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

Diff of /code/trunk/ChangeLog

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

revision 13 by nigel, Sat Feb 24 21:38:21 2007 UTC revision 563 by ph10, Sun Oct 31 14:15:04 2010 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.11 10-Oct-2010
5    ------------------------
6    
7    1.  (*THEN) was not working properly if there were untried alternatives prior
8        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
9        backtracked to try for "b" instead of moving to the next alternative branch
10        at the same level (in this case, to look for "c"). The Perl documentation
11        is clear that when (*THEN) is backtracked onto, it goes to the "next
12        alternative in the innermost enclosing group".
13    
14    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
15        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
16        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
17        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
18        (*THEN).
19    
20    3.  If \s appeared in a character class, it removed the VT character from
21        the class, even if it had been included by some previous item, for example
22        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
23        of \s, but is part of the POSIX "space" class.)
24    
25    4.  A partial match never returns an empty string (because you can always
26        match an empty string at the end of the subject); however the checking for
27        an empty string was starting at the "start of match" point. This has been
28        changed to the "earliest inspected character" point, because the returned
29        data for a partial match starts at this character. This means that, for
30        example, /(?<=abc)def/ gives a partial match for the subject "abc"
31        (previously it gave "no match").
32    
33    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
34        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
35        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
36        has an implication that the given string is incomplete (because a partial
37        match is preferred over a full match). For this reason, these items now
38        give a partial match in this situation. [Aside: previously, the one case
39        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
40        match rather than a full match, which was wrong by the old rules, but is
41        now correct.]
42    
43    6.  There was a bug in the handling of #-introduced comments, recognized when
44        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
45        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
46        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
47        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
48        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
49        places in pcre_compile().
50    
51    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
52        comments when looking ahead for named forward references to subpatterns,
53        the only newline sequence it recognized was NL. It now handles newlines
54        according to the set newline convention.
55    
56    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
57        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
58        cater for a lack of strerror(). These oversights have been fixed.
59    
60    9.  Added --match-limit and --recursion-limit to pcregrep.
61    
62    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
63    
64    
65    Version 8.10 25-Jun-2010
66    ------------------------
67    
68    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
69        THEN.
70    
71    2.  (*ACCEPT) was not working when inside an atomic group.
72    
73    3.  Inside a character class, \B is treated as a literal by default, but
74        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
75        causes the error). The code is unchanged, but I tidied the documentation.
76    
77    4.  Inside a character class, PCRE always treated \R and \X as literals,
78        whereas Perl faults them if its -w option is set. I have changed PCRE so
79        that it faults them when PCRE_EXTRA is set.
80    
81    5.  Added support for \N, which always matches any character other than
82        newline. (It is the same as "." when PCRE_DOTALL is not set.)
83    
84    6.  When compiling pcregrep with newer versions of gcc which may have
85        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
86        declared with attribute warn_unused_result" were given. Just casting the
87        result to (void) does not stop the warnings; a more elaborate fudge is
88        needed. I've used a macro to implement this.
89    
90    7.  Minor change to pcretest.c to avoid a compiler warning.
91    
92    8.  Added four artifical Unicode properties to help with an option to make
93        \s etc use properties (see next item). The new properties are: Xan
94        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
95    
96    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
97        use Unicode properties. (*UCP) at the start of a pattern can be used to set
98        this option. Modified pcretest to add /W to test this facility. Added
99        REG_UCP to make it available via the POSIX interface.
100    
101    10. Added --line-buffered to pcregrep.
102    
103    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
104        studied, and the match started with a letter with a code point greater than
105        127 whose first byte was different to the first byte of the other case of
106        the letter, the other case of this starting letter was not recognized
107        (#976).
108    
109    12. If a pattern that was studied started with a repeated Unicode property
110        test, for example, \p{Nd}+, there was the theoretical possibility of
111        setting up an incorrect bitmap of starting bytes, but fortunately it could
112        not have actually happened in practice until change 8 above was made (it
113        added property types that matched character-matching opcodes).
114    
115    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
116        possible starting bytes for non-anchored patterns.
117    
118    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
119        \R, and also a number of cases that involve Unicode properties, both
120        explicit and implicit when PCRE_UCP is set.
121    
122    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
123        input, it could crash or give wrong results if characters with values
124        greater than 0xc0 were present in the subject string. (Detail: it assumed
125        UTF-8 input when processing these items.)
126    
127    16. Added a lot of (int) casts to avoid compiler warnings in systems where
128        size_t is 64-bit (#991).
129    
130    17. Added a check for running out of memory when PCRE is compiled with
131        --disable-stack-for-recursion (#990).
132    
133    18. If the last data line in a file for pcretest does not have a newline on
134        the end, a newline was missing in the output.
135    
136    19. The default pcre_chartables.c file recognizes only ASCII characters (values
137        less than 128) in its various bitmaps. However, there is a facility for
138        generating tables according to the current locale when PCRE is compiled. It
139        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
140        characters, are recognized by isspace() and therefore were getting set in
141        these tables, and indeed these tables seem to approximate to ISO 8859. This
142        caused a problem in UTF-8 mode when pcre_study() was used to create a list
143        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
144        which of course cannot start UTF-8 characters. I have changed the code so
145        that only real ASCII characters (less than 128) and the correct starting
146        bytes for UTF-8 encodings are set for characters greater than 127 when in
147        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
148        altogether.)
149    
150    20. Added the /T option to pcretest so as to be able to run tests with non-
151        standard character tables, thus making it possible to include the tests
152        used for 19 above in the standard set of tests.
153    
154    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
155        reference to a subpattern the other side of a comment that contains an
156        opening parenthesis caused either an internal compiling error, or a
157        reference to the wrong subpattern.
158    
159    
160    Version 8.02 19-Mar-2010
161    ------------------------
162    
163    1.  The Unicode data tables have been updated to Unicode 5.2.0.
164    
165    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
166        configured.
167    
168    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
169        original author of that file, following a query about its status.
170    
171    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
172        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
173    
174    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
175        quantifier applied to a forward-referencing subroutine call, could compile
176        incorrect code or give the error "internal error: previously-checked
177        referenced subpattern not found".
178    
179    6.  Both MS Visual Studio and Symbian OS have problems with initializing
180        variables to point to external functions. For these systems, therefore,
181        pcre_malloc etc. are now initialized to local functions that call the
182        relevant global functions.
183    
184    7.  There were two entries missing in the vectors called coptable and poptable
185        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
186        I've fixed the data, and added a kludgy way of testing at compile time that
187        the lengths are correct (equal to the number of opcodes).
188    
189    8.  Following on from 7, I added a similar kludge to check the length of the
190        eint vector in pcreposix.c.
191    
192    9.  Error texts for pcre_compile() are held as one long string to avoid too
193        much relocation at load time. To find a text, the string is searched,
194        counting zeros. There was no check for running off the end of the string,
195        which could happen if a new error number was added without updating the
196        string.
197    
198    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
199    
200    11. \K was not working if it appeared in an atomic group or in a group that
201        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
202        \K is "not well defined" if used in an assertion. PCRE now accepts it if
203        the assertion is positive, but not if it is negative.
204    
205    12. Change 11 fortuitously reduced the size of the stack frame used in the
206        "match()" function of pcre_exec.c by one pointer. Forthcoming
207        implementation of support for (*MARK) will need an extra pointer on the
208        stack; I have reserved it now, so that the stack frame size does not
209        decrease.
210    
211    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
212        item in branch that calls a recursion is a subroutine call - as in the
213        second branch in the above example - was incorrectly given the compile-
214        time error "recursive call could loop indefinitely" because pcre_compile()
215        was not correctly checking the subroutine for matching a non-empty string.
216    
217    14. The checks for overrunning compiling workspace could trigger after an
218        overrun had occurred. This is a "should never occur" error, but it can be
219        triggered by pathological patterns such as hundreds of nested parentheses.
220        The checks now trigger 100 bytes before the end of the workspace.
221    
222    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
223    
224    
225    Version 8.01 19-Jan-2010
226    ------------------------
227    
228    1.  If a pattern contained a conditional subpattern with only one branch (in
229        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
230        computed the wrong minimum data length (which is of course zero for such
231        subpatterns). This could cause incorrect "no match" results.
232    
233    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
234        the pattern is reset in the first branch, pcre_compile() failed with
235        "internal error: code overflow at offset...". This happened only when
236        the reset was to the original external option setting. (An optimization
237        abstracts leading options settings into an external setting, which was the
238        cause of this.)
239    
240    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
241        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
242        assertion pattern did not match (meaning that the assertion was true), it
243        was incorrectly treated as false if the SKIP had been reached during the
244        matching. This also applied to assertions used as conditions.
245    
246    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
247        assertion subpattern, including such a pattern used as a condition,
248        unpredictable results occurred, instead of the error return
249        PCRE_ERROR_DFA_UITEM.
250    
251    5.  The C++ GlobalReplace function was not working like Perl for the special
252        situation when an empty string is matched. It now does the fancy magic
253        stuff that is necessary.
254    
255    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
256        removed. (These were left over from very, very early versions of PCRE.)
257    
258    7.  Some cosmetic changes to the code to make life easier when compiling it
259        as part of something else:
260    
261        (a) Change DEBUG to PCRE_DEBUG.
262    
263        (b) In pcre_compile(), rename the member of the "branch_chain" structure
264            called "current" as "current_branch", to prevent a collision with the
265            Linux macro when compiled as a kernel module.
266    
267        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
268            prevent a collision with the Linux macro when compiled as a kernel
269            module.
270    
271    8.  In pcre_compile() there are some checks for integer overflows that used to
272        cast potentially large values to (double). This has been changed to that
273        when building, a check for int64_t is made, and if it is found, it is used
274        instead, thus avoiding the use of floating point arithmetic. (There is no
275        other use of FP in PCRE.) If int64_t is not found, the fallback is to
276        double.
277    
278    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
279        2005 (difference between two addresses compared to an unsigned value).
280    
281    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
282        custom one, because of the following reported problem in Windows:
283    
284          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
285              under Win32.
286          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
287              therefore missing the function definition.
288          - The compiler thus generates a "C" signature for the test function.
289          - The linker fails to find the "C" function.
290          - PCRE fails to configure if asked to do so against libbz2.
291    
292    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
293        messages were output:
294    
295          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
296          rerunning libtoolize, to keep the correct libtool macros in-tree.
297          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
298    
299        I have done both of these things.
300    
301    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
302        most of the time, it *can* run out if it is given a pattern that contains a
303        runaway infinite recursion. I updated the discussion in the pcrestack man
304        page.
305    
306    13. Now that we have gone to the x.xx style of version numbers, the minor
307        version may start with zero. Using 08 or 09 is a bad idea because users
308        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
309        interpreted as invalid octal numbers. I've updated the previous comment in
310        configure.ac, and also added a check that gives an error if 08 or 09 are
311        used.
312    
313    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
314        causing partial matching to fail when the end of the subject matched \W
315        in a UTF-8 pattern where \W was quantified with a minimum of 3.
316    
317    15. There were some discrepancies between the declarations in pcre_internal.h
318        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
319        their definitions. The declarations used "const uschar *" and the
320        definitions used USPTR. Even though USPTR is normally defined as "const
321        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
322        reported that: "This difference in casting confuses some C++ compilers, for
323        example, SunCC recognizes above declarations as different functions and
324        generates broken code for hbpcre." I have changed the declarations to use
325        USPTR.
326    
327    16. GNU libtool is named differently on some systems. The autogen.sh script now
328        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
329        (FreeBSD).
330    
331    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
332        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
333        comment: "Figure out how to create a longlong from a string: strtoll and
334        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
335        instance, but it only takes 2 args instead of 3!"
336    
337    18. A subtle bug concerned with back references has been fixed by a change of
338        specification, with a corresponding code fix. A pattern such as
339        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
340        refers, was giving matches when it shouldn't. For example, xa=xaaa would
341        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
342        same bug. Such groups have to be quantified to be useful, or contained
343        inside another quantified group. (If there's no repetition, the reference
344        can never match.) The problem arises because, having left the group and
345        moved on to the rest of the pattern, a later failure that backtracks into
346        the group uses the captured value from the final iteration of the group
347        rather than the correct earlier one. I have fixed this in PCRE by forcing
348        any group that contains a reference to itself to be an atomic group; that
349        is, there cannot be any backtracking into it once it has completed. This is
350        similar to recursive and subroutine calls.
351    
352    
353    Version 8.00 19-Oct-09
354    ----------------------
355    
356    1.  The table for translating pcre_compile() error codes into POSIX error codes
357        was out-of-date, and there was no check on the pcre_compile() error code
358        being within the table. This could lead to an OK return being given in
359        error.
360    
361    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
362        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
363        in a Windows environment.
364    
365    3.  The pcregrep --count option prints the count for each file even when it is
366        zero, as does GNU grep. However, pcregrep was also printing all files when
367        --files-with-matches was added. Now, when both options are given, it prints
368        counts only for those files that have at least one match. (GNU grep just
369        prints the file name in this circumstance, but including the count seems
370        more useful - otherwise, why use --count?) Also ensured that the
371        combination -clh just lists non-zero counts, with no names.
372    
373    4.  The long form of the pcregrep -F option was incorrectly implemented as
374        --fixed_strings instead of --fixed-strings. This is an incompatible change,
375        but it seems right to fix it, and I didn't think it was worth preserving
376        the old behaviour.
377    
378    5.  The command line items --regex=pattern and --regexp=pattern were not
379        recognized by pcregrep, which required --regex pattern or --regexp pattern
380        (with a space rather than an '='). The man page documented the '=' forms,
381        which are compatible with GNU grep; these now work.
382    
383    6.  No libpcreposix.pc file was created for pkg-config; there was just
384        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
385    
386    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
387        when UCP support is not needed, by modifying the Python script that
388        generates it from Unicode data files. This should not matter if the module
389        is correctly used as a library, but I received one complaint about 50K of
390        unwanted data. My guess is that the person linked everything into his
391        program rather than using a library. Anyway, it does no harm.
392    
393    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
394        was a minimum greater than 1 for a wide character in a possessive
395        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
396        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
397        character. Chaos in the form of incorrect output or a compiling loop could
398        result.
399    
400    9.  The restrictions on what a pattern can contain when partial matching is
401        requested for pcre_exec() have been removed. All patterns can now be
402        partially matched by this function. In addition, if there are at least two
403        slots in the offset vector, the offset of the earliest inspected character
404        for the match and the offset of the end of the subject are set in them when
405        PCRE_ERROR_PARTIAL is returned.
406    
407    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
408        synonymous with PCRE_PARTIAL, for backwards compatibility, and
409        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
410        and may be more useful for multi-segment matching.
411    
412    11. Partial matching with pcre_exec() is now more intuitive. A partial match
413        used to be given if ever the end of the subject was reached; now it is
414        given only if matching could not proceed because another character was
415        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
416        string "Z", which now yields "no match" instead of "partial match". In the
417        case of pcre_dfa_exec(), "no match" is given if every matching path for the
418        final character ended with (*FAIL).
419    
420    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
421        if the pattern had a "must contain" character that was already found in the
422        earlier partial match, unless partial matching was again requested. For
423        example, with the pattern /dog.(body)?/, the "must contain" character is
424        "g". If the first part-match was for the string "dog", restarting with
425        "sbody" failed. This bug has been fixed.
426    
427    13. The string returned by pcre_dfa_exec() after a partial match has been
428        changed so that it starts at the first inspected character rather than the
429        first character of the match. This makes a difference only if the pattern
430        starts with a lookbehind assertion or \b or \B (\K is not supported by
431        pcre_dfa_exec()). It's an incompatible change, but it makes the two
432        matching functions compatible, and I think it's the right thing to do.
433    
434    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
435        so that the demonstration program is easily available in environments where
436        PCRE has not been installed from source.
437    
438    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
439        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
440        library.
441    
442    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
443        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
444        is not the first non-POSIX option to be added. Clearly some people find
445        these options useful.
446    
447    17. If a caller to the POSIX matching function regexec() passes a non-zero
448        value for nmatch with a NULL value for pmatch, the value of
449        nmatch is forced to zero.
