixed length for look
1147        behind). There were bugs in these functions when a repeated \p or \P was
1148        present in the pattern. These operators have additional parameters compared
1149        with \d, etc, and these were not being taken into account when moving along
1150        the compiled data. Specifically:
1151    
1152        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1153            length.
1154    
1155        (b) An item such as \pL+ within a repeated group could cause crashes or
1156            loops.
1157    
1158        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1159            "reference to non-existent subpattern" error.
1160    
1161        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1162    
1163    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1164        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1165    
1166    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1167    
1168    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1169        character were causing crashes (broken optimization).
1170    
1171    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1172        \p or \P) caused a compile-time loop.
1173    
1174    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1175        break. For example, the unstudied pattern [\r\n]A does not match the string
1176        "\r\nA" because change 7.0/46 below moves the current point on by two
1177        characters after failing to match at the start. However, the pattern \nA
1178        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1179        the same is true. There doesn't seem any very clean way out of this, but
1180        what I have chosen to do makes the common cases work: PCRE now takes note
1181        of whether there can be an explicit match for \r or \n anywhere in the
1182        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1183        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1184        pattern has explicit CR or LF references.
1185    
1186    29. Added (*CR) etc for changing newline setting at start of pattern.
1187    
1188    
1189    Version 7.2 19-Jun-07
1190    ---------------------
1191    
1192     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1193        which is apparently normally available under Windows.
1194    
1195     2. Re-jig the pcregrep tests with different newline settings in an attempt
1196        to make them independent of the local environment's newline setting.
1197    
1198     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1199    
1200     4. Some of the "internals" tests were previously cut out when the link size
1201        was not 2, because the output contained actual offsets. The recent new
1202        "Z" feature of pcretest means that these can be cut out, making the tests
1203        usable with all link sizes.
1204    
1205     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1206        stack recursion. This gives a massive performance boost under BSD, but just
1207        a small improvement under Linux. However, it saves one field in the frame
1208        in all cases.
1209    
1210     6. Added more features from the forthcoming Perl 5.10:
1211    
1212        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1213            recursion call. It refers to the nth most recently opened parentheses.
1214    
1215        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1216            to be opened parentheses.
1217    
1218        (c) Conditions that refer to capturing parentheses can be specified
1219            relatively, for example, (?(-2)... or (?(+3)...
1220    
1221        (d) \K resets the start of the current match so that everything before
1222            is not part of it.
1223    
1224        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1225    
1226        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1227            reference syntax.
1228    
1229        (g) (?| introduces a group in which the numbering of parentheses in each
1230            alternative starts with the same number.
1231    
1232        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1233    
1234     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1235        PCRE_INFO_JCHANGED.
1236    
1237     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1238        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1239        for detecting groups that can match an empty string.
1240    
1241     9. A pattern with a very large number of alternatives (more than several
1242        hundred) was running out of internal workspace during the pre-compile
1243        phase, where pcre_compile() figures out how much memory will be needed. A
1244        bit of new cunning has reduced the workspace needed for groups with
1245        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1246        workspace instead of running out of the 4096 that are available.
1247    
1248    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1249    
1250    11. Applied patch from Google to remove an optimization that didn't quite work.
1251        The report of the bug said:
1252    
1253          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1254          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1255          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1256    
1257    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1258        it matched the wrong number of bytes.
1259    
1260    
1261    Version 7.1 24-Apr-07
1262    ---------------------
1263    
1264     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1265        that is more "standard", making use of automake and other Autotools. There
1266        is some re-arrangement of the files and adjustment of comments consequent
1267        on this.
1268    
1269     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1270        for recursive directory scanning broke on some systems because the files
1271        are not scanned in any specific order and on different systems the order
1272        was different. A call to "sort" has been inserted into RunGrepTest for the
1273        approprate test as a short-term fix. In the longer term there may be an
1274        alternative.
1275    
1276     3. I had an email from Eric Raymond about problems translating some of PCRE's
1277        man pages to HTML (despite the fact that I distribute HTML pages, some
1278        people do their own conversions for various reasons). The problems
1279        concerned the use of low-level troff macros .br and .in. I have therefore
1280        removed all such uses from the man pages (some were redundant, some could
1281        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1282        HTML has been updated to handle .nf/.fi and to complain if it encounters
1283        .br or .in.
1284    
1285     4. Updated comments in configure.ac that get placed in config.h.in and also
1286        arranged for config.h to be included in the distribution, with the name
1287        config.h.generic, for the benefit of those who have to compile without
1288        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1289    
1290     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1291        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1292        makevp.bat for the current PCRE, using the additional files
1293        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1294    
1295     6. A Windows user reported a minor discrepancy with test 2, which turned out
1296        to be caused by a trailing space on an input line that had got lost in his
1297        copy. The trailing space was an accident, so I've just removed it.
1298    
1299     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1300        that is needed.
1301    
1302     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1303        as "const" (a) because they are and (b) because it helps the PHP
1304        maintainers who have recently made a script to detect big data structures
1305        in the php code that should be moved to the .rodata section. I remembered
1306        to update Builducptable as well, so it won't revert if ucptable.h is ever
1307        re-created.
1308    
1309     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1310        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1311        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1312        support is not required. This saves 1.5-2K of code, which is important in
1313        some applications.
1314    
1315        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1316        so as not to refer to the tables, even though these functions will never be
1317        called when UTF-8 support is disabled. Otherwise there are problems with a
1318        shared library.
1319    
1320    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1321    
1322        (a) It was defining its arguments as char * instead of void *.
1323    
1324        (b) It was assuming that all moves were upwards in memory; this was true
1325            a long time ago when I wrote it, but is no longer the case.
1326    
1327        The emulated memove() is provided for those environments that have neither
1328        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1329        is clearly not the case, as these two bugs were recently reported.
1330    
1331    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1332        and Detrail to create the HTML documentation, the .txt form of the man
1333        pages, and it removes trailing spaces from listed files. It also creates
1334        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1335        case, it wraps all the #defines with #ifndefs. This script should be run
1336        before "make dist".
1337    
1338    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1339        with Unicode property support.
1340    
1341        (a) For a maximizing quantifier, if the two different cases of the
1342            character were of different lengths in their UTF-8 codings (there are
1343            some cases like this - I found 11), and the matching function had to
1344            back up over a mixture of the two cases, it incorrectly assumed they
1345            were both the same length.
1346    
1347        (b) When PCRE was configured to use the heap rather than the stack for
1348            recursion during matching, it was not correctly preserving the data for
1349            the other case of a UTF-8 character when checking ahead for a match
1350            while processing a minimizing repeat. If the check also involved
1351            matching a wide character, but failed, corruption could cause an
1352            erroneous result when trying to check for a repeat of the original
1353            character.
1354    
1355    13. Some tidying changes to the testing mechanism:
1356    
1357        (a) The RunTest script now detects the internal link size and whether there
1358            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1359            values substituted by "configure". (The RunGrepTest script already did
1360            this for UTF-8.) The configure.ac script no longer substitutes the
1361            relevant variables.
1362    
1363        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1364            with length and offset values. This means that the output is different
1365            for different internal link sizes. Test 2 is skipped for link sizes
1366            other than 2 because of this, bypassing the problem. Unfortunately,
1367            there was also a test in test 3 (the locale tests) that used /B and
1368            failed for link sizes other than 2. Rather than cut the whole test out,
1369            I have added a new /Z option to pcretest that replaces the length and
1370            offset values with spaces. This is now used to make test 3 independent
1371            of link size. (Test 2 will be tidied up later.)
1372    
1373    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1374        segmentation fault instead of returning the appropriate error message.
1375    
1376    15. In multiline mode when the newline sequence was set to "any", the pattern
1377        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1378        This doesn't seem right; it now treats the CRLF combination as the line
1379        ending, and so does not match in that case. It's only a pattern such as ^$
1380        that would hit this one: something like ^ABC$ would have failed after \r
1381        and then tried again after \r\n.
1382    
1383    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1384        in an attempt to make files that differ only in their line terminators
1385        compare equal. This works on Linux.
1386    
1387    17. Under certain error circumstances pcregrep might try to free random memory
1388        as it exited. This is now fixed, thanks to valgrind.
1389    
1390    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1391        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1392        was because its rules for how to advance for /g after matching an empty
1393        string at the end of a line did not allow for this case. They now check for
1394        it specially.
1395    
1396    20. pcretest is supposed to handle patterns and data of any length, by
1397        extending its buffers when necessary. It was getting this wrong when the
1398        buffer for a data line had to be extended.
1399    
1400    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1401        CRLF as a newline sequence.
1402    
1403    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1404        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1405        I have nevertheless tidied it up.
1406    
1407    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1408    
1409    24. Added a man page for pcre-config.
1410    
1411    
1412    Version 7.0 19-Dec-06
1413    ---------------------
1414    
1415     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1416        moving to gcc 4.1.1.
1417    
1418     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1419        sys/time.h, which is documented as needed for this function. It doesn't
1420        seem to matter on Linux, but it showed up on some releases of OS X.
1421    
1422     3. It seems that there are systems where bytes whose values are greater than
1423        127 match isprint() in the "C" locale. The "C" locale should be the
1424        default when a C program starts up. In most systems, only ASCII printing
1425        characters match isprint(). This difference caused the output from pcretest
1426        to vary, making some of the tests fail. I have changed pcretest so that:
1427    
1428        (a) When it is outputting text in the compiled version of a pattern, bytes
1429            other than 32-126 are always shown as hex escapes.
1430    
1431        (b) When it is outputting text that is a matched part of a subject string,
1432            it does the same, unless a different locale has been set for the match
1433            (using the /L modifier). In this case, it uses isprint() to decide.
1434    
1435     4. Fixed a major bug that caused incorrect computation of the amount of memory
1436        required for a compiled pattern when options that changed within the
1437        pattern affected the logic of the preliminary scan that determines the
1438        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1439        that the computed length was too small. The symptoms of this bug were
1440        either the PCRE error "internal error: code overflow" from pcre_compile(),
1441        or a glibc crash with a message such as "pcretest: free(): invalid next
1442        size (fast)". Examples of patterns that provoked this bug (shown in
1443        pcretest format) are:
1444    
1445          /(?-x: )/x
1446          /(?x)(?-x: \s*#\s*)/
1447          /((?i)[\x{c0}])/8
1448          /(?i:[\x{c0}])/8
1449    
1450        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1451        is now done differently.
1452    
1453     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1454        wrapper classes; (b) implement a new function in the C++ scanner that is
1455        more efficient than the old way of doing things because it avoids levels of
1456        recursion in the regex matching; (c) add a paragraph to the documentation
1457        for the FullMatch() function.
1458    
1459     6. The escape sequence \n was being treated as whatever was defined as
1460        "newline". Not only was this contrary to the documentation, which states
1461        that \n is character 10 (hex 0A), but it also went horribly wrong when
1462        "newline" was defined as CRLF. This has been fixed.
