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

Diff of /code/trunk/ChangeLog

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

revision 158 by ph10, Tue May 1 10:30:14 2007 UTC revision 571 by ph10, Tue Nov 16 17:51:37 2010 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Version 8.11 10-Oct-2010
5    ------------------------
6    
7    1.  (*THEN) was not working properly if there were untried alternatives prior
8        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
9        backtracked to try for "b" instead of moving to the next alternative branch
10        at the same level (in this case, to look for "c"). The Perl documentation
11        is clear that when (*THEN) is backtracked onto, it goes to the "next
12        alternative in the innermost enclosing group".
13    
14    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
15        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
16        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
17        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
18        (*THEN).
19    
20    3.  If \s appeared in a character class, it removed the VT character from
21        the class, even if it had been included by some previous item, for example
22        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
23        of \s, but is part of the POSIX "space" class.)
24    
25    4.  A partial match never returns an empty string (because you can always
26        match an empty string at the end of the subject); however the checking for
27        an empty string was starting at the "start of match" point. This has been
28        changed to the "earliest inspected character" point, because the returned
29        data for a partial match starts at this character. This means that, for
30        example, /(?<=abc)def/ gives a partial match for the subject "abc"
31        (previously it gave "no match").
32    
33    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
34        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
35        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
36        has an implication that the given string is incomplete (because a partial
37        match is preferred over a full match). For this reason, these items now
38        give a partial match in this situation. [Aside: previously, the one case
39        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
40        match rather than a full match, which was wrong by the old rules, but is
41        now correct.]
42    
43    6.  There was a bug in the handling of #-introduced comments, recognized when
44        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
45        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
46        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
47        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
48        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
49        places in pcre_compile().
50    
51    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
52        comments when looking ahead for named forward references to subpatterns,
53        the only newline sequence it recognized was NL. It now handles newlines
54        according to the set newline convention.
55    
56    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
57        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
58        cater for a lack of strerror(). These oversights have been fixed.
59    
60    9.  Added --match-limit and --recursion-limit to pcregrep.
61    
62    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
63    
64    11. When the -o option was used, pcregrep was setting a return code of 1, even
65        when matches were found, and --line-buffered was not being honoured.
66    
67    12. Added an optional parentheses number to the -o and --only-matching options
68        of pcregrep.
69    
70    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
71        can match an empty string. The code to do it in pcretest and pcredemo
72        needed fixing:
73    
74        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
75            only one byte after an empty string match just before CRLF (this case
76            just got forgotten; "any" and "anycrlf" were OK).
77    
78        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
79            mode when an empty string match preceded an ASCII character followed by
80            a non-ASCII character. (The code for advancing by one character rather
81            than one byte was nonsense.)
82    
83        (c) The pcredemo.c sample program did not have any code at all to handle
84            the cases when CRLF is a valid newline sequence.
85    
86    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
87        as a starting offset was within the subject string. There is now a new
88        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
89        negative or greater than the length of the string. In order to test this,
90        pcretest is extended to allow the setting of negative starting offsets.
91    
92    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
93        starting offset points to the beginning of a UTF-8 character was
94        unnecessarily clumsy. I tidied it up.
95    
96    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
97        bad UTF-8 sequence and one that is incomplete.
98    
99    17. Nobody had reported that the --include_dir option, which was added in
100        release 7.7 should have been called --include-dir (hyphen, not underscore)
101        for compatibility with GNU grep. I have changed it to --include-dir, but
102        left --include_dir as an undocumented synonym, and the same for
103        --exclude-dir, though that is not available in GNU grep, at least as of
104        release 2.5.4.
105    
106    
107    Version 8.10 25-Jun-2010
108    ------------------------
109    
110    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
111        THEN.
112    
113    2.  (*ACCEPT) was not working when inside an atomic group.
114    
115    3.  Inside a character class, \B is treated as a literal by default, but
116        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
117        causes the error). The code is unchanged, but I tidied the documentation.
118    
119    4.  Inside a character class, PCRE always treated \R and \X as literals,
120        whereas Perl faults them if its -w option is set. I have changed PCRE so
121        that it faults them when PCRE_EXTRA is set.
122    
123    5.  Added support for \N, which always matches any character other than
124        newline. (It is the same as "." when PCRE_DOTALL is not set.)
125    
126    6.  When compiling pcregrep with newer versions of gcc which may have
127        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
128        declared with attribute warn_unused_result" were given. Just casting the
129        result to (void) does not stop the warnings; a more elaborate fudge is
130        needed. I've used a macro to implement this.
131    
132    7.  Minor change to pcretest.c to avoid a compiler warning.
133    
134    8.  Added four artifical Unicode properties to help with an option to make
135        \s etc use properties (see next item). The new properties are: Xan
136        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
137    
138    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
139        use Unicode properties. (*UCP) at the start of a pattern can be used to set
140        this option. Modified pcretest to add /W to test this facility. Added
141        REG_UCP to make it available via the POSIX interface.
142    
143    10. Added --line-buffered to pcregrep.
144    
145    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
146        studied, and the match started with a letter with a code point greater than
147        127 whose first byte was different to the first byte of the other case of
148        the letter, the other case of this starting letter was not recognized
149        (#976).
150    
151    12. If a pattern that was studied started with a repeated Unicode property
152        test, for example, \p{Nd}+, there was the theoretical possibility of
153        setting up an incorrect bitmap of starting bytes, but fortunately it could
154        not have actually happened in practice until change 8 above was made (it
155        added property types that matched character-matching opcodes).
156    
157    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
158        possible starting bytes for non-anchored patterns.
159    
160    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
161        \R, and also a number of cases that involve Unicode properties, both
162        explicit and implicit when PCRE_UCP is set.
163    
164    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
165        input, it could crash or give wrong results if characters with values
166        greater than 0xc0 were present in the subject string. (Detail: it assumed
167        UTF-8 input when processing these items.)
168    
169    16. Added a lot of (int) casts to avoid compiler warnings in systems where
170        size_t is 64-bit (#991).
