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

Diff of /code/trunk/ChangeLog

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

revision 35 by nigel, Sat Feb 24 21:39:05 2007 UTC revision 584 by ph10, Wed Jan 12 11:41:55 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.12 12-Jan-2011
5    ------------------------
6    
7    1.  Fixed some typos in the markup of the man pages, and wrote a script that
8        checks for such things as part of the documentation building process.
9    
10    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
11        --match-limit and --recursion-limit options (added for 8.11). In
12        particular, this made one of the standard tests crash. (The integer value
13        went into the wrong half of a long int.)
14    
15    
16    Version 8.11 10-Dec-2010
17    ------------------------
18    
19    1.  (*THEN) was not working properly if there were untried alternatives prior
20        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
21        backtracked to try for "b" instead of moving to the next alternative branch
22        at the same level (in this case, to look for "c"). The Perl documentation
23        is clear that when (*THEN) is backtracked onto, it goes to the "next
24        alternative in the innermost enclosing group".
25    
26    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
27        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
28        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
29        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
30        (*THEN).
31    
32    3.  If \s appeared in a character class, it removed the VT character from
33        the class, even if it had been included by some previous item, for example
34        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
35        of \s, but is part of the POSIX "space" class.)
36    
37    4.  A partial match never returns an empty string (because you can always
38        match an empty string at the end of the subject); however the checking for
39        an empty string was starting at the "start of match" point. This has been
40        changed to the "earliest inspected character" point, because the returned
41        data for a partial match starts at this character. This means that, for
42        example, /(?<=abc)def/ gives a partial match for the subject "abc"
43        (previously it gave "no match").
44    
45    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
46        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
47        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
48        has an implication that the given string is incomplete (because a partial
49        match is preferred over a full match). For this reason, these items now
50        give a partial match in this situation. [Aside: previously, the one case
51        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
52        match rather than a full match, which was wrong by the old rules, but is
53        now correct.]
54    
55    6.  There was a bug in the handling of #-introduced comments, recognized when
56        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
57        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
58        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
59        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
60        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
61        places in pcre_compile().
62    
63    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
64        comments when looking ahead for named forward references to subpatterns,
65        the only newline sequence it recognized was NL. It now handles newlines
66        according to the set newline convention.
67    
68    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
69        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
70        cater for a lack of strerror(). These oversights have been fixed.
71    
72    9.  Added --match-limit and --recursion-limit to pcregrep.
73    
74    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
75    
76    11. When the -o option was used, pcregrep was setting a return code of 1, even
77        when matches were found, and --line-buffered was not being honoured.
78    
79    12. Added an optional parentheses number to the -o and --only-matching options
80        of pcregrep.
81    
82    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
83        can match an empty string. The code to do it in pcretest and pcredemo
84        needed fixing:
85    
86        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
87            only one byte after an empty string match just before CRLF (this case
88            just got forgotten; "any" and "anycrlf" were OK).
89    
90        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
91            mode when an empty string match preceded an ASCII character followed by
92            a non-ASCII character. (The code for advancing by one character rather
93            than one byte was nonsense.)
94    
95        (c) The pcredemo.c sample program did not have any code at all to handle
96            the cases when CRLF is a valid newline sequence.
97    
98    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
99        as a starting offset was within the subject string. There is now a new
100        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
101        negative or greater than the length of the string. In order to test this,
102        pcretest is extended to allow the setting of negative starting offsets.
103    
104    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
105        starting offset points to the beginning of a UTF-8 character was
106        unnecessarily clumsy. I tidied it up.
107    
108    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
109        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
110    
111    17. Nobody had reported that the --include_dir option, which was added in
112        release 7.7 should have been called --include-dir (hyphen, not underscore)
113        for compatibility with GNU grep. I have changed it to --include-dir, but
114        left --include_dir as an undocumented synonym, and the same for
115        --exclude-dir, though that is not available in GNU grep, at least as of
116        release 2.5.4.
117    
118    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
119        characters from a string of bytes) have been redefined so as not to use
120        loops, in order to improve performance in some environments. At the same
121        time, I abstracted some of the common code into auxiliary macros to save
122        repetition (this should not affect the compiled code).
123    
124    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
125        compile-time error is now given if \c is not followed by an ASCII
126        character, that is, a byte less than 128. (In EBCDIC mode, the code is
127        different, and any byte value is allowed.)
128    
129    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
130        START_OPTIMIZE option, which is now allowed at compile time - but just
131        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
132        to pcregrep and other applications that have no direct access to PCRE
133        options. The new /Y option in pcretest sets this option when calling
134        pcre_compile().
135    
136    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
137        back references. Groups containing recursive back references were forced to
138        be atomic by that change, but in the case of named groups, the amount of
139        memory required was incorrectly computed, leading to "Failed: internal
140        error: code overflow". This has been fixed.
141    
142    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
143        pcretest.c, to avoid build problems in some Borland environments.
144    
145    
146    Version 8.10 25-Jun-2010
147    ------------------------
148    
149    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
150        THEN.
151    
152    2.  (*ACCEPT) was not working when inside an atomic group.
153    
154    3.  Inside a character class, \B is treated as a literal by default, but
155        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
156        causes the error). The code is unchanged, but I tidied the documentation.
157    
158    4.  Inside a character class, PCRE always treated \R and \X as literals,
159        whereas Perl faults them if its -w option is set. I have changed PCRE so
160        that it faults them when PCRE_EXTRA is set.
161    
162    5.  Added support for \N, which always matches any character other than
163        newline. (It is the same as "." when PCRE_DOTALL is not set.)
164    
165    6.  When compiling pcregrep with newer versions of gcc which may have
166        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
167        declared with attribute warn_unused_result" were given. Just casting the
168        result to (void) does not stop the warnings; a more elaborate fudge is
169        needed. I've used a macro to implement this.
170    
171    7.  Minor change to pcretest.c to avoid a compiler warning.
172    
173    8.  Added four artifical Unicode properties to help with an option to make
174        \s etc use properties (see next item). The new properties are: Xan
175        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
176    
177    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
178        use Unicode properties. (*UCP) at the start of a pattern can be used to set
179        this option. Modified pcretest to add /W to test this facility. Added
180        REG_UCP to make it available via the POSIX interface.
181    
182    10. Added --line-buffered to pcregrep.
183    
184    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
185        studied, and the match started with a letter with a code point greater than
186        127 whose first byte was different to the first byte of the other case of
187        the letter, the other case of this starting letter was not recognized
188        (#976).
189    
190    12. If a pattern that was studied started with a repeated Unicode property
191        test, for example, \p{Nd}+, there was the theoretical possibility of
192        setting up an incorrect bitmap of starting bytes, but fortunately it could
193        not have actually happened in practice until change 8 above was made (it
194        added property types that matched character-matching opcodes).
195    
196    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
197        possible starting bytes for non-anchored patterns.
198    
199    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
200        \R, and also a number of cases that involve Unicode properties, both
201        explicit and implicit when PCRE_UCP is set.
202    
203    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
204        input, it could crash or give wrong results if characters with values
205        greater than 0xc0 were present in the subject string. (Detail: it assumed
206        UTF-8 input when processing these items.)
207    
208    16. Added a lot of (int) casts to avoid compiler warnings in systems where
209        size_t is 64-bit (#991).
210    
211    17. Added a check for running out of memory when PCRE is compiled with
212        --disable-stack-for-recursion (#990).
213    
214    18. If the last data line in a file for pcretest does not have a newline on
215        the end, a newline was missing in the output.
216    
217    19. The default pcre_chartables.c file recognizes only ASCII characters (values
218        less than 128) in its various bitmaps. However, there is a facility for
219        generating tables according to the current locale when PCRE is compiled. It
220        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
221        characters, are recognized by isspace() and therefore were getting set in
222        these tables, and indeed these tables seem to approximate to ISO 8859. This
223        caused a problem in UTF-8 mode when pcre_study() was used to create a list
224        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
225        which of course cannot start UTF-8 characters. I have changed the code so
226        that only real ASCII characters (less than 128) and the correct starting
227        bytes for UTF-8 encodings are set for characters greater than 127 when in
228        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
229        altogether.)
230    
231    20. Added the /T option to pcretest so as to be able to run tests with non-
232        standard character tables, thus making it possible to include the tests
233        used for 19 above in the standard set of tests.
234    
235    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
236        reference to a subpattern the other side of a comment that contains an
237        opening parenthesis caused either an internal compiling error, or a
238        reference to the wrong subpattern.
239    
240    
241    Version 8.02 19-Mar-2010
242    ------------------------
243    
244    1.  The Unicode data tables have been updated to Unicode 5.2.0.
245    
246    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
247        configured.
248    
249    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
250        original author of that file, following a query about its status.
251    
252    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
253        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
254    
255    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
256        quantifier applied to a forward-referencing subroutine call, could compile
257        incorrect code or give the error "internal error: previously-checked
258        referenced subpattern not found".
259    
260    6.  Both MS Visual Studio and Symbian OS have problems with initializing
261        variables to point to external functions. For these systems, therefore,
262        pcre_malloc etc. are now initialized to local functions that call the
263        relevant global functions.
264    
265    7.  There were two entries missing in the vectors called coptable and poptable
266        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
267        I've fixed the data, and added a kludgy way of testing at compile time that
268        the lengths are correct (equal to the number of opcodes).
269    
270    8.  Following on from 7, I added a similar kludge to check the length of the
271        eint vector in pcreposix.c.
272    
273    9.  Error texts for pcre_compile() are held as one long string to avoid too
274        much relocation at load time. To find a text, the string is searched,
275        counting zeros. There was no check for running off the end of the string,
276        which could happen if a new error number was added without updating the
277        string.
278    
279    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
280    
281    11. \K was not working if it appeared in an atomic group or in a group that
282        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
283        \K is "not well defined" if used in an assertion. PCRE now accepts it if
284        the assertion is positive, but not if it is negative.
285    
286    12. Change 11 fortuitously reduced the size of the stack frame used in the
287        "match()" function of pcre_exec.c by one pointer. Forthcoming
288        implementation of support for (*MARK) will need an extra pointer on the
289        stack; I have reserved it now, so that the stack frame size does not
290        decrease.
291    
292    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
293        item in branch that calls a recursion is a subroutine call - as in the
294        second branch in the above example - was incorrectly given the compile-
295        time error "recursive call could loop indefinitely" because pcre_compile()
296        was not correctly checking the subroutine for matching a non-empty string.
297    
298    14. The checks for overrunning compiling workspace could trigger after an
299        overrun had occurred. This is a "should never occur" error, but it can be
300        triggered by pathological patterns such as hundreds of nested parentheses.
301        The checks now trigger 100 bytes before the end of the workspace.
302    
303    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
304    
305    
306    Version 8.01 19-Jan-2010
307    ------------------------
308    
309    1.  If a pattern contained a conditional subpattern with only one branch (in
310        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
311        computed the wrong minimum data length (which is of course zero for such
312        subpatterns). This could cause incorrect "no match" results.
313    
314    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
315        the pattern is reset in the first branch, pcre_compile() failed with
316        "internal error: code overflow at offset...". This happened only when
317        the reset was to the original external option setting. (An optimization
318        abstracts leading options settings into an external setting, which was the
319        cause of this.)
320    
321    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
322        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
323        assertion pattern did not match (meaning that the assertion was true), it
324        was incorrectly treated as false if the SKIP had been reached during the
325        matching. This also applied to assertions used as conditions.
326    
327    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
328        assertion subpattern, including such a pattern used as a condition,
329        unpredictable results occurred, instead of the error return
330        PCRE_ERROR_DFA_UITEM.
331    
332    5.  The C++ GlobalReplace function was not working like Perl for the special
333        situation when an empty string is matched. It now does the fancy magic
334        stuff that is necessary.
335    
336    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
337        removed. (These were left over from very, very early versions of PCRE.)
338    
339    7.  Some cosmetic changes to the code to make life easier when compiling it
340        as part of something else:
341    
342        (a) Change DEBUG to PCRE_DEBUG.
343    
344        (b) In pcre_compile(), rename the member of the "branch_chain" structure
345            called "current" as "current_branch", to prevent a collision with the
346            Linux macro when compiled as a kernel module.
347    
348        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
349            prevent a collision with the Linux macro when compiled as a kernel
350            module.
351    
352    8.  In pcre_compile() there are some checks for integer overflows that used to
353        cast potentially large values to (double). This has been changed to that
354        when building, a check for int64_t is made, and if it is found, it is used
355        instead, thus avoiding the use of floating point arithmetic. (There is no
356        other use of FP in PCRE.) If int64_t is not found, the fallback is to
357        double.
358    
359    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
360        2005 (difference between two addresses compared to an unsigned value).
361    
362    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
363        custom one, because of the following reported problem in Windows:
364    
365          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
366              under Win32.
367          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
368              therefore missing the function definition.
369          - The compiler thus generates a "C" signature for the test function.
370          - The linker fails to find the "C" function.
371          - PCRE fails to configure if asked to do so against libbz2.
372    
373    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
374        messages were output:
375    
376          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
377          rerunning libtoolize, to keep the correct libtool macros in-tree.
378          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
379    
380        I have done both of these things.
381    
382    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
383        most of the time, it *can* run out if it is given a pattern that contains a
384        runaway infinite recursion. I updated the discussion in the pcrestack man
385        page.
386    
387    13. Now that we have gone to the x.xx style of version numbers, the minor
388        version may start with zero. Using 08 or 09 is a bad idea because users
389        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
390        interpreted as invalid octal numbers. I've updated the previous comment in
391        configure.ac, and also added a check that gives an error if 08 or 09 are
392        used.
393    
394    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
395        causing partial matching to fail when the end of the subject matched \W
396        in a UTF-8 pattern where \W was quantified with a minimum of 3.
397    
398    15. There were some discrepancies between the declarations in pcre_internal.h
399        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
400        their definitions. The declarations used "const uschar *" and the
401        definitions used USPTR. Even though USPTR is normally defined as "const
402        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
403        reported that: "This difference in casting confuses some C++ compilers, for
404        example, SunCC recognizes above declarations as different functions and
405        generates broken code for hbpcre." I have changed the declarations to use
406        USPTR.
407    
408    16. GNU libtool is named differently on some systems. The autogen.sh script now
409        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
410        (FreeBSD).
411    
412    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
413        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
414        comment: "Figure out how to create a longlong from a string: strtoll and
415        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
416        instance, but it only takes 2 args instead of 3!"
417    
418    18. A subtle bug concerned with back references has been fixed by a change of
419        specification, with a corresponding code fix. A pattern such as
420        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
421        refers, was giving matches when it shouldn't. For example, xa=xaaa would
422        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
423        same bug. Such groups have to be quantified to be useful, or contained
424        inside another quantified group. (If there's no repetition, the reference
425        can never match.) The problem arises because, having left the group and
426        moved on to the rest of the pattern, a later failure that backtracks into
427        the group uses the captured value from the final iteration of the group
428        rather than the correct earlier one. I have fixed this in PCRE by forcing
429        any group that contains a reference to itself to be an atomic group; that
430        is, there cannot be any backtracking into it once it has completed. This is
431        similar to recursive and subroutine calls.
432    
433    
434    Version 8.00 19-Oct-09
435    ----------------------
436    
437    1.  The table for translating pcre_compile() error codes into POSIX error codes
438        was out-of-date, and there was no check on the pcre_compile() error code
439        being within the table. This could lead to an OK return being given in
440        error.
441    
442    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
443        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
444        in a Windows environment.
