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

Diff of /code/trunk/ChangeLog

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

revision 7 by nigel, Sat Feb 24 21:38:09 2007 UTC revision 561 by ph10, Sat Oct 30 18:37:47 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    
63    Version 8.10 25-Jun-2010
64    ------------------------
65    
66    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
67        THEN.
68    
69    2.  (*ACCEPT) was not working when inside an atomic group.
70    
71    3.  Inside a character class, \B is treated as a literal by default, but
72        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
73        causes the error). The code is unchanged, but I tidied the documentation.
74    
75    4.  Inside a character class, PCRE always treated \R and \X as literals,
76        whereas Perl faults them if its -w option is set. I have changed PCRE so
77        that it faults them when PCRE_EXTRA is set.
78    
79    5.  Added support for \N, which always matches any character other than
80        newline. (It is the same as "." when PCRE_DOTALL is not set.)
81    
82    6.  When compiling pcregrep with newer versions of gcc which may have
83        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
84        declared with attribute warn_unused_result" were given. Just casting the
85        result to (void) does not stop the warnings; a more elaborate fudge is
86        needed. I've used a macro to implement this.
87    
88    7.  Minor change to pcretest.c to avoid a compiler warning.
89    
90    8.  Added four artifical Unicode properties to help with an option to make
91        \s etc use properties (see next item). The new properties are: Xan
92        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
93    
94    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
95        use Unicode properties. (*UCP) at the start of a pattern can be used to set
96        this option. Modified pcretest to add /W to test this facility. Added
97        REG_UCP to make it available via the POSIX interface.
98    
99    10. Added --line-buffered to pcregrep.
100    
101    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
102        studied, and the match started with a letter with a code point greater than
103        127 whose first byte was different to the first byte of the other case of
104        the letter, the other case of this starting letter was not recognized
105        (#976).
106    
107    12. If a pattern that was studied started with a repeated Unicode property
108        test, for example, \p{Nd}+, there was the theoretical possibility of
109        setting up an incorrect bitmap of starting bytes, but fortunately it could
110        not have actually happened in practice until change 8 above was made (it
111        added property types that matched character-matching opcodes).
112    
113    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
114        possible starting bytes for non-anchored patterns.
115    
116    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
117        \R, and also a number of cases that involve Unicode properties, both
118        explicit and implicit when PCRE_UCP is set.
119    
120    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
121        input, it could crash or give wrong results if characters with values
122        greater than 0xc0 were present in the subject string. (Detail: it assumed
123        UTF-8 input when processing these items.)
124    
125    16. Added a lot of (int) casts to avoid compiler warnings in systems where
126        size_t is 64-bit (#991).
127    
128    17. Added a check for running out of memory when PCRE is compiled with
129        --disable-stack-for-recursion (#990).
130    
131    18. If the last data line in a file for pcretest does not have a newline on
132        the end, a newline was missing in the output.
133    
134    19. The default pcre_chartables.c file recognizes only ASCII characters (values
135        less than 128) in its various bitmaps. However, there is a facility for
136        generating tables according to the current locale when PCRE is compiled. It
137        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
138        characters, are recognized by isspace() and therefore were getting set in
139        these tables, and indeed these tables seem to approximate to ISO 8859. This
140        caused a problem in UTF-8 mode when pcre_study() was used to create a list
141        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
142        which of course cannot start UTF-8 characters. I have changed the code so
143        that only real ASCII characters (less than 128) and the correct starting
144        bytes for UTF-8 encodings are set for characters greater than 127 when in
145        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
146        altogether.)
147    
148    20. Added the /T option to pcretest so as to be able to run tests with non-
149        standard character tables, thus making it possible to include the tests
150        used for 19 above in the standard set of tests.
151    
152    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
153        reference to a subpattern the other side of a comment that contains an
154        opening parenthesis caused either an internal compiling error, or a
155        reference to the wrong subpattern.
156    
157    
158    Version 8.02 19-Mar-2010
159    ------------------------
160    
161    1.  The Unicode data tables have been updated to Unicode 5.2.0.
162    
163    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
164        configured.
165    
166    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
167        original author of that file, following a query about its status.
168    
169    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
170        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
171    
172    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
173        quantifier applied to a forward-referencing subroutine call, could compile
174        incorrect code or give the error "internal error: previously-checked
175        referenced subpattern not found".
176    
177    6.  Both MS Visual Studio and Symbian OS have problems with initializing
178        variables to point to external functions. For these systems, therefore,
179        pcre_malloc etc. are now initialized to local functions that call the
180        relevant global functions.
181    
182    7.  There were two entries missing in the vectors called coptable and poptable
183        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
184        I've fixed the data, and added a kludgy way of testing at compile time that
185        the lengths are correct (equal to the number of opcodes).
186    
187    8.  Following on from 7, I added a similar kludge to check the length of the
188        eint vector in pcreposix.c.
189    
190    9.  Error texts for pcre_compile() are held as one long string to avoid too
191        much relocation at load time. To find a text, the string is searched,
192        counting zeros. There was no check for running off the end of the string,
193        which could happen if a new error number was added without updating the
194        string.
195    
196    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
197    
198    11. \K was not working if it appeared in an atomic group or in a group that
199        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
200        \K is "not well defined" if used in an assertion. PCRE now accepts it if
201        the assertion is positive, but not if it is negative.
202    
203    12. Change 11 fortuitously reduced the size of the stack frame used in the
204        "match()" function of pcre_exec.c by one pointer. Forthcoming
205        implementation of support for (*MARK) will need an extra pointer on the
206        stack; I have reserved it now, so that the stack frame size does not
207        decrease.
208    
209    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
210        item in branch that calls a recursion is a subroutine call - as in the
211        second branch in the above example - was incorrectly given the compile-
212        time error "recursive call could loop indefinitely" because pcre_compile()
213        was not correctly checking the subroutine for matching a non-empty string.
214    
215    14. The checks for overrunning compiling workspace could trigger after an
216        overrun had occurred. This is a "should never occur" error, but it can be
217        triggered by pathological patterns such as hundreds of nested parentheses.
218        The checks now trigger 100 bytes before the end of the workspace.
219    
220    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
221    
222    
223    Version 8.01 19-Jan-2010
224    ------------------------
225    
226    1.  If a pattern contained a conditional subpattern with only one branch (in
227        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
228        computed the wrong minimum data length (which is of course zero for such
229        subpatterns). This could cause incorrect "no match" results.
230    
231    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
232        the pattern is reset in the first branch, pcre_compile() failed with
233        "internal error: code overflow at offset...". This happened only when
234        the reset was to the original external option setting. (An optimization
235        abstracts leading options settings into an external setting, which was the
236        cause of this.)
237    
238    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
239        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
240        assertion pattern did not match (meaning that the assertion was true), it
241        was incorrectly treated as false if the SKIP had been reached during the
242        matching. This also applied to assertions used as conditions.
243    
244    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
245        assertion subpattern, including such a pattern used as a condition,
246        unpredictable results occurred, instead of the error return
247        PCRE_ERROR_DFA_UITEM.
248    
249    5.  The C++ GlobalReplace function was not working like Perl for the special
250        situation when an empty string is matched. It now does the fancy magic
251        stuff that is necessary.
252    
253    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
254        removed. (These were left over from very, very early versions of PCRE.)
255    
256    7.  Some cosmetic changes to the code to make life easier when compiling it
257        as part of something else:
258    
259        (a) Change DEBUG to PCRE_DEBUG.
260    
261        (b) In pcre_compile(), rename the member of the "branch_chain" structure
262            called "current" as "current_branch", to prevent a collision with the
263            Linux macro when compiled as a kernel module.
264    
265        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
266            prevent a collision with the Linux macro when compiled as a kernel
267            module.
268    
269    8.  In pcre_compile() there are some checks for integer overflows that used to
270        cast potentially large values to (double). This has been changed to that
271        when building, a check for int64_t is made, and if it is found, it is used
272        instead, thus avoiding the use of floating point arithmetic. (There is no
273        other use of FP in PCRE.) If int64_t is not found, the fallback is to
274        double.
275    
276    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
277        2005 (difference between two addresses compared to an unsigned value).
278    
279    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
280        custom one, because of the following reported problem in Windows:
281    
282          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
283              under Win32.
284          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
285              therefore missing the function definition.
286          - The compiler thus generates a "C" signature for the test function.
287          - The linker fails to find the "C" function.
288          - PCRE fails to configure if asked to do so against libbz2.
289    
290    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
291        messages were output:
292    
293          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
294          rerunning libtoolize, to keep the correct libtool macros in-tree.
295          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
296    
297        I have done both of these things.
298    
299    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
300        most of the time, it *can* run out if it is given a pattern that contains a
301        runaway infinite recursion. I updated the discussion in the pcrestack man
302        page.
303    
304    13. Now that we have gone to the x.xx style of version numbers, the minor
305        version may start with zero. Using 08 or 09 is a bad idea because users
306        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
307        interpreted as invalid octal numbers. I've updated the previous comment in
308        configure.ac, and also added a check that gives an error if 08 or 09 are
309        used.
310    
311    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
312        causing partial matching to fail when the end of the subject matched \W
313        in a UTF-8 pattern where \W was quantified with a minimum of 3.
314    
315    15. There were some discrepancies between the declarations in pcre_internal.h
316        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
317        their definitions. The declarations used "const uschar *" and the
318        definitions used USPTR. Even though USPTR is normally defined as "const
319        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
320        reported that: "This difference in casting confuses some C++ compilers, for
321        example, SunCC recognizes above declarations as different functions and
322        generates broken code for hbpcre." I have changed the declarations to use
323        USPTR.
324    
325    16. GNU libtool is named differently on some systems. The autogen.sh script now
326        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
327        (FreeBSD).
328    
329    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
330        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
331        comment: "Figure out how to create a longlong from a string: strtoll and
332        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
333        instance, but it only takes 2 args instead of 3!"
334    
335    18. A subtle bug concerned with back references has been fixed by a change of
336        specification, with a corresponding code fix. A pattern such as
337        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
338        refers, was giving matches when it shouldn't. For example, xa=xaaa would
339        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
340        same bug. Such groups have to be quantified to be useful, or contained
341        inside another quantified group. (If there's no repetition, the reference
342        can never match.) The problem arises because, having left the group and
343        moved on to the rest of the pattern, a later failure that backtracks into
344        the group uses the captured value from the final iteration of the group
345        rather than the correct earlier one. I have fixed this in PCRE by forcing
346        any group that contains a reference to itself to be an atomic group; that
347        is, there cannot be any backtracking into it once it has completed. This is
348        similar to recursive and subroutine calls.
349    
350    
351    Version 8.00 19-Oct-09
352    ----------------------
353    
354    1.  The table for translating pcre_compile() error codes into POSIX error codes
355        was out-of-date, and there was no check on the pcre_compile() error code
356        being within the table. This could lead to an OK return being given in
357        error.
358    
359    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
360        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
361        in a Windows environment.
362    
363    3.  The pcregrep --count option prints the count for each file even when it is
364        zero, as does GNU grep. However, pcregrep was also printing all files when
365        --files-with-matches was added. Now, when both options are given, it prints
366        counts only for those files that have at least one match. (GNU grep just
367        prints the file name in this circumstance, but including the count seems
368        more useful - otherwise, why use --count?) Also ensured that the
369        combination -clh just lists non-zero counts, with no names.
370    
371    4.  The long form of the pcregrep -F option was incorrectly implemented as
372        --fixed_strings instead of --fixed-strings. This is an incompatible change,
373        but it seems right to fix it, and I didn't think it was worth preserving
374        the old behaviour.
375    
376    5.  The command line items --regex=pattern and --regexp=pattern were not
377        recognized by pcregrep, which required --regex pattern or --regexp pattern
378        (with a space rather than an '='). The man page documented the '=' forms,
379        which are compatible with GNU grep; these now work.
380    
381    6.  No libpcreposix.pc file was created for pkg-config; there was just
382        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
383    
384    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
385        when UCP support is not needed, by modifying the Python script that
386        generates it from Unicode data files. This should not matter if the module
387        is correctly used as a library, but I received one complaint about 50K of
388        unwanted data. My guess is that the person linked everything into his
389        program rather than using a library. Anyway, it does no harm.
390    
391    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
392        was a minimum greater than 1 for a wide character in a possessive
393        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
394        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
395        character. Chaos in the form of incorrect output or a compiling loop could
396        result.
397    
398    9.  The restrictions on what a pattern can contain when partial matching is
399        requested for pcre_exec() have been removed. All patterns can now be
400        partially matched by this function. In addition, if there are at least two
401        slots in the offset vector, the offset of the earliest inspected character
402        for the match and the offset of the end of the subject are set in them when
403        PCRE_ERROR_PARTIAL is returned.
404    
405    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
406        synonymous with PCRE_PARTIAL, for backwards compatibility, and
407        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
408        and may be more useful for multi-segment matching.
409    
410    11. Partial matching with pcre_exec() is now more intuitive. A partial match
411        used to be given if ever the end of the subject was reached; now it is
412        given only if matching could not proceed because another character was
413        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
414        string "Z", which now yields "no match" instead of "partial match". In the
415        case of pcre_dfa_exec(), "no match" is given if every matching path for the
416        final character ended with (*FAIL).
417    
418    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
419        if the pattern had a "must contain" character that was already found in the
420        earlier partial match, unless partial matching was again requested. For
421        example, with the pattern /dog.(body)?/, the "must contain" character is
422        "g". If the first part-match was for the string "dog", restarting with
423        "sbody" failed. This bug has been fixed.
424    
425    13. The string returned by pcre_dfa_exec() after a partial match has been
426        changed so that it starts at the first inspected character rather than the
427        first character of the match. This makes a difference only if the pattern
428        starts with a lookbehind assertion or \b or \B (\K is not supported by
429        pcre_dfa_exec()). It's an incompatible change, but it makes the two
430        matching functions compatible, and I think it's the right thing to do.
431    
432    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
433        so that the demonstration program is easily available in environments where
434        PCRE has not been installed from source.
435    
436    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
437        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
438        library.
439    
440    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
441        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
442        is not the first non-POSIX option to be added. Clearly some people find
443        these options useful.
444    
445    17. If a caller to the POSIX matching function regexec() passes a non-zero
446        value for nmatch with a NULL value for pmatch, the value of
447        nmatch is forced to zero.
448    
449    18. RunGrepTest did not have a test for the availability of the -u option of
450        the diff command, as RunTest does. It now checks in the same way as
451        RunTest, and also checks for the -b option.
452    
453    19. If an odd number of negated classes containing just a single character
454        interposed, within parentheses, between a forward reference to a named
455        subpattern and the definition of the subpattern, compilation crashed with
456        an internal error, complaining that it could not find the referenced
457        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
458        [The bug was that it was starting one character too far in when skipping
459        over the character class, thus treating the ] as data rather than
460        terminating the class. This meant it could skip too much.]
461    
462    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
463        /g option in pcretest when the pattern contains \K, which makes it possible
464        to have an empty string match not at the start, even when the pattern is
465        anchored. Updated pcretest and pcredemo to use this option.
466    
467    21. If the maximum number of capturing subpatterns in a recursion was greater
468        than the maximum at the outer level, the higher number was returned, but
469        with unset values at the outer level. The correct (outer level) value is
470        now given.
471    
472    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
473        PCRE did not set those parentheses (unlike Perl). I have now found a way to
474        make it do so. The string so far is captured, making this feature
475        compatible with Perl.
476    
477    23. The tests have been re-organized, adding tests 11 and 12, to make it
478        possible to check the Perl 5.10 features against Perl 5.10.
479    
480    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
481        pattern matches a fixed length string. PCRE did not allow this; now it
482        does. Neither allows recursion.
483    
484    25. I finally figured out how to implement a request to provide the minimum
485        length of subject string that was needed in order to match a given pattern.
486        (It was back references and recursion that I had previously got hung up
487        on.) This code has now been added to pcre_study(); it finds a lower bound
488        to the length of subject needed. It is not necessarily the greatest lower
489        bound, but using it to avoid searching strings that are too short does give
490        some useful speed-ups. The value is available to calling programs via
491        pcre_fullinfo().
492    
493    26. While implementing 25, I discovered to my embarrassment that pcretest had
494        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
495        study optimizations had never been tested with that matching function.
