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

Diff of /code/trunk/ChangeLog

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

revision 63 by nigel, Sat Feb 24 21:40:03 2007 UTC revision 595 by ph10, Mon May 2 10:33:29 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 4.00 17-Feb-03  Version 8.13 30-Apr-2011
5    ------------------------
6    
7    1.  The Unicode data tables have been updated to Unicode 6.0.0.
8    
9    2.  Two minor typos in pcre_internal.h have been fixed.
10    
11    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
12        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
13        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
14    
15    4.  There were a number of related bugs in the code for matching backrefences
16        caselessly in UTF-8 mode when codes for the characters concerned were
17        different numbers of bytes. For example, U+023A and U+2C65 are an upper
18        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
19        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
20        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
21        2-byte code at the end of the subject (it thought there wasn't enough data
22        left).
23    
24    
25    Version 8.12 15-Jan-2011
26    ------------------------
27    
28    1.  Fixed some typos in the markup of the man pages, and wrote a script that
29        checks for such things as part of the documentation building process.
30    
31    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
32        --match-limit and --recursion-limit options (added for 8.11). In
33        particular, this made one of the standard tests fail. (The integer value
34        went into the wrong half of a long int.)
35    
36    3.  If the --colour option was given to pcregrep with -v (invert match), it
37        did strange things, either producing crazy output, or crashing. It should,
38        of course, ignore a request for colour when reporting lines that do not
39        match.
40    
41    4.  Another pcregrep bug caused similar problems if --colour was specified with
42        -M (multiline) and the pattern match finished with a line ending.
43    
44    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
45        matched in multiline mode, the following line was shown as part of the
46        match. This seems wrong, so I have changed it.
47    
48    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
49        the check for further matches in the same line (so they could be coloured)
50        to overrun the end of the current line. If another match was found, it was
51        incorrectly shown (and then shown again when found in the next line).
52    
53    7.  If pcregrep was compiled under Windows, there was a reference to the
54        function pcregrep_exit() before it was defined. I am assuming this was
55        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
56        reported by a user. I've moved the definition above the reference.
57    
58    
59    Version 8.11 10-Dec-2010
60    ------------------------
61    
62    1.  (*THEN) was not working properly if there were untried alternatives prior
63        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
64        backtracked to try for "b" instead of moving to the next alternative branch
65        at the same level (in this case, to look for "c"). The Perl documentation
66        is clear that when (*THEN) is backtracked onto, it goes to the "next
67        alternative in the innermost enclosing group".
68    
69    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
70        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
71        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
72        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
73        (*THEN).
74    
75    3.  If \s appeared in a character class, it removed the VT character from
76        the class, even if it had been included by some previous item, for example
77        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
78        of \s, but is part of the POSIX "space" class.)
79    
80    4.  A partial match never returns an empty string (because you can always
81        match an empty string at the end of the subject); however the checking for
82        an empty string was starting at the "start of match" point. This has been
83        changed to the "earliest inspected character" point, because the returned
84        data for a partial match starts at this character. This means that, for
85        example, /(?<=abc)def/ gives a partial match for the subject "abc"
86        (previously it gave "no match").
87    
88    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
89        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
90        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
91        has an implication that the given string is incomplete (because a partial
92        match is preferred over a full match). For this reason, these items now
93        give a partial match in this situation. [Aside: previously, the one case
94        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
95        match rather than a full match, which was wrong by the old rules, but is
96        now correct.]
97    
98    6.  There was a bug in the handling of #-introduced comments, recognized when
99        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
100        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
101        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
102        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
103        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
104        places in pcre_compile().
105    
106    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
107        comments when looking ahead for named forward references to subpatterns,
108        the only newline sequence it recognized was NL. It now handles newlines
109        according to the set newline convention.
110    
111    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
112        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
113        cater for a lack of strerror(). These oversights have been fixed.
114    
115    9.  Added --match-limit and --recursion-limit to pcregrep.
116    
117    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
118    
119    11. When the -o option was used, pcregrep was setting a return code of 1, even
120        when matches were found, and --line-buffered was not being honoured.
121    
122    12. Added an optional parentheses number to the -o and --only-matching options
123        of pcregrep.
124    
125    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
126        can match an empty string. The code to do it in pcretest and pcredemo
127        needed fixing:
128    
129        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
130            only one byte after an empty string match just before CRLF (this case
131            just got forgotten; "any" and "anycrlf" were OK).
132    
133        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
134            mode when an empty string match preceded an ASCII character followed by
135            a non-ASCII character. (The code for advancing by one character rather
136            than one byte was nonsense.)
137    
138        (c) The pcredemo.c sample program did not have any code at all to handle
139            the cases when CRLF is a valid newline sequence.
140    
141    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
142        as a starting offset was within the subject string. There is now a new
143        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
144        negative or greater than the length of the string. In order to test this,
145        pcretest is extended to allow the setting of negative starting offsets.
146    
147    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
148        starting offset points to the beginning of a UTF-8 character was
149        unnecessarily clumsy. I tidied it up.
150    
151    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
152        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
153    
154    17. Nobody had reported that the --include_dir option, which was added in
155        release 7.7 should have been called --include-dir (hyphen, not underscore)
156        for compatibility with GNU grep. I have changed it to --include-dir, but
157        left --include_dir as an undocumented synonym, and the same for
158        --exclude-dir, though that is not available in GNU grep, at least as of
159        release 2.5.4.
160    
161    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
162        characters from a string of bytes) have been redefined so as not to use
163        loops, in order to improve performance in some environments. At the same
164        time, I abstracted some of the common code into auxiliary macros to save
165        repetition (this should not affect the compiled code).
166    
167    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
168        compile-time error is now given if \c is not followed by an ASCII
169        character, that is, a byte less than 128. (In EBCDIC mode, the code is
170        different, and any byte value is allowed.)
171    
172    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
173        START_OPTIMIZE option, which is now allowed at compile time - but just
174        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
175        to pcregrep and other applications that have no direct access to PCRE
176        options. The new /Y option in pcretest sets this option when calling
177        pcre_compile().
178    
179    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
180        back references. Groups containing recursive back references were forced to
181        be atomic by that change, but in the case of named groups, the amount of
182        memory required was incorrectly computed, leading to "Failed: internal
183        error: code overflow". This has been fixed.
184    
185    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
186        pcretest.c, to avoid build problems in some Borland environments.
187    
188    
189    Version 8.10 25-Jun-2010
190    ------------------------
191    
192    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
193        THEN.
194    
195    2.  (*ACCEPT) was not working when inside an atomic group.
196    
197    3.  Inside a character class, \B is treated as a literal by default, but
198        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
199        causes the error). The code is unchanged, but I tidied the documentation.
200    
201    4.  Inside a character class, PCRE always treated \R and \X as literals,
202        whereas Perl faults them if its -w option is set. I have changed PCRE so
203        that it faults them when PCRE_EXTRA is set.
204    
205    5.  Added support for \N, which always matches any character other than
206        newline. (It is the same as "." when PCRE_DOTALL is not set.)
207    
208    6.  When compiling pcregrep with newer versions of gcc which may have
209        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
210        declared with attribute warn_unused_result" were given. Just casting the
211        result to (void) does not stop the warnings; a more elaborate fudge is
212        needed. I've used a macro to implement this.
213    
214    7.  Minor change to pcretest.c to avoid a compiler warning.
215    
216    8.  Added four artifical Unicode properties to help with an option to make
217        \s etc use properties (see next item). The new properties are: Xan
218        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
219    
220    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
221        use Unicode properties. (*UCP) at the start of a pattern can be used to set
222        this option. Modified pcretest to add /W to test this facility. Added
223        REG_UCP to make it available via the POSIX interface.
224    
225    10. Added --line-buffered to pcregrep.
226    
227    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
228        studied, and the match started with a letter with a code point greater than
229        127 whose first byte was different to the first byte of the other case of
230        the letter, the other case of this starting letter was not recognized
231        (#976).
232    
233    12. If a pattern that was studied started with a repeated Unicode property
234        test, for example, \p{Nd}+, there was the theoretical possibility of
235        setting up an incorrect bitmap of starting bytes, but fortunately it could
236        not have actually happened in practice until change 8 above was made (it
237        added property types that matched character-matching opcodes).
238    
239    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
240        possible starting bytes for non-anchored patterns.
241    
242    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
243        \R, and also a number of cases that involve Unicode properties, both
244        explicit and implicit when PCRE_UCP is set.
245    
246    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
247        input, it could crash or give wrong results if characters with values
248        greater than 0xc0 were present in the subject string. (Detail: it assumed
249        UTF-8 input when processing these items.)
250    
251    16. Added a lot of (int) casts to avoid compiler warnings in systems where
252        size_t is 64-bit (#991).
253    
254    17. Added a check for running out of memory when PCRE is compiled with
255        --disable-stack-for-recursion (#990).
256    
257    18. If the last data line in a file for pcretest does not have a newline on
258        the end, a newline was missing in the output.
259    
260    19. The default pcre_chartables.c file recognizes only ASCII characters (values
261        less than 128) in its various bitmaps. However, there is a facility for
262        generating tables according to the current locale when PCRE is compiled. It
263        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
264        characters, are recognized by isspace() and therefore were getting set in
265        these tables, and indeed these tables seem to approximate to ISO 8859. This
266        caused a problem in UTF-8 mode when pcre_study() was used to create a list
267        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
268        which of course cannot start UTF-8 characters. I have changed the code so
269        that only real ASCII characters (less than 128) and the correct starting
270        bytes for UTF-8 encodings are set for characters greater than 127 when in
271        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
272        altogether.)
273    
274    20. Added the /T option to pcretest so as to be able to run tests with non-
275        standard character tables, thus making it possible to include the tests
276        used for 19 above in the standard set of tests.
277    
278    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
279        reference to a subpattern the other side of a comment that contains an
280        opening parenthesis caused either an internal compiling error, or a
281        reference to the wrong subpattern.
282    
283    
284    Version 8.02 19-Mar-2010
285    ------------------------
286    
287    1.  The Unicode data tables have been updated to Unicode 5.2.0.
288    
289    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
290        configured.
291    
292    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
293        original author of that file, following a query about its status.
294    
295    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
296        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
297    
298    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
299        quantifier applied to a forward-referencing subroutine call, could compile
300        incorrect code or give the error "internal error: previously-checked
301        referenced subpattern not found".
302    
303    6.  Both MS Visual Studio and Symbian OS have problems with initializing
304        variables to point to external functions. For these systems, therefore,
305        pcre_malloc etc. are now initialized to local functions that call the
306        relevant global functions.
307    
308    7.  There were two entries missing in the vectors called coptable and poptable
309        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
310        I've fixed the data, and added a kludgy way of testing at compile time that
311        the lengths are correct (equal to the number of opcodes).
312    
313    8.  Following on from 7, I added a similar kludge to check the length of the
314        eint vector in pcreposix.c.
315    
316    9.  Error texts for pcre_compile() are held as one long string to avoid too
317        much relocation at load time. To find a text, the string is searched,
318        counting zeros. There was no check for running off the end of the string,
319        which could happen if a new error number was added without updating the
320        string.
321    
322    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
323    
324    11. \K was not working if it appeared in an atomic group or in a group that
325        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
326        \K is "not well defined" if used in an assertion. PCRE now accepts it if
327        the assertion is positive, but not if it is negative.
328    
329    12. Change 11 fortuitously reduced the size of the stack frame used in the
330        "match()" function of pcre_exec.c by one pointer. Forthcoming
331        implementation of support for (*MARK) will need an extra pointer on the
332        stack; I have reserved it now, so that the stack frame size does not
333        decrease.
334    
335    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
336        item in branch that calls a recursion is a subroutine call - as in the
337        second branch in the above example - was incorrectly given the compile-
338        time error "recursive call could loop indefinitely" because pcre_compile()
339        was not correctly checking the subroutine for matching a non-empty string.
340    
341    14. The checks for overrunning compiling workspace could trigger after an
342        overrun had occurred. This is a "should never occur" error, but it can be
343        triggered by pathological patterns such as hundreds of nested parentheses.
344        The checks now trigger 100 bytes before the end of the workspace.
345    
346    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
347    
348    
349    Version 8.01 19-Jan-2010
350    ------------------------
351    
352    1.  If a pattern contained a conditional subpattern with only one branch (in
353        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
354        computed the wrong minimum data length (which is of course zero for such
355        subpatterns). This could cause incorrect "no match" results.
356    
357    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
358        the pattern is reset in the first branch, pcre_compile() failed with
359        "internal error: code overflow at offset...". This happened only when
360        the reset was to the original external option setting. (An optimization
361        abstracts leading options settings into an external setting, which was the
362        cause of this.)
363    
364    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
365        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
366        assertion pattern did not match (meaning that the assertion was true), it
367        was incorrectly treated as false if the SKIP had been reached during the
368        matching. This also applied to assertions used as conditions.
369    
370    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
371        assertion subpattern, including such a pattern used as a condition,
372        unpredictable results occurred, instead of the error return
373        PCRE_ERROR_DFA_UITEM.
374    
375    5.  The C++ GlobalReplace function was not working like Perl for the special
376        situation when an empty string is matched. It now does the fancy magic
377        stuff that is necessary.
378    
379    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
380        removed. (These were left over from very, very early versions of PCRE.)