445    
446    3.  The pcregrep --count option prints the count for each file even when it is
447        zero, as does GNU grep. However, pcregrep was also printing all files when
448        --files-with-matches was added. Now, when both options are given, it prints
449        counts only for those files that have at least one match. (GNU grep just
450        prints the file name in this circumstance, but including the count seems
451        more useful - otherwise, why use --count?) Also ensured that the
452        combination -clh just lists non-zero counts, with no names.
453    
454    4.  The long form of the pcregrep -F option was incorrectly implemented as
455        --fixed_strings instead of --fixed-strings. This is an incompatible change,
456        but it seems right to fix it, and I didn't think it was worth preserving
457        the old behaviour.
458    
459    5.  The command line items --regex=pattern and --regexp=pattern were not
460        recognized by pcregrep, which required --regex pattern or --regexp pattern
461        (with a space rather than an '='). The man page documented the '=' forms,
462        which are compatible with GNU grep; these now work.
463    
464    6.  No libpcreposix.pc file was created for pkg-config; there was just
465        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
466    
467    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
468        when UCP support is not needed, by modifying the Python script that
469        generates it from Unicode data files. This should not matter if the module
470        is correctly used as a library, but I received one complaint about 50K of
471        unwanted data. My guess is that the person linked everything into his
472        program rather than using a library. Anyway, it does no harm.
473    
474    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
475        was a minimum greater than 1 for a wide character in a possessive
476        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
477        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
478        character. Chaos in the form of incorrect output or a compiling loop could
479        result.
480    
481    9.  The restrictions on what a pattern can contain when partial matching is
482        requested for pcre_exec() have been removed. All patterns can now be
483        partially matched by this function. In addition, if there are at least two
484        slots in the offset vector, the offset of the earliest inspected character
485        for the match and the offset of the end of the subject are set in them when
486        PCRE_ERROR_PARTIAL is returned.
487    
488    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
489        synonymous with PCRE_PARTIAL, for backwards compatibility, and
490        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
491        and may be more useful for multi-segment matching.
492    
493    11. Partial matching with pcre_exec() is now more intuitive. A partial match
494        used to be given if ever the end of the subject was reached; now it is
495        given only if matching could not proceed because another character was
496        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
497        string "Z", which now yields "no match" instead of "partial match". In the
498        case of pcre_dfa_exec(), "no match" is given if every matching path for the
499        final character ended with (*FAIL).
500    
501    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
502        if the pattern had a "must contain" character that was already found in the
503        earlier partial match, unless partial matching was again requested. For
504        example, with the pattern /dog.(body)?/, the "must contain" character is
505        "g". If the first part-match was for the string "dog", restarting with
506        "sbody" failed. This bug has been fixed.
507    
508    13. The string returned by pcre_dfa_exec() after a partial match has been
509        changed so that it starts at the first inspected character rather than the
510        first character of the match. This makes a difference only if the pattern
511        starts with a lookbehind assertion or \b or \B (\K is not supported by
512        pcre_dfa_exec()). It's an incompatible change, but it makes the two
513        matching functions compatible, and I think it's the right thing to do.
514    
515    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
516        so that the demonstration program is easily available in environments where
517        PCRE has not been installed from source.
518    
519    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
520        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
521        library.
522    
523    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
524        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
525        is not the first non-POSIX option to be added. Clearly some people find
526        these options useful.
527    
528    17. If a caller to the POSIX matching function regexec() passes a non-zero
529        value for nmatch with a NULL value for pmatch, the value of
530        nmatch is forced to zero.
531    
532    18. RunGrepTest did not have a test for the availability of the -u option of
533        the diff command, as RunTest does. It now checks in the same way as
534        RunTest, and also checks for the -b option.
535    
536    19. If an odd number of negated classes containing just a single character
537        interposed, within parentheses, between a forward reference to a named
538        subpattern and the definition of the subpattern, compilation crashed with
539        an internal error, complaining that it could not find the referenced
540        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
541        [The bug was that it was starting one character too far in when skipping
542        over the character class, thus treating the ] as data rather than
543        terminating the class. This meant it could skip too much.]
544    
545    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
546        /g option in pcretest when the pattern contains \K, which makes it possible
547        to have an empty string match not at the start, even when the pattern is
548        anchored. Updated pcretest and pcredemo to use this option.
549    
550    21. If the maximum number of capturing subpatterns in a recursion was greater
551        than the maximum at the outer level, the higher number was returned, but
552        with unset values at the outer level. The correct (outer level) value is
553        now given.
554    
555    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
556        PCRE did not set those parentheses (unlike Perl). I have now found a way to
557        make it do so. The string so far is captured, making this feature
558        compatible with Perl.
559    
560    23. The tests have been re-organized, adding tests 11 and 12, to make it
561        possible to check the Perl 5.10 features against Perl 5.10.
562    
563    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
564        pattern matches a fixed length string. PCRE did not allow this; now it
565        does. Neither allows recursion.
566    
567    25. I finally figured out how to implement a request to provide the minimum
568        length of subject string that was needed in order to match a given pattern.
569        (It was back references and recursion that I had previously got hung up
570        on.) This code has now been added to pcre_study(); it finds a lower bound
571        to the length of subject needed. It is not necessarily the greatest lower
572        bound, but using it to avoid searching strings that are too short does give
573        some useful speed-ups. The value is available to calling programs via
574        pcre_fullinfo().
575    
576    26. While implementing 25, I discovered to my embarrassment that pcretest had
577        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
578        study optimizations had never been tested with that matching function.
579        Oops. What is worse, even when it was passed study data, there was a bug in
580        pcre_dfa_exec() that meant it never actually used it. Double oops. There
581        were also very few tests of studied patterns with pcre_dfa_exec().
582    
583    27. If (?| is used to create subpatterns with duplicate numbers, they are now
584        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
585        on the other side of the coin, they are no longer allowed to have different
586        names, because these cannot be distinguished in PCRE, and this has caused
587        confusion. (This is a difference from Perl.)
588    
589    28. When duplicate subpattern names are present (necessarily with different
590        numbers, as required by 27 above), and a test is made by name in a
591        conditional pattern, either for a subpattern having been matched, or for
592        recursion in such a pattern, all the associated numbered subpatterns are
593        tested, and the overall condition is true if the condition is true for any
594        one of them. This is the way Perl works, and is also more like the way
595        testing by number works.
596    
597    
598    Version 7.9 11-Apr-09
599    ---------------------
600    
601    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
602        (pcretest), all targets were linked against these libraries. This included
603        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
604        libraries. This caused unwanted dependencies to be created. This problem
605        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
606        pcretest is linked with readline.
607    
608    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
609        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
610        moved outside it again, because FALSE and TRUE are already defined in AIX,
611        but BOOL is not.
612    
613    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
614        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
615    
616    4.  The pcregrep documentation said spaces were inserted as well as colons (or
617        hyphens) following file names and line numbers when outputting matching
618        lines. This is not true; no spaces are inserted. I have also clarified the
619        wording for the --colour (or --color) option.
620    
621    5.  In pcregrep, when --colour was used with -o, the list of matching strings
622        was not coloured; this is different to GNU grep, so I have changed it to be
623        the same.
624    
625    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
626        each matching line was coloured. Now it goes on to look for further matches
627        of any of the test patterns, which is the same behaviour as GNU grep.
628    
629    7.  A pattern that could match an empty string could cause pcregrep to loop; it
630        doesn't make sense to accept an empty string match in pcregrep, so I have
631        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
632        seems to be how GNU grep behaves.
633    
634    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
635        start or after a newline", because the conditional assertion was not being
636        correctly handled. The rule now is that both the assertion and what follows
637        in the first alternative must satisfy the test.
638    
639    9.  If auto-callout was enabled in a pattern with a conditional group whose
640        condition was an assertion, PCRE could crash during matching, both with
641        pcre_exec() and pcre_dfa_exec().
642    
643    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
644        used for matching.
645    
646    11. Unicode property support in character classes was not working for
647        characters (bytes) greater than 127 when not in UTF-8 mode.
648    
649    12. Added the -M command line option to pcretest.
650    
651    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
652    
653    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
654    
655    16. Added comments and documentation about mis-use of no_arg in the C++
656        wrapper.
657    
658    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
659        from Martin Jerabek that uses macro names for all relevant character and
660        string constants.
661    
662    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
663        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
664        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
665        these, but not everybody uses configure.
666    
667    19. A conditional group that had only one branch was not being correctly
668        recognized as an item that could match an empty string. This meant that an
669        enclosing group might also not be so recognized, causing infinite looping
670        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
671        with the subject "ab", where knowledge that the repeated group can match
672        nothing is needed in order to break the loop.
673    
674    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
675        exec(), but without supplying a callout function, matching went wrong.
676    
677    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
678        leak if the size of the offset vector was greater than 30. When the vector
679        is smaller, the saved offsets during recursion go onto a local stack
680        vector, but for larger vectors malloc() is used. It was failing to free
681        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
682        error, in fact).
683    
684    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
685        heapframe that is used only when UTF-8 support is enabled. This caused no
686        problem, but was untidy.
687    
688    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
689        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
690        included within another project.
691    
692    24. Steven Van Ingelgem's patches to add more options to the CMake support,
693        slightly modified by me:
694    
695          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
696              not building pcregrep.
697    
698          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
699              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
700    
701    25. Forward references, both numeric and by name, in patterns that made use of
702        duplicate group numbers, could behave incorrectly or give incorrect errors,
703        because when scanning forward to find the reference group, PCRE was not
704        taking into account the duplicate group numbers. A pattern such as
705        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
706    
707    26. Changed a few more instances of "const unsigned char *" to USPTR, making
708        the feature of a custom pointer more persuasive (as requested by a user).
709    
710    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
711        pcretest.c, inside #ifndefs, because it seems they are sometimes already
712        pre-defined.
713    
714    28. Added support for (*UTF8) at the start of a pattern.
715    
716    29. Arrange for flags added by the "release type" setting in CMake to be shown
717        in the configuration summary.
718    
719    
720    Version 7.8 05-Sep-08
721    ---------------------
722    
723    1.  Replaced UCP searching code with optimized version as implemented for Ad
724        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
725        stage table and inline lookup instead of a function, giving speed ups of 2
726        to 5 times on some simple patterns that I tested. Permission was given to
727        distribute the MultiStage2.py script that generates the tables (it's not in
728        the tarball, but is in the Subversion repository).
729    
730    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
731        scripts.
732    
733    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
734        a group with a zero qualifier. The result of the study could be incorrect,
735        or the function might crash, depending on the pattern.
736    
737    4.  Caseless matching was not working for non-ASCII characters in back
738        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
739        It now works when Unicode Property Support is available.
740    
741    5.  In pcretest, an escape such as \x{de} in the data was always generating
742        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
743        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
744        truncation.
745    
746    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
747    
748    7.  Added two (int) casts to pcregrep when printing the difference of two
749        pointers, in case they are 64-bit values.
750    
751    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
752        test 2 if it fails.
753    
754    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
755        and a #define of that name to empty if it is not externally set. This is to
756        allow users of MSVC to set it if necessary.
757    
758    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
759        the convenience functions in the pcre_get.c source file.
760    
761    11. An option change at the start of a pattern that had top-level alternatives
762        could cause overwriting and/or a crash. This command provoked a crash in
763        some environments:
764    
765          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
766    
767        This potential security problem was recorded as CVE-2008-2371.
768    
769    12. For a pattern where the match had to start at the beginning or immediately
770        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
771        pcre_dfa_exec() could read past the end of the passed subject if there was
772        no match. To help with detecting such bugs (e.g. with valgrind), I modified
773        pcretest so that it places the subject at the end of its malloc-ed buffer.
774    
775    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
776        exec() might read past the end of the data buffer in UTF-8 mode.
777    
778    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
779        the data contained the byte 0x85 as part of a UTF-8 character within its
780        first line. This applied both to normal and DFA matching.
781    
782    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
783        /^[^d]*?$/8 failed to match "abc".
784    
785    16. Added a missing copyright notice to pcrecpp_internal.h.
786    
787    17. Make it more clear in the documentation that values returned from
788        pcre_exec() in ovector are byte offsets, not character counts.
789    
790    18. Tidied a few places to stop certain compilers from issuing warnings.
791    
792    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
793        supplied by Stefan Weber. I made a further small update for 7.8 because
794        there is a change of source arrangements: the pcre_searchfuncs.c module is
795        replaced by pcre_ucd.c.
796    
797    
798    Version 7.7 07-May-08
799    ---------------------
800    
801    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
802        a string to a long long, pretend we don't even have a long long." This is
803        done by checking for the strtoq, strtoll, and _strtoi64 functions.
804    
805    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
806        pre-7.6 versions, which defined a global no_arg variable instead of putting
807        it in the RE class. (See also #8 below.)
808    
809    3.  Remove a line of dead code, identified by coverity and reported by Nuno
810        Lopes.
811    
812    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
813    
814        (1) The include/exclude patterns were being applied to the whole pathnames
815            of files, instead of just to the final components.
816    
817        (2) If there was more than one level of directory, the subdirectories were
818            skipped unless they satisfied the include/exclude conditions. This is
819            inconsistent with GNU grep (and could even be seen as contrary to the
820            pcregrep specification - which I improved to make it absolutely clear).
821            The action now is always to scan all levels of directory, and just
822            apply the include/exclude patterns to regular files.
823    
824    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
825        --exclude_dir in the tests to avoid scanning .svn directories.
826    
827    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
828        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
829        doesn't support NULs in patterns.
830    
831    7.  Added some missing "const"s to declarations of static tables in
832        pcre_compile.c and pcre_dfa_exec.c.
833    
834    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
835        caused by fix #2  above. (Subsequently also a second patch to fix the
836        first patch. And a third patch - this was a messy problem.)
837    
838    9.  Applied Craig's patch to remove the use of push_back().
839    
840    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
841        matching function regexec().
842    
843    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
844        which, however, unlike Perl's \g{...}, are subroutine calls, not back
845        references. PCRE supports relative numbers with this syntax (I don't think
846        Oniguruma does).
847    
848    12. Previously, a group with a zero repeat such as (...){0} was completely
849        omitted from the compiled regex. However, this means that if the group
850        was called as a subroutine from elsewhere in the pattern, things went wrong
851        (an internal error was given). Such groups are now left in the compiled
852        pattern, with a new opcode that causes them to be skipped at execution
853        time.
854    
855    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
856        to the way PCRE behaves:
857    
858        (a) A lone ] character is dis-allowed (Perl treats it as data).
859    
860        (b) A back reference to an unmatched subpattern matches an empty string
861            (Perl fails the current match path).
862    
863        (c) A data ] in a character class must be notated as \] because if the
864            first data character in a class is ], it defines an empty class. (In
865            Perl it is not possible to have an empty class.) The empty class []
866            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
867            The negative empty class [^] matches any one character, independently
868            of the DOTALL setting.
869    
870    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
871        non-existent subpattern following a character class starting with ']' and
872        containing () gave an internal compiling error instead of "reference to
873        non-existent subpattern". Fortunately, when the pattern did exist, the
874        compiled code was correct. (When scanning forwards to check for the
875        existencd of the subpattern, it was treating the data ']' as terminating
876        the class, so got the count wrong. When actually compiling, the reference
877        was subsequently set up correctly.)
878    
879    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
880        it was being rejected as not supported by pcre_dfa_exec(), even though
881        other assertions are supported. I have made pcre_dfa_exec() support
882        (*FAIL).
883    
884    16. The implementation of 13c above involved the invention of a new opcode,
885        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
886        cannot be changed at match time, I realized I could make a small
887        improvement to matching performance by compiling OP_ALLANY instead of
888        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
889        on the OP_ANY path.