171    
172    17. Added a check for running out of memory when PCRE is compiled with
173        --disable-stack-for-recursion (#990).
174    
175    18. If the last data line in a file for pcretest does not have a newline on
176        the end, a newline was missing in the output.
177    
178    19. The default pcre_chartables.c file recognizes only ASCII characters (values
179        less than 128) in its various bitmaps. However, there is a facility for
180        generating tables according to the current locale when PCRE is compiled. It
181        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
182        characters, are recognized by isspace() and therefore were getting set in
183        these tables, and indeed these tables seem to approximate to ISO 8859. This
184        caused a problem in UTF-8 mode when pcre_study() was used to create a list
185        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
186        which of course cannot start UTF-8 characters. I have changed the code so
187        that only real ASCII characters (less than 128) and the correct starting
188        bytes for UTF-8 encodings are set for characters greater than 127 when in
189        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
190        altogether.)
191    
192    20. Added the /T option to pcretest so as to be able to run tests with non-
193        standard character tables, thus making it possible to include the tests
194        used for 19 above in the standard set of tests.
195    
196    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
197        reference to a subpattern the other side of a comment that contains an
198        opening parenthesis caused either an internal compiling error, or a
199        reference to the wrong subpattern.
200    
201    
202    Version 8.02 19-Mar-2010
203    ------------------------
204    
205    1.  The Unicode data tables have been updated to Unicode 5.2.0.
206    
207    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
208        configured.
209    
210    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
211        original author of that file, following a query about its status.
212    
213    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
214        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
215    
216    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
217        quantifier applied to a forward-referencing subroutine call, could compile
218        incorrect code or give the error "internal error: previously-checked
219        referenced subpattern not found".
220    
221    6.  Both MS Visual Studio and Symbian OS have problems with initializing
222        variables to point to external functions. For these systems, therefore,
223        pcre_malloc etc. are now initialized to local functions that call the
224        relevant global functions.
225    
226    7.  There were two entries missing in the vectors called coptable and poptable
227        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
228        I've fixed the data, and added a kludgy way of testing at compile time that
229        the lengths are correct (equal to the number of opcodes).
230    
231    8.  Following on from 7, I added a similar kludge to check the length of the
232        eint vector in pcreposix.c.
233    
234    9.  Error texts for pcre_compile() are held as one long string to avoid too
235        much relocation at load time. To find a text, the string is searched,
236        counting zeros. There was no check for running off the end of the string,
237        which could happen if a new error number was added without updating the
238        string.
239    
240    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
241    
242    11. \K was not working if it appeared in an atomic group or in a group that
243        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
244        \K is "not well defined" if used in an assertion. PCRE now accepts it if
245        the assertion is positive, but not if it is negative.
246    
247    12. Change 11 fortuitously reduced the size of the stack frame used in the
248        "match()" function of pcre_exec.c by one pointer. Forthcoming
249        implementation of support for (*MARK) will need an extra pointer on the
250        stack; I have reserved it now, so that the stack frame size does not
251        decrease.
252    
253    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
254        item in branch that calls a recursion is a subroutine call - as in the
255        second branch in the above example - was incorrectly given the compile-
256        time error "recursive call could loop indefinitely" because pcre_compile()
257        was not correctly checking the subroutine for matching a non-empty string.
258    
259    14. The checks for overrunning compiling workspace could trigger after an
260        overrun had occurred. This is a "should never occur" error, but it can be
261        triggered by pathological patterns such as hundreds of nested parentheses.
262        The checks now trigger 100 bytes before the end of the workspace.
263    
264    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
265    
266    
267    Version 8.01 19-Jan-2010
268    ------------------------
269    
270    1.  If a pattern contained a conditional subpattern with only one branch (in
271        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
272        computed the wrong minimum data length (which is of course zero for such
273        subpatterns). This could cause incorrect "no match" results.
274    
275    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
276        the pattern is reset in the first branch, pcre_compile() failed with
277        "internal error: code overflow at offset...". This happened only when
278        the reset was to the original external option setting. (An optimization
279        abstracts leading options settings into an external setting, which was the
280        cause of this.)
281    
282    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
283        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
284        assertion pattern did not match (meaning that the assertion was true), it
285        was incorrectly treated as false if the SKIP had been reached during the
286        matching. This also applied to assertions used as conditions.
287    
288    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
289        assertion subpattern, including such a pattern used as a condition,
290        unpredictable results occurred, instead of the error return
291        PCRE_ERROR_DFA_UITEM.
292    
293    5.  The C++ GlobalReplace function was not working like Perl for the special
294        situation when an empty string is matched. It now does the fancy magic
295        stuff that is necessary.
296    
297    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
298        removed. (These were left over from very, very early versions of PCRE.)
299    
300    7.  Some cosmetic changes to the code to make life easier when compiling it
301        as part of something else:
302    
303        (a) Change DEBUG to PCRE_DEBUG.
304    
305        (b) In pcre_compile(), rename the member of the "branch_chain" structure
306            called "current" as "current_branch", to prevent a collision with the
307            Linux macro when compiled as a kernel module.
308    
309        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
310            prevent a collision with the Linux macro when compiled as a kernel
311            module.
312    
313    8.  In pcre_compile() there are some checks for integer overflows that used to
314        cast potentially large values to (double). This has been changed to that
315        when building, a check for int64_t is made, and if it is found, it is used
316        instead, thus avoiding the use of floating point arithmetic. (There is no
317        other use of FP in PCRE.) If int64_t is not found, the fallback is to
318        double.
319    
320    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
321        2005 (difference between two addresses compared to an unsigned value).
322    
323    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
324        custom one, because of the following reported problem in Windows:
325    
326          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
327              under Win32.
328          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
329              therefore missing the function definition.
330          - The compiler thus generates a "C" signature for the test function.
331          - The linker fails to find the "C" function.
332          - PCRE fails to configure if asked to do so against libbz2.