381    
382    7.  Some cosmetic changes to the code to make life easier when compiling it
383        as part of something else:
384    
385        (a) Change DEBUG to PCRE_DEBUG.
386    
387        (b) In pcre_compile(), rename the member of the "branch_chain" structure
388            called "current" as "current_branch", to prevent a collision with the
389            Linux macro when compiled as a kernel module.
390    
391        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
392            prevent a collision with the Linux macro when compiled as a kernel
393            module.
394    
395    8.  In pcre_compile() there are some checks for integer overflows that used to
396        cast potentially large values to (double). This has been changed to that
397        when building, a check for int64_t is made, and if it is found, it is used
398        instead, thus avoiding the use of floating point arithmetic. (There is no
399        other use of FP in PCRE.) If int64_t is not found, the fallback is to
400        double.
401    
402    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
403        2005 (difference between two addresses compared to an unsigned value).
404    
405    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
406        custom one, because of the following reported problem in Windows:
407    
408          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
409              under Win32.
410          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
411              therefore missing the function definition.
412          - The compiler thus generates a "C" signature for the test function.
413          - The linker fails to find the "C" function.
414          - PCRE fails to configure if asked to do so against libbz2.
415    
416    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
417        messages were output:
418    
419          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
420          rerunning libtoolize, to keep the correct libtool macros in-tree.
421          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
422    
423        I have done both of these things.
424    
425    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
426        most of the time, it *can* run out if it is given a pattern that contains a
427        runaway infinite recursion. I updated the discussion in the pcrestack man
428        page.
429    
430    13. Now that we have gone to the x.xx style of version numbers, the minor
431        version may start with zero. Using 08 or 09 is a bad idea because users
432        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
433        interpreted as invalid octal numbers. I've updated the previous comment in
434        configure.ac, and also added a check that gives an error if 08 or 09 are
435        used.
436    
437    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
438        causing partial matching to fail when the end of the subject matched \W
439        in a UTF-8 pattern where \W was quantified with a minimum of 3.
440    
441    15. There were some discrepancies between the declarations in pcre_internal.h
442        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
443        their definitions. The declarations used "const uschar *" and the
444        definitions used USPTR. Even though USPTR is normally defined as "const
445        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
446        reported that: "This difference in casting confuses some C++ compilers, for
447        example, SunCC recognizes above declarations as different functions and
448        generates broken code for hbpcre." I have changed the declarations to use
449        USPTR.
450    
451    16. GNU libtool is named differently on some systems. The autogen.sh script now
452        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
453        (FreeBSD).
454    
455    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
456        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
457        comment: "Figure out how to create a longlong from a string: strtoll and
458        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
459        instance, but it only takes 2 args instead of 3!"
460    
461    18. A subtle bug concerned with back references has been fixed by a change of
462        specification, with a corresponding code fix. A pattern such as
463        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
464        refers, was giving matches when it shouldn't. For example, xa=xaaa would
465        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
466        same bug. Such groups have to be quantified to be useful, or contained
467        inside another quantified group. (If there's no repetition, the reference
468        can never match.) The problem arises because, having left the group and
469        moved on to the rest of the pattern, a later failure that backtracks into
470        the group uses the captured value from the final iteration of the group
471        rather than the correct earlier one. I have fixed this in PCRE by forcing
472        any group that contains a reference to itself to be an atomic group; that
473        is, there cannot be any backtracking into it once it has completed. This is
474        similar to recursive and subroutine calls.
475    
476    
477    Version 8.00 19-Oct-09
478  ----------------------  ----------------------
479    
480    1.  The table for translating pcre_compile() error codes into POSIX error codes
481        was out-of-date, and there was no check on the pcre_compile() error code
482        being within the table. This could lead to an OK return being given in
483        error.
484    
485    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
486        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
487        in a Windows environment.
488    
489    3.  The pcregrep --count option prints the count for each file even when it is
490        zero, as does GNU grep. However, pcregrep was also printing all files when
491        --files-with-matches was added. Now, when both options are given, it prints
492        counts only for those files that have at least one match. (GNU grep just
493        prints the file name in this circumstance, but including the count seems
494        more useful - otherwise, why use --count?) Also ensured that the
495        combination -clh just lists non-zero counts, with no names.
496    
497    4.  The long form of the pcregrep -F option was incorrectly implemented as
498        --fixed_strings instead of --fixed-strings. This is an incompatible change,
499        but it seems right to fix it, and I didn't think it was worth preserving
500        the old behaviour.
501    
502    5.  The command line items --regex=pattern and --regexp=pattern were not
503        recognized by pcregrep, which required --regex pattern or --regexp pattern
504        (with a space rather than an '='). The man page documented the '=' forms,
505        which are compatible with GNU grep; these now work.
506    
507    6.  No libpcreposix.pc file was created for pkg-config; there was just
508        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
509    
510    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
511        when UCP support is not needed, by modifying the Python script that
512        generates it from Unicode data files. This should not matter if the module
513        is correctly used as a library, but I received one complaint about 50K of
514        unwanted data. My guess is that the person linked everything into his
515        program rather than using a library. Anyway, it does no harm.
516    
517    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
518        was a minimum greater than 1 for a wide character in a possessive
519        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
520        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
521        character. Chaos in the form of incorrect output or a compiling loop could
522        result.
523    
524    9.  The restrictions on what a pattern can contain when partial matching is
525        requested for pcre_exec() have been removed. All patterns can now be
526        partially matched by this function. In addition, if there are at least two
527        slots in the offset vector, the offset of the earliest inspected character
528        for the match and the offset of the end of the subject are set in them when
529        PCRE_ERROR_PARTIAL is returned.
530    
531    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
532        synonymous with PCRE_PARTIAL, for backwards compatibility, and
533        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
534        and may be more useful for multi-segment matching.
535    
536    11. Partial matching with pcre_exec() is now more intuitive. A partial match
537        used to be given if ever the end of the subject was reached; now it is
538        given only if matching could not proceed because another character was
539        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
540        string "Z", which now yields "no match" instead of "partial match". In the
541        case of pcre_dfa_exec(), "no match" is given if every matching path for the
542        final character ended with (*FAIL).
543    
544    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
545        if the pattern had a "must contain" character that was already found in the
546        earlier partial match, unless partial matching was again requested. For
547        example, with the pattern /dog.(body)?/, the "must contain" character is
548        "g". If the first part-match was for the string "dog", restarting with
549        "sbody" failed. This bug has been fixed.
550    
551    13. The string returned by pcre_dfa_exec() after a partial match has been
552        changed so that it starts at the first inspected character rather than the
553        first character of the match. This makes a difference only if the pattern
554        starts with a lookbehind assertion or \b or \B (\K is not supported by
555        pcre_dfa_exec()). It's an incompatible change, but it makes the two
556        matching functions compatible, and I think it's the right thing to do.
557    
558    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
559        so that the demonstration program is easily available in environments where
560        PCRE has not been installed from source.
561    
562    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
563        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
564        library.
565    
566    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
567        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
568        is not the first non-POSIX option to be added. Clearly some people find
569        these options useful.
570    
571    17. If a caller to the POSIX matching function regexec() passes a non-zero
572        value for nmatch with a NULL value for pmatch, the value of
573        nmatch is forced to zero.
574    
575    18. RunGrepTest did not have a test for the availability of the -u option of
576        the diff command, as RunTest does. It now checks in the same way as
577        RunTest, and also checks for the -b option.
578    
579    19. If an odd number of negated classes containing just a single character
580        interposed, within parentheses, between a forward reference to a named
581        subpattern and the definition of the subpattern, compilation crashed with
582        an internal error, complaining that it could not find the referenced
583        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
584        [The bug was that it was starting one character too far in when skipping
585        over the character class, thus treating the ] as data rather than
586        terminating the class. This meant it could skip too much.]
587    
588    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
589        /g option in pcretest when the pattern contains \K, which makes it possible
590        to have an empty string match not at the start, even when the pattern is
591        anchored. Updated pcretest and pcredemo to use this option.
592    
593    21. If the maximum number of capturing subpatterns in a recursion was greater
594        than the maximum at the outer level, the higher number was returned, but
595        with unset values at the outer level. The correct (outer level) value is
596        now given.
597    
598    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
599        PCRE did not set those parentheses (unlike Perl). I have now found a way to
600        make it do so. The string so far is captured, making this feature
601        compatible with Perl.
602    
603    23. The tests have been re-organized, adding tests 11 and 12, to make it
604        possible to check the Perl 5.10 features against Perl 5.10.
605    
606    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
607        pattern matches a fixed length string. PCRE did not allow this; now it
608        does. Neither allows recursion.
609    
610    25. I finally figured out how to implement a request to provide the minimum
611        length of subject string that was needed in order to match a given pattern.
612        (It was back references and recursion that I had previously got hung up
613        on.) This code has now been added to pcre_study(); it finds a lower bound
614        to the length of subject needed. It is not necessarily the greatest lower
615        bound, but using it to avoid searching strings that are too short does give
616        some useful speed-ups. The value is available to calling programs via
617        pcre_fullinfo().
618    
619    26. While implementing 25, I discovered to my embarrassment that pcretest had
620        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
621        study optimizations had never been tested with that matching function.
622        Oops. What is worse, even when it was passed study data, there was a bug in
623        pcre_dfa_exec() that meant it never actually used it. Double oops. There
624        were also very few tests of studied patterns with pcre_dfa_exec().
625    
626    27. If (?| is used to create subpatterns with duplicate numbers, they are now
627        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
628        on the other side of the coin, they are no longer allowed to have different
629        names, because these cannot be distinguished in PCRE, and this has caused
630        confusion. (This is a difference from Perl.)
631    
632    28. When duplicate subpattern names are present (necessarily with different
633        numbers, as required by 27 above), and a test is made by name in a
634        conditional pattern, either for a subpattern having been matched, or for
635        recursion in such a pattern, all the associated numbered subpatterns are
636        tested, and the overall condition is true if the condition is true for any
637        one of them. This is the way Perl works, and is also more like the way
638        testing by number works.
639    
640    
641    Version 7.9 11-Apr-09
642    ---------------------
643    
644    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
645        (pcretest), all targets were linked against these libraries. This included
646        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
647        libraries. This caused unwanted dependencies to be created. This problem
648        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
649        pcretest is linked with readline.
650    
651    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
652        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
653        moved outside it again, because FALSE and TRUE are already defined in AIX,
654        but BOOL is not.
655    
656    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
657        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
658    
659    4.  The pcregrep documentation said spaces were inserted as well as colons (or
660        hyphens) following file names and line numbers when outputting matching
661        lines. This is not true; no spaces are inserted. I have also clarified the
662        wording for the --colour (or --color) option.
663    
664    5.  In pcregrep, when --colour was used with -o, the list of matching strings
665        was not coloured; this is different to GNU grep, so I have changed it to be
666        the same.
667    
668    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
669        each matching line was coloured. Now it goes on to look for further matches
670        of any of the test patterns, which is the same behaviour as GNU grep.
671    
672    7.  A pattern that could match an empty string could cause pcregrep to loop; it
673        doesn't make sense to accept an empty string match in pcregrep, so I have
674        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
675        seems to be how GNU grep behaves.
676    
677    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
678        start or after a newline", because the conditional assertion was not being
679        correctly handled. The rule now is that both the assertion and what follows
680        in the first alternative must satisfy the test.
681    
682    9.  If auto-callout was enabled in a pattern with a conditional group whose
683        condition was an assertion, PCRE could crash during matching, both with
684        pcre_exec() and pcre_dfa_exec().
685    
686    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
687        used for matching.
688    
689    11. Unicode property support in character classes was not working for
690        characters (bytes) greater than 127 when not in UTF-8 mode.
691    
692    12. Added the -M command line option to pcretest.
693    
694    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
695    
696    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
697    
698    16. Added comments and documentation about mis-use of no_arg in the C++
699        wrapper.
700    
701    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
702        from Martin Jerabek that uses macro names for all relevant character and
703        string constants.
704    
705    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
706        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
707        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
708        these, but not everybody uses configure.
709    
710    19. A conditional group that had only one branch was not being correctly
711        recognized as an item that could match an empty string. This meant that an
712        enclosing group might also not be so recognized, causing infinite looping
713        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
714        with the subject "ab", where knowledge that the repeated group can match
715        nothing is needed in order to break the loop.
716    
717    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
718        exec(), but without supplying a callout function, matching went wrong.
719    
720    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
721        leak if the size of the offset vector was greater than 30. When the vector
722        is smaller, the saved offsets during recursion go onto a local stack
723        vector, but for larger vectors malloc() is used. It was failing to free
724        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
725        error, in fact).
726    
727    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
728        heapframe that is used only when UTF-8 support is enabled. This caused no
729        problem, but was untidy.
730    
731    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
732        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
733        included within another project.
734    
735    24. Steven Van Ingelgem's patches to add more options to the CMake support,
736        slightly modified by me:
737    
738          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
739              not building pcregrep.
740    
741          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
742              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
743    
744    25. Forward references, both numeric and by name, in patterns that made use of
745        duplicate group numbers, could behave incorrectly or give incorrect errors,
746        because when scanning forward to find the reference group, PCRE was not
747        taking into account the duplicate group numbers. A pattern such as
748        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
749    
750    26. Changed a few more instances of "const unsigned char *" to USPTR, making
751        the feature of a custom pointer more persuasive (as requested by a user).
752    
753    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
754        pcretest.c, inside #ifndefs, because it seems they are sometimes already
755        pre-defined.