890    
891    17. Compiling pcretest on Windows with readline support failed without the
892        following two fixes: (1) Make the unistd.h include conditional on
893        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
894    
895    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
896        ncurses library to be included for pcretest when ReadLine support is
897        requested, but also to allow for it to be overridden. This patch came from
898        Daniel Bergström.
899    
900    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
901        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
902        any errors with the current Unicode tables. Thanks to Peter Kankowski for
903        spotting this.
904    
905    
906    Version 7.6 28-Jan-08
907    ---------------------
908    
909    1.  A character class containing a very large number of characters with
910        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
911        overflow.
912    
913    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
914        HAVE_LONG_LONG is not defined.
915    
916    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
917        bring them up to date and include new features. This patch includes:
918    
919        - Fixed PH's badly added libz and libbz2 support.
920        - Fixed a problem with static linking.
921        - Added pcredemo. [But later removed - see 7 below.]
922        - Fixed dftables problem and added an option.
923        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
924            HAVE_LONG_LONG.
925        - Added readline support for pcretest.
926        - Added an listing of the option settings after cmake has run.
927    
928    4.  A user submitted a patch to Makefile that makes it easy to create
929        "pcre.dll" under mingw when using Configure/Make. I added stuff to
930        Makefile.am that cause it to include this special target, without
931        affecting anything else. Note that the same mingw target plus all
932        the other distribution libraries and programs are now supported
933        when configuring with CMake (see 6 below) instead of with
934        Configure/Make.
935    
936    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
937        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
938        exported in the Windows port". It has not yet been confirmed that the patch
939        solves the problem, but it does no harm.
940    
941    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
942        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
943        with CMake, and also correct the comment about stack recursion.
944    
945    7.  Remove the automatic building of pcredemo from the ./configure system and
946        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
947        of a program that users should build themselves after PCRE is installed, so
948        building it automatically is not really right. What is more, it gave
949        trouble in some build environments.
950    
951    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
952    
953    
954    Version 7.5 10-Jan-08
955    ---------------------
956    
957    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
958        values in parens when parsing an RE using the C++ wrapper."
959    
960    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
961        Characters greater than 255 were excluded from the class instead of being
962        included.
963    
964    3.  The same bug as (2) above applied to negated POSIX classes such as
965        [:^space:].
966    
967    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
968        defined or documented. It seems to have been a typo for PCRE_STATIC, so
969        I have changed it.
970    
971    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
972        first named subpattern) and a construct such as (?&a) would reference the
973        first named subpattern whose name started with "a" (in other words, the
974        length check was missing). Both these problems are fixed. "Subpattern name
975        expected" is now given for (?&) (a zero-length name), and this patch also
976        makes it give the same error for \k'' (previously it complained that that
977        was a reference to a non-existent subpattern).
978    
979    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
980        this is right because (?- can be followed by option settings as well as by
981        digits. I have, however, made the messages clearer.
982    
983    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
984        than the number used in the conditional) now cause a compile-time error.
985        This is actually not compatible with Perl, which accepts such patterns, but
986        treats the conditional as always being FALSE (as PCRE used to), but it
987        seems to me that giving a diagnostic is better.
988    
989    8.  Change "alphameric" to the more common word "alphanumeric" in comments
990        and messages.
991    
992    9.  Fix two occurrences of "backslash" in comments that should have been
993        "backspace".
994    
995    10. Remove two redundant lines of code that can never be obeyed (their function
996        was moved elsewhere).
997    
998    11. The program that makes PCRE's Unicode character property table had a bug
999        which caused it to generate incorrect table entries for sequences of
1000        characters that have the same character type, but are in different scripts.
1001        It amalgamated them into a single range, with the script of the first of
1002        them. In other words, some characters were in the wrong script. There were
1003        thirteen such cases, affecting characters in the following ranges:
1004    
1005          U+002b0 - U+002c1
1006          U+0060c - U+0060d
1007          U+0061e - U+00612
1008          U+0064b - U+0065e
1009          U+0074d - U+0076d
1010          U+01800 - U+01805
1011          U+01d00 - U+01d77
1012          U+01d9b - U+01dbf
1013          U+0200b - U+0200f
1014          U+030fc - U+030fe
1015          U+03260 - U+0327f
1016          U+0fb46 - U+0fbb1
1017          U+10450 - U+1049d
1018    
1019    12. The -o option (show only the matching part of a line) for pcregrep was not
1020        compatible with GNU grep in that, if there was more than one match in a
1021        line, it showed only the first of them. It now behaves in the same way as
1022        GNU grep.
1023    
1024    13. If the -o and -v options were combined for pcregrep, it printed a blank
1025        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1026        does the same. The return code can be used to tell if there were any
1027        non-matching lines.
1028    
1029    14. Added --file-offsets and --line-offsets to pcregrep.
1030    
1031    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1032        infinitely looping recursion. The bug was that positive lookaheads were not
1033        being skipped when checking for a possible empty match (negative lookaheads
1034        and both kinds of lookbehind were skipped).
1035    
1036    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1037        inclusion of <windows.h> to before rather than after the definition of
1038        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1039    
1040    17. Specifying a possessive quantifier with a specific limit for a Unicode
1041        character property caused pcre_compile() to compile bad code, which led at
1042        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1043        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1044        caused the error; without that there was no problem.
1045    
1046    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1047    
1048    19. Added --enable-pcretest-libreadline.
1049    
1050    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1051        RE::GlobalReplace(). As a result, the number of replacements returned was
1052        double what it should be. I removed one of the increments, but Craig sent a
1053        later patch that removed the other one (the right fix) and added unit tests
1054        that check the return values (which was not done before).
1055    
1056    21. Several CMake things:
1057    
1058        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1059            the names libpcre and libpcreposix, not just pcre and pcreposix.
1060    
1061        (2) The above change means that pcretest and pcregrep are now correctly
1062            linked with the newly-built libraries, not previously installed ones.
1063    
1064        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1065    
1066    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1067        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1068        UTF-8 newline character). The key issue is that the pattern starts .*;
1069        this means that the match must be either at the beginning, or after a
1070        newline. The bug was in the code for advancing after a failed match and
1071        checking that the new position followed a newline. It was not taking
1072        account of UTF-8 characters correctly.
1073    
1074    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1075        character classes. PCRE was not treating the sequence [:...:] as a
1076        character class unless the ... were all letters. Perl, however, seems to
1077        allow any characters between [: and :], though of course it rejects as
1078        unknown any "names" that contain non-letters, because all the known class
1079        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1080        for example, whereas PCRE did not - it did not recognize a POSIX character
1081        class. This seemed a bit dangerous, so the code has been changed to be
1082        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1083        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1084        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1085        Perl does, and where it didn't before.
1086    
1087    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1088        Windows environments %n is disabled by default.
1089    
1090    
1091    Version 7.4 21-Sep-07
1092    ---------------------
1093    
1094    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1095        means that a class such as [\s] counted as "explicit reference to CR or
1096        LF". That isn't really right - the whole point of the change was to try to
1097        help when there was an actual mention of one of the two characters. So now
1098        the change happens only if \r or \n (or a literal CR or LF) character is
1099        encountered.
1100    
1101    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1102        of both had grown to the point where there were only 3 bits left.
1103        Fortunately, there was spare space in the data structure, and so I have
1104        moved the internal flags into a new 16-bit field to free up more option
1105        bits.
1106    
1107    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1108        but did not set the internal JCHANGED flag - either of these is enough to
1109        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1110        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1111        start sets both bits.
1112    
1113    4.  Added options (at build time, compile time, exec time) to change \R from
1114        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1115    
1116    5.  doc/pcresyntax.html was missing from the distribution.
1117    
1118    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1119        compatibility, even though it is no longer used.
1120    
1121    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1122        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1123        windows.h file is present (where different names are used). [This was
1124        reversed later after testing - see 16 below.]
1125    
1126    8.  Changed all #include <config.h> to #include "config.h". There were also
1127        some further <pcre.h> cases that I changed to "pcre.h".
1128    
1129    9.  When pcregrep was used with the --colour option, it missed the line ending
1130        sequence off the lines that it output.
1131    
1132    10. It was pointed out to me that arrays of string pointers cause lots of
1133        relocations when a shared library is dynamically loaded. A technique of
1134        using a single long string with a table of offsets can drastically reduce
1135        these. I have refactored PCRE in four places to do this. The result is
1136        dramatic:
1137    
1138          Originally:                          290
1139          After changing UCP table:            187
1140          After changing error message table:   43
1141          After changing table of "verbs"       36
1142          After changing table of Posix names   22
1143    
1144        Thanks to the folks working on Gregex for glib for this insight.
1145    
1146    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1147        unicode-properties was also set.
1148    
1149    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1150    
1151    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1152        checked only for CRLF.
1153    
1154    14. Added casts to pcretest.c to avoid compiler warnings.
1155    
1156    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1157    
1158    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1159        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1160        entirely. This removes changes made in 7 above.
1161    
1162    17. The CMake files have been updated, and there is now more information about
1163        building with CMake in the NON-UNIX-USE document.
1164    
1165    
1166    Version 7.3 28-Aug-07
1167    ---------------------
1168    
1169     1. In the rejigging of the build system that eventually resulted in 7.1, the
1170        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1171        brackets there is not right, since it causes compilers to look for an
1172        installed pcre.h, not the version that is in the source that is being
1173        compiled (which of course may be different). I have changed it back to:
1174    
1175          #include "pcre.h"
1176    
1177        I have a vague recollection that the change was concerned with compiling in
1178        different directories, but in the new build system, that is taken care of
1179        by the VPATH setting the Makefile.
1180    
1181     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1182        when the subject happened to end in the byte 0x85 (e.g. if the last
1183        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1184        characters but of course it shouldn't be taken as a newline when it is part
1185        of another character. The bug was that, for an unlimited repeat of . in
1186        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1187        characters when looking for a newline.
1188    
1189     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1190    
1191     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1192        in debug output.
1193    
1194     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1195        long printing in the pcrecpp unittest when running under MinGW.
1196    
1197     6. ESC_K was left out of the EBCDIC table.
1198    
1199     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1200        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1201        limit also applies to "virtual nesting" when a pattern is recursive, and in
1202        this case 1000 isn't so big. I have been able to remove this limit at the
1203        expense of backing off one optimization in certain circumstances. Normally,
1204        when pcre_exec() would call its internal match() function recursively and
1205        immediately return the result unconditionally, it uses a "tail recursion"
1206        feature to save stack. However, when a subpattern that can match an empty
1207        string has an unlimited repetition quantifier, it no longer makes this
1208        optimization. That gives it a stack frame in which to save the data for
1209        checking that an empty string has been matched. Previously this was taken
1210        from the 1000-entry workspace that had been reserved. So now there is no
1211        explicit limit, but more stack is used.
1212    
1213     8. Applied Daniel's patches to solve problems with the import/export magic
1214        syntax that is required for Windows, and which was going wrong for the
1215        pcreposix and pcrecpp parts of the library. These were overlooked when this
1216        problem was solved for the main library.
1217    
1218     9. There were some crude static tests to avoid integer overflow when computing
1219        the size of patterns that contain repeated groups with explicit upper
1220        limits. As the maximum quantifier is 65535, the maximum group length was
1221        set at 30,000 so that the product of these two numbers did not overflow a
1222        32-bit integer. However, it turns out that people want to use groups that
1223        are longer than 30,000 bytes (though not repeat them that many times).
1224        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1225        made it possible to implement the integer overflow checks in a much more
1226        dynamic way, which I have now done. The artificial limitation on group
1227        length has been removed - we now have only the limit on the total length of
1228        the compiled pattern, which depends on the LINK_SIZE setting.
1229    
1230    10. Fixed a bug in the documentation for get/copy named substring when
1231        duplicate names are permitted. If none of the named substrings are set, the
1232        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1233        empty string.
1234    
1235    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1236        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1237        because the ] is interpreted as the first data character and the
1238        terminating ] is not found. PCRE has been made compatible with Perl in this
1239        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1240        cause memory overwriting.
1241    
1242    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1243        string has been matched (to stop an infinite loop). It was not recognizing
1244        a conditional subpattern that could match an empty string if that
1245        subpattern was within another subpattern. For example, it looped when
1246        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1247        condition was not nested. This bug has been fixed.
1248    
1249    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1250        past the start of the subject in the presence of bytes with the top bit
1251        set, for example "\x8aBCD".
1252    
1253    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1254        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1255    
1256    14. Optimized (?!) to (*FAIL).
1257    
1258    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1259        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1260        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1261        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1262        does: it's just the validity check that is more restrictive.
1263    
1264    16. Inserted checks for integer overflows during escape sequence (backslash)
1265        processing, and also fixed erroneous offset values for syntax errors during
1266        backslash processing.
1267    
1268    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1269        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1270    
1271    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1272        caused an overrun.
1273    
1274    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1275        something other than just ASCII characters) inside a group that had an
1276        unlimited repeat caused a loop at compile time (while checking to see
1277        whether the group could match an empty string).
1278    
1279    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1280        [\P{Any}] did so. (Error in the code for printing property names.)
1281    
1282    21. An orphan \E inside a character class could cause a crash.
1283    
1284    22. A repeated capturing bracket such as (A)? could cause a wild memory
1285        reference during compilation.
1286    
1287    23. There are several functions in pcre_compile() that scan along a compiled
1288        expression for various reasons (e.g. to see if it's fixed length for look
1289        behind). There were bugs in these functions when a repeated \p or \P was
1290        present in the pattern. These operators have additional parameters compared
1291        with \d, etc, and these were not being taken into account when moving along
1292        the compiled data. Specifically:
1293    
1294        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1295            length.
1296    
1297        (b) An item such as \pL+ within a repeated group could cause crashes or
1298            loops.
1299    
1300        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1301            "reference to non-existent subpattern" error.
1302    
1303        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1304    
1305    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1306        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1307    
1308    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1309    
1310    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1311        character were causing crashes (broken optimization).
1312    
1313    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1314        \p or \P) caused a compile-time loop.
1315    
1316    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1317        break. For example, the unstudied pattern [\r\n]A does not match the string
1318        "\r\nA" because change 7.0/46 below moves the current point on by two
1319        characters after failing to match at the start. However, the pattern \nA
1320        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1321        the same is true. There doesn't seem any very clean way out of this, but
1322        what I have chosen to do makes the common cases work: PCRE now takes note
1323        of whether there can be an explicit match for \r or \n anywhere in the
1324        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1325        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1326        pattern has explicit CR or LF references.
1327    
1328    29. Added (*CR) etc for changing newline setting at start of pattern.
1329    
1330    
1331    Version 7.2 19-Jun-07
1332    ---------------------
1333    
1334     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1335        which is apparently normally available under Windows.
1336    
1337     2. Re-jig the pcregrep tests with different newline settings in an attempt
1338        to make them independent of the local environment's newline setting.
1339    
1340     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1341    
1342     4. Some of the "internals" tests were previously cut out when the link size
1343        was not 2, because the output contained actual offsets. The recent new
1344        "Z" feature of pcretest means that these can be cut out, making the tests
1345        usable with all link sizes.
1346    
1347     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1348        stack recursion. This gives a massive performance boost under BSD, but just
1349        a small improvement under Linux. However, it saves one field in the frame
1350        in all cases.
1351    
1352     6. Added more features from the forthcoming Perl 5.10:
1353    
1354        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1355            recursion call. It refers to the nth most recently opened parentheses.
1356    
1357        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1358            to be opened parentheses.
1359    
1360        (c) Conditions that refer to capturing parentheses can be specified
1361            relatively, for example, (?(-2)... or (?(+3)...
1362    
1363        (d) \K resets the start of the current match so that everything before
1364            is not part of it.
1365    
1366        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1367    
1368        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1369            reference syntax.
1370    
1371        (g) (?| introduces a group in which the numbering of parentheses in each
1372            alternative starts with the same number.
