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

Diff of /code/trunk/ChangeLog

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

revision 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 482 by ph10, Mon Jan 4 15:55:46 2010 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.01 11-Dec-09
5    ----------------------
6    
7    1.  If a pattern contained a conditional subpattern with only one branch (in
8        particular, this includes all (DEFINE) patterns), a call to pcre_study()
9        computed the wrong minimum data length (which is of course zero for such
10        subpatterns).
11    
12    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
13        the pattern is reset in the first branch, pcre_compile() failed with
14        "internal error: code overflow at offset...". This happened only when
15        the reset was to the original external option setting. (An optimization
16        abstracts leading options settings into an external setting, which was the
17        cause of this.)
18    
19    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
20        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
21        assertion pattern did not match (meaning that the assertion was true), it
22        was incorrectly treated as false if the SKIP had been reached during the
23        matching. This also applied to assertions used as conditions.
24    
25    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
26        assertion subpattern, including such a pattern used as a condition,
27        unpredictable results occurred, instead of the error return
28        PCRE_ERROR_DFA_UITEM.
29    
30    5.  The C++ GlobalReplace function was not working like Perl for the special
31        situation when an empty string is matched. It now does the fancy magic
32        stuff that is necessary.
33    
34    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
35        removed. (These were left over from very, very early versions of PCRE.)
36    
37    7.  Some cosmetic changes to the code to make life easier when compiling it
38        as part of something else:
39    
40        (a) Change DEBUG to PCRE_DEBUG.
41    
42        (b) In pcre_compile(), rename the member of the "branch_chain" structure
43            called "current" as "current_branch", to prevent a collision with the
44            Linux macro when compiled as a kernel module.
45    
46        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
47            prevent a collision with the Linux macro when compiled as a kernel
48            module.
49    
50    8.  In pcre_compile() there are some checks for integer overflows that used to
51        cast potentially large values to (double). This has been changed to that
52        when building, a check for int64_t is made, and if it is found, it is used
53        instead, thus avoiding the use of floating point arithmetic. (There is no
54        other use of FP in PCRE.) If int64_t is not found, the fallback is to
55        double.
56    
57    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
58        2005 (difference between two addresses compared to an unsigned value).
59    
60    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
61        custom one, because of the following reported problem in Windows:
62    
63          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
64            under Win32.
65          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
66            therefore missing the function definition.
67          - The compiler thus generates a "C" signature for the test function.
68          - The linker fails to find the "C" function.
69          - PCRE fails to configure if asked to do so against libbz2.
70    
71    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
72        messages were output:
73    
74          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
75          rerunning libtoolize, to keep the correct libtool macros in-tree.
76          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
77    
78        I have done both of these things.
79    
80    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
81        most of the time, it *can* run out if it is given a pattern that contains a
82        runaway infinite recursion. I updated the discussion in the pcrestack man
83        page.
84    
85    13. Now that we have gone to the x.xx style of version numbers, the minor
86        version may start with zero. Using 08 or 09 is a bad idea because users
87        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
88        interpreted as invalid octal numbers. I've updated the previous comment in
89        configure.ac, and also added a check that gives an error if 08 or 09 are
90        used.
91    
92    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
93        causing partial matching to fail where the end of the subject matched \W.
94    
95    
96    
97    Version 8.00 19-Oct-09
98    ----------------------
99    
100    1.  The table for translating pcre_compile() error codes into POSIX error codes
101        was out-of-date, and there was no check on the pcre_compile() error code
102        being within the table. This could lead to an OK return being given in
103        error.
104    
105    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
106        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
107        in a Windows environment.
108    
109    3.  The pcregrep --count option prints the count for each file even when it is
110        zero, as does GNU grep. However, pcregrep was also printing all files when
111        --files-with-matches was added. Now, when both options are given, it prints
112        counts only for those files that have at least one match. (GNU grep just
113        prints the file name in this circumstance, but including the count seems
114        more useful - otherwise, why use --count?) Also ensured that the
115        combination -clh just lists non-zero counts, with no names.
116    
117    4.  The long form of the pcregrep -F option was incorrectly implemented as
118        --fixed_strings instead of --fixed-strings. This is an incompatible change,
119        but it seems right to fix it, and I didn't think it was worth preserving
120        the old behaviour.