333    
334    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
335        messages were output:
336    
337          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
338          rerunning libtoolize, to keep the correct libtool macros in-tree.
339          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
340    
341        I have done both of these things.
342    
343    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
344        most of the time, it *can* run out if it is given a pattern that contains a
345        runaway infinite recursion. I updated the discussion in the pcrestack man
346        page.
347    
348    13. Now that we have gone to the x.xx style of version numbers, the minor
349        version may start with zero. Using 08 or 09 is a bad idea because users
350        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
351        interpreted as invalid octal numbers. I've updated the previous comment in
352        configure.ac, and also added a check that gives an error if 08 or 09 are
353        used.
354    
355    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
356        causing partial matching to fail when the end of the subject matched \W
357        in a UTF-8 pattern where \W was quantified with a minimum of 3.
358    
359    15. There were some discrepancies between the declarations in pcre_internal.h
360        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
361        their definitions. The declarations used "const uschar *" and the
362        definitions used USPTR. Even though USPTR is normally defined as "const
363        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
364        reported that: "This difference in casting confuses some C++ compilers, for
365        example, SunCC recognizes above declarations as different functions and
366        generates broken code for hbpcre." I have changed the declarations to use
367        USPTR.
368    
369    16. GNU libtool is named differently on some systems. The autogen.sh script now
370        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
371        (FreeBSD).
372    
373    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
374        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
375        comment: "Figure out how to create a longlong from a string: strtoll and
376        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
377        instance, but it only takes 2 args instead of 3!"
378    
379    18. A subtle bug concerned with back references has been fixed by a change of
380        specification, with a corresponding code fix. A pattern such as
381        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
382        refers, was giving matches when it shouldn't. For example, xa=xaaa would
383        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
384        same bug. Such groups have to be quantified to be useful, or contained
385        inside another quantified group. (If there's no repetition, the reference
386        can never match.) The problem arises because, having left the group and
387        moved on to the rest of the pattern, a later failure that backtracks into
388        the group uses the captured value from the final iteration of the group
389        rather than the correct earlier one. I have fixed this in PCRE by forcing
390        any group that contains a reference to itself to be an atomic group; that
391        is, there cannot be any backtracking into it once it has completed. This is
392        similar to recursive and subroutine calls.
393    
394    
395    Version 8.00 19-Oct-09
396    ----------------------
397    
398    1.  The table for translating pcre_compile() error codes into POSIX error codes
399        was out-of-date, and there was no check on the pcre_compile() error code
400        being within the table. This could lead to an OK return being given in
401        error.
402    
403    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
404        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
405        in a Windows environment.
406    
407    3.  The pcregrep --count option prints the count for each file even when it is
408        zero, as does GNU grep. However, pcregrep was also printing all files when
409        --files-with-matches was added. Now, when both options are given, it prints
410        counts only for those files that have at least one match. (GNU grep just
411        prints the file name in this circumstance, but including the count seems
412        more useful - otherwise, why use --count?) Also ensured that the
413        combination -clh just lists non-zero counts, with no names.
414    
415    4.  The long form of the pcregrep -F option was incorrectly implemented as
416        --fixed_strings instead of --fixed-strings. This is an incompatible change,
417        but it seems right to fix it, and I didn't think it was worth preserving
418        the old behaviour.
419    
420    5.  The command line items --regex=pattern and --regexp=pattern were not
421        recognized by pcregrep, which required --regex pattern or --regexp pattern
422        (with a space rather than an '='). The man page documented the '=' forms,
423        which are compatible with GNU grep; these now work.
424    
425    6.  No libpcreposix.pc file was created for pkg-config; there was just
426        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
427    
428    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
429        when UCP support is not needed, by modifying the Python script that
430        generates it from Unicode data files. This should not matter if the module
431        is correctly used as a library, but I received one complaint about 50K of
432        unwanted data. My guess is that the person linked everything into his
433        program rather than using a library. Anyway, it does no harm.
434    
435    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
436        was a minimum greater than 1 for a wide character in a possessive
437        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
438        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
439        character. Chaos in the form of incorrect output or a compiling loop could
440        result.
441    
442    9.  The restrictions on what a pattern can contain when partial matching is
443        requested for pcre_exec() have been removed. All patterns can now be
444        partially matched by this function. In addition, if there are at least two
445        slots in the offset vector, the offset of the earliest inspected character
446        for the match and the offset of the end of the subject are set in them when
447        PCRE_ERROR_PARTIAL is returned.
448    
449    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
450        synonymous with PCRE_PARTIAL, for backwards compatibility, and
451        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
452        and may be more useful for multi-segment matching.
453    
454    11. Partial matching with pcre_exec() is now more intuitive. A partial match
455        used to be given if ever the end of the subject was reached; now it is
456        given only if matching could not proceed because another character was
457        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
458        string "Z", which now yields "no match" instead of "partial match". In the
459        case of pcre_dfa_exec(), "no match" is given if every matching path for the
460        final character ended with (*FAIL).
461    
462    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
463        if the pattern had a "must contain" character that was already found in the
464        earlier partial match, unless partial matching was again requested. For
465        example, with the pattern /dog.(body)?/, the "must contain" character is
466        "g". If the first part-match was for the string "dog", restarting with
467        "sbody" failed. This bug has been fixed.
468    
469    13. The string returned by pcre_dfa_exec() after a partial match has been
470        changed so that it starts at the first inspected character rather than the
471        first character of the match. This makes a difference only if the pattern
472        starts with a lookbehind assertion or \b or \B (\K is not supported by
473        pcre_dfa_exec()). It's an incompatible change, but it makes the two
474        matching functions compatible, and I think it's the right thing to do.
475    
476    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
477        so that the demonstration program is easily available in environments where
478        PCRE has not been installed from source.
479    
480    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
481        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
482        library.
483    
484    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
485        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
486        is not the first non-POSIX option to be added. Clearly some people find
487        these options useful.
488    
489    17. If a caller to the POSIX matching function regexec() passes a non-zero
490        value for nmatch with a NULL value for pmatch, the value of
491        nmatch is forced to zero.