1373    
1374        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1375    
1376     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1377        PCRE_INFO_JCHANGED.
1378    
1379     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1380        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1381        for detecting groups that can match an empty string.
1382    
1383     9. A pattern with a very large number of alternatives (more than several
1384        hundred) was running out of internal workspace during the pre-compile
1385        phase, where pcre_compile() figures out how much memory will be needed. A
1386        bit of new cunning has reduced the workspace needed for groups with
1387        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1388        workspace instead of running out of the 4096 that are available.
1389    
1390    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1391    
1392    11. Applied patch from Google to remove an optimization that didn't quite work.
1393        The report of the bug said:
1394    
1395          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1396          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1397          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1398    
1399    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1400        it matched the wrong number of bytes.
1401    
1402    
1403    Version 7.1 24-Apr-07
1404    ---------------------
1405    
1406     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1407        that is more "standard", making use of automake and other Autotools. There
1408        is some re-arrangement of the files and adjustment of comments consequent
1409        on this.
1410    
1411     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1412        for recursive directory scanning broke on some systems because the files
1413        are not scanned in any specific order and on different systems the order
1414        was different. A call to "sort" has been inserted into RunGrepTest for the
1415        approprate test as a short-term fix. In the longer term there may be an
1416        alternative.
1417    
1418     3. I had an email from Eric Raymond about problems translating some of PCRE's
1419        man pages to HTML (despite the fact that I distribute HTML pages, some
1420        people do their own conversions for various reasons). The problems
1421        concerned the use of low-level troff macros .br and .in. I have therefore
1422        removed all such uses from the man pages (some were redundant, some could
1423        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1424        HTML has been updated to handle .nf/.fi and to complain if it encounters
1425        .br or .in.
1426    
1427     4. Updated comments in configure.ac that get placed in config.h.in and also
1428        arranged for config.h to be included in the distribution, with the name
1429        config.h.generic, for the benefit of those who have to compile without
1430        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1431    
1432     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1433        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1434        makevp.bat for the current PCRE, using the additional files
1435        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1436    
1437     6. A Windows user reported a minor discrepancy with test 2, which turned out
1438        to be caused by a trailing space on an input line that had got lost in his
1439        copy. The trailing space was an accident, so I've just removed it.
1440    
1441     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1442        that is needed.
1443    
1444     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1445        as "const" (a) because they are and (b) because it helps the PHP
1446        maintainers who have recently made a script to detect big data structures
1447        in the php code that should be moved to the .rodata section. I remembered
1448        to update Builducptable as well, so it won't revert if ucptable.h is ever
1449        re-created.
1450    
1451     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1452        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1453        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1454        support is not required. This saves 1.5-2K of code, which is important in
1455        some applications.
1456    
1457        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1458        so as not to refer to the tables, even though these functions will never be
1459        called when UTF-8 support is disabled. Otherwise there are problems with a
1460        shared library.
1461    
1462    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1463    
1464        (a) It was defining its arguments as char * instead of void *.
1465    
1466        (b) It was assuming that all moves were upwards in memory; this was true
1467            a long time ago when I wrote it, but is no longer the case.
1468    
1469        The emulated memove() is provided for those environments that have neither
1470        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1471        is clearly not the case, as these two bugs were recently reported.
1472    
1473    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1474        and Detrail to create the HTML documentation, the .txt form of the man
1475        pages, and it removes trailing spaces from listed files. It also creates
1476        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1477        case, it wraps all the #defines with #ifndefs. This script should be run
1478        before "make dist".
1479    
1480    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1481        with Unicode property support.
1482    
1483        (a) For a maximizing quantifier, if the two different cases of the
1484            character were of different lengths in their UTF-8 codings (there are
1485            some cases like this - I found 11), and the matching function had to
1486            back up over a mixture of the two cases, it incorrectly assumed they
1487            were both the same length.
1488    
1489        (b) When PCRE was configured to use the heap rather than the stack for
1490            recursion during matching, it was not correctly preserving the data for
1491            the other case of a UTF-8 character when checking ahead for a match
1492            while processing a minimizing repeat. If the check also involved
1493            matching a wide character, but failed, corruption could cause an
1494            erroneous result when trying to check for a repeat of the original
1495            character.
1496    
1497    13. Some tidying changes to the testing mechanism:
1498    
1499        (a) The RunTest script now detects the internal link size and whether there
1500            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1501            values substituted by "configure". (The RunGrepTest script already did
1502            this for UTF-8.) The configure.ac script no longer substitutes the
1503            relevant variables.
1504    
1505        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1506            with length and offset values. This means that the output is different
1507            for different internal link sizes. Test 2 is skipped for link sizes
1508            other than 2 because of this, bypassing the problem. Unfortunately,
1509            there was also a test in test 3 (the locale tests) that used /B and
1510            failed for link sizes other than 2. Rather than cut the whole test out,
1511            I have added a new /Z option to pcretest that replaces the length and
1512            offset values with spaces. This is now used to make test 3 independent
1513            of link size. (Test 2 will be tidied up later.)
1514    
1515    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1516        segmentation fault instead of returning the appropriate error message.
1517    
1518    15. In multiline mode when the newline sequence was set to "any", the pattern
1519        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1520        This doesn't seem right; it now treats the CRLF combination as the line
1521        ending, and so does not match in that case. It's only a pattern such as ^$
1522        that would hit this one: something like ^ABC$ would have failed after \r
1523        and then tried again after \r\n.
1524    
1525    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1526        in an attempt to make files that differ only in their line terminators
1527        compare equal. This works on Linux.
1528    
1529    17. Under certain error circumstances pcregrep might try to free random memory
1530        as it exited. This is now fixed, thanks to valgrind.
1531    
1532    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1533        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1534        was because its rules for how to advance for /g after matching an empty
1535        string at the end of a line did not allow for this case. They now check for
1536        it specially.
1537    
1538    20. pcretest is supposed to handle patterns and data of any length, by
1539        extending its buffers when necessary. It was getting this wrong when the
1540        buffer for a data line had to be extended.
1541    
1542    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1543        CRLF as a newline sequence.
1544    
1545    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1546        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1547        I have nevertheless tidied it up.
1548    
1549    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1550    
1551    24. Added a man page for pcre-config.
1552    
1553    
1554    Version 7.0 19-Dec-06
1555    ---------------------
1556    
1557     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1558        moving to gcc 4.1.1.
1559    
1560     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1561        sys/time.h, which is documented as needed for this function. It doesn't
1562        seem to matter on Linux, but it showed up on some releases of OS X.
1563    
1564     3. It seems that there are systems where bytes whose values are greater than
1565        127 match isprint() in the "C" locale. The "C" locale should be the
1566        default when a C program starts up. In most systems, only ASCII printing
1567        characters match isprint(). This difference caused the output from pcretest
1568        to vary, making some of the tests fail. I have changed pcretest so that:
1569    
1570        (a) When it is outputting text in the compiled version of a pattern, bytes
1571            other than 32-126 are always shown as hex escapes.
1572    
1573        (b) When it is outputting text that is a matched part of a subject string,
1574            it does the same, unless a different locale has been set for the match
1575            (using the /L modifier). In this case, it uses isprint() to decide.
1576    
1577     4. Fixed a major bug that caused incorrect computation of the amount of memory
1578        required for a compiled pattern when options that changed within the
1579        pattern affected the logic of the preliminary scan that determines the
1580        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1581        that the computed length was too small. The symptoms of this bug were
1582        either the PCRE error "internal error: code overflow" from pcre_compile(),
1583        or a glibc crash with a message such as "pcretest: free(): invalid next
1584        size (fast)". Examples of patterns that provoked this bug (shown in
1585        pcretest format) are:
1586    
1587          /(?-x: )/x
1588          /(?x)(?-x: \s*#\s*)/
1589          /((?i)[\x{c0}])/8
1590          /(?i:[\x{c0}])/8
1591    
1592        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1593        is now done differently.
1594    
1595     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1596        wrapper classes; (b) implement a new function in the C++ scanner that is
1597        more efficient than the old way of doing things because it avoids levels of
1598        recursion in the regex matching; (c) add a paragraph to the documentation
1599        for the FullMatch() function.
1600    
1601     6. The escape sequence \n was being treated as whatever was defined as
1602        "newline". Not only was this contrary to the documentation, which states
1603        that \n is character 10 (hex 0A), but it also went horribly wrong when
1604        "newline" was defined as CRLF. This has been fixed.
1605    
1606     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1607        was being set to -1 for the "end of line" case (supposedly a value that no
1608        character can have). Though this value is never used (the check for end of
1609        line is "zero bytes in current character"), it caused compiler complaints.
1610        I've changed it to 0xffffffff.
1611    
1612     8. In pcre_version.c, the version string was being built by a sequence of
1613        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1614        string (as it is for production releases) called a macro with an empty
1615        argument. The C standard says the result of this is undefined. The gcc
1616        compiler treats it as an empty string (which was what was wanted) but it is
1617        reported that Visual C gives an error. The source has been hacked around to
1618        avoid this problem.
1619    
1620     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1621        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1622        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1623        of them did).
1624    
1625    10. Originally, pcretest opened its input and output without "b"; then I was
1626        told that "b" was needed in some environments, so it was added for release
1627        5.0 to both the input and output. (It makes no difference on Unix-like
1628        systems.) Later I was told that it is wrong for the input on Windows. I've
1629        now abstracted the modes into two macros, to make it easier to fiddle with
1630        them, and removed "b" from the input mode under Windows.
1631    
1632    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1633    
1634    12. Added -help and --help to pcretest as an official way of being reminded
1635        of the options.
1636    
1637    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1638        and pcrecpp.cc because they annoy compilers at high warning levels.
1639    
1640    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1641    
1642    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1643        scripts are not C programs :-) and which was not noticed because it works
1644        on Linux.
1645    
1646    16. pcretest is supposed to handle any length of pattern and data line (as one
1647        line or as a continued sequence of lines) by extending its input buffer if
1648        necessary. This feature was broken for very long pattern lines, leading to
1649        a string of junk being passed to pcre_compile() if the pattern was longer
1650        than about 50K.
1651    
1652    17. I have done a major re-factoring of the way pcre_compile() computes the
1653        amount of memory needed for a compiled pattern. Previously, there was code
1654        that made a preliminary scan of the pattern in order to do this. That was
1655        OK when PCRE was new, but as the facilities have expanded, it has become
1656        harder and harder to keep it in step with the real compile phase, and there
1657        have been a number of bugs (see for example, 4 above). I have now found a
1658        cunning way of running the real compile function in a "fake" mode that
1659        enables it to compute how much memory it would need, while actually only
1660        ever using a few hundred bytes of working memory and without too many
1661        tests of the mode. This should make future maintenance and development
1662        easier. A side effect of this work is that the limit of 200 on the nesting
1663        depth of parentheses has been removed (though this was never a serious
1664        limitation, I suspect). However, there is a downside: pcre_compile() now
1665        runs more slowly than before (30% or more, depending on the pattern). I
1666        hope this isn't a big issue. There is no effect on runtime performance.
1667    
1668    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1669        newline (only possible for the last line of a file) and it was a
1670        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1671    
1672    19. Added additional timing features to pcretest. (1) The -tm option now times
1673        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1674        separate command line item, by a number that specifies the number of
1675        repeats to use when timing. The default is 50000; this gives better
1676        precision, but takes uncomfortably long for very large patterns.
1677    
1678    20. Extended pcre_study() to be more clever in cases where a branch of a
1679        subpattern has no definite first character. For example, (a*|b*)[cd] would
1680        previously give no result from pcre_study(). Now it recognizes that the
1681        first character must be a, b, c, or d.
1682    
1683    21. There was an incorrect error "recursive call could loop indefinitely" if
1684        a subpattern (or the entire pattern) that was being tested for matching an
1685        empty string contained only one non-empty item after a nested subpattern.
1686        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1687        incorrectly, because the \d was being skipped in the check.
1688    
1689    22. The pcretest program now has a new pattern option /B and a command line
1690        option -b, which is equivalent to adding /B to every pattern. This causes
1691        it to show the compiled bytecode, without the additional information that
1692        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1693        is the same as /B/I).
1694    
1695    23. A new optimization is now able automatically to treat some sequences such
1696        as a*b as a*+b. More specifically, if something simple (such as a character
1697        or a simple class like \d) has an unlimited quantifier, and is followed by
1698        something that cannot possibly match the quantified thing, the quantifier
1699        is automatically "possessified".
1700    
1701    24. A recursive reference to a subpattern whose number was greater than 39
1702        went wrong under certain circumstances in UTF-8 mode. This bug could also
1703        have affected the operation of pcre_study().
1704    
1705    25. Realized that a little bit of performance could be had by replacing
1706        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1707    
1708    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1709    
1710    27. Possessive quantifiers such as a++ were previously implemented by turning
1711        them into atomic groups such as ($>a+). Now they have their own opcodes,
1712        which improves performance. This includes the automatically created ones
1713        from 23 above.
1714    
1715    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1716        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1717        the first matched character to be a colon. This applied both to named and
1718        numbered groups.
1719    
1720    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1721    
1722    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1723        building PCRE on Windows easier, so I have included it in the distribution.
1724    
1725    31. There is now a check in pcretest against a ridiculously large number being
1726        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1727        loop, the loop is abandoned.
1728    
1729    32. Forward references to subpatterns in conditions such as (?(2)...) where
1730        subpattern 2 is defined later cause pcre_compile() to search forwards in
1731        the pattern for the relevant set of parentheses. This search went wrong
1732        when there were unescaped parentheses in a character class, parentheses
1733        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1734    
1735    33. "Subroutine" calls and backreferences were previously restricted to
1736        referencing subpatterns earlier in the regex. This restriction has now
1737        been removed.
1738    
1739    34. Added a number of extra features that are going to be in Perl 5.10. On the
1740        whole, these are just syntactic alternatives for features that PCRE had
1741        previously implemented using the Python syntax or my own invention. The
1742        other formats are all retained for compatibility.
1743    
1744        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1745            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1746            also .NET compatible.
1747    
1748        (b) A recursion or subroutine call to a named group can now be defined as
1749            (?&name) as well as (?P>name).
1750    
1751        (c) A backreference to a named group can now be defined as \k<name> or
1752            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1753            5.10, are also .NET compatible.
1754    
1755        (d) A conditional reference to a named group can now use the syntax
1756            (?(<name>) or (?('name') as well as (?(name).
1757    
1758        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1759            groups (named and numbered) that are never evaluated inline, but can be
1760            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1761            is always false. There may be only one alternative in such a group.
1762    
1763        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1764            as the simple (?(R). The condition is true only if the most recent
1765            recursion is that of the given number or name. It does not search out
1766            through the entire recursion stack.
1767    
1768        (g) The escape \gN or \g{N} has been added, where N is a positive or
1769            negative number, specifying an absolute or relative reference.
1770    
1771    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1772        some "unreachable code" warnings.
1773    
1774    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1775        things, this adds five new scripts.
1776    
1777    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1778        There were also incompatibilities regarding the handling of \Q..\E inside
1779        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1780        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1781    
1782    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1783        matches an empty string, and forcibly breaks the loop. There were bugs in
1784        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1785        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1786        separate and independent bugs (that affected different cases) have been
1787        fixed.
1788    
1789    39. Refactored the code to abolish the use of different opcodes for small
1790        capturing bracket numbers. This is a tidy that I avoided doing when I
1791        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1792        The new approach is not only tidier, it makes it possible to reduce the
1793        memory needed to fix the previous bug (38).
1794    
1795    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1796        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1797        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1798        mode.
1799    
1800    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1801        report.
1802    
1803    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1804        copy construction and assignment in the C++ wrapper.