496        Oops. What is worse, even when it was passed study data, there was a bug in
497        pcre_dfa_exec() that meant it never actually used it. Double oops. There
498        were also very few tests of studied patterns with pcre_dfa_exec().
499    
500    27. If (?| is used to create subpatterns with duplicate numbers, they are now
501        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
502        on the other side of the coin, they are no longer allowed to have different
503        names, because these cannot be distinguished in PCRE, and this has caused
504        confusion. (This is a difference from Perl.)
505    
506    28. When duplicate subpattern names are present (necessarily with different
507        numbers, as required by 27 above), and a test is made by name in a
508        conditional pattern, either for a subpattern having been matched, or for
509        recursion in such a pattern, all the associated numbered subpatterns are
510        tested, and the overall condition is true if the condition is true for any
511        one of them. This is the way Perl works, and is also more like the way
512        testing by number works.
513    
514    
515    Version 7.9 11-Apr-09
516    ---------------------
517    
518    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
519        (pcretest), all targets were linked against these libraries. This included
520        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
521        libraries. This caused unwanted dependencies to be created. This problem
522        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
523        pcretest is linked with readline.
524    
525    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
526        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
527        moved outside it again, because FALSE and TRUE are already defined in AIX,
528        but BOOL is not.
529    
530    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
531        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
532    
533    4.  The pcregrep documentation said spaces were inserted as well as colons (or
534        hyphens) following file names and line numbers when outputting matching
535        lines. This is not true; no spaces are inserted. I have also clarified the
536        wording for the --colour (or --color) option.
537    
538    5.  In pcregrep, when --colour was used with -o, the list of matching strings
539        was not coloured; this is different to GNU grep, so I have changed it to be
540        the same.
541    
542    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
543        each matching line was coloured. Now it goes on to look for further matches
544        of any of the test patterns, which is the same behaviour as GNU grep.
545    
546    7.  A pattern that could match an empty string could cause pcregrep to loop; it
547        doesn't make sense to accept an empty string match in pcregrep, so I have
548        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
549        seems to be how GNU grep behaves.
550    
551    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
552        start or after a newline", because the conditional assertion was not being
553        correctly handled. The rule now is that both the assertion and what follows
554        in the first alternative must satisfy the test.
555    
556    9.  If auto-callout was enabled in a pattern with a conditional group whose
557        condition was an assertion, PCRE could crash during matching, both with
558        pcre_exec() and pcre_dfa_exec().
559    
560    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
561        used for matching.
562    
563    11. Unicode property support in character classes was not working for
564        characters (bytes) greater than 127 when not in UTF-8 mode.
565    
566    12. Added the -M command line option to pcretest.
567    
568    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
569    
570    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
571    
572    16. Added comments and documentation about mis-use of no_arg in the C++
573        wrapper.
574    
575    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
576        from Martin Jerabek that uses macro names for all relevant character and
577        string constants.
578    
579    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
580        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
581        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
582        these, but not everybody uses configure.
583    
584    19. A conditional group that had only one branch was not being correctly
585        recognized as an item that could match an empty string. This meant that an
586        enclosing group might also not be so recognized, causing infinite looping
587        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
588        with the subject "ab", where knowledge that the repeated group can match
589        nothing is needed in order to break the loop.
590    
591    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
592        exec(), but without supplying a callout function, matching went wrong.
593    
594    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
595        leak if the size of the offset vector was greater than 30. When the vector
596        is smaller, the saved offsets during recursion go onto a local stack
597        vector, but for larger vectors malloc() is used. It was failing to free
598        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
599        error, in fact).
600    
601    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
602        heapframe that is used only when UTF-8 support is enabled. This caused no
603        problem, but was untidy.
604    
605    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
606        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
607        included within another project.
608    
609    24. Steven Van Ingelgem's patches to add more options to the CMake support,
610        slightly modified by me:
611    
612          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
613              not building pcregrep.
614    
615          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
616              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
617    
618    25. Forward references, both numeric and by name, in patterns that made use of
619        duplicate group numbers, could behave incorrectly or give incorrect errors,
620        because when scanning forward to find the reference group, PCRE was not
621        taking into account the duplicate group numbers. A pattern such as
622        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
623    
624    26. Changed a few more instances of "const unsigned char *" to USPTR, making
625        the feature of a custom pointer more persuasive (as requested by a user).
626    
627    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
628        pcretest.c, inside #ifndefs, because it seems they are sometimes already
629        pre-defined.
630    
631    28. Added support for (*UTF8) at the start of a pattern.
632    
633    29. Arrange for flags added by the "release type" setting in CMake to be shown
634        in the configuration summary.
635    
636    
637    Version 7.8 05-Sep-08
638    ---------------------
639    
640    1.  Replaced UCP searching code with optimized version as implemented for Ad
641        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
642        stage table and inline lookup instead of a function, giving speed ups of 2
643        to 5 times on some simple patterns that I tested. Permission was given to
644        distribute the MultiStage2.py script that generates the tables (it's not in
645        the tarball, but is in the Subversion repository).
646    
647    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
648        scripts.
649    
650    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
651        a group with a zero qualifier. The result of the study could be incorrect,
652        or the function might crash, depending on the pattern.
653    
654    4.  Caseless matching was not working for non-ASCII characters in back
655        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
656        It now works when Unicode Property Support is available.
657    
658    5.  In pcretest, an escape such as \x{de} in the data was always generating
659        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
660        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
661        truncation.
662    
663    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
664    
665    7.  Added two (int) casts to pcregrep when printing the difference of two
666        pointers, in case they are 64-bit values.
667    
668    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
669        test 2 if it fails.
670    
671    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
672        and a #define of that name to empty if it is not externally set. This is to
673        allow users of MSVC to set it if necessary.
674    
675    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
676        the convenience functions in the pcre_get.c source file.
677    
678    11. An option change at the start of a pattern that had top-level alternatives
679        could cause overwriting and/or a crash. This command provoked a crash in
680        some environments:
681    
682          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
683    
684        This potential security problem was recorded as CVE-2008-2371.
685    
686    12. For a pattern where the match had to start at the beginning or immediately
687        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
688        pcre_dfa_exec() could read past the end of the passed subject if there was
689        no match. To help with detecting such bugs (e.g. with valgrind), I modified
690        pcretest so that it places the subject at the end of its malloc-ed buffer.
691    
692    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
693        exec() might read past the end of the data buffer in UTF-8 mode.
694    
695    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
696        the data contained the byte 0x85 as part of a UTF-8 character within its
697        first line. This applied both to normal and DFA matching.
698    
699    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
700        /^[^d]*?$/8 failed to match "abc".
701    
702    16. Added a missing copyright notice to pcrecpp_internal.h.
703    
704    17. Make it more clear in the documentation that values returned from
705        pcre_exec() in ovector are byte offsets, not character counts.
706    
707    18. Tidied a few places to stop certain compilers from issuing warnings.
708    
709    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
710        supplied by Stefan Weber. I made a further small update for 7.8 because
711        there is a change of source arrangements: the pcre_searchfuncs.c module is
712        replaced by pcre_ucd.c.
713    
714    
715    Version 7.7 07-May-08
716    ---------------------
717    
718    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
719        a string to a long long, pretend we don't even have a long long." This is
720        done by checking for the strtoq, strtoll, and _strtoi64 functions.
721    
722    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
723        pre-7.6 versions, which defined a global no_arg variable instead of putting
724        it in the RE class. (See also #8 below.)
725    
726    3.  Remove a line of dead code, identified by coverity and reported by Nuno
727        Lopes.
728    
729    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
730    
731        (1) The include/exclude patterns were being applied to the whole pathnames
732            of files, instead of just to the final components.
733    
734        (2) If there was more than one level of directory, the subdirectories were
735            skipped unless they satisfied the include/exclude conditions. This is
736            inconsistent with GNU grep (and could even be seen as contrary to the
737            pcregrep specification - which I improved to make it absolutely clear).
738            The action now is always to scan all levels of directory, and just
739            apply the include/exclude patterns to regular files.
740    
741    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
742        --exclude_dir in the tests to avoid scanning .svn directories.
743    
744    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
745        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
746        doesn't support NULs in patterns.
747    
748    7.  Added some missing "const"s to declarations of static tables in
749        pcre_compile.c and pcre_dfa_exec.c.
750    
751    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
752        caused by fix #2  above. (Subsequently also a second patch to fix the
753        first patch. And a third patch - this was a messy problem.)
754    
755    9.  Applied Craig's patch to remove the use of push_back().
756    
757    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
758        matching function regexec().
759    
760    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
761        which, however, unlike Perl's \g{...}, are subroutine calls, not back
762        references. PCRE supports relative numbers with this syntax (I don't think
763        Oniguruma does).
764    
765    12. Previously, a group with a zero repeat such as (...){0} was completely
766        omitted from the compiled regex. However, this means that if the group
767        was called as a subroutine from elsewhere in the pattern, things went wrong
768        (an internal error was given). Such groups are now left in the compiled
769        pattern, with a new opcode that causes them to be skipped at execution
770        time.
771    
772    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
773        to the way PCRE behaves:
774    
775        (a) A lone ] character is dis-allowed (Perl treats it as data).
776    
777        (b) A back reference to an unmatched subpattern matches an empty string
778            (Perl fails the current match path).
779    
780        (c) A data ] in a character class must be notated as \] because if the
781            first data character in a class is ], it defines an empty class. (In
782            Perl it is not possible to have an empty class.) The empty class []
783            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
784            The negative empty class [^] matches any one character, independently
785            of the DOTALL setting.
786    
787    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
788        non-existent subpattern following a character class starting with ']' and
789        containing () gave an internal compiling error instead of "reference to
790        non-existent subpattern". Fortunately, when the pattern did exist, the
791        compiled code was correct. (When scanning forwards to check for the
792        existencd of the subpattern, it was treating the data ']' as terminating
793        the class, so got the count wrong. When actually compiling, the reference
794        was subsequently set up correctly.)
795    
796    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
797        it was being rejected as not supported by pcre_dfa_exec(), even though
798        other assertions are supported. I have made pcre_dfa_exec() support
799        (*FAIL).
800    
801    16. The implementation of 13c above involved the invention of a new opcode,
802        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
803        cannot be changed at match time, I realized I could make a small
804        improvement to matching performance by compiling OP_ALLANY instead of
805        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
806        on the OP_ANY path.
807    
808    17. Compiling pcretest on Windows with readline support failed without the
809        following two fixes: (1) Make the unistd.h include conditional on
810        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
811    
812    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
813        ncurses library to be included for pcretest when ReadLine support is
814        requested, but also to allow for it to be overridden. This patch came from
815        Daniel Bergström.
816    
817    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
818        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
819        any errors with the current Unicode tables. Thanks to Peter Kankowski for
820        spotting this.
821    
822    
823    Version 7.6 28-Jan-08
824    ---------------------
825    
826    1.  A character class containing a very large number of characters with
827        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
828        overflow.
829    
830    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
831        HAVE_LONG_LONG is not defined.
832    
833    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
834        bring them up to date and include new features. This patch includes:
835    
836        - Fixed PH's badly added libz and libbz2 support.
837        - Fixed a problem with static linking.
838        - Added pcredemo. [But later removed - see 7 below.]
839        - Fixed dftables problem and added an option.
840        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
841            HAVE_LONG_LONG.
842        - Added readline support for pcretest.
843        - Added an listing of the option settings after cmake has run.
844    
845    4.  A user submitted a patch to Makefile that makes it easy to create
846        "pcre.dll" under mingw when using Configure/Make. I added stuff to
847        Makefile.am that cause it to include this special target, without
848        affecting anything else. Note that the same mingw target plus all
849        the other distribution libraries and programs are now supported
850        when configuring with CMake (see 6 below) instead of with
851        Configure/Make.
852    
853    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
854        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
855        exported in the Windows port". It has not yet been confirmed that the patch
856        solves the problem, but it does no harm.
857    
858    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
859        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
860        with CMake, and also correct the comment about stack recursion.
861    
862    7.  Remove the automatic building of pcredemo from the ./configure system and
863        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
864        of a program that users should build themselves after PCRE is installed, so
865        building it automatically is not really right. What is more, it gave
866        trouble in some build environments.
867    
868    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
869    
870    
871    Version 7.5 10-Jan-08
872    ---------------------
873    
874    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
875        values in parens when parsing an RE using the C++ wrapper."
876    
877    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
878        Characters greater than 255 were excluded from the class instead of being
879        included.
880    
881    3.  The same bug as (2) above applied to negated POSIX classes such as
882        [:^space:].
883    
884    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
885        defined or documented. It seems to have been a typo for PCRE_STATIC, so
886        I have changed it.
887    
888    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
889        first named subpattern) and a construct such as (?&a) would reference the
890        first named subpattern whose name started with "a" (in other words, the
891        length check was missing). Both these problems are fixed. "Subpattern name
892        expected" is now given for (?&) (a zero-length name), and this patch also
893        makes it give the same error for \k'' (previously it complained that that
894        was a reference to a non-existent subpattern).
895    
896    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
897        this is right because (?- can be followed by option settings as well as by
898        digits. I have, however, made the messages clearer.
899    
900    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
901        than the number used in the conditional) now cause a compile-time error.
902        This is actually not compatible with Perl, which accepts such patterns, but
903        treats the conditional as always being FALSE (as PCRE used to), but it
904        seems to me that giving a diagnostic is better.
905    
906    8.  Change "alphameric" to the more common word "alphanumeric" in comments
907        and messages.
908    
909    9.  Fix two occurrences of "backslash" in comments that should have been
910        "backspace".
911    
912    10. Remove two redundant lines of code that can never be obeyed (their function
913        was moved elsewhere).
914    
915    11. The program that makes PCRE's Unicode character property table had a bug
916        which caused it to generate incorrect table entries for sequences of
917        characters that have the same character type, but are in different scripts.
918        It amalgamated them into a single range, with the script of the first of
919        them. In other words, some characters were in the wrong script. There were
920        thirteen such cases, affecting characters in the following ranges:
921    
922          U+002b0 - U+002c1
923          U+0060c - U+0060d
924          U+0061e - U+00612
925          U+0064b - U+0065e
926          U+0074d - U+0076d
927          U+01800 - U+01805
928          U+01d00 - U+01d77
929          U+01d9b - U+01dbf
930          U+0200b - U+0200f
931          U+030fc - U+030fe
932          U+03260 - U+0327f
933          U+0fb46 - U+0fbb1
934          U+10450 - U+1049d
935    
936    12. The -o option (show only the matching part of a line) for pcregrep was not
937        compatible with GNU grep in that, if there was more than one match in a
938        line, it showed only the first of them. It now behaves in the same way as
939        GNU grep.
940    
941    13. If the -o and -v options were combined for pcregrep, it printed a blank
942        line for every non-matching line. GNU grep prints nothing, and pcregrep now
943        does the same. The return code can be used to tell if there were any
944        non-matching lines.
945    
946    14. Added --file-offsets and --line-offsets to pcregrep.
947    
948    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
949        infinitely looping recursion. The bug was that positive lookaheads were not
950        being skipped when checking for a possible empty match (negative lookaheads
951        and both kinds of lookbehind were skipped).
952    
953    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
954        inclusion of <windows.h> to before rather than after the definition of
955        INVALID_FILE_ATTRIBUTES (patch from David Byron).
956    
957    17. Specifying a possessive quantifier with a specific limit for a Unicode
958        character property caused pcre_compile() to compile bad code, which led at
959        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
960        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
961        caused the error; without that there was no problem.
962    
963    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
964    
965    19. Added --enable-pcretest-libreadline.
966    
967    20. In pcrecpp.cc, the variable 'count' was incremented twice in
968        RE::GlobalReplace(). As a result, the number of replacements returned was
969        double what it should be. I removed one of the increments, but Craig sent a
970        later patch that removed the other one (the right fix) and added unit tests
971        that check the return values (which was not done before).
972    
973    21. Several CMake things:
974    
975        (1) Arranged that, when cmake is used on Unix, the libraries end up with
976            the names libpcre and libpcreposix, not just pcre and pcreposix.
977    
978        (2) The above change means that pcretest and pcregrep are now correctly
979            linked with the newly-built libraries, not previously installed ones.
980    
981        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
982    
983    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
984        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
985        UTF-8 newline character). The key issue is that the pattern starts .*;
986        this means that the match must be either at the beginning, or after a
987        newline. The bug was in the code for advancing after a failed match and
988        checking that the new position followed a newline. It was not taking
989        account of UTF-8 characters correctly.