1463    
1464     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1465        was being set to -1 for the "end of line" case (supposedly a value that no
1466        character can have). Though this value is never used (the check for end of
1467        line is "zero bytes in current character"), it caused compiler complaints.
1468        I've changed it to 0xffffffff.
1469    
1470     8. In pcre_version.c, the version string was being built by a sequence of
1471        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1472        string (as it is for production releases) called a macro with an empty
1473        argument. The C standard says the result of this is undefined. The gcc
1474        compiler treats it as an empty string (which was what was wanted) but it is
1475        reported that Visual C gives an error. The source has been hacked around to
1476        avoid this problem.
1477    
1478     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1479        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1480        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1481        of them did).
1482    
1483    10. Originally, pcretest opened its input and output without "b"; then I was
1484        told that "b" was needed in some environments, so it was added for release
1485        5.0 to both the input and output. (It makes no difference on Unix-like
1486        systems.) Later I was told that it is wrong for the input on Windows. I've
1487        now abstracted the modes into two macros, to make it easier to fiddle with
1488        them, and removed "b" from the input mode under Windows.
1489    
1490    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1491    
1492    12. Added -help and --help to pcretest as an official way of being reminded
1493        of the options.
1494    
1495    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1496        and pcrecpp.cc because they annoy compilers at high warning levels.
1497    
1498    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1499    
1500    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1501        scripts are not C programs :-) and which was not noticed because it works
1502        on Linux.
1503    
1504    16. pcretest is supposed to handle any length of pattern and data line (as one
1505        line or as a continued sequence of lines) by extending its input buffer if
1506        necessary. This feature was broken for very long pattern lines, leading to
1507        a string of junk being passed to pcre_compile() if the pattern was longer
1508        than about 50K.
1509    
1510    17. I have done a major re-factoring of the way pcre_compile() computes the
1511        amount of memory needed for a compiled pattern. Previously, there was code
1512        that made a preliminary scan of the pattern in order to do this. That was
1513        OK when PCRE was new, but as the facilities have expanded, it has become
1514        harder and harder to keep it in step with the real compile phase, and there
1515        have been a number of bugs (see for example, 4 above). I have now found a
1516        cunning way of running the real compile function in a "fake" mode that
1517        enables it to compute how much memory it would need, while actually only
1518        ever using a few hundred bytes of working memory and without too many
1519        tests of the mode. This should make future maintenance and development
1520        easier. A side effect of this work is that the limit of 200 on the nesting
1521        depth of parentheses has been removed (though this was never a serious
1522        limitation, I suspect). However, there is a downside: pcre_compile() now
1523        runs more slowly than before (30% or more, depending on the pattern). I
1524        hope this isn't a big issue. There is no effect on runtime performance.
1525    
1526    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1527        newline (only possible for the last line of a file) and it was a
1528        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1529    
1530    19. Added additional timing features to pcretest. (1) The -tm option now times
1531        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1532        separate command line item, by a number that specifies the number of
1533        repeats to use when timing. The default is 50000; this gives better
1534        precision, but takes uncomfortably long for very large patterns.
1535    
1536    20. Extended pcre_study() to be more clever in cases where a branch of a
1537        subpattern has no definite first character. For example, (a*|b*)[cd] would
1538        previously give no result from pcre_study(). Now it recognizes that the
1539        first character must be a, b, c, or d.
1540    
1541    21. There was an incorrect error "recursive call could loop indefinitely" if
1542        a subpattern (or the entire pattern) that was being tested for matching an
1543        empty string contained only one non-empty item after a nested subpattern.
1544        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1545        incorrectly, because the \d was being skipped in the check.
1546    
1547    22. The pcretest program now has a new pattern option /B and a command line
1548        option -b, which is equivalent to adding /B to every pattern. This causes
1549        it to show the compiled bytecode, without the additional information that
1550        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1551        is the same as /B/I).
1552    
1553    23. A new optimization is now able automatically to treat some sequences such
1554        as a*b as a*+b. More specifically, if something simple (such as a character
1555        or a simple class like \d) has an unlimited quantifier, and is followed by
1556        something that cannot possibly match the quantified thing, the quantifier
1557        is automatically "possessified".
1558    
1559    24. A recursive reference to a subpattern whose number was greater than 39
1560        went wrong under certain circumstances in UTF-8 mode. This bug could also
1561        have affected the operation of pcre_study().
1562    
1563    25. Realized that a little bit of performance could be had by replacing
1564        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1565    
1566    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1567    
1568    27. Possessive quantifiers such as a++ were previously implemented by turning
1569        them into atomic groups such as ($>a+). Now they have their own opcodes,
1570        which improves performance. This includes the automatically created ones
1571        from 23 above.
1572    
1573    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1574        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1575        the first matched character to be a colon. This applied both to named and
1576        numbered groups.
1577    
1578    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1579    
1580    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1581        building PCRE on Windows easier, so I have included it in the distribution.
1582    
1583    31. There is now a check in pcretest against a ridiculously large number being
1584        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1585        loop, the loop is abandoned.
1586    
1587    32. Forward references to subpatterns in conditions such as (?(2)...) where
1588        subpattern 2 is defined later cause pcre_compile() to search forwards in
1589        the pattern for the relevant set of parentheses. This search went wrong
1590        when there were unescaped parentheses in a character class, parentheses
1591        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1592    
1593    33. "Subroutine" calls and backreferences were previously restricted to
1594        referencing subpatterns earlier in the regex. This restriction has now
1595        been removed.
1596    
1597    34. Added a number of extra features that are going to be in Perl 5.10. On the
1598        whole, these are just syntactic alternatives for features that PCRE had
1599        previously implemented using the Python syntax or my own invention. The
1600        other formats are all retained for compatibility.
1601    
1602        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1603            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1604            also .NET compatible.
1605    
1606        (b) A recursion or subroutine call to a named group can now be defined as
1607            (?&name) as well as (?P>name).
1608    
1609        (c) A backreference to a named group can now be defined as \k<name> or
1610            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1611            5.10, are also .NET compatible.
1612    
1613        (d) A conditional reference to a named group can now use the syntax
1614            (?(<name>) or (?('name') as well as (?(name).
1615    
1616        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1617            groups (named and numbered) that are never evaluated inline, but can be
1618            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1619            is always false. There may be only one alternative in such a group.
1620    
1621        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1622            as the simple (?(R). The condition is true only if the most recent
1623            recursion is that of the given number or name. It does not search out
1624            through the entire recursion stack.
1625    
1626        (g) The escape \gN or \g{N} has been added, where N is a positive or
1627            negative number, specifying an absolute or relative reference.
1628    
1629    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1630        some "unreachable code" warnings.
1631    
1632    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1633        things, this adds five new scripts.
1634    
1635    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1636        There were also incompatibilities regarding the handling of \Q..\E inside
1637        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1638        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1639    
1640    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1641        matches an empty string, and forcibly breaks the loop. There were bugs in
1642        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1643        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1644        separate and independent bugs (that affected different cases) have been
1645        fixed.
1646    
1647    39. Refactored the code to abolish the use of different opcodes for small
1648        capturing bracket numbers. This is a tidy that I avoided doing when I
1649        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1650        The new approach is not only tidier, it makes it possible to reduce the
1651        memory needed to fix the previous bug (38).
1652    
1653    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1654        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1655        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1656        mode.
1657    
1658    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1659        report.
1660    
1661    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1662        copy construction and assignment in the C++ wrapper.
1663    
1664    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1665        couple of bugs that could have given wrong results in the "--newline=crlf"
1666        case.
1667    
1668    44. Added a number of casts and did some reorganization of signed/unsigned int
1669        variables following suggestions from Dair Grant. Also renamed the variable
1670        "this" as "item" because it is a C++ keyword.
1671    
1672    45. Arranged for dftables to add
1673    
1674          #include "pcre_internal.h"
1675    
1676        to pcre_chartables.c because without it, gcc 4.x may remove the array
1677        definition from the final binary if PCRE is built into a static library and
1678        dead code stripping is activated.
1679    
1680    46. For an unanchored pattern, if a match attempt fails at the start of a
1681        newline sequence, and the newline setting is CRLF or ANY, and the next two
1682        characters are CRLF, advance by two characters instead of one.
1683    
1684    
1685    Version 6.7 04-Jul-06
1686    ---------------------
1687    
1688     1. In order to handle tests when input lines are enormously long, pcretest has
1689        been re-factored so that it automatically extends its buffers when
1690        necessary. The code is crude, but this _is_ just a test program. The
1691        default size has been increased from 32K to 50K.
1692    
1693     2. The code in pcre_study() was using the value of the re argument before
1694        testing it for NULL. (Of course, in any sensible call of the function, it
1695        won't be NULL.)
1696    
1697     3. The memmove() emulation function in pcre_internal.h, which is used on
1698        systems that lack both memmove() and bcopy() - that is, hardly ever -
1699        was missing a "static" storage class specifier.
1700    
1701     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1702        containing an extended class (one that cannot be represented by a bitmap
1703        because it contains high-valued characters or Unicode property items, e.g.
1704        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1705        pattern, but PCRE should not loop if you do not (it no longer does).
1706        [Detail: two cases were found: (a) a repeated subpattern containing an
1707        extended class; (b) a recursive reference to a subpattern that followed a
1708        previous extended class. It wasn't skipping over the extended class
1709        correctly when UTF-8 mode was not set.]
1710    
1711     5. A negated single-character class was not being recognized as fixed-length
1712        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1713        compile error "lookbehind assertion is not fixed length".
1714    
1715     6. The RunPerlTest auxiliary script was showing an unexpected difference
1716        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1717        write a Perl script that can interpret lines of an input file either as
1718        byte characters or as UTF-8, which is what "perltest" was being required to
1719        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1720        can't do is switch easily at run time between having the "use utf8;" pragma
1721        or not. In the end, I fudged it by using the RunPerlTest script to insert
1722        "use utf8;" explicitly for the UTF-8 tests.
1723    
1724     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1725        the end of the subject string, contrary to the documentation and to what
1726        Perl does. This was true of both matching functions. Now it matches only at
1727        the start of the subject and immediately after *internal* newlines.
1728    
1729     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1730        a pointer to an int instead of a pointer to an unsigned long int. This
1731        caused problems on 64-bit systems.
1732    
1733     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1734        instance of the 'standard' template library not being so standard".
1735    
1736    10. There was no check on the number of named subpatterns nor the maximum
1737        length of a subpattern name. The product of these values is used to compute
1738        the size of the memory block for a compiled pattern. By supplying a very
1739        long subpattern name and a large number of named subpatterns, the size
1740        computation could be caused to overflow. This is now prevented by limiting
1741        the length of names to 32 characters, and the number of named subpatterns
1742        to 10,000.
1743    
1744    11. Subpatterns that are repeated with specific counts have to be replicated in
1745        the compiled pattern. The size of memory for this was computed from the
1746        length of the subpattern and the repeat count. The latter is limited to
1747        65535, but there was no limit on the former, meaning that integer overflow
1748        could in principle occur. The compiled length of a repeated subpattern is
1749        now limited to 30,000 bytes in order to prevent this.