121    
122    5.  The command line items --regex=pattern and --regexp=pattern were not
123        recognized by pcregrep, which required --regex pattern or --regexp pattern
124        (with a space rather than an '='). The man page documented the '=' forms,
125        which are compatible with GNU grep; these now work.
126    
127    6.  No libpcreposix.pc file was created for pkg-config; there was just
128        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
129    
130    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
131        when UCP support is not needed, by modifying the Python script that
132        generates it from Unicode data files. This should not matter if the module
133        is correctly used as a library, but I received one complaint about 50K of
134        unwanted data. My guess is that the person linked everything into his
135        program rather than using a library. Anyway, it does no harm.
136    
137    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
138        was a minimum greater than 1 for a wide character in a possessive
139        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
140        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
141        character. Chaos in the form of incorrect output or a compiling loop could
142        result.
143    
144    9.  The restrictions on what a pattern can contain when partial matching is
145        requested for pcre_exec() have been removed. All patterns can now be
146        partially matched by this function. In addition, if there are at least two
147        slots in the offset vector, the offset of the earliest inspected character
148        for the match and the offset of the end of the subject are set in them when
149        PCRE_ERROR_PARTIAL is returned.
150    
151    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
152        synonymous with PCRE_PARTIAL, for backwards compatibility, and
153        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
154        and may be more useful for multi-segment matching.
155    
156    11. Partial matching with pcre_exec() is now more intuitive. A partial match
157        used to be given if ever the end of the subject was reached; now it is
158        given only if matching could not proceed because another character was
159        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
160        string "Z", which now yields "no match" instead of "partial match". In the
161        case of pcre_dfa_exec(), "no match" is given if every matching path for the
162        final character ended with (*FAIL).
163    
164    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
165        if the pattern had a "must contain" character that was already found in the
166        earlier partial match, unless partial matching was again requested. For
167        example, with the pattern /dog.(body)?/, the "must contain" character is
168        "g". If the first part-match was for the string "dog", restarting with
169        "sbody" failed. This bug has been fixed.
170    
171    13. The string returned by pcre_dfa_exec() after a partial match has been
172        changed so that it starts at the first inspected character rather than the
173        first character of the match. This makes a difference only if the pattern
174        starts with a lookbehind assertion or \b or \B (\K is not supported by
175        pcre_dfa_exec()). It's an incompatible change, but it makes the two
176        matching functions compatible, and I think it's the right thing to do.
177    
178    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
179        so that the demonstration program is easily available in environments where
180        PCRE has not been installed from source.
181    
182    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
183        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
184        library.
185    
186    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
187        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
188        is not the first non-POSIX option to be added. Clearly some people find
189        these options useful.
190    
191    17. If a caller to the POSIX matching function regexec() passes a non-zero
192        value for nmatch with a NULL value for pmatch, the value of
193        nmatch is forced to zero.
194    
195    18. RunGrepTest did not have a test for the availability of the -u option of
196        the diff command, as RunTest does. It now checks in the same way as
197        RunTest, and also checks for the -b option.
198    
199    19. If an odd number of negated classes containing just a single character
200        interposed, within parentheses, between a forward reference to a named
201        subpattern and the definition of the subpattern, compilation crashed with
202        an internal error, complaining that it could not find the referenced
203        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
204        [The bug was that it was starting one character too far in when skipping
205        over the character class, thus treating the ] as data rather than
206        terminating the class. This meant it could skip too much.]
207    
208    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
209        /g option in pcretest when the pattern contains \K, which makes it possible
210        to have an empty string match not at the start, even when the pattern is
211        anchored. Updated pcretest and pcredemo to use this option.
212    
213    21. If the maximum number of capturing subpatterns in a recursion was greater
214        than the maximum at the outer level, the higher number was returned, but
215        with unset values at the outer level. The correct (outer level) value is
216        now given.
217    
218    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
219        PCRE did not set those parentheses (unlike Perl). I have now found a way to
220        make it do so. The string so far is captured, making this feature
221        compatible with Perl.
222    
223    23. The tests have been re-organized, adding tests 11 and 12, to make it
224        possible to check the Perl 5.10 features against Perl 5.10.
225    
226    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
227        pattern matches a fixed length string. PCRE did not allow this; now it
228        does. Neither allows recursion.