990    
991    23. PCRE was behaving differently from Perl in the way it recognized POSIX
992        character classes. PCRE was not treating the sequence [:...:] as a
993        character class unless the ... were all letters. Perl, however, seems to
994        allow any characters between [: and :], though of course it rejects as
995        unknown any "names" that contain non-letters, because all the known class
996        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
997        for example, whereas PCRE did not - it did not recognize a POSIX character
998        class. This seemed a bit dangerous, so the code has been changed to be
999        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1000        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1001        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1002        Perl does, and where it didn't before.
1003    
1004    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1005        Windows environments %n is disabled by default.
1006    
1007    
1008    Version 7.4 21-Sep-07
1009    ---------------------
1010    
1011    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1012        means that a class such as [\s] counted as "explicit reference to CR or
1013        LF". That isn't really right - the whole point of the change was to try to
1014        help when there was an actual mention of one of the two characters. So now
1015        the change happens only if \r or \n (or a literal CR or LF) character is
1016        encountered.
1017    
1018    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1019        of both had grown to the point where there were only 3 bits left.
1020        Fortunately, there was spare space in the data structure, and so I have
1021        moved the internal flags into a new 16-bit field to free up more option
1022        bits.
1023    
1024    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1025        but did not set the internal JCHANGED flag - either of these is enough to
1026        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1027        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1028        start sets both bits.
1029    
1030    4.  Added options (at build time, compile time, exec time) to change \R from
1031        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1032    
1033    5.  doc/pcresyntax.html was missing from the distribution.
1034    
1035    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1036        compatibility, even though it is no longer used.
1037    
1038    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1039        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1040        windows.h file is present (where different names are used). [This was
1041        reversed later after testing - see 16 below.]
1042    
1043    8.  Changed all #include <config.h> to #include "config.h". There were also
1044        some further <pcre.h> cases that I changed to "pcre.h".
1045    
1046    9.  When pcregrep was used with the --colour option, it missed the line ending
1047        sequence off the lines that it output.
1048    
1049    10. It was pointed out to me that arrays of string pointers cause lots of
1050        relocations when a shared library is dynamically loaded. A technique of
1051        using a single long string with a table of offsets can drastically reduce
1052        these. I have refactored PCRE in four places to do this. The result is
1053        dramatic:
1054    
1055          Originally:                          290
1056          After changing UCP table:            187
1057          After changing error message table:   43
1058          After changing table of "verbs"       36
1059          After changing table of Posix names   22
1060    
1061        Thanks to the folks working on Gregex for glib for this insight.
1062    
1063    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1064        unicode-properties was also set.
1065    
1066    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1067    
1068    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1069        checked only for CRLF.
1070    
1071    14. Added casts to pcretest.c to avoid compiler warnings.
1072    
1073    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1074    
1075    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1076        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1077        entirely. This removes changes made in 7 above.
1078    
1079    17. The CMake files have been updated, and there is now more information about
1080        building with CMake in the NON-UNIX-USE document.
1081    
1082    
1083    Version 7.3 28-Aug-07
1084    ---------------------
1085    
1086     1. In the rejigging of the build system that eventually resulted in 7.1, the
1087        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1088        brackets there is not right, since it causes compilers to look for an
1089        installed pcre.h, not the version that is in the source that is being
1090        compiled (which of course may be different). I have changed it back to:
1091    
1092          #include "pcre.h"
1093    
1094        I have a vague recollection that the change was concerned with compiling in
1095        different directories, but in the new build system, that is taken care of
1096        by the VPATH setting the Makefile.
1097    
1098     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1099        when the subject happened to end in the byte 0x85 (e.g. if the last
1100        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1101        characters but of course it shouldn't be taken as a newline when it is part
1102        of another character. The bug was that, for an unlimited repeat of . in
1103        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1104        characters when looking for a newline.
1105    
1106     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1107    
1108     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1109        in debug output.
1110    
1111     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1112        long printing in the pcrecpp unittest when running under MinGW.
1113    
1114     6. ESC_K was left out of the EBCDIC table.
1115    
1116     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1117        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1118        limit also applies to "virtual nesting" when a pattern is recursive, and in
1119        this case 1000 isn't so big. I have been able to remove this limit at the
1120        expense of backing off one optimization in certain circumstances. Normally,
1121        when pcre_exec() would call its internal match() function recursively and
1122        immediately return the result unconditionally, it uses a "tail recursion"
1123        feature to save stack. However, when a subpattern that can match an empty
1124        string has an unlimited repetition quantifier, it no longer makes this
1125        optimization. That gives it a stack frame in which to save the data for
1126        checking that an empty string has been matched. Previously this was taken
1127        from the 1000-entry workspace that had been reserved. So now there is no
1128        explicit limit, but more stack is used.
1129    
1130     8. Applied Daniel's patches to solve problems with the import/export magic
1131        syntax that is required for Windows, and which was going wrong for the
1132        pcreposix and pcrecpp parts of the library. These were overlooked when this
1133        problem was solved for the main library.
1134    
1135     9. There were some crude static tests to avoid integer overflow when computing
1136        the size of patterns that contain repeated groups with explicit upper
1137        limits. As the maximum quantifier is 65535, the maximum group length was
1138        set at 30,000 so that the product of these two numbers did not overflow a
1139        32-bit integer. However, it turns out that people want to use groups that
1140        are longer than 30,000 bytes (though not repeat them that many times).
1141        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1142        made it possible to implement the integer overflow checks in a much more
1143        dynamic way, which I have now done. The artificial limitation on group
1144        length has been removed - we now have only the limit on the total length of
1145        the compiled pattern, which depends on the LINK_SIZE setting.
1146    
1147    10. Fixed a bug in the documentation for get/copy named substring when
1148        duplicate names are permitted. If none of the named substrings are set, the
1149        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1150        empty string.
1151    
1152    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1153        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1154        because the ] is interpreted as the first data character and the
1155        terminating ] is not found. PCRE has been made compatible with Perl in this
1156        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1157        cause memory overwriting.
1158    
1159    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1160        string has been matched (to stop an infinite loop). It was not recognizing
1161        a conditional subpattern that could match an empty string if that
1162        subpattern was within another subpattern. For example, it looped when
1163        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1164        condition was not nested. This bug has been fixed.
1165    
1166    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1167        past the start of the subject in the presence of bytes with the top bit
1168        set, for example "\x8aBCD".
1169    
1170    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1171        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1172    
1173    14. Optimized (?!) to (*FAIL).
1174    
1175    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1176        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1177        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1178        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1179        does: it's just the validity check that is more restrictive.
1180    
1181    16. Inserted checks for integer overflows during escape sequence (backslash)
1182        processing, and also fixed erroneous offset values for syntax errors during
1183        backslash processing.
1184    
1185    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1186        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1187    
1188    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1189        caused an overrun.
1190    
1191    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1192        something other than just ASCII characters) inside a group that had an
1193        unlimited repeat caused a loop at compile time (while checking to see
1194        whether the group could match an empty string).
1195    
1196    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1197        [\P{Any}] did so. (Error in the code for printing property names.)
1198    
1199    21. An orphan \E inside a character class could cause a crash.
1200    
1201    22. A repeated capturing bracket such as (A)? could cause a wild memory
1202        reference during compilation.
1203    
1204    23. There are several functions in pcre_compile() that scan along a compiled
1205        expression for various reasons (e.g. to see if it's fixed length for look
1206        behind). There were bugs in these functions when a repeated \p or \P was
1207        present in the pattern. These operators have additional parameters compared
1208        with \d, etc, and these were not being taken into account when moving along
1209        the compiled data. Specifically:
1210    
1211        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1212            length.
1213    
1214        (b) An item such as \pL+ within a repeated group could cause crashes or
1215            loops.
1216    
1217        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1218            "reference to non-existent subpattern" error.
1219    
1220        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1221    
1222    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1223        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1224    
1225    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1226    
1227    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1228        character were causing crashes (broken optimization).
1229    
1230    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1231        \p or \P) caused a compile-time loop.
1232    
1233    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1234        break. For example, the unstudied pattern [\r\n]A does not match the string
1235        "\r\nA" because change 7.0/46 below moves the current point on by two
1236        characters after failing to match at the start. However, the pattern \nA
1237        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1238        the same is true. There doesn't seem any very clean way out of this, but
1239        what I have chosen to do makes the common cases work: PCRE now takes note
1240        of whether there can be an explicit match for \r or \n anywhere in the
1241        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1242        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1243        pattern has explicit CR or LF references.
1244    
1245    29. Added (*CR) etc for changing newline setting at start of pattern.
1246    
1247    
1248    Version 7.2 19-Jun-07
1249    ---------------------
1250    
1251     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1252        which is apparently normally available under Windows.
1253    
1254     2. Re-jig the pcregrep tests with different newline settings in an attempt
1255        to make them independent of the local environment's newline setting.
1256    
1257     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1258    
1259     4. Some of the "internals" tests were previously cut out when the link size
1260        was not 2, because the output contained actual offsets. The recent new
1261        "Z" feature of pcretest means that these can be cut out, making the tests
1262        usable with all link sizes.
1263    
1264     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1265        stack recursion. This gives a massive performance boost under BSD, but just
1266        a small improvement under Linux. However, it saves one field in the frame
1267        in all cases.
1268    
1269     6. Added more features from the forthcoming Perl 5.10:
1270    
1271        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1272            recursion call. It refers to the nth most recently opened parentheses.
1273    
1274        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1275            to be opened parentheses.
1276    
1277        (c) Conditions that refer to capturing parentheses can be specified
1278            relatively, for example, (?(-2)... or (?(+3)...
1279    
1280        (d) \K resets the start of the current match so that everything before
1281            is not part of it.
1282    
1283        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1284    
1285        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1286            reference syntax.
1287    
1288        (g) (?| introduces a group in which the numbering of parentheses in each
1289            alternative starts with the same number.
1290    
1291        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1292    
1293     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1294        PCRE_INFO_JCHANGED.
1295    
1296     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1297        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1298        for detecting groups that can match an empty string.
1299    
1300     9. A pattern with a very large number of alternatives (more than several
1301        hundred) was running out of internal workspace during the pre-compile
1302        phase, where pcre_compile() figures out how much memory will be needed. A
1303        bit of new cunning has reduced the workspace needed for groups with
1304        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1305        workspace instead of running out of the 4096 that are available.
1306    
1307    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1308    
1309    11. Applied patch from Google to remove an optimization that didn't quite work.
1310        The report of the bug said:
1311    
1312          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1313          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1314          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1315    
1316    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1317        it matched the wrong number of bytes.
1318    
1319    
1320    Version 7.1 24-Apr-07
1321    ---------------------
1322    
1323     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1324        that is more "standard", making use of automake and other Autotools. There
1325        is some re-arrangement of the files and adjustment of comments consequent
1326        on this.
1327    
1328     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1329        for recursive directory scanning broke on some systems because the files
1330        are not scanned in any specific order and on different systems the order
1331        was different. A call to "sort" has been inserted into RunGrepTest for the
1332        approprate test as a short-term fix. In the longer term there may be an
1333        alternative.
1334    
1335     3. I had an email from Eric Raymond about problems translating some of PCRE's
1336        man pages to HTML (despite the fact that I distribute HTML pages, some
1337        people do their own conversions for various reasons). The problems
1338        concerned the use of low-level troff macros .br and .in. I have therefore
1339        removed all such uses from the man pages (some were redundant, some could
1340        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1341        HTML has been updated to handle .nf/.fi and to complain if it encounters
1342        .br or .in.
1343    
1344     4. Updated comments in configure.ac that get placed in config.h.in and also
1345        arranged for config.h to be included in the distribution, with the name
1346        config.h.generic, for the benefit of those who have to compile without
1347        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1348    
1349     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1350        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1351        makevp.bat for the current PCRE, using the additional files
1352        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1353    
1354     6. A Windows user reported a minor discrepancy with test 2, which turned out
1355        to be caused by a trailing space on an input line that had got lost in his
1356        copy. The trailing space was an accident, so I've just removed it.
1357    
1358     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1359        that is needed.
1360    
1361     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1362        as "const" (a) because they are and (b) because it helps the PHP
1363        maintainers who have recently made a script to detect big data structures
1364        in the php code that should be moved to the .rodata section. I remembered
1365        to update Builducptable as well, so it won't revert if ucptable.h is ever
1366        re-created.
1367    
1368     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1369        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1370        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1371        support is not required. This saves 1.5-2K of code, which is important in
1372        some applications.
1373    
1374        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1375        so as not to refer to the tables, even though these functions will never be
1376        called when UTF-8 support is disabled. Otherwise there are problems with a
1377        shared library.
1378    
1379    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1380    
1381        (a) It was defining its arguments as char * instead of void *.
1382    
1383        (b) It was assuming that all moves were upwards in memory; this was true
1384            a long time ago when I wrote it, but is no longer the case.
1385    
1386        The emulated memove() is provided for those environments that have neither
1387        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1388        is clearly not the case, as these two bugs were recently reported.
1389    
1390    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1391        and Detrail to create the HTML documentation, the .txt form of the man
1392        pages, and it removes trailing spaces from listed files. It also creates
1393        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1394        case, it wraps all the #defines with #ifndefs. This script should be run
1395        before "make dist".
1396    
1397    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1398        with Unicode property support.
1399    
1400        (a) For a maximizing quantifier, if the two different cases of the
1401            character were of different lengths in their UTF-8 codings (there are
1402            some cases like this - I found 11), and the matching function had to
1403            back up over a mixture of the two cases, it incorrectly assumed they
1404            were both the same length.
1405    
1406        (b) When PCRE was configured to use the heap rather than the stack for
1407            recursion during matching, it was not correctly preserving the data for
1408            the other case of a UTF-8 character when checking ahead for a match
1409            while processing a minimizing repeat. If the check also involved
1410            matching a wide character, but failed, corruption could cause an
1411            erroneous result when trying to check for a repeat of the original
1412            character.
1413    
1414    13. Some tidying changes to the testing mechanism:
1415    
1416        (a) The RunTest script now detects the internal link size and whether there
1417            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1418            values substituted by "configure". (The RunGrepTest script already did
1419            this for UTF-8.) The configure.ac script no longer substitutes the
1420            relevant variables.
1421    
1422        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1423            with length and offset values. This means that the output is different
1424            for different internal link sizes. Test 2 is skipped for link sizes
1425            other than 2 because of this, bypassing the problem. Unfortunately,
1426            there was also a test in test 3 (the locale tests) that used /B and
1427            failed for link sizes other than 2. Rather than cut the whole test out,
1428            I have added a new /Z option to pcretest that replaces the length and
1429            offset values with spaces. This is now used to make test 3 independent
1430            of link size. (Test 2 will be tidied up later.)
1431    
1432    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1433        segmentation fault instead of returning the appropriate error message.
1434    
1435    15. In multiline mode when the newline sequence was set to "any", the pattern
1436        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1437        This doesn't seem right; it now treats the CRLF combination as the line
1438        ending, and so does not match in that case. It's only a pattern such as ^$
1439        that would hit this one: something like ^ABC$ would have failed after \r
1440        and then tried again after \r\n.
1441    
1442    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1443        in an attempt to make files that differ only in their line terminators
1444        compare equal. This works on Linux.
1445    
1446    17. Under certain error circumstances pcregrep might try to free random memory
1447        as it exited. This is now fixed, thanks to valgrind.
1448    
1449    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1450        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1451        was because its rules for how to advance for /g after matching an empty
1452        string at the end of a line did not allow for this case. They now check for
1453        it specially.
1454    
1455    20. pcretest is supposed to handle patterns and data of any length, by
1456        extending its buffers when necessary. It was getting this wrong when the
1457        buffer for a data line had to be extended.
1458    
1459    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1460        CRLF as a newline sequence.
1461    
1462    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1463        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1464        I have nevertheless tidied it up.
1465    
1466    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1467    
1468    24. Added a man page for pcre-config.
1469    
1470    
1471    Version 7.0 19-Dec-06
1472    ---------------------
1473    
1474     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1475        moving to gcc 4.1.1.
1476    
1477     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1478        sys/time.h, which is documented as needed for this function. It doesn't
1479        seem to matter on Linux, but it showed up on some releases of OS X.