756    
757    28. Added support for (*UTF8) at the start of a pattern.
758    
759    29. Arrange for flags added by the "release type" setting in CMake to be shown
760        in the configuration summary.
761    
762    
763    Version 7.8 05-Sep-08
764    ---------------------
765    
766    1.  Replaced UCP searching code with optimized version as implemented for Ad
767        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
768        stage table and inline lookup instead of a function, giving speed ups of 2
769        to 5 times on some simple patterns that I tested. Permission was given to
770        distribute the MultiStage2.py script that generates the tables (it's not in
771        the tarball, but is in the Subversion repository).
772    
773    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
774        scripts.
775    
776    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
777        a group with a zero qualifier. The result of the study could be incorrect,
778        or the function might crash, depending on the pattern.
779    
780    4.  Caseless matching was not working for non-ASCII characters in back
781        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
782        It now works when Unicode Property Support is available.
783    
784    5.  In pcretest, an escape such as \x{de} in the data was always generating
785        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
786        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
787        truncation.
788    
789    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
790    
791    7.  Added two (int) casts to pcregrep when printing the difference of two
792        pointers, in case they are 64-bit values.
793    
794    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
795        test 2 if it fails.
796    
797    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
798        and a #define of that name to empty if it is not externally set. This is to
799        allow users of MSVC to set it if necessary.
800    
801    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
802        the convenience functions in the pcre_get.c source file.
803    
804    11. An option change at the start of a pattern that had top-level alternatives
805        could cause overwriting and/or a crash. This command provoked a crash in
806        some environments:
807    
808          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
809    
810        This potential security problem was recorded as CVE-2008-2371.
811    
812    12. For a pattern where the match had to start at the beginning or immediately
813        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
814        pcre_dfa_exec() could read past the end of the passed subject if there was
815        no match. To help with detecting such bugs (e.g. with valgrind), I modified
816        pcretest so that it places the subject at the end of its malloc-ed buffer.
817    
818    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
819        exec() might read past the end of the data buffer in UTF-8 mode.
820    
821    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
822        the data contained the byte 0x85 as part of a UTF-8 character within its
823        first line. This applied both to normal and DFA matching.
824    
825    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
826        /^[^d]*?$/8 failed to match "abc".
827    
828    16. Added a missing copyright notice to pcrecpp_internal.h.
829    
830    17. Make it more clear in the documentation that values returned from
831        pcre_exec() in ovector are byte offsets, not character counts.
832    
833    18. Tidied a few places to stop certain compilers from issuing warnings.
834    
835    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
836        supplied by Stefan Weber. I made a further small update for 7.8 because
837        there is a change of source arrangements: the pcre_searchfuncs.c module is
838        replaced by pcre_ucd.c.
839    
840    
841    Version 7.7 07-May-08
842    ---------------------
843    
844    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
845        a string to a long long, pretend we don't even have a long long." This is
846        done by checking for the strtoq, strtoll, and _strtoi64 functions.
847    
848    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
849        pre-7.6 versions, which defined a global no_arg variable instead of putting
850        it in the RE class. (See also #8 below.)
851    
852    3.  Remove a line of dead code, identified by coverity and reported by Nuno
853        Lopes.
854    
855    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
856    
857        (1) The include/exclude patterns were being applied to the whole pathnames
858            of files, instead of just to the final components.
859    
860        (2) If there was more than one level of directory, the subdirectories were
861            skipped unless they satisfied the include/exclude conditions. This is
862            inconsistent with GNU grep (and could even be seen as contrary to the
863            pcregrep specification - which I improved to make it absolutely clear).
864            The action now is always to scan all levels of directory, and just
865            apply the include/exclude patterns to regular files.
866    
867    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
868        --exclude_dir in the tests to avoid scanning .svn directories.
869    
870    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
871        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
872        doesn't support NULs in patterns.
873    
874    7.  Added some missing "const"s to declarations of static tables in
875        pcre_compile.c and pcre_dfa_exec.c.
876    
877    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
878        caused by fix #2  above. (Subsequently also a second patch to fix the
879        first patch. And a third patch - this was a messy problem.)
880    
881    9.  Applied Craig's patch to remove the use of push_back().
882    
883    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
884        matching function regexec().
885    
886    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
887        which, however, unlike Perl's \g{...}, are subroutine calls, not back
888        references. PCRE supports relative numbers with this syntax (I don't think
889        Oniguruma does).
890    
891    12. Previously, a group with a zero repeat such as (...){0} was completely
892        omitted from the compiled regex. However, this means that if the group
893        was called as a subroutine from elsewhere in the pattern, things went wrong
894        (an internal error was given). Such groups are now left in the compiled
895        pattern, with a new opcode that causes them to be skipped at execution
896        time.
897    
898    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
899        to the way PCRE behaves:
900    
901        (a) A lone ] character is dis-allowed (Perl treats it as data).
902    
903        (b) A back reference to an unmatched subpattern matches an empty string
904            (Perl fails the current match path).
905    
906        (c) A data ] in a character class must be notated as \] because if the
907            first data character in a class is ], it defines an empty class. (In
908            Perl it is not possible to have an empty class.) The empty class []
909            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
910            The negative empty class [^] matches any one character, independently
911            of the DOTALL setting.
912    
913    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
914        non-existent subpattern following a character class starting with ']' and
915        containing () gave an internal compiling error instead of "reference to
916        non-existent subpattern". Fortunately, when the pattern did exist, the
917        compiled code was correct. (When scanning forwards to check for the
918        existencd of the subpattern, it was treating the data ']' as terminating
919        the class, so got the count wrong. When actually compiling, the reference
920        was subsequently set up correctly.)
921    
922    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
923        it was being rejected as not supported by pcre_dfa_exec(), even though
924        other assertions are supported. I have made pcre_dfa_exec() support
925        (*FAIL).
926    
927    16. The implementation of 13c above involved the invention of a new opcode,
928        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
929        cannot be changed at match time, I realized I could make a small
930        improvement to matching performance by compiling OP_ALLANY instead of
931        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
932        on the OP_ANY path.
933    
934    17. Compiling pcretest on Windows with readline support failed without the
935        following two fixes: (1) Make the unistd.h include conditional on
936        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
937    
938    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
939        ncurses library to be included for pcretest when ReadLine support is
940        requested, but also to allow for it to be overridden. This patch came from
941        Daniel Bergström.
942    
943    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
944        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
945        any errors with the current Unicode tables. Thanks to Peter Kankowski for
946        spotting this.
947    
948    
949    Version 7.6 28-Jan-08
950    ---------------------
951    
952    1.  A character class containing a very large number of characters with
953        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
954        overflow.
955    
956    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
957        HAVE_LONG_LONG is not defined.
958    
959    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
960        bring them up to date and include new features. This patch includes:
961    
962        - Fixed PH's badly added libz and libbz2 support.
963        - Fixed a problem with static linking.
964        - Added pcredemo. [But later removed - see 7 below.]
965        - Fixed dftables problem and added an option.
966        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
967            HAVE_LONG_LONG.
968        - Added readline support for pcretest.
969        - Added an listing of the option settings after cmake has run.
970    
971    4.  A user submitted a patch to Makefile that makes it easy to create
972        "pcre.dll" under mingw when using Configure/Make. I added stuff to
973        Makefile.am that cause it to include this special target, without
974        affecting anything else. Note that the same mingw target plus all
975        the other distribution libraries and programs are now supported
976        when configuring with CMake (see 6 below) instead of with
977        Configure/Make.
978    
979    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
980        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
981        exported in the Windows port". It has not yet been confirmed that the patch
982        solves the problem, but it does no harm.
983    
984    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
985        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
986        with CMake, and also correct the comment about stack recursion.
987    
988    7.  Remove the automatic building of pcredemo from the ./configure system and
989        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
990        of a program that users should build themselves after PCRE is installed, so
991        building it automatically is not really right. What is more, it gave
992        trouble in some build environments.
993    
994    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
995    
996    
997    Version 7.5 10-Jan-08
998    ---------------------
999    
1000    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1001        values in parens when parsing an RE using the C++ wrapper."
1002    
1003    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1004        Characters greater than 255 were excluded from the class instead of being
1005        included.
1006    
1007    3.  The same bug as (2) above applied to negated POSIX classes such as
1008        [:^space:].
1009    
1010    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1011        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1012        I have changed it.
1013    
1014    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1015        first named subpattern) and a construct such as (?&a) would reference the
1016        first named subpattern whose name started with "a" (in other words, the
1017        length check was missing). Both these problems are fixed. "Subpattern name
1018        expected" is now given for (?&) (a zero-length name), and this patch also
1019        makes it give the same error for \k'' (previously it complained that that
1020        was a reference to a non-existent subpattern).
1021    
1022    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1023        this is right because (?- can be followed by option settings as well as by
1024        digits. I have, however, made the messages clearer.
1025    
1026    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1027        than the number used in the conditional) now cause a compile-time error.
1028        This is actually not compatible with Perl, which accepts such patterns, but
1029        treats the conditional as always being FALSE (as PCRE used to), but it
1030        seems to me that giving a diagnostic is better.
1031    
1032    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1033        and messages.
1034    
1035    9.  Fix two occurrences of "backslash" in comments that should have been
1036        "backspace".
1037    
1038    10. Remove two redundant lines of code that can never be obeyed (their function
1039        was moved elsewhere).
1040    
1041    11. The program that makes PCRE's Unicode character property table had a bug
1042        which caused it to generate incorrect table entries for sequences of
1043        characters that have the same character type, but are in different scripts.
1044        It amalgamated them into a single range, with the script of the first of
1045        them. In other words, some characters were in the wrong script. There were
1046        thirteen such cases, affecting characters in the following ranges:
1047    
1048          U+002b0 - U+002c1
1049          U+0060c - U+0060d
1050          U+0061e - U+00612
1051          U+0064b - U+0065e
1052          U+0074d - U+0076d
1053          U+01800 - U+01805
1054          U+01d00 - U+01d77
1055          U+01d9b - U+01dbf
1056          U+0200b - U+0200f
1057          U+030fc - U+030fe
1058          U+03260 - U+0327f
1059          U+0fb46 - U+0fbb1
1060          U+10450 - U+1049d
1061    
1062    12. The -o option (show only the matching part of a line) for pcregrep was not
1063        compatible with GNU grep in that, if there was more than one match in a
1064        line, it showed only the first of them. It now behaves in the same way as
1065        GNU grep.
1066    
1067    13. If the -o and -v options were combined for pcregrep, it printed a blank
1068        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1069        does the same. The return code can be used to tell if there were any
1070        non-matching lines.
1071    
1072    14. Added --file-offsets and --line-offsets to pcregrep.
1073    
1074    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1075        infinitely looping recursion. The bug was that positive lookaheads were not
1076        being skipped when checking for a possible empty match (negative lookaheads
1077        and both kinds of lookbehind were skipped).
1078    
1079    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1080        inclusion of <windows.h> to before rather than after the definition of
1081        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1082    
1083    17. Specifying a possessive quantifier with a specific limit for a Unicode
1084        character property caused pcre_compile() to compile bad code, which led at
1085        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1086        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1087        caused the error; without that there was no problem.
1088    
1089    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1090    
1091    19. Added --enable-pcretest-libreadline.
1092    
1093    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1094        RE::GlobalReplace(). As a result, the number of replacements returned was
1095        double what it should be. I removed one of the increments, but Craig sent a
1096        later patch that removed the other one (the right fix) and added unit tests
1097        that check the return values (which was not done before).
1098    
1099    21. Several CMake things:
1100    
1101        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1102            the names libpcre and libpcreposix, not just pcre and pcreposix.
1103    
1104        (2) The above change means that pcretest and pcregrep are now correctly
1105            linked with the newly-built libraries, not previously installed ones.
1106    
1107        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1108    
1109    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1110        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1111        UTF-8 newline character). The key issue is that the pattern starts .*;
1112        this means that the match must be either at the beginning, or after a
1113        newline. The bug was in the code for advancing after a failed match and
1114        checking that the new position followed a newline. It was not taking
1115        account of UTF-8 characters correctly.
1116    
1117    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1118        character classes. PCRE was not treating the sequence [:...:] as a
1119        character class unless the ... were all letters. Perl, however, seems to
1120        allow any characters between [: and :], though of course it rejects as
1121        unknown any "names" that contain non-letters, because all the known class
1122        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1123        for example, whereas PCRE did not - it did not recognize a POSIX character
1124        class. This seemed a bit dangerous, so the code has been changed to be
1125        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1126        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1127        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1128        Perl does, and where it didn't before.
1129    
1130    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1131        Windows environments %n is disabled by default.
1132    
1133    
1134    Version 7.4 21-Sep-07
1135    ---------------------
1136    
1137    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1138        means that a class such as [\s] counted as "explicit reference to CR or
1139        LF". That isn't really right - the whole point of the change was to try to
1140        help when there was an actual mention of one of the two characters. So now
1141        the change happens only if \r or \n (or a literal CR or LF) character is
1142        encountered.
1143    
1144    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1145        of both had grown to the point where there were only 3 bits left.
1146        Fortunately, there was spare space in the data structure, and so I have
1147        moved the internal flags into a new 16-bit field to free up more option
1148        bits.
1149    
1150    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1151        but did not set the internal JCHANGED flag - either of these is enough to
1152        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1153        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1154        start sets both bits.