492    
493    18. RunGrepTest did not have a test for the availability of the -u option of
494        the diff command, as RunTest does. It now checks in the same way as
495        RunTest, and also checks for the -b option.
496    
497    19. If an odd number of negated classes containing just a single character
498        interposed, within parentheses, between a forward reference to a named
499        subpattern and the definition of the subpattern, compilation crashed with
500        an internal error, complaining that it could not find the referenced
501        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
502        [The bug was that it was starting one character too far in when skipping
503        over the character class, thus treating the ] as data rather than
504        terminating the class. This meant it could skip too much.]
505    
506    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
507        /g option in pcretest when the pattern contains \K, which makes it possible
508        to have an empty string match not at the start, even when the pattern is
509        anchored. Updated pcretest and pcredemo to use this option.
510    
511    21. If the maximum number of capturing subpatterns in a recursion was greater
512        than the maximum at the outer level, the higher number was returned, but
513        with unset values at the outer level. The correct (outer level) value is
514        now given.
515    
516    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
517        PCRE did not set those parentheses (unlike Perl). I have now found a way to
518        make it do so. The string so far is captured, making this feature
519        compatible with Perl.
520    
521    23. The tests have been re-organized, adding tests 11 and 12, to make it
522        possible to check the Perl 5.10 features against Perl 5.10.
523    
524    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
525        pattern matches a fixed length string. PCRE did not allow this; now it
526        does. Neither allows recursion.
527    
528    25. I finally figured out how to implement a request to provide the minimum
529        length of subject string that was needed in order to match a given pattern.
530        (It was back references and recursion that I had previously got hung up
531        on.) This code has now been added to pcre_study(); it finds a lower bound
532        to the length of subject needed. It is not necessarily the greatest lower
533        bound, but using it to avoid searching strings that are too short does give
534        some useful speed-ups. The value is available to calling programs via
535        pcre_fullinfo().
536    
537    26. While implementing 25, I discovered to my embarrassment that pcretest had
538        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
539        study optimizations had never been tested with that matching function.
540        Oops. What is worse, even when it was passed study data, there was a bug in
541        pcre_dfa_exec() that meant it never actually used it. Double oops. There
542        were also very few tests of studied patterns with pcre_dfa_exec().
543    
544    27. If (?| is used to create subpatterns with duplicate numbers, they are now
545        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
546        on the other side of the coin, they are no longer allowed to have different
547        names, because these cannot be distinguished in PCRE, and this has caused
548        confusion. (This is a difference from Perl.)
549    
550    28. When duplicate subpattern names are present (necessarily with different
551        numbers, as required by 27 above), and a test is made by name in a
552        conditional pattern, either for a subpattern having been matched, or for
553        recursion in such a pattern, all the associated numbered subpatterns are
554        tested, and the overall condition is true if the condition is true for any
555        one of them. This is the way Perl works, and is also more like the way
556        testing by number works.
557    
558    
559    Version 7.9 11-Apr-09
560    ---------------------
561    
562    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
563        (pcretest), all targets were linked against these libraries. This included
564        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
565        libraries. This caused unwanted dependencies to be created. This problem
566        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
567        pcretest is linked with readline.
568    
569    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
570        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
571        moved outside it again, because FALSE and TRUE are already defined in AIX,
572        but BOOL is not.
573    
574    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
575        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
576    
577    4.  The pcregrep documentation said spaces were inserted as well as colons (or
578        hyphens) following file names and line numbers when outputting matching
579        lines. This is not true; no spaces are inserted. I have also clarified the
580        wording for the --colour (or --color) option.
581    
582    5.  In pcregrep, when --colour was used with -o, the list of matching strings
583        was not coloured; this is different to GNU grep, so I have changed it to be
584        the same.
585    
586    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
587        each matching line was coloured. Now it goes on to look for further matches
588        of any of the test patterns, which is the same behaviour as GNU grep.
589    
590    7.  A pattern that could match an empty string could cause pcregrep to loop; it
591        doesn't make sense to accept an empty string match in pcregrep, so I have
592        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
593        seems to be how GNU grep behaves.
594    
595    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
596        start or after a newline", because the conditional assertion was not being
597        correctly handled. The rule now is that both the assertion and what follows
598        in the first alternative must satisfy the test.
599    
600    9.  If auto-callout was enabled in a pattern with a conditional group whose
601        condition was an assertion, PCRE could crash during matching, both with
602        pcre_exec() and pcre_dfa_exec().
603    
604    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
605        used for matching.
606    
607    11. Unicode property support in character classes was not working for
608        characters (bytes) greater than 127 when not in UTF-8 mode.
609    
610    12. Added the -M command line option to pcretest.
611    
612    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
613    
614    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
615    
616    16. Added comments and documentation about mis-use of no_arg in the C++
617        wrapper.
618    
619    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
620        from Martin Jerabek that uses macro names for all relevant character and
621        string constants.
622    
623    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
624        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
625        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
626        these, but not everybody uses configure.
627    
628    19. A conditional group that had only one branch was not being correctly
629        recognized as an item that could match an empty string. This meant that an
630        enclosing group might also not be so recognized, causing infinite looping
631        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
632        with the subject "ab", where knowledge that the repeated group can match
633        nothing is needed in order to break the loop.
634    
635    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
636        exec(), but without supplying a callout function, matching went wrong.
637    
638    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
639        leak if the size of the offset vector was greater than 30. When the vector
640        is smaller, the saved offsets during recursion go onto a local stack
641        vector, but for larger vectors malloc() is used. It was failing to free
642        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
643        error, in fact).
644    
645    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
646        heapframe that is used only when UTF-8 support is enabled. This caused no
647        problem, but was untidy.
648    
649    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
650        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
651        included within another project.
652    
653    24. Steven Van Ingelgem's patches to add more options to the CMake support,
654        slightly modified by me:
655    
656          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
657              not building pcregrep.