1480    
1481     3. It seems that there are systems where bytes whose values are greater than
1482        127 match isprint() in the "C" locale. The "C" locale should be the
1483        default when a C program starts up. In most systems, only ASCII printing
1484        characters match isprint(). This difference caused the output from pcretest
1485        to vary, making some of the tests fail. I have changed pcretest so that:
1486    
1487        (a) When it is outputting text in the compiled version of a pattern, bytes
1488            other than 32-126 are always shown as hex escapes.
1489    
1490        (b) When it is outputting text that is a matched part of a subject string,
1491            it does the same, unless a different locale has been set for the match
1492            (using the /L modifier). In this case, it uses isprint() to decide.
1493    
1494     4. Fixed a major bug that caused incorrect computation of the amount of memory
1495        required for a compiled pattern when options that changed within the
1496        pattern affected the logic of the preliminary scan that determines the
1497        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1498        that the computed length was too small. The symptoms of this bug were
1499        either the PCRE error "internal error: code overflow" from pcre_compile(),
1500        or a glibc crash with a message such as "pcretest: free(): invalid next
1501        size (fast)". Examples of patterns that provoked this bug (shown in
1502        pcretest format) are:
1503    
1504          /(?-x: )/x
1505          /(?x)(?-x: \s*#\s*)/
1506          /((?i)[\x{c0}])/8
1507          /(?i:[\x{c0}])/8
1508    
1509        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1510        is now done differently.
1511    
1512     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1513        wrapper classes; (b) implement a new function in the C++ scanner that is
1514        more efficient than the old way of doing things because it avoids levels of
1515        recursion in the regex matching; (c) add a paragraph to the documentation
1516        for the FullMatch() function.
1517    
1518     6. The escape sequence \n was being treated as whatever was defined as
1519        "newline". Not only was this contrary to the documentation, which states
1520        that \n is character 10 (hex 0A), but it also went horribly wrong when
1521        "newline" was defined as CRLF. This has been fixed.
1522    
1523     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1524        was being set to -1 for the "end of line" case (supposedly a value that no
1525        character can have). Though this value is never used (the check for end of
1526        line is "zero bytes in current character"), it caused compiler complaints.
1527        I've changed it to 0xffffffff.
1528    
1529     8. In pcre_version.c, the version string was being built by a sequence of
1530        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1531        string (as it is for production releases) called a macro with an empty
1532        argument. The C standard says the result of this is undefined. The gcc
1533        compiler treats it as an empty string (which was what was wanted) but it is
1534        reported that Visual C gives an error. The source has been hacked around to
1535        avoid this problem.
1536    
1537     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1538        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1539        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1540        of them did).
1541    
1542    10. Originally, pcretest opened its input and output without "b"; then I was
1543        told that "b" was needed in some environments, so it was added for release
1544        5.0 to both the input and output. (It makes no difference on Unix-like
1545        systems.) Later I was told that it is wrong for the input on Windows. I've
1546        now abstracted the modes into two macros, to make it easier to fiddle with
1547        them, and removed "b" from the input mode under Windows.
1548    
1549    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1550    
1551    12. Added -help and --help to pcretest as an official way of being reminded
1552        of the options.
1553    
1554    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1555        and pcrecpp.cc because they annoy compilers at high warning levels.
1556    
1557    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1558    
1559    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1560        scripts are not C programs :-) and which was not noticed because it works
1561        on Linux.
1562    
1563    16. pcretest is supposed to handle any length of pattern and data line (as one
1564        line or as a continued sequence of lines) by extending its input buffer if
1565        necessary. This feature was broken for very long pattern lines, leading to
1566        a string of junk being passed to pcre_compile() if the pattern was longer
1567        than about 50K.
1568    
1569    17. I have done a major re-factoring of the way pcre_compile() computes the
1570        amount of memory needed for a compiled pattern. Previously, there was code
1571        that made a preliminary scan of the pattern in order to do this. That was
1572        OK when PCRE was new, but as the facilities have expanded, it has become
1573        harder and harder to keep it in step with the real compile phase, and there
1574        have been a number of bugs (see for example, 4 above). I have now found a
1575        cunning way of running the real compile function in a "fake" mode that
1576        enables it to compute how much memory it would need, while actually only
1577        ever using a few hundred bytes of working memory and without too many
1578        tests of the mode. This should make future maintenance and development
1579        easier. A side effect of this work is that the limit of 200 on the nesting
1580        depth of parentheses has been removed (though this was never a serious
1581        limitation, I suspect). However, there is a downside: pcre_compile() now
1582        runs more slowly than before (30% or more, depending on the pattern). I
1583        hope this isn't a big issue. There is no effect on runtime performance.
1584    
1585    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1586        newline (only possible for the last line of a file) and it was a
1587        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1588    
1589    19. Added additional timing features to pcretest. (1) The -tm option now times
1590        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1591        separate command line item, by a number that specifies the number of
1592        repeats to use when timing. The default is 50000; this gives better
1593        precision, but takes uncomfortably long for very large patterns.
1594    
1595    20. Extended pcre_study() to be more clever in cases where a branch of a
1596        subpattern has no definite first character. For example, (a*|b*)[cd] would
1597        previously give no result from pcre_study(). Now it recognizes that the
1598        first character must be a, b, c, or d.
1599    
1600    21. There was an incorrect error "recursive call could loop indefinitely" if
1601        a subpattern (or the entire pattern) that was being tested for matching an
1602        empty string contained only one non-empty item after a nested subpattern.
1603        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1604        incorrectly, because the \d was being skipped in the check.
1605    
1606    22. The pcretest program now has a new pattern option /B and a command line
1607        option -b, which is equivalent to adding /B to every pattern. This causes
1608        it to show the compiled bytecode, without the additional information that
1609        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1610        is the same as /B/I).
1611    
1612    23. A new optimization is now able automatically to treat some sequences such
1613        as a*b as a*+b. More specifically, if something simple (such as a character
1614        or a simple class like \d) has an unlimited quantifier, and is followed by
1615        something that cannot possibly match the quantified thing, the quantifier
1616        is automatically "possessified".
1617    
1618    24. A recursive reference to a subpattern whose number was greater than 39
1619        went wrong under certain circumstances in UTF-8 mode. This bug could also
1620        have affected the operation of pcre_study().
1621    
1622    25. Realized that a little bit of performance could be had by replacing
1623        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1624    
1625    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1626    
1627    27. Possessive quantifiers such as a++ were previously implemented by turning
1628        them into atomic groups such as ($>a+). Now they have their own opcodes,
1629        which improves performance. This includes the automatically created ones
1630        from 23 above.
1631    
1632    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1633        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1634        the first matched character to be a colon. This applied both to named and
1635        numbered groups.
1636    
1637    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1638    
1639    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1640        building PCRE on Windows easier, so I have included it in the distribution.
1641    
1642    31. There is now a check in pcretest against a ridiculously large number being
1643        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1644        loop, the loop is abandoned.
1645    
1646    32. Forward references to subpatterns in conditions such as (?(2)...) where
1647        subpattern 2 is defined later cause pcre_compile() to search forwards in
1648        the pattern for the relevant set of parentheses. This search went wrong
1649        when there were unescaped parentheses in a character class, parentheses
1650        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1651    
1652    33. "Subroutine" calls and backreferences were previously restricted to
1653        referencing subpatterns earlier in the regex. This restriction has now
1654        been removed.
1655    
1656    34. Added a number of extra features that are going to be in Perl 5.10. On the
1657        whole, these are just syntactic alternatives for features that PCRE had
1658        previously implemented using the Python syntax or my own invention. The
1659        other formats are all retained for compatibility.
1660    
1661        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1662            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1663            also .NET compatible.
1664    
1665        (b) A recursion or subroutine call to a named group can now be defined as
1666            (?&name) as well as (?P>name).
1667    
1668        (c) A backreference to a named group can now be defined as \k<name> or
1669            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1670            5.10, are also .NET compatible.
1671    
1672        (d) A conditional reference to a named group can now use the syntax
1673            (?(<name>) or (?('name') as well as (?(name).
1674    
1675        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1676            groups (named and numbered) that are never evaluated inline, but can be
1677            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1678            is always false. There may be only one alternative in such a group.
1679    
1680        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1681            as the simple (?(R). The condition is true only if the most recent
1682            recursion is that of the given number or name. It does not search out
1683            through the entire recursion stack.
1684    
1685        (g) The escape \gN or \g{N} has been added, where N is a positive or
1686            negative number, specifying an absolute or relative reference.
1687    
1688    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1689        some "unreachable code" warnings.
1690    
1691    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1692        things, this adds five new scripts.
1693    
1694    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1695        There were also incompatibilities regarding the handling of \Q..\E inside
1696        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1697        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1698    
1699    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1700        matches an empty string, and forcibly breaks the loop. There were bugs in
1701        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1702        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1703        separate and independent bugs (that affected different cases) have been
1704        fixed.
1705    
1706    39. Refactored the code to abolish the use of different opcodes for small
1707        capturing bracket numbers. This is a tidy that I avoided doing when I
1708        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1709        The new approach is not only tidier, it makes it possible to reduce the
1710        memory needed to fix the previous bug (38).
1711    
1712    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1713        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1714        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1715        mode.
1716    
1717    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1718        report.
1719    
1720    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1721        copy construction and assignment in the C++ wrapper.
1722    
1723    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1724        couple of bugs that could have given wrong results in the "--newline=crlf"
1725        case.
1726    
1727    44. Added a number of casts and did some reorganization of signed/unsigned int
1728        variables following suggestions from Dair Grant. Also renamed the variable
1729        "this" as "item" because it is a C++ keyword.
1730    
1731    45. Arranged for dftables to add
1732    
1733          #include "pcre_internal.h"
1734    
1735        to pcre_chartables.c because without it, gcc 4.x may remove the array
1736        definition from the final binary if PCRE is built into a static library and
1737        dead code stripping is activated.
1738    
1739    46. For an unanchored pattern, if a match attempt fails at the start of a
1740        newline sequence, and the newline setting is CRLF or ANY, and the next two
1741        characters are CRLF, advance by two characters instead of one.
1742    
1743    
1744    Version 6.7 04-Jul-06
1745    ---------------------
1746    
1747     1. In order to handle tests when input lines are enormously long, pcretest has
1748        been re-factored so that it automatically extends its buffers when
1749        necessary. The code is crude, but this _is_ just a test program. The
1750        default size has been increased from 32K to 50K.
1751    
1752     2. The code in pcre_study() was using the value of the re argument before
1753        testing it for NULL. (Of course, in any sensible call of the function, it
1754        won't be NULL.)
1755    
1756     3. The memmove() emulation function in pcre_internal.h, which is used on
1757        systems that lack both memmove() and bcopy() - that is, hardly ever -
1758        was missing a "static" storage class specifier.
1759    
1760     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1761        containing an extended class (one that cannot be represented by a bitmap
1762        because it contains high-valued characters or Unicode property items, e.g.
1763        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1764        pattern, but PCRE should not loop if you do not (it no longer does).
1765        [Detail: two cases were found: (a) a repeated subpattern containing an
1766        extended class; (b) a recursive reference to a subpattern that followed a
1767        previous extended class. It wasn't skipping over the extended class
1768        correctly when UTF-8 mode was not set.]
1769    
1770     5. A negated single-character class was not being recognized as fixed-length
1771        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1772        compile error "lookbehind assertion is not fixed length".
1773    
1774     6. The RunPerlTest auxiliary script was showing an unexpected difference
1775        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1776        write a Perl script that can interpret lines of an input file either as
1777        byte characters or as UTF-8, which is what "perltest" was being required to
1778        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1779        can't do is switch easily at run time between having the "use utf8;" pragma
1780        or not. In the end, I fudged it by using the RunPerlTest script to insert
1781        "use utf8;" explicitly for the UTF-8 tests.
1782    
1783     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1784        the end of the subject string, contrary to the documentation and to what
1785        Perl does. This was true of both matching functions. Now it matches only at
1786        the start of the subject and immediately after *internal* newlines.
1787    
1788     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1789        a pointer to an int instead of a pointer to an unsigned long int. This
1790        caused problems on 64-bit systems.
1791    
1792     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1793        instance of the 'standard' template library not being so standard".
1794    
1795    10. There was no check on the number of named subpatterns nor the maximum
1796        length of a subpattern name. The product of these values is used to compute
1797        the size of the memory block for a compiled pattern. By supplying a very
1798        long subpattern name and a large number of named subpatterns, the size
1799        computation could be caused to overflow. This is now prevented by limiting
1800        the length of names to 32 characters, and the number of named subpatterns
1801        to 10,000.
1802    
1803    11. Subpatterns that are repeated with specific counts have to be replicated in
1804        the compiled pattern. The size of memory for this was computed from the
1805        length of the subpattern and the repeat count. The latter is limited to
1806        65535, but there was no limit on the former, meaning that integer overflow
1807        could in principle occur. The compiled length of a repeated subpattern is
1808        now limited to 30,000 bytes in order to prevent this.
1809    
1810    12. Added the optional facility to have named substrings with the same name.
1811    
1812    13. Added the ability to use a named substring as a condition, using the
1813        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1814        are numbers (not recommended). Forward references are permitted.
1815    
1816    14. Added forward references in named backreferences (if you see what I mean).
1817    
1818    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1819        pattern could run off the end of the subject. For example, the pattern
1820        "(?s)(.{1,5})"8 did this with the subject "ab".
1821    
1822    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1823        PCRE_CASELESS was set when matching characters that were quantified with ?
1824        or *.
1825    
1826    17. A character class other than a single negated character that had a minimum
1827        but no maximum quantifier - for example [ab]{6,} - was not handled
1828        correctly by pce_dfa_exec(). It would match only one character.
1829    
1830    18. A valid (though odd) pattern that looked like a POSIX character
1831        class but used an invalid character after [ (for example [[,abc,]]) caused
1832        pcre_compile() to give the error "Failed: internal error: code overflow" or
1833        in some cases to crash with a glibc free() error. This could even happen if
1834        the pattern terminated after [[ but there just happened to be a sequence of
1835        letters, a binary zero, and a closing ] in the memory that followed.
1836    
1837    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1838        over the years. Originally (before any Unicode support), just the bottom 8
1839        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1840        output from "man perlunicode" includes this:
1841    
1842          The regular expression compiler produces polymorphic opcodes.  That
1843          is, the pattern adapts to the data and automatically switches to
1844          the Unicode character scheme when presented with Unicode data--or
1845          instead uses a traditional byte scheme when presented with byte
1846          data.
1847    
1848        Sadly, a wide octal escape does not cause a switch, and in a string with
1849        no other multibyte characters, these octal escapes are treated as before.
1850        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1851        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1852        Unicode string.
1853    
1854        I have not perpetrated such confusion in PCRE. Up till now, it took just
1855        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1856        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1857        translate to the appropriate multibyte character.
1858    
1859    29. Applied some refactoring to reduce the number of warnings from Microsoft
1860        and Borland compilers. This has included removing the fudge introduced
1861        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1862        a warning about an unused variable.
1863    
1864    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1865        characters since release 4.0, because Perl (from release 5.004) does not.
1866        [Or at least, is documented not to: some releases seem to be in conflict
1867        with the documentation.] However, when a pattern was studied with
1868        pcre_study() and all its branches started with \s, PCRE still included VT
1869        as a possible starting character. Of course, this did no harm; it just
1870        caused an unnecessary match attempt.
1871    
1872    22. Removed a now-redundant internal flag bit that recorded the fact that case
1873        dependency changed within the pattern. This was once needed for "required
1874        byte" processing, but is no longer used. This recovers a now-scarce options
1875        bit. Also moved the least significant internal flag bit to the most-
1876        significant bit of the word, which was not previously used (hangover from
1877        the days when it was an int rather than a uint) to free up another bit for
1878        the future.
1879    
1880    23. Added support for CRLF line endings as well as CR and LF. As well as the
1881        default being selectable at build time, it can now be changed at runtime
1882        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1883        specify that it is scanning data with non-default line endings.
1884    
1885    24. Changed the definition of CXXLINK to make it agree with the definition of
1886        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1887    
1888    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1889        recursions. This makes a big different to stack usage for some patterns.
1890    
1891    26. If a subpattern containing a named recursion or subroutine reference such
1892        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1893        the space required for the compiled pattern went wrong and gave too small a
1894        value. Depending on the environment, this could lead to "Failed: internal
1895        error: code overflow at offset 49" or "glibc detected double free or
1896        corruption" errors.