1155    
1156    4.  Added options (at build time, compile time, exec time) to change \R from
1157        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1158    
1159    5.  doc/pcresyntax.html was missing from the distribution.
1160    
1161    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1162        compatibility, even though it is no longer used.
1163    
1164    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1165        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1166        windows.h file is present (where different names are used). [This was
1167        reversed later after testing - see 16 below.]
1168    
1169    8.  Changed all #include <config.h> to #include "config.h". There were also
1170        some further <pcre.h> cases that I changed to "pcre.h".
1171    
1172    9.  When pcregrep was used with the --colour option, it missed the line ending
1173        sequence off the lines that it output.
1174    
1175    10. It was pointed out to me that arrays of string pointers cause lots of
1176        relocations when a shared library is dynamically loaded. A technique of
1177        using a single long string with a table of offsets can drastically reduce
1178        these. I have refactored PCRE in four places to do this. The result is
1179        dramatic:
1180    
1181          Originally:                          290
1182          After changing UCP table:            187
1183          After changing error message table:   43
1184          After changing table of "verbs"       36
1185          After changing table of Posix names   22
1186    
1187        Thanks to the folks working on Gregex for glib for this insight.
1188    
1189    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1190        unicode-properties was also set.
1191    
1192    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1193    
1194    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1195        checked only for CRLF.
1196    
1197    14. Added casts to pcretest.c to avoid compiler warnings.
1198    
1199    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1200    
1201    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1202        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1203        entirely. This removes changes made in 7 above.
1204    
1205    17. The CMake files have been updated, and there is now more information about
1206        building with CMake in the NON-UNIX-USE document.
1207    
1208    
1209    Version 7.3 28-Aug-07
1210    ---------------------
1211    
1212     1. In the rejigging of the build system that eventually resulted in 7.1, the
1213        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1214        brackets there is not right, since it causes compilers to look for an
1215        installed pcre.h, not the version that is in the source that is being
1216        compiled (which of course may be different). I have changed it back to:
1217    
1218          #include "pcre.h"
1219    
1220        I have a vague recollection that the change was concerned with compiling in
1221        different directories, but in the new build system, that is taken care of
1222        by the VPATH setting the Makefile.
1223    
1224     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1225        when the subject happened to end in the byte 0x85 (e.g. if the last
1226        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1227        characters but of course it shouldn't be taken as a newline when it is part
1228        of another character. The bug was that, for an unlimited repeat of . in
1229        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1230        characters when looking for a newline.
1231    
1232     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1233    
1234     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1235        in debug output.
1236    
1237     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1238        long printing in the pcrecpp unittest when running under MinGW.
1239    
1240     6. ESC_K was left out of the EBCDIC table.
1241    
1242     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1243        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1244        limit also applies to "virtual nesting" when a pattern is recursive, and in
1245        this case 1000 isn't so big. I have been able to remove this limit at the
1246        expense of backing off one optimization in certain circumstances. Normally,
1247        when pcre_exec() would call its internal match() function recursively and
1248        immediately return the result unconditionally, it uses a "tail recursion"
1249        feature to save stack. However, when a subpattern that can match an empty
1250        string has an unlimited repetition quantifier, it no longer makes this
1251        optimization. That gives it a stack frame in which to save the data for
1252        checking that an empty string has been matched. Previously this was taken
1253        from the 1000-entry workspace that had been reserved. So now there is no
1254        explicit limit, but more stack is used.
1255    
1256     8. Applied Daniel's patches to solve problems with the import/export magic
1257        syntax that is required for Windows, and which was going wrong for the
1258        pcreposix and pcrecpp parts of the library. These were overlooked when this
1259        problem was solved for the main library.
1260    
1261     9. There were some crude static tests to avoid integer overflow when computing
1262        the size of patterns that contain repeated groups with explicit upper
1263        limits. As the maximum quantifier is 65535, the maximum group length was
1264        set at 30,000 so that the product of these two numbers did not overflow a
1265        32-bit integer. However, it turns out that people want to use groups that
1266        are longer than 30,000 bytes (though not repeat them that many times).
1267        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1268        made it possible to implement the integer overflow checks in a much more
1269        dynamic way, which I have now done. The artificial limitation on group
1270        length has been removed - we now have only the limit on the total length of
1271        the compiled pattern, which depends on the LINK_SIZE setting.
1272    
1273    10. Fixed a bug in the documentation for get/copy named substring when
1274        duplicate names are permitted. If none of the named substrings are set, the
1275        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1276        empty string.
1277    
1278    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1279        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1280        because the ] is interpreted as the first data character and the
1281        terminating ] is not found. PCRE has been made compatible with Perl in this
1282        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1283        cause memory overwriting.
1284    
1285    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1286        string has been matched (to stop an infinite loop). It was not recognizing
1287        a conditional subpattern that could match an empty string if that
1288        subpattern was within another subpattern. For example, it looped when
1289        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1290        condition was not nested. This bug has been fixed.
1291    
1292    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1293        past the start of the subject in the presence of bytes with the top bit
1294        set, for example "\x8aBCD".
1295    
1296    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1297        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1298    
1299    14. Optimized (?!) to (*FAIL).
1300    
1301    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1302        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1303        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1304        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1305        does: it's just the validity check that is more restrictive.
1306    
1307    16. Inserted checks for integer overflows during escape sequence (backslash)
1308        processing, and also fixed erroneous offset values for syntax errors during
1309        backslash processing.
1310    
1311    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1312        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1313    
1314    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1315        caused an overrun.
1316    
1317    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1318        something other than just ASCII characters) inside a group that had an
1319        unlimited repeat caused a loop at compile time (while checking to see
1320        whether the group could match an empty string).
1321    
1322    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1323        [\P{Any}] did so. (Error in the code for printing property names.)
1324    
1325    21. An orphan \E inside a character class could cause a crash.
1326    
1327    22. A repeated capturing bracket such as (A)? could cause a wild memory
1328        reference during compilation.
1329    
1330    23. There are several functions in pcre_compile() that scan along a compiled
1331        expression for various reasons (e.g. to see if it's fixed length for look
1332        behind). There were bugs in these functions when a repeated \p or \P was
1333        present in the pattern. These operators have additional parameters compared
1334        with \d, etc, and these were not being taken into account when moving along
1335        the compiled data. Specifically:
1336    
1337        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1338            length.
1339    
1340        (b) An item such as \pL+ within a repeated group could cause crashes or
1341            loops.
1342    
1343        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1344            "reference to non-existent subpattern" error.
1345    
1346        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1347    
1348    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1349        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1350    
1351    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1352    
1353    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1354        character were causing crashes (broken optimization).
1355    
1356    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1357        \p or \P) caused a compile-time loop.
1358    
1359    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1360        break. For example, the unstudied pattern [\r\n]A does not match the string
1361        "\r\nA" because change 7.0/46 below moves the current point on by two
1362        characters after failing to match at the start. However, the pattern \nA
1363        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1364        the same is true. There doesn't seem any very clean way out of this, but
1365        what I have chosen to do makes the common cases work: PCRE now takes note
1366        of whether there can be an explicit match for \r or \n anywhere in the
1367        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1368        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1369        pattern has explicit CR or LF references.
1370    
1371    29. Added (*CR) etc for changing newline setting at start of pattern.
1372    
1373    
1374    Version 7.2 19-Jun-07
1375    ---------------------
1376    
1377     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1378        which is apparently normally available under Windows.
1379    
1380     2. Re-jig the pcregrep tests with different newline settings in an attempt
1381        to make them independent of the local environment's newline setting.
1382    
1383     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1384    
1385     4. Some of the "internals" tests were previously cut out when the link size
1386        was not 2, because the output contained actual offsets. The recent new
1387        "Z" feature of pcretest means that these can be cut out, making the tests
1388        usable with all link sizes.
1389    
1390     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1391        stack recursion. This gives a massive performance boost under BSD, but just
1392        a small improvement under Linux. However, it saves one field in the frame
1393        in all cases.
1394    
1395     6. Added more features from the forthcoming Perl 5.10:
1396    
1397        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1398            recursion call. It refers to the nth most recently opened parentheses.
1399    
1400        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1401            to be opened parentheses.
1402    
1403        (c) Conditions that refer to capturing parentheses can be specified
1404            relatively, for example, (?(-2)... or (?(+3)...
1405    
1406        (d) \K resets the start of the current match so that everything before
1407            is not part of it.
1408    
1409        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1410    
1411        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1412            reference syntax.
1413    
1414        (g) (?| introduces a group in which the numbering of parentheses in each
1415            alternative starts with the same number.
1416    
1417        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1418    
1419     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1420        PCRE_INFO_JCHANGED.
1421    
1422     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1423        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1424        for detecting groups that can match an empty string.
1425    
1426     9. A pattern with a very large number of alternatives (more than several
1427        hundred) was running out of internal workspace during the pre-compile
1428        phase, where pcre_compile() figures out how much memory will be needed. A
1429        bit of new cunning has reduced the workspace needed for groups with
1430        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1431        workspace instead of running out of the 4096 that are available.
1432    
1433    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1434    
1435    11. Applied patch from Google to remove an optimization that didn't quite work.
1436        The report of the bug said:
1437    
1438          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1439          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1440          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1441    
1442    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1443        it matched the wrong number of bytes.
1444    
1445    
1446    Version 7.1 24-Apr-07
1447    ---------------------
1448    
1449     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1450        that is more "standard", making use of automake and other Autotools. There
1451        is some re-arrangement of the files and adjustment of comments consequent
1452        on this.
1453    
1454     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1455        for recursive directory scanning broke on some systems because the files
1456        are not scanned in any specific order and on different systems the order
1457        was different. A call to "sort" has been inserted into RunGrepTest for the
1458        approprate test as a short-term fix. In the longer term there may be an
1459        alternative.
1460    
1461     3. I had an email from Eric Raymond about problems translating some of PCRE's
1462        man pages to HTML (despite the fact that I distribute HTML pages, some
1463        people do their own conversions for various reasons). The problems
1464        concerned the use of low-level troff macros .br and .in. I have therefore
1465        removed all such uses from the man pages (some were redundant, some could
1466        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1467        HTML has been updated to handle .nf/.fi and to complain if it encounters
1468        .br or .in.
1469    
1470     4. Updated comments in configure.ac that get placed in config.h.in and also
1471        arranged for config.h to be included in the distribution, with the name
1472        config.h.generic, for the benefit of those who have to compile without
1473        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1474    
1475     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1476        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1477        makevp.bat for the current PCRE, using the additional files
1478        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1479    
1480     6. A Windows user reported a minor discrepancy with test 2, which turned out
1481        to be caused by a trailing space on an input line that had got lost in his
1482        copy. The trailing space was an accident, so I've just removed it.
1483    
1484     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1485        that is needed.
1486    
1487     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1488        as "const" (a) because they are and (b) because it helps the PHP
1489        maintainers who have recently made a script to detect big data structures
1490        in the php code that should be moved to the .rodata section. I remembered
1491        to update Builducptable as well, so it won't revert if ucptable.h is ever
1492        re-created.
1493    
1494     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1495        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1496        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1497        support is not required. This saves 1.5-2K of code, which is important in
1498        some applications.
1499    
1500        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1501        so as not to refer to the tables, even though these functions will never be
1502        called when UTF-8 support is disabled. Otherwise there are problems with a
1503        shared library.
1504    
1505    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1506    
1507        (a) It was defining its arguments as char * instead of void *.
1508    
1509        (b) It was assuming that all moves were upwards in memory; this was true
1510            a long time ago when I wrote it, but is no longer the case.
1511    
1512        The emulated memove() is provided for those environments that have neither
1513        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1514        is clearly not the case, as these two bugs were recently reported.
1515    
1516    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1517        and Detrail to create the HTML documentation, the .txt form of the man
1518        pages, and it removes trailing spaces from listed files. It also creates
1519        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1520        case, it wraps all the #defines with #ifndefs. This script should be run
1521        before "make dist".
1522    
1523    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1524        with Unicode property support.
1525    
1526        (a) For a maximizing quantifier, if the two different cases of the
1527            character were of different lengths in their UTF-8 codings (there are
1528            some cases like this - I found 11), and the matching function had to
1529            back up over a mixture of the two cases, it incorrectly assumed they
1530            were both the same length.
1531    
1532        (b) When PCRE was configured to use the heap rather than the stack for
1533            recursion during matching, it was not correctly preserving the data for
1534            the other case of a UTF-8 character when checking ahead for a match
1535            while processing a minimizing repeat. If the check also involved
1536            matching a wide character, but failed, corruption could cause an
1537            erroneous result when trying to check for a repeat of the original
1538            character.
1539    
1540    13. Some tidying changes to the testing mechanism:
1541    
1542        (a) The RunTest script now detects the internal link size and whether there
1543            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1544            values substituted by "configure". (The RunGrepTest script already did
1545            this for UTF-8.) The configure.ac script no longer substitutes the
1546            relevant variables.
1547    
1548        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1549            with length and offset values. This means that the output is different
1550            for different internal link sizes. Test 2 is skipped for link sizes
1551            other than 2 because of this, bypassing the problem. Unfortunately,
1552            there was also a test in test 3 (the locale tests) that used /B and
1553            failed for link sizes other than 2. Rather than cut the whole test out,
1554            I have added a new /Z option to pcretest that replaces the length and
1555            offset values with spaces. This is now used to make test 3 independent
1556            of link size. (Test 2 will be tidied up later.)