658    
659          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
660              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
661    
662    25. Forward references, both numeric and by name, in patterns that made use of
663        duplicate group numbers, could behave incorrectly or give incorrect errors,
664        because when scanning forward to find the reference group, PCRE was not
665        taking into account the duplicate group numbers. A pattern such as
666        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
667    
668    26. Changed a few more instances of "const unsigned char *" to USPTR, making
669        the feature of a custom pointer more persuasive (as requested by a user).
670    
671    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
672        pcretest.c, inside #ifndefs, because it seems they are sometimes already
673        pre-defined.
674    
675    28. Added support for (*UTF8) at the start of a pattern.
676    
677    29. Arrange for flags added by the "release type" setting in CMake to be shown
678        in the configuration summary.
679    
680    
681    Version 7.8 05-Sep-08
682    ---------------------
683    
684    1.  Replaced UCP searching code with optimized version as implemented for Ad
685        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
686        stage table and inline lookup instead of a function, giving speed ups of 2
687        to 5 times on some simple patterns that I tested. Permission was given to
688        distribute the MultiStage2.py script that generates the tables (it's not in
689        the tarball, but is in the Subversion repository).
690    
691    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
692        scripts.
693    
694    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
695        a group with a zero qualifier. The result of the study could be incorrect,
696        or the function might crash, depending on the pattern.
697    
698    4.  Caseless matching was not working for non-ASCII characters in back
699        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
700        It now works when Unicode Property Support is available.
701    
702    5.  In pcretest, an escape such as \x{de} in the data was always generating
703        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
704        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
705        truncation.
706    
707    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
708    
709    7.  Added two (int) casts to pcregrep when printing the difference of two
710        pointers, in case they are 64-bit values.
711    
712    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
713        test 2 if it fails.
714    
715    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
716        and a #define of that name to empty if it is not externally set. This is to
717        allow users of MSVC to set it if necessary.
718    
719    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
720        the convenience functions in the pcre_get.c source file.
721    
722    11. An option change at the start of a pattern that had top-level alternatives
723        could cause overwriting and/or a crash. This command provoked a crash in
724        some environments:
725    
726          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
727    
728        This potential security problem was recorded as CVE-2008-2371.
729    
730    12. For a pattern where the match had to start at the beginning or immediately
731        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
732        pcre_dfa_exec() could read past the end of the passed subject if there was
733        no match. To help with detecting such bugs (e.g. with valgrind), I modified
734        pcretest so that it places the subject at the end of its malloc-ed buffer.
735    
736    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
737        exec() might read past the end of the data buffer in UTF-8 mode.
738    
739    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
740        the data contained the byte 0x85 as part of a UTF-8 character within its
741        first line. This applied both to normal and DFA matching.
742    
743    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
744        /^[^d]*?$/8 failed to match "abc".
745    
746    16. Added a missing copyright notice to pcrecpp_internal.h.
747    
748    17. Make it more clear in the documentation that values returned from
749        pcre_exec() in ovector are byte offsets, not character counts.
750    
751    18. Tidied a few places to stop certain compilers from issuing warnings.
752    
753    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
754        supplied by Stefan Weber. I made a further small update for 7.8 because
755        there is a change of source arrangements: the pcre_searchfuncs.c module is
756        replaced by pcre_ucd.c.
757    
758    
759    Version 7.7 07-May-08
760    ---------------------
761    
762    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
763        a string to a long long, pretend we don't even have a long long." This is
764        done by checking for the strtoq, strtoll, and _strtoi64 functions.
765    
766    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
767        pre-7.6 versions, which defined a global no_arg variable instead of putting
768        it in the RE class. (See also #8 below.)
769    
770    3.  Remove a line of dead code, identified by coverity and reported by Nuno
771        Lopes.
772    
773    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
774    
775        (1) The include/exclude patterns were being applied to the whole pathnames
776            of files, instead of just to the final components.
777    
778        (2) If there was more than one level of directory, the subdirectories were
779            skipped unless they satisfied the include/exclude conditions. This is
780            inconsistent with GNU grep (and could even be seen as contrary to the
781            pcregrep specification - which I improved to make it absolutely clear).
782            The action now is always to scan all levels of directory, and just
783            apply the include/exclude patterns to regular files.
784    
785    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
786        --exclude_dir in the tests to avoid scanning .svn directories.
787    
788    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
789        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
790        doesn't support NULs in patterns.
791    
792    7.  Added some missing "const"s to declarations of static tables in
793        pcre_compile.c and pcre_dfa_exec.c.
794    
795    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
796        caused by fix #2  above. (Subsequently also a second patch to fix the
797        first patch. And a third patch - this was a messy problem.)
798    
799    9.  Applied Craig's patch to remove the use of push_back().
800    
801    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
802        matching function regexec().
803    
804    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
805        which, however, unlike Perl's \g{...}, are subroutine calls, not back
806        references. PCRE supports relative numbers with this syntax (I don't think
807        Oniguruma does).
808    
809    12. Previously, a group with a zero repeat such as (...){0} was completely
810        omitted from the compiled regex. However, this means that if the group
811        was called as a subroutine from elsewhere in the pattern, things went wrong
812        (an internal error was given). Such groups are now left in the compiled
813        pattern, with a new opcode that causes them to be skipped at execution
814        time.
815    
816    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
817        to the way PCRE behaves:
818    
819        (a) A lone ] character is dis-allowed (Perl treats it as data).
820    
821        (b) A back reference to an unmatched subpattern matches an empty string
822            (Perl fails the current match path).
823    
824        (c) A data ] in a character class must be notated as \] because if the
825            first data character in a class is ], it defines an empty class. (In
826            Perl it is not possible to have an empty class.) The empty class []
827            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
828            The negative empty class [^] matches any one character, independently
829            of the DOTALL setting.