229    
230    25. I finally figured out how to implement a request to provide the minimum
231        length of subject string that was needed in order to match a given pattern.
232        (It was back references and recursion that I had previously got hung up
233        on.) This code has now been added to pcre_study(); it finds a lower bound
234        to the length of subject needed. It is not necessarily the greatest lower
235        bound, but using it to avoid searching strings that are too short does give
236        some useful speed-ups. The value is available to calling programs via
237        pcre_fullinfo().
238    
239    26. While implementing 25, I discovered to my embarrassment that pcretest had
240        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
241        study optimizations had never been tested with that matching function.
242        Oops. What is worse, even when it was passed study data, there was a bug in
243        pcre_dfa_exec() that meant it never actually used it. Double oops. There
244        were also very few tests of studied patterns with pcre_dfa_exec().
245    
246    27. If (?| is used to create subpatterns with duplicate numbers, they are now
247        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
248        on the other side of the coin, they are no longer allowed to have different
249        names, because these cannot be distinguished in PCRE, and this has caused
250        confusion. (This is a difference from Perl.)
251    
252    28. When duplicate subpattern names are present (necessarily with different
253        numbers, as required by 27 above), and a test is made by name in a
254        conditional pattern, either for a subpattern having been matched, or for
255        recursion in such a pattern, all the associated numbered subpatterns are
256        tested, and the overall condition is true if the condition is true for any
257        one of them. This is the way Perl works, and is also more like the way
258        testing by number works.
259    
260    
261    Version 7.9 11-Apr-09
262    ---------------------
263    
264    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
265        (pcretest), all targets were linked against these libraries. This included
266        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
267        libraries. This caused unwanted dependencies to be created. This problem
268        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
269        pcretest is linked with readline.
270    
271    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
272        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
273        moved outside it again, because FALSE and TRUE are already defined in AIX,
274        but BOOL is not.
275    
276    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
277        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
278    
279    4.  The pcregrep documentation said spaces were inserted as well as colons (or
280        hyphens) following file names and line numbers when outputting matching
281        lines. This is not true; no spaces are inserted. I have also clarified the
282        wording for the --colour (or --color) option.
283    
284    5.  In pcregrep, when --colour was used with -o, the list of matching strings
285        was not coloured; this is different to GNU grep, so I have changed it to be
286        the same.
287    
288    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
289        each matching line was coloured. Now it goes on to look for further matches
290        of any of the test patterns, which is the same behaviour as GNU grep.
291    
292    7.  A pattern that could match an empty string could cause pcregrep to loop; it
293        doesn't make sense to accept an empty string match in pcregrep, so I have
294        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
295        seems to be how GNU grep behaves.
296    
297    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
298        start or after a newline", because the conditional assertion was not being
299        correctly handled. The rule now is that both the assertion and what follows
300        in the first alternative must satisfy the test.
301    
302    9.  If auto-callout was enabled in a pattern with a conditional group whose
303        condition was an assertion, PCRE could crash during matching, both with
304        pcre_exec() and pcre_dfa_exec().
305    
306    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
307        used for matching.
308    
309    11. Unicode property support in character classes was not working for
310        characters (bytes) greater than 127 when not in UTF-8 mode.
311    
312    12. Added the -M command line option to pcretest.
313    
314    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
315    
316    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
317    
318    16. Added comments and documentation about mis-use of no_arg in the C++
319        wrapper.
320    
321    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
322        from Martin Jerabek that uses macro names for all relevant character and
323        string constants.
324    
325    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
326        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
327        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
328        these, but not everybody uses configure.
329    
330    19. A conditional group that had only one branch was not being correctly
331        recognized as an item that could match an empty string. This meant that an
332        enclosing group might also not be so recognized, causing infinite looping
333        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
334        with the subject "ab", where knowledge that the repeated group can match
335        nothing is needed in order to break the loop.
336    
337    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
338        exec(), but without supplying a callout function, matching went wrong.
339    
340    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
341        leak if the size of the offset vector was greater than 30. When the vector
342        is smaller, the saved offsets during recursion go onto a local stack
343        vector, but for larger vectors malloc() is used. It was failing to free
344        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
345        error, in fact).
346    
347    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
348        heapframe that is used only when UTF-8 support is enabled. This caused no
349        problem, but was untidy.