1805    
1806    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1807        couple of bugs that could have given wrong results in the "--newline=crlf"
1808        case.
1809    
1810    44. Added a number of casts and did some reorganization of signed/unsigned int
1811        variables following suggestions from Dair Grant. Also renamed the variable
1812        "this" as "item" because it is a C++ keyword.
1813    
1814    45. Arranged for dftables to add
1815    
1816          #include "pcre_internal.h"
1817    
1818        to pcre_chartables.c because without it, gcc 4.x may remove the array
1819        definition from the final binary if PCRE is built into a static library and
1820        dead code stripping is activated.
1821    
1822    46. For an unanchored pattern, if a match attempt fails at the start of a
1823        newline sequence, and the newline setting is CRLF or ANY, and the next two
1824        characters are CRLF, advance by two characters instead of one.
1825    
1826    
1827    Version 6.7 04-Jul-06
1828    ---------------------
1829    
1830     1. In order to handle tests when input lines are enormously long, pcretest has
1831        been re-factored so that it automatically extends its buffers when
1832        necessary. The code is crude, but this _is_ just a test program. The
1833        default size has been increased from 32K to 50K.
1834    
1835     2. The code in pcre_study() was using the value of the re argument before
1836        testing it for NULL. (Of course, in any sensible call of the function, it
1837        won't be NULL.)
1838    
1839     3. The memmove() emulation function in pcre_internal.h, which is used on
1840        systems that lack both memmove() and bcopy() - that is, hardly ever -
1841        was missing a "static" storage class specifier.
1842    
1843     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1844        containing an extended class (one that cannot be represented by a bitmap
1845        because it contains high-valued characters or Unicode property items, e.g.
1846        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1847        pattern, but PCRE should not loop if you do not (it no longer does).
1848        [Detail: two cases were found: (a) a repeated subpattern containing an
1849        extended class; (b) a recursive reference to a subpattern that followed a
1850        previous extended class. It wasn't skipping over the extended class
1851        correctly when UTF-8 mode was not set.]
1852    
1853     5. A negated single-character class was not being recognized as fixed-length
1854        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1855        compile error "lookbehind assertion is not fixed length".
1856    
1857     6. The RunPerlTest auxiliary script was showing an unexpected difference
1858        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1859        write a Perl script that can interpret lines of an input file either as
1860        byte characters or as UTF-8, which is what "perltest" was being required to
1861        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1862        can't do is switch easily at run time between having the "use utf8;" pragma
1863        or not. In the end, I fudged it by using the RunPerlTest script to insert
1864        "use utf8;" explicitly for the UTF-8 tests.
1865    
1866     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1867        the end of the subject string, contrary to the documentation and to what
1868        Perl does. This was true of both matching functions. Now it matches only at
1869        the start of the subject and immediately after *internal* newlines.
1870    
1871     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1872        a pointer to an int instead of a pointer to an unsigned long int. This
1873        caused problems on 64-bit systems.
1874    
1875     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1876        instance of the 'standard' template library not being so standard".
1877    
1878    10. There was no check on the number of named subpatterns nor the maximum
1879        length of a subpattern name. The product of these values is used to compute
1880        the size of the memory block for a compiled pattern. By supplying a very
1881        long subpattern name and a large number of named subpatterns, the size
1882        computation could be caused to overflow. This is now prevented by limiting
1883        the length of names to 32 characters, and the number of named subpatterns
1884        to 10,000.
1885    
1886    11. Subpatterns that are repeated with specific counts have to be replicated in
1887        the compiled pattern. The size of memory for this was computed from the
1888        length of the subpattern and the repeat count. The latter is limited to
1889        65535, but there was no limit on the former, meaning that integer overflow
1890        could in principle occur. The compiled length of a repeated subpattern is
1891        now limited to 30,000 bytes in order to prevent this.
1892    
1893    12. Added the optional facility to have named substrings with the same name.
1894    
1895    13. Added the ability to use a named substring as a condition, using the
1896        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1897        are numbers (not recommended). Forward references are permitted.
1898    
1899    14. Added forward references in named backreferences (if you see what I mean).
1900    
1901    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1902        pattern could run off the end of the subject. For example, the pattern
1903        "(?s)(.{1,5})"8 did this with the subject "ab".
1904    
1905    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1906        PCRE_CASELESS was set when matching characters that were quantified with ?
1907        or *.
1908    
1909    17. A character class other than a single negated character that had a minimum
1910        but no maximum quantifier - for example [ab]{6,} - was not handled
1911        correctly by pce_dfa_exec(). It would match only one character.
1912    
1913    18. A valid (though odd) pattern that looked like a POSIX character
1914        class but used an invalid character after [ (for example [[,abc,]]) caused
1915        pcre_compile() to give the error "Failed: internal error: code overflow" or
1916        in some cases to crash with a glibc free() error. This could even happen if
1917        the pattern terminated after [[ but there just happened to be a sequence of
1918        letters, a binary zero, and a closing ] in the memory that followed.
1919    
1920    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1921        over the years. Originally (before any Unicode support), just the bottom 8
1922        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1923        output from "man perlunicode" includes this:
1924    
1925          The regular expression compiler produces polymorphic opcodes.  That
1926          is, the pattern adapts to the data and automatically switches to
1927          the Unicode character scheme when presented with Unicode data--or
1928          instead uses a traditional byte scheme when presented with byte
1929          data.
1930    
1931        Sadly, a wide octal escape does not cause a switch, and in a string with
1932        no other multibyte characters, these octal escapes are treated as before.
1933        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1934        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1935        Unicode string.
1936    
1937        I have not perpetrated such confusion in PCRE. Up till now, it took just
1938        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1939        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1940        translate to the appropriate multibyte character.
1941    
1942    29. Applied some refactoring to reduce the number of warnings from Microsoft
1943        and Borland compilers. This has included removing the fudge introduced
1944        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1945        a warning about an unused variable.
1946    
1947    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1948        characters since release 4.0, because Perl (from release 5.004) does not.
1949        [Or at least, is documented not to: some releases seem to be in conflict
1950        with the documentation.] However, when a pattern was studied with
1951        pcre_study() and all its branches started with \s, PCRE still included VT
1952        as a possible starting character. Of course, this did no harm; it just
1953        caused an unnecessary match attempt.
1954    
1955    22. Removed a now-redundant internal flag bit that recorded the fact that case
1956        dependency changed within the pattern. This was once needed for "required
1957        byte" processing, but is no longer used. This recovers a now-scarce options
1958        bit. Also moved the least significant internal flag bit to the most-
1959        significant bit of the word, which was not previously used (hangover from
1960        the days when it was an int rather than a uint) to free up another bit for
1961        the future.
1962    
1963    23. Added support for CRLF line endings as well as CR and LF. As well as the
1964        default being selectable at build time, it can now be changed at runtime
1965        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1966        specify that it is scanning data with non-default line endings.
1967    
1968    24. Changed the definition of CXXLINK to make it agree with the definition of
1969        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1970    
1971    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1972        recursions. This makes a big different to stack usage for some patterns.
1973    
1974    26. If a subpattern containing a named recursion or subroutine reference such
1975        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1976        the space required for the compiled pattern went wrong and gave too small a
1977        value. Depending on the environment, this could lead to "Failed: internal
1978        error: code overflow at offset 49" or "glibc detected double free or
1979        corruption" errors.
1980    
1981    27. Applied patches from Google (a) to support the new newline modes and (b) to
1982        advance over multibyte UTF-8 characters in GlobalReplace.
1983    
1984    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1985        difference for some implementation of PCRE in some Windows version.
1986    
1987    29. Added some extra testing facilities to pcretest:
1988    
1989        \q<number>   in a data line sets the "match limit" value
1990        \Q<number>   in a data line sets the "match recursion limt" value
1991        -S <number>  sets the stack size, where <number> is in megabytes
1992    
1993        The -S option isn't available for Windows.
1994    
1995    
1996    Version 6.6 06-Feb-06
1997    ---------------------
1998    
1999     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2000        in pcreposix.h. I have copied the definition from pcre.h.
2001    
2002     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2003        because pcre.h is no longer a built file.
2004    
2005     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2006        not normally included in the compiled code.
2007    
2008    
2009    Version 6.5 01-Feb-06
2010    ---------------------
2011    
2012     1. When using the partial match feature with pcre_dfa_exec(), it was not
2013        anchoring the second and subsequent partial matches at the new starting
2014        point. This could lead to incorrect results. For example, with the pattern
2015        /1234/, partially matching against "123" and then "a4" gave a match.
2016    
2017     2. Changes to pcregrep:
2018    
2019        (a) All non-match returns from pcre_exec() were being treated as failures
2020            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2021            error message is output. Some extra information is given for the
2022            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2023            probably the only errors that are likely to be caused by users (by
2024            specifying a regex that has nested indefinite repeats, for instance).
2025            If there are more than 20 of these errors, pcregrep is abandoned.
2026    
2027        (b) A binary zero was treated as data while matching, but terminated the
2028            output line if it was written out. This has been fixed: binary zeroes
2029            are now no different to any other data bytes.
2030    
2031        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2032            used to set a locale for matching. The --locale=xxxx long option has
2033            been added (no short equivalent) to specify a locale explicitly on the
2034            pcregrep command, overriding the environment variables.
2035    
2036        (d) When -B was used with -n, some line numbers in the output were one less
2037            than they should have been.
2038    
2039        (e) Added the -o (--only-matching) option.
2040    
2041        (f) If -A or -C was used with -c (count only), some lines of context were
2042            accidentally printed for the final match.
2043    
2044        (g) Added the -H (--with-filename) option.
2045    
2046        (h) The combination of options -rh failed to suppress file names for files
2047            that were found from directory arguments.
2048    
2049        (i) Added the -D (--devices) and -d (--directories) options.
2050    
2051        (j) Added the -F (--fixed-strings) option.
2052    
2053        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2054    
2055        (l) Added the --colo(u)r option.
2056    
2057        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2058            is not present by default.
2059    
2060     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2061        items such as (?R) or (?1), when the recursion could match a number of
2062        alternatives. If it matched one of the alternatives, but subsequently,
2063        outside the recursion, there was a failure, the code tried to back up into
2064        the recursion. However, because of the way PCRE is implemented, this is not
2065        possible, and the result was an incorrect result from the match.
2066    
2067        In order to prevent this happening, the specification of recursion has
2068        been changed so that all such subpatterns are automatically treated as
2069        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2070    
2071     4. I had overlooked the fact that, in some locales, there are characters for
2072        which isalpha() is true but neither isupper() nor islower() are true. In
2073        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2074        and ordfeminine) are like this. This affected the treatment of \w and \W
2075        when they appeared in character classes, but not when they appeared outside
2076        a character class. The bit map for "word" characters is now created
2077        separately from the results of isalnum() instead of just taking it from the
2078        upper, lower, and digit maps. (Plus the underscore character, of course.)
2079    
2080     5. The above bug also affected the handling of POSIX character classes such as
2081        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2082        permanent tables. Instead, the bit maps for such a class were previously
2083        created as the appropriate unions of the upper, lower, and digit bitmaps.
2084        Now they are created by subtraction from the [[:word:]] class, which has
2085        its own bitmap.
2086    
2087     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2088        It is created by subtracting the vertical space characters (\x09, \x0a,
2089        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2090        subtraction was done in the overall bitmap for a character class, meaning
2091        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2092        be recognized. This bug has been fixed.
2093    
2094     7. Patches from the folks at Google:
2095    
2096          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2097          real life, but is still worth protecting against".
2098    
2099          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2100          regular expressions".
2101    
2102          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2103          have it.
2104    
2105          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2106          "configure" and the latter not, in order to fix a problem somebody had
2107          with compiling the Arg class on HP-UX.
2108    
2109          (e) Improve the error-handling of the C++ wrapper a little bit.
2110    
2111          (f) New tests for checking recursion limiting.
2112    
2113     8. The pcre_memmove() function, which is used only if the environment does not
2114        have a standard memmove() function (and is therefore rarely compiled),
2115        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2116        returning a result (though PCRE never actually uses the result).
2117    
2118     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2119        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2120        returned instead of calling malloc() with an overflowing number that would
2121        most likely cause subsequent chaos.
2122    
2123    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2124    
2125    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2126        with this option is matched, the nmatch and pmatch options of regexec() are
2127        ignored.
2128    
2129    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2130        provided in case anyone wants to the the POSIX interface with UTF-8
2131        strings.
2132    
2133    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2134        C++ linking (needed for some HP-UX environments).
2135    
2136    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2137        (unused parameter) and in the pcre_printint() function (omitted "default"
2138        switch label when the default is to do nothing).
2139    
2140    15. Added some code to make it possible, when PCRE is compiled as a C++
2141        library, to replace subject pointers for pcre_exec() with a smart pointer
2142        class, thus making it possible to process discontinuous strings.
2143    
2144    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2145        much the same function. They were added by different people who were trying
2146        to make PCRE easy to compile on non-Unix systems. It has been suggested
2147        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2148        for compiling on Windows systems. I have therefore replaced it with
2149        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2150        defaults to "extern" for C or "extern C" for C++, which works fine on
2151        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2152        SCOPE with something explicit in config.h. In addition:
2153    
2154        (a) pcreposix.h still had just "extern" instead of either of these macros;
2155            I have replaced it with PCRE_DATA_SCOPE.
2156    
2157        (b) Functions such as _pcre_xclass(), which are internal to the library,
2158            but external in the C sense, all had PCRE_EXPORT in their definitions.
2159            This is apparently wrong for the Windows case, so I have removed it.
2160            (It makes no difference on Unix-like systems.)
2161    
2162    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2163        of recursive calls to match(). This is different to MATCH_LIMIT because
2164        that limits the total number of calls to match(), not all of which increase
2165        the depth of recursion. Limiting the recursion depth limits the amount of
2166        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2167        when PCRE is compiled, and changed at run time. A patch from Google adds
2168        this functionality to the C++ interface.
2169    
2170    18. Changes to the handling of Unicode character properties:
2171    
2172        (a) Updated the table to Unicode 4.1.0.
2173    
2174        (b) Recognize characters that are not in the table as "Cn" (undefined).
2175    
2176        (c) I revised the way the table is implemented to a much improved format
2177            which includes recognition of ranges. It now supports the ranges that
2178            are defined in UnicodeData.txt, and it also amalgamates other
2179            characters into ranges. This has reduced the number of entries in the
2180            table from around 16,000 to around 3,000, thus reducing its size
2181            considerably. I realized I did not need to use a tree structure after
2182            all - a binary chop search is just as efficient. Having reduced the
2183            number of entries, I extended their size from 6 bytes to 8 bytes to
2184            allow for more data.
2185    
2186        (d) Added support for Unicode script names via properties such as \p{Han}.
2187    
2188    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2189        matching that character.
2190    
2191    20. When matching a repeated Unicode property with a minimum greater than zero,
2192        (for example \pL{2,}), PCRE could look past the end of the subject if it
2193        reached it while seeking the minimum number of characters. This could
2194        happen only if some of the characters were more than one byte long, because
2195        there is a check for at least the minimum number of bytes.
2196    
2197    21. Refactored the implementation of \p and \P so as to be more general, to
2198        allow for more different types of property in future. This has changed the
2199        compiled form incompatibly. Anybody with saved compiled patterns that use
2200        \p or \P will have to recompile them.
2201    
2202    22. Added "Any" and "L&" to the supported property types.
2203    
2204    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2205        but give a compile time error if the value is greater than 0xff.
2206    
2207    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2208        accidentally not being installed or uninstalled.
2209    
2210    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2211        made were to insert the current release number. This seemed silly, because
2212        it made things harder for people building PCRE on systems that don't run
2213        "configure". I have turned pcre.h into a distributed file, no longer built
2214        by "configure", with the version identification directly included. There is
2215        no longer a pcre.h.in file.