450    
451    18. RunGrepTest did not have a test for the availability of the -u option of
452        the diff command, as RunTest does. It now checks in the same way as
453        RunTest, and also checks for the -b option.
454    
455    19. If an odd number of negated classes containing just a single character
456        interposed, within parentheses, between a forward reference to a named
457        subpattern and the definition of the subpattern, compilation crashed with
458        an internal error, complaining that it could not find the referenced
459        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
460        [The bug was that it was starting one character too far in when skipping
461        over the character class, thus treating the ] as data rather than
462        terminating the class. This meant it could skip too much.]
463    
464    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
465        /g option in pcretest when the pattern contains \K, which makes it possible
466        to have an empty string match not at the start, even when the pattern is
467        anchored. Updated pcretest and pcredemo to use this option.
468    
469    21. If the maximum number of capturing subpatterns in a recursion was greater
470        than the maximum at the outer level, the higher number was returned, but
471        with unset values at the outer level. The correct (outer level) value is
472        now given.
473    
474    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
475        PCRE did not set those parentheses (unlike Perl). I have now found a way to
476        make it do so. The string so far is captured, making this feature
477        compatible with Perl.
478    
479    23. The tests have been re-organized, adding tests 11 and 12, to make it
480        possible to check the Perl 5.10 features against Perl 5.10.
481    
482    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
483        pattern matches a fixed length string. PCRE did not allow this; now it
484        does. Neither allows recursion.
485    
486    25. I finally figured out how to implement a request to provide the minimum
487        length of subject string that was needed in order to match a given pattern.
488        (It was back references and recursion that I had previously got hung up
489        on.) This code has now been added to pcre_study(); it finds a lower bound
490        to the length of subject needed. It is not necessarily the greatest lower
491        bound, but using it to avoid searching strings that are too short does give
492        some useful speed-ups. The value is available to calling programs via
493        pcre_fullinfo().
494    
495    26. While implementing 25, I discovered to my embarrassment that pcretest had
496        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
497        study optimizations had never been tested with that matching function.
498        Oops. What is worse, even when it was passed study data, there was a bug in
499        pcre_dfa_exec() that meant it never actually used it. Double oops. There
500        were also very few tests of studied patterns with pcre_dfa_exec().
501    
502    27. If (?| is used to create subpatterns with duplicate numbers, they are now
503        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
504        on the other side of the coin, they are no longer allowed to have different
505        names, because these cannot be distinguished in PCRE, and this has caused
506        confusion. (This is a difference from Perl.)
507    
508    28. When duplicate subpattern names are present (necessarily with different
509        numbers, as required by 27 above), and a test is made by name in a
510        conditional pattern, either for a subpattern having been matched, or for
511        recursion in such a pattern, all the associated numbered subpatterns are
512        tested, and the overall condition is true if the condition is true for any
513        one of them. This is the way Perl works, and is also more like the way
514        testing by number works.
515    
516    
517    Version 7.9 11-Apr-09
518    ---------------------
519    
520    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
521        (pcretest), all targets were linked against these libraries. This included
522        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
523        libraries. This caused unwanted dependencies to be created. This problem
524        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
525        pcretest is linked with readline.
526    
527    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
528        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
529        moved outside it again, because FALSE and TRUE are already defined in AIX,
530        but BOOL is not.
531    
532    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
533        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
534    
535    4.  The pcregrep documentation said spaces were inserted as well as colons (or
536        hyphens) following file names and line numbers when outputting matching
537        lines. This is not true; no spaces are inserted. I have also clarified the
538        wording for the --colour (or --color) option.
539    
540    5.  In pcregrep, when --colour was used with -o, the list of matching strings
541        was not coloured; this is different to GNU grep, so I have changed it to be
542        the same.
543    
544    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
545        each matching line was coloured. Now it goes on to look for further matches
546        of any of the test patterns, which is the same behaviour as GNU grep.
547    
548    7.  A pattern that could match an empty string could cause pcregrep to loop; it
549        doesn't make sense to accept an empty string match in pcregrep, so I have
550        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
551        seems to be how GNU grep behaves.
552    
553    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
554        start or after a newline", because the conditional assertion was not being
555        correctly handled. The rule now is that both the assertion and what follows
556        in the first alternative must satisfy the test.
557    
558    9.  If auto-callout was enabled in a pattern with a conditional group whose
559        condition was an assertion, PCRE could crash during matching, both with
560        pcre_exec() and pcre_dfa_exec().
561    
562    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
563        used for matching.
564    
565    11. Unicode property support in character classes was not working for
566        characters (bytes) greater than 127 when not in UTF-8 mode.
567    
568    12. Added the -M command line option to pcretest.
569    
570    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
571    
572    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
573    
574    16. Added comments and documentation about mis-use of no_arg in the C++
575        wrapper.
576    
577    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
578        from Martin Jerabek that uses macro names for all relevant character and
579        string constants.
580    
581    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
582        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
583        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
584        these, but not everybody uses configure.
585    
586    19. A conditional group that had only one branch was not being correctly
587        recognized as an item that could match an empty string. This meant that an
588        enclosing group might also not be so recognized, causing infinite looping
589        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
590        with the subject "ab", where knowledge that the repeated group can match
591        nothing is needed in order to break the loop.
592    
593    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
594        exec(), but without supplying a callout function, matching went wrong.
595    
596    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
597        leak if the size of the offset vector was greater than 30. When the vector
598        is smaller, the saved offsets during recursion go onto a local stack
599        vector, but for larger vectors malloc() is used. It was failing to free
600        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
601        error, in fact).
602    
603    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
604        heapframe that is used only when UTF-8 support is enabled. This caused no
605        problem, but was untidy.
606    
607    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
608        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
609        included within another project.
610    
611    24. Steven Van Ingelgem's patches to add more options to the CMake support,
612        slightly modified by me:
613    
614          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
615              not building pcregrep.
616    
617          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
618              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
619    
620    25. Forward references, both numeric and by name, in patterns that made use of
621        duplicate group numbers, could behave incorrectly or give incorrect errors,
622        because when scanning forward to find the reference group, PCRE was not
623        taking into account the duplicate group numbers. A pattern such as
624        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
625    
626    26. Changed a few more instances of "const unsigned char *" to USPTR, making
627        the feature of a custom pointer more persuasive (as requested by a user).
628    
629    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
630        pcretest.c, inside #ifndefs, because it seems they are sometimes already
631        pre-defined.
632    
633    28. Added support for (*UTF8) at the start of a pattern.
634    
635    29. Arrange for flags added by the "release type" setting in CMake to be shown
636        in the configuration summary.
637    
638    
639    Version 7.8 05-Sep-08
640    ---------------------
641    
642    1.  Replaced UCP searching code with optimized version as implemented for Ad
643        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
644        stage table and inline lookup instead of a function, giving speed ups of 2
645        to 5 times on some simple patterns that I tested. Permission was given to
646        distribute the MultiStage2.py script that generates the tables (it's not in
647        the tarball, but is in the Subversion repository).
648    
649    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
650        scripts.
651    
652    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
653        a group with a zero qualifier. The result of the study could be incorrect,
654        or the function might crash, depending on the pattern.
655    
656    4.  Caseless matching was not working for non-ASCII characters in back
657        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
658        It now works when Unicode Property Support is available.
659    
660    5.  In pcretest, an escape such as \x{de} in the data was always generating
661        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
662        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
663        truncation.
664    
665    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
666    
667    7.  Added two (int) casts to pcregrep when printing the difference of two
668        pointers, in case they are 64-bit values.
669    
670    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
671        test 2 if it fails.
672    
673    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
674        and a #define of that name to empty if it is not externally set. This is to
675        allow users of MSVC to set it if necessary.
676    
677    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
678        the convenience functions in the pcre_get.c source file.
679    
680    11. An option change at the start of a pattern that had top-level alternatives
681        could cause overwriting and/or a crash. This command provoked a crash in
682        some environments:
683    
684          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
685    
686        This potential security problem was recorded as CVE-2008-2371.
687    
688    12. For a pattern where the match had to start at the beginning or immediately
689        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
690        pcre_dfa_exec() could read past the end of the passed subject if there was
691        no match. To help with detecting such bugs (e.g. with valgrind), I modified
692        pcretest so that it places the subject at the end of its malloc-ed buffer.
693    
694    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
695        exec() might read past the end of the data buffer in UTF-8 mode.
696    
697    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
698        the data contained the byte 0x85 as part of a UTF-8 character within its
699        first line. This applied both to normal and DFA matching.
700    
701    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
702        /^[^d]*?$/8 failed to match "abc".
703    
704    16. Added a missing copyright notice to pcrecpp_internal.h.
705    
706    17. Make it more clear in the documentation that values returned from
707        pcre_exec() in ovector are byte offsets, not character counts.
708    
709    18. Tidied a few places to stop certain compilers from issuing warnings.
710    
711    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
712        supplied by Stefan Weber. I made a further small update for 7.8 because
713        there is a change of source arrangements: the pcre_searchfuncs.c module is
714        replaced by pcre_ucd.c.
715    
716    
717    Version 7.7 07-May-08
718    ---------------------
719    
720    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
721        a string to a long long, pretend we don't even have a long long." This is
722        done by checking for the strtoq, strtoll, and _strtoi64 functions.
723    
724    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
725        pre-7.6 versions, which defined a global no_arg variable instead of putting
726        it in the RE class. (See also #8 below.)
727    
728    3.  Remove a line of dead code, identified by coverity and reported by Nuno
729        Lopes.
730    
731    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
732    
733        (1) The include/exclude patterns were being applied to the whole pathnames
734            of files, instead of just to the final components.
735    
736        (2) If there was more than one level of directory, the subdirectories were
737            skipped unless they satisfied the include/exclude conditions. This is
738            inconsistent with GNU grep (and could even be seen as contrary to the
739            pcregrep specification - which I improved to make it absolutely clear).
740            The action now is always to scan all levels of directory, and just
741            apply the include/exclude patterns to regular files.
742    
743    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
744        --exclude_dir in the tests to avoid scanning .svn directories.
745    
746    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
747        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
748        doesn't support NULs in patterns.
749    
750    7.  Added some missing "const"s to declarations of static tables in
751        pcre_compile.c and pcre_dfa_exec.c.
752    
753    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
754        caused by fix #2  above. (Subsequently also a second patch to fix the
755        first patch. And a third patch - this was a messy problem.)
756    
757    9.  Applied Craig's patch to remove the use of push_back().
758    
759    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
760        matching function regexec().
761    
762    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
763        which, however, unlike Perl's \g{...}, are subroutine calls, not back
764        references. PCRE supports relative numbers with this syntax (I don't think
765        Oniguruma does).
766    
767    12. Previously, a group with a zero repeat such as (...){0} was completely
768        omitted from the compiled regex. However, this means that if the group
769        was called as a subroutine from elsewhere in the pattern, things went wrong
770        (an internal error was given). Such groups are now left in the compiled
771        pattern, with a new opcode that causes them to be skipped at execution
772        time.
773    
774    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
775        to the way PCRE behaves:
776    
777        (a) A lone ] character is dis-allowed (Perl treats it as data).
778    
779        (b) A back reference to an unmatched subpattern matches an empty string
780            (Perl fails the current match path).
781    
782        (c) A data ] in a character class must be notated as \] because if the
783            first data character in a class is ], it defines an empty class. (In
784            Perl it is not possible to have an empty class.) The empty class []
785            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
786            The negative empty class [^] matches any one character, independently
787            of the DOTALL setting.
788    
789    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
790        non-existent subpattern following a character class starting with ']' and
791        containing () gave an internal compiling error instead of "reference to
792        non-existent subpattern". Fortunately, when the pattern did exist, the
793        compiled code was correct. (When scanning forwards to check for the
794        existencd of the subpattern, it was treating the data ']' as terminating
795        the class, so got the count wrong. When actually compiling, the reference
796        was subsequently set up correctly.)
797    
798    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
799        it was being rejected as not supported by pcre_dfa_exec(), even though
800        other assertions are supported. I have made pcre_dfa_exec() support
801        (*FAIL).
802    
803    16. The implementation of 13c above involved the invention of a new opcode,
804        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
805        cannot be changed at match time, I realized I could make a small
806        improvement to matching performance by compiling OP_ALLANY instead of
807        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
808        on the OP_ANY path.
809    
810    17. Compiling pcretest on Windows with readline support failed without the
811        following two fixes: (1) Make the unistd.h include conditional on
812        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
813    
814    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
815        ncurses library to be included for pcretest when ReadLine support is
816        requested, but also to allow for it to be overridden. This patch came from
817        Daniel Bergström.
818    
819    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
820        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
821        any errors with the current Unicode tables. Thanks to Peter Kankowski for
822        spotting this.
823    
824    
825    Version 7.6 28-Jan-08
826    ---------------------
827    
828    1.  A character class containing a very large number of characters with
829        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
830        overflow.
831    
832    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
833        HAVE_LONG_LONG is not defined.
834    
835    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
836        bring them up to date and include new features. This patch includes:
837    
838        - Fixed PH's badly added libz and libbz2 support.
839        - Fixed a problem with static linking.
840        - Added pcredemo. [But later removed - see 7 below.]
841        - Fixed dftables problem and added an option.
842        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
843            HAVE_LONG_LONG.
844        - Added readline support for pcretest.
845        - Added an listing of the option settings after cmake has run.
846    
847    4.  A user submitted a patch to Makefile that makes it easy to create
848        "pcre.dll" under mingw when using Configure/Make. I added stuff to
849        Makefile.am that cause it to include this special target, without
850        affecting anything else. Note that the same mingw target plus all
851        the other distribution libraries and programs are now supported
852        when configuring with CMake (see 6 below) instead of with
853        Configure/Make.
854    
855    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
856        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
857        exported in the Windows port". It has not yet been confirmed that the patch
858        solves the problem, but it does no harm.
859    
860    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
861        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
862        with CMake, and also correct the comment about stack recursion.
863    
864    7.  Remove the automatic building of pcredemo from the ./configure system and
865        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
866        of a program that users should build themselves after PCRE is installed, so
867        building it automatically is not really right. What is more, it gave
868        trouble in some build environments.
869    
870    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
871    
872    
873    Version 7.5 10-Jan-08
874    ---------------------
875    
876    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
877        values in parens when parsing an RE using the C++ wrapper."
878    
879    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
880        Characters greater than 255 were excluded from the class instead of being
881        included.
882    
883    3.  The same bug as (2) above applied to negated POSIX classes such as
884        [:^space:].
885    
886    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
887        defined or documented. It seems to have been a typo for PCRE_STATIC, so
888        I have changed it.
889    
890    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
891        first named subpattern) and a construct such as (?&a) would reference the
892        first named subpattern whose name started with "a" (in other words, the
893        length check was missing). Both these problems are fixed. "Subpattern name
894        expected" is now given for (?&) (a zero-length name), and this patch also
895        makes it give the same error for \k'' (previously it complained that that
896        was a reference to a non-existent subpattern).
897    
898    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
899        this is right because (?- can be followed by option settings as well as by
900        digits. I have, however, made the messages clearer.
901    
902    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
903        than the number used in the conditional) now cause a compile-time error.
904        This is actually not compatible with Perl, which accepts such patterns, but
905        treats the conditional as always being FALSE (as PCRE used to), but it
906        seems to me that giving a diagnostic is better.
907    
908    8.  Change "alphameric" to the more common word "alphanumeric" in comments
909        and messages.
910    
911    9.  Fix two occurrences of "backslash" in comments that should have been
912        "backspace".
913    
914    10. Remove two redundant lines of code that can never be obeyed (their function
915        was moved elsewhere).
916    
917    11. The program that makes PCRE's Unicode character property table had a bug
918        which caused it to generate incorrect table entries for sequences of
919        characters that have the same character type, but are in different scripts.