1897    
1898    27. Applied patches from Google (a) to support the new newline modes and (b) to
1899        advance over multibyte UTF-8 characters in GlobalReplace.
1900    
1901    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1902        difference for some implementation of PCRE in some Windows version.
1903    
1904    29. Added some extra testing facilities to pcretest:
1905    
1906        \q<number>   in a data line sets the "match limit" value
1907        \Q<number>   in a data line sets the "match recursion limt" value
1908        -S <number>  sets the stack size, where <number> is in megabytes
1909    
1910        The -S option isn't available for Windows.
1911    
1912    
1913    Version 6.6 06-Feb-06
1914    ---------------------
1915    
1916     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1917        in pcreposix.h. I have copied the definition from pcre.h.
1918    
1919     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1920        because pcre.h is no longer a built file.
1921    
1922     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1923        not normally included in the compiled code.
1924    
1925    
1926    Version 6.5 01-Feb-06
1927    ---------------------
1928    
1929     1. When using the partial match feature with pcre_dfa_exec(), it was not
1930        anchoring the second and subsequent partial matches at the new starting
1931        point. This could lead to incorrect results. For example, with the pattern
1932        /1234/, partially matching against "123" and then "a4" gave a match.
1933    
1934     2. Changes to pcregrep:
1935    
1936        (a) All non-match returns from pcre_exec() were being treated as failures
1937            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1938            error message is output. Some extra information is given for the
1939            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1940            probably the only errors that are likely to be caused by users (by
1941            specifying a regex that has nested indefinite repeats, for instance).
1942            If there are more than 20 of these errors, pcregrep is abandoned.
1943    
1944        (b) A binary zero was treated as data while matching, but terminated the
1945            output line if it was written out. This has been fixed: binary zeroes
1946            are now no different to any other data bytes.
1947    
1948        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1949            used to set a locale for matching. The --locale=xxxx long option has
1950            been added (no short equivalent) to specify a locale explicitly on the
1951            pcregrep command, overriding the environment variables.
1952    
1953        (d) When -B was used with -n, some line numbers in the output were one less
1954            than they should have been.
1955    
1956        (e) Added the -o (--only-matching) option.
1957    
1958        (f) If -A or -C was used with -c (count only), some lines of context were
1959            accidentally printed for the final match.
1960    
1961        (g) Added the -H (--with-filename) option.
1962    
1963        (h) The combination of options -rh failed to suppress file names for files
1964            that were found from directory arguments.
1965    
1966        (i) Added the -D (--devices) and -d (--directories) options.
1967    
1968        (j) Added the -F (--fixed-strings) option.
1969    
1970        (k) Allow "-" to be used as a file name for -f as well as for a data file.
1971    
1972        (l) Added the --colo(u)r option.
1973    
1974        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1975            is not present by default.
1976    
1977     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1978        items such as (?R) or (?1), when the recursion could match a number of
1979        alternatives. If it matched one of the alternatives, but subsequently,
1980        outside the recursion, there was a failure, the code tried to back up into
1981        the recursion. However, because of the way PCRE is implemented, this is not
1982        possible, and the result was an incorrect result from the match.
1983    
1984        In order to prevent this happening, the specification of recursion has
1985        been changed so that all such subpatterns are automatically treated as
1986        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1987    
1988     4. I had overlooked the fact that, in some locales, there are characters for
1989        which isalpha() is true but neither isupper() nor islower() are true. In
1990        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1991        and ordfeminine) are like this. This affected the treatment of \w and \W
1992        when they appeared in character classes, but not when they appeared outside
1993        a character class. The bit map for "word" characters is now created
1994        separately from the results of isalnum() instead of just taking it from the
1995        upper, lower, and digit maps. (Plus the underscore character, of course.)
1996    
1997     5. The above bug also affected the handling of POSIX character classes such as
1998        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1999        permanent tables. Instead, the bit maps for such a class were previously
2000        created as the appropriate unions of the upper, lower, and digit bitmaps.
2001        Now they are created by subtraction from the [[:word:]] class, which has
2002        its own bitmap.
2003    
2004     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2005        It is created by subtracting the vertical space characters (\x09, \x0a,
2006        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2007        subtraction was done in the overall bitmap for a character class, meaning
2008        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2009        be recognized. This bug has been fixed.
2010    
2011     7. Patches from the folks at Google:
2012    
2013          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2014          real life, but is still worth protecting against".
2015    
2016          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2017          regular expressions".
2018    
2019          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2020          have it.
2021    
2022          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2023          "configure" and the latter not, in order to fix a problem somebody had
2024          with compiling the Arg class on HP-UX.
2025    
2026          (e) Improve the error-handling of the C++ wrapper a little bit.
2027    
2028          (f) New tests for checking recursion limiting.
2029    
2030     8. The pcre_memmove() function, which is used only if the environment does not
2031        have a standard memmove() function (and is therefore rarely compiled),
2032        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2033        returning a result (though PCRE never actually uses the result).
2034    
2035     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2036        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2037        returned instead of calling malloc() with an overflowing number that would
2038        most likely cause subsequent chaos.
2039    
2040    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2041    
2042    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2043        with this option is matched, the nmatch and pmatch options of regexec() are
2044        ignored.
2045    
2046    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2047        provided in case anyone wants to the the POSIX interface with UTF-8
2048        strings.
2049    
2050    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2051        C++ linking (needed for some HP-UX environments).
2052    
2053    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2054        (unused parameter) and in the pcre_printint() function (omitted "default"
2055        switch label when the default is to do nothing).
2056    
2057    15. Added some code to make it possible, when PCRE is compiled as a C++
2058        library, to replace subject pointers for pcre_exec() with a smart pointer
2059        class, thus making it possible to process discontinuous strings.
2060    
2061    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2062        much the same function. They were added by different people who were trying
2063        to make PCRE easy to compile on non-Unix systems. It has been suggested
2064        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2065        for compiling on Windows systems. I have therefore replaced it with
2066        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2067        defaults to "extern" for C or "extern C" for C++, which works fine on
2068        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2069        SCOPE with something explicit in config.h. In addition:
2070    
2071        (a) pcreposix.h still had just "extern" instead of either of these macros;
2072            I have replaced it with PCRE_DATA_SCOPE.
2073    
2074        (b) Functions such as _pcre_xclass(), which are internal to the library,
2075            but external in the C sense, all had PCRE_EXPORT in their definitions.
2076            This is apparently wrong for the Windows case, so I have removed it.
2077            (It makes no difference on Unix-like systems.)
2078    
2079    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2080        of recursive calls to match(). This is different to MATCH_LIMIT because
2081        that limits the total number of calls to match(), not all of which increase
2082        the depth of recursion. Limiting the recursion depth limits the amount of
2083        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2084        when PCRE is compiled, and changed at run time. A patch from Google adds
2085        this functionality to the C++ interface.
2086    
2087    18. Changes to the handling of Unicode character properties:
2088    
2089        (a) Updated the table to Unicode 4.1.0.
2090    
2091        (b) Recognize characters that are not in the table as "Cn" (undefined).
2092    
2093        (c) I revised the way the table is implemented to a much improved format
2094            which includes recognition of ranges. It now supports the ranges that
2095            are defined in UnicodeData.txt, and it also amalgamates other
2096            characters into ranges. This has reduced the number of entries in the
2097            table from around 16,000 to around 3,000, thus reducing its size
2098            considerably. I realized I did not need to use a tree structure after
2099            all - a binary chop search is just as efficient. Having reduced the
2100            number of entries, I extended their size from 6 bytes to 8 bytes to
2101            allow for more data.
2102    
2103        (d) Added support for Unicode script names via properties such as \p{Han}.
2104    
2105    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2106        matching that character.
2107    
2108    20. When matching a repeated Unicode property with a minimum greater than zero,
2109        (for example \pL{2,}), PCRE could look past the end of the subject if it
2110        reached it while seeking the minimum number of characters. This could
2111        happen only if some of the characters were more than one byte long, because
2112        there is a check for at least the minimum number of bytes.
2113    
2114    21. Refactored the implementation of \p and \P so as to be more general, to
2115        allow for more different types of property in future. This has changed the
2116        compiled form incompatibly. Anybody with saved compiled patterns that use
2117        \p or \P will have to recompile them.
2118    
2119    22. Added "Any" and "L&" to the supported property types.
2120    
2121    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2122        but give a compile time error if the value is greater than 0xff.
2123    
2124    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2125        accidentally not being installed or uninstalled.
2126    
2127    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2128        made were to insert the current release number. This seemed silly, because
2129        it made things harder for people building PCRE on systems that don't run
2130        "configure". I have turned pcre.h into a distributed file, no longer built
2131        by "configure", with the version identification directly included. There is
2132        no longer a pcre.h.in file.
2133    
2134        However, this change necessitated a change to the pcre-config script as
2135        well. It is built from pcre-config.in, and one of the substitutions was the
2136        release number. I have updated configure.ac so that ./configure now finds
2137        the release number by grepping pcre.h.
2138    
2139    26. Added the ability to run the tests under valgrind.
2140    
2141    
2142    Version 6.4 05-Sep-05
2143    ---------------------
2144    
2145     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2146        "--" to be printed when multiple files were scanned, even when none of the
2147        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2148        consider it to be a bug, and have restored the previous behaviour.
2149    
2150     2. A couple of code tidies to get rid of compiler warnings.
2151    
2152     3. The pcretest program used to cheat by referring to symbols in the library
2153        whose names begin with _pcre_. These are internal symbols that are not
2154        really supposed to be visible externally, and in some environments it is
2155        possible to suppress them. The cheating is now confined to including
2156        certain files from the library's source, which is a bit cleaner.
2157    
2158     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2159        file's purpose clearer.
2160    
2161     5. Reorganized pcre_ucp_findchar().
2162    
2163    
2164    Version 6.3 15-Aug-05
2165    ---------------------
2166    
2167     1. The file libpcre.pc.in did not have general read permission in the tarball.
2168    
2169     2. There were some problems when building without C++ support:
2170    
2171        (a) If C++ support was not built, "make install" and "make test" still
2172            tried to test it.
2173    
2174        (b) There were problems when the value of CXX was explicitly set. Some
2175            changes have been made to try to fix these, and ...
2176    
2177        (c) --disable-cpp can now be used to explicitly disable C++ support.
2178    
2179        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2180            backslash in a target when C++ was disabled. This confuses some
2181            versions of "make", apparently. Using an intermediate variable solves
2182            this. (Same for CPP_LOBJ.)
2183    
2184     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2185        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2186        necessary on certain architectures.
2187    
2188     4. Added a setting of -export-symbols-regex to the link command to remove
2189        those symbols that are exported in the C sense, but actually are local
2190        within the library, and not documented. Their names all begin with
2191        "_pcre_". This is not a perfect job, because (a) we have to except some
2192        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2193        available (and never for static libraries). I have made a note to try to
2194        find a way round (a) in the future.
2195    
2196    
2197    Version 6.2 01-Aug-05
2198    ---------------------
2199    
2200     1. There was no test for integer overflow of quantifier values. A construction
2201        such as {1111111111111111} would give undefined results. What is worse, if
2202        a minimum quantifier for a parenthesized subpattern overflowed and became
2203        negative, the calculation of the memory size went wrong. This could have
2204        led to memory overwriting.
2205    
2206     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2207    
2208     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2209        operating environments where this matters.
2210    
2211     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2212        PCRE options from within the C++ wrapper.
2213    
2214     5. Named capturing subpatterns were not being correctly counted when a pattern
2215        was compiled. This caused two problems: (a) If there were more than 100
2216        such subpatterns, the calculation of the memory needed for the whole
2217        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2218        back references of the form \12, where the number was greater than 9, were
2219        not recognized as back references, even though there were sufficient
2220        previous subpatterns.
2221    
2222     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2223        versions of gcc, e.g. 2.95.4.
2224    
2225    
2226    Version 6.1 21-Jun-05
2227    ---------------------
2228    
2229     1. There was one reference to the variable "posix" in pcretest.c that was not
2230        surrounded by "#if !defined NOPOSIX".
2231    
2232     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2233        the cross-check on the old pcre_info() function, for the benefit of the
2234        cut-down version of PCRE that is currently imported into Exim.
2235    
2236     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2237        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2238        patterns that start (?i) but I don't think that matters. The use of (?i) is
2239        just an example; this all applies to the other options as well.
2240    
2241     4. Since libtool seems to echo the compile commands it is issuing, the output
2242        from "make" can be reduced a bit by putting "@" in front of each libtool
2243        compile command.
2244    
2245     5. Patch from the folks at Google for configure.in to be a bit more thorough
2246        in checking for a suitable C++ installation before trying to compile the
2247        C++ stuff. This should fix a reported problem when a compiler was present,
2248        but no suitable headers.
2249    
2250     6. The man pages all had just "PCRE" as their title. I have changed them to
2251        be the relevant file name. I have also arranged that these names are
2252        retained in the file doc/pcre.txt, which is a concatenation in text format
2253        of all the man pages except the little individual ones for each function.
2254    
2255     7. The NON-UNIX-USE file had not been updated for the different set of source
2256        files that come with release 6. I also added a few comments about the C++
2257        wrapper.
2258    
2259    
2260    Version 6.0 07-Jun-05
2261    ---------------------
2262    
2263     1. Some minor internal re-organization to help with my DFA experiments.
2264    
2265     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2266        didn't matter for the library itself when fully configured, but did matter
2267        when compiling without UCP support, or within Exim, where the ucp files are
2268        not imported.
2269    
2270     3. Refactoring of the library code to split up the various functions into
2271        different source modules. The addition of the new DFA matching code (see
2272        below) to a single monolithic source would have made it really too
2273        unwieldy, quite apart from causing all the code to be include in a
2274        statically linked application, when only some functions are used. This is
2275        relevant even without the DFA addition now that patterns can be compiled in
2276        one application and matched in another.
2277    
2278        The downside of splitting up is that there have to be some external
2279        functions and data tables that are used internally in different modules of
2280        the library but which are not part of the API. These have all had their
2281        names changed to start with "_pcre_" so that they are unlikely to clash
2282        with other external names.
2283    
2284     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2285        a different (DFA) algorithm. Although it is slower than the original
2286        function, it does have some advantages for certain types of matching
2287        problem.
2288    
2289     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2290        including restarting after a partial match.
2291    
2292     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2293        defined when compiling for Windows was sent to me. I have put it into the
2294        code, though I have no means of testing or verifying it.
2295    
2296     7. Added the pcre_refcount() auxiliary function.
2297    
2298     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2299        match before or at the first newline in the subject string. In pcretest,
2300        the /f option on a pattern can be used to set this.
2301    
2302     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2303        would behave wrongly. This has been present in PCRE since release 4.0.
2304    
2305    10. A number of changes to the pcregrep command:
2306    
2307        (a) Refactored how -x works; insert ^(...)$ instead of setting
2308            PCRE_ANCHORED and checking the length, in preparation for adding
2309            something similar for -w.
2310    
2311        (b) Added the -w (match as a word) option.
2312    
2313        (c) Refactored the way lines are read and buffered so as to have more
2314            than one at a time available.
2315    
2316        (d) Implemented a pcregrep test script.
2317    
2318        (e) Added the -M (multiline match) option. This allows patterns to match
2319            over several lines of the subject. The buffering ensures that at least
2320            8K, or the rest of the document (whichever is the shorter) is available
2321            for matching (and similarly the previous 8K for lookbehind assertions).
2322    
2323        (f) Changed the --help output so that it now says
2324    
2325              -w, --word-regex(p)
2326    
2327            instead of two lines, one with "regex" and the other with "regexp"
2328            because that confused at least one person since the short forms are the
2329            same. (This required a bit of code, as the output is generated
2330            automatically from a table. It wasn't just a text change.)
2331    
2332        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2333            option but starts with a hyphen. Could be a pattern or a path name
2334            starting with a hyphen, for instance.
2335    
2336        (h) "-" can be given as a file name to represent stdin.
2337    
2338        (i) When file names are being printed, "(standard input)" is used for
2339            the standard input, for compatibility with GNU grep. Previously
2340            "<stdin>" was used.
2341    
2342        (j) The option --label=xxx can be used to supply a name to be used for
2343            stdin when file names are being printed. There is no short form.
2344    
2345        (k) Re-factored the options decoding logic because we are going to add
2346            two more options that take data. Such options can now be given in four
2347            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2348    
2349        (l) Added the -A, -B, and -C options for requesting that lines of context
2350            around matches be printed.