1557    
1558    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1559        segmentation fault instead of returning the appropriate error message.
1560    
1561    15. In multiline mode when the newline sequence was set to "any", the pattern
1562        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1563        This doesn't seem right; it now treats the CRLF combination as the line
1564        ending, and so does not match in that case. It's only a pattern such as ^$
1565        that would hit this one: something like ^ABC$ would have failed after \r
1566        and then tried again after \r\n.
1567    
1568    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1569        in an attempt to make files that differ only in their line terminators
1570        compare equal. This works on Linux.
1571    
1572    17. Under certain error circumstances pcregrep might try to free random memory
1573        as it exited. This is now fixed, thanks to valgrind.
1574    
1575    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1576        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1577        was because its rules for how to advance for /g after matching an empty
1578        string at the end of a line did not allow for this case. They now check for
1579        it specially.
1580    
1581    20. pcretest is supposed to handle patterns and data of any length, by
1582        extending its buffers when necessary. It was getting this wrong when the
1583        buffer for a data line had to be extended.
1584    
1585    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1586        CRLF as a newline sequence.
1587    
1588    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1589        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1590        I have nevertheless tidied it up.
1591    
1592    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1593    
1594    24. Added a man page for pcre-config.
1595    
1596    
1597    Version 7.0 19-Dec-06
1598    ---------------------
1599    
1600     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1601        moving to gcc 4.1.1.
1602    
1603     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1604        sys/time.h, which is documented as needed for this function. It doesn't
1605        seem to matter on Linux, but it showed up on some releases of OS X.
1606    
1607     3. It seems that there are systems where bytes whose values are greater than
1608        127 match isprint() in the "C" locale. The "C" locale should be the
1609        default when a C program starts up. In most systems, only ASCII printing
1610        characters match isprint(). This difference caused the output from pcretest
1611        to vary, making some of the tests fail. I have changed pcretest so that:
1612    
1613        (a) When it is outputting text in the compiled version of a pattern, bytes
1614            other than 32-126 are always shown as hex escapes.
1615    
1616        (b) When it is outputting text that is a matched part of a subject string,
1617            it does the same, unless a different locale has been set for the match
1618            (using the /L modifier). In this case, it uses isprint() to decide.
1619    
1620     4. Fixed a major bug that caused incorrect computation of the amount of memory
1621        required for a compiled pattern when options that changed within the
1622        pattern affected the logic of the preliminary scan that determines the
1623        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1624        that the computed length was too small. The symptoms of this bug were
1625        either the PCRE error "internal error: code overflow" from pcre_compile(),
1626        or a glibc crash with a message such as "pcretest: free(): invalid next
1627        size (fast)". Examples of patterns that provoked this bug (shown in
1628        pcretest format) are:
1629    
1630          /(?-x: )/x
1631          /(?x)(?-x: \s*#\s*)/
1632          /((?i)[\x{c0}])/8
1633          /(?i:[\x{c0}])/8
1634    
1635        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1636        is now done differently.
1637    
1638     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1639        wrapper classes; (b) implement a new function in the C++ scanner that is
1640        more efficient than the old way of doing things because it avoids levels of
1641        recursion in the regex matching; (c) add a paragraph to the documentation
1642        for the FullMatch() function.
1643    
1644     6. The escape sequence \n was being treated as whatever was defined as
1645        "newline". Not only was this contrary to the documentation, which states
1646        that \n is character 10 (hex 0A), but it also went horribly wrong when
1647        "newline" was defined as CRLF. This has been fixed.
1648    
1649     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1650        was being set to -1 for the "end of line" case (supposedly a value that no
1651        character can have). Though this value is never used (the check for end of
1652        line is "zero bytes in current character"), it caused compiler complaints.
1653        I've changed it to 0xffffffff.
1654    
1655     8. In pcre_version.c, the version string was being built by a sequence of
1656        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1657        string (as it is for production releases) called a macro with an empty
1658        argument. The C standard says the result of this is undefined. The gcc
1659        compiler treats it as an empty string (which was what was wanted) but it is
1660        reported that Visual C gives an error. The source has been hacked around to
1661        avoid this problem.
1662    
1663     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1664        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1665        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1666        of them did).
1667    
1668    10. Originally, pcretest opened its input and output without "b"; then I was
1669        told that "b" was needed in some environments, so it was added for release
1670        5.0 to both the input and output. (It makes no difference on Unix-like
1671        systems.) Later I was told that it is wrong for the input on Windows. I've
1672        now abstracted the modes into two macros, to make it easier to fiddle with
1673        them, and removed "b" from the input mode under Windows.
1674    
1675    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1676    
1677    12. Added -help and --help to pcretest as an official way of being reminded
1678        of the options.
1679    
1680    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1681        and pcrecpp.cc because they annoy compilers at high warning levels.
1682    
1683    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1684    
1685    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1686        scripts are not C programs :-) and which was not noticed because it works
1687        on Linux.
1688    
1689    16. pcretest is supposed to handle any length of pattern and data line (as one
1690        line or as a continued sequence of lines) by extending its input buffer if
1691        necessary. This feature was broken for very long pattern lines, leading to
1692        a string of junk being passed to pcre_compile() if the pattern was longer
1693        than about 50K.
1694    
1695    17. I have done a major re-factoring of the way pcre_compile() computes the
1696        amount of memory needed for a compiled pattern. Previously, there was code
1697        that made a preliminary scan of the pattern in order to do this. That was
1698        OK when PCRE was new, but as the facilities have expanded, it has become
1699        harder and harder to keep it in step with the real compile phase, and there
1700        have been a number of bugs (see for example, 4 above). I have now found a
1701        cunning way of running the real compile function in a "fake" mode that
1702        enables it to compute how much memory it would need, while actually only
1703        ever using a few hundred bytes of working memory and without too many
1704        tests of the mode. This should make future maintenance and development
1705        easier. A side effect of this work is that the limit of 200 on the nesting
1706        depth of parentheses has been removed (though this was never a serious
1707        limitation, I suspect). However, there is a downside: pcre_compile() now
1708        runs more slowly than before (30% or more, depending on the pattern). I
1709        hope this isn't a big issue. There is no effect on runtime performance.
1710    
1711    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1712        newline (only possible for the last line of a file) and it was a
1713        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1714    
1715    19. Added additional timing features to pcretest. (1) The -tm option now times
1716        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1717        separate command line item, by a number that specifies the number of
1718        repeats to use when timing. The default is 50000; this gives better
1719        precision, but takes uncomfortably long for very large patterns.
1720    
1721    20. Extended pcre_study() to be more clever in cases where a branch of a
1722        subpattern has no definite first character. For example, (a*|b*)[cd] would
1723        previously give no result from pcre_study(). Now it recognizes that the
1724        first character must be a, b, c, or d.
1725    
1726    21. There was an incorrect error "recursive call could loop indefinitely" if
1727        a subpattern (or the entire pattern) that was being tested for matching an
1728        empty string contained only one non-empty item after a nested subpattern.
1729        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1730        incorrectly, because the \d was being skipped in the check.
1731    
1732    22. The pcretest program now has a new pattern option /B and a command line
1733        option -b, which is equivalent to adding /B to every pattern. This causes
1734        it to show the compiled bytecode, without the additional information that
1735        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1736        is the same as /B/I).
1737    
1738    23. A new optimization is now able automatically to treat some sequences such
1739        as a*b as a*+b. More specifically, if something simple (such as a character
1740        or a simple class like \d) has an unlimited quantifier, and is followed by
1741        something that cannot possibly match the quantified thing, the quantifier
1742        is automatically "possessified".
1743    
1744    24. A recursive reference to a subpattern whose number was greater than 39
1745        went wrong under certain circumstances in UTF-8 mode. This bug could also
1746        have affected the operation of pcre_study().
1747    
1748    25. Realized that a little bit of performance could be had by replacing
1749        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1750    
1751    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1752    
1753    27. Possessive quantifiers such as a++ were previously implemented by turning
1754        them into atomic groups such as ($>a+). Now they have their own opcodes,
1755        which improves performance. This includes the automatically created ones
1756        from 23 above.
1757    
1758    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1759        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1760        the first matched character to be a colon. This applied both to named and
1761        numbered groups.
1762    
1763    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1764    
1765    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1766        building PCRE on Windows easier, so I have included it in the distribution.
1767    
1768    31. There is now a check in pcretest against a ridiculously large number being
1769        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1770        loop, the loop is abandoned.
1771    
1772    32. Forward references to subpatterns in conditions such as (?(2)...) where
1773        subpattern 2 is defined later cause pcre_compile() to search forwards in
1774        the pattern for the relevant set of parentheses. This search went wrong
1775        when there were unescaped parentheses in a character class, parentheses
1776        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1777    
1778    33. "Subroutine" calls and backreferences were previously restricted to
1779        referencing subpatterns earlier in the regex. This restriction has now
1780        been removed.
1781    
1782    34. Added a number of extra features that are going to be in Perl 5.10. On the
1783        whole, these are just syntactic alternatives for features that PCRE had
1784        previously implemented using the Python syntax or my own invention. The
1785        other formats are all retained for compatibility.
1786    
1787        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1788            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1789            also .NET compatible.
1790    
1791        (b) A recursion or subroutine call to a named group can now be defined as
1792            (?&name) as well as (?P>name).
1793    
1794        (c) A backreference to a named group can now be defined as \k<name> or
1795            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1796            5.10, are also .NET compatible.
1797    
1798        (d) A conditional reference to a named group can now use the syntax
1799            (?(<name>) or (?('name') as well as (?(name).
1800    
1801        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1802            groups (named and numbered) that are never evaluated inline, but can be
1803            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1804            is always false. There may be only one alternative in such a group.
1805    
1806        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1807            as the simple (?(R). The condition is true only if the most recent
1808            recursion is that of the given number or name. It does not search out
1809            through the entire recursion stack.
1810    
1811        (g) The escape \gN or \g{N} has been added, where N is a positive or
1812            negative number, specifying an absolute or relative reference.
1813    
1814    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1815        some "unreachable code" warnings.
1816    
1817    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1818        things, this adds five new scripts.
1819    
1820    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1821        There were also incompatibilities regarding the handling of \Q..\E inside
1822        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1823        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1824    
1825    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1826        matches an empty string, and forcibly breaks the loop. There were bugs in
1827        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1828        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1829        separate and independent bugs (that affected different cases) have been
1830        fixed.
1831    
1832    39. Refactored the code to abolish the use of different opcodes for small
1833        capturing bracket numbers. This is a tidy that I avoided doing when I
1834        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1835        The new approach is not only tidier, it makes it possible to reduce the
1836        memory needed to fix the previous bug (38).
1837    
1838    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1839        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1840        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1841        mode.
1842    
1843    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1844        report.
1845    
1846    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1847        copy construction and assignment in the C++ wrapper.
1848    
1849    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1850        couple of bugs that could have given wrong results in the "--newline=crlf"
1851        case.
1852    
1853    44. Added a number of casts and did some reorganization of signed/unsigned int
1854        variables following suggestions from Dair Grant. Also renamed the variable
1855        "this" as "item" because it is a C++ keyword.
1856    
1857    45. Arranged for dftables to add
1858    
1859          #include "pcre_internal.h"
1860    
1861        to pcre_chartables.c because without it, gcc 4.x may remove the array
1862        definition from the final binary if PCRE is built into a static library and
1863        dead code stripping is activated.
1864    
1865    46. For an unanchored pattern, if a match attempt fails at the start of a
1866        newline sequence, and the newline setting is CRLF or ANY, and the next two
1867        characters are CRLF, advance by two characters instead of one.
1868    
1869    
1870    Version 6.7 04-Jul-06
1871    ---------------------
1872    
1873     1. In order to handle tests when input lines are enormously long, pcretest has
1874        been re-factored so that it automatically extends its buffers when
1875        necessary. The code is crude, but this _is_ just a test program. The
1876        default size has been increased from 32K to 50K.
1877    
1878     2. The code in pcre_study() was using the value of the re argument before
1879        testing it for NULL. (Of course, in any sensible call of the function, it
1880        won't be NULL.)
1881    
1882     3. The memmove() emulation function in pcre_internal.h, which is used on
1883        systems that lack both memmove() and bcopy() - that is, hardly ever -
1884        was missing a "static" storage class specifier.
1885    
1886     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1887        containing an extended class (one that cannot be represented by a bitmap
1888        because it contains high-valued characters or Unicode property items, e.g.
1889        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1890        pattern, but PCRE should not loop if you do not (it no longer does).
1891        [Detail: two cases were found: (a) a repeated subpattern containing an
1892        extended class; (b) a recursive reference to a subpattern that followed a
1893        previous extended class. It wasn't skipping over the extended class
1894        correctly when UTF-8 mode was not set.]
1895    
1896     5. A negated single-character class was not being recognized as fixed-length
1897        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1898        compile error "lookbehind assertion is not fixed length".
1899    
1900     6. The RunPerlTest auxiliary script was showing an unexpected difference
1901        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1902        write a Perl script that can interpret lines of an input file either as
1903        byte characters or as UTF-8, which is what "perltest" was being required to
1904        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1905        can't do is switch easily at run time between having the "use utf8;" pragma
1906        or not. In the end, I fudged it by using the RunPerlTest script to insert
1907        "use utf8;" explicitly for the UTF-8 tests.