920        It amalgamated them into a single range, with the script of the first of
921        them. In other words, some characters were in the wrong script. There were
922        thirteen such cases, affecting characters in the following ranges:
923    
924          U+002b0 - U+002c1
925          U+0060c - U+0060d
926          U+0061e - U+00612
927          U+0064b - U+0065e
928          U+0074d - U+0076d
929          U+01800 - U+01805
930          U+01d00 - U+01d77
931          U+01d9b - U+01dbf
932          U+0200b - U+0200f
933          U+030fc - U+030fe
934          U+03260 - U+0327f
935          U+0fb46 - U+0fbb1
936          U+10450 - U+1049d
937    
938    12. The -o option (show only the matching part of a line) for pcregrep was not
939        compatible with GNU grep in that, if there was more than one match in a
940        line, it showed only the first of them. It now behaves in the same way as
941        GNU grep.
942    
943    13. If the -o and -v options were combined for pcregrep, it printed a blank
944        line for every non-matching line. GNU grep prints nothing, and pcregrep now
945        does the same. The return code can be used to tell if there were any
946        non-matching lines.
947    
948    14. Added --file-offsets and --line-offsets to pcregrep.
949    
950    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
951        infinitely looping recursion. The bug was that positive lookaheads were not
952        being skipped when checking for a possible empty match (negative lookaheads
953        and both kinds of lookbehind were skipped).
954    
955    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
956        inclusion of <windows.h> to before rather than after the definition of
957        INVALID_FILE_ATTRIBUTES (patch from David Byron).
958    
959    17. Specifying a possessive quantifier with a specific limit for a Unicode
960        character property caused pcre_compile() to compile bad code, which led at
961        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
962        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
963        caused the error; without that there was no problem.
964    
965    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
966    
967    19. Added --enable-pcretest-libreadline.
968    
969    20. In pcrecpp.cc, the variable 'count' was incremented twice in
970        RE::GlobalReplace(). As a result, the number of replacements returned was
971        double what it should be. I removed one of the increments, but Craig sent a
972        later patch that removed the other one (the right fix) and added unit tests
973        that check the return values (which was not done before).
974    
975    21. Several CMake things:
976    
977        (1) Arranged that, when cmake is used on Unix, the libraries end up with
978            the names libpcre and libpcreposix, not just pcre and pcreposix.
979    
980        (2) The above change means that pcretest and pcregrep are now correctly
981            linked with the newly-built libraries, not previously installed ones.
982    
983        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
984    
985    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
986        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
987        UTF-8 newline character). The key issue is that the pattern starts .*;
988        this means that the match must be either at the beginning, or after a
989        newline. The bug was in the code for advancing after a failed match and
990        checking that the new position followed a newline. It was not taking
991        account of UTF-8 characters correctly.
992    
993    23. PCRE was behaving differently from Perl in the way it recognized POSIX
994        character classes. PCRE was not treating the sequence [:...:] as a
995        character class unless the ... were all letters. Perl, however, seems to
996        allow any characters between [: and :], though of course it rejects as
997        unknown any "names" that contain non-letters, because all the known class
998        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
999        for example, whereas PCRE did not - it did not recognize a POSIX character
1000        class. This seemed a bit dangerous, so the code has been changed to be
1001        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1002        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1003        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1004        Perl does, and where it didn't before.
1005    
1006    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1007        Windows environments %n is disabled by default.
1008    
1009    
1010    Version 7.4 21-Sep-07
1011    ---------------------
1012    
1013    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1014        means that a class such as [\s] counted as "explicit reference to CR or
1015        LF". That isn't really right - the whole point of the change was to try to
1016        help when there was an actual mention of one of the two characters. So now
1017        the change happens only if \r or \n (or a literal CR or LF) character is
1018        encountered.
1019    
1020    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1021        of both had grown to the point where there were only 3 bits left.
1022        Fortunately, there was spare space in the data structure, and so I have
1023        moved the internal flags into a new 16-bit field to free up more option
1024        bits.
1025    
1026    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1027        but did not set the internal JCHANGED flag - either of these is enough to
1028        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1029        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1030        start sets both bits.
1031    
1032    4.  Added options (at build time, compile time, exec time) to change \R from
1033        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1034    
1035    5.  doc/pcresyntax.html was missing from the distribution.
1036    
1037    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1038        compatibility, even though it is no longer used.
1039    
1040    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1041        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1042        windows.h file is present (where different names are used). [This was
1043        reversed later after testing - see 16 below.]
1044    
1045    8.  Changed all #include <config.h> to #include "config.h". There were also
1046        some further <pcre.h> cases that I changed to "pcre.h".
1047    
1048    9.  When pcregrep was used with the --colour option, it missed the line ending
1049        sequence off the lines that it output.
1050    
1051    10. It was pointed out to me that arrays of string pointers cause lots of
1052        relocations when a shared library is dynamically loaded. A technique of
1053        using a single long string with a table of offsets can drastically reduce
1054        these. I have refactored PCRE in four places to do this. The result is
1055        dramatic:
1056    
1057          Originally:                          290
1058          After changing UCP table:            187
1059          After changing error message table:   43
1060          After changing table of "verbs"       36
1061          After changing table of Posix names   22
1062    
1063        Thanks to the folks working on Gregex for glib for this insight.
1064    
1065    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1066        unicode-properties was also set.
1067    
1068    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1069    
1070    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1071        checked only for CRLF.
1072    
1073    14. Added casts to pcretest.c to avoid compiler warnings.
1074    
1075    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1076    
1077    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1078        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1079        entirely. This removes changes made in 7 above.
1080    
1081    17. The CMake files have been updated, and there is now more information about
1082        building with CMake in the NON-UNIX-USE document.
1083    
1084    
1085    Version 7.3 28-Aug-07
1086    ---------------------
1087    
1088     1. In the rejigging of the build system that eventually resulted in 7.1, the
1089        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1090        brackets there is not right, since it causes compilers to look for an
1091        installed pcre.h, not the version that is in the source that is being
1092        compiled (which of course may be different). I have changed it back to:
1093    
1094          #include "pcre.h"
1095    
1096        I have a vague recollection that the change was concerned with compiling in
1097        different directories, but in the new build system, that is taken care of
1098        by the VPATH setting the Makefile.
1099    
1100     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1101        when the subject happened to end in the byte 0x85 (e.g. if the last
1102        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1103        characters but of course it shouldn't be taken as a newline when it is part
1104        of another character. The bug was that, for an unlimited repeat of . in
1105        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1106        characters when looking for a newline.
1107    
1108     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1109    
1110     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1111        in debug output.
1112    
1113     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1114        long printing in the pcrecpp unittest when running under MinGW.
1115    
1116     6. ESC_K was left out of the EBCDIC table.
1117    
1118     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1119        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1120        limit also applies to "virtual nesting" when a pattern is recursive, and in
1121        this case 1000 isn't so big. I have been able to remove this limit at the
1122        expense of backing off one optimization in certain circumstances. Normally,
1123        when pcre_exec() would call its internal match() function recursively and
1124        immediately return the result unconditionally, it uses a "tail recursion"
1125        feature to save stack. However, when a subpattern that can match an empty
1126        string has an unlimited repetition quantifier, it no longer makes this
1127        optimization. That gives it a stack frame in which to save the data for
1128        checking that an empty string has been matched. Previously this was taken
1129        from the 1000-entry workspace that had been reserved. So now there is no
1130        explicit limit, but more stack is used.
1131    
1132     8. Applied Daniel's patches to solve problems with the import/export magic
1133        syntax that is required for Windows, and which was going wrong for the
1134        pcreposix and pcrecpp parts of the library. These were overlooked when this
1135        problem was solved for the main library.
1136    
1137     9. There were some crude static tests to avoid integer overflow when computing
1138        the size of patterns that contain repeated groups with explicit upper
1139        limits. As the maximum quantifier is 65535, the maximum group length was
1140        set at 30,000 so that the product of these two numbers did not overflow a
1141        32-bit integer. However, it turns out that people want to use groups that
1142        are longer than 30,000 bytes (though not repeat them that many times).
1143        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1144        made it possible to implement the integer overflow checks in a much more
1145        dynamic way, which I have now done. The artificial limitation on group
1146        length has been removed - we now have only the limit on the total length of
1147        the compiled pattern, which depends on the LINK_SIZE setting.
1148    
1149    10. Fixed a bug in the documentation for get/copy named substring when
1150        duplicate names are permitted. If none of the named substrings are set, the
1151        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1152        empty string.
1153    
1154    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1155        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1156        because the ] is interpreted as the first data character and the
1157        terminating ] is not found. PCRE has been made compatible with Perl in this
1158        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1159        cause memory overwriting.
1160    
1161    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1162        string has been matched (to stop an infinite loop). It was not recognizing
1163        a conditional subpattern that could match an empty string if that
1164        subpattern was within another subpattern. For example, it looped when
1165        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1166        condition was not nested. This bug has been fixed.
1167    
1168    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1169        past the start of the subject in the presence of bytes with the top bit
1170        set, for example "\x8aBCD".
1171    
1172    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1173        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1174    
1175    14. Optimized (?!) to (*FAIL).
1176    
1177    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1178        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1179        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1180        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1181        does: it's just the validity check that is more restrictive.
1182    
1183    16. Inserted checks for integer overflows during escape sequence (backslash)
1184        processing, and also fixed erroneous offset values for syntax errors during
1185        backslash processing.
1186    
1187    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1188        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1189    
1190    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1191        caused an overrun.
1192    
1193    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1194        something other than just ASCII characters) inside a group that had an
1195        unlimited repeat caused a loop at compile time (while checking to see
1196        whether the group could match an empty string).
1197    
1198    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1199        [\P{Any}] did so. (Error in the code for printing property names.)
1200    
1201    21. An orphan \E inside a character class could cause a crash.
1202    
1203    22. A repeated capturing bracket such as (A)? could cause a wild memory
1204        reference during compilation.
1205    
1206    23. There are several functions in pcre_compile() that scan along a compiled
1207        expression for various reasons (e.g. to see if it's fixed length for look
1208        behind). There were bugs in these functions when a repeated \p or \P was
1209        present in the pattern. These operators have additional parameters compared
1210        with \d, etc, and these were not being taken into account when moving along
1211        the compiled data. Specifically:
1212    
1213        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1214            length.
1215    
1216        (b) An item such as \pL+ within a repeated group could cause crashes or
1217            loops.
1218    
1219        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1220            "reference to non-existent subpattern" error.
1221    
1222        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1223    
1224    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1225        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1226    
1227    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1228    
1229    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1230        character were causing crashes (broken optimization).
1231    
1232    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1233        \p or \P) caused a compile-time loop.
1234    
1235    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1236        break. For example, the unstudied pattern [\r\n]A does not match the string
1237        "\r\nA" because change 7.0/46 below moves the current point on by two
1238        characters after failing to match at the start. However, the pattern \nA
1239        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1240        the same is true. There doesn't seem any very clean way out of this, but
1241        what I have chosen to do makes the common cases work: PCRE now takes note
1242        of whether there can be an explicit match for \r or \n anywhere in the
1243        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1244        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1245        pattern has explicit CR or LF references.
1246    
1247    29. Added (*CR) etc for changing newline setting at start of pattern.
1248    
1249    
1250    Version 7.2 19-Jun-07
1251    ---------------------
1252    
1253     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1254        which is apparently normally available under Windows.
1255    
1256     2. Re-jig the pcregrep tests with different newline settings in an attempt
1257        to make them independent of the local environment's newline setting.
1258    
1259     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1260    
1261     4. Some of the "internals" tests were previously cut out when the link size
1262        was not 2, because the output contained actual offsets. The recent new
1263        "Z" feature of pcretest means that these can be cut out, making the tests
1264        usable with all link sizes.
1265    
1266     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1267        stack recursion. This gives a massive performance boost under BSD, but just
1268        a small improvement under Linux. However, it saves one field in the frame
1269        in all cases.
1270    
1271     6. Added more features from the forthcoming Perl 5.10:
1272    
1273        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1274            recursion call. It refers to the nth most recently opened parentheses.
1275    
1276        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1277            to be opened parentheses.
1278    
1279        (c) Conditions that refer to capturing parentheses can be specified
1280            relatively, for example, (?(-2)... or (?(+3)...
1281    
1282        (d) \K resets the start of the current match so that everything before
1283            is not part of it.
1284    
1285        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1286    
1287        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1288            reference syntax.
1289    
1290        (g) (?| introduces a group in which the numbering of parentheses in each
1291            alternative starts with the same number.
1292    
1293        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1294    
1295     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1296        PCRE_INFO_JCHANGED.
1297    
1298     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1299        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1300        for detecting groups that can match an empty string.
1301    
1302     9. A pattern with a very large number of alternatives (more than several
1303        hundred) was running out of internal workspace during the pre-compile
1304        phase, where pcre_compile() figures out how much memory will be needed. A
1305        bit of new cunning has reduced the workspace needed for groups with
1306        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1307        workspace instead of running out of the 4096 that are available.
1308    
1309    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1310    
1311    11. Applied patch from Google to remove an optimization that didn't quite work.
1312        The report of the bug said:
1313    
1314          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1315          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1316          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1317    
1318    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1319        it matched the wrong number of bytes.
1320    
1321    
1322    Version 7.1 24-Apr-07
1323    ---------------------
1324    
1325     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1326        that is more "standard", making use of automake and other Autotools. There
1327        is some re-arrangement of the files and adjustment of comments consequent
1328        on this.
1329    
1330     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1331        for recursive directory scanning broke on some systems because the files
1332        are not scanned in any specific order and on different systems the order
1333        was different. A call to "sort" has been inserted into RunGrepTest for the
1334        approprate test as a short-term fix. In the longer term there may be an
1335        alternative.
1336    
1337     3. I had an email from Eric Raymond about problems translating some of PCRE's
1338        man pages to HTML (despite the fact that I distribute HTML pages, some
1339        people do their own conversions for various reasons). The problems
1340        concerned the use of low-level troff macros .br and .in. I have therefore
1341        removed all such uses from the man pages (some were redundant, some could
1342        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1343        HTML has been updated to handle .nf/.fi and to complain if it encounters
1344        .br or .in.
1345    
1346     4. Updated comments in configure.ac that get placed in config.h.in and also
1347        arranged for config.h to be included in the distribution, with the name
1348        config.h.generic, for the benefit of those who have to compile without
1349        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1350    
1351     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1352        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1353        makevp.bat for the current PCRE, using the additional files
1354        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1355    
1356     6. A Windows user reported a minor discrepancy with test 2, which turned out
1357        to be caused by a trailing space on an input line that had got lost in his
1358        copy. The trailing space was an accident, so I've just removed it.
1359    
1360     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1361        that is needed.
1362    
1363     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1364        as "const" (a) because they are and (b) because it helps the PHP
1365        maintainers who have recently made a script to detect big data structures
1366        in the php code that should be moved to the .rodata section. I remembered
1367        to update Builducptable as well, so it won't revert if ucptable.h is ever
1368        re-created.
1369    
1370     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1371        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1372        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1373        support is not required. This saves 1.5-2K of code, which is important in
1374        some applications.
1375    
1376        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1377        so as not to refer to the tables, even though these functions will never be
1378        called when UTF-8 support is disabled. Otherwise there are problems with a
1379        shared library.
1380    
1381    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1382    
1383        (a) It was defining its arguments as char * instead of void *.
1384    
1385        (b) It was assuming that all moves were upwards in memory; this was true
1386            a long time ago when I wrote it, but is no longer the case.
1387    
1388        The emulated memove() is provided for those environments that have neither
1389        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1390        is clearly not the case, as these two bugs were recently reported.
1391    
1392    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1393        and Detrail to create the HTML documentation, the .txt form of the man
1394        pages, and it removes trailing spaces from listed files. It also creates
1395        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1396        case, it wraps all the #defines with #ifndefs. This script should be run
1397        before "make dist".
1398    
1399    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1400        with Unicode property support.
1401    
1402        (a) For a maximizing quantifier, if the two different cases of the
1403            character were of different lengths in their UTF-8 codings (there are
1404            some cases like this - I found 11), and the matching function had to
1405            back up over a mixture of the two cases, it incorrectly assumed they
1406            were both the same length.
1407    
1408        (b) When PCRE was configured to use the heap rather than the stack for
1409            recursion during matching, it was not correctly preserving the data for
1410            the other case of a UTF-8 character when checking ahead for a match
1411            while processing a minimizing repeat. If the check also involved
1412            matching a wide character, but failed, corruption could cause an
1413            erroneous result when trying to check for a repeat of the original
1414            character.