2216    
2217        However, this change necessitated a change to the pcre-config script as
2218        well. It is built from pcre-config.in, and one of the substitutions was the
2219        release number. I have updated configure.ac so that ./configure now finds
2220        the release number by grepping pcre.h.
2221    
2222    26. Added the ability to run the tests under valgrind.
2223    
2224    
2225    Version 6.4 05-Sep-05
2226    ---------------------
2227    
2228     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2229        "--" to be printed when multiple files were scanned, even when none of the
2230        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2231        consider it to be a bug, and have restored the previous behaviour.
2232    
2233     2. A couple of code tidies to get rid of compiler warnings.
2234    
2235     3. The pcretest program used to cheat by referring to symbols in the library
2236        whose names begin with _pcre_. These are internal symbols that are not
2237        really supposed to be visible externally, and in some environments it is
2238        possible to suppress them. The cheating is now confined to including
2239        certain files from the library's source, which is a bit cleaner.
2240    
2241     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2242        file's purpose clearer.
2243    
2244     5. Reorganized pcre_ucp_findchar().
2245    
2246    
2247    Version 6.3 15-Aug-05
2248    ---------------------
2249    
2250     1. The file libpcre.pc.in did not have general read permission in the tarball.
2251    
2252     2. There were some problems when building without C++ support:
2253    
2254        (a) If C++ support was not built, "make install" and "make test" still
2255            tried to test it.
2256    
2257        (b) There were problems when the value of CXX was explicitly set. Some
2258            changes have been made to try to fix these, and ...
2259    
2260        (c) --disable-cpp can now be used to explicitly disable C++ support.
2261    
2262        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2263            backslash in a target when C++ was disabled. This confuses some
2264            versions of "make", apparently. Using an intermediate variable solves
2265            this. (Same for CPP_LOBJ.)
2266    
2267     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2268        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2269        necessary on certain architectures.
2270    
2271     4. Added a setting of -export-symbols-regex to the link command to remove
2272        those symbols that are exported in the C sense, but actually are local
2273        within the library, and not documented. Their names all begin with
2274        "_pcre_". This is not a perfect job, because (a) we have to except some
2275        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2276        available (and never for static libraries). I have made a note to try to
2277        find a way round (a) in the future.
2278    
2279    
2280    Version 6.2 01-Aug-05
2281    ---------------------
2282    
2283     1. There was no test for integer overflow of quantifier values. A construction
2284        such as {1111111111111111} would give undefined results. What is worse, if
2285        a minimum quantifier for a parenthesized subpattern overflowed and became
2286        negative, the calculation of the memory size went wrong. This could have
2287        led to memory overwriting.
2288    
2289     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2290    
2291     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2292        operating environments where this matters.
2293    
2294     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2295        PCRE options from within the C++ wrapper.
2296    
2297     5. Named capturing subpatterns were not being correctly counted when a pattern
2298        was compiled. This caused two problems: (a) If there were more than 100
2299        such subpatterns, the calculation of the memory needed for the whole
2300        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2301        back references of the form \12, where the number was greater than 9, were
2302        not recognized as back references, even though there were sufficient
2303        previous subpatterns.
2304    
2305     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2306        versions of gcc, e.g. 2.95.4.
2307    
2308    
2309    Version 6.1 21-Jun-05
2310    ---------------------
2311    
2312     1. There was one reference to the variable "posix" in pcretest.c that was not
2313        surrounded by "#if !defined NOPOSIX".
2314    
2315     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2316        the cross-check on the old pcre_info() function, for the benefit of the
2317        cut-down version of PCRE that is currently imported into Exim.
2318    
2319     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2320        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2321        patterns that start (?i) but I don't think that matters. The use of (?i) is
2322        just an example; this all applies to the other options as well.
2323    
2324     4. Since libtool seems to echo the compile commands it is issuing, the output
2325        from "make" can be reduced a bit by putting "@" in front of each libtool
2326        compile command.
2327    
2328     5. Patch from the folks at Google for configure.in to be a bit more thorough
2329        in checking for a suitable C++ installation before trying to compile the
2330        C++ stuff. This should fix a reported problem when a compiler was present,
2331        but no suitable headers.
2332    
2333     6. The man pages all had just "PCRE" as their title. I have changed them to
2334        be the relevant file name. I have also arranged that these names are
2335        retained in the file doc/pcre.txt, which is a concatenation in text format
2336        of all the man pages except the little individual ones for each function.
2337    
2338     7. The NON-UNIX-USE file had not been updated for the different set of source
2339        files that come with release 6. I also added a few comments about the C++
2340        wrapper.
2341    
2342    
2343    Version 6.0 07-Jun-05
2344    ---------------------
2345    
2346     1. Some minor internal re-organization to help with my DFA experiments.
2347    
2348     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2349        didn't matter for the library itself when fully configured, but did matter
2350        when compiling without UCP support, or within Exim, where the ucp files are
2351        not imported.
2352    
2353     3. Refactoring of the library code to split up the various functions into
2354        different source modules. The addition of the new DFA matching code (see
2355        below) to a single monolithic source would have made it really too
2356        unwieldy, quite apart from causing all the code to be include in a
2357        statically linked application, when only some functions are used. This is
2358        relevant even without the DFA addition now that patterns can be compiled in
2359        one application and matched in another.
2360    
2361        The downside of splitting up is that there have to be some external
2362        functions and data tables that are used internally in different modules of
2363        the library but which are not part of the API. These have all had their
2364        names changed to start with "_pcre_" so that they are unlikely to clash
2365        with other external names.
2366    
2367     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2368        a different (DFA) algorithm. Although it is slower than the original
2369        function, it does have some advantages for certain types of matching
2370        problem.
2371    
2372     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2373        including restarting after a partial match.
2374    
2375     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2376        defined when compiling for Windows was sent to me. I have put it into the
2377        code, though I have no means of testing or verifying it.
2378    
2379     7. Added the pcre_refcount() auxiliary function.
2380    
2381     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2382        match before or at the first newline in the subject string. In pcretest,
2383        the /f option on a pattern can be used to set this.
2384    
2385     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2386        would behave wrongly. This has been present in PCRE since release 4.0.
2387    
2388    10. A number of changes to the pcregrep command:
2389    
2390        (a) Refactored how -x works; insert ^(...)$ instead of setting
2391            PCRE_ANCHORED and checking the length, in preparation for adding
2392            something similar for -w.
2393    
2394        (b) Added the -w (match as a word) option.
2395    
2396        (c) Refactored the way lines are read and buffered so as to have more
2397            than one at a time available.
2398    
2399        (d) Implemented a pcregrep test script.
2400    
2401        (e) Added the -M (multiline match) option. This allows patterns to match
2402            over several lines of the subject. The buffering ensures that at least
2403            8K, or the rest of the document (whichever is the shorter) is available
2404            for matching (and similarly the previous 8K for lookbehind assertions).
2405    
2406        (f) Changed the --help output so that it now says
2407    
2408              -w, --word-regex(p)
2409    
2410            instead of two lines, one with "regex" and the other with "regexp"
2411            because that confused at least one person since the short forms are the
2412            same. (This required a bit of code, as the output is generated
2413            automatically from a table. It wasn't just a text change.)
2414    
2415        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2416            option but starts with a hyphen. Could be a pattern or a path name
2417            starting with a hyphen, for instance.
2418    
2419        (h) "-" can be given as a file name to represent stdin.
2420    
2421        (i) When file names are being printed, "(standard input)" is used for
2422            the standard input, for compatibility with GNU grep. Previously
2423            "<stdin>" was used.
2424    
2425        (j) The option --label=xxx can be used to supply a name to be used for
2426            stdin when file names are being printed. There is no short form.
2427    
2428        (k) Re-factored the options decoding logic because we are going to add
2429            two more options that take data. Such options can now be given in four
2430            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2431    
2432        (l) Added the -A, -B, and -C options for requesting that lines of context
2433            around matches be printed.
2434    
2435        (m) Added the -L option to print the names of files that do not contain
2436            any matching lines, that is, the complement of -l.
2437    
2438        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2439            continue to scan other files.
2440    
2441        (o) The -s option was incorrectly implemented. For compatibility with other
2442            greps, it now suppresses the error message for a non-existent or non-
2443            accessible file (but not the return code). There is a new option called
2444            -q that suppresses the output of matching lines, which was what -s was
2445            previously doing.
2446    
2447        (p) Added --include and --exclude options to specify files for inclusion
2448            and exclusion when recursing.
2449    
2450    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2451        Hopefully, it now does.
2452    
2453    12. Missing cast in pcre_study().
2454    
2455    13. Added an "uninstall" target to the makefile.
2456    
2457    14. Replaced "extern" in the function prototypes in Makefile.in with
2458        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2459        world, but is set differently for Windows.
2460    
2461    15. Added a second compiling function called pcre_compile2(). The only
2462        difference is that it has an extra argument, which is a pointer to an
2463        integer error code. When there is a compile-time failure, this is set
2464        non-zero, in addition to the error test pointer being set to point to an
2465        error message. The new argument may be NULL if no error number is required
2466        (but then you may as well call pcre_compile(), which is now just a
2467        wrapper). This facility is provided because some applications need a
2468        numeric error indication, but it has also enabled me to tidy up the way
2469        compile-time errors are handled in the POSIX wrapper.
2470    
2471    16. Added VPATH=.libs to the makefile; this should help when building with one
2472        prefix path and installing with another. (Or so I'm told by someone who
2473        knows more about this stuff than I do.)
2474    
2475    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2476        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2477        match everything, including newlines. This is not POSIX-compatible, but
2478        somebody wanted the feature. From pcretest it can be activated by using
2479        both the P and the s flags.
2480    
2481    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2482    
2483    19. libpcre.pc was being incorrectly installed as executable.
2484    
2485    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2486        it now also looks for '\r' so that it will work unmodified on Windows.
2487    
2488    21. Added Google's contributed C++ wrapper to the distribution.
2489    
2490    22. Added some untidy missing memory free() calls in pcretest, to keep
2491        Electric Fence happy when testing.
2492    
2493    
2494    
2495    Version 5.0 13-Sep-04
2496    ---------------------
2497    
2498     1. Internal change: literal characters are no longer packed up into items
2499        containing multiple characters in a single byte-string. Each character
2500        is now matched using a separate opcode. However, there may be more than one
2501        byte in the character in UTF-8 mode.
2502    
2503     2. The pcre_callout_block structure has two new fields: pattern_position and
2504        next_item_length. These contain the offset in the pattern to the next match
2505        item, and its length, respectively.
2506    
2507     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2508        insertion of callouts before each pattern item. Added the /C option to
2509        pcretest to make use of this.
2510    
2511     4. On the advice of a Windows user, the lines
2512    
2513          #if defined(_WIN32) || defined(WIN32)
2514          _setmode( _fileno( stdout ), 0x8000 );
2515          #endif  /* defined(_WIN32) || defined(WIN32) */
2516    
2517        have been added to the source of pcretest. This apparently does useful
2518        magic in relation to line terminators.
2519    
2520     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2521        for the benefit of those environments where the "b" makes a difference.
2522    
2523     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2524        to know about it. I have put a hack into configure.in that adds in code
2525        to set GCC=yes if CC=icc. This seems to end up at a point in the
2526        generated configure script that is early enough to affect the setting of
2527        compiler options, which is what is needed, but I have no means of testing
2528        whether it really works. (The user who reported this had patched the
2529        generated configure script, which of course I cannot do.)
2530    
2531        LATER: After change 22 below (new libtool files), the configure script
2532        seems to know about icc (and also ecc). Therefore, I have commented out
2533        this hack in configure.in.
2534    
2535     7. Added support for pkg-config (2 patches were sent in).
2536    
2537     8. Negated POSIX character classes that used a combination of internal tables
2538        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2539        [[:^ascii]]. Typically, they would match almost any characters. The other
2540        POSIX classes were not broken in this way.
2541    
2542     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2543        to find the match, as PCRE was deluded into thinking that the match had to
2544        start at the start point or following a newline. The same bug applied to
2545        patterns with negative forward assertions or any backward assertions
2546        preceding ".*" at the start, unless the pattern required a fixed first
2547        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2548    
2549    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2550        starting at the last subject character, bytes beyond the end of the subject
2551        string were read.
2552    
2553    11. Renamed the variable "class" as "classbits" to make life easier for C++
2554        users. (Previously there was a macro definition, but it apparently wasn't
2555        enough.)
2556    
2557    12. Added the new field "tables" to the extra data so that tables can be passed
2558        in at exec time, or the internal tables can be re-selected. This allows
2559        a compiled regex to be saved and re-used at a later time by a different
2560        program that might have everything at different addresses.
2561    
2562    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2563        -R library as well as a -L library.
2564    
2565    14. The debugging options of pcretest (-d on the command line or D on a
2566        pattern) showed incorrect output for anything following an extended class
2567        that contained multibyte characters and which was followed by a quantifier.
2568    
2569    15. Added optional support for general category Unicode character properties
2570        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2571        support. It adds about 90K to the size of the library. The meanings of the
2572        inbuilt class escapes such as \d and \s have NOT been changed.
2573    
2574    16. Updated pcredemo.c to include calls to free() to release the memory for the
2575        compiled pattern.
2576    
2577    17. The generated file chartables.c was being created in the source directory
2578        instead of in the building directory. This caused the build to fail if the
2579        source directory was different from the building directory, and was
2580        read-only.
2581    
2582    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2583        file. No doubt somebody will tell me if they don't make sense... Also added
2584        Dan Mooney's comments about building on OpenVMS.
2585    
2586    19. Added support for partial matching via the PCRE_PARTIAL option for
2587        pcre_exec() and the \P data escape in pcretest.
2588    
2589    20. Extended pcretest with 3 new pattern features:
2590    
2591        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2592              write the compiled pattern to the file whose name is "rest-of-line".
2593              This is a straight binary dump of the data, with the saved pointer to
2594              the character tables forced to be NULL. The study data, if any, is
2595              written too. After writing, pcretest reads a new pattern.
2596    
2597        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2598              compiled pattern from the given file. There must not be any
2599              occurrences of "<" in the file name (pretty unlikely); if there are,
2600              pcretest will instead treat the initial "<" as a pattern delimiter.
2601              After reading in the pattern, pcretest goes on to read data lines as
2602              usual.
2603    
2604        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2605              and 16-bit fields in a compiled pattern, to simulate a pattern that
2606              was compiled on a host of opposite endianness.
2607    
2608    21. The pcre-exec() function can now cope with patterns that were compiled on
2609        hosts of opposite endianness, with this restriction:
2610    
2611          As for any compiled expression that is saved and used later, the tables
2612          pointer field cannot be preserved; the extra_data field in the arguments
2613          to pcre_exec() should be used to pass in a tables address if a value
2614          other than the default internal tables were used at compile time.
2615    
2616    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2617        now diagnosed as an error. Previously, most of the time, a negative number
2618        would have been treated as zero, but if in addition "ovector" was passed as
2619        NULL, a crash could occur.
2620    
2621    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2622        new versions from the libtool 1.5 distribution (the last one is a copy of
2623        a file called libtool.m4). This seems to have fixed the need to patch
2624        "configure" to support Darwin 1.3 (which I used to do). However, I still
2625        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2626        workstation).
2627    
2628    24. Changed the PCRE licence to be the more standard "BSD" licence.
2629    
2630    
2631    Version 4.5 01-Dec-03
2632    ---------------------
2633    
2634     1. There has been some re-arrangement of the code for the match() function so
2635        that it can be compiled in a version that does not call itself recursively.