1908    
1909     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1910        the end of the subject string, contrary to the documentation and to what
1911        Perl does. This was true of both matching functions. Now it matches only at
1912        the start of the subject and immediately after *internal* newlines.
1913    
1914     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1915        a pointer to an int instead of a pointer to an unsigned long int. This
1916        caused problems on 64-bit systems.
1917    
1918     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1919        instance of the 'standard' template library not being so standard".
1920    
1921    10. There was no check on the number of named subpatterns nor the maximum
1922        length of a subpattern name. The product of these values is used to compute
1923        the size of the memory block for a compiled pattern. By supplying a very
1924        long subpattern name and a large number of named subpatterns, the size
1925        computation could be caused to overflow. This is now prevented by limiting
1926        the length of names to 32 characters, and the number of named subpatterns
1927        to 10,000.
1928    
1929    11. Subpatterns that are repeated with specific counts have to be replicated in
1930        the compiled pattern. The size of memory for this was computed from the
1931        length of the subpattern and the repeat count. The latter is limited to
1932        65535, but there was no limit on the former, meaning that integer overflow
1933        could in principle occur. The compiled length of a repeated subpattern is
1934        now limited to 30,000 bytes in order to prevent this.
1935    
1936    12. Added the optional facility to have named substrings with the same name.
1937    
1938    13. Added the ability to use a named substring as a condition, using the
1939        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1940        are numbers (not recommended). Forward references are permitted.
1941    
1942    14. Added forward references in named backreferences (if you see what I mean).
1943    
1944    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1945        pattern could run off the end of the subject. For example, the pattern
1946        "(?s)(.{1,5})"8 did this with the subject "ab".
1947    
1948    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1949        PCRE_CASELESS was set when matching characters that were quantified with ?
1950        or *.
1951    
1952    17. A character class other than a single negated character that had a minimum
1953        but no maximum quantifier - for example [ab]{6,} - was not handled
1954        correctly by pce_dfa_exec(). It would match only one character.
1955    
1956    18. A valid (though odd) pattern that looked like a POSIX character
1957        class but used an invalid character after [ (for example [[,abc,]]) caused
1958        pcre_compile() to give the error "Failed: internal error: code overflow" or
1959        in some cases to crash with a glibc free() error. This could even happen if
1960        the pattern terminated after [[ but there just happened to be a sequence of
1961        letters, a binary zero, and a closing ] in the memory that followed.
1962    
1963    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1964        over the years. Originally (before any Unicode support), just the bottom 8
1965        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1966        output from "man perlunicode" includes this:
1967    
1968          The regular expression compiler produces polymorphic opcodes.  That
1969          is, the pattern adapts to the data and automatically switches to
1970          the Unicode character scheme when presented with Unicode data--or
1971          instead uses a traditional byte scheme when presented with byte
1972          data.
1973    
1974        Sadly, a wide octal escape does not cause a switch, and in a string with
1975        no other multibyte characters, these octal escapes are treated as before.
1976        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1977        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1978        Unicode string.
1979    
1980        I have not perpetrated such confusion in PCRE. Up till now, it took just
1981        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1982        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1983        translate to the appropriate multibyte character.
1984    
1985    29. Applied some refactoring to reduce the number of warnings from Microsoft
1986        and Borland compilers. This has included removing the fudge introduced
1987        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1988        a warning about an unused variable.
1989    
1990    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1991        characters since release 4.0, because Perl (from release 5.004) does not.
1992        [Or at least, is documented not to: some releases seem to be in conflict
1993        with the documentation.] However, when a pattern was studied with
1994        pcre_study() and all its branches started with \s, PCRE still included VT
1995        as a possible starting character. Of course, this did no harm; it just
1996        caused an unnecessary match attempt.
1997    
1998    22. Removed a now-redundant internal flag bit that recorded the fact that case
1999        dependency changed within the pattern. This was once needed for "required
2000        byte" processing, but is no longer used. This recovers a now-scarce options
2001        bit. Also moved the least significant internal flag bit to the most-
2002        significant bit of the word, which was not previously used (hangover from
2003        the days when it was an int rather than a uint) to free up another bit for
2004        the future.
2005    
2006    23. Added support for CRLF line endings as well as CR and LF. As well as the
2007        default being selectable at build time, it can now be changed at runtime
2008        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2009        specify that it is scanning data with non-default line endings.
2010    
2011    24. Changed the definition of CXXLINK to make it agree with the definition of
2012        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2013    
2014    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2015        recursions. This makes a big different to stack usage for some patterns.
2016    
2017    26. If a subpattern containing a named recursion or subroutine reference such
2018        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2019        the space required for the compiled pattern went wrong and gave too small a
2020        value. Depending on the environment, this could lead to "Failed: internal
2021        error: code overflow at offset 49" or "glibc detected double free or
2022        corruption" errors.
2023    
2024    27. Applied patches from Google (a) to support the new newline modes and (b) to
2025        advance over multibyte UTF-8 characters in GlobalReplace.
2026    
2027    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2028        difference for some implementation of PCRE in some Windows version.
2029    
2030    29. Added some extra testing facilities to pcretest:
2031    
2032        \q<number>   in a data line sets the "match limit" value
2033        \Q<number>   in a data line sets the "match recursion limt" value
2034        -S <number>  sets the stack size, where <number> is in megabytes
2035    
2036        The -S option isn't available for Windows.
2037    
2038    
2039    Version 6.6 06-Feb-06
2040    ---------------------
2041    
2042     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2043        in pcreposix.h. I have copied the definition from pcre.h.
2044    
2045     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2046        because pcre.h is no longer a built file.
2047    
2048     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2049        not normally included in the compiled code.
2050    
2051    
2052    Version 6.5 01-Feb-06
2053    ---------------------
2054    
2055     1. When using the partial match feature with pcre_dfa_exec(), it was not
2056        anchoring the second and subsequent partial matches at the new starting
2057        point. This could lead to incorrect results. For example, with the pattern
2058        /1234/, partially matching against "123" and then "a4" gave a match.
2059    
2060     2. Changes to pcregrep:
2061    
2062        (a) All non-match returns from pcre_exec() were being treated as failures
2063            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2064            error message is output. Some extra information is given for the
2065            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2066            probably the only errors that are likely to be caused by users (by
2067            specifying a regex that has nested indefinite repeats, for instance).
2068            If there are more than 20 of these errors, pcregrep is abandoned.
2069    
2070        (b) A binary zero was treated as data while matching, but terminated the
2071            output line if it was written out. This has been fixed: binary zeroes
2072            are now no different to any other data bytes.
2073    
2074        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2075            used to set a locale for matching. The --locale=xxxx long option has
2076            been added (no short equivalent) to specify a locale explicitly on the
2077            pcregrep command, overriding the environment variables.
2078    
2079        (d) When -B was used with -n, some line numbers in the output were one less
2080            than they should have been.
2081    
2082        (e) Added the -o (--only-matching) option.
2083    
2084        (f) If -A or -C was used with -c (count only), some lines of context were
2085            accidentally printed for the final match.
2086    
2087        (g) Added the -H (--with-filename) option.
2088    
2089        (h) The combination of options -rh failed to suppress file names for files
2090            that were found from directory arguments.
2091    
2092        (i) Added the -D (--devices) and -d (--directories) options.
2093    
2094        (j) Added the -F (--fixed-strings) option.
2095    
2096        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2097    
2098        (l) Added the --colo(u)r option.
2099    
2100        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2101            is not present by default.
2102    
2103     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2104        items such as (?R) or (?1), when the recursion could match a number of
2105        alternatives. If it matched one of the alternatives, but subsequently,
2106        outside the recursion, there was a failure, the code tried to back up into
2107        the recursion. However, because of the way PCRE is implemented, this is not
2108        possible, and the result was an incorrect result from the match.
2109    
2110        In order to prevent this happening, the specification of recursion has
2111        been changed so that all such subpatterns are automatically treated as
2112        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2113    
2114     4. I had overlooked the fact that, in some locales, there are characters for
2115        which isalpha() is true but neither isupper() nor islower() are true. In
2116        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2117        and ordfeminine) are like this. This affected the treatment of \w and \W
2118        when they appeared in character classes, but not when they appeared outside
2119        a character class. The bit map for "word" characters is now created
2120        separately from the results of isalnum() instead of just taking it from the
2121        upper, lower, and digit maps. (Plus the underscore character, of course.)
2122    
2123     5. The above bug also affected the handling of POSIX character classes such as
2124        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2125        permanent tables. Instead, the bit maps for such a class were previously
2126        created as the appropriate unions of the upper, lower, and digit bitmaps.
2127        Now they are created by subtraction from the [[:word:]] class, which has
2128        its own bitmap.
2129    
2130     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2131        It is created by subtracting the vertical space characters (\x09, \x0a,
2132        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2133        subtraction was done in the overall bitmap for a character class, meaning
2134        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2135        be recognized. This bug has been fixed.
2136    
2137     7. Patches from the folks at Google:
2138    
2139          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2140          real life, but is still worth protecting against".
2141    
2142          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2143          regular expressions".
2144    
2145          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2146          have it.
2147    
2148          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2149          "configure" and the latter not, in order to fix a problem somebody had
2150          with compiling the Arg class on HP-UX.
2151    
2152          (e) Improve the error-handling of the C++ wrapper a little bit.
2153    
2154          (f) New tests for checking recursion limiting.
2155    
2156     8. The pcre_memmove() function, which is used only if the environment does not
2157        have a standard memmove() function (and is therefore rarely compiled),
2158        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2159        returning a result (though PCRE never actually uses the result).
2160    
2161     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2162        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2163        returned instead of calling malloc() with an overflowing number that would
2164        most likely cause subsequent chaos.
2165    
2166    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2167    
2168    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2169        with this option is matched, the nmatch and pmatch options of regexec() are
2170        ignored.
2171    
2172    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2173        provided in case anyone wants to the the POSIX interface with UTF-8
2174        strings.
2175    
2176    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2177        C++ linking (needed for some HP-UX environments).
2178    
2179    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2180        (unused parameter) and in the pcre_printint() function (omitted "default"
2181        switch label when the default is to do nothing).
2182    
2183    15. Added some code to make it possible, when PCRE is compiled as a C++
2184        library, to replace subject pointers for pcre_exec() with a smart pointer
2185        class, thus making it possible to process discontinuous strings.
2186    
2187    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2188        much the same function. They were added by different people who were trying
2189        to make PCRE easy to compile on non-Unix systems. It has been suggested
2190        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2191        for compiling on Windows systems. I have therefore replaced it with
2192        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2193        defaults to "extern" for C or "extern C" for C++, which works fine on
2194        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2195        SCOPE with something explicit in config.h. In addition:
2196    
2197        (a) pcreposix.h still had just "extern" instead of either of these macros;
2198            I have replaced it with PCRE_DATA_SCOPE.
2199    
2200        (b) Functions such as _pcre_xclass(), which are internal to the library,
2201            but external in the C sense, all had PCRE_EXPORT in their definitions.
2202            This is apparently wrong for the Windows case, so I have removed it.
2203            (It makes no difference on Unix-like systems.)
2204    
2205    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2206        of recursive calls to match(). This is different to MATCH_LIMIT because
2207        that limits the total number of calls to match(), not all of which increase
2208        the depth of recursion. Limiting the recursion depth limits the amount of
2209        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2210        when PCRE is compiled, and changed at run time. A patch from Google adds
2211        this functionality to the C++ interface.
2212    
2213    18. Changes to the handling of Unicode character properties:
2214    
2215        (a) Updated the table to Unicode 4.1.0.
2216    
2217        (b) Recognize characters that are not in the table as "Cn" (undefined).
2218    
2219        (c) I revised the way the table is implemented to a much improved format
2220            which includes recognition of ranges. It now supports the ranges that
2221            are defined in UnicodeData.txt, and it also amalgamates other
2222            characters into ranges. This has reduced the number of entries in the
2223            table from around 16,000 to around 3,000, thus reducing its size
2224            considerably. I realized I did not need to use a tree structure after
2225            all - a binary chop search is just as efficient. Having reduced the
2226            number of entries, I extended their size from 6 bytes to 8 bytes to
2227            allow for more data.
2228    
2229        (d) Added support for Unicode script names via properties such as \p{Han}.
2230    
2231    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2232        matching that character.
2233    
2234    20. When matching a repeated Unicode property with a minimum greater than zero,
2235        (for example \pL{2,}), PCRE could look past the end of the subject if it
2236        reached it while seeking the minimum number of characters. This could
2237        happen only if some of the characters were more than one byte long, because
2238        there is a check for at least the minimum number of bytes.
2239    
2240    21. Refactored the implementation of \p and \P so as to be more general, to
2241        allow for more different types of property in future. This has changed the
2242        compiled form incompatibly. Anybody with saved compiled patterns that use
2243        \p or \P will have to recompile them.
2244    
2245    22. Added "Any" and "L&" to the supported property types.
2246    
2247    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2248        but give a compile time error if the value is greater than 0xff.
2249    
2250    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2251        accidentally not being installed or uninstalled.
2252    
2253    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2254        made were to insert the current release number. This seemed silly, because
2255        it made things harder for people building PCRE on systems that don't run
2256        "configure". I have turned pcre.h into a distributed file, no longer built
2257        by "configure", with the version identification directly included. There is
2258        no longer a pcre.h.in file.