1750    
1751    12. Added the optional facility to have named substrings with the same name.
1752    
1753    13. Added the ability to use a named substring as a condition, using the
1754        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1755        are numbers (not recommended). Forward references are permitted.
1756    
1757    14. Added forward references in named backreferences (if you see what I mean).
1758    
1759    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1760        pattern could run off the end of the subject. For example, the pattern
1761        "(?s)(.{1,5})"8 did this with the subject "ab".
1762    
1763    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1764        PCRE_CASELESS was set when matching characters that were quantified with ?
1765        or *.
1766    
1767    17. A character class other than a single negated character that had a minimum
1768        but no maximum quantifier - for example [ab]{6,} - was not handled
1769        correctly by pce_dfa_exec(). It would match only one character.
1770    
1771    18. A valid (though odd) pattern that looked like a POSIX character
1772        class but used an invalid character after [ (for example [[,abc,]]) caused
1773        pcre_compile() to give the error "Failed: internal error: code overflow" or
1774        in some cases to crash with a glibc free() error. This could even happen if
1775        the pattern terminated after [[ but there just happened to be a sequence of
1776        letters, a binary zero, and a closing ] in the memory that followed.
1777    
1778    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1779        over the years. Originally (before any Unicode support), just the bottom 8
1780        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1781        output from "man perlunicode" includes this:
1782    
1783          The regular expression compiler produces polymorphic opcodes.  That
1784          is, the pattern adapts to the data and automatically switches to
1785          the Unicode character scheme when presented with Unicode data--or
1786          instead uses a traditional byte scheme when presented with byte
1787          data.
1788    
1789        Sadly, a wide octal escape does not cause a switch, and in a string with
1790        no other multibyte characters, these octal escapes are treated as before.
1791        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1792        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1793        Unicode string.
1794    
1795        I have not perpetrated such confusion in PCRE. Up till now, it took just
1796        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1797        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1798        translate to the appropriate multibyte character.
1799    
1800    29. Applied some refactoring to reduce the number of warnings from Microsoft
1801        and Borland compilers. This has included removing the fudge introduced
1802        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1803        a warning about an unused variable.
1804    
1805    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1806        characters since release 4.0, because Perl (from release 5.004) does not.
1807        [Or at least, is documented not to: some releases seem to be in conflict
1808        with the documentation.] However, when a pattern was studied with
1809        pcre_study() and all its branches started with \s, PCRE still included VT
1810        as a possible starting character. Of course, this did no harm; it just
1811        caused an unnecessary match attempt.
1812    
1813    22. Removed a now-redundant internal flag bit that recorded the fact that case
1814        dependency changed within the pattern. This was once needed for "required
1815        byte" processing, but is no longer used. This recovers a now-scarce options
1816        bit. Also moved the least significant internal flag bit to the most-
1817        significant bit of the word, which was not previously used (hangover from
1818        the days when it was an int rather than a uint) to free up another bit for
1819        the future.
1820    
1821    23. Added support for CRLF line endings as well as CR and LF. As well as the
1822        default being selectable at build time, it can now be changed at runtime
1823        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1824        specify that it is scanning data with non-default line endings.
1825    
1826    24. Changed the definition of CXXLINK to make it agree with the definition of
1827        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1828    
1829    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1830        recursions. This makes a big different to stack usage for some patterns.
1831    
1832    26. If a subpattern containing a named recursion or subroutine reference such
1833        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1834        the space required for the compiled pattern went wrong and gave too small a
1835        value. Depending on the environment, this could lead to "Failed: internal
1836        error: code overflow at offset 49" or "glibc detected double free or
1837        corruption" errors.
1838    
1839    27. Applied patches from Google (a) to support the new newline modes and (b) to
1840        advance over multibyte UTF-8 characters in GlobalReplace.
1841    
1842    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1843        difference for some implementation of PCRE in some Windows version.
1844    
1845    29. Added some extra testing facilities to pcretest:
1846    
1847        \q<number>   in a data line sets the "match limit" value
1848        \Q<number>   in a data line sets the "match recursion limt" value
1849        -S <number>  sets the stack size, where <number> is in megabytes
1850    
1851        The -S option isn't available for Windows.
1852    
1853    
1854    Version 6.6 06-Feb-06
1855    ---------------------
1856    
1857     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1858        in pcreposix.h. I have copied the definition from pcre.h.
1859    
1860     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1861        because pcre.h is no longer a built file.
1862    
1863     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1864        not normally included in the compiled code.
1865    
1866    
1867    Version 6.5 01-Feb-06
1868    ---------------------
1869    
1870     1. When using the partial match feature with pcre_dfa_exec(), it was not
1871        anchoring the second and subsequent partial matches at the new starting
1872        point. This could lead to incorrect results. For example, with the pattern
1873        /1234/, partially matching against "123" and then "a4" gave a match.
1874    
1875     2. Changes to pcregrep:
1876    
1877        (a) All non-match returns from pcre_exec() were being treated as failures
1878            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1879            error message is output. Some extra information is given for the
1880            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1881            probably the only errors that are likely to be caused by users (by
1882            specifying a regex that has nested indefinite repeats, for instance).
1883            If there are more than 20 of these errors, pcregrep is abandoned.
1884    
1885        (b) A binary zero was treated as data while matching, but terminated the
1886            output line if it was written out. This has been fixed: binary zeroes
1887            are now no different to any other data bytes.
1888    
1889        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1890            used to set a locale for matching. The --locale=xxxx long option has
1891            been added (no short equivalent) to specify a locale explicitly on the
1892            pcregrep command, overriding the environment variables.
1893    
1894        (d) When -B was used with -n, some line numbers in the output were one less
1895            than they should have been.
1896    
1897        (e) Added the -o (--only-matching) option.
1898    
1899        (f) If -A or -C was used with -c (count only), some lines of context were
1900            accidentally printed for the final match.
1901    
1902        (g) Added the -H (--with-filename) option.
1903    
1904        (h) The combination of options -rh failed to suppress file names for files
1905            that were found from directory arguments.
1906    
1907        (i) Added the -D (--devices) and -d (--directories) options.
1908    
1909        (j) Added the -F (--fixed-strings) option.
1910    
1911        (k) Allow "-" to be used as a file name for -f as well as for a data file.
1912    
1913        (l) Added the --colo(u)r option.
1914    
1915        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1916            is not present by default.
1917    
1918     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1919        items such as (?R) or (?1), when the recursion could match a number of
1920        alternatives. If it matched one of the alternatives, but subsequently,
1921        outside the recursion, there was a failure, the code tried to back up into
1922        the recursion. However, because of the way PCRE is implemented, this is not
1923        possible, and the result was an incorrect result from the match.
1924    
1925        In order to prevent this happening, the specification of recursion has
1926        been changed so that all such subpatterns are automatically treated as
1927        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1928    
1929     4. I had overlooked the fact that, in some locales, there are characters for
1930        which isalpha() is true but neither isupper() nor islower() are true. In
1931        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1932        and ordfeminine) are like this. This affected the treatment of \w and \W
1933        when they appeared in character classes, but not when they appeared outside
1934        a character class. The bit map for "word" characters is now created
1935        separately from the results of isalnum() instead of just taking it from the
1936        upper, lower, and digit maps. (Plus the underscore character, of course.)
1937    
1938     5. The above bug also affected the handling of POSIX character classes such as
1939        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1940        permanent tables. Instead, the bit maps for such a class were previously
1941        created as the appropriate unions of the upper, lower, and digit bitmaps.
1942        Now they are created by subtraction from the [[:word:]] class, which has
1943        its own bitmap.
1944    
1945     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1946        It is created by subtracting the vertical space characters (\x09, \x0a,
1947        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1948        subtraction was done in the overall bitmap for a character class, meaning
1949        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1950        be recognized. This bug has been fixed.
1951    
1952     7. Patches from the folks at Google:
1953    
1954          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1955          real life, but is still worth protecting against".
1956    
1957          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1958          regular expressions".
1959    
1960          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1961          have it.
1962    
1963          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1964          "configure" and the latter not, in order to fix a problem somebody had
1965          with compiling the Arg class on HP-UX.
1966    
1967          (e) Improve the error-handling of the C++ wrapper a little bit.
1968    
1969          (f) New tests for checking recursion limiting.
1970    
1971     8. The pcre_memmove() function, which is used only if the environment does not
1972        have a standard memmove() function (and is therefore rarely compiled),
1973        contained two bugs: (a) use of int instead of size_t, and (b) it was not
1974        returning a result (though PCRE never actually uses the result).
1975    
1976     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1977        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1978        returned instead of calling malloc() with an overflowing number that would
1979        most likely cause subsequent chaos.
1980    
1981    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1982    
1983    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1984        with this option is matched, the nmatch and pmatch options of regexec() are
1985        ignored.
1986    
1987    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1988        provided in case anyone wants to the the POSIX interface with UTF-8
1989        strings.
1990    
1991    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1992        C++ linking (needed for some HP-UX environments).
1993    
1994    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1995        (unused parameter) and in the pcre_printint() function (omitted "default"
1996        switch label when the default is to do nothing).
1997    
1998    15. Added some code to make it possible, when PCRE is compiled as a C++
1999        library, to replace subject pointers for pcre_exec() with a smart pointer
2000        class, thus making it possible to process discontinuous strings.
2001    
2002    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2003        much the same function. They were added by different people who were trying
2004        to make PCRE easy to compile on non-Unix systems. It has been suggested
2005        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2006        for compiling on Windows systems. I have therefore replaced it with
2007        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2008        defaults to "extern" for C or "extern C" for C++, which works fine on
2009        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2010        SCOPE with something explicit in config.h. In addition:
2011    
2012        (a) pcreposix.h still had just "extern" instead of either of these macros;
2013            I have replaced it with PCRE_DATA_SCOPE.
2014    
2015        (b) Functions such as _pcre_xclass(), which are internal to the library,
2016            but external in the C sense, all had PCRE_EXPORT in their definitions.
2017            This is apparently wrong for the Windows case, so I have removed it.
2018            (It makes no difference on Unix-like systems.)
2019    
2020    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2021        of recursive calls to match(). This is different to MATCH_LIMIT because
2022        that limits the total number of calls to match(), not all of which increase
2023        the depth of recursion. Limiting the recursion depth limits the amount of
2024        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2025        when PCRE is compiled, and changed at run time. A patch from Google adds
2026        this functionality to the C++ interface.
2027    
2028    18. Changes to the handling of Unicode character properties:
2029    
2030        (a) Updated the table to Unicode 4.1.0.
2031    
2032        (b) Recognize characters that are not in the table as "Cn" (undefined).