830    
831    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
832        non-existent subpattern following a character class starting with ']' and
833        containing () gave an internal compiling error instead of "reference to
834        non-existent subpattern". Fortunately, when the pattern did exist, the
835        compiled code was correct. (When scanning forwards to check for the
836        existencd of the subpattern, it was treating the data ']' as terminating
837        the class, so got the count wrong. When actually compiling, the reference
838        was subsequently set up correctly.)
839    
840    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
841        it was being rejected as not supported by pcre_dfa_exec(), even though
842        other assertions are supported. I have made pcre_dfa_exec() support
843        (*FAIL).
844    
845    16. The implementation of 13c above involved the invention of a new opcode,
846        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
847        cannot be changed at match time, I realized I could make a small
848        improvement to matching performance by compiling OP_ALLANY instead of
849        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
850        on the OP_ANY path.
851    
852    17. Compiling pcretest on Windows with readline support failed without the
853        following two fixes: (1) Make the unistd.h include conditional on
854        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
855    
856    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
857        ncurses library to be included for pcretest when ReadLine support is
858        requested, but also to allow for it to be overridden. This patch came from
859        Daniel Bergström.
860    
861    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
862        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
863        any errors with the current Unicode tables. Thanks to Peter Kankowski for
864        spotting this.
865    
866    
867    Version 7.6 28-Jan-08
868    ---------------------
869    
870    1.  A character class containing a very large number of characters with
871        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
872        overflow.
873    
874    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
875        HAVE_LONG_LONG is not defined.
876    
877    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
878        bring them up to date and include new features. This patch includes:
879    
880        - Fixed PH's badly added libz and libbz2 support.
881        - Fixed a problem with static linking.
882        - Added pcredemo. [But later removed - see 7 below.]
883        - Fixed dftables problem and added an option.
884        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
885            HAVE_LONG_LONG.
886        - Added readline support for pcretest.
887        - Added an listing of the option settings after cmake has run.
888    
889    4.  A user submitted a patch to Makefile that makes it easy to create
890        "pcre.dll" under mingw when using Configure/Make. I added stuff to
891        Makefile.am that cause it to include this special target, without
892        affecting anything else. Note that the same mingw target plus all
893        the other distribution libraries and programs are now supported
894        when configuring with CMake (see 6 below) instead of with
895        Configure/Make.
896    
897    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
898        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
899        exported in the Windows port". It has not yet been confirmed that the patch
900        solves the problem, but it does no harm.
901    
902    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
903        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
904        with CMake, and also correct the comment about stack recursion.
905    
906    7.  Remove the automatic building of pcredemo from the ./configure system and
907        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
908        of a program that users should build themselves after PCRE is installed, so
909        building it automatically is not really right. What is more, it gave
910        trouble in some build environments.
911    
912    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
913    
914    
915    Version 7.5 10-Jan-08
916    ---------------------
917    
918    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
919        values in parens when parsing an RE using the C++ wrapper."
920    
921    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
922        Characters greater than 255 were excluded from the class instead of being
923        included.
924    
925    3.  The same bug as (2) above applied to negated POSIX classes such as
926        [:^space:].
927    
928    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
929        defined or documented. It seems to have been a typo for PCRE_STATIC, so
930        I have changed it.
931    
932    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
933        first named subpattern) and a construct such as (?&a) would reference the
934        first named subpattern whose name started with "a" (in other words, the
935        length check was missing). Both these problems are fixed. "Subpattern name
936        expected" is now given for (?&) (a zero-length name), and this patch also
937        makes it give the same error for \k'' (previously it complained that that
938        was a reference to a non-existent subpattern).
939    
940    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
941        this is right because (?- can be followed by option settings as well as by
942        digits. I have, however, made the messages clearer.
943    
944    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
945        than the number used in the conditional) now cause a compile-time error.
946        This is actually not compatible with Perl, which accepts such patterns, but
947        treats the conditional as always being FALSE (as PCRE used to), but it
948        seems to me that giving a diagnostic is better.
949    
950    8.  Change "alphameric" to the more common word "alphanumeric" in comments
951        and messages.
952    
953    9.  Fix two occurrences of "backslash" in comments that should have been
954        "backspace".
955    
956    10. Remove two redundant lines of code that can never be obeyed (their function
957        was moved elsewhere).
958    
959    11. The program that makes PCRE's Unicode character property table had a bug
960        which caused it to generate incorrect table entries for sequences of
961        characters that have the same character type, but are in different scripts.
962        It amalgamated them into a single range, with the script of the first of
963        them. In other words, some characters were in the wrong script. There were
964        thirteen such cases, affecting characters in the following ranges:
965    
966          U+002b0 - U+002c1
967          U+0060c - U+0060d
968          U+0061e - U+00612
969          U+0064b - U+0065e
970          U+0074d - U+0076d
971          U+01800 - U+01805
972          U+01d00 - U+01d77
973          U+01d9b - U+01dbf
974          U+0200b - U+0200f
975          U+030fc - U+030fe
976          U+03260 - U+0327f
977          U+0fb46 - U+0fbb1
978          U+10450 - U+1049d
979    
980    12. The -o option (show only the matching part of a line) for pcregrep was not
981        compatible with GNU grep in that, if there was more than one match in a
982        line, it showed only the first of them. It now behaves in the same way as
983        GNU grep.
984    
985    13. If the -o and -v options were combined for pcregrep, it printed a blank
986        line for every non-matching line. GNU grep prints nothing, and pcregrep now
987        does the same. The return code can be used to tell if there were any
988        non-matching lines.
989    
990    14. Added --file-offsets and --line-offsets to pcregrep.
991    
992    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
993        infinitely looping recursion. The bug was that positive lookaheads were not
994        being skipped when checking for a possible empty match (negative lookaheads
995        and both kinds of lookbehind were skipped).
996    
997    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
998        inclusion of <windows.h> to before rather than after the definition of
999        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1000    
1001    17. Specifying a possessive quantifier with a specific limit for a Unicode
1002        character property caused pcre_compile() to compile bad code, which led at
1003        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1004        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1005        caused the error; without that there was no problem.
1006    
1007    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1008    
1009    19. Added --enable-pcretest-libreadline.