1415    
1416    13. Some tidying changes to the testing mechanism:
1417    
1418        (a) The RunTest script now detects the internal link size and whether there
1419            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1420            values substituted by "configure". (The RunGrepTest script already did
1421            this for UTF-8.) The configure.ac script no longer substitutes the
1422            relevant variables.
1423    
1424        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1425            with length and offset values. This means that the output is different
1426            for different internal link sizes. Test 2 is skipped for link sizes
1427            other than 2 because of this, bypassing the problem. Unfortunately,
1428            there was also a test in test 3 (the locale tests) that used /B and
1429            failed for link sizes other than 2. Rather than cut the whole test out,
1430            I have added a new /Z option to pcretest that replaces the length and
1431            offset values with spaces. This is now used to make test 3 independent
1432            of link size. (Test 2 will be tidied up later.)
1433    
1434    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1435        segmentation fault instead of returning the appropriate error message.
1436    
1437    15. In multiline mode when the newline sequence was set to "any", the pattern
1438        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1439        This doesn't seem right; it now treats the CRLF combination as the line
1440        ending, and so does not match in that case. It's only a pattern such as ^$
1441        that would hit this one: something like ^ABC$ would have failed after \r
1442        and then tried again after \r\n.
1443    
1444    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1445        in an attempt to make files that differ only in their line terminators
1446        compare equal. This works on Linux.
1447    
1448    17. Under certain error circumstances pcregrep might try to free random memory
1449        as it exited. This is now fixed, thanks to valgrind.
1450    
1451    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1452        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1453        was because its rules for how to advance for /g after matching an empty
1454        string at the end of a line did not allow for this case. They now check for
1455        it specially.
1456    
1457    20. pcretest is supposed to handle patterns and data of any length, by
1458        extending its buffers when necessary. It was getting this wrong when the
1459        buffer for a data line had to be extended.
1460    
1461    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1462        CRLF as a newline sequence.
1463    
1464    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1465        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1466        I have nevertheless tidied it up.
1467    
1468    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1469    
1470    24. Added a man page for pcre-config.
1471    
1472    
1473    Version 7.0 19-Dec-06
1474    ---------------------
1475    
1476     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1477        moving to gcc 4.1.1.
1478    
1479     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1480        sys/time.h, which is documented as needed for this function. It doesn't
1481        seem to matter on Linux, but it showed up on some releases of OS X.
1482    
1483     3. It seems that there are systems where bytes whose values are greater than
1484        127 match isprint() in the "C" locale. The "C" locale should be the
1485        default when a C program starts up. In most systems, only ASCII printing
1486        characters match isprint(). This difference caused the output from pcretest
1487        to vary, making some of the tests fail. I have changed pcretest so that:
1488    
1489        (a) When it is outputting text in the compiled version of a pattern, bytes
1490            other than 32-126 are always shown as hex escapes.
1491    
1492        (b) When it is outputting text that is a matched part of a subject string,
1493            it does the same, unless a different locale has been set for the match
1494            (using the /L modifier). In this case, it uses isprint() to decide.
1495    
1496     4. Fixed a major bug that caused incorrect computation of the amount of memory
1497        required for a compiled pattern when options that changed within the
1498        pattern affected the logic of the preliminary scan that determines the
1499        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1500        that the computed length was too small. The symptoms of this bug were
1501        either the PCRE error "internal error: code overflow" from pcre_compile(),
1502        or a glibc crash with a message such as "pcretest: free(): invalid next
1503        size (fast)". Examples of patterns that provoked this bug (shown in
1504        pcretest format) are:
1505    
1506          /(?-x: )/x
1507          /(?x)(?-x: \s*#\s*)/
1508          /((?i)[\x{c0}])/8
1509          /(?i:[\x{c0}])/8
1510    
1511        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1512        is now done differently.
1513    
1514     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1515        wrapper classes; (b) implement a new function in the C++ scanner that is
1516        more efficient than the old way of doing things because it avoids levels of
1517        recursion in the regex matching; (c) add a paragraph to the documentation
1518        for the FullMatch() function.
1519    
1520     6. The escape sequence \n was being treated as whatever was defined as
1521        "newline". Not only was this contrary to the documentation, which states
1522        that \n is character 10 (hex 0A), but it also went horribly wrong when
1523        "newline" was defined as CRLF. This has been fixed.
1524    
1525     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1526        was being set to -1 for the "end of line" case (supposedly a value that no
1527        character can have). Though this value is never used (the check for end of
1528        line is "zero bytes in current character"), it caused compiler complaints.
1529        I've changed it to 0xffffffff.
1530    
1531     8. In pcre_version.c, the version string was being built by a sequence of
1532        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1533        string (as it is for production releases) called a macro with an empty
1534        argument. The C standard says the result of this is undefined. The gcc
1535        compiler treats it as an empty string (which was what was wanted) but it is
1536        reported that Visual C gives an error. The source has been hacked around to
1537        avoid this problem.
1538    
1539     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1540        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1541        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1542        of them did).
1543    
1544    10. Originally, pcretest opened its input and output without "b"; then I was
1545        told that "b" was needed in some environments, so it was added for release
1546        5.0 to both the input and output. (It makes no difference on Unix-like
1547        systems.) Later I was told that it is wrong for the input on Windows. I've
1548        now abstracted the modes into two macros, to make it easier to fiddle with
1549        them, and removed "b" from the input mode under Windows.
1550    
1551    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1552    
1553    12. Added -help and --help to pcretest as an official way of being reminded
1554        of the options.
1555    
1556    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1557        and pcrecpp.cc because they annoy compilers at high warning levels.
1558    
1559    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1560    
1561    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1562        scripts are not C programs :-) and which was not noticed because it works
1563        on Linux.
1564    
1565    16. pcretest is supposed to handle any length of pattern and data line (as one
1566        line or as a continued sequence of lines) by extending its input buffer if
1567        necessary. This feature was broken for very long pattern lines, leading to
1568        a string of junk being passed to pcre_compile() if the pattern was longer
1569        than about 50K.
1570    
1571    17. I have done a major re-factoring of the way pcre_compile() computes the
1572        amount of memory needed for a compiled pattern. Previously, there was code
1573        that made a preliminary scan of the pattern in order to do this. That was
1574        OK when PCRE was new, but as the facilities have expanded, it has become
1575        harder and harder to keep it in step with the real compile phase, and there
1576        have been a number of bugs (see for example, 4 above). I have now found a
1577        cunning way of running the real compile function in a "fake" mode that
1578        enables it to compute how much memory it would need, while actually only
1579        ever using a few hundred bytes of working memory and without too many
1580        tests of the mode. This should make future maintenance and development
1581        easier. A side effect of this work is that the limit of 200 on the nesting
1582        depth of parentheses has been removed (though this was never a serious
1583        limitation, I suspect). However, there is a downside: pcre_compile() now
1584        runs more slowly than before (30% or more, depending on the pattern). I
1585        hope this isn't a big issue. There is no effect on runtime performance.
1586    
1587    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1588        newline (only possible for the last line of a file) and it was a
1589        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1590    
1591    19. Added additional timing features to pcretest. (1) The -tm option now times
1592        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1593        separate command line item, by a number that specifies the number of
1594        repeats to use when timing. The default is 50000; this gives better
1595        precision, but takes uncomfortably long for very large patterns.
1596    
1597    20. Extended pcre_study() to be more clever in cases where a branch of a
1598        subpattern has no definite first character. For example, (a*|b*)[cd] would
1599        previously give no result from pcre_study(). Now it recognizes that the
1600        first character must be a, b, c, or d.
1601    
1602    21. There was an incorrect error "recursive call could loop indefinitely" if
1603        a subpattern (or the entire pattern) that was being tested for matching an
1604        empty string contained only one non-empty item after a nested subpattern.
1605        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1606        incorrectly, because the \d was being skipped in the check.
1607    
1608    22. The pcretest program now has a new pattern option /B and a command line
1609        option -b, which is equivalent to adding /B to every pattern. This causes
1610        it to show the compiled bytecode, without the additional information that
1611        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1612        is the same as /B/I).
1613    
1614    23. A new optimization is now able automatically to treat some sequences such
1615        as a*b as a*+b. More specifically, if something simple (such as a character
1616        or a simple class like \d) has an unlimited quantifier, and is followed by
1617        something that cannot possibly match the quantified thing, the quantifier
1618        is automatically "possessified".
1619    
1620    24. A recursive reference to a subpattern whose number was greater than 39
1621        went wrong under certain circumstances in UTF-8 mode. This bug could also
1622        have affected the operation of pcre_study().
1623    
1624    25. Realized that a little bit of performance could be had by replacing
1625        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1626    
1627    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1628    
1629    27. Possessive quantifiers such as a++ were previously implemented by turning
1630        them into atomic groups such as ($>a+). Now they have their own opcodes,
1631        which improves performance. This includes the automatically created ones
1632        from 23 above.
1633    
1634    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1635        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1636        the first matched character to be a colon. This applied both to named and
1637        numbered groups.
1638    
1639    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1640    
1641    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1642        building PCRE on Windows easier, so I have included it in the distribution.
1643    
1644    31. There is now a check in pcretest against a ridiculously large number being
1645        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1646        loop, the loop is abandoned.
1647    
1648    32. Forward references to subpatterns in conditions such as (?(2)...) where
1649        subpattern 2 is defined later cause pcre_compile() to search forwards in
1650        the pattern for the relevant set of parentheses. This search went wrong
1651        when there were unescaped parentheses in a character class, parentheses
1652        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1653    
1654    33. "Subroutine" calls and backreferences were previously restricted to
1655        referencing subpatterns earlier in the regex. This restriction has now
1656        been removed.
1657    
1658    34. Added a number of extra features that are going to be in Perl 5.10. On the
1659        whole, these are just syntactic alternatives for features that PCRE had
1660        previously implemented using the Python syntax or my own invention. The
1661        other formats are all retained for compatibility.
1662    
1663        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1664            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1665            also .NET compatible.
1666    
1667        (b) A recursion or subroutine call to a named group can now be defined as
1668            (?&name) as well as (?P>name).
1669    
1670        (c) A backreference to a named group can now be defined as \k<name> or
1671            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1672            5.10, are also .NET compatible.
1673    
1674        (d) A conditional reference to a named group can now use the syntax
1675            (?(<name>) or (?('name') as well as (?(name).
1676    
1677        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1678            groups (named and numbered) that are never evaluated inline, but can be
1679            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1680            is always false. There may be only one alternative in such a group.
1681    
1682        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1683            as the simple (?(R). The condition is true only if the most recent
1684            recursion is that of the given number or name. It does not search out
1685            through the entire recursion stack.
1686    
1687        (g) The escape \gN or \g{N} has been added, where N is a positive or
1688            negative number, specifying an absolute or relative reference.
1689    
1690    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1691        some "unreachable code" warnings.
1692    
1693    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1694        things, this adds five new scripts.
1695    
1696    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1697        There were also incompatibilities regarding the handling of \Q..\E inside
1698        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1699        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1700    
1701    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1702        matches an empty string, and forcibly breaks the loop. There were bugs in
1703        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1704        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1705        separate and independent bugs (that affected different cases) have been
1706        fixed.
1707    
1708    39. Refactored the code to abolish the use of different opcodes for small
1709        capturing bracket numbers. This is a tidy that I avoided doing when I
1710        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1711        The new approach is not only tidier, it makes it possible to reduce the
1712        memory needed to fix the previous bug (38).
1713    
1714    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1715        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1716        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1717        mode.
1718    
1719    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1720        report.
1721    
1722    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1723        copy construction and assignment in the C++ wrapper.
1724    
1725    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1726        couple of bugs that could have given wrong results in the "--newline=crlf"
1727        case.
1728    
1729    44. Added a number of casts and did some reorganization of signed/unsigned int
1730        variables following suggestions from Dair Grant. Also renamed the variable
1731        "this" as "item" because it is a C++ keyword.
1732    
1733    45. Arranged for dftables to add
1734    
1735          #include "pcre_internal.h"
1736    
1737        to pcre_chartables.c because without it, gcc 4.x may remove the array
1738        definition from the final binary if PCRE is built into a static library and
1739        dead code stripping is activated.
1740    
1741    46. For an unanchored pattern, if a match attempt fails at the start of a
1742        newline sequence, and the newline setting is CRLF or ANY, and the next two
1743        characters are CRLF, advance by two characters instead of one.
1744    
1745    
1746    Version 6.7 04-Jul-06
1747    ---------------------
1748    
1749     1. In order to handle tests when input lines are enormously long, pcretest has
1750        been re-factored so that it automatically extends its buffers when
1751        necessary. The code is crude, but this _is_ just a test program. The
1752        default size has been increased from 32K to 50K.
1753    
1754     2. The code in pcre_study() was using the value of the re argument before
1755        testing it for NULL. (Of course, in any sensible call of the function, it
1756        won't be NULL.)
1757    
1758     3. The memmove() emulation function in pcre_internal.h, which is used on
1759        systems that lack both memmove() and bcopy() - that is, hardly ever -
1760        was missing a "static" storage class specifier.
1761    
1762     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1763        containing an extended class (one that cannot be represented by a bitmap
1764        because it contains high-valued characters or Unicode property items, e.g.
1765        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1766        pattern, but PCRE should not loop if you do not (it no longer does).
1767        [Detail: two cases were found: (a) a repeated subpattern containing an
1768        extended class; (b) a recursive reference to a subpattern that followed a
1769        previous extended class. It wasn't skipping over the extended class
1770        correctly when UTF-8 mode was not set.]
1771    
1772     5. A negated single-character class was not being recognized as fixed-length
1773        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1774        compile error "lookbehind assertion is not fixed length".
1775    
1776     6. The RunPerlTest auxiliary script was showing an unexpected difference
1777        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1778        write a Perl script that can interpret lines of an input file either as
1779        byte characters or as UTF-8, which is what "perltest" was being required to
1780        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1781        can't do is switch easily at run time between having the "use utf8;" pragma
1782        or not. In the end, I fudged it by using the RunPerlTest script to insert
1783        "use utf8;" explicitly for the UTF-8 tests.
1784    
1785     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1786        the end of the subject string, contrary to the documentation and to what
1787        Perl does. This was true of both matching functions. Now it matches only at
1788        the start of the subject and immediately after *internal* newlines.
1789    
1790     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1791        a pointer to an int instead of a pointer to an unsigned long int. This
1792        caused problems on 64-bit systems.
1793    
1794     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1795        instance of the 'standard' template library not being so standard".
1796    
1797    10. There was no check on the number of named subpatterns nor the maximum
1798        length of a subpattern name. The product of these values is used to compute
1799        the size of the memory block for a compiled pattern. By supplying a very
1800        long subpattern name and a large number of named subpatterns, the size
1801        computation could be caused to overflow. This is now prevented by limiting
1802        the length of names to 32 characters, and the number of named subpatterns
1803        to 10,000.
1804    
1805    11. Subpatterns that are repeated with specific counts have to be replicated in
1806        the compiled pattern. The size of memory for this was computed from the
1807        length of the subpattern and the repeat count. The latter is limited to
1808        65535, but there was no limit on the former, meaning that integer overflow
1809        could in principle occur. The compiled length of a repeated subpattern is
1810        now limited to 30,000 bytes in order to prevent this.
1811    
1812    12. Added the optional facility to have named substrings with the same name.
1813    
1814    13. Added the ability to use a named substring as a condition, using the
1815        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1816        are numbers (not recommended). Forward references are permitted.
1817    
1818    14. Added forward references in named backreferences (if you see what I mean).
1819    
1820    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1821        pattern could run off the end of the subject. For example, the pattern
1822        "(?s)(.{1,5})"8 did this with the subject "ab".
1823    
1824    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1825        PCRE_CASELESS was set when matching characters that were quantified with ?
1826        or *.
1827    
1828    17. A character class other than a single negated character that had a minimum
1829        but no maximum quantifier - for example [ab]{6,} - was not handled
1830        correctly by pce_dfa_exec(). It would match only one character.