2259    
2260        However, this change necessitated a change to the pcre-config script as
2261        well. It is built from pcre-config.in, and one of the substitutions was the
2262        release number. I have updated configure.ac so that ./configure now finds
2263        the release number by grepping pcre.h.
2264    
2265    26. Added the ability to run the tests under valgrind.
2266    
2267    
2268    Version 6.4 05-Sep-05
2269    ---------------------
2270    
2271     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2272        "--" to be printed when multiple files were scanned, even when none of the
2273        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2274        consider it to be a bug, and have restored the previous behaviour.
2275    
2276     2. A couple of code tidies to get rid of compiler warnings.
2277    
2278     3. The pcretest program used to cheat by referring to symbols in the library
2279        whose names begin with _pcre_. These are internal symbols that are not
2280        really supposed to be visible externally, and in some environments it is
2281        possible to suppress them. The cheating is now confined to including
2282        certain files from the library's source, which is a bit cleaner.
2283    
2284     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2285        file's purpose clearer.
2286    
2287     5. Reorganized pcre_ucp_findchar().
2288    
2289    
2290    Version 6.3 15-Aug-05
2291    ---------------------
2292    
2293     1. The file libpcre.pc.in did not have general read permission in the tarball.
2294    
2295     2. There were some problems when building without C++ support:
2296    
2297        (a) If C++ support was not built, "make install" and "make test" still
2298            tried to test it.
2299    
2300        (b) There were problems when the value of CXX was explicitly set. Some
2301            changes have been made to try to fix these, and ...
2302    
2303        (c) --disable-cpp can now be used to explicitly disable C++ support.
2304    
2305        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2306            backslash in a target when C++ was disabled. This confuses some
2307            versions of "make", apparently. Using an intermediate variable solves
2308            this. (Same for CPP_LOBJ.)
2309    
2310     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2311        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2312        necessary on certain architectures.
2313    
2314     4. Added a setting of -export-symbols-regex to the link command to remove
2315        those symbols that are exported in the C sense, but actually are local
2316        within the library, and not documented. Their names all begin with
2317        "_pcre_". This is not a perfect job, because (a) we have to except some
2318        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2319        available (and never for static libraries). I have made a note to try to
2320        find a way round (a) in the future.
2321    
2322    
2323    Version 6.2 01-Aug-05
2324    ---------------------
2325    
2326     1. There was no test for integer overflow of quantifier values. A construction
2327        such as {1111111111111111} would give undefined results. What is worse, if
2328        a minimum quantifier for a parenthesized subpattern overflowed and became
2329        negative, the calculation of the memory size went wrong. This could have
2330        led to memory overwriting.
2331    
2332     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2333    
2334     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2335        operating environments where this matters.
2336    
2337     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2338        PCRE options from within the C++ wrapper.
2339    
2340     5. Named capturing subpatterns were not being correctly counted when a pattern
2341        was compiled. This caused two problems: (a) If there were more than 100
2342        such subpatterns, the calculation of the memory needed for the whole
2343        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2344        back references of the form \12, where the number was greater than 9, were
2345        not recognized as back references, even though there were sufficient
2346        previous subpatterns.
2347    
2348     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2349        versions of gcc, e.g. 2.95.4.
2350    
2351    
2352    Version 6.1 21-Jun-05
2353    ---------------------
2354    
2355     1. There was one reference to the variable "posix" in pcretest.c that was not
2356        surrounded by "#if !defined NOPOSIX".
2357    
2358     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2359        the cross-check on the old pcre_info() function, for the benefit of the
2360        cut-down version of PCRE that is currently imported into Exim.
2361    
2362     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2363        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2364        patterns that start (?i) but I don't think that matters. The use of (?i) is
2365        just an example; this all applies to the other options as well.
2366    
2367     4. Since libtool seems to echo the compile commands it is issuing, the output
2368        from "make" can be reduced a bit by putting "@" in front of each libtool
2369        compile command.
2370    
2371     5. Patch from the folks at Google for configure.in to be a bit more thorough
2372        in checking for a suitable C++ installation before trying to compile the
2373        C++ stuff. This should fix a reported problem when a compiler was present,
2374        but no suitable headers.
2375    
2376     6. The man pages all had just "PCRE" as their title. I have changed them to
2377        be the relevant file name. I have also arranged that these names are
2378        retained in the file doc/pcre.txt, which is a concatenation in text format
2379        of all the man pages except the little individual ones for each function.
2380    
2381     7. The NON-UNIX-USE file had not been updated for the different set of source
2382        files that come with release 6. I also added a few comments about the C++
2383        wrapper.
2384    
2385    
2386    Version 6.0 07-Jun-05
2387    ---------------------
2388    
2389     1. Some minor internal re-organization to help with my DFA experiments.
2390    
2391     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2392        didn't matter for the library itself when fully configured, but did matter
2393        when compiling without UCP support, or within Exim, where the ucp files are
2394        not imported.
2395    
2396     3. Refactoring of the library code to split up the various functions into
2397        different source modules. The addition of the new DFA matching code (see
2398        below) to a single monolithic source would have made it really too
2399        unwieldy, quite apart from causing all the code to be include in a
2400        statically linked application, when only some functions are used. This is
2401        relevant even without the DFA addition now that patterns can be compiled in
2402        one application and matched in another.
2403    
2404        The downside of splitting up is that there have to be some external
2405        functions and data tables that are used internally in different modules of
2406        the library but which are not part of the API. These have all had their
2407        names changed to start with "_pcre_" so that they are unlikely to clash
2408        with other external names.
2409    
2410     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2411        a different (DFA) algorithm. Although it is slower than the original
2412        function, it does have some advantages for certain types of matching
2413        problem.
2414    
2415     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2416        including restarting after a partial match.
2417    
2418     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2419        defined when compiling for Windows was sent to me. I have put it into the
2420        code, though I have no means of testing or verifying it.
2421    
2422     7. Added the pcre_refcount() auxiliary function.
2423    
2424     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2425        match before or at the first newline in the subject string. In pcretest,
2426        the /f option on a pattern can be used to set this.
2427    
2428     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2429        would behave wrongly. This has been present in PCRE since release 4.0.
2430    
2431    10. A number of changes to the pcregrep command:
2432    
2433        (a) Refactored how -x works; insert ^(...)$ instead of setting
2434            PCRE_ANCHORED and checking the length, in preparation for adding
2435            something similar for -w.
2436    
2437        (b) Added the -w (match as a word) option.
2438    
2439        (c) Refactored the way lines are read and buffered so as to have more
2440            than one at a time available.
2441    
2442        (d) Implemented a pcregrep test script.
2443    
2444        (e) Added the -M (multiline match) option. This allows patterns to match
2445            over several lines of the subject. The buffering ensures that at least
2446            8K, or the rest of the document (whichever is the shorter) is available
2447            for matching (and similarly the previous 8K for lookbehind assertions).
2448    
2449        (f) Changed the --help output so that it now says
2450    
2451              -w, --word-regex(p)
2452    
2453            instead of two lines, one with "regex" and the other with "regexp"
2454            because that confused at least one person since the short forms are the
2455            same. (This required a bit of code, as the output is generated
2456            automatically from a table. It wasn't just a text change.)
2457    
2458        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2459            option but starts with a hyphen. Could be a pattern or a path name
2460            starting with a hyphen, for instance.
2461    
2462        (h) "-" can be given as a file name to represent stdin.
2463    
2464        (i) When file names are being printed, "(standard input)" is used for
2465            the standard input, for compatibility with GNU grep. Previously
2466            "<stdin>" was used.
2467    
2468        (j) The option --label=xxx can be used to supply a name to be used for
2469            stdin when file names are being printed. There is no short form.
2470    
2471        (k) Re-factored the options decoding logic because we are going to add
2472            two more options that take data. Such options can now be given in four
2473            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2474    
2475        (l) Added the -A, -B, and -C options for requesting that lines of context
2476            around matches be printed.
2477    
2478        (m) Added the -L option to print the names of files that do not contain
2479            any matching lines, that is, the complement of -l.
2480    
2481        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2482            continue to scan other files.
2483    
2484        (o) The -s option was incorrectly implemented. For compatibility with other
2485            greps, it now suppresses the error message for a non-existent or non-
2486            accessible file (but not the return code). There is a new option called
2487            -q that suppresses the output of matching lines, which was what -s was
2488            previously doing.
2489    
2490        (p) Added --include and --exclude options to specify files for inclusion
2491            and exclusion when recursing.
2492    
2493    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2494        Hopefully, it now does.
2495    
2496    12. Missing cast in pcre_study().
2497    
2498    13. Added an "uninstall" target to the makefile.
2499    
2500    14. Replaced "extern" in the function prototypes in Makefile.in with
2501        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2502        world, but is set differently for Windows.
2503    
2504    15. Added a second compiling function called pcre_compile2(). The only
2505        difference is that it has an extra argument, which is a pointer to an
2506        integer error code. When there is a compile-time failure, this is set
2507        non-zero, in addition to the error test pointer being set to point to an
2508        error message. The new argument may be NULL if no error number is required
2509        (but then you may as well call pcre_compile(), which is now just a
2510        wrapper). This facility is provided because some applications need a
2511        numeric error indication, but it has also enabled me to tidy up the way
2512        compile-time errors are handled in the POSIX wrapper.
2513    
2514    16. Added VPATH=.libs to the makefile; this should help when building with one
2515        prefix path and installing with another. (Or so I'm told by someone who
2516        knows more about this stuff than I do.)
2517    
2518    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2519        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2520        match everything, including newlines. This is not POSIX-compatible, but
2521        somebody wanted the feature. From pcretest it can be activated by using
2522        both the P and the s flags.
2523    
2524    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2525    
2526    19. libpcre.pc was being incorrectly installed as executable.
2527    
2528    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2529        it now also looks for '\r' so that it will work unmodified on Windows.
2530    
2531    21. Added Google's contributed C++ wrapper to the distribution.
2532    
2533    22. Added some untidy missing memory free() calls in pcretest, to keep
2534        Electric Fence happy when testing.
2535    
2536    
2537    
2538    Version 5.0 13-Sep-04
2539    ---------------------
2540    
2541     1. Internal change: literal characters are no longer packed up into items
2542        containing multiple characters in a single byte-string. Each character
2543        is now matched using a separate opcode. However, there may be more than one
2544        byte in the character in UTF-8 mode.
2545    
2546     2. The pcre_callout_block structure has two new fields: pattern_position and
2547        next_item_length. These contain the offset in the pattern to the next match
2548        item, and its length, respectively.
2549    
2550     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2551        insertion of callouts before each pattern item. Added the /C option to
2552        pcretest to make use of this.
2553    
2554     4. On the advice of a Windows user, the lines
2555    
2556          #if defined(_WIN32) || defined(WIN32)
2557          _setmode( _fileno( stdout ), 0x8000 );
2558          #endif  /* defined(_WIN32) || defined(WIN32) */
2559    
2560        have been added to the source of pcretest. This apparently does useful
2561        magic in relation to line terminators.
2562    
2563     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2564        for the benefit of those environments where the "b" makes a difference.
2565    
2566     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2567        to know about it. I have put a hack into configure.in that adds in code
2568        to set GCC=yes if CC=icc. This seems to end up at a point in the
2569        generated configure script that is early enough to affect the setting of
2570        compiler options, which is what is needed, but I have no means of testing
2571        whether it really works. (The user who reported this had patched the
2572        generated configure script, which of course I cannot do.)
2573    
2574        LATER: After change 22 below (new libtool files), the configure script
2575        seems to know about icc (and also ecc). Therefore, I have commented out
2576        this hack in configure.in.
2577    
2578     7. Added support for pkg-config (2 patches were sent in).
2579    
2580     8. Negated POSIX character classes that used a combination of internal tables
2581        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2582        [[:^ascii]]. Typically, they would match almost any characters. The other
2583        POSIX classes were not broken in this way.
2584    
2585     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2586        to find the match, as PCRE was deluded into thinking that the match had to
2587        start at the start point or following a newline. The same bug applied to
2588        patterns with negative forward assertions or any backward assertions
2589        preceding ".*" at the start, unless the pattern required a fixed first
2590        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2591    
2592    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2593        starting at the last subject character, bytes beyond the end of the subject
2594        string were read.
2595    
2596    11. Renamed the variable "class" as "classbits" to make life easier for C++
2597        users. (Previously there was a macro definition, but it apparently wasn't
2598        enough.)
2599    
2600    12. Added the new field "tables" to the extra data so that tables can be passed
2601        in at exec time, or the internal tables can be re-selected. This allows
2602        a compiled regex to be saved and re-used at a later time by a different
2603        program that might have everything at different addresses.
2604    
2605    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2606        -R library as well as a -L library.
2607    
2608    14. The debugging options of pcretest (-d on the command line or D on a
2609        pattern) showed incorrect output for anything following an extended class
2610        that contained multibyte characters and which was followed by a quantifier.
2611    
2612    15. Added optional support for general category Unicode character properties
2613        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2614        support. It adds about 90K to the size of the library. The meanings of the
2615        inbuilt class escapes such as \d and \s have NOT been changed.
2616    
2617    16. Updated pcredemo.c to include calls to free() to release the memory for the
2618        compiled pattern.
2619    
2620    17. The generated file chartables.c was being created in the source directory
2621        instead of in the building directory. This caused the build to fail if the
2622        source directory was different from the building directory, and was
2623        read-only.