2033    
2034        (c) I revised the way the table is implemented to a much improved format
2035            which includes recognition of ranges. It now supports the ranges that
2036            are defined in UnicodeData.txt, and it also amalgamates other
2037            characters into ranges. This has reduced the number of entries in the
2038            table from around 16,000 to around 3,000, thus reducing its size
2039            considerably. I realized I did not need to use a tree structure after
2040            all - a binary chop search is just as efficient. Having reduced the
2041            number of entries, I extended their size from 6 bytes to 8 bytes to
2042            allow for more data.
2043    
2044        (d) Added support for Unicode script names via properties such as \p{Han}.
2045    
2046    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2047        matching that character.
2048    
2049    20. When matching a repeated Unicode property with a minimum greater than zero,
2050        (for example \pL{2,}), PCRE could look past the end of the subject if it
2051        reached it while seeking the minimum number of characters. This could
2052        happen only if some of the characters were more than one byte long, because
2053        there is a check for at least the minimum number of bytes.
2054    
2055    21. Refactored the implementation of \p and \P so as to be more general, to
2056        allow for more different types of property in future. This has changed the
2057        compiled form incompatibly. Anybody with saved compiled patterns that use
2058        \p or \P will have to recompile them.
2059    
2060    22. Added "Any" and "L&" to the supported property types.
2061    
2062    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2063        but give a compile time error if the value is greater than 0xff.
2064    
2065    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2066        accidentally not being installed or uninstalled.
2067    
2068    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2069        made were to insert the current release number. This seemed silly, because
2070        it made things harder for people building PCRE on systems that don't run
2071        "configure". I have turned pcre.h into a distributed file, no longer built
2072        by "configure", with the version identification directly included. There is
2073        no longer a pcre.h.in file.
2074    
2075        However, this change necessitated a change to the pcre-config script as
2076        well. It is built from pcre-config.in, and one of the substitutions was the
2077        release number. I have updated configure.ac so that ./configure now finds
2078        the release number by grepping pcre.h.
2079    
2080    26. Added the ability to run the tests under valgrind.
2081    
2082    
2083    Version 6.4 05-Sep-05
2084    ---------------------
2085    
2086     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2087        "--" to be printed when multiple files were scanned, even when none of the
2088        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2089        consider it to be a bug, and have restored the previous behaviour.
2090    
2091     2. A couple of code tidies to get rid of compiler warnings.
2092    
2093     3. The pcretest program used to cheat by referring to symbols in the library
2094        whose names begin with _pcre_. These are internal symbols that are not
2095        really supposed to be visible externally, and in some environments it is
2096        possible to suppress them. The cheating is now confined to including
2097        certain files from the library's source, which is a bit cleaner.
2098    
2099     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2100        file's purpose clearer.
2101    
2102     5. Reorganized pcre_ucp_findchar().
2103    
2104    
2105    Version 6.3 15-Aug-05
2106    ---------------------
2107    
2108     1. The file libpcre.pc.in did not have general read permission in the tarball.
2109    
2110     2. There were some problems when building without C++ support:
2111    
2112        (a) If C++ support was not built, "make install" and "make test" still
2113            tried to test it.
2114    
2115        (b) There were problems when the value of CXX was explicitly set. Some
2116            changes have been made to try to fix these, and ...
2117    
2118        (c) --disable-cpp can now be used to explicitly disable C++ support.
2119    
2120        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2121            backslash in a target when C++ was disabled. This confuses some
2122            versions of "make", apparently. Using an intermediate variable solves
2123            this. (Same for CPP_LOBJ.)
2124    
2125     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2126        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2127        necessary on certain architectures.
2128    
2129     4. Added a setting of -export-symbols-regex to the link command to remove
2130        those symbols that are exported in the C sense, but actually are local
2131        within the library, and not documented. Their names all begin with
2132        "_pcre_". This is not a perfect job, because (a) we have to except some
2133        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2134        available (and never for static libraries). I have made a note to try to
2135        find a way round (a) in the future.
2136    
2137    
2138    Version 6.2 01-Aug-05
2139    ---------------------
2140    
2141     1. There was no test for integer overflow of quantifier values. A construction
2142        such as {1111111111111111} would give undefined results. What is worse, if
2143        a minimum quantifier for a parenthesized subpattern overflowed and became
2144        negative, the calculation of the memory size went wrong. This could have
2145        led to memory overwriting.
2146    
2147     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2148    
2149     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2150        operating environments where this matters.
2151    
2152     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2153        PCRE options from within the C++ wrapper.
2154    
2155     5. Named capturing subpatterns were not being correctly counted when a pattern
2156        was compiled. This caused two problems: (a) If there were more than 100
2157        such subpatterns, the calculation of the memory needed for the whole
2158        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2159        back references of the form \12, where the number was greater than 9, were
2160        not recognized as back references, even though there were sufficient
2161        previous subpatterns.
2162    
2163     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2164        versions of gcc, e.g. 2.95.4.
2165    
2166    
2167    Version 6.1 21-Jun-05
2168    ---------------------
2169    
2170     1. There was one reference to the variable "posix" in pcretest.c that was not
2171        surrounded by "#if !defined NOPOSIX".
2172    
2173     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2174        the cross-check on the old pcre_info() function, for the benefit of the
2175        cut-down version of PCRE that is currently imported into Exim.
2176    
2177     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2178        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2179        patterns that start (?i) but I don't think that matters. The use of (?i) is
2180        just an example; this all applies to the other options as well.
2181    
2182     4. Since libtool seems to echo the compile commands it is issuing, the output
2183        from "make" can be reduced a bit by putting "@" in front of each libtool
2184        compile command.
2185    
2186     5. Patch from the folks at Google for configure.in to be a bit more thorough
2187        in checking for a suitable C++ installation before trying to compile the
2188        C++ stuff. This should fix a reported problem when a compiler was present,
2189        but no suitable headers.
2190    
2191     6. The man pages all had just "PCRE" as their title. I have changed them to
2192        be the relevant file name. I have also arranged that these names are
2193        retained in the file doc/pcre.txt, which is a concatenation in text format
2194        of all the man pages except the little individual ones for each function.
2195    
2196     7. The NON-UNIX-USE file had not been updated for the different set of source
2197        files that come with release 6. I also added a few comments about the C++
2198        wrapper.
2199    
2200    
2201    Version 6.0 07-Jun-05
2202    ---------------------
2203    
2204     1. Some minor internal re-organization to help with my DFA experiments.
2205    
2206     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2207        didn't matter for the library itself when fully configured, but did matter
2208        when compiling without UCP support, or within Exim, where the ucp files are
2209        not imported.
2210    
2211     3. Refactoring of the library code to split up the various functions into
2212        different source modules. The addition of the new DFA matching code (see
2213        below) to a single monolithic source would have made it really too
2214        unwieldy, quite apart from causing all the code to be include in a
2215        statically linked application, when only some functions are used. This is
2216        relevant even without the DFA addition now that patterns can be compiled in
2217        one application and matched in another.
2218    
2219        The downside of splitting up is that there have to be some external
2220        functions and data tables that are used internally in different modules of
2221        the library but which are not part of the API. These have all had their
2222        names changed to start with "_pcre_" so that they are unlikely to clash
2223        with other external names.
2224    
2225     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2226        a different (DFA) algorithm. Although it is slower than the original
2227        function, it does have some advantages for certain types of matching
2228        problem.
2229    
2230     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2231        including restarting after a partial match.
2232    
2233     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2234        defined when compiling for Windows was sent to me. I have put it into the
2235        code, though I have no means of testing or verifying it.
2236    
2237     7. Added the pcre_refcount() auxiliary function.
2238    
2239     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2240        match before or at the first newline in the subject string. In pcretest,
2241        the /f option on a pattern can be used to set this.
2242    
2243     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2244        would behave wrongly. This has been present in PCRE since release 4.0.
2245    
2246    10. A number of changes to the pcregrep command:
2247    
2248        (a) Refactored how -x works; insert ^(...)$ instead of setting
2249            PCRE_ANCHORED and checking the length, in preparation for adding
2250            something similar for -w.
2251    
2252        (b) Added the -w (match as a word) option.
2253    
2254        (c) Refactored the way lines are read and buffered so as to have more
2255            than one at a time available.
2256    
2257        (d) Implemented a pcregrep test script.
2258    
2259        (e) Added the -M (multiline match) option. This allows patterns to match
2260            over several lines of the subject. The buffering ensures that at least
2261            8K, or the rest of the document (whichever is the shorter) is available
2262            for matching (and similarly the previous 8K for lookbehind assertions).
2263    
2264        (f) Changed the --help output so that it now says
2265    
2266              -w, --word-regex(p)
2267    
2268            instead of two lines, one with "regex" and the other with "regexp"
2269            because that confused at least one person since the short forms are the
2270            same. (This required a bit of code, as the output is generated
2271            automatically from a table. It wasn't just a text change.)
2272    
2273        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2274            option but starts with a hyphen. Could be a pattern or a path name
2275            starting with a hyphen, for instance.
2276    
2277        (h) "-" can be given as a file name to represent stdin.
2278    
2279        (i) When file names are being printed, "(standard input)" is used for
2280            the standard input, for compatibility with GNU grep. Previously
2281            "<stdin>" was used.
2282    
2283        (j) The option --label=xxx can be used to supply a name to be used for
2284            stdin when file names are being printed. There is no short form.
2285    
2286        (k) Re-factored the options decoding logic because we are going to add
2287            two more options that take data. Such options can now be given in four
2288            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2289    
2290        (l) Added the -A, -B, and -C options for requesting that lines of context
2291            around matches be printed.
2292    
2293        (m) Added the -L option to print the names of files that do not contain
2294            any matching lines, that is, the complement of -l.
2295    
2296        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2297            continue to scan other files.
2298    
2299        (o) The -s option was incorrectly implemented. For compatibility with other
2300            greps, it now suppresses the error message for a non-existent or non-
2301            accessible file (but not the return code). There is a new option called
2302            -q that suppresses the output of matching lines, which was what -s was
2303            previously doing.
2304    
2305        (p) Added --include and --exclude options to specify files for inclusion
2306            and exclusion when recursing.
2307    
2308    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2309        Hopefully, it now does.
2310    
2311    12. Missing cast in pcre_study().
2312    
2313    13. Added an "uninstall" target to the makefile.
2314    
2315    14. Replaced "extern" in the function prototypes in Makefile.in with
2316        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2317        world, but is set differently for Windows.
2318    
2319    15. Added a second compiling function called pcre_compile2(). The only
2320        difference is that it has an extra argument, which is a pointer to an
2321        integer error code. When there is a compile-time failure, this is set
2322        non-zero, in addition to the error test pointer being set to point to an
2323        error message. The new argument may be NULL if no error number is required
2324        (but then you may as well call pcre_compile(), which is now just a
2325        wrapper). This facility is provided because some applications need a
2326        numeric error indication, but it has also enabled me to tidy up the way
2327        compile-time errors are handled in the POSIX wrapper.
2328    
2329    16. Added VPATH=.libs to the makefile; this should help when building with one
2330        prefix path and installing with another. (Or so I'm told by someone who
2331        knows more about this stuff than I do.)