2636        Instead, it keeps those local variables that need separate instances for
2637        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2638        needs to "recurse". Keeping track of where control must go is done by means
2639        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2640        hide most of the details from the main code, and operates only if
2641        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2642        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2643        operating.
2644    
2645        To make it easier for callers to provide specially tailored get/free
2646        functions for this usage, two new functions, pcre_stack_malloc, and
2647        pcre_stack_free, are used. They are always called in strict stacking order,
2648        and the size of block requested is always the same.
2649    
2650        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2651        PCRE has been compiled to use the stack or the heap for recursion. The
2652        -C option of pcretest uses this to show which version is compiled.
2653    
2654        A new data escape \S, is added to pcretest; it causes the amounts of store
2655        obtained and freed by both kinds of malloc/free at match time to be added
2656        to the output.
2657    
2658     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2659        what's available on my current Linux desktop machine.
2660    
2661     3. When matching a UTF-8 string, the test for a valid string at the start has
2662        been extended. If start_offset is not zero, PCRE now checks that it points
2663        to a byte that is the start of a UTF-8 character. If not, it returns
2664        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2665        this is necessary because there may be backward assertions in the pattern.
2666        When matching the same subject several times, it may save resources to use
2667        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2668    
2669     4. The code for checking the validity of UTF-8 strings has been tightened so
2670        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2671        containing "overlong sequences".
2672    
2673     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2674        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2675        should have been "&", but it just so happened that all the cases this let
2676        through by mistake were picked up later in the function.
2677    
2678     6. I had used a variable called "isblank" - this is a C99 function, causing
2679        some compilers to warn. To avoid this, I renamed it (as "blankclass").
2680    
2681     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2682        prompting; (b) run "./pcretest /dev/null" at the start of the test script
2683        so the version is shown; (c) stop "make test" echoing "./RunTest".
2684    
2685     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2686    
2687     9. The prototype for memmove() for systems that don't have it was using
2688        size_t, but the inclusion of the header that defines size_t was later. I've
2689        moved the #includes for the C headers earlier to avoid this.
2690    
2691    10. Added some adjustments to the code to make it easier to compiler on certain
2692        special systems:
2693    
2694          (a) Some "const" qualifiers were missing.
2695          (b) Added the macro EXPORT before all exported functions; by default this
2696              is defined to be empty.
2697          (c) Changed the dftables auxiliary program (that builds chartables.c) so
2698              that it reads its output file name as an argument instead of writing
2699              to the standard output and assuming this can be redirected.
2700    
2701    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2702        class containing characters with values greater than 255, PCRE compilation
2703        went into a loop.
2704    
2705    12. A recursive reference to a subpattern that was within another subpattern
2706        that had a minimum quantifier of zero caused PCRE to crash. For example,
2707        (x(y(?2))z)? provoked this bug with a subject that got as far as the
2708        recursion. If the recursively-called subpattern itself had a zero repeat,
2709        that was OK.
2710    
2711    13. In pcretest, the buffer for reading a data line was set at 30K, but the
2712        buffer into which it was copied (for escape processing) was still set at
2713        1024, so long lines caused crashes.
2714    
2715    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2716        "internal error: code overflow...". This applied to any character class
2717        that was followed by a possessive quantifier.
2718    
2719    15. Modified the Makefile to add libpcre.la as a prerequisite for
2720        libpcreposix.la because I was told this is needed for a parallel build to
2721        work.
2722    
2723    16. If a pattern that contained .* following optional items at the start was
2724        studied, the wrong optimizing data was generated, leading to matching
2725        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2726        matching string must start with a or b or c. The correct conclusion for
2727        this pattern is that a match can start with any character.
2728    
2729    
2730    Version 4.4 13-Aug-03
2731    ---------------------
2732    
2733     1. In UTF-8 mode, a character class containing characters with values between
2734        127 and 255 was not handled correctly if the compiled pattern was studied.
2735        In fixing this, I have also improved the studying algorithm for such
2736        classes (slightly).
2737    
2738     2. Three internal functions had redundant arguments passed to them. Removal
2739        might give a very teeny performance improvement.
2740    
2741     3. Documentation bug: the value of the capture_top field in a callout is *one
2742        more than* the number of the hightest numbered captured substring.
2743    
2744     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2745        in incorrectly linking with a previously installed version. They now link
2746        explicitly with libpcre.la.
2747    
2748     5. configure.in no longer needs to recognize Cygwin specially.
2749    
2750     6. A problem in pcre.in for Windows platforms is fixed.
2751    
2752     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2753        pcretest, it used to include the size of the study block as part of its
2754        output. Unfortunately, the structure contains a field that has a different
2755        size on different hardware architectures. This meant that the tests that
2756        showed this size failed. As the block is currently always of a fixed size,
2757        this information isn't actually particularly useful in pcretest output, so
2758        I have just removed it.
2759    
2760     8. Three pre-processor statements accidentally did not start in column 1.
2761        Sadly, there are *still* compilers around that complain, even though
2762        standard C has not required this for well over a decade. Sigh.
2763    
2764     9. In pcretest, the code for checking callouts passed small integers in the
2765        callout_data field, which is a void * field. However, some picky compilers
2766        complained about the casts involved for this on 64-bit systems. Now
2767        pcretest passes the address of the small integer instead, which should get
2768        rid of the warnings.
2769    
2770    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2771        both compile and run time, and gives an error if an invalid UTF-8 sequence
2772        is found. There is a option for disabling this check in cases where the
2773        string is known to be correct and/or the maximum performance is wanted.
2774    
2775    11. In response to a bug report, I changed one line in Makefile.in from
2776    
2777            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2778        to
2779            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2780    
2781        to look similar to other lines, but I have no way of telling whether this
2782        is the right thing to do, as I do not use Windows. No doubt I'll get told
2783        if it's wrong...
2784    
2785    
2786    Version 4.3 21-May-03
2787    ---------------------
2788    
2789    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2790       Makefile.
2791    
2792    2. Some refactoring to improve the quality of the code:
2793    
2794       (i)   The utf8_table... variables are now declared "const".
2795    
2796       (ii)  The code for \cx, which used the "case flipping" table to upper case
2797             lower case letters, now just substracts 32. This is ASCII-specific,
2798             but the whole concept of \cx is ASCII-specific, so it seems
2799             reasonable.
2800    
2801       (iii) PCRE was using its character types table to recognize decimal and
2802             hexadecimal digits in the pattern. This is silly, because it handles
2803             only 0-9, a-f, and A-F, but the character types table is locale-
2804             specific, which means strange things might happen. A private
2805             table is now used for this - though it costs 256 bytes, a table is
2806             much faster than multiple explicit tests. Of course, the standard
2807             character types table is still used for matching digits in subject
2808             strings against \d.
2809    
2810       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2811             ending in _t are). So I've renamed it as ESC_tee.
2812    
2813    3. The first argument for regexec() in the POSIX wrapper should have been
2814       defined as "const".
2815    
2816    4. Changed pcretest to use malloc() for its buffers so that they can be
2817       Electric Fenced for debugging.
2818    
2819    5. There were several places in the code where, in UTF-8 mode, PCRE would try
2820       to read one or more bytes before the start of the subject string. Often this
2821       had no effect on PCRE's behaviour, but in some circumstances it could
2822       provoke a segmentation fault.
2823    
2824    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2825       to try to read one or more bytes before the start of the subject string.
2826    
2827    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2828       UTF-8 support could misbehave in various ways if the subject string
2829       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2830       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2831       back over UTF-8 characters.)
2832    
2833    
2834    Version 4.2 14-Apr-03
2835    ---------------------
2836    
2837    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2838    
2839    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2840         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2841         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2842         [WIN_PREFIX]: new variable, "cyg" for Cygwin
2843         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2844           and BUILD_EXEEXT
2845         Note: automatic setting of the BUILD variables is not yet working
2846         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2847           compile-time but not at link-time
2848         [LINK]: use for linking executables only
2849         make different versions for Windows and non-Windows
2850         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2851           libraries
2852         [LINK_FOR_BUILD]: new variable
2853         [OBJEXT]: use throughout
2854         [EXEEXT]: use throughout
2855         <winshared>: new target
2856         <wininstall>: new target
2857         <dftables.o>: use native compiler
2858         <dftables>: use native linker
2859         <install>: handle Windows platform correctly
2860         <clean>: ditto
2861         <check>: ditto
2862         copy DLL to top builddir before testing
2863    
2864       As part of these changes, -no-undefined was removed again. This was reported
2865       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2866       in any case.
2867    
2868    3. Some tidies to get rid of compiler warnings:
2869    
2870       . In the match_data structure, match_limit was an unsigned long int, whereas
2871         match_call_count was an int. I've made them both unsigned long ints.
2872    
2873       . In pcretest the fact that a const uschar * doesn't automatically cast to
2874         a void * provoked a warning.
2875    
2876       . Turning on some more compiler warnings threw up some "shadow" variables
2877         and a few more missing casts.
2878    
2879    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2880       option, a class that contained a single character with a value between 128
2881       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2882    
2883    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2884       option, a class that contained several characters, but with at least one
2885       whose value was between 128 and 255 caused PCRE to crash.
2886    
2887    
2888    Version 4.1 12-Mar-03
2889    ---------------------
2890    
2891    1. Compiling with gcc -pedantic found a couple of places where casts were
2892    needed, and a string in dftables.c that was longer than standard compilers are
2893    required to support.
2894    
2895    2. Compiling with Sun's compiler found a few more places where the code could
2896    be tidied up in order to avoid warnings.
2897    
2898    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2899    first of these names is deprecated in the latest Autoconf in favour of the name
2900    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2901    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2902    analogy I have changed it to CFLAGS_FOR_BUILD.
2903    
2904    4. Added -no-undefined to the linking command in the Makefile, because this is
2905    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2906    linking step for the pcreposix library.
2907    
2908    5. PCRE was failing to diagnose the case of two named groups with the same
2909    name.
2910    
2911    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2912    literal character that is needed in the subject for a match, and scans along to
2913    ensure that it is present before embarking on the full matching process. This
2914    saves time in cases of nested unlimited repeats that are never going to match.
2915    Problem: the scan can take a lot of time if the subject is very long (e.g.
2916    megabytes), thus penalizing straightforward matches. It is now done only if the
2917    amount of subject to be scanned is less than 1000 bytes.
2918    
2919    7. A lesser problem with the same optimization is that it was recording the
2920    first character of an anchored pattern as "needed", thus provoking a search
2921    right along the subject, even when the first match of the pattern was going to
2922    fail. The "needed" character is now not set for anchored patterns, unless it
2923    follows something in the pattern that is of non-fixed length. Thus, it still
2924    fulfils its original purpose of finding quick non-matches in cases of nested
2925    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2926    
2927    
2928    Version 4.0 17-Feb-03
2929    ---------------------
2930    
2931    1. If a comment in an extended regex that started immediately after a meta-item
2932    extended to the end of string, PCRE compiled incorrect data. This could lead to
2933    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2934    
2935    2. Moved to autoconf 2.53 and libtool 1.4.2.
2936    
2937    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2938    the special perltest8 script is no longer needed - all the tests can be run
2939    from a single perltest script.
2940    
2941    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2942    by \s. It has now been removed in PCRE. This means it isn't recognized as
2943    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2944    class [:space:] *does* include VT, thereby creating a mess.
2945    
2946    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2947    space and tab.
2948    
2949    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2950    its new features into the main test script, reducing the number of scripts.
2951    
2952    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2953    were backward compatible, and made the (?i) apply to the whole pattern, as if
2954    /i were given. Now it behaves more logically, and applies the option setting
2955    only to what follows. PCRE has been changed to follow suit. However, if it
2956    finds options settings right at the start of the pattern, it extracts them into
2957    the global options, as before. Thus, they show up in the info data.
2958    
2959    8. Added support for the \Q...\E escape sequence. Characters in between are
2960    treated as literals. This is slightly different from Perl in that $ and @ are
2961    also handled as literals inside the quotes. In Perl, they will cause variable
2962    interpolation. Note the following examples:
2963    
2964        Pattern            PCRE matches      Perl matches
2965    
2966        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
2967        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
2968        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
2969    
2970    For compatibility with Perl, \Q...\E sequences are recognized inside character
2971    classes as well as outside them.
2972    
2973    9. Re-organized 3 code statements in pcretest to avoid "overflow in
2974    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2975    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2976    signed/unsigned warnings.
2977    
2978    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2979    option for pcretest, so I've replaced it by a simple function that does just
2980    that job.
2981    
2982    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2983    "pcregrep -".
2984    
2985    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2986    Java package. This provides some syntactic sugar for simple cases of what my
2987    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2988    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2989    item, you can use this simplified notation. Note that only makes sense with
2990    greedy quantifiers. Consequently, the use of the possessive quantifier forces
2991    greediness, whatever the setting of the PCRE_UNGREEDY option.
2992    
2993    13. A change of greediness default within a pattern was not taking effect at
2994    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2995    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2996    was abstracted outside.
2997    
2998    14. PCRE now supports the \G assertion. It is true when the current matching
2999    position is at the start point of the match. This differs from \A when the
3000    starting offset is non-zero. Used with the /g option of pcretest (or similar
3001    code), it works in the same way as it does for Perl's /g option. If all
3002    alternatives of a regex begin with \G, the expression is anchored to the start
3003    match position, and the "anchored" flag is set in the compiled expression.
3004    
3005    15. Some bugs concerning the handling of certain option changes within patterns
3006    have been fixed. These applied to options other than (?ims). For example,
3007    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3008    been the other way round. Some of this was related to change 7 above.
3009    
3010    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3011    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3012    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3013    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3014    
3015    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3016    mode. Unlike ".", it always matches newline, whatever the setting of
3017    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3018    assertions. Perl allows it, but it doesn't (in general) work because it can't
3019    calculate the length of the lookbehind. At least, that's the case for Perl
3020    5.8.0 - I've been told they are going to document that it doesn't work in
3021    future.
3022    
3023    18. Added an error diagnosis for escapes that PCRE does not support: these are
3024    \L, \l, \N, \P, \p, \U, \u, and \X.
3025    
3026    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3027    reading past the end of the pattern in cases such as /[abcd/.
3028    
3029    20. PCRE was getting more memory than necessary for patterns with classes that
3030    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3031    
3032    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3033    compiling PCRE for use with Virtual Pascal.
3034    
3035    22. Small fix to the Makefile to make it work properly if the build is done
3036    outside the source tree.
3037    
3038    23. Added a new extension: a condition to go with recursion. If a conditional
3039    subpattern starts with (?(R) the "true" branch is used if recursion has
3040    happened, whereas the "false" branch is used only at the top level.
3041    
3042    24. When there was a very long string of literal characters (over 255 bytes
3043    without UTF support, over 250 bytes with UTF support), the computation of how
3044    much memory was required could be incorrect, leading to segfaults or other
3045    strange effects.
3046    
3047    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3048    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3049    there was a subsequent back reference to those brackets. This meant that, for
3050    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3051    possible to check for precisely this case. All we can do is abandon the
3052    optimization if .* occurs inside capturing brackets when there are any back
3053    references whatsoever. (See below for a better fix that came later.)
3054    
3055    26. The handling of the optimization for finding the first character of a
3056    non-anchored pattern, and for finding a character that is required later in the
3057    match were failing in some cases. This didn't break the matching; it just
3058    failed to optimize when it could. The way this is done has been re-implemented.