1831    
1832    18. A valid (though odd) pattern that looked like a POSIX character
1833        class but used an invalid character after [ (for example [[,abc,]]) caused
1834        pcre_compile() to give the error "Failed: internal error: code overflow" or
1835        in some cases to crash with a glibc free() error. This could even happen if
1836        the pattern terminated after [[ but there just happened to be a sequence of
1837        letters, a binary zero, and a closing ] in the memory that followed.
1838    
1839    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1840        over the years. Originally (before any Unicode support), just the bottom 8
1841        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1842        output from "man perlunicode" includes this:
1843    
1844          The regular expression compiler produces polymorphic opcodes.  That
1845          is, the pattern adapts to the data and automatically switches to
1846          the Unicode character scheme when presented with Unicode data--or
1847          instead uses a traditional byte scheme when presented with byte
1848          data.
1849    
1850        Sadly, a wide octal escape does not cause a switch, and in a string with
1851        no other multibyte characters, these octal escapes are treated as before.
1852        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1853        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1854        Unicode string.
1855    
1856        I have not perpetrated such confusion in PCRE. Up till now, it took just
1857        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1858        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1859        translate to the appropriate multibyte character.
1860    
1861    29. Applied some refactoring to reduce the number of warnings from Microsoft
1862        and Borland compilers. This has included removing the fudge introduced
1863        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1864        a warning about an unused variable.
1865    
1866    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1867        characters since release 4.0, because Perl (from release 5.004) does not.
1868        [Or at least, is documented not to: some releases seem to be in conflict
1869        with the documentation.] However, when a pattern was studied with
1870        pcre_study() and all its branches started with \s, PCRE still included VT
1871        as a possible starting character. Of course, this did no harm; it just
1872        caused an unnecessary match attempt.
1873    
1874    22. Removed a now-redundant internal flag bit that recorded the fact that case
1875        dependency changed within the pattern. This was once needed for "required
1876        byte" processing, but is no longer used. This recovers a now-scarce options
1877        bit. Also moved the least significant internal flag bit to the most-
1878        significant bit of the word, which was not previously used (hangover from
1879        the days when it was an int rather than a uint) to free up another bit for
1880        the future.
1881    
1882    23. Added support for CRLF line endings as well as CR and LF. As well as the
1883        default being selectable at build time, it can now be changed at runtime
1884        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1885        specify that it is scanning data with non-default line endings.
1886    
1887    24. Changed the definition of CXXLINK to make it agree with the definition of
1888        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1889    
1890    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1891        recursions. This makes a big different to stack usage for some patterns.
1892    
1893    26. If a subpattern containing a named recursion or subroutine reference such
1894        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1895        the space required for the compiled pattern went wrong and gave too small a
1896        value. Depending on the environment, this could lead to "Failed: internal
1897        error: code overflow at offset 49" or "glibc detected double free or
1898        corruption" errors.
1899    
1900    27. Applied patches from Google (a) to support the new newline modes and (b) to
1901        advance over multibyte UTF-8 characters in GlobalReplace.
1902    
1903    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1904        difference for some implementation of PCRE in some Windows version.
1905    
1906    29. Added some extra testing facilities to pcretest:
1907    
1908        \q<number>   in a data line sets the "match limit" value
1909        \Q<number>   in a data line sets the "match recursion limt" value
1910        -S <number>  sets the stack size, where <number> is in megabytes
1911    
1912        The -S option isn't available for Windows.
1913    
1914    
1915    Version 6.6 06-Feb-06
1916    ---------------------
1917    
1918     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1919        in pcreposix.h. I have copied the definition from pcre.h.
1920    
1921     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1922        because pcre.h is no longer a built file.
1923    
1924     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1925        not normally included in the compiled code.
1926    
1927    
1928    Version 6.5 01-Feb-06
1929    ---------------------
1930    
1931     1. When using the partial match feature with pcre_dfa_exec(), it was not
1932        anchoring the second and subsequent partial matches at the new starting
1933        point. This could lead to incorrect results. For example, with the pattern
1934        /1234/, partially matching against "123" and then "a4" gave a match.
1935    
1936     2. Changes to pcregrep:
1937    
1938        (a) All non-match returns from pcre_exec() were being treated as failures
1939            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1940            error message is output. Some extra information is given for the
1941            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1942            probably the only errors that are likely to be caused by users (by
1943            specifying a regex that has nested indefinite repeats, for instance).
1944            If there are more than 20 of these errors, pcregrep is abandoned.
1945    
1946        (b) A binary zero was treated as data while matching, but terminated the
1947            output line if it was written out. This has been fixed: binary zeroes
1948            are now no different to any other data bytes.
1949    
1950        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1951            used to set a locale for matching. The --locale=xxxx long option has
1952            been added (no short equivalent) to specify a locale explicitly on the
1953            pcregrep command, overriding the environment variables.
1954    
1955        (d) When -B was used with -n, some line numbers in the output were one less
1956            than they should have been.
1957    
1958        (e) Added the -o (--only-matching) option.
1959    
1960        (f) If -A or -C was used with -c (count only), some lines of context were
1961            accidentally printed for the final match.
1962    
1963        (g) Added the -H (--with-filename) option.
1964    
1965        (h) The combination of options -rh failed to suppress file names for files
1966            that were found from directory arguments.
1967    
1968        (i) Added the -D (--devices) and -d (--directories) options.
1969    
1970        (j) Added the -F (--fixed-strings) option.
1971    
1972        (k) Allow "-" to be used as a file name for -f as well as for a data file.
1973    
1974        (l) Added the --colo(u)r option.
1975    
1976        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1977            is not present by default.
1978    
1979     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1980        items such as (?R) or (?1), when the recursion could match a number of
1981        alternatives. If it matched one of the alternatives, but subsequently,
1982        outside the recursion, there was a failure, the code tried to back up into
1983        the recursion. However, because of the way PCRE is implemented, this is not
1984        possible, and the result was an incorrect result from the match.
1985    
1986        In order to prevent this happening, the specification of recursion has
1987        been changed so that all such subpatterns are automatically treated as
1988        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1989    
1990     4. I had overlooked the fact that, in some locales, there are characters for
1991        which isalpha() is true but neither isupper() nor islower() are true. In
1992        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1993        and ordfeminine) are like this. This affected the treatment of \w and \W
1994        when they appeared in character classes, but not when they appeared outside
1995        a character class. The bit map for "word" characters is now created
1996        separately from the results of isalnum() instead of just taking it from the
1997        upper, lower, and digit maps. (Plus the underscore character, of course.)
1998    
1999     5. The above bug also affected the handling of POSIX character classes such as
2000        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2001        permanent tables. Instead, the bit maps for such a class were previously
2002        created as the appropriate unions of the upper, lower, and digit bitmaps.
2003        Now they are created by subtraction from the [[:word:]] class, which has
2004        its own bitmap.
2005    
2006     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2007        It is created by subtracting the vertical space characters (\x09, \x0a,
2008        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2009        subtraction was done in the overall bitmap for a character class, meaning
2010        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2011        be recognized. This bug has been fixed.
2012    
2013     7. Patches from the folks at Google:
2014    
2015          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2016          real life, but is still worth protecting against".
2017    
2018          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2019          regular expressions".
2020    
2021          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2022          have it.
2023    
2024          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2025          "configure" and the latter not, in order to fix a problem somebody had
2026          with compiling the Arg class on HP-UX.
2027    
2028          (e) Improve the error-handling of the C++ wrapper a little bit.
2029    
2030          (f) New tests for checking recursion limiting.
2031    
2032     8. The pcre_memmove() function, which is used only if the environment does not
2033        have a standard memmove() function (and is therefore rarely compiled),
2034        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2035        returning a result (though PCRE never actually uses the result).
2036    
2037     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2038        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2039        returned instead of calling malloc() with an overflowing number that would
2040        most likely cause subsequent chaos.
2041    
2042    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2043    
2044    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2045        with this option is matched, the nmatch and pmatch options of regexec() are
2046        ignored.
2047    
2048    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2049        provided in case anyone wants to the the POSIX interface with UTF-8
2050        strings.
2051    
2052    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2053        C++ linking (needed for some HP-UX environments).
2054    
2055    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2056        (unused parameter) and in the pcre_printint() function (omitted "default"
2057        switch label when the default is to do nothing).
2058    
2059    15. Added some code to make it possible, when PCRE is compiled as a C++
2060        library, to replace subject pointers for pcre_exec() with a smart pointer
2061        class, thus making it possible to process discontinuous strings.
2062    
2063    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2064        much the same function. They were added by different people who were trying
2065        to make PCRE easy to compile on non-Unix systems. It has been suggested
2066        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2067        for compiling on Windows systems. I have therefore replaced it with
2068        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2069        defaults to "extern" for C or "extern C" for C++, which works fine on
2070        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2071        SCOPE with something explicit in config.h. In addition:
2072    
2073        (a) pcreposix.h still had just "extern" instead of either of these macros;
2074            I have replaced it with PCRE_DATA_SCOPE.
2075    
2076        (b) Functions such as _pcre_xclass(), which are internal to the library,
2077            but external in the C sense, all had PCRE_EXPORT in their definitions.
2078            This is apparently wrong for the Windows case, so I have removed it.
2079            (It makes no difference on Unix-like systems.)
2080    
2081    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2082        of recursive calls to match(). This is different to MATCH_LIMIT because
2083        that limits the total number of calls to match(), not all of which increase
2084        the depth of recursion. Limiting the recursion depth limits the amount of
2085        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2086        when PCRE is compiled, and changed at run time. A patch from Google adds
2087        this functionality to the C++ interface.
2088    
2089    18. Changes to the handling of Unicode character properties:
2090    
2091        (a) Updated the table to Unicode 4.1.0.
2092    
2093        (b) Recognize characters that are not in the table as "Cn" (undefined).
2094    
2095        (c) I revised the way the table is implemented to a much improved format
2096            which includes recognition of ranges. It now supports the ranges that
2097            are defined in UnicodeData.txt, and it also amalgamates other
2098            characters into ranges. This has reduced the number of entries in the
2099            table from around 16,000 to around 3,000, thus reducing its size
2100            considerably. I realized I did not need to use a tree structure after
2101            all - a binary chop search is just as efficient. Having reduced the
2102            number of entries, I extended their size from 6 bytes to 8 bytes to
2103            allow for more data.
2104    
2105        (d) Added support for Unicode script names via properties such as \p{Han}.
2106    
2107    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2108        matching that character.
2109    
2110    20. When matching a repeated Unicode property with a minimum greater than zero,
2111        (for example \pL{2,}), PCRE could look past the end of the subject if it
2112        reached it while seeking the minimum number of characters. This could
2113        happen only if some of the characters were more than one byte long, because
2114        there is a check for at least the minimum number of bytes.
2115    
2116    21. Refactored the implementation of \p and \P so as to be more general, to
2117        allow for more different types of property in future. This has changed the
2118        compiled form incompatibly. Anybody with saved compiled patterns that use
2119        \p or \P will have to recompile them.
2120    
2121    22. Added "Any" and "L&" to the supported property types.
2122    
2123    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2124        but give a compile time error if the value is greater than 0xff.
2125    
2126    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2127        accidentally not being installed or uninstalled.
2128    
2129    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2130        made were to insert the current release number. This seemed silly, because
2131        it made things harder for people building PCRE on systems that don't run
2132        "configure". I have turned pcre.h into a distributed file, no longer built
2133        by "configure", with the version identification directly included. There is
2134        no longer a pcre.h.in file.
2135    
2136        However, this change necessitated a change to the pcre-config script as
2137        well. It is built from pcre-config.in, and one of the substitutions was the
2138        release number. I have updated configure.ac so that ./configure now finds
2139        the release number by grepping pcre.h.
2140    
2141    26. Added the ability to run the tests under valgrind.
2142    
2143    
2144    Version 6.4 05-Sep-05
2145    ---------------------
2146    
2147     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2148        "--" to be printed when multiple files were scanned, even when none of the
2149        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2150        consider it to be a bug, and have restored the previous behaviour.
2151    
2152     2. A couple of code tidies to get rid of compiler warnings.
2153    
2154     3. The pcretest program used to cheat by referring to symbols in the library
2155        whose names begin with _pcre_. These are internal symbols that are not
2156        really supposed to be visible externally, and in some environments it is
2157        possible to suppress them. The cheating is now confined to including
2158        certain files from the library's source, which is a bit cleaner.
2159    
2160     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2161        file's purpose clearer.
2162    
2163     5. Reorganized pcre_ucp_findchar().
2164    
2165    
2166    Version 6.3 15-Aug-05
2167    ---------------------
2168    
2169     1. The file libpcre.pc.in did not have general read permission in the tarball.
2170    
2171     2. There were some problems when building without C++ support:
2172    
2173        (a) If C++ support was not built, "make install" and "make test" still
2174            tried to test it.
2175    
2176        (b) There were problems when the value of CXX was explicitly set. Some
2177            changes have been made to try to fix these, and ...
2178    
2179        (c) --disable-cpp can now be used to explicitly disable C++ support.
2180    
2181        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2182            backslash in a target when C++ was disabled. This confuses some
2183            versions of "make", apparently. Using an intermediate variable solves
2184            this. (Same for CPP_LOBJ.)
2185    
2186     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2187        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2188        necessary on certain architectures.
2189    
2190     4. Added a setting of -export-symbols-regex to the link command to remove
2191        those symbols that are exported in the C sense, but actually are local
2192        within the library, and not documented. Their names all begin with
2193        "_pcre_". This is not a perfect job, because (a) we have to except some
2194        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2195        available (and never for static libraries). I have made a note to try to
2196        find a way round (a) in the future.
2197    
2198    
2199    Version 6.2 01-Aug-05
2200    ---------------------
2201    
2202     1. There was no test for integer overflow of quantifier values. A construction
2203        such as {1111111111111111} would give undefined results. What is worse, if
2204        a minimum quantifier for a parenthesized subpattern overflowed and became
2205        negative, the calculation of the memory size went wrong. This could have
2206        led to memory overwriting.
2207    
2208     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2209    
2210     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2211        operating environments where this matters.
2212    
2213     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2214        PCRE options from within the C++ wrapper.
2215    
2216     5. Named capturing subpatterns were not being correctly counted when a pattern
2217        was compiled. This caused two problems: (a) If there were more than 100
2218        such subpatterns, the calculation of the memory needed for the whole
2219        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2220        back references of the form \12, where the number was greater than 9, were
2221        not recognized as back references, even though there were sufficient
2222        previous subpatterns.
2223    
2224     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2225        versions of gcc, e.g. 2.95.4.
2226    
2227    
2228    Version 6.1 21-Jun-05
2229    ---------------------
2230    
2231     1. There was one reference to the variable "posix" in pcretest.c that was not
2232        surrounded by "#if !defined NOPOSIX".
2233    
2234     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2235        the cross-check on the old pcre_info() function, for the benefit of the
2236        cut-down version of PCRE that is currently imported into Exim.
2237    
2238     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2239        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2240        patterns that start (?i) but I don't think that matters. The use of (?i) is
2241        just an example; this all applies to the other options as well.
2242    
2243     4. Since libtool seems to echo the compile commands it is issuing, the output
2244        from "make" can be reduced a bit by putting "@" in front of each libtool
2245        compile command.
2246    
2247     5. Patch from the folks at Google for configure.in to be a bit more thorough
2248        in checking for a suitable C++ installation before trying to compile the
2249        C++ stuff. This should fix a reported problem when a compiler was present,
2250        but no suitable headers.
2251    
2252     6. The man pages all had just "PCRE" as their title. I have changed them to
2253        be the relevant file name. I have also arranged that these names are
2254        retained in the file doc/pcre.txt, which is a concatenation in text format
2255        of all the man pages except the little individual ones for each function.
2256    
2257     7. The NON-UNIX-USE file had not been updated for the different set of source
2258        files that come with release 6. I also added a few comments about the C++
2259        wrapper.
2260    
2261    
2262    Version 6.0 07-Jun-05
2263    ---------------------
2264    
2265     1. Some minor internal re-organization to help with my DFA experiments.
2266    
2267     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2268        didn't matter for the library itself when fully configured, but did matter
2269        when compiling without UCP support, or within Exim, where the ucp files are
2270        not imported.