2624    
2625    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2626        file. No doubt somebody will tell me if they don't make sense... Also added
2627        Dan Mooney's comments about building on OpenVMS.
2628    
2629    19. Added support for partial matching via the PCRE_PARTIAL option for
2630        pcre_exec() and the \P data escape in pcretest.
2631    
2632    20. Extended pcretest with 3 new pattern features:
2633    
2634        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2635              write the compiled pattern to the file whose name is "rest-of-line".
2636              This is a straight binary dump of the data, with the saved pointer to
2637              the character tables forced to be NULL. The study data, if any, is
2638              written too. After writing, pcretest reads a new pattern.
2639    
2640        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2641              compiled pattern from the given file. There must not be any
2642              occurrences of "<" in the file name (pretty unlikely); if there are,
2643              pcretest will instead treat the initial "<" as a pattern delimiter.
2644              After reading in the pattern, pcretest goes on to read data lines as
2645              usual.
2646    
2647        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2648              and 16-bit fields in a compiled pattern, to simulate a pattern that
2649              was compiled on a host of opposite endianness.
2650    
2651    21. The pcre-exec() function can now cope with patterns that were compiled on
2652        hosts of opposite endianness, with this restriction:
2653    
2654          As for any compiled expression that is saved and used later, the tables
2655          pointer field cannot be preserved; the extra_data field in the arguments
2656          to pcre_exec() should be used to pass in a tables address if a value
2657          other than the default internal tables were used at compile time.
2658    
2659    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2660        now diagnosed as an error. Previously, most of the time, a negative number
2661        would have been treated as zero, but if in addition "ovector" was passed as
2662        NULL, a crash could occur.
2663    
2664    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2665        new versions from the libtool 1.5 distribution (the last one is a copy of
2666        a file called libtool.m4). This seems to have fixed the need to patch
2667        "configure" to support Darwin 1.3 (which I used to do). However, I still
2668        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2669        workstation).
2670    
2671    24. Changed the PCRE licence to be the more standard "BSD" licence.
2672    
2673    
2674    Version 4.5 01-Dec-03
2675    ---------------------
2676    
2677     1. There has been some re-arrangement of the code for the match() function so
2678        that it can be compiled in a version that does not call itself recursively.
2679        Instead, it keeps those local variables that need separate instances for
2680        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2681        needs to "recurse". Keeping track of where control must go is done by means
2682        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2683        hide most of the details from the main code, and operates only if
2684        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2685        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2686        operating.
2687    
2688        To make it easier for callers to provide specially tailored get/free
2689        functions for this usage, two new functions, pcre_stack_malloc, and
2690        pcre_stack_free, are used. They are always called in strict stacking order,
2691        and the size of block requested is always the same.
2692    
2693        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2694        PCRE has been compiled to use the stack or the heap for recursion. The
2695        -C option of pcretest uses this to show which version is compiled.
2696    
2697        A new data escape \S, is added to pcretest; it causes the amounts of store
2698        obtained and freed by both kinds of malloc/free at match time to be added
2699        to the output.
2700    
2701     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2702        what's available on my current Linux desktop machine.
2703    
2704     3. When matching a UTF-8 string, the test for a valid string at the start has
2705        been extended. If start_offset is not zero, PCRE now checks that it points
2706        to a byte that is the start of a UTF-8 character. If not, it returns
2707        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2708        this is necessary because there may be backward assertions in the pattern.
2709        When matching the same subject several times, it may save resources to use
2710        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2711    
2712     4. The code for checking the validity of UTF-8 strings has been tightened so
2713        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2714        containing "overlong sequences".
2715    
2716     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2717        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2718        should have been "&", but it just so happened that all the cases this let
2719        through by mistake were picked up later in the function.
2720    
2721     6. I had used a variable called "isblank" - this is a C99 function, causing
2722        some compilers to warn. To avoid this, I renamed it (as "blankclass").
2723    
2724     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2725        prompting; (b) run "./pcretest /dev/null" at the start of the test script
2726        so the version is shown; (c) stop "make test" echoing "./RunTest".
2727    
2728     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2729    
2730     9. The prototype for memmove() for systems that don't have it was using
2731        size_t, but the inclusion of the header that defines size_t was later. I've
2732        moved the #includes for the C headers earlier to avoid this.
2733    
2734    10. Added some adjustments to the code to make it easier to compiler on certain
2735        special systems:
2736    
2737          (a) Some "const" qualifiers were missing.
2738          (b) Added the macro EXPORT before all exported functions; by default this
2739              is defined to be empty.
2740          (c) Changed the dftables auxiliary program (that builds chartables.c) so
2741              that it reads its output file name as an argument instead of writing
2742              to the standard output and assuming this can be redirected.
2743    
2744    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2745        class containing characters with values greater than 255, PCRE compilation
2746        went into a loop.
2747    
2748    12. A recursive reference to a subpattern that was within another subpattern
2749        that had a minimum quantifier of zero caused PCRE to crash. For example,
2750        (x(y(?2))z)? provoked this bug with a subject that got as far as the
2751        recursion. If the recursively-called subpattern itself had a zero repeat,
2752        that was OK.
2753    
2754    13. In pcretest, the buffer for reading a data line was set at 30K, but the
2755        buffer into which it was copied (for escape processing) was still set at
2756        1024, so long lines caused crashes.
2757    
2758    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2759        "internal error: code overflow...". This applied to any character class
2760        that was followed by a possessive quantifier.
2761    
2762    15. Modified the Makefile to add libpcre.la as a prerequisite for
2763        libpcreposix.la because I was told this is needed for a parallel build to
2764        work.
2765    
2766    16. If a pattern that contained .* following optional items at the start was
2767        studied, the wrong optimizing data was generated, leading to matching
2768        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2769        matching string must start with a or b or c. The correct conclusion for
2770        this pattern is that a match can start with any character.
2771    
2772    
2773    Version 4.4 13-Aug-03
2774    ---------------------
2775    
2776     1. In UTF-8 mode, a character class containing characters with values between
2777        127 and 255 was not handled correctly if the compiled pattern was studied.
2778        In fixing this, I have also improved the studying algorithm for such
2779        classes (slightly).
2780    
2781     2. Three internal functions had redundant arguments passed to them. Removal
2782        might give a very teeny performance improvement.
2783    
2784     3. Documentation bug: the value of the capture_top field in a callout is *one
2785        more than* the number of the hightest numbered captured substring.
2786    
2787     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2788        in incorrectly linking with a previously installed version. They now link
2789        explicitly with libpcre.la.
2790    
2791     5. configure.in no longer needs to recognize Cygwin specially.
2792    
2793     6. A problem in pcre.in for Windows platforms is fixed.
2794    
2795     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2796        pcretest, it used to include the size of the study block as part of its
2797        output. Unfortunately, the structure contains a field that has a different
2798        size on different hardware architectures. This meant that the tests that
2799        showed this size failed. As the block is currently always of a fixed size,
2800        this information isn't actually particularly useful in pcretest output, so
2801        I have just removed it.
2802    
2803     8. Three pre-processor statements accidentally did not start in column 1.
2804        Sadly, there are *still* compilers around that complain, even though
2805        standard C has not required this for well over a decade. Sigh.
2806    
2807     9. In pcretest, the code for checking callouts passed small integers in the
2808        callout_data field, which is a void * field. However, some picky compilers
2809        complained about the casts involved for this on 64-bit systems. Now
2810        pcretest passes the address of the small integer instead, which should get
2811        rid of the warnings.
2812    
2813    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2814        both compile and run time, and gives an error if an invalid UTF-8 sequence
2815        is found. There is a option for disabling this check in cases where the
2816        string is known to be correct and/or the maximum performance is wanted.
2817    
2818    11. In response to a bug report, I changed one line in Makefile.in from
2819    
2820            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2821        to
2822            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2823    
2824        to look similar to other lines, but I have no way of telling whether this
2825        is the right thing to do, as I do not use Windows. No doubt I'll get told
2826        if it's wrong...
2827    
2828    
2829    Version 4.3 21-May-03
2830    ---------------------
2831    
2832    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2833       Makefile.
2834    
2835    2. Some refactoring to improve the quality of the code:
2836    
2837       (i)   The utf8_table... variables are now declared "const".
2838    
2839       (ii)  The code for \cx, which used the "case flipping" table to upper case
2840             lower case letters, now just substracts 32. This is ASCII-specific,
2841             but the whole concept of \cx is ASCII-specific, so it seems
2842             reasonable.
2843    
2844       (iii) PCRE was using its character types table to recognize decimal and
2845             hexadecimal digits in the pattern. This is silly, because it handles
2846             only 0-9, a-f, and A-F, but the character types table is locale-
2847             specific, which means strange things might happen. A private
2848             table is now used for this - though it costs 256 bytes, a table is
2849             much faster than multiple explicit tests. Of course, the standard
2850             character types table is still used for matching digits in subject
2851             strings against \d.
2852    
2853       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2854             ending in _t are). So I've renamed it as ESC_tee.
2855    
2856    3. The first argument for regexec() in the POSIX wrapper should have been
2857       defined as "const".
2858    
2859    4. Changed pcretest to use malloc() for its buffers so that they can be
2860       Electric Fenced for debugging.
2861    
2862    5. There were several places in the code where, in UTF-8 mode, PCRE would try
2863       to read one or more bytes before the start of the subject string. Often this
2864       had no effect on PCRE's behaviour, but in some circumstances it could
2865       provoke a segmentation fault.
2866    
2867    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2868       to try to read one or more bytes before the start of the subject string.
2869    
2870    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2871       UTF-8 support could misbehave in various ways if the subject string
2872       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2873       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2874       back over UTF-8 characters.)
2875    
2876    
2877    Version 4.2 14-Apr-03
2878    ---------------------
2879    
2880    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2881    
2882    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2883         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2884         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2885         [WIN_PREFIX]: new variable, "cyg" for Cygwin
2886         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2887           and BUILD_EXEEXT
2888         Note: automatic setting of the BUILD variables is not yet working
2889         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2890           compile-time but not at link-time
2891         [LINK]: use for linking executables only
2892         make different versions for Windows and non-Windows
2893         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2894           libraries
2895         [LINK_FOR_BUILD]: new variable
2896         [OBJEXT]: use throughout
2897         [EXEEXT]: use throughout
2898         <winshared>: new target
2899         <wininstall>: new target
2900         <dftables.o>: use native compiler
2901         <dftables>: use native linker
2902         <install>: handle Windows platform correctly
2903         <clean>: ditto
2904         <check>: ditto
2905         copy DLL to top builddir before testing
2906    
2907       As part of these changes, -no-undefined was removed again. This was reported
2908       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2909       in any case.
2910    
2911    3. Some tidies to get rid of compiler warnings:
2912    
2913       . In the match_data structure, match_limit was an unsigned long int, whereas
2914         match_call_count was an int. I've made them both unsigned long ints.
2915    
2916       . In pcretest the fact that a const uschar * doesn't automatically cast to
2917         a void * provoked a warning.
2918    
2919       . Turning on some more compiler warnings threw up some "shadow" variables
2920         and a few more missing casts.
2921    
2922    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2923       option, a class that contained a single character with a value between 128
2924       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2925    
2926    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2927       option, a class that contained several characters, but with at least one
2928       whose value was between 128 and 255 caused PCRE to crash.
2929    
2930    
2931    Version 4.1 12-Mar-03
2932    ---------------------
2933    
2934    1. Compiling with gcc -pedantic found a couple of places where casts were
2935    needed, and a string in dftables.c that was longer than standard compilers are
2936    required to support.
2937    
2938    2. Compiling with Sun's compiler found a few more places where the code could
2939    be tidied up in order to avoid warnings.
2940    
2941    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2942    first of these names is deprecated in the latest Autoconf in favour of the name
2943    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2944    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2945    analogy I have changed it to CFLAGS_FOR_BUILD.
2946    
2947    4. Added -no-undefined to the linking command in the Makefile, because this is
2948    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2949    linking step for the pcreposix library.
2950    
2951    5. PCRE was failing to diagnose the case of two named groups with the same
2952    name.
2953    
2954    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2955    literal character that is needed in the subject for a match, and scans along to
2956    ensure that it is present before embarking on the full matching process. This
2957    saves time in cases of nested unlimited repeats that are never going to match.
2958    Problem: the scan can take a lot of time if the subject is very long (e.g.
2959    megabytes), thus penalizing straightforward matches. It is now done only if the
2960    amount of subject to be scanned is less than 1000 bytes.
2961    
2962    7. A lesser problem with the same optimization is that it was recording the
2963    first character of an anchored pattern as "needed", thus provoking a search
2964    right along the subject, even when the first match of the pattern was going to
2965    fail. The "needed" character is now not set for anchored patterns, unless it
2966    follows something in the pattern that is of non-fixed length. Thus, it still
2967    fulfils its original purpose of finding quick non-matches in cases of nested
2968    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2969    
2970    
2971    Version 4.0 17-Feb-03
2972    ---------------------
2973    
2974  1. If a comment in an extended regex that started immediately after a meta-item  1. If a comment in an extended regex that started immediately after a meta-item
2975  extended to the end of string, PCRE compiled incorrect data. This could lead to  extended to the end of string, PCRE compiled incorrect data. This could lead to
2976  all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.  all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.

Legend:
Removed from v.63  
changed lines
  Added in v.595

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12