1010    
1011    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1012        RE::GlobalReplace(). As a result, the number of replacements returned was
1013        double what it should be. I removed one of the increments, but Craig sent a
1014        later patch that removed the other one (the right fix) and added unit tests
1015        that check the return values (which was not done before).
1016    
1017    21. Several CMake things:
1018    
1019        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1020            the names libpcre and libpcreposix, not just pcre and pcreposix.
1021    
1022        (2) The above change means that pcretest and pcregrep are now correctly
1023            linked with the newly-built libraries, not previously installed ones.
1024    
1025        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1026    
1027    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1028        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1029        UTF-8 newline character). The key issue is that the pattern starts .*;
1030        this means that the match must be either at the beginning, or after a
1031        newline. The bug was in the code for advancing after a failed match and
1032        checking that the new position followed a newline. It was not taking
1033        account of UTF-8 characters correctly.
1034    
1035    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1036        character classes. PCRE was not treating the sequence [:...:] as a
1037        character class unless the ... were all letters. Perl, however, seems to
1038        allow any characters between [: and :], though of course it rejects as
1039        unknown any "names" that contain non-letters, because all the known class
1040        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1041        for example, whereas PCRE did not - it did not recognize a POSIX character
1042        class. This seemed a bit dangerous, so the code has been changed to be
1043        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1044        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1045        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1046        Perl does, and where it didn't before.
1047    
1048    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1049        Windows environments %n is disabled by default.
1050    
1051    
1052    Version 7.4 21-Sep-07
1053    ---------------------
1054    
1055    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1056        means that a class such as [\s] counted as "explicit reference to CR or
1057        LF". That isn't really right - the whole point of the change was to try to
1058        help when there was an actual mention of one of the two characters. So now
1059        the change happens only if \r or \n (or a literal CR or LF) character is
1060        encountered.
1061    
1062    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1063        of both had grown to the point where there were only 3 bits left.
1064        Fortunately, there was spare space in the data structure, and so I have
1065        moved the internal flags into a new 16-bit field to free up more option
1066        bits.
1067    
1068    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1069        but did not set the internal JCHANGED flag - either of these is enough to
1070        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1071        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1072        start sets both bits.
1073    
1074    4.  Added options (at build time, compile time, exec time) to change \R from
1075        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1076    
1077    5.  doc/pcresyntax.html was missing from the distribution.
1078    
1079    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1080        compatibility, even though it is no longer used.
1081    
1082    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1083        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1084        windows.h file is present (where different names are used). [This was
1085        reversed later after testing - see 16 below.]
1086    
1087    8.  Changed all #include <config.h> to #include "config.h". There were also
1088        some further <pcre.h> cases that I changed to "pcre.h".
1089    
1090    9.  When pcregrep was used with the --colour option, it missed the line ending
1091        sequence off the lines that it output.
1092    
1093    10. It was pointed out to me that arrays of string pointers cause lots of
1094        relocations when a shared library is dynamically loaded. A technique of
1095        using a single long string with a table of offsets can drastically reduce
1096        these. I have refactored PCRE in four places to do this. The result is
1097        dramatic:
1098    
1099          Originally:                          290
1100          After changing UCP table:            187
1101          After changing error message table:   43
1102          After changing table of "verbs"       36
1103          After changing table of Posix names   22
1104    
1105        Thanks to the folks working on Gregex for glib for this insight.
1106    
1107    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1108        unicode-properties was also set.
1109    
1110    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1111    
1112    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1113        checked only for CRLF.
1114    
1115    14. Added casts to pcretest.c to avoid compiler warnings.
1116    
1117    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1118    
1119    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1120        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1121        entirely. This removes changes made in 7 above.
1122    
1123    17. The CMake files have been updated, and there is now more information about
1124        building with CMake in the NON-UNIX-USE document.
1125    
1126    
1127    Version 7.3 28-Aug-07
1128    ---------------------
1129    
1130     1. In the rejigging of the build system that eventually resulted in 7.1, the
1131        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1132        brackets there is not right, since it causes compilers to look for an
1133        installed pcre.h, not the version that is in the source that is being
1134        compiled (which of course may be different). I have changed it back to:
1135    
1136          #include "pcre.h"
1137    
1138        I have a vague recollection that the change was concerned with compiling in
1139        different directories, but in the new build system, that is taken care of
1140        by the VPATH setting the Makefile.
1141    
1142     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1143        when the subject happened to end in the byte 0x85 (e.g. if the last
1144        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1145        characters but of course it shouldn't be taken as a newline when it is part
1146        of another character. The bug was that, for an unlimited repeat of . in
1147        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1148        characters when looking for a newline.
1149    
1150     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1151    
1152     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1153        in debug output.
1154    
1155     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1156        long printing in the pcrecpp unittest when running under MinGW.
1157    
1158     6. ESC_K was left out of the EBCDIC table.
1159    
1160     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1161        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1162        limit also applies to "virtual nesting" when a pattern is recursive, and in
1163        this case 1000 isn't so big. I have been able to remove this limit at the
1164        expense of backing off one optimization in certain circumstances. Normally,
1165        when pcre_exec() would call its internal match() function recursively and
1166        immediately return the result unconditionally, it uses a "tail recursion"
1167        feature to save stack. However, when a subpattern that can match an empty
1168        string has an unlimited repetition quantifier, it no longer makes this
1169        optimization. That gives it a stack frame in which to save the data for
1170        checking that an empty string has been matched. Previously this was taken
1171        from the 1000-entry workspace that had been reserved. So now there is no
1172        explicit limit, but more stack is used.
1173    
1174     8. Applied Daniel's patches to solve problems with the import/export magic
1175        syntax that is required for Windows, and which was going wrong for the
1176        pcreposix and pcrecpp parts of the library. These were overlooked when this
1177        problem was solved for the main library.