2271    
2272     3. Refactoring of the library code to split up the various functions into
2273        different source modules. The addition of the new DFA matching code (see
2274        below) to a single monolithic source would have made it really too
2275        unwieldy, quite apart from causing all the code to be include in a
2276        statically linked application, when only some functions are used. This is
2277        relevant even without the DFA addition now that patterns can be compiled in
2278        one application and matched in another.
2279    
2280        The downside of splitting up is that there have to be some external
2281        functions and data tables that are used internally in different modules of
2282        the library but which are not part of the API. These have all had their
2283        names changed to start with "_pcre_" so that they are unlikely to clash
2284        with other external names.
2285    
2286     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2287        a different (DFA) algorithm. Although it is slower than the original
2288        function, it does have some advantages for certain types of matching
2289        problem.
2290    
2291     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2292        including restarting after a partial match.
2293    
2294     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2295        defined when compiling for Windows was sent to me. I have put it into the
2296        code, though I have no means of testing or verifying it.
2297    
2298     7. Added the pcre_refcount() auxiliary function.
2299    
2300     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2301        match before or at the first newline in the subject string. In pcretest,
2302        the /f option on a pattern can be used to set this.
2303    
2304     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2305        would behave wrongly. This has been present in PCRE since release 4.0.
2306    
2307    10. A number of changes to the pcregrep command:
2308    
2309        (a) Refactored how -x works; insert ^(...)$ instead of setting
2310            PCRE_ANCHORED and checking the length, in preparation for adding
2311            something similar for -w.
2312    
2313        (b) Added the -w (match as a word) option.
2314    
2315        (c) Refactored the way lines are read and buffered so as to have more
2316            than one at a time available.
2317    
2318        (d) Implemented a pcregrep test script.
2319    
2320        (e) Added the -M (multiline match) option. This allows patterns to match
2321            over several lines of the subject. The buffering ensures that at least
2322            8K, or the rest of the document (whichever is the shorter) is available
2323            for matching (and similarly the previous 8K for lookbehind assertions).
2324    
2325        (f) Changed the --help output so that it now says
2326    
2327              -w, --word-regex(p)
2328    
2329            instead of two lines, one with "regex" and the other with "regexp"
2330            because that confused at least one person since the short forms are the
2331            same. (This required a bit of code, as the output is generated
2332            automatically from a table. It wasn't just a text change.)
2333    
2334        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2335            option but starts with a hyphen. Could be a pattern or a path name
2336            starting with a hyphen, for instance.
2337    
2338        (h) "-" can be given as a file name to represent stdin.
2339    
2340        (i) When file names are being printed, "(standard input)" is used for
2341            the standard input, for compatibility with GNU grep. Previously
2342            "<stdin>" was used.
2343    
2344        (j) The option --label=xxx can be used to supply a name to be used for
2345            stdin when file names are being printed. There is no short form.
2346    
2347        (k) Re-factored the options decoding logic because we are going to add
2348            two more options that take data. Such options can now be given in four
2349            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2350    
2351        (l) Added the -A, -B, and -C options for requesting that lines of context
2352            around matches be printed.
2353    
2354        (m) Added the -L option to print the names of files that do not contain
2355            any matching lines, that is, the complement of -l.
2356    
2357        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2358            continue to scan other files.
2359    
2360        (o) The -s option was incorrectly implemented. For compatibility with other
2361            greps, it now suppresses the error message for a non-existent or non-
2362            accessible file (but not the return code). There is a new option called
2363            -q that suppresses the output of matching lines, which was what -s was
2364            previously doing.
2365    
2366        (p) Added --include and --exclude options to specify files for inclusion
2367            and exclusion when recursing.
2368    
2369    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2370        Hopefully, it now does.
2371    
2372    12. Missing cast in pcre_study().
2373    
2374    13. Added an "uninstall" target to the makefile.
2375    
2376    14. Replaced "extern" in the function prototypes in Makefile.in with
2377        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2378        world, but is set differently for Windows.
2379    
2380    15. Added a second compiling function called pcre_compile2(). The only
2381        difference is that it has an extra argument, which is a pointer to an
2382        integer error code. When there is a compile-time failure, this is set
2383        non-zero, in addition to the error test pointer being set to point to an
2384        error message. The new argument may be NULL if no error number is required
2385        (but then you may as well call pcre_compile(), which is now just a
2386        wrapper). This facility is provided because some applications need a
2387        numeric error indication, but it has also enabled me to tidy up the way
2388        compile-time errors are handled in the POSIX wrapper.
2389    
2390    16. Added VPATH=.libs to the makefile; this should help when building with one
2391        prefix path and installing with another. (Or so I'm told by someone who
2392        knows more about this stuff than I do.)
2393    
2394    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2395        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2396        match everything, including newlines. This is not POSIX-compatible, but
2397        somebody wanted the feature. From pcretest it can be activated by using
2398        both the P and the s flags.
2399    
2400    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2401    
2402    19. libpcre.pc was being incorrectly installed as executable.
2403    
2404    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2405        it now also looks for '\r' so that it will work unmodified on Windows.
2406    
2407    21. Added Google's contributed C++ wrapper to the distribution.
2408    
2409    22. Added some untidy missing memory free() calls in pcretest, to keep
2410        Electric Fence happy when testing.
2411    
2412    
2413    
2414    Version 5.0 13-Sep-04
2415    ---------------------
2416    
2417     1. Internal change: literal characters are no longer packed up into items
2418        containing multiple characters in a single byte-string. Each character
2419        is now matched using a separate opcode. However, there may be more than one
2420        byte in the character in UTF-8 mode.
2421    
2422     2. The pcre_callout_block structure has two new fields: pattern_position and
2423        next_item_length. These contain the offset in the pattern to the next match
2424        item, and its length, respectively.
2425    
2426     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2427        insertion of callouts before each pattern item. Added the /C option to
2428        pcretest to make use of this.
2429    
2430     4. On the advice of a Windows user, the lines
2431    
2432          #if defined(_WIN32) || defined(WIN32)
2433          _setmode( _fileno( stdout ), 0x8000 );
2434          #endif  /* defined(_WIN32) || defined(WIN32) */
2435    
2436        have been added to the source of pcretest. This apparently does useful
2437        magic in relation to line terminators.
2438    
2439     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2440        for the benefit of those environments where the "b" makes a difference.
2441    
2442     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2443        to know about it. I have put a hack into configure.in that adds in code
2444        to set GCC=yes if CC=icc. This seems to end up at a point in the
2445        generated configure script that is early enough to affect the setting of
2446        compiler options, which is what is needed, but I have no means of testing
2447        whether it really works. (The user who reported this had patched the
2448        generated configure script, which of course I cannot do.)
2449    
2450        LATER: After change 22 below (new libtool files), the configure script
2451        seems to know about icc (and also ecc). Therefore, I have commented out
2452        this hack in configure.in.
2453    
2454     7. Added support for pkg-config (2 patches were sent in).
2455    
2456     8. Negated POSIX character classes that used a combination of internal tables
2457        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2458        [[:^ascii]]. Typically, they would match almost any characters. The other
2459        POSIX classes were not broken in this way.
2460    
2461     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2462        to find the match, as PCRE was deluded into thinking that the match had to
2463        start at the start point or following a newline. The same bug applied to
2464        patterns with negative forward assertions or any backward assertions
2465        preceding ".*" at the start, unless the pattern required a fixed first
2466        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2467    
2468    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2469        starting at the last subject character, bytes beyond the end of the subject
2470        string were read.
2471    
2472    11. Renamed the variable "class" as "classbits" to make life easier for C++
2473        users. (Previously there was a macro definition, but it apparently wasn't
2474        enough.)
2475    
2476    12. Added the new field "tables" to the extra data so that tables can be passed
2477        in at exec time, or the internal tables can be re-selected. This allows
2478        a compiled regex to be saved and re-used at a later time by a different
2479        program that might have everything at different addresses.
2480    
2481    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2482        -R library as well as a -L library.
2483    
2484    14. The debugging options of pcretest (-d on the command line or D on a
2485        pattern) showed incorrect output for anything following an extended class
2486        that contained multibyte characters and which was followed by a quantifier.
2487    
2488    15. Added optional support for general category Unicode character properties
2489        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2490        support. It adds about 90K to the size of the library. The meanings of the
2491        inbuilt class escapes such as \d and \s have NOT been changed.
2492    
2493    16. Updated pcredemo.c to include calls to free() to release the memory for the
2494        compiled pattern.
2495    
2496    17. The generated file chartables.c was being created in the source directory
2497        instead of in the building directory. This caused the build to fail if the
2498        source directory was different from the building directory, and was
2499        read-only.
2500    
2501    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2502        file. No doubt somebody will tell me if they don't make sense... Also added
2503        Dan Mooney's comments about building on OpenVMS.
2504    
2505    19. Added support for partial matching via the PCRE_PARTIAL option for
2506        pcre_exec() and the \P data escape in pcretest.
2507    
2508    20. Extended pcretest with 3 new pattern features:
2509    
2510        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2511              write the compiled pattern to the file whose name is "rest-of-line".
2512              This is a straight binary dump of the data, with the saved pointer to
2513              the character tables forced to be NULL. The study data, if any, is
2514              written too. After writing, pcretest reads a new pattern.
2515    
2516        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2517              compiled pattern from the given file. There must not be any
2518              occurrences of "<" in the file name (pretty unlikely); if there are,
2519              pcretest will instead treat the initial "<" as a pattern delimiter.
2520              After reading in the pattern, pcretest goes on to read data lines as
2521              usual.
2522    
2523        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2524              and 16-bit fields in a compiled pattern, to simulate a pattern that
2525              was compiled on a host of opposite endianness.
2526    
2527    21. The pcre-exec() function can now cope with patterns that were compiled on
2528        hosts of opposite endianness, with this restriction:
2529    
2530          As for any compiled expression that is saved and used later, the tables
2531          pointer field cannot be preserved; the extra_data field in the arguments
2532          to pcre_exec() should be used to pass in a tables address if a value
2533          other than the default internal tables were used at compile time.
2534    
2535    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2536        now diagnosed as an error. Previously, most of the time, a negative number
2537        would have been treated as zero, but if in addition "ovector" was passed as
2538        NULL, a crash could occur.
2539    
2540    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2541        new versions from the libtool 1.5 distribution (the last one is a copy of
2542        a file called libtool.m4). This seems to have fixed the need to patch
2543        "configure" to support Darwin 1.3 (which I used to do). However, I still
2544        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2545        workstation).
2546    
2547    24. Changed the PCRE licence to be the more standard "BSD" licence.
2548    
2549    
2550    Version 4.5 01-Dec-03
2551    ---------------------
2552    
2553     1. There has been some re-arrangement of the code for the match() function so
2554        that it can be compiled in a version that does not call itself recursively.
2555        Instead, it keeps those local variables that need separate instances for
2556        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2557        needs to "recurse". Keeping track of where control must go is done by means
2558        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2559        hide most of the details from the main code, and operates only if
2560        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2561        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2562        operating.
2563    
2564        To make it easier for callers to provide specially tailored get/free
2565        functions for this usage, two new functions, pcre_stack_malloc, and
2566        pcre_stack_free, are used. They are always called in strict stacking order,
2567        and the size of block requested is always the same.
2568    
2569        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2570        PCRE has been compiled to use the stack or the heap for recursion. The
2571        -C option of pcretest uses this to show which version is compiled.
2572    
2573        A new data escape \S, is added to pcretest; it causes the amounts of store
2574        obtained and freed by both kinds of malloc/free at match time to be added
2575        to the output.
2576    
2577     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2578        what's available on my current Linux desktop machine.
2579    
2580     3. When matching a UTF-8 string, the test for a valid string at the start has
2581        been extended. If start_offset is not zero, PCRE now checks that it points
2582        to a byte that is the start of a UTF-8 character. If not, it returns
2583        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2584        this is necessary because there may be backward assertions in the pattern.
2585        When matching the same subject several times, it may save resources to use
2586        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2587    
2588     4. The code for checking the validity of UTF-8 strings has been tightened so
2589        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2590        containing "overlong sequences".
2591    
2592     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2593        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2594        should have been "&", but it just so happened that all the cases this let
2595        through by mistake were picked up later in the function.
2596    
2597     6. I had used a variable called "isblank" - this is a C99 function, causing
2598        some compilers to warn. To avoid this, I renamed it (as "blankclass").
2599    
2600     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2601        prompting; (b) run "./pcretest /dev/null" at the start of the test script
2602        so the version is shown; (c) stop "make test" echoing "./RunTest".
2603    
2604     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2605    
2606     9. The prototype for memmove() for systems that don't have it was using
2607        size_t, but the inclusion of the header that defines size_t was later. I've
2608        moved the #includes for the C headers earlier to avoid this.
2609    
2610    10. Added some adjustments to the code to make it easier to compiler on certain
2611        special systems:
2612    
2613          (a) Some "const" qualifiers were missing.
2614          (b) Added the macro EXPORT before all exported functions; by default this
2615              is defined to be empty.
2616          (c) Changed the dftables auxiliary program (that builds chartables.c) so
2617              that it reads its output file name as an argument instead of writing
2618              to the standard output and assuming this can be redirected.
2619    
2620    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2621        class containing characters with values greater than 255, PCRE compilation
2622        went into a loop.
2623    
2624    12. A recursive reference to a subpattern that was within another subpattern
2625        that had a minimum quantifier of zero caused PCRE to crash. For example,
2626        (x(y(?2))z)? provoked this bug with a subject that got as far as the
2627        recursion. If the recursively-called subpattern itself had a zero repeat,
2628        that was OK.
2629    
2630    13. In pcretest, the buffer for reading a data line was set at 30K, but the
2631        buffer into which it was copied (for escape processing) was still set at
2632        1024, so long lines caused crashes.
2633    
2634    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2635        "internal error: code overflow...". This applied to any character class
2636        that was followed by a possessive quantifier.
2637    
2638    15. Modified the Makefile to add libpcre.la as a prerequisite for
2639        libpcreposix.la because I was told this is needed for a parallel build to
2640        work.
2641    
2642    16. If a pattern that contained .* following optional items at the start was
2643        studied, the wrong optimizing data was generated, leading to matching
2644        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2645        matching string must start with a or b or c. The correct conclusion for
2646        this pattern is that a match can start with any character.
2647    
2648    
2649    Version 4.4 13-Aug-03
2650    ---------------------
2651    
2652     1. In UTF-8 mode, a character class containing characters with values between
2653        127 and 255 was not handled correctly if the compiled pattern was studied.
2654        In fixing this, I have also improved the studying algorithm for such
2655        classes (slightly).
2656    
2657     2. Three internal functions had redundant arguments passed to them. Removal
2658        might give a very teeny performance improvement.
2659    
2660     3. Documentation bug: the value of the capture_top field in a callout is *one
2661        more than* the number of the hightest numbered captured substring.
2662    
2663     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2664        in incorrectly linking with a previously installed version. They now link
2665        explicitly with libpcre.la.
2666    
2667     5. configure.in no longer needs to recognize Cygwin specially.
2668    
2669     6. A problem in pcre.in for Windows platforms is fixed.
2670    
2671     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2672        pcretest, it used to include the size of the study block as part of its
2673        output. Unfortunately, the structure contains a field that has a different
2674        size on different hardware architectures. This meant that the tests that
2675        showed this size failed. As the block is currently always of a fixed size,
2676        this information isn't actually particularly useful in pcretest output, so
2677        I have just removed it.
2678    
2679     8. Three pre-processor statements accidentally did not start in column 1.
2680        Sadly, there are *still* compilers around that complain, even though
2681        standard C has not required this for well over a decade. Sigh.
2682    
2683     9. In pcretest, the code for checking callouts passed small integers in the
2684        callout_data field, which is a void * field. However, some picky compilers
2685        complained about the casts involved for this on 64-bit systems. Now
2686        pcretest passes the address of the small integer instead, which should get
2687        rid of the warnings.
2688    
2689    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2690        both compile and run time, and gives an error if an invalid UTF-8 sequence
2691        is found. There is a option for disabling this check in cases where the
2692        string is known to be correct and/or the maximum performance is wanted.