2351    
2352        (m) Added the -L option to print the names of files that do not contain
2353            any matching lines, that is, the complement of -l.
2354    
2355        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2356            continue to scan other files.
2357    
2358        (o) The -s option was incorrectly implemented. For compatibility with other
2359            greps, it now suppresses the error message for a non-existent or non-
2360            accessible file (but not the return code). There is a new option called
2361            -q that suppresses the output of matching lines, which was what -s was
2362            previously doing.
2363    
2364        (p) Added --include and --exclude options to specify files for inclusion
2365            and exclusion when recursing.
2366    
2367    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2368        Hopefully, it now does.
2369    
2370    12. Missing cast in pcre_study().
2371    
2372    13. Added an "uninstall" target to the makefile.
2373    
2374    14. Replaced "extern" in the function prototypes in Makefile.in with
2375        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2376        world, but is set differently for Windows.
2377    
2378    15. Added a second compiling function called pcre_compile2(). The only
2379        difference is that it has an extra argument, which is a pointer to an
2380        integer error code. When there is a compile-time failure, this is set
2381        non-zero, in addition to the error test pointer being set to point to an
2382        error message. The new argument may be NULL if no error number is required
2383        (but then you may as well call pcre_compile(), which is now just a
2384        wrapper). This facility is provided because some applications need a
2385        numeric error indication, but it has also enabled me to tidy up the way
2386        compile-time errors are handled in the POSIX wrapper.
2387    
2388    16. Added VPATH=.libs to the makefile; this should help when building with one
2389        prefix path and installing with another. (Or so I'm told by someone who
2390        knows more about this stuff than I do.)
2391    
2392    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2393        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2394        match everything, including newlines. This is not POSIX-compatible, but
2395        somebody wanted the feature. From pcretest it can be activated by using
2396        both the P and the s flags.
2397    
2398    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2399    
2400    19. libpcre.pc was being incorrectly installed as executable.
2401    
2402    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2403        it now also looks for '\r' so that it will work unmodified on Windows.
2404    
2405    21. Added Google's contributed C++ wrapper to the distribution.
2406    
2407    22. Added some untidy missing memory free() calls in pcretest, to keep
2408        Electric Fence happy when testing.
2409    
2410    
2411    
2412    Version 5.0 13-Sep-04
2413    ---------------------
2414    
2415     1. Internal change: literal characters are no longer packed up into items
2416        containing multiple characters in a single byte-string. Each character
2417        is now matched using a separate opcode. However, there may be more than one
2418        byte in the character in UTF-8 mode.
2419    
2420     2. The pcre_callout_block structure has two new fields: pattern_position and
2421        next_item_length. These contain the offset in the pattern to the next match
2422        item, and its length, respectively.
2423    
2424     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2425        insertion of callouts before each pattern item. Added the /C option to
2426        pcretest to make use of this.
2427    
2428     4. On the advice of a Windows user, the lines
2429    
2430          #if defined(_WIN32) || defined(WIN32)
2431          _setmode( _fileno( stdout ), 0x8000 );
2432          #endif  /* defined(_WIN32) || defined(WIN32) */
2433    
2434        have been added to the source of pcretest. This apparently does useful
2435        magic in relation to line terminators.
2436    
2437     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2438        for the benefit of those environments where the "b" makes a difference.
2439    
2440     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2441        to know about it. I have put a hack into configure.in that adds in code
2442        to set GCC=yes if CC=icc. This seems to end up at a point in the
2443        generated configure script that is early enough to affect the setting of
2444        compiler options, which is what is needed, but I have no means of testing
2445        whether it really works. (The user who reported this had patched the
2446        generated configure script, which of course I cannot do.)
2447    
2448        LATER: After change 22 below (new libtool files), the configure script
2449        seems to know about icc (and also ecc). Therefore, I have commented out
2450        this hack in configure.in.
2451    
2452     7. Added support for pkg-config (2 patches were sent in).
2453    
2454     8. Negated POSIX character classes that used a combination of internal tables
2455        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2456        [[:^ascii]]. Typically, they would match almost any characters. The other
2457        POSIX classes were not broken in this way.
2458    
2459     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2460        to find the match, as PCRE was deluded into thinking that the match had to
2461        start at the start point or following a newline. The same bug applied to
2462        patterns with negative forward assertions or any backward assertions
2463        preceding ".*" at the start, unless the pattern required a fixed first
2464        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2465    
2466    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2467        starting at the last subject character, bytes beyond the end of the subject
2468        string were read.
2469    
2470    11. Renamed the variable "class" as "classbits" to make life easier for C++
2471        users. (Previously there was a macro definition, but it apparently wasn't
2472        enough.)
2473    
2474    12. Added the new field "tables" to the extra data so that tables can be passed
2475        in at exec time, or the internal tables can be re-selected. This allows
2476        a compiled regex to be saved and re-used at a later time by a different
2477        program that might have everything at different addresses.
2478    
2479    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2480        -R library as well as a -L library.
2481    
2482    14. The debugging options of pcretest (-d on the command line or D on a
2483        pattern) showed incorrect output for anything following an extended class
2484        that contained multibyte characters and which was followed by a quantifier.
2485    
2486    15. Added optional support for general category Unicode character properties
2487        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2488        support. It adds about 90K to the size of the library. The meanings of the
2489        inbuilt class escapes such as \d and \s have NOT been changed.
2490    
2491    16. Updated pcredemo.c to include calls to free() to release the memory for the
2492        compiled pattern.
2493    
2494    17. The generated file chartables.c was being created in the source directory
2495        instead of in the building directory. This caused the build to fail if the
2496        source directory was different from the building directory, and was
2497        read-only.
2498    
2499    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2500        file. No doubt somebody will tell me if they don't make sense... Also added
2501        Dan Mooney's comments about building on OpenVMS.
2502    
2503    19. Added support for partial matching via the PCRE_PARTIAL option for
2504        pcre_exec() and the \P data escape in pcretest.
2505    
2506    20. Extended pcretest with 3 new pattern features:
2507    
2508        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2509              write the compiled pattern to the file whose name is "rest-of-line".
2510              This is a straight binary dump of the data, with the saved pointer to
2511              the character tables forced to be NULL. The study data, if any, is
2512              written too. After writing, pcretest reads a new pattern.
2513    
2514        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2515              compiled pattern from the given file. There must not be any
2516              occurrences of "<" in the file name (pretty unlikely); if there are,
2517              pcretest will instead treat the initial "<" as a pattern delimiter.
2518              After reading in the pattern, pcretest goes on to read data lines as
2519              usual.
2520    
2521        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2522              and 16-bit fields in a compiled pattern, to simulate a pattern that
2523              was compiled on a host of opposite endianness.
2524    
2525    21. The pcre-exec() function can now cope with patterns that were compiled on
2526        hosts of opposite endianness, with this restriction:
2527    
2528          As for any compiled expression that is saved and used later, the tables
2529          pointer field cannot be preserved; the extra_data field in the arguments
2530          to pcre_exec() should be used to pass in a tables address if a value
2531          other than the default internal tables were used at compile time.
2532    
2533    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2534        now diagnosed as an error. Previously, most of the time, a negative number
2535        would have been treated as zero, but if in addition "ovector" was passed as
2536        NULL, a crash could occur.
2537    
2538    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2539        new versions from the libtool 1.5 distribution (the last one is a copy of
2540        a file called libtool.m4). This seems to have fixed the need to patch
2541        "configure" to support Darwin 1.3 (which I used to do). However, I still
2542        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2543        workstation).
2544    
2545    24. Changed the PCRE licence to be the more standard "BSD" licence.
2546    
2547    
2548    Version 4.5 01-Dec-03
2549    ---------------------
2550    
2551     1. There has been some re-arrangement of the code for the match() function so
2552        that it can be compiled in a version that does not call itself recursively.
2553        Instead, it keeps those local variables that need separate instances for
2554        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2555        needs to "recurse". Keeping track of where control must go is done by means
2556        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2557        hide most of the details from the main code, and operates only if
2558        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2559        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2560        operating.
2561    
2562        To make it easier for callers to provide specially tailored get/free
2563        functions for this usage, two new functions, pcre_stack_malloc, and
2564        pcre_stack_free, are used. They are always called in strict stacking order,
2565        and the size of block requested is always the same.
2566    
2567        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2568        PCRE has been compiled to use the stack or the heap for recursion. The
2569        -C option of pcretest uses this to show which version is compiled.
2570    
2571        A new data escape \S, is added to pcretest; it causes the amounts of store
2572        obtained and freed by both kinds of malloc/free at match time to be added
2573        to the output.
2574    
2575     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2576        what's available on my current Linux desktop machine.
2577    
2578     3. When matching a UTF-8 string, the test for a valid string at the start has
2579        been extended. If start_offset is not zero, PCRE now checks that it points
2580        to a byte that is the start of a UTF-8 character. If not, it returns
2581        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2582        this is necessary because there may be backward assertions in the pattern.
2583        When matching the same subject several times, it may save resources to use
2584        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2585    
2586     4. The code for checking the validity of UTF-8 strings has been tightened so
2587        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2588        containing "overlong sequences".
2589    
2590     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2591        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2592        should have been "&", but it just so happened that all the cases this let
2593        through by mistake were picked up later in the function.
2594    
2595     6. I had used a variable called "isblank" - this is a C99 function, causing
2596        some compilers to warn. To avoid this, I renamed it (as "blankclass").
2597    
2598     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2599        prompting; (b) run "./pcretest /dev/null" at the start of the test script
2600        so the version is shown; (c) stop "make test" echoing "./RunTest".
2601    
2602     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2603    
2604     9. The prototype for memmove() for systems that don't have it was using
2605        size_t, but the inclusion of the header that defines size_t was later. I've
2606        moved the #includes for the C headers earlier to avoid this.
2607    
2608    10. Added some adjustments to the code to make it easier to compiler on certain
2609        special systems:
2610    
2611          (a) Some "const" qualifiers were missing.
2612          (b) Added the macro EXPORT before all exported functions; by default this
2613              is defined to be empty.
2614          (c) Changed the dftables auxiliary program (that builds chartables.c) so
2615              that it reads its output file name as an argument instead of writing
2616              to the standard output and assuming this can be redirected.
2617    
2618    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2619        class containing characters with values greater than 255, PCRE compilation
2620        went into a loop.
2621    
2622    12. A recursive reference to a subpattern that was within another subpattern
2623        that had a minimum quantifier of zero caused PCRE to crash. For example,
2624        (x(y(?2))z)? provoked this bug with a subject that got as far as the
2625        recursion. If the recursively-called subpattern itself had a zero repeat,
2626        that was OK.
2627    
2628    13. In pcretest, the buffer for reading a data line was set at 30K, but the
2629        buffer into which it was copied (for escape processing) was still set at
2630        1024, so long lines caused crashes.
2631    
2632    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2633        "internal error: code overflow...". This applied to any character class
2634        that was followed by a possessive quantifier.
2635    
2636    15. Modified the Makefile to add libpcre.la as a prerequisite for
2637        libpcreposix.la because I was told this is needed for a parallel build to
2638        work.
2639    
2640    16. If a pattern that contained .* following optional items at the start was
2641        studied, the wrong optimizing data was generated, leading to matching
2642        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2643        matching string must start with a or b or c. The correct conclusion for
2644        this pattern is that a match can start with any character.
2645    
2646    
2647    Version 4.4 13-Aug-03
2648    ---------------------
2649    
2650     1. In UTF-8 mode, a character class containing characters with values between
2651        127 and 255 was not handled correctly if the compiled pattern was studied.
2652        In fixing this, I have also improved the studying algorithm for such
2653        classes (slightly).
2654    
2655     2. Three internal functions had redundant arguments passed to them. Removal
2656        might give a very teeny performance improvement.
2657    
2658     3. Documentation bug: the value of the capture_top field in a callout is *one
2659        more than* the number of the hightest numbered captured substring.
2660    
2661     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2662        in incorrectly linking with a previously installed version. They now link
2663        explicitly with libpcre.la.
2664    
2665     5. configure.in no longer needs to recognize Cygwin specially.
2666    
2667     6. A problem in pcre.in for Windows platforms is fixed.
2668    
2669     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2670        pcretest, it used to include the size of the study block as part of its
2671        output. Unfortunately, the structure contains a field that has a different
2672        size on different hardware architectures. This meant that the tests that
2673        showed this size failed. As the block is currently always of a fixed size,
2674        this information isn't actually particularly useful in pcretest output, so
2675        I have just removed it.
2676    
2677     8. Three pre-processor statements accidentally did not start in column 1.
2678        Sadly, there are *still* compilers around that complain, even though
2679        standard C has not required this for well over a decade. Sigh.
2680    
2681     9. In pcretest, the code for checking callouts passed small integers in the
2682        callout_data field, which is a void * field. However, some picky compilers
2683        complained about the casts involved for this on 64-bit systems. Now
2684        pcretest passes the address of the small integer instead, which should get
2685        rid of the warnings.
2686    
2687    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2688        both compile and run time, and gives an error if an invalid UTF-8 sequence
2689        is found. There is a option for disabling this check in cases where the
2690        string is known to be correct and/or the maximum performance is wanted.
2691    
2692    11. In response to a bug report, I changed one line in Makefile.in from
2693    
2694            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2695        to
2696            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2697    
2698        to look similar to other lines, but I have no way of telling whether this
2699        is the right thing to do, as I do not use Windows. No doubt I'll get told
2700        if it's wrong...
2701    
2702    
2703    Version 4.3 21-May-03
2704    ---------------------
2705    
2706    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2707       Makefile.
2708    
2709    2. Some refactoring to improve the quality of the code:
2710    
2711       (i)   The utf8_table... variables are now declared "const".
2712    
2713       (ii)  The code for \cx, which used the "case flipping" table to upper case
2714             lower case letters, now just substracts 32. This is ASCII-specific,
2715             but the whole concept of \cx is ASCII-specific, so it seems
2716             reasonable.
2717    
2718       (iii) PCRE was using its character types table to recognize decimal and
2719             hexadecimal digits in the pattern. This is silly, because it handles
2720             only 0-9, a-f, and A-F, but the character types table is locale-
2721             specific, which means strange things might happen. A private
2722             table is now used for this - though it costs 256 bytes, a table is
2723             much faster than multiple explicit tests. Of course, the standard
2724             character types table is still used for matching digits in subject
2725             strings against \d.
2726    
2727       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2728             ending in _t are). So I've renamed it as ESC_tee.
2729    
2730    3. The first argument for regexec() in the POSIX wrapper should have been
2731       defined as "const".
2732    
2733    4. Changed pcretest to use malloc() for its buffers so that they can be
2734       Electric Fenced for debugging.
2735    
2736    5. There were several places in the code where, in UTF-8 mode, PCRE would try
2737       to read one or more bytes before the start of the subject string. Often this
2738       had no effect on PCRE's behaviour, but in some circumstances it could
2739       provoke a segmentation fault.
2740    
2741    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2742       to try to read one or more bytes before the start of the subject string.
2743    
2744    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2745       UTF-8 support could misbehave in various ways if the subject string
2746       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2747       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2748       back over UTF-8 characters.)
2749    
2750    
2751    Version 4.2 14-Apr-03
2752    ---------------------
2753    
2754    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2755    
2756    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2757         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2758         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2759         [WIN_PREFIX]: new variable, "cyg" for Cygwin
2760         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2761           and BUILD_EXEEXT
2762         Note: automatic setting of the BUILD variables is not yet working
2763         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2764           compile-time but not at link-time
2765         [LINK]: use for linking executables only
2766         make different versions for Windows and non-Windows
2767         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2768           libraries
2769         [LINK_FOR_BUILD]: new variable
2770         [OBJEXT]: use throughout
2771         [EXEEXT]: use throughout
2772         <winshared>: new target
2773         <wininstall>: new target
2774         <dftables.o>: use native compiler
2775         <dftables>: use native linker
2776         <install>: handle Windows platform correctly
2777         <clean>: ditto
2778         <check>: ditto
2779         copy DLL to top builddir before testing
2780    
2781       As part of these changes, -no-undefined was removed again. This was reported
2782       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2783       in any case.
2784    
2785    3. Some tidies to get rid of compiler warnings:
2786    
2787       . In the match_data structure, match_limit was an unsigned long int, whereas
2788         match_call_count was an int. I've made them both unsigned long ints.