350    
351    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
352        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
353        included within another project.
354    
355    24. Steven Van Ingelgem's patches to add more options to the CMake support,
356        slightly modified by me:
357    
358          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
359              not building pcregrep.
360    
361          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
362              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
363    
364    25. Forward references, both numeric and by name, in patterns that made use of
365        duplicate group numbers, could behave incorrectly or give incorrect errors,
366        because when scanning forward to find the reference group, PCRE was not
367        taking into account the duplicate group numbers. A pattern such as
368        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
369    
370    26. Changed a few more instances of "const unsigned char *" to USPTR, making
371        the feature of a custom pointer more persuasive (as requested by a user).
372    
373    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
374        pcretest.c, inside #ifndefs, because it seems they are sometimes already
375        pre-defined.
376    
377    28. Added support for (*UTF8) at the start of a pattern.
378    
379    29. Arrange for flags added by the "release type" setting in CMake to be shown
380        in the configuration summary.
381    
382    
383    Version 7.8 05-Sep-08
384    ---------------------
385    
386    1.  Replaced UCP searching code with optimized version as implemented for Ad
387        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
388        stage table and inline lookup instead of a function, giving speed ups of 2
389        to 5 times on some simple patterns that I tested. Permission was given to
390        distribute the MultiStage2.py script that generates the tables (it's not in
391        the tarball, but is in the Subversion repository).
392    
393    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
394        scripts.
395    
396    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
397        a group with a zero qualifier. The result of the study could be incorrect,
398        or the function might crash, depending on the pattern.
399    
400    4.  Caseless matching was not working for non-ASCII characters in back
401        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
402        It now works when Unicode Property Support is available.
403    
404    5.  In pcretest, an escape such as \x{de} in the data was always generating
405        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
406        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
407        truncation.
408    
409    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
410    
411    7.  Added two (int) casts to pcregrep when printing the difference of two
412        pointers, in case they are 64-bit values.
413    
414    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
415        test 2 if it fails.
416    
417    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
418        and a #define of that name to empty if it is not externally set. This is to
419        allow users of MSVC to set it if necessary.
420    
421    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
422        the convenience functions in the pcre_get.c source file.
423    
424    11. An option change at the start of a pattern that had top-level alternatives
425        could cause overwriting and/or a crash. This command provoked a crash in
426        some environments:
427    
428          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
429    
430        This potential security problem was recorded as CVE-2008-2371.
431    
432    12. For a pattern where the match had to start at the beginning or immediately
433        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
434        pcre_dfa_exec() could read past the end of the passed subject if there was
435        no match. To help with detecting such bugs (e.g. with valgrind), I modified
436        pcretest so that it places the subject at the end of its malloc-ed buffer.
437    
438    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
439        exec() might read past the end of the data buffer in UTF-8 mode.
440    
441    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
442        the data contained the byte 0x85 as part of a UTF-8 character within its
443        first line. This applied both to normal and DFA matching.
444    
445    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
446        /^[^d]*?$/8 failed to match "abc".
447    
448    16. Added a missing copyright notice to pcrecpp_internal.h.
449    
450    17. Make it more clear in the documentation that values returned from
451        pcre_exec() in ovector are byte offsets, not character counts.
452    
453    18. Tidied a few places to stop certain compilers from issuing warnings.
454    
455    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
456        supplied by Stefan Weber. I made a further small update for 7.8 because
457        there is a change of source arrangements: the pcre_searchfuncs.c module is
458        replaced by pcre_ucd.c.
459    
460    
461    Version 7.7 07-May-08
462    ---------------------
463    
464    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
465        a string to a long long, pretend we don't even have a long long." This is
466        done by checking for the strtoq, strtoll, and _strtoi64 functions.
467    
468    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
469        pre-7.6 versions, which defined a global no_arg variable instead of putting
470        it in the RE class. (See also #8 below.)
471    
472    3.  Remove a line of dead code, identified by coverity and reported by Nuno
473        Lopes.
474    
475    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
476    
477        (1) The include/exclude patterns were being applied to the whole pathnames
478            of files, instead of just to the final components.
479    
480        (2) If there was more than one level of directory, the subdirectories were
481            skipped unless they satisfied the include/exclude conditions. This is
482            inconsistent with GNU grep (and could even be seen as contrary to the
483            