2332    
2333    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2334        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2335        match everything, including newlines. This is not POSIX-compatible, but
2336        somebody wanted the feature. From pcretest it can be activated by using
2337        both the P and the s flags.
2338    
2339    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2340    
2341    19. libpcre.pc was being incorrectly installed as executable.
2342    
2343    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2344        it now also looks for '\r' so that it will work unmodified on Windows.
2345    
2346    21. Added Google's contributed C++ wrapper to the distribution.
2347    
2348    22. Added some untidy missing memory free() calls in pcretest, to keep
2349        Electric Fence happy when testing.
2350    
2351    
2352    
2353    Version 5.0 13-Sep-04
2354    ---------------------
2355    
2356     1. Internal change: literal characters are no longer packed up into items
2357        containing multiple characters in a single byte-string. Each character
2358        is now matched using a separate opcode. However, there may be more than one
2359        byte in the character in UTF-8 mode.
2360    
2361     2. The pcre_callout_block structure has two new fields: pattern_position and
2362        next_item_length. These contain the offset in the pattern to the next match
2363        item, and its length, respectively.
2364    
2365     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2366        insertion of callouts before each pattern item. Added the /C option to
2367        pcretest to make use of this.
2368    
2369     4. On the advice of a Windows user, the lines
2370    
2371          #if defined(_WIN32) || defined(WIN32)
2372          _setmode( _fileno( stdout ), 0x8000 );
2373          #endif  /* defined(_WIN32) || defined(WIN32) */
2374    
2375        have been added to the source of pcretest. This apparently does useful
2376        magic in relation to line terminators.
2377    
2378     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2379        for the benefit of those environments where the "b" makes a difference.
2380    
2381     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2382        to know about it. I have put a hack into configure.in that adds in code
2383        to set GCC=yes if CC=icc. This seems to end up at a point in the
2384        generated configure script that is early enough to affect the setting of
2385        compiler options, which is what is needed, but I have no means of testing
2386        whether it really works. (The user who reported this had patched the
2387        generated configure script, which of course I cannot do.)
2388    
2389        LATER: After change 22 below (new libtool files), the configure script
2390        seems to know about icc (and also ecc). Therefore, I have commented out
2391        this hack in configure.in.
2392    
2393     7. Added support for pkg-config (2 patches were sent in).
2394    
2395     8. Negated POSIX character classes that used a combination of internal tables
2396        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2397        [[:^ascii]]. Typically, they would match almost any characters. The other
2398        POSIX classes were not broken in this way.
2399    
2400     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2401        to find the match, as PCRE was deluded into thinking that the match had to
2402        start at the start point or following a newline. The same bug applied to
2403        patterns with negative forward assertions or any backward assertions
2404        preceding ".*" at the start, unless the pattern required a fixed first
2405        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2406    
2407    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2408        starting at the last subject character, bytes beyond the end of the subject
2409        string were read.
2410    
2411    11. Renamed the variable "class" as "classbits" to make life easier for C++
2412        users. (Previously there was a macro definition, but it apparently wasn't
2413        enough.)
2414    
2415    12. Added the new field "tables" to the extra data so that tables can be passed
2416        in at exec time, or the internal tables can be re-selected. This allows
2417        a compiled regex to be saved and re-used at a later time by a different
2418        program that might have everything at different addresses.
2419    
2420    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2421        -R library as well as a -L library.
2422    
2423    14. The debugging options of pcretest (-d on the command line or D on a
2424        pattern) showed incorrect output for anything following an extended class
2425        that contained multibyte characters and which was followed by a quantifier.
2426    
2427    15. Added optional support for general category Unicode character properties
2428        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2429        support. It adds about 90K to the size of the library. The meanings of the
2430        inbuilt class escapes such as \d and \s have NOT been changed.
2431    
2432    16. Updated pcredemo.c to include calls to free() to release the memory for the
2433        compiled pattern.
2434    
2435    17. The generated file chartables.c was being created in the source directory
2436        instead of in the building directory. This caused the build to fail if the
2437        source directory was different from the building directory, and was
2438        read-only.
2439    
2440    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2441        file. No doubt somebody will tell me if they don't make sense... Also added
2442        Dan Mooney's comments about building on OpenVMS.
2443    
2444    19. Added support for partial matching via the PCRE_PARTIAL option for
2445        pcre_exec() and the \P data escape in pcretest.
2446    
2447    20. Extended pcretest with 3 new pattern features:
2448    
2449        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2450              write the compiled pattern to the file whose name is "rest-of-line".
2451              This is a straight binary dump of the data, with the saved pointer to
2452              the character tables forced to be NULL. The study data, if any, is
2453              written too. After writing, pcretest reads a new pattern.
2454    
2455        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2456              compiled pattern from the given file. There must not be any
2457              occurrences of "<" in the file name (pretty unlikely); if there are,
2458              pcretest will instead treat the initial "<" as a pattern delimiter.
2459              After reading in the pattern, pcretest goes on to read data lines as
2460              usual.
2461    
2462        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2463              and 16-bit fields in a compiled pattern, to simulate a pattern that
2464              was compiled on a host of opposite endianness.
2465    
2466    21. The pcre-exec() function can now cope with patterns that were compiled on
2467        hosts of opposite endianness, with this restriction:
2468    
2469          As for any compiled expression that is saved and used later, the tables
2470          pointer field cannot be preserved; the extra_data field in the arguments
2471          to pcre_exec() should be used to pass in a tables address if a value
2472          other than the default internal tables were used at compile time.
2473    
2474    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2475        now diagnosed as an error. Previously, most of the time, a negative number
2476        would have been treated as zero, but if in addition "ovector" was passed as
2477        NULL, a crash could occur.
2478    
2479    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2480        new versions from the libtool 1.5 distribution (the last one is a copy of
2481        a file called libtool.m4). This seems to have fixed the need to patch
2482        "configure" to support Darwin 1.3 (which I used to do). However, I still
2483        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2484        workstation).
2485    
2486    24. Changed the PCRE licence to be the more standard "BSD" licence.
2487    
2488    
2489    Version 4.5 01-Dec-03
2490    ---------------------
2491    
2492     1. There has been some re-arrangement of the code for the match() function so
2493        that it can be compiled in a version that does not call itself recursively.
2494        Instead, it keeps those local variables that need separate instances for
2495        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2496        needs to "recurse". Keeping track of where control must go is done by means
2497        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2498        hide most of the details from the main code, and operates only if
2499        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2500        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2501        operating.
2502    
2503        To make it easier for callers to provide specially tailored get/free
2504        functions for this usage, two new functions, pcre_stack_malloc, and
2505        pcre_stack_free, are used. They are always called in strict stacking order,
2506        and the size of block requested is always the same.
2507    
2508        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2509        PCRE has been compiled to use the stack or the heap for recursion. The
2510        -C option of pcretest uses this to show which version is compiled.
2511    
2512        A new data escape \S, is added to pcretest; it causes the amounts of store
2513        obtained and freed by both kinds of malloc/free at match time to be added
2514        to the output.
2515    
2516     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2517        what's available on my current Linux desktop machine.
2518    
2519     3. When matching a UTF-8 string, the test for a valid string at the start has
2520        been extended. If start_offset is not zero, PCRE now checks that it points
2521        to a byte that is the start of a UTF-8 character. If not, it returns
2522        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2523        this is necessary because there may be backward assertions in the pattern.
2524        When matching the same subject several times, it may save resources to use
2525        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2526    
2527     4. The code for checking the validity of UTF-8 strings has been tightened so
2528        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2529        containing "overlong sequences".
2530    
2531     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2532        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2533        should have been "&", but it just so happened that all the cases this let
2534        through by mistake were picked up later in the function.
2535    
2536     6. I had used a variable called "isblank" - this is a C99 function, causing
2537        some compilers to warn. To avoid this, I renamed it (as "blankclass").
2538    
2539     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2540        prompting; (b) run "./pcretest /dev/null" at the start of the test script
2541        so the version is shown; (c) stop "make test" echoing "./RunTest".
2542    
2543     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2544    
2545     9. The prototype for memmove() for systems that don't have it was using
2546        size_t, but the inclusion of the header that defines size_t was later. I've
2547        moved the #includes for the C headers earlier to avoid this.
2548    
2549    10. Added some adjustments to the code to make it easier to compiler on certain
2550        special systems:
2551    
2552          (a) Some "const" qualifiers were missing.
2553          (b) Added the macro EXPORT before all exported functions; by default this
2554              is defined to be empty.
2555          (c) Changed the dftables auxiliary program (that builds chartables.c) so
2556              that it reads its output file name as an argument instead of writing
2557              to the standard output and assuming this can be redirected.
2558    
2559    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2560        class containing characters with values greater than 255, PCRE compilation
2561        went into a loop.
2562    
2563    12. A recursive reference to a subpattern that was within another subpattern
2564        that had a minimum quantifier of zero caused PCRE to crash. For example,
2565        (x(y(?2))z)? provoked this bug with a subject that got as far as the
2566        recursion. If the recursively-called subpattern itself had a zero repeat,
2567        that was OK.
2568    
2569    13. In pcretest, the buffer for reading a data line was set at 30K, but the
2570        buffer into which it was copied (for escape processing) was still set at
2571        1024, so long lines caused crashes.
2572    
2573    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2574        "internal error: code overflow...". This applied to any character class
2575        that was followed by a possessive quantifier.
2576    
2577    15. Modified the Makefile to add libpcre.la as a prerequisite for
2578        libpcreposix.la because I was told this is needed for a parallel build to
2579        work.
2580    
2581    16. If a pattern that contained .* following optional items at the start was
2582        studied, the wrong optimizing data was generated, leading to matching
2583        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2584        matching string must start with a or b or c. The correct conclusion for
2585        this pattern is that a match can start with any character.
2586    
2587    
2588    Version 4.4 13-Aug-03
2589    ---------------------
2590    
2591     1. In UTF-8 mode, a character class containing characters with values between
2592        127 and 255 was not handled correctly if the compiled pattern was studied.
2593        In fixing this, I have also improved the studying algorithm for such
2594        classes (slightly).
2595    
2596     2. Three internal functions had redundant arguments passed to them. Removal
2597        might give a very teeny performance improvement.
2598    
2599     3. Documentation bug: the value of the capture_top field in a callout is *one
2600        more than* the number of the hightest numbered captured substring.
2601    
2602     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2603        in incorrectly linking with a previously installed version. They now link
2604        explicitly with libpcre.la.
2605    
2606     5. configure.in no longer needs to recognize Cygwin specially.
2607    
2608     6. A problem in pcre.in for Windows platforms is fixed.
2609    
2610     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2611        pcretest, it used to include the size of the study block as part of its
2612        output. Unfortunately, the structure contains a field that has a different
2613        size on different hardware architectures. This meant that the tests that
2614        showed this size failed. As the block is currently always of a fixed size,
2615        this information isn't actually particularly useful in pcretest output, so
2616        I have just removed it.
2617    
2618     8. Three pre-processor statements accidentally did not start in column 1.