1178    
1179     9. There were some crude static tests to avoid integer overflow when computing
1180        the size of patterns that contain repeated groups with explicit upper
1181        limits. As the maximum quantifier is 65535, the maximum group length was
1182        set at 30,000 so that the product of these two numbers did not overflow a
1183        32-bit integer. However, it turns out that people want to use groups that
1184        are longer than 30,000 bytes (though not repeat them that many times).
1185        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1186        made it possible to implement the integer overflow checks in a much more
1187        dynamic way, which I have now done. The artificial limitation on group
1188        length has been removed - we now have only the limit on the total length of
1189        the compiled pattern, which depends on the LINK_SIZE setting.
1190    
1191    10. Fixed a bug in the documentation for get/copy named substring when
1192        duplicate names are permitted. If none of the named substrings are set, the
1193        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1194        empty string.
1195    
1196    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1197        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1198        because the ] is interpreted as the first data character and the
1199        terminating ] is not found. PCRE has been made compatible with Perl in this
1200        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1201        cause memory overwriting.
1202    
1203    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1204        string has been matched (to stop an infinite loop). It was not recognizing
1205        a conditional subpattern that could match an empty string if that
1206        subpattern was within another subpattern. For example, it looped when
1207        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1208        condition was not nested. This bug has been fixed.
1209    
1210    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1211        past the start of the subject in the presence of bytes with the top bit
1212        set, for example "\x8aBCD".
1213    
1214    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1215        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1216    
1217    14. Optimized (?!) to (*FAIL).
1218    
1219    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1220        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1221        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1222        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1223        does: it's just the validity check that is more restrictive.
1224    
1225    16. Inserted checks for integer overflows during escape sequence (backslash)
1226        processing, and also fixed erroneous offset values for syntax errors during
1227        backslash processing.
1228    
1229    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1230        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1231    
1232    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1233        caused an overrun.
1234    
1235    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1236        something other than just ASCII characters) inside a group that had an
1237        unlimited repeat caused a loop at compile time (while checking to see
1238        whether the group could match an empty string).
1239    
1240    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1241        [\P{Any}] did so. (Error in the code for printing property names.)
1242    
1243    21. An orphan \E inside a character class could cause a crash.
1244    
1245    22. A repeated capturing bracket such as (A)? could cause a wild memory
1246        reference during compilation.
1247    
1248    23. There are several functions in pcre_compile() that scan along a compiled
1249        expression for various reasons (e.g. to see if it's fixed length for look
1250        behind). There were bugs in these functions when a repeated \p or \P was
1251        present in the pattern. These operators have additional parameters compared
1252        with \d, etc, and these were not being taken into account when moving along
1253        the compiled data. Specifically:
1254    
1255        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1256            length.
1257    
1258        (b) An item such as \pL+ within a repeated group could cause crashes or
1259            loops.
1260    
1261        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1262            "reference to non-existent subpattern" error.
1263    
1264        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1265    
1266    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1267        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1268    
1269    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1270    
1271    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1272        character were causing crashes (broken optimization).
1273    
1274    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1275        \p or \P) caused a compile-time loop.
1276    
1277    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1278        break. For example, the unstudied pattern [\r\n]A does not match the string
1279        "\r\nA" because change 7.0/46 below moves the current point on by two
1280        characters after failing to match at the start. However, the pattern \nA
1281        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1282        the same is true. There doesn't seem any very clean way out of this, but
1283        what I have chosen to do makes the common cases work: PCRE now takes note
1284        of whether there can be an explicit match for \r or \n anywhere in the
1285        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1286        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1287        pattern has explicit CR or LF references.
1288    
1289    29. Added (*CR) etc for changing newline setting at start of pattern.
1290    
1291    
1292    Version 7.2 19-Jun-07
1293  ---------------------  ---------------------
1294    
1295   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1296      which is apparently normally available under Windows.      which is apparently normally available under Windows.
1297    
1298   2. Re-jig the pcregrep tests with different newline settings in an attempt   2. Re-jig the pcregrep tests with different newline settings in an attempt
1299      to make them independent of the local environment's newline setting.      to make them independent of the local environment's newline setting.
1300    
1301     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1302    
1303     4. Some of the "internals" tests were previously cut out when the link size
1304        was not 2, because the output contained actual offsets. The recent new
1305        "Z" feature of pcretest means that these can be cut out, making the tests
1306        usable with all link sizes.
1307    
1308     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1309        stack recursion. This gives a massive performance boost under BSD, but just
1310        a small improvement under Linux. However, it saves one field in the frame
1311        in all cases.
1312    
1313     6. Added more features from the forthcoming Perl 5.10:
1314    
1315        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1316            recursion call. It refers to the nth most recently opened parentheses.
1317    
1318        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1319            to be opened parentheses.
1320    
1321        (c) Conditions that refer to capturing parentheses can be specified
1322            relatively, for example, (?(-2)... or (?(+3)...
1323    
1324        (d) \K resets the start of the current match so that everything before
1325            is not part of it.
1326    
1327        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1328    
1329        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1330            reference syntax.
1331    
1332        (g) (?| introduces a group in which the numbering of parentheses in each
1333            alternative starts with the same number.
1334    
1335        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1336    
1337     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1338        PCRE_INFO_JCHANGED.
1339    
1340     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1341        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1342        for detecting groups that can match an empty string.
1343    
1344     9. A pattern with a very large number of alternatives (more than several
1345        hundred) was running out of internal workspace during the pre-compile
1346        phase, where pcre_compile() figures out how much memory will be needed. A
1347        bit of new cunning has reduced the workspace needed for groups with
1348        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1349        workspace instead of running out of the 4096 that are available.
1350    
1351    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1352    
1353    11. Applied patch from Google to remove an optimization that didn't quite work.
1354        The report of the bug said:
1355    
1356          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1357          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1358          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1359    
1360    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1361        it matched the wrong number of bytes.
1362    
1363    
1364  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
1365  ---------------------  ---------------------

Legend:
Removed from v.158  
changed lines
  Added in v.571

  ViewVC Help
Powered by ViewVC 1.1.5