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

Diff of /code/trunk/ChangeLog

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

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