2693    
2694    11. In response to a bug report, I changed one line in Makefile.in from
2695    
2696            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2697        to
2698            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2699    
2700        to look similar to other lines, but I have no way of telling whether this
2701        is the right thing to do, as I do not use Windows. No doubt I'll get told
2702        if it's wrong...
2703    
2704    
2705    Version 4.3 21-May-03
2706    ---------------------
2707    
2708    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2709       Makefile.
2710    
2711    2. Some refactoring to improve the quality of the code:
2712    
2713       (i)   The utf8_table... variables are now declared "const".
2714    
2715       (ii)  The code for \cx, which used the "case flipping" table to upper case
2716             lower case letters, now just substracts 32. This is ASCII-specific,
2717             but the whole concept of \cx is ASCII-specific, so it seems
2718             reasonable.
2719    
2720       (iii) PCRE was using its character types table to recognize decimal and
2721             hexadecimal digits in the pattern. This is silly, because it handles
2722             only 0-9, a-f, and A-F, but the character types table is locale-
2723             specific, which means strange things might happen. A private
2724             table is now used for this - though it costs 256 bytes, a table is
2725             much faster than multiple explicit tests. Of course, the standard
2726             character types table is still used for matching digits in subject
2727             strings against \d.
2728    
2729       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2730             ending in _t are). So I've renamed it as ESC_tee.
2731    
2732    3. The first argument for regexec() in the POSIX wrapper should have been
2733       defined as "const".
2734    
2735    4. Changed pcretest to use malloc() for its buffers so that they can be
2736       Electric Fenced for debugging.
2737    
2738    5. There were several places in the code where, in UTF-8 mode, PCRE would try
2739       to read one or more bytes before the start of the subject string. Often this
2740       had no effect on PCRE's behaviour, but in some circumstances it could
2741       provoke a segmentation fault.
2742    
2743    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2744       to try to read one or more bytes before the start of the subject string.
2745    
2746    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2747       UTF-8 support could misbehave in various ways if the subject string
2748       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2749       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2750       back over UTF-8 characters.)
2751    
2752    
2753    Version 4.2 14-Apr-03
2754    ---------------------
2755    
2756    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2757    
2758    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2759         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2760         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2761         [WIN_PREFIX]: new variable, "cyg" for Cygwin
2762         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2763           and BUILD_EXEEXT
2764         Note: automatic setting of the BUILD variables is not yet working
2765         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2766           compile-time but not at link-time
2767         [LINK]: use for linking executables only
2768         make different versions for Windows and non-Windows
2769         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2770           libraries
2771         [LINK_FOR_BUILD]: new variable
2772         [OBJEXT]: use throughout
2773         [EXEEXT]: use throughout
2774         <winshared>: new target
2775         <wininstall>: new target
2776         <dftables.o>: use native compiler
2777         <dftables>: use native linker
2778         <install>: handle Windows platform correctly
2779         <clean>: ditto
2780         <check>: ditto
2781         copy DLL to top builddir before testing
2782    
2783       As part of these changes, -no-undefined was removed again. This was reported
2784       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2785       in any case.
2786    
2787    3. Some tidies to get rid of compiler warnings:
2788    
2789       . In the match_data structure, match_limit was an unsigned long int, whereas
2790         match_call_count was an int. I've made them both unsigned long ints.
2791    
2792       . In pcretest the fact that a const uschar * doesn't automatically cast to
2793         a void * provoked a warning.
2794    
2795       . Turning on some more compiler warnings threw up some "shadow" variables
2796         and a few more missing casts.
2797    
2798    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2799       option, a class that contained a single character with a value between 128
2800       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2801    
2802    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2803       option, a class that contained several characters, but with at least one
2804       whose value was between 128 and 255 caused PCRE to crash.
2805    
2806    
2807    Version 4.1 12-Mar-03
2808    ---------------------
2809    
2810    1. Compiling with gcc -pedantic found a couple of places where casts were
2811    needed, and a string in dftables.c that was longer than standard compilers are
2812    required to support.
2813    
2814    2. Compiling with Sun's compiler found a few more places where the code could
2815    be tidied up in order to avoid warnings.
2816    
2817    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2818    first of these names is deprecated in the latest Autoconf in favour of the name
2819    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2820    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2821    analogy I have changed it to CFLAGS_FOR_BUILD.
2822    
2823    4. Added -no-undefined to the linking command in the Makefile, because this is
2824    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2825    linking step for the pcreposix library.
2826    
2827    5. PCRE was failing to diagnose the case of two named groups with the same
2828    name.
2829    
2830    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2831    literal character that is needed in the subject for a match, and scans along to
2832    ensure that it is present before embarking on the full matching process. This
2833    saves time in cases of nested unlimited repeats that are never going to match.
2834    Problem: the scan can take a lot of time if the subject is very long (e.g.
2835    megabytes), thus penalizing straightforward matches. It is now done only if the
2836    amount of subject to be scanned is less than 1000 bytes.
2837    
2838    7. A lesser problem with the same optimization is that it was recording the
2839    first character of an anchored pattern as "needed", thus provoking a search
2840    right along the subject, even when the first match of the pattern was going to
2841    fail. The "needed" character is now not set for anchored patterns, unless it
2842    follows something in the pattern that is of non-fixed length. Thus, it still
2843    fulfils its original purpose of finding quick non-matches in cases of nested
2844    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2845    
2846    
2847    Version 4.0 17-Feb-03
2848    ---------------------
2849    
2850    1. If a comment in an extended regex that started immediately after a meta-item
2851    extended to the end of string, PCRE compiled incorrect data. This could lead to
2852    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2853    
2854    2. Moved to autoconf 2.53 and libtool 1.4.2.
2855    
2856    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2857    the special perltest8 script is no longer needed - all the tests can be run
2858    from a single perltest script.
2859    
2860    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2861    by \s. It has now been removed in PCRE. This means it isn't recognized as
2862    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2863    class [:space:] *does* include VT, thereby creating a mess.
2864    
2865    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2866    space and tab.
2867    
2868    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2869    its new features into the main test script, reducing the number of scripts.
2870    
2871    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2872    were backward compatible, and made the (?i) apply to the whole pattern, as if
2873    /i were given. Now it behaves more logically, and applies the option setting
2874    only to what follows. PCRE has been changed to follow suit. However, if it
2875    finds options settings right at the start of the pattern, it extracts them into
2876    the global options, as before. Thus, they show up in the info data.
2877    
2878    8. Added support for the \Q...\E escape sequence. Characters in between are
2879    treated as literals. This is slightly different from Perl in that $ and @ are
2880    also handled as literals inside the quotes. In Perl, they will cause variable
2881    interpolation. Note the following examples:
2882    
2883        Pattern            PCRE matches      Perl matches
2884    
2885        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
2886        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
2887        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
2888    
2889    For compatibility with Perl, \Q...\E sequences are recognized inside character
2890    classes as well as outside them.
2891    
2892    9. Re-organized 3 code statements in pcretest to avoid "overflow in
2893    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2894    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2895    signed/unsigned warnings.
2896    
2897    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2898    option for pcretest, so I've replaced it by a simple function that does just
2899    that job.
2900    
2901    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2902    "pcregrep -".
2903    
2904    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2905    Java package. This provides some syntactic sugar for simple cases of what my
2906    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2907    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2908    item, you can use this simplified notation. Note that only makes sense with
2909    greedy quantifiers. Consequently, the use of the possessive quantifier forces
2910    greediness, whatever the setting of the PCRE_UNGREEDY option.
2911    
2912    13. A change of greediness default within a pattern was not taking effect at
2913    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2914    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2915    was abstracted outside.
2916    
2917    14. PCRE now supports the \G assertion. It is true when the current matching
2918    position is at the start point of the match. This differs from \A when the
2919    starting offset is non-zero. Used with the /g option of pcretest (or similar
2920    code), it works in the same way as it does for Perl's /g option. If all
2921    alternatives of a regex begin with \G, the expression is anchored to the start
2922    match position, and the "anchored" flag is set in the compiled expression.
2923    
2924    15. Some bugs concerning the handling of certain option changes within patterns
2925    have been fixed. These applied to options other than (?ims). For example,
2926    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2927    been the other way round. Some of this was related to change 7 above.
2928    
2929    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2930    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2931    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2932    POSIX classes only within a class (e.g. /[[:alpha:]]/).
2933    
2934    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2935    mode. Unlike ".", it always matches newline, whatever the setting of
2936    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2937    assertions. Perl allows it, but it doesn't (in general) work because it can't
2938    calculate the length of the lookbehind. At least, that's the case for Perl
2939    5.8.0 - I've been told they are going to document that it doesn't work in
2940    future.
2941    
2942    18. Added an error diagnosis for escapes that PCRE does not support: these are
2943    \L, \l, \N, \P, \p, \U, \u, and \X.
2944    
2945    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2946    reading past the end of the pattern in cases such as /[abcd/.
2947    
2948    20. PCRE was getting more memory than necessary for patterns with classes that
2949    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2950    
2951    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2952    compiling PCRE for use with Virtual Pascal.
2953    
2954    22. Small fix to the Makefile to make it work properly if the build is done
2955    outside the source tree.
2956    
2957    23. Added a new extension: a condition to go with recursion. If a conditional
2958    subpattern starts with (?(R) the "true" branch is used if recursion has
2959    happened, whereas the "false" branch is used only at the top level.
2960    
2961    24. When there was a very long string of literal characters (over 255 bytes
2962    without UTF support, over 250 bytes with UTF support), the computation of how
2963    much memory was required could be incorrect, leading to segfaults or other
2964    strange effects.
2965    
2966    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2967    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2968    there was a subsequent back reference to those brackets. This meant that, for
2969    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2970    possible to check for precisely this case. All we can do is abandon the
2971    optimization if .* occurs inside capturing brackets when there are any back
2972    references whatsoever. (See below for a better fix that came later.)
2973    
2974    26. The handling of the optimization for finding the first character of a
2975    non-anchored pattern, and for finding a character that is required later in the
2976    match were failing in some cases. This didn't break the matching; it just
2977    failed to optimize when it could. The way this is done has been re-implemented.
2978    
2979    27. Fixed typo in error message for invalid (?R item (it said "(?p").
2980    
2981    28. Added a new feature that provides some of the functionality that Perl
2982    provides with (?{...}). The facility is termed a "callout". The way it is done
2983    in PCRE is for the caller to provide an optional function, by setting
2984    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2985    global variable. By default it is unset, which disables all calling out. To get
2986    the function called, the regex must include (?C) at appropriate points. This
2987    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2988    This provides a means of identifying different callout points. When PCRE
2989    reaches such a point in the regex, if pcre_callout has been set, the external
2990    function is called. It is provided with data in a structure called
2991    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2992    matching continues; if it returns a non-zero value, the match at the current
2993    point fails. However, backtracking will occur if possible. [This was changed
2994    later and other features added - see item 49 below.]
2995    
2996    29. pcretest is upgraded to test the callout functionality. It provides a
2997    callout function that displays information. By default, it shows the start of
2998    the match and the current position in the text. There are some new data escapes
2999    to vary what happens:
3000    
3001        \C+         in addition, show current contents of captured substrings
3002        \C-         do not supply a callout function
3003        \C!n        return 1 when callout number n is reached
3004        \C!n!m      return 1 when callout number n is reached for the mth time
3005    
3006    30. If pcregrep was called with the -l option and just a single file name, it
3007    output "<stdin>" if a match was found, instead of the file name.
3008    
3009    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3010    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3011    pcre_exec(). This saves a malloc/free per call. The default value of
3012    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3013    when configuring.
3014    
3015    32. The default maximum size of a compiled pattern is 64K. There have been a
3016    few cases of people hitting this limit. The code now uses macros to handle the
3017    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3018    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3019    configuring. Tests 2 and 5 work only with 2-byte links because they output
3020    debugging information about compiled patterns.
3021    
3022    33. Internal code re-arrangements:
3023    
3024    (a) Moved the debugging function for printing out a compiled regex into
3025        its own source file (printint.c) and used #include to pull it into
3026        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3027        separate copies.
3028    
3029    (b) Defined the list of op-code names for debugging as a macro in
3030        internal.h so that it is next to the definition of the opcodes.
3031    
3032    (c) Defined a table of op-code lengths for simpler skipping along compiled
3033        code. This is again a macro in internal.h so that it is next to the
3034        definition of the opcodes.
3035    
3036    34. Added support for recursive calls to individual subpatterns, along the
3037    lines of Robin Houston's patch (but implemented somewhat differently).
3038    
3039    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3040    allow it to read and process whole directories in Win32. This code was
3041    contributed by Lionel Fourquaux; it has not been tested by me.
3042    
3043    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3044    used to name a group. Names consist of alphanumerics and underscores, and must
3045    be unique. Back references use the syntax (?P=name) and recursive calls use
3046    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3047    numbers. The function pcre_fullinfo() can be used after compilation to extract
3048    a name/number map. There are three relevant calls:
3049    
3050      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3051      PCRE_INFO_NAMECOUNT            yields the number of entries
3052      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3053    
3054    The map is a vector of fixed-size entries. The size of each entry depends on
3055    the length of the longest name used. The first two bytes of each entry are the
3056    group number, most significant byte first. There follows the corresponding
3057    name, zero terminated. The names are in alphabetical order.
3058    
3059    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3060    case instead of 255. Making it the same both with and without UTF-8 support
3061    means that the same test output works with both.
3062    
3063    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3064    calling malloc() with a zero argument.
3065    
3066    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3067    optimization. I've improved things by keeping a bitmap of backreferences with
3068    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3069    fact referenced, the optimization can be applied. It is unlikely that a
3070    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3071    the match to follow \n) will appear inside brackets with a number greater than
3072    31, but if it does, any back reference > 31 suppresses the optimization.
3073    
3074    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3075    of disabling numbered capturing parentheses. Any opening parenthesis that is
3076    not followed by ? behaves as if it were followed by ?: but named parentheses
3077    can still be used for capturing (and they will acquire numbers in the usual
3078    way).
3079    
3080    41. Redesigned the return codes from the match() function into yes/no/error so
3081    that errors can be passed back from deep inside the nested calls. A malloc
3082    failure while inside a recursive subpattern call now causes the
3083    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3084    
3085    42. It is now possible to set a limit on the number of times the match()
3086    function is called in a call to pcre_exec(). This facility makes it possible to
3087    limit the amount of recursion and backtracking, though not in a directly
3088    obvious way, because the match() function is used in a number of different
3089    circumstances. The count starts from zero for each position in the subject
3090    string (for non-anchored patterns). The default limit is, for compatibility, a
3091    large number, namely 10 000 000. You can change this in two ways:
3092    
3093    (a) When configuring PCRE before making, you can use --with-match-limit=n
3094        to set a default value for the compiled library.
3095    
3096    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3097        a different value is set. See 45 below.
3098    
3099    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3100    
3101    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3102    of things that can be changed at compile time. The first argument specifies
3103    what is wanted and the second points to where the information is to be placed.
3104    The current list of available information is:
3105    
3106      PCRE_CONFIG_UTF8
3107    
3108    The output is an integer that is set to one if UTF-8 support is available;
3109    otherwise it is set to zero.
3110    
3111      PCRE_CONFIG_NEWLINE
3112    
3113    The output is an integer that it set to the value of the code that is used for
3114    newline. It is either LF (10) or CR (13).
3115    
3116      PCRE_CONFIG_LINK_SIZE
3117    
3118    The output is an integer that contains the number of bytes used for internal
3119    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3120    
3121      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3122    
3123    The output is an integer that contains the threshold above which the POSIX
3124    interface uses malloc() for output vectors. See item 31 above.
3125    
3126      PCRE_CONFIG_MATCH_LIMIT
3127    
3128    The output is an unsigned integer that contains the default limit of the number
3129    of match() calls in a pcre_exec() execution. See 42 above.
3130    
3131    44. pcretest has been upgraded by the addition of the -C option. This causes it
3132    to extract all the available output from the new pcre_config() function, and to
3133    output it. The program then exits immediately.
3134    
3135    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3136    order to support additional features. One way would have been to define
3137    pcre_exec2() (for example) with extra arguments, but this would not have been
3138    extensible, and would also have required all calls to the original function to
3139    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3140    is used for passing in "extra" data from pcre_study().