2619        Sadly, there are *still* compilers around that complain, even though
2620        standard C has not required this for well over a decade. Sigh.
2621    
2622     9. In pcretest, the code for checking callouts passed small integers in the
2623        callout_data field, which is a void * field. However, some picky compilers
2624        complained about the casts involved for this on 64-bit systems. Now
2625        pcretest passes the address of the small integer instead, which should get
2626        rid of the warnings.
2627    
2628    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2629        both compile and run time, and gives an error if an invalid UTF-8 sequence
2630        is found. There is a option for disabling this check in cases where the
2631        string is known to be correct and/or the maximum performance is wanted.
2632    
2633    11. In response to a bug report, I changed one line in Makefile.in from
2634    
2635            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2636        to
2637            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2638    
2639        to look similar to other lines, but I have no way of telling whether this
2640        is the right thing to do, as I do not use Windows. No doubt I'll get told
2641        if it's wrong...
2642    
2643    
2644    Version 4.3 21-May-03
2645    ---------------------
2646    
2647    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2648       Makefile.
2649    
2650    2. Some refactoring to improve the quality of the code:
2651    
2652       (i)   The utf8_table... variables are now declared "const".
2653    
2654       (ii)  The code for \cx, which used the "case flipping" table to upper case
2655             lower case letters, now just substracts 32. This is ASCII-specific,
2656             but the whole concept of \cx is ASCII-specific, so it seems
2657             reasonable.
2658    
2659       (iii) PCRE was using its character types table to recognize decimal and
2660             hexadecimal digits in the pattern. This is silly, because it handles
2661             only 0-9, a-f, and A-F, but the character types table is locale-
2662             specific, which means strange things might happen. A private
2663             table is now used for this - though it costs 256 bytes, a table is
2664             much faster than multiple explicit tests. Of course, the standard
2665             character types table is still used for matching digits in subject
2666             strings against \d.
2667    
2668       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2669             ending in _t are). So I've renamed it as ESC_tee.
2670    
2671    3. The first argument for regexec() in the POSIX wrapper should have been
2672       defined as "const".
2673    
2674    4. Changed pcretest to use malloc() for its buffers so that they can be
2675       Electric Fenced for debugging.
2676    
2677    5. There were several places in the code where, in UTF-8 mode, PCRE would try
2678       to read one or more bytes before the start of the subject string. Often this
2679       had no effect on PCRE's behaviour, but in some circumstances it could
2680       provoke a segmentation fault.
2681    
2682    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2683       to try to read one or more bytes before the start of the subject string.
2684    
2685    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2686       UTF-8 support could misbehave in various ways if the subject string
2687       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2688       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2689       back over UTF-8 characters.)
2690    
2691    
2692    Version 4.2 14-Apr-03
2693    ---------------------
2694    
2695    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2696    
2697    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2698         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2699         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2700         [WIN_PREFIX]: new variable, "cyg" for Cygwin
2701         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2702           and BUILD_EXEEXT
2703         Note: automatic setting of the BUILD variables is not yet working
2704         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2705           compile-time but not at link-time
2706         [LINK]: use for linking executables only
2707         make different versions for Windows and non-Windows
2708         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2709           libraries
2710         [LINK_FOR_BUILD]: new variable
2711         [OBJEXT]: use throughout
2712         [EXEEXT]: use throughout
2713         <winshared>: new target
2714         <wininstall>: new target
2715         <dftables.o>: use native compiler
2716         <dftables>: use native linker
2717         <install>: handle Windows platform correctly
2718         <clean>: ditto
2719         <check>: ditto
2720         copy DLL to top builddir before testing
2721    
2722       As part of these changes, -no-undefined was removed again. This was reported
2723       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2724       in any case.
2725    
2726    3. Some tidies to get rid of compiler warnings:
2727    
2728       . In the match_data structure, match_limit was an unsigned long int, whereas
2729         match_call_count was an int. I've made them both unsigned long ints.
2730    
2731       . In pcretest the fact that a const uschar * doesn't automatically cast to
2732         a void * provoked a warning.
2733    
2734       . Turning on some more compiler warnings threw up some "shadow" variables
2735         and a few more missing casts.
2736    
2737    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2738       option, a class that contained a single character with a value between 128
2739       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2740    
2741    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2742       option, a class that contained several characters, but with at least one
2743       whose value was between 128 and 255 caused PCRE to crash.
2744    
2745    
2746    Version 4.1 12-Mar-03
2747    ---------------------
2748    
2749    1. Compiling with gcc -pedantic found a couple of places where casts were
2750    needed, and a string in dftables.c that was longer than standard compilers are
2751    required to support.
2752    
2753    2. Compiling with Sun's compiler found a few more places where the code could
2754    be tidied up in order to avoid warnings.
2755    
2756    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2757    first of these names is deprecated in the latest Autoconf in favour of the name
2758    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2759    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2760    analogy I have changed it to CFLAGS_FOR_BUILD.
2761    
2762    4. Added -no-undefined to the linking command in the Makefile, because this is
2763    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2764    linking step for the pcreposix library.
2765    
2766    5. PCRE was failing to diagnose the case of two named groups with the same
2767    name.
2768    
2769    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2770    literal character that is needed in the subject for a match, and scans along to
2771    ensure that it is present before embarking on the full matching process. This
2772    saves time in cases of nested unlimited repeats that are never going to match.
2773    Problem: the scan can take a lot of time if the subject is very long (e.g.
2774    megabytes), thus penalizing straightforward matches. It is now done only if the
2775    amount of subject to be scanned is less than 1000 bytes.
2776    
2777    7. A lesser problem with the same optimization is that it was recording the
2778    first character of an anchored pattern as "needed", thus provoking a search
2779    right along the subject, even when the first match of the pattern was going to
2780    fail. The "needed" character is now not set for anchored patterns, unless it
2781    follows something in the pattern that is of non-fixed length. Thus, it still
2782    fulfils its original purpose of finding quick non-matches in cases of nested
2783    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2784    
2785    
2786    Version 4.0 17-Feb-03
2787    ---------------------
2788    
2789    1. If a comment in an extended regex that started immediately after a meta-item
2790    extended to the end of string, PCRE compiled incorrect data. This could lead to
2791    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2792    
2793    2. Moved to autoconf 2.53 and libtool 1.4.2.
2794    
2795    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2796    the special perltest8 script is no longer needed - all the tests can be run
2797    from a single perltest script.
2798    
2799    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2800    by \s. It has now been removed in PCRE. This means it isn't recognized as
2801    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2802    class [:space:] *does* include VT, thereby creating a mess.
2803    
2804    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2805    space and tab.
2806    
2807    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2808    its new features into the main test script, reducing the number of scripts.
2809    
2810    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2811    were backward compatible, and made the (?i) apply to the whole pattern, as if
2812    /i were given. Now it behaves more logically, and applies the option setting
2813    only to what follows. PCRE has been changed to follow suit. However, if it
2814    finds options settings right at the start of the pattern, it extracts them into
2815    the global options, as before. Thus, they show up in the info data.
2816    
2817    8. Added support for the \Q...\E escape sequence. Characters in between are
2818    treated as literals. This is slightly different from Perl in that $ and @ are
2819    also handled as literals inside the quotes. In Perl, they will cause variable
2820    interpolation. Note the following examples:
2821    
2822        Pattern            PCRE matches      Perl matches
2823    
2824        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
2825        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
2826        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
2827    
2828    For compatibility with Perl, \Q...\E sequences are recognized inside character
2829    classes as well as outside them.
2830    
2831    9. Re-organized 3 code statements in pcretest to avoid "overflow in
2832    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2833    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2834    signed/unsigned warnings.
2835    
2836    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2837    option for pcretest, so I've replaced it by a simple function that does just
2838    that job.
2839    
2840    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2841    "pcregrep -".
2842    
2843    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2844    Java package. This provides some syntactic sugar for simple cases of what my
2845    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2846    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2847    item, you can use this simplified notation. Note that only makes sense with
2848    greedy quantifiers. Consequently, the use of the possessive quantifier forces
2849    greediness, whatever the setting of the PCRE_UNGREEDY option.
2850    
2851    13. A change of greediness default within a pattern was not taking effect at
2852    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2853    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2854    was abstracted outside.
2855    
2856    14. PCRE now supports the \G assertion. It is true when the current matching
2857    position is at the start point of the match. This differs from \A when the
2858    starting offset is non-zero. Used with the /g option of pcretest (or similar
2859    code), it works in the same way as it does for Perl's /g option. If all
2860    alternatives of a regex begin with \G, the expression is anchored to the start
2861    match position, and the "anchored" flag is set in the compiled expression.
2862    
2863    15. Some bugs concerning the handling of certain option changes within patterns
2864    have been fixed. These applied to options other than (?ims). For example,
2865    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2866    been the other way round. Some of this was related to change 7 above.
2867    
2868    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2869    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2870    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2871    POSIX classes only within a class (e.g. /[[:alpha:]]/).
2872    
2873    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2874    mode. Unlike ".", it always matches newline, whatever the setting of
2875    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2876    assertions. Perl allows it, but it doesn't (in general) work because it can't
2877    calculate the length of the lookbehind. At least, that's the case for Perl
2878    5.8.0 - I've been told they are going to document that it doesn't work in
2879    future.
2880    
2881    18. Added an error diagnosis for escapes that PCRE does not support: these are
2882    \L, \l, \N, \P, \p, \U, \u, and \X.
2883    
2884    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2885    reading past the end of the pattern in cases such as /[abcd/.
2886    
2887    20. PCRE was getting more memory than necessary for patterns with classes that
2888    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2889    
2890    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2891    compiling PCRE for use with Virtual Pascal.
2892    
2893    22. Small fix to the Makefile to make it work properly if the build is done
2894    outside the source tree.
2895    
2896    23. Added a new extension: a condition to go with recursion. If a conditional
2897    subpattern starts with (?(R) the "true" branch is used if recursion has
2898    happened, whereas the "false" branch is used only at the top level.
2899    
2900    24. When there was a very long string of literal characters (over 255 bytes
2901    without UTF support, over 250 bytes with UTF support), the computation of how
2902    much memory was required could be incorrect, leading to segfaults or other
2903    strange effects.
2904    
2905    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2906    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2907    there was a subsequent back reference to those brackets. This meant that, for
2908    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2909    possible to check for precisely this case. All we can do is abandon the
2910    optimization if .* occurs inside capturing brackets when there are any back
2911    references whatsoever. (See below for a better fix that came later.)
2912    
2913    26. The handling of the optimization for finding the first character of a
2914    non-anchored pattern, and for finding a character that is required later in the
2915    match were failing in some cases. This didn't break the matching; it just
2916    failed to optimize when it could. The way this is done has been re-implemented.
2917    
2918    27. Fixed typo in error message for invalid (?R item (it said "(?p").