3059    
3060    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3061    
3062    28. Added a new feature that provides some of the functionality that Perl
3063    provides with (?{...}). The facility is termed a "callout". The way it is done
3064    in PCRE is for the caller to provide an optional function, by setting
3065    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3066    global variable. By default it is unset, which disables all calling out. To get
3067    the function called, the regex must include (?C) at appropriate points. This
3068    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3069    This provides a means of identifying different callout points. When PCRE
3070    reaches such a point in the regex, if pcre_callout has been set, the external
3071    function is called. It is provided with data in a structure called
3072    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3073    matching continues; if it returns a non-zero value, the match at the current
3074    point fails. However, backtracking will occur if possible. [This was changed
3075    later and other features added - see item 49 below.]
3076    
3077    29. pcretest is upgraded to test the callout functionality. It provides a
3078    callout function that displays information. By default, it shows the start of
3079    the match and the current position in the text. There are some new data escapes
3080    to vary what happens:
3081    
3082        \C+         in addition, show current contents of captured substrings
3083        \C-         do not supply a callout function
3084        \C!n        return 1 when callout number n is reached
3085        \C!n!m      return 1 when callout number n is reached for the mth time
3086    
3087    30. If pcregrep was called with the -l option and just a single file name, it
3088    output "<stdin>" if a match was found, instead of the file name.
3089    
3090    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3091    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3092    pcre_exec(). This saves a malloc/free per call. The default value of
3093    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3094    when configuring.
3095    
3096    32. The default maximum size of a compiled pattern is 64K. There have been a
3097    few cases of people hitting this limit. The code now uses macros to handle the
3098    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3099    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3100    configuring. Tests 2 and 5 work only with 2-byte links because they output
3101    debugging information about compiled patterns.
3102    
3103    33. Internal code re-arrangements:
3104    
3105    (a) Moved the debugging function for printing out a compiled regex into
3106        its own source file (printint.c) and used #include to pull it into
3107        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3108        separate copies.
3109    
3110    (b) Defined the list of op-code names for debugging as a macro in
3111        internal.h so that it is next to the definition of the opcodes.
3112    
3113    (c) Defined a table of op-code lengths for simpler skipping along compiled
3114        code. This is again a macro in internal.h so that it is next to the
3115        definition of the opcodes.
3116    
3117    34. Added support for recursive calls to individual subpatterns, along the
3118    lines of Robin Houston's patch (but implemented somewhat differently).
3119    
3120    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3121    allow it to read and process whole directories in Win32. This code was
3122    contributed by Lionel Fourquaux; it has not been tested by me.
3123    
3124    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3125    used to name a group. Names consist of alphanumerics and underscores, and must
3126    be unique. Back references use the syntax (?P=name) and recursive calls use
3127    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3128    numbers. The function pcre_fullinfo() can be used after compilation to extract
3129    a name/number map. There are three relevant calls:
3130    
3131      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3132      PCRE_INFO_NAMECOUNT            yields the number of entries
3133      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3134    
3135    The map is a vector of fixed-size entries. The size of each entry depends on
3136    the length of the longest name used. The first two bytes of each entry are the
3137    group number, most significant byte first. There follows the corresponding
3138    name, zero terminated. The names are in alphabetical order.
3139    
3140    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3141    case instead of 255. Making it the same both with and without UTF-8 support
3142    means that the same test output works with both.
3143    
3144    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3145    calling malloc() with a zero argument.
3146    
3147    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3148    optimization. I've improved things by keeping a bitmap of backreferences with
3149    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3150    fact referenced, the optimization can be applied. It is unlikely that a
3151    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3152    the match to follow \n) will appear inside brackets with a number greater than
3153    31, but if it does, any back reference > 31 suppresses the optimization.
3154    
3155    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3156    of disabling numbered capturing parentheses. Any opening parenthesis that is
3157    not followed by ? behaves as if it were followed by ?: but named parentheses
3158    can still be used for capturing (and they will acquire numbers in the usual
3159    way).
3160    
3161    41. Redesigned the return codes from the match() function into yes/no/error so
3162    that errors can be passed back from deep inside the nested calls. A malloc
3163    failure while inside a recursive subpattern call now causes the
3164    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3165    
3166    42. It is now possible to set a limit on the number of times the match()
3167    function is called in a call to pcre_exec(). This facility makes it possible to
3168    limit the amount of recursion and backtracking, though not in a directly
3169    obvious way, because the match() function is used in a number of different
3170    circumstances. The count starts from zero for each position in the subject
3171    string (for non-anchored patterns). The default limit is, for compatibility, a
3172    large number, namely 10 000 000. You can change this in two ways:
3173    
3174    (a) When configuring PCRE before making, you can use --with-match-limit=n
3175        to set a default value for the compiled library.
3176    
3177    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3178        a different value is set. See 45 below.
3179    
3180    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3181    
3182    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3183    of things that can be changed at compile time. The first argument specifies
3184    what is wanted and the second points to where the information is to be placed.
3185    The current list of available information is:
3186    
3187      PCRE_CONFIG_UTF8
3188    
3189    The output is an integer that is set to one if UTF-8 support is available;
3190    otherwise it is set to zero.
3191    
3192      PCRE_CONFIG_NEWLINE
3193    
3194    The output is an integer that it set to the value of the code that is used for
3195    newline. It is either LF (10) or CR (13).
3196    
3197      PCRE_CONFIG_LINK_SIZE
3198    
3199    The output is an integer that contains the number of bytes used for internal
3200    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3201    
3202      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3203    
3204    The output is an integer that contains the threshold above which the POSIX
3205    interface uses malloc() for output vectors. See item 31 above.
3206    
3207      PCRE_CONFIG_MATCH_LIMIT
3208    
3209    The output is an unsigned integer that contains the default limit of the number
3210    of match() calls in a pcre_exec() execution. See 42 above.
3211    
3212    44. pcretest has been upgraded by the addition of the -C option. This causes it
3213    to extract all the available output from the new pcre_config() function, and to
3214    output it. The program then exits immediately.
3215    
3216    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3217    order to support additional features. One way would have been to define
3218    pcre_exec2() (for example) with extra arguments, but this would not have been
3219    extensible, and would also have required all calls to the original function to
3220    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3221    is used for passing in "extra" data from pcre_study().
3222    
3223    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3224    contains the following fields:
3225    
3226      flags         a bitmap indicating which of the following fields are set
3227      study_data    opaque data from pcre_study()
3228      match_limit   a way of specifying a limit on match() calls for a specific
3229                      call to pcre_exec()
3230      callout_data  data for callouts (see 49 below)
3231    
3232    The flag bits are also defined in pcre.h, and are
3233    
3234      PCRE_EXTRA_STUDY_DATA
3235      PCRE_EXTRA_MATCH_LIMIT
3236      PCRE_EXTRA_CALLOUT_DATA
3237    
3238    The pcre_study() function now returns one of these new pcre_extra blocks, with
3239    the actual study data pointed to by the study_data field, and the
3240    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3241    before. That is, this change is entirely upwards-compatible and requires no
3242    change to existing code.
3243    
3244    If you want to pass in additional data to pcre_exec(), you can either place it
3245    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3246    block.
3247    
3248    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3249    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3250    times with different match limits, until it finds the minimum value needed for
3251    pcre_exec() to complete. The value is then output. This can be instructive; for
3252    most simple matches the number is quite small, but for pathological cases it
3253    gets very large very quickly.
3254    
3255    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3256    returns the size of the data block pointed to by the study_data field in a
3257    pcre_extra block, that is, the value that was passed as the argument to
3258    pcre_malloc() when PCRE was getting memory in which to place the information
3259    created by pcre_study(). The fourth argument should point to a size_t variable.
3260    pcretest has been extended so that this information is shown after a successful
3261    pcre_study() call when information about the compiled regex is being displayed.
3262    
3263    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3264    because what follows is always an absolute path. (Later: it turns out that this
3265    is more than cosmetic for MinGW, because it doesn't like empty path
3266    components.)
3267    
3268    49. Some changes have been made to the callout feature (see 28 above):
3269    
3270    (i)  A callout function now has three choices for what it returns:
3271    
3272           0  =>  success, carry on matching
3273         > 0  =>  failure at this point, but backtrack if possible
3274         < 0  =>  serious error, return this value from pcre_exec()
3275    
3276         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3277         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3278         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3279         use by callout functions. It will never be used by PCRE itself.
3280    
3281    (ii) The pcre_extra structure (see 45 above) has a void * field called
3282         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3283         pcre_callout_block structure has a field of the same name. The contents of
3284         the field passed in the pcre_extra structure are passed to the callout
3285         function in the corresponding field in the callout block. This makes it
3286         easier to use the same callout-containing regex from multiple threads. For
3287         testing, the pcretest program has a new data escape
3288    
3289           \C*n        pass the number n (may be negative) as callout_data
3290    
3291         If the callout function in pcretest receives a non-zero value as
3292         callout_data, it returns that value.
3293    
3294    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3295    there were some redundant $(CFLAGS) in commands that are now specified as
3296    $(LINK), which already includes $(CFLAGS).
3297    
3298    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3299    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3300    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3301    one-byte characters throughout. Note that case-insensitive matching applies
3302    only to characters whose values are less than 256. PCRE doesn't support the
3303    notion of cases for higher-valued characters.
3304    
3305    (i)   A character class whose characters are all within 0-255 is handled as
3306          a bit map, and the map is inverted for negative classes. Previously, a
3307          character > 255 always failed to match such a class; however it should
3308          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3309    
3310    (ii)  A negated character class with a single character < 255 is coded as
3311          "not this character" (OP_NOT). This wasn't working properly when the test
3312          character was multibyte, either singly or repeated.
3313    
3314    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3315          mode, for example: \x{100}{2,3}.
3316    
3317    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3318          singly or repeated) now correctly test multibyte characters. However,
3319          PCRE doesn't recognize any characters with values greater than 255 as
3320          digits, spaces, or word characters. Such characters always match \D, \S,
3321          and \W, and never match \d, \s, or \w.
3322    
3323    (v)   Classes may now contain characters and character ranges with values
3324          greater than 255. For example: [ab\x{100}-\x{400}].
3325    
3326    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3327          PCRE in UTF-8 mode.
3328    
3329    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3330    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3331    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3332    value.)
3333    
3334    53. The single man page has become too large. I have therefore split it up into
3335    a number of separate man pages. These also give rise to individual HTML pages;
3336    these are now put in a separate directory, and there is an index.html page that
3337    lists them all. Some hyperlinking between the pages has been installed.
3338    
3339    54. Added convenience functions for handling named capturing parentheses.
3340    
3341    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3342    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3343    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3344    are faulted.
3345    
3346    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3347    calling configure. These values are used when compiling the dftables.c program
3348    which is run to generate the source of the default character tables. They
3349    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3350    you will need to set these values.
3351    
3352    57. Updated the building process for Windows DLL, as provided by Fred Cox.
3353    
3354    
3355    Version 3.9 02-Jan-02
3356    ---------------------
3357    
3358    1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3359    
3360    2. If --disable-static was given, the building process failed when trying to
3361    build pcretest and pcregrep. (For some reason it was using libtool to compile
3362    them, which is not right, as they aren't part of the library.)
3363    
3364    
3365    Version 3.8 18-Dec-01
3366    ---------------------
3367    
3368    1. The experimental UTF-8 code was completely screwed up. It was packing the
3369    bytes in the wrong order. How dumb can you get?
3370    
3371    
3372    Version 3.7 29-Oct-01
3373    ---------------------
3374    
3375    1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3376    This caused pcretest, when used on the test data, to segfault. Unfortunately,
3377    this didn't happen under Solaris 8, where I normally test things.
3378    
3379    2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3380    doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3381    isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3382    this fix an hour or so after the initial 3.7 release.)
3383    
3384    
3385    Version 3.6 23-Oct-01
3386    ---------------------
3387    
3388    1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3389    offsets passed as NULL with zero offset count.
3390    
3391    2. The config.guess and config.sub files had not been updated when I moved to
3392    the latest autoconf.
3393    
3394    
3395    Version 3.5 15-Aug-01
3396    ---------------------
3397    
3398    1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3399    had been forgotten.
3400    
3401    2. By using declared but undefined structures, we can avoid using "void"
3402    definitions in pcre.h while keeping the internal definitions of the structures
3403    private.
3404    
3405    3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3406    user point of view, this means that both static and shared libraries are built
3407    by default, but this can be individually controlled. More of the work of
3408    handling this static/shared cases is now inside libtool instead of PCRE's make
3409    file.
3410    
3411    4. The pcretest utility is now installed along with pcregrep because it is
3412    useful for users (to test regexs) and by doing this, it automatically gets
3413    relinked by libtool. The documentation has been turned into a man page, so
3414    there are now .1, .txt, and .html versions in /doc.
3415    
3416    5. Upgrades to pcregrep:
3417       (i)   Added long-form option names like gnu grep.
3418       (ii)  Added --help to list all options with an explanatory phrase.
3419       (iii) Added -r, --recursive to recurse into sub-directories.
3420       (iv)  Added -f, --file to read patterns from a file.
3421    
3422    6. pcre_exec() was referring to its "code" argument before testing that
3423    argument for NULL (and giving an error if it was NULL).
3424    
3425    7. Upgraded Makefile.in to allow for compiling in a different directory from
3426    the source directory.
3427    
3428    8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3429    options bits, the pointer it was passed was to an int instead of to an unsigned
3430    long int. This mattered only on 64-bit systems.
3431    
3432    9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3433    generated) instead of pcre.in, which it its source. Also made the same change
3434    in several of the .c files.
3435    
3436    10. A new release of gcc defines printf() as a macro, which broke pcretest
3437    because it had an ifdef in the middle of a string argument for printf(). Fixed
3438    by using separate calls to printf().
3439    
3440    11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3441    script, to force use of CR or LF instead of \n in the source. On non-Unix
3442    systems, the value can be set in config.h.
3443    
3444    12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3445    absolute limit. Changed the text of the error message to make this clear, and
3446    likewise updated the man page.
3447    
3448    13. The limit of 99 on the number of capturing subpatterns has been removed.
3449    The new limit is 65535, which I hope will not be a "real" limit.
3450    
3451    
3452    Version 3.4 22-Aug-00
3453    ---------------------
3454    
3455    1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3456    
3457    2. Diagnose condition (?(0) as an error instead of crashing on matching.
3458    
3459    
3460    Version 3.3 01-Aug-00
3461    ---------------------
3462    
3463    1. If an octal character was given, but the value was greater than \377, it
3464    was not getting masked to the least significant bits, as documented. This could
3465    lead to crashes in some systems.
3466    
3467    2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3468    the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3469    
3470    3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3471    These just pass their arguments on to (pcre_free)(), but they are provided
3472    because some uses of PCRE bind it to non-C systems that can call its functions,
3473    but cannot call free() or pcre_free() directly.
3474    
3475    4. Add "make test" as a synonym for "make check". Corrected some comments in
3476    the Makefile.
3477    
3478    5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3479    Makefile.
3480    
3481    6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3482    command called pgrep for grepping around the active processes.
3483    
3484    7. Added the beginnings of support for UTF-8 character strings.
3485    
3486    8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3487    RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3488    the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3489    out for the ar command.)
3490    
3491    
3492    Version 3.2 12-May-00
3493    ---------------------
3494    
3495    This is purely a bug fixing release.
3496    
3497    1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3498    of ZA. This was just one example of several cases that could provoke this bug,
3499    which was introduced by change 9 of version 2.00. The code for breaking
3500    infinite loops after an iteration that matches an empty string was't working
3501    correctly.
3502    
3503    2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3504    when matched against abbab (for example). After matching an empty string, it
3505    wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3506    caused it to match further down the string than it should.
3507    
3508    3. The code contained an inclusion of sys/types.h. It isn't clear why this
3509    was there because it doesn't seem to be needed, and it causes trouble on some
3510    systems, as it is not a Standard C header. It has been removed.
3511    
3512    4. Made 4 silly changes to the source to avoid stupid compiler warnings that
<