3141    
3142    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3143    contains the following fields:
3144    
3145      flags         a bitmap indicating which of the following fields are set
3146      study_data    opaque data from pcre_study()
3147      match_limit   a way of specifying a limit on match() calls for a specific
3148                      call to pcre_exec()
3149      callout_data  data for callouts (see 49 below)
3150    
3151    The flag bits are also defined in pcre.h, and are
3152    
3153      PCRE_EXTRA_STUDY_DATA
3154      PCRE_EXTRA_MATCH_LIMIT
3155      PCRE_EXTRA_CALLOUT_DATA
3156    
3157    The pcre_study() function now returns one of these new pcre_extra blocks, with
3158    the actual study data pointed to by the study_data field, and the
3159    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3160    before. That is, this change is entirely upwards-compatible and requires no
3161    change to existing code.
3162    
3163    If you want to pass in additional data to pcre_exec(), you can either place it
3164    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3165    block.
3166    
3167    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3168    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3169    times with different match limits, until it finds the minimum value needed for
3170    pcre_exec() to complete. The value is then output. This can be instructive; for
3171    most simple matches the number is quite small, but for pathological cases it
3172    gets very large very quickly.
3173    
3174    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3175    returns the size of the data block pointed to by the study_data field in a
3176    pcre_extra block, that is, the value that was passed as the argument to
3177    pcre_malloc() when PCRE was getting memory in which to place the information
3178    created by pcre_study(). The fourth argument should point to a size_t variable.
3179    pcretest has been extended so that this information is shown after a successful
3180    pcre_study() call when information about the compiled regex is being displayed.
3181    
3182    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3183    because what follows is always an absolute path. (Later: it turns out that this
3184    is more than cosmetic for MinGW, because it doesn't like empty path
3185    components.)
3186    
3187    49. Some changes have been made to the callout feature (see 28 above):
3188    
3189    (i)  A callout function now has three choices for what it returns:
3190    
3191           0  =>  success, carry on matching
3192         > 0  =>  failure at this point, but backtrack if possible
3193         < 0  =>  serious error, return this value from pcre_exec()
3194    
3195         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3196         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3197         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3198         use by callout functions. It will never be used by PCRE itself.
3199    
3200    (ii) The pcre_extra structure (see 45 above) has a void * field called
3201         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3202         pcre_callout_block structure has a field of the same name. The contents of
3203         the field passed in the pcre_extra structure are passed to the callout
3204         function in the corresponding field in the callout block. This makes it
3205         easier to use the same callout-containing regex from multiple threads. For
3206         testing, the pcretest program has a new data escape
3207    
3208           \C*n        pass the number n (may be negative) as callout_data
3209    
3210         If the callout function in pcretest receives a non-zero value as
3211         callout_data, it returns that value.
3212    
3213    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3214    there were some redundant $(CFLAGS) in commands that are now specified as
3215    $(LINK), which already includes $(CFLAGS).
3216    
3217    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3218    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3219    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3220    one-byte characters throughout. Note that case-insensitive matching applies
3221    only to characters whose values are less than 256. PCRE doesn't support the
3222    notion of cases for higher-valued characters.
3223    
3224    (i)   A character class whose characters are all within 0-255 is handled as
3225          a bit map, and the map is inverted for negative classes. Previously, a
3226          character > 255 always failed to match such a class; however it should
3227          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3228    
3229    (ii)  A negated character class with a single character < 255 is coded as
3230          "not this character" (OP_NOT). This wasn't working properly when the test
3231          character was multibyte, either singly or repeated.
3232    
3233    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3234          mode, for example: \x{100}{2,3}.
3235    
3236    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3237          singly or repeated) now correctly test multibyte characters. However,
3238          PCRE doesn't recognize any characters with values greater than 255 as
3239          digits, spaces, or word characters. Such characters always match \D, \S,
3240          and \W, and never match \d, \s, or \w.
3241    
3242    (v)   Classes may now contain characters and character ranges with values
3243          greater than 255. For example: [ab\x{100}-\x{400}].
3244    
3245    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3246          PCRE in UTF-8 mode.
3247    
3248    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3249    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3250    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3251    value.)
3252    
3253    53. The single man page has become too large. I have therefore split it up into
3254    a number of separate man pages. These also give rise to individual HTML pages;
3255    these are now put in a separate directory, and there is an index.html page that
3256    lists them all. Some hyperlinking between the pages has been installed.
3257    
3258    54. Added convenience functions for handling named capturing parentheses.
3259    
3260    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3261    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3262    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3263    are faulted.
3264    
3265    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3266    calling configure. These values are used when compiling the dftables.c program
3267    which is run to generate the source of the default character tables. They
3268    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3269    you will need to set these values.
3270    
3271    57. Updated the building process for Windows DLL, as provided by Fred Cox.
3272    
3273    
3274    Version 3.9 02-Jan-02
3275    ---------------------
3276    
3277    1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3278    
3279    2. If --disable-static was given, the building process failed when trying to
3280    build pcretest and pcregrep. (For some reason it was using libtool to compile
3281    them, which is not right, as they aren't part of the library.)
3282    
3283    
3284    Version 3.8 18-Dec-01
3285    ---------------------
3286    
3287    1. The experimental UTF-8 code was completely screwed up. It was packing the
3288    bytes in the wrong order. How dumb can you get?
3289    
3290    
3291    Version 3.7 29-Oct-01
3292    ---------------------
3293    
3294    1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3295    This caused pcretest, when used on the test data, to segfault. Unfortunately,
3296    this didn't happen under Solaris 8, where I normally test things.
3297    
3298    2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3299    doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3300    isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3301    this fix an hour or so after the initial 3.7 release.)
3302    
3303    
3304    Version 3.6 23-Oct-01
3305    ---------------------
3306    
3307    1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3308    offsets passed as NULL with zero offset count.
3309    
3310    2. The config.guess and config.sub files had not been updated when I moved to
3311    the latest autoconf.
3312    
3313    
3314    Version 3.5 15-Aug-01
3315    ---------------------
3316    
3317    1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3318    had been forgotten.
3319    
3320    2. By using declared but undefined structures, we can avoid using "void"
3321    definitions in pcre.h while keeping the internal definitions of the structures
3322    private.
3323    
3324    3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3325    user point of view, this means that both static and shared libraries are built
3326    by default, but this can be individually controlled. More of the work of
3327    handling this static/shared cases is now inside libtool instead of PCRE's make
3328    file.
3329    
3330    4. The pcretest utility is now installed along with pcregrep because it is
3331    useful for users (to test regexs) and by doing this, it automatically gets
3332    relinked by libtool. The documentation has been turned into a man page, so
3333    there are now .1, .txt, and .html versions in /doc.
3334    
3335    5. Upgrades to pcregrep:
3336       (i)   Added long-form option names like gnu grep.
3337       (ii)  Added --help to list all options with an explanatory phrase.
3338       (iii) Added -r, --recursive to recurse into sub-directories.
3339       (iv)  Added -f, --file to read patterns from a file.
3340    
3341    6. pcre_exec() was referring to its "code" argument before testing that
3342    argument for NULL (and giving an error if it was NULL).
3343    
3344    7. Upgraded Makefile.in to allow for compiling in a different directory from
3345    the source directory.
3346    
3347    8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3348    options bits, the pointer it was passed was to an int instead of to an unsigned
3349    long int. This mattered only on 64-bit systems.
3350    
3351    9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3352    generated) instead of pcre.in, which it its source. Also made the same change
3353    in several of the .c files.
3354    
3355    10. A new release of gcc defines printf() as a macro, which broke pcretest
3356    because it had an ifdef in the middle of a string argument for printf(). Fixed
3357    by using separate calls to printf().
3358    
3359    11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3360    script, to force use of CR or LF instead of \n in the source. On non-Unix
3361    systems, the value can be set in config.h.
3362    
3363    12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3364    absolute limit. Changed the text of the error message to make this clear, and
3365    likewise updated the man page.
3366    
3367    13. The limit of 99 on the number of capturing subpatterns has been removed.
3368    The new limit is 65535, which I hope will not be a "real" limit.
3369    
3370    
3371    Version 3.4 22-Aug-00
3372    ---------------------
3373    
3374    1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3375    
3376    2. Diagnose condition (?(0) as an error instead of crashing on matching.
3377    
3378    
3379    Version 3.3 01-Aug-00
3380    ---------------------
3381    
3382    1. If an octal character was given, but the value was greater than \377, it
3383    was not getting masked to the least significant bits, as documented. This could
3384    lead to crashes in some systems.
3385    
3386    2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3387    the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3388    
3389    3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3390    These just pass their arguments on to (pcre_free)(), but they are provided
3391    because some uses of PCRE bind it to non-C systems that can call its functions,
3392    but cannot call free() or pcre_free() directly.
3393    
3394    4. Add "make test" as a synonym for "make check". Corrected some comments in
3395    the Makefile.
3396    
3397    5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3398    Makefile.
3399    
3400    6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3401    command called pgrep for grepping around the active processes.
3402    
3403    7. Added the beginnings of support for UTF-8 character strings.
3404    
3405    8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3406    RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3407    the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3408    out for the ar command.)
3409    
3410    
3411    Version 3.2 12-May-00
3412    ---------------------
3413    
3414    This is purely a bug fixing release.
3415    
3416    1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3417    of ZA. This was just one example of several cases that could provoke this bug,
3418    which was introduced by change 9 of version 2.00. The code for breaking
3419    infinite loops after an iteration that matches an empty string was't working
3420    correctly.
3421    
3422    2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3423    when matched against abbab (for example). After matching an empty string, it
3424    wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3425    caused it to match further down the string than it should.
3426    
3427    3. The code contained an inclusion of sys/types.h. It isn't clear why this
3428    was there because it doesn't seem to be needed, and it causes trouble on some
3429    systems, as it is not a Standard C header. It has been removed.
3430    
3431    4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3432    were reported on the Macintosh. The changes were from
3433    
3434      while ((c = *(++ptr)) != 0 && c != '\n');
3435    to
3436      while ((c = *(++ptr)) != 0 && c != '\n') ;
3437    
3438    Totally extraordinary, but if that's what it takes...
3439    
3440    5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3441    available. Added HAVE_BCOPY and an autoconf test for it; if neither
3442    HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3443    assumes the way PCRE uses memmove() (always moving upwards).
3444    
3445    6. PCRE is being used in one environment where strchr() is not available. There
3446    was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3447    faster code anyway.
3448    
3449    
3450    Version 3.1 09-Feb-00
3451    ---------------------
3452    
3453    The only change in this release is the fixing of some bugs in Makefile.in for
3454    the "install" target:
3455    
3456    (1) It was failing to install pcreposix.h.
3457    
3458    (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3459    
3460    
3461    Version 3.0 01-Feb-00
3462    ---------------------
3463    
3464    1. Add support for the /+ modifier to perltest (to output $` like it does in
3465    pcretest).
3466    
3467    2. Add support for the /g modifier to perltest.
3468    
3469    3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3470    matches null strings.
3471    
3472    4. Fix perltest so that it doesn't do unwanted things when fed an empty
3473    pattern. Perl treats empty patterns specially - it reuses the most recent
3474    pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3475    effect.
3476    
3477    5. The POSIX interface was broken in that it was just handing over the POSIX
3478    captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3479    required a bigger vector, with some working space on the end. This means that
3480    the POSIX wrapper now has to get and free some memory, and copy the results.
3481    
3482    6. Added some simple autoconf support, placing the test data and the
3483    documentation in separate directories, re-organizing some of the
3484    information files, and making it build pcre-config (a GNU standard). Also added
3485    libtool support for building PCRE as a shared library, which is now the
3486    default.
3487    
3488    7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3489    09 are not valid octal constants. Single digits will be used for minor values
3490    less than 10.
3491    
3492    8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3493    existing programs that set these in the POSIX interface can use PCRE without
3494    modification.
3495    
3496    9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3497    return all that pcre_info() returns, plus additional data. The pcre_info()
3498    function is retained for compatibility, but is considered to be obsolete.
3499    
3500    10. Added experimental recursion feature (?R) to handle one common case that
3501    Perl 5.6 will be able to do with (?p{...}).
3502    
3503    11. Added support for POSIX character classes like [:alpha:], which Perl is
3504    adopting.
3505    
3506    
3507    Version 2.08 31-Aug-99
3508    ----------------------
3509    
3510    1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3511    trying to match at the startoffset position, but instead was moving forward to
3512    the next newline as if a previous match had failed.
3513    
3514    2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3515    and could get into a loop if a null string was matched other than at the start
3516    of the subject.
3517    
3518    3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3519    be distinguished at compile time, and for completeness also added PCRE_DATE.
3520    
3521    5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3522    in GnuWin32 environments.
3523    
3524    
3525    Version 2.07 29-Jul-99
3526    ----------------------
3527    
3528    1. The documentation is now supplied in plain text form and HTML as well as in
3529    the form of man page sources.
3530    
3531    2. C++ compilers don't like assigning (void *) values to other pointer types.
3532    In particular this affects malloc(). Although there is no problem in Standard
3533    C, I've put in casts to keep C++ compilers happy.
3534    
3535    3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3536    should be (const char *).
3537    
3538    4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3539    be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3540    However, I haven't made this a standard facility. The documentation doesn't
3541    mention it, and the Makefile doesn't support it.
3542    
3543    5. The Makefile now contains an "install" target, with editable destinations at
3544    the top of the file. The pcretest program is not installed.
3545    
3546    6. pgrep -V now gives the PCRE version number and date.
3547    
3548    7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3549    causing the entire string to be ignored, instead of just the last character.
3550    
3551    8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3552    non-matching string, it can take a very, very long time, even for strings of
3553    quite modest length, because of the nested recursion. PCRE now does better in
3554    some of these cases. It does this by remembering the last required literal
3555    character in the pattern, and pre-searching the subject to ensure it is present
3556    before running the real match. In other words, it applies a heuristic to detect
3557    some types of certain failure quickly, and in the above example, if presented
3558    with a string that has no trailing " it gives "no match" very quickly.
3559    
3560    9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3561    other alternatives are tried instead.
3562    
3563    
3564    Version 2.06 09-Jun-99
3565    ----------------------
3566    
3567    1. Change pcretest's output for amount of store used to show just the code
3568    space, because the remainder (the data block) varies in size between 32-bit and
3569    64-bit systems.
3570    
3571    2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3572    start matching at. This allows lookbehinds to work when searching for multiple
3573    occurrences in a string.
3574    
3575    3. Added additional options to pcretest for testing multiple occurrences:
3576    
3577       /+   outputs the rest of the string that follows a match
3578       /g   loops for multiple occurrences, using the new startoffset argument
3579       /G   loops for multiple occurrences by passing an incremented pointer
3580    
3581    4. PCRE wasn't doing the "first character" optimization for patterns starting
3582    with \b or \B, though it was doing it for other lookbehind assertions. That is,
3583    it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3584    the letter 'x'. On long subject strings, this gives a significant speed-up.
3585    
3586    
3587    Version 2.05 21-Apr-99
3588    ----------------------
3589    
3590    1. Changed the type of magic_number from int to long int so that it works
3591    properly on 16-bit systems.
3592    
3593    2. Fixed a bug which caused patterns starting with .* not to work correctly
3594    when the subject string contained newline characters. PCRE was assuming
3595    anchoring for such patterns in all cases, which is not correct because .* will
3596    not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3597    DOTALL is set at top level; otherwise it knows that patterns starting with .*
3598    must be retried after every newline in the subject.
3599    
3600    
3601    Version 2.04 18-Feb-99
3602    ----------------------
3603    
3604    1. For parenthesized subpatterns with repeats whose minimum was zero, the
3605    computation of the store needed to hold the pattern was incorrect (too large).
3606    If such patterns were nested a few deep, this could multiply and become a real
3607    problem.
3608    
3609    2. Added /M option to pcretest to show the memory requirement of a specific
3610    pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3611    
3612    3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3613    compiled in such a way that the backtracking after subsequent failure was
3614    pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3615    ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3616    
3617    
3618    Version 2.03 02-Feb-99
3619    ----------------------
3620    
3621    1. Fixed typo and small mistake in man page.
3622    
3623