2919    
2920    28. Added a new feature that provides some of the functionality that Perl
2921    provides with (?{...}). The facility is termed a "callout". The way it is done
2922    in PCRE is for the caller to provide an optional function, by setting
2923    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2924    global variable. By default it is unset, which disables all calling out. To get
2925    the function called, the regex must include (?C) at appropriate points. This
2926    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2927    This provides a means of identifying different callout points. When PCRE
2928    reaches such a point in the regex, if pcre_callout has been set, the external
2929    function is called. It is provided with data in a structure called
2930    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2931    matching continues; if it returns a non-zero value, the match at the current
2932    point fails. However, backtracking will occur if possible. [This was changed
2933    later and other features added - see item 49 below.]
2934    
2935    29. pcretest is upgraded to test the callout functionality. It provides a
2936    callout function that displays information. By default, it shows the start of
2937    the match and the current position in the text. There are some new data escapes
2938    to vary what happens:
2939    
2940        \C+         in addition, show current contents of captured substrings
2941        \C-         do not supply a callout function
2942        \C!n        return 1 when callout number n is reached
2943        \C!n!m      return 1 when callout number n is reached for the mth time
2944    
2945    30. If pcregrep was called with the -l option and just a single file name, it
2946    output "<stdin>" if a match was found, instead of the file name.
2947    
2948    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2949    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2950    pcre_exec(). This saves a malloc/free per call. The default value of
2951    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2952    when configuring.
2953    
2954    32. The default maximum size of a compiled pattern is 64K. There have been a
2955    few cases of people hitting this limit. The code now uses macros to handle the
2956    storing of links as offsets within the compiled pattern. It defaults to 2-byte
2957    links, but this can be changed to 3 or 4 bytes by --with-link-size when
2958    configuring. Tests 2 and 5 work only with 2-byte links because they output
2959    debugging information about compiled patterns.
2960    
2961    33. Internal code re-arrangements:
2962    
2963    (a) Moved the debugging function for printing out a compiled regex into
2964        its own source file (printint.c) and used #include to pull it into
2965        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2966        separate copies.
2967    
2968    (b) Defined the list of op-code names for debugging as a macro in
2969        internal.h so that it is next to the definition of the opcodes.
2970    
2971    (c) Defined a table of op-code lengths for simpler skipping along compiled
2972        code. This is again a macro in internal.h so that it is next to the
2973        definition of the opcodes.
2974    
2975    34. Added support for recursive calls to individual subpatterns, along the
2976    lines of Robin Houston's patch (but implemented somewhat differently).
2977    
2978    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2979    allow it to read and process whole directories in Win32. This code was
2980    contributed by Lionel Fourquaux; it has not been tested by me.
2981    
2982    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2983    used to name a group. Names consist of alphanumerics and underscores, and must
2984    be unique. Back references use the syntax (?P=name) and recursive calls use
2985    (?P>name) which is a PCRE extension to the Python extension. Groups still have
2986    numbers. The function pcre_fullinfo() can be used after compilation to extract
2987    a name/number map. There are three relevant calls:
2988    
2989      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
2990      PCRE_INFO_NAMECOUNT            yields the number of entries
2991      PCRE_INFO_NAMETABLE            yields a pointer to the map.
2992    
2993    The map is a vector of fixed-size entries. The size of each entry depends on
2994    the length of the longest name used. The first two bytes of each entry are the
2995    group number, most significant byte first. There follows the corresponding
2996    name, zero terminated. The names are in alphabetical order.
2997    
2998    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2999    case instead of 255. Making it the same both with and without UTF-8 support
3000    means that the same test output works with both.
3001    
3002    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3003    calling malloc() with a zero argument.
3004    
3005    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3006    optimization. I've improved things by keeping a bitmap of backreferences with
3007    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3008    fact referenced, the optimization can be applied. It is unlikely that a
3009    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3010    the match to follow \n) will appear inside brackets with a number greater than
3011    31, but if it does, any back reference > 31 suppresses the optimization.
3012    
3013    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3014    of disabling numbered capturing parentheses. Any opening parenthesis that is
3015    not followed by ? behaves as if it were followed by ?: but named parentheses
3016    can still be used for capturing (and they will acquire numbers in the usual
3017    way).
3018    
3019    41. Redesigned the return codes from the match() function into yes/no/error so
3020    that errors can be passed back from deep inside the nested calls. A malloc
3021    failure while inside a recursive subpattern call now causes the
3022    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3023    
3024    42. It is now possible to set a limit on the number of times the match()
3025    function is called in a call to pcre_exec(). This facility makes it possible to
3026    limit the amount of recursion and backtracking, though not in a directly
3027    obvious way, because the match() function is used in a number of different
3028    circumstances. The count starts from zero for each position in the subject
3029    string (for non-anchored patterns). The default limit is, for compatibility, a
3030    large number, namely 10 000 000. You can change this in two ways:
3031    
3032    (a) When configuring PCRE before making, you can use --with-match-limit=n
3033        to set a default value for the compiled library.
3034    
3035    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3036        a different value is set. See 45 below.
3037    
3038    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3039    
3040    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3041    of things that can be changed at compile time. The first argument specifies
3042    what is wanted and the second points to where the information is to be placed.
3043    The current list of available information is:
3044    
3045      PCRE_CONFIG_UTF8
3046    
3047    The output is an integer that is set to one if UTF-8 support is available;
3048    otherwise it is set to zero.
3049    
3050      PCRE_CONFIG_NEWLINE
3051    
3052    The output is an integer that it set to the value of the code that is used for
3053    newline. It is either LF (10) or CR (13).
3054    
3055      PCRE_CONFIG_LINK_SIZE
3056    
3057    The output is an integer that contains the number of bytes used for internal
3058    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3059    
3060      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3061    
3062    The output is an integer that contains the threshold above which the POSIX
3063    interface uses malloc() for output vectors. See item 31 above.
3064    
3065      PCRE_CONFIG_MATCH_LIMIT
3066    
3067    The output is an unsigned integer that contains the default limit of the number
3068    of match() calls in a pcre_exec() execution. See 42 above.
3069    
3070    44. pcretest has been upgraded by the addition of the -C option. This causes it
3071    to extract all the available output from the new pcre_config() function, and to
3072    output it. The program then exits immediately.
3073    
3074    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3075    order to support additional features. One way would have been to define
3076    pcre_exec2() (for example) with extra arguments, but this would not have been
3077    extensible, and would also have required all calls to the original function to
3078    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3079    is used for passing in "extra" data from pcre_study().
3080    
3081    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3082    contains the following fields:
3083    
3084      flags         a bitmap indicating which of the following fields are set
3085      study_data    opaque data from pcre_study()
3086      match_limit   a way of specifying a limit on match() calls for a specific
3087                      call to pcre_exec()
3088      callout_data  data for callouts (see 49 below)
3089    
3090    The flag bits are also defined in pcre.h, and are
3091    
3092      PCRE_EXTRA_STUDY_DATA
3093      PCRE_EXTRA_MATCH_LIMIT
3094      PCRE_EXTRA_CALLOUT_DATA
3095    
3096    The pcre_study() function now returns one of these new pcre_extra blocks, with
3097    the actual study data pointed to by the study_data field, and the
3098    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3099    before. That is, this change is entirely upwards-compatible and requires no
3100    change to existing code.
3101    
3102    If you want to pass in additional data to pcre_exec(), you can either place it
3103    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3104    block.
3105    
3106    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3107    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3108    times with different match limits, until it finds the minimum value needed for
3109    pcre_exec() to complete. The value is then output. This can be instructive; for
3110    most simple matches the number is quite small, but for pathological cases it
3111    gets very large very quickly.
3112    
3113    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3114    returns the size of the data block pointed to by the study_data field in a
3115    pcre_extra block, that is, the value that was passed as the argument to
3116    pcre_malloc() when PCRE was getting memory in which to place the information
3117    created by pcre_study(). The fourth argument should point to a size_t variable.
3118    pcretest has been extended so that this information is shown after a successful
3119    pcre_study() call when information about the compiled regex is being displayed.
3120    
3121    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3122    because what follows is always an absolute path. (Later: it turns out that this
3123    is more than cosmetic for MinGW, because it doesn't like empty path
3124    components.)
3125    
3126    49. Some changes have been made to the callout feature (see 28 above):
3127    
3128    (i)  A callout function now has three choices for what it returns:
3129    
3130           0  =>  success, carry on matching
3131         > 0  =>  failure at this point, but backtrack if possible
3132         < 0  =>  serious error, return this value from pcre_exec()
3133    
3134         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3135         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3136         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3137         use by callout functions. It will never be used by PCRE itself.
3138    
3139    (ii) The pcre_extra structure (see 45 above) has a void * field called
3140         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3141         pcre_callout_block structure has a field of the same name. The contents of
3142         the field passed in the pcre_extra structure are passed to the callout
3143         function in the corresponding field in the callout block. This makes it
3144         easier to use the same callout-containing regex from multiple threads. For
3145         testing, the pcretest program has a new data escape
3146    
3147           \C*n        pass the number n (may be negative) as callout_data
3148    
3149         If the callout function in pcretest receives a non-zero value as
3150         callout_data, it returns that value.
3151    
3152    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3153    there were some redundant $(CFLAGS) in commands that are now specified as
3154    $(LINK), which already includes $(CFLAGS).
3155    
3156    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3157    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3158    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3159    one-byte characters throughout. Note that case-insensitive matching applies
3160    only to characters whose values are less than 256. PCRE doesn't support the
3161    notion of cases for higher-valued characters.
3162    
3163    (i)   A character class whose characters are all within 0-255 is handled as
3164          a bit map, and the map is inverted for negative classes. Previously, a
3165          character > 255 always failed to match such a class; however it should
3166          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3167    
3168    (ii)  A negated character class with a single character < 255 is coded as
3169          "not this character" (OP_NOT). This wasn't working properly when the test
3170          character was multibyte, either singly or repeated.
3171    
3172    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3173          mode, for example: \x{100}{2,3}.
3174    
3175    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3176          singly or repeated) now correctly test multibyte characters. However,
3177          PCRE doesn't recognize any characters with values greater than 255 as
3178          digits, spaces, or word characters. Such characters always match \D, \S,
3179          and \W, and never match \d, \s, or \w.
3180    
3181    (v)   Classes may now contain characters and character ranges with values
3182          greater than 255. For example: [ab\x{100}-\x{400}].
3183    
3184    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3185          PCRE in UTF-8 mode.
3186    
3187    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3188    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3189    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3190    value.)
3191    
3192    53. The single man page has become too large. I have therefore split it up into
3193    a number of separate man pages. These also give rise to individual HTML pages;
3194    these are now put in a separate directory, and there is an index.html page that
3195    lists them all. Some hyperlinking between the pages has been installed.
3196    
3197    54. Added convenience functions for handling named capturing parentheses.
3198    
3199    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3200    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3201    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3202    are faulted.
3203    
3204    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3205    calling configure. These values are used when compiling the dftables.c program
3206    which is run to generate the source of the default character tables. They
3207    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3208    you will need to set these values.