2789    
2790       . In pcretest the fact that a const uschar * doesn't automatically cast to
2791         a void * provoked a warning.
2792    
2793       . Turning on some more compiler warnings threw up some "shadow" variables
2794         and a few more missing casts.
2795    
2796    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2797       option, a class that contained a single character with a value between 128
2798       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2799    
2800    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2801       option, a class that contained several characters, but with at least one
2802       whose value was between 128 and 255 caused PCRE to crash.
2803    
2804    
2805    Version 4.1 12-Mar-03
2806    ---------------------
2807    
2808    1. Compiling with gcc -pedantic found a couple of places where casts were
2809    needed, and a string in dftables.c that was longer than standard compilers are
2810    required to support.
2811    
2812    2. Compiling with Sun's compiler found a few more places where the code could
2813    be tidied up in order to avoid warnings.
2814    
2815    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2816    first of these names is deprecated in the latest Autoconf in favour of the name
2817    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2818    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2819    analogy I have changed it to CFLAGS_FOR_BUILD.
2820    
2821    4. Added -no-undefined to the linking command in the Makefile, because this is
2822    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2823    linking step for the pcreposix library.
2824    
2825    5. PCRE was failing to diagnose the case of two named groups with the same
2826    name.
2827    
2828    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2829    literal character that is needed in the subject for a match, and scans along to
2830    ensure that it is present before embarking on the full matching process. This
2831    saves time in cases of nested unlimited repeats that are never going to match.
2832    Problem: the scan can take a lot of time if the subject is very long (e.g.
2833    megabytes), thus penalizing straightforward matches. It is now done only if the
2834    amount of subject to be scanned is less than 1000 bytes.
2835    
2836    7. A lesser problem with the same optimization is that it was recording the
2837    first character of an anchored pattern as "needed", thus provoking a search
2838    right along the subject, even when the first match of the pattern was going to
2839    fail. The "needed" character is now not set for anchored patterns, unless it
2840    follows something in the pattern that is of non-fixed length. Thus, it still
2841    fulfils its original purpose of finding quick non-matches in cases of nested
2842    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2843    
2844    
2845    Version 4.0 17-Feb-03
2846    ---------------------
2847    
2848    1. If a comment in an extended regex that started immediately after a meta-item
2849    extended to the end of string, PCRE compiled incorrect data. This could lead to
2850    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2851    
2852    2. Moved to autoconf 2.53 and libtool 1.4.2.
2853    
2854    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2855    the special perltest8 script is no longer needed - all the tests can be run
2856    from a single perltest script.
2857    
2858    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2859    by \s. It has now been removed in PCRE. This means it isn't recognized as
2860    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2861    class [:space:] *does* include VT, thereby creating a mess.
2862    
2863    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2864    space and tab.
2865    
2866    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2867    its new features into the main test script, reducing the number of scripts.
2868    
2869    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2870    were backward compatible, and made the (?i) apply to the whole pattern, as if
2871    /i were given. Now it behaves more logically, and applies the option setting
2872    only to what follows. PCRE has been changed to follow suit. However, if it
2873    finds options settings right at the start of the pattern, it extracts them into
2874    the global options, as before. Thus, they show up in the info data.
2875    
2876    8. Added support for the \Q...\E escape sequence. Characters in between are
2877    treated as literals. This is slightly different from Perl in that $ and @ are
2878    also handled as literals inside the quotes. In Perl, they will cause variable
2879    interpolation. Note the following examples:
2880    
2881        Pattern            PCRE matches      Perl matches
2882    
2883        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
2884        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
2885        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
2886    
2887    For compatibility with Perl, \Q...\E sequences are recognized inside character
2888    classes as well as outside them.
2889    
2890    9. Re-organized 3 code statements in pcretest to avoid "overflow in
2891    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2892    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2893    signed/unsigned warnings.
2894    
2895    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2896    option for pcretest, so I've replaced it by a simple function that does just
2897    that job.
2898    
2899    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2900    "pcregrep -".
2901    
2902    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2903    Java package. This provides some syntactic sugar for simple cases of what my
2904    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2905    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2906    item, you can use this simplified notation. Note that only makes sense with
2907    greedy quantifiers. Consequently, the use of the possessive quantifier forces
2908    greediness, whatever the setting of the PCRE_UNGREEDY option.
2909    
2910    13. A change of greediness default within a pattern was not taking effect at
2911    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2912    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2913    was abstracted outside.
2914    
2915    14. PCRE now supports the \G assertion. It is true when the current matching
2916    position is at the start point of the match. This differs from \A when the
2917    starting offset is non-zero. Used with the /g option of pcretest (or similar
2918    code), it works in the same way as it does for Perl's /g option. If all
2919    alternatives of a regex begin with \G, the expression is anchored to the start
2920    match position, and the "anchored" flag is set in the compiled expression.
2921    
2922    15. Some bugs concerning the handling of certain option changes within patterns
2923    have been fixed. These applied to options other than (?ims). For example,
2924    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2925    been the other way round. Some of this was related to change 7 above.
2926    
2927    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2928    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2929    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2930    POSIX classes only within a class (e.g. /[[:alpha:]]/).
2931    
2932    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2933    mode. Unlike ".", it always matches newline, whatever the setting of
2934    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2935    assertions. Perl allows it, but it doesn't (in general) work because it can't
2936    calculate the length of the lookbehind. At least, that's the case for Perl
2937    5.8.0 - I've been told they are going to document that it doesn't work in
2938    future.
2939    
2940    18. Added an error diagnosis for escapes that PCRE does not support: these are
2941    \L, \l, \N, \P, \p, \U, \u, and \X.
2942    
2943    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2944    reading past the end of the pattern in cases such as /[abcd/.
2945    
2946    20. PCRE was getting more memory than necessary for patterns with classes that
2947    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2948    
2949    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2950    compiling PCRE for use with Virtual Pascal.
2951    
2952    22. Small fix to the Makefile to make it work properly if the build is done
2953    outside the source tree.
2954    
2955    23. Added a new extension: a condition to go with recursion. If a conditional
2956    subpattern starts with (?(R) the "true" branch is used if recursion has
2957    happened, whereas the "false" branch is used only at the top level.
2958    
2959    24. When there was a very long string of literal characters (over 255 bytes
2960    without UTF support, over 250 bytes with UTF support), the computation of how
2961    much memory was required could be incorrect, leading to segfaults or other
2962    strange effects.
2963    
2964    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2965    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2966    there was a subsequent back reference to those brackets. This meant that, for
2967    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2968    possible to check for precisely this case. All we can do is abandon the
2969    optimization if .* occurs inside capturing brackets when there are any back
2970    references whatsoever. (See below for a better fix that came later.)
2971    
2972    26. The handling of the optimization for finding the first character of a
2973    non-anchored pattern, and for finding a character that is required later in the
2974    match were failing in some cases. This didn't break the matching; it just
2975    failed to optimize when it could. The way this is done has been re-implemented.
2976    
2977    27. Fixed typo in error message for invalid (?R item (it said "(?p").
2978    
2979    28. Added a new feature that provides some of the functionality that Perl
2980    provides with (?{...}). The facility is termed a "callout". The way it is done
2981    in PCRE is for the caller to provide an optional function, by setting
2982    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2983    global variable. By default it is unset, which disables all calling out. To get
2984    the function called, the regex must include (?C) at appropriate points. This
2985    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2986    This provides a means of identifying different callout points. When PCRE
2987    reaches such a point in the regex, if pcre_callout has been set, the external
2988    function is called. It is provided with data in a structure called
2989    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2990    matching continues; if it returns a non-zero value, the match at the current
2991    point fails. However, backtracking will occur if possible. [This was changed
2992    later and other features added - see item 49 below.]
2993    
2994    29. pcretest is upgraded to test the callout functionality. It provides a
2995    callout function that displays information. By default, it shows the start of
2996    the match and the current position in the text. There are some new data escapes
2997    to vary what happens:
2998    
2999        \C+         in addition, show current contents of captured substrings
3000        \C-         do not supply a callout function
3001        \C!n        return 1 when callout number n is reached
3002        \C!n!m      return 1 when callout number n is reached for the mth time
3003    
3004    30. If pcregrep was called with the -l option and just a single file name, it
3005    output "<stdin>" if a match was found, instead of the file name.
3006    
3007    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3008    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3009    pcre_exec(). This saves a malloc/free per call. The default value of
3010    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3011    when configuring.
3012    
3013    32. The default maximum size of a compiled pattern is 64K. There have been a
3014    few cases of people hitting this limit. The code now uses macros to handle the
3015    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3016    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3017    configuring. Tests 2 and 5 work only with 2-byte links because they output
3018    debugging information about compiled patterns.
3019    
3020    33. Internal code re-arrangements:
3021    
3022    (a) Moved the debugging function for printing out a compiled regex into
3023        its own source file (printint.c) and used #include to pull it into
3024        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3025        separate copies.
3026    
3027    (b) Defined the list of op-code names for debugging as a macro in
3028        internal.h so that it is next to the definition of the opcodes.
3029    
3030    (c) Defined a table of op-code lengths for simpler skipping along compiled
3031        code. This is again a macro in internal.h so that it is next to the
3032        definition of the opcodes.
3033    
3034    34. Added support for recursive calls to individual subpatterns, along the
3035    lines of Robin Houston's patch (but implemented somewhat differently).
3036    
3037    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3038    allow it to read and process whole directories in Win32. This code was
3039    contributed by Lionel Fourquaux; it has not been tested by me.
3040    
3041    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3042    used to name a group. Names consist of alphanumerics and underscores, and must
3043    be unique. Back references use the syntax (?P=name) and recursive calls use
3044    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3045    numbers. The function pcre_fullinfo() can be used after compilation to extract
3046    a name/number map. There are three relevant calls:
3047    
3048      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3049      PCRE_INFO_NAMECOUNT            yields the number of entries
3050      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3051    
3052    The map is a vector of fixed-size entries. The size of each entry depends on
3053    the length of the longest name used. The first two bytes of each entry are the
3054    group number, most significant byte first. There follows the corresponding
3055    name, zero terminated. The names are in alphabetical order.
3056    
3057    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3058    case instead of 255. Making it the same both with and without UTF-8 support
3059    means that the same test output works with both.
3060    
3061    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3062    calling malloc() with a zero argument.
3063    
3064    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3065    optimization. I've improved things by keeping a bitmap of backreferences with
3066    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3067    fact referenced, the optimization can be applied. It is unlikely that a
3068    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3069    the match to follow \n) will appear inside brackets with a number greater than
3070    31, but if it does, any back reference > 31 suppresses the optimization.
3071    
3072    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3073    of disabling numbered capturing parentheses. Any opening parenthesis that is
3074    not followed by ? behaves as if it were followed by ?: but named parentheses
3075    can still be used for capturing (and they will acquire numbers in the usual
3076    way).
3077    
3078    41. Redesigned the return codes from the match() function into yes/no/error so
3079    that errors can be passed back from deep inside the nested calls. A malloc
3080    failure while inside a recursive subpattern call now causes the
3081    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3082    
3083    42. It is now possible to set a limit on the number of times the match()
3084    function is called in a call to pcre_exec(). This facility makes it possible to
3085    limit the amount of recursion and backtracking, though not in a directly
3086    obvious way, because the match() function is used in a number of different
3087    circumstances. The count starts from zero for each position in the subject
3088    string (for non-anchored patterns). The default limit is, for compatibility, a
3089    large number, namely 10 000 000. You can change this in two ways:
3090    
3091    (a) When configuring PCRE before making, you can use --with-match-limit=n
3092        to set a default value for the compiled library.
3093    
3094    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3095        a different value is set. See 45 below.
3096    
3097    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3098    
3099    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3100    of things that can be changed at compile time. The first argument specifies
3101    what is wanted and the second points to where the information is to be placed.
3102    The current list of available information is:
3103    
3104      PCRE_CONFIG_UTF8
3105    
3106    The output is an integer that is set to one if UTF-8 support is available;
3107    otherwise it is set to zero.
3108    
3109      PCRE_CONFIG_NEWLINE
3110    
3111    The output is an integer that it set to the value of the code that is used for
3112    newline. It is either LF (10) or CR (13).
3113    
3114      PCRE_CONFIG_LINK_SIZE
3115    
3116    The output is an integer that contains the number of bytes used for internal
3117    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3118    
3119      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3120    
3121    The output is an integer that contains the threshold above which the POSIX
3122    interface uses malloc() for output vectors. See item 31 above.
3123    
3124      PCRE_CONFIG_MATCH_LIMIT
3125    
3126    The output is an unsigned integer that contains the default limit of the number
3127    of match() calls in a pcre_exec() execution. See 42 above.
3128    
3129    44. pcretest has been upgraded by the addition of the -C option. This causes it
3130    to extract all the available output from the new pcre_config() function, and to
3131    output it. The program then exits immediately.
3132    
3133    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3134    order to support additional features. One way would have been to define
3135    pcre_exec2() (for example) with extra arguments, but this would not have been
3136    extensible, and would also have required all calls to the original function to
3137    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3138    is used for passing in "extra" data from pcre_study().
3139    
3140    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3141    contains the following fields:
3142    
3143      flags         a bitmap indicating which of the following fields are set
3144      study_data    opaque data from pcre_study()
3145      match_limit   a way of specifying a limit on match() calls for a specific
3146                      call to pcre_exec()
3147      callout_data  data for callouts (see 49 below)
3148    
3149    The flag bits are also defined in pcre.h, and are
3150    
3151      PCRE_EXTRA_STUDY_DATA
3152      PCRE_EXTRA_MATCH_LIMIT
3153      PCRE_EXTRA_CALLOUT_DATA
3154    
3155    The pcre_study() function now returns one of these new pcre_extra blocks, with
3156    the actual study data pointed to by the study_data field, and the
3157    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3158    before. That is, this change is entirely upwards-compatible and requires no
3159    change to existing code.
3160    
3161    If you want to pass in additional data to pcre_exec(), you can either place it
3162    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3163    block.
3164    
3165    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3166    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3167    times with different match limits, until it finds the minimum value needed for
3168    pcre_exec() to complete. The value is then output. This can be instructive; for
3169    most simple matches the number is quite small, but for pathological cases it
3170    gets very large very quickly.
3171    
3172    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3173    returns the size of the data block pointed to by the study_data field in a
3174    pcre_extra block, that is, the value that was passed as the argument to
3175    pcre_malloc() when PCRE was getting memory in which to place the information
3176    created by pcre_study(). The fourth argument should point to a size_t variable.
3177    pcretest has been extended so that this information is shown after a successful
3178    pcre_study() call when information about the compiled regex is being displayed.
3179    
3180    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3181    because what follows is always an absolute path. (Later: it turns out that this
3182    is more than cosmetic for MinGW, because it doesn't like empty path
3183    components.)
3184    
3185    49. Some changes have been made to the callout feature (see 28 above):
3186    
3187    (i)  A callout function now has three choices for what it returns:
3188    
3189           0  =>  success, carry on matching
3190         > 0  =>  failure at this point, but backtrack if possible
3191         < 0  =>  serious error, return this value from pcre_exec()
3192    
3193         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3194         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3195         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3196         use by callout functions. It will never be used by PCRE itself.
3197    
3198    (ii) The pcre_extra structure (see 45 above) has a void * field called
3199         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3200         pcre_callout_block structure has a field of the same name. The contents of
3201         the field passed in the pcre_extra structure are passed to the callout
3202         function in the corresponding field in the callout block. This makes it
3203         easier to use the same callout-containing regex from multiple threads. For
3204         testing, the pcretest program has a new data escape
3205    
3206           \C*n        pass the number n (may be negative) as callout_data
3207    
3208         If the callout function in pcretest receives a non-zero value as
3209         callout_data, it returns that value.
3210    
3211    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3212    there were some redundant $(CFLAGS) in commands that are now specified as
3213    $(LINK), which already includes $(CFLAGS).
3214    
3215    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3216    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3217    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3218    one-byte characters throughout. Note that case-insensitive matching applies
3219    only to characters whose values are less than 256. PCRE doesn't support the
3220    notion of cases for higher-valued characters.
3221    
3222    (i)   A character class whose characters are all within 0-255 is handled as
3223          a bit map, and the map is inverted for negative classes. Previously, a
3224          character > 255 always failed to match such a class; however it should
3225          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3226    
3227    (ii)  A negated character class with a single character < 255 is coded as
3228          "not this character" (OP_NOT). This wasn't working properly when the test
3229          character was multibyte, either singly or repeated.