3209    
3210    57. Updated the building process for Windows DLL, as provided by Fred Cox.
3211    
3212    
3213    Version 3.9 02-Jan-02
3214    ---------------------
3215    
3216    1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3217    
3218    2. If --disable-static was given, the building process failed when trying to
3219    build pcretest and pcregrep. (For some reason it was using libtool to compile
3220    them, which is not right, as they aren't part of the library.)
3221    
3222    
3223    Version 3.8 18-Dec-01
3224    ---------------------
3225    
3226    1. The experimental UTF-8 code was completely screwed up. It was packing the
3227    bytes in the wrong order. How dumb can you get?
3228    
3229    
3230    Version 3.7 29-Oct-01
3231    ---------------------
3232    
3233    1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3234    This caused pcretest, when used on the test data, to segfault. Unfortunately,
3235    this didn't happen under Solaris 8, where I normally test things.
3236    
3237    2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3238    doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3239    isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3240    this fix an hour or so after the initial 3.7 release.)
3241    
3242    
3243    Version 3.6 23-Oct-01
3244    ---------------------
3245    
3246    1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3247    offsets passed as NULL with zero offset count.
3248    
3249    2. The config.guess and config.sub files had not been updated when I moved to
3250    the latest autoconf.
3251    
3252    
3253    Version 3.5 15-Aug-01
3254    ---------------------
3255    
3256    1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3257    had been forgotten.
3258    
3259    2. By using declared but undefined structures, we can avoid using "void"
3260    definitions in pcre.h while keeping the internal definitions of the structures
3261    private.
3262    
3263    3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3264    user point of view, this means that both static and shared libraries are built
3265    by default, but this can be individually controlled. More of the work of
3266    handling this static/shared cases is now inside libtool instead of PCRE's make
3267    file.
3268    
3269    4. The pcretest utility is now installed along with pcregrep because it is
3270    useful for users (to test regexs) and by doing this, it automatically gets
3271    relinked by libtool. The documentation has been turned into a man page, so
3272    there are now .1, .txt, and .html versions in /doc.
3273    
3274    5. Upgrades to pcregrep:
3275       (i)   Added long-form option names like gnu grep.
3276       (ii)  Added --help to list all options with an explanatory phrase.
3277       (iii) Added -r, --recursive to recurse into sub-directories.
3278       (iv)  Added -f, --file to read patterns from a file.
3279    
3280    6. pcre_exec() was referring to its "code" argument before testing that
3281    argument for NULL (and giving an error if it was NULL).
3282    
3283    7. Upgraded Makefile.in to allow for compiling in a different directory from
3284    the source directory.
3285    
3286    8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3287    options bits, the pointer it was passed was to an int instead of to an unsigned
3288    long int. This mattered only on 64-bit systems.
3289    
3290    9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3291    generated) instead of pcre.in, which it its source. Also made the same change
3292    in several of the .c files.
3293    
3294    10. A new release of gcc defines printf() as a macro, which broke pcretest
3295    because it had an ifdef in the middle of a string argument for printf(). Fixed
3296    by using separate calls to printf().
3297    
3298    11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3299    script, to force use of CR or LF instead of \n in the source. On non-Unix
3300    systems, the value can be set in config.h.
3301    
3302    12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3303    absolute limit. Changed the text of the error message to make this clear, and
3304    likewise updated the man page.
3305    
3306    13. The limit of 99 on the number of capturing subpatterns has been removed.
3307    The new limit is 65535, which I hope will not be a "real" limit.
3308    
3309    
3310    Version 3.4 22-Aug-00
3311    ---------------------
3312    
3313    1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3314    
3315    2. Diagnose condition (?(0) as an error instead of crashing on matching.
3316    
3317    
3318    Version 3.3 01-Aug-00
3319    ---------------------
3320    
3321    1. If an octal character was given, but the value was greater than \377, it
3322    was not getting masked to the least significant bits, as documented. This could
3323    lead to crashes in some systems.
3324    
3325    2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3326    the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3327    
3328    3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3329    These just pass their arguments on to (pcre_free)(), but they are provided
3330    because some uses of PCRE bind it to non-C systems that can call its functions,
3331    but cannot call free() or pcre_free() directly.
3332    
3333    4. Add "make test" as a synonym for "make check". Corrected some comments in
3334    the Makefile.
3335    
3336    5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3337    Makefile.
3338    
3339    6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3340    command called pgrep for grepping around the active processes.
3341    
3342    7. Added the beginnings of support for UTF-8 character strings.
3343    
3344    8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3345    RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3346    the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3347    out for the ar command.)
3348    
3349    
3350    Version 3.2 12-May-00
3351    ---------------------
3352    
3353    This is purely a bug fixing release.
3354    
3355    1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3356    of ZA. This was just one example of several cases that could provoke this bug,
3357    which was introduced by change 9 of version 2.00. The code for breaking
3358    infinite loops after an iteration that matches an empty string was't working
3359    correctly.
3360    
3361    2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3362    when matched against abbab (for example). After matching an empty string, it
3363    wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3364    caused it to match further down the string than it should.
3365    
3366    3. The code contained an inclusion of sys/types.h. It isn't clear why this
3367    was there because it doesn't seem to be needed, and it causes trouble on some
3368    systems, as it is not a Standard C header. It has been removed.
3369    
3370    4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3371    were reported on the Macintosh. The changes were from
3372    
3373      while ((c = *(++ptr)) != 0 && c != '\n');
3374    to
3375      while ((c = *(++ptr)) != 0 && c != '\n') ;
3376    
3377    Totally extraordinary, but if that's what it takes...
3378    
3379    5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3380    available. Added HAVE_BCOPY and an autoconf test for it; if neither
3381    HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3382    assumes the way PCRE uses memmove() (always moving upwards).
3383    
3384    6. PCRE is being used in one environment where strchr() is not available. There
3385    was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3386    faster code anyway.
3387    
3388    
3389    Version 3.1 09-Feb-00
3390    ---------------------
3391    
3392    The only change in this release is the fixing of some bugs in Makefile.in for
3393    the "install" target:
3394    
3395    (1) It was failing to install pcreposix.h.
3396    
3397    (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3398    
3399    
3400    Version 3.0 01-Feb-00
3401    ---------------------
3402    
3403    1. Add support for the /+ modifier to perltest (to output $` like it does in
3404    pcretest).
3405    
3406    2. Add support for the /g modifier to perltest.
3407    
3408    3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3409    matches null strings.
3410    
3411    4. Fix perltest so that it doesn't do unwanted things when fed an empty
3412    pattern. Perl treats empty patterns specially - it reuses the most recent
3413    pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3414    effect.
3415    
3416    5. The POSIX interface was broken in that it was just handing over the POSIX
3417    captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3418    required a bigger vector, with some working space on the end. This means that
3419    the POSIX wrapper now has to get and free some memory, and copy the results.
3420    
3421    6. Added some simple autoconf support, placing the test data and the
3422    documentation in separate directories, re-organizing some of the
3423    information files, and making it build pcre-config (a GNU standard). Also added
3424    libtool support for building PCRE as a shared library, which is now the
3425    default.
3426    
3427    7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3428    09 are not valid octal constants. Single digits will be used for minor values
3429    less than 10.
3430    
3431    8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3432    existing programs that set these in the POSIX interface can use PCRE without
3433    modification.
3434    
3435    9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3436    return all that pcre_info() returns, plus additional data. The pcre_info()
3437    function is retained for compatibility, but is considered to be obsolete.
3438    
3439    10. Added experimental recursion feature (?R) to handle one common case that
3440    Perl 5.6 will be able to do with (?p{...}).
3441    
3442    11. Added support for POSIX character classes like [:alpha:], which Perl is
3443    adopting.
3444    
3445    
3446    Version 2.08 31-Aug-99
3447    ----------------------
3448    
3449    1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3450    trying to match at the startoffset position, but instead was moving forward to
3451    the next newline as if a previous match had failed.
3452    
3453    2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3454    and could get into a loop if a null string was matched other than at the start
3455    of the subject.
3456    
3457    3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3458    be distinguished at compile time, and for completeness also added PCRE_DATE.
3459    
3460    5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3461    in GnuWin32 environments.
3462    
3463    
3464    Version 2.07 29-Jul-99
3465    ----------------------
3466    
3467    1. The documentation is now supplied in plain text form and HTML as well as in
3468    the form of man page sources.
3469    
3470    2. C++ compilers don't like assigning (void *) values to other pointer types.
3471    In particular this affects malloc(). Although there is no problem in Standard
3472    C, I've put in casts to keep C++ compilers happy.
3473    
3474    3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3475    should be (const char *).
3476    
3477    4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3478    be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3479    However, I haven't made this a standard facility. The documentation doesn't
3480    mention it, and the Makefile doesn't support it.
3481    
3482    5. The Makefile now contains an "install" target, with editable destinations at
3483    the top of the file. The pcretest program is not installed.
3484    
3485    6. pgrep -V now gives the PCRE version number and date.
3486    
3487    7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3488    causing the entire string to be ignored, instead of just the last character.
3489    
3490    8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3491    non-matching string, it can take a very, very long time, even for strings of
3492    quite modest length, because of the nested recursion. PCRE now does better in
3493    some of these cases. It does this by remembering the last required literal
3494    character in the pattern, and pre-searching the subject to ensure it is present
3495    before running the real match. In other words, it applies a heuristic to detect
3496    some types of certain failure quickly, and in the above example, if presented
3497    with a string that has no trailing " it gives "no match" very quickly.
3498    
3499    9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3500    other alternatives are tried instead.
3501    
3502    
3503    Version 2.06 09-Jun-99
3504    ----------------------
3505    
3506    1. Change pcretest's output for amount of store used to show just the code
3507    space, because the remainder (the data block) varies in size between 32-bit and
3508    64-bit systems.
3509    
3510    2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3511    start matching at. This allows lookbehinds to work when searching for multiple
3512    occurrences in a string.
3513    
3514    3. Added additional options to pcretest for testing multiple occurrences:
3515    
3516       /+   outputs the rest of the string that follows a match
3517       /g   loops for multiple occurrences, using the new startoffset argument
3518       /G   loops for multiple occurrences by passing an incremented pointer
3519    
3520    4. PCRE wasn't doing the "first character" optimization for patterns starting
3521    with \b or \B, though it was doing it for other lookbehind assertions. That is,
3522    it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3523    the letter 'x'. On long subject strings, this gives a significant speed-up.
3524    
3525    
3526    Version 2.05 21-Apr-99
3527    ----------------------
3528    
3529    1. Changed the type of magic_number from int to long int so that it works
3530    properly on 16-bit systems.
3531    
3532    2. Fixed a bug which caused patterns starting with .* not to work correctly
3533    when the subject string contained newline characters. PCRE was assuming
3534    anchoring for such patterns in all cases, which is not correct because .* will
3535    not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3536    DOTALL is set at top level; otherwise it knows that patterns starting with .*
3537    must be retried after every newline in the subject.