3230    
3231    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3232          mode, for example: \x{100}{2,3}.
3233    
3234    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3235          singly or repeated) now correctly test multibyte characters. However,
3236          PCRE doesn't recognize any characters with values greater than 255 as
3237          digits, spaces, or word characters. Such characters always match \D, \S,
3238          and \W, and never match \d, \s, or \w.
3239    
3240    (v)   Classes may now contain characters and character ranges with values
3241          greater than 255. For example: [ab\x{100}-\x{400}].
3242    
3243    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3244          PCRE in UTF-8 mode.
3245    
3246    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3247    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3248    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3249    value.)
3250    
3251    53. The single man page has become too large. I have therefore split it up into
3252    a number of separate man pages. These also give rise to individual HTML pages;
3253    these are now put in a separate directory, and there is an index.html page that
3254    lists them all. Some hyperlinking between the pages has been installed.
3255    
3256    54. Added convenience functions for handling named capturing parentheses.
3257    
3258    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3259    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3260    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3261    are faulted.
3262    
3263    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3264    calling configure. These values are used when compiling the dftables.c program
3265    which is run to generate the source of the default character tables. They
3266    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3267    you will need to set these values.
3268    
3269    57. Updated the building process for Windows DLL, as provided by Fred Cox.
3270    
3271    
3272    Version 3.9 02-Jan-02
3273    ---------------------
3274    
3275    1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3276    
3277    2. If --disable-static was given, the building process failed when trying to
3278    build pcretest and pcregrep. (For some reason it was using libtool to compile
3279    them, which is not right, as they aren't part of the library.)
3280    
3281    
3282    Version 3.8 18-Dec-01
3283    ---------------------
3284    
3285    1. The experimental UTF-8 code was completely screwed up. It was packing the
3286    bytes in the wrong order. How dumb can you get?
3287    
3288    
3289    Version 3.7 29-Oct-01
3290    ---------------------
3291    
3292    1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3293    This caused pcretest, when used on the test data, to segfault. Unfortunately,
3294    this didn't happen under Solaris 8, where I normally test things.
3295    
3296    2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3297    doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3298    isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3299    this fix an hour or so after the initial 3.7 release.)
3300    
3301    
3302    Version 3.6 23-Oct-01
3303    ---------------------
3304    
3305    1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3306    offsets passed as NULL with zero offset count.
3307    
3308    2. The config.guess and config.sub files had not been updated when I moved to
3309    the latest autoconf.
3310    
3311    
3312    Version 3.5 15-Aug-01
3313    ---------------------
3314    
3315    1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3316    had been forgotten.
3317    
3318    2. By using declared but undefined structures, we can avoid using "void"
3319    definitions in pcre.h while keeping the internal definitions of the structures
3320    private.
3321    
3322    3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3323    user point of view, this means that both static and shared libraries are built
3324    by default, but this can be individually controlled. More of the work of
3325    handling this static/shared cases is now inside libtool instead of PCRE's make
3326    file.
3327    
3328    4. The pcretest utility is now installed along with pcregrep because it is
3329    useful for users (to test regexs) and by doing this, it automatically gets
3330    relinked by libtool. The documentation has been turned into a man page, so
3331    there are now .1, .txt, and .html versions in /doc.
3332    
3333    5. Upgrades to pcregrep:
3334       (i)   Added long-form option names like gnu grep.
3335       (ii)  Added --help to list all options with an explanatory phrase.
3336       (iii) Added -r, --recursive to recurse into sub-directories.
3337       (iv)  Added -f, --file to read patterns from a file.
3338    
3339    6. pcre_exec() was referring to its "code" argument before testing that
3340    argument for NULL (and giving an error if it was NULL).
3341    
3342    7. Upgraded Makefile.in to allow for compiling in a different directory from
3343    the source directory.
3344    
3345    8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3346    options bits, the pointer it was passed was to an int instead of to an unsigned
3347    long int. This mattered only on 64-bit systems.
3348    
3349    9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3350    generated) instead of pcre.in, which it its source. Also made the same change
3351    in several of the .c files.
3352    
3353    10. A new release of gcc defines printf() as a macro, which broke pcretest
3354    because it had an ifdef in the middle of a string argument for printf(). Fixed
3355    by using separate calls to printf().
3356    
3357    11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3358    script, to force use of CR or LF instead of \n in the source. On non-Unix
3359    systems, the value can be set in config.h.
3360    
3361    12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3362    absolute limit. Changed the text of the error message to make this clear, and
3363    likewise updated the man page.
3364    
3365    13. The limit of 99 on the number of capturing subpatterns has been removed.
3366    The new limit is 65535, which I hope will not be a "real" limit.
3367    
3368    
3369    Version 3.4 22-Aug-00
3370    ---------------------
3371    
3372    1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3373    
3374    2. Diagnose condition (?(0) as an error instead of crashing on matching.
3375    
3376    
3377    Version 3.3 01-Aug-00
3378    ---------------------
3379    
3380    1. If an octal character was given, but the value was greater than \377, it
3381    was not getting masked to the least significant bits, as documented. This could
3382    lead to crashes in some systems.
3383    
3384    2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3385    the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3386    
3387    3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3388    These just pass their arguments on to (pcre_free)(), but they are provided
3389    because some uses of PCRE bind it to non-C systems that can call its functions,
3390    but cannot call free() or pcre_free() directly.
3391    
3392    4. Add "make test" as a synonym for "make check". Corrected some comments in
3393    the Makefile.
3394    
3395    5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3396    Makefile.
3397    
3398    6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3399    command called pgrep for grepping around the active processes.
3400    
3401    7. Added the beginnings of support for UTF-8 character strings.
3402    
3403    8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3404    RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3405    the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3406    out for the ar command.)
3407    
3408    
3409    Version 3.2 12-May-00
3410    ---------------------
3411    
3412    This is purely a bug fixing release.
3413    
3414    1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3415    of ZA. This was just one example of several cases that could provoke this bug,
3416    which was introduced by change 9 of version 2.00. The code for breaking
3417    infinite loops after an iteration that matches an empty string was't working
3418    correctly.
3419    
3420    2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3421    when matched against abbab (for example). After matching an empty string, it
3422    wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3423    caused it to match further down the string than it should.
3424    
3425    3. The code contained an inclusion of sys/types.h. It isn't clear why this
3426    was there because it doesn't seem to be needed, and it causes trouble on some
3427    systems, as it is not a Standard C header. It has been removed.
3428    
3429    4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3430    were reported on the Macintosh. The changes were from
3431    
3432      while ((c = *(++ptr)) != 0 && c != '\n');
3433    to
3434      while ((c = *(++ptr)) != 0 && c != '\n') ;
3435    
3436    Totally extraordinary, but if that's what it takes...
3437    
3438    5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3439    available. Added HAVE_BCOPY and an autoconf test for it; if neither
3440    HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3441    assumes the way PCRE uses memmove() (always moving upwards).
3442    
3443    6. PCRE is being used in one environment where strchr() is not available. There
3444    was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3445    faster code anyway.
3446    
3447    
3448    Version 3.1 09-Feb-00
3449    ---------------------
3450    
3451    The only change in this release is the fixing of some bugs in Makefile.in for
3452    the "install" target:
3453    
3454    (1) It was failing to install pcreposix.h.
3455    
3456    (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3457    
3458    
3459    Version 3.0 01-Feb-00
3460    ---------------------
3461    
3462    1. Add support for the /+ modifier to perltest (to output $` like it does in
3463    pcretest).
3464    
3465    2. Add support for the /g modifier to perltest.
3466    
3467    3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3468    matches null strings.
3469    
3470    4. Fix perltest so that it doesn't do unwanted things when fed an empty
3471    pattern. Perl treats empty patterns specially - it reuses the most recent
3472    pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3473    effect.
3474    
3475    5. The POSIX interface was broken in that it was just handing over the POSIX
3476    captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3477    required a bigger vector, with some working space on the end. This means that
3478    the POSIX wrapper now has to get and free some memory, and copy the results.
3479    
3480    6. Added some simple autoconf support, placing the test data and the
3481    documentation in separate directories, re-organizing some of the
3482    information files, and making it build pcre-config (a GNU standard). Also added
3483    libtool support for building PCRE as a shared library, which is now the
3484    default.
3485    
3486    7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3487    09 are not valid octal constants. Single digits will be used for minor values
3488    less than 10.
3489    
3490    8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3491    existing programs that set these in the POSIX interface can use PCRE without
3492    modification.
3493    
3494    9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3495    return all that pcre_info() returns, plus additional data. The pcre_info()
3496    function is retained for compatibility, but is considered to be obsolete.
3497    
3498    10. Added experimental recursion feature (?R) to handle one common case that
3499    Perl 5.6 will be able to do with (?p{...}).
3500    
3501    11. Added support for POSIX character classes like [:alpha:], which Perl is
3502    adopting.
3503    
3504    
3505    Version 2.08 31-Aug-99
3506    ----------------------
3507    
3508    1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3509    trying to match at the startoffset position, but instead was moving forward to
3510    the next newline as if a previous match had failed.
3511    
3512    2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3513    and could get into a loop if a null string was matched other than at the start
3514    of the subject.
3515    
3516    3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3517    be distinguished at compile time, and for completeness also added PCRE_DATE.
3518    
3519    5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3520    in GnuWin32 environments.
3521    
3522    
3523    Version 2.07 29-Jul-99
3524    ----------------------
3525    
3526    1. The documentation is now supplied in plain text form and HTML as well as in
3527    the form of man page sources.
3528    
3529    2. C++ compilers don't like assigning (void *) values to other pointer types.
3530    In particular this affects malloc(). Although there is no problem in Standard
3531    C, I've put in casts to keep C++ compilers happy.
3532    
3533    3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3534    should be (const char *).
3535    
3536    4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3537    be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3538    However, I haven't made this a standard facility. The documentation doesn't
3539    mention it, and the Makefile doesn't support it.
3540    
3541    5. The Makefile now contains an "install" target, with editable destinations at
3542    the top of the file. The pcretest program is not installed.
3543    
3544    6. pgrep -V now gives the PCRE version number and date.
3545    
3546    7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3547    causing the entire string to be ignored, instead of just the last character.
3548    
3549    8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3550    non-matching string, it can take a very, very long time, even for strings of
3551    quite modest length, because of the nested recursion. PCRE now does better in
3552    some of these cases. It does this by remembering the last required literal
3553    character in the pattern, and pre-searching the subject to ensure it is present
3554    before running the real match. In other words, it applies a heuristic to detect
3555    some types of certain failure quickly, and in the above example, if presented
3556    with a string that has no trailing " it gives "no match" very quickly.
3557    
3558    9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3559    other alternatives are tried instead.
3560    
3561    
3562    Version 2.06 09-Jun-99
3563    ----------------------
3564    
3565    1. Change pcretest's output for amount of store used to show just the code
3566    space, because the remainder (the data block) varies in size between 32-bit and
3567    64-bit systems.
3568    
3569    2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3570    start matching at. This allows lookbehinds to work when searching for multiple
3571    occurrences in a string.
3572    
3573    3. Added additional options to pcretest for testing multiple occurrences:
3574    
3575       /+   outputs the rest of the string that follows a match
3576       /g   loops for multiple occurrences, using the new startoffset argument
3577       /G   loops for multiple occurrences by passing an incremented pointer
3578    
3579    4. PCRE wasn't doing the "first character" optimization for patterns starting
3580    with \b or \B, though it was doing it for other lookbehind assertions. That is,
3581    it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3582    the letter 'x'. On long subject strings, this gives a significant speed-up.
3583    
3584    
3585    Version 2.05 21-Apr-99
3586    ----------------------
3587    
3588    1. Changed the type of magic_number from int to long int so that it works
3589    properly on 16-bit systems.
3590    
3591    2. Fixed a bug which caused patterns starting with .* not to work correctly
3592    when the subject string contained newline characters. PCRE was assuming
3593    anchoring for such patterns in all cases, which is not correct because .* will
3594    not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3595    DOTALL is set at top level; otherwise it knows that patterns starting with .*
3596    must be retried after every newline in the subject.
3597    
3598    
3599    Version 2.04 18-Feb-99
3600    ----------------------
3601    
3602    1. For parenthesized subpatterns with repeats whose minimum was zero, the
3603    computation of the store needed to hold the pattern was incorrect (too large).
3604    If such patterns were nested a few deep, this could multiply and become a real
3605    problem.
3606    
3607    2. Added /M option to pcretest to show the memory requirement of a specific
3608    pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3609    
3610    3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3611    compiled in such a way that the backtracking after subsequent failure was
3612    pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3613    ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3614    
3615    
3616    Version 2.03 02-Feb-99
3617    ----------------------
3618    
3619    1. Fixed typo and small mistake in man page.
3620    
3621    2. Added 4th condition (GPL supersedes if conflict) and created separate
3622    LICENCE file containing the conditions.
3623    
3624    3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3625    Perl, that is the internal \ allows the delimiter to be included in the
3626    pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3627    the final delimiter, add \ to the end of the pattern (to test the error).
3628    
3629    4. Added the convenience functions for extracting substrings after a successful
3630    match. Updated pcretest to make it able to test these functions.
3631    
3632    
3633    Version 2.02 14-Jan-99
3634    ----------------------
3635    
3636    1. Initialized the working variables associated with each extraction so that
3637    their saving and restoring doesn't refer to uninitialized store.
3638    
3639    2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3640    compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3641    fix the problem.
3642    
3643    3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3644    calls, and wasn't printing the correct value for compiling calls. Increased the
3645    default value of LOOPREPEAT, and the number of significant figures in the
3646    times.
3647    
3648    4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3649    
3650    5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3651    a building problem on Windows NT with a FAT file system.
3652    
3653    
3654    Version 2.01 21-Oct-98
3655    ----------------------
3656    
3657    1. Changed the API for pcre_compile() to allow for the provision of a pointer
3658    to character tables built by pcre_maketables() in the current locale. If NULL
3659    is passed, the default tables are used.
3660    
3661    
3662    Version 2.00 24-Sep-98
3663    ----------------------
3664    
3665    1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3666    it any more.
3667    
3668    2. Allow quantification of (?>) groups, and make it work correctly.
3669    
3670    3. The first character computation wasn't working for (?>) groups.
3671    
3672    4. Correct the implementation of \Z (it is permitted to match on the \n at the
3673    end of the subject) and add 5.005's \z, which really does match only at the
3674    very end of the subject.
3675    
3676    5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3677    
3678    6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3679    DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3680    localized options. All options to pcre_study() were also removed.
3681    
3682    7. Add other new features from 5.005:
3683    
3684       $(?<=           positive lookbehind
3685       $(?<!           negative lookbehind
3686       (?imsx-imsx)    added the unsetting capability
3687                       such a setting is global if at outer level; local otherwise
3688       (?imsx-imsx:)   non-capturing groups with option setting
3689       (?(cond)re|re)  conditional pattern matching
3690    
3691       A backreference to itself in a repeated group matches the previous
3692       captured string.
3693    
3694    8. General tidying up of studying (both automatic and via "study")
3695    consequential on the addition of new assertions.
3696    
3697    9. As in 5.005, unlimited repeated groups that could match an empty substring
3698    are no longer faulted at compile time. Instead, the loop is forcibly broken at
3699    runtime if any iteration does actually match an empty substring.
3700    
3701    10. Include the RunTest script in the distribution.
3702    
3703    11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3704    discrepancies, some of which were old and were also with respect to 5.004. They
3705    have now been fixed.
3706    
3707    
3708    Version 1.09 28-Apr-98
3709    ----------------------
3710    
3711    1. A negated single character class followed by a quantifier with a minimum
3712    value of one (e.g.  [^x]{1,6}  ) was not compiled correctly. This could lead to
3713    program crashes, or just wrong answers. This did not apply to negated classes
3714    containing more than one character, or to minima other than one.
3715    
3716    
3717    Version 1.08 27-Mar-98
3718    ----------------------
3719    
3720    1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3721    
3722    2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3723    latter must appear before anything that relies on it in the pattern.
3724    
3725    
3726    Version 1.07 16-Feb-98
3727    ----------------------
3728    
3729    1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3730    repeat of a potentially empty string).
3731    
3732    
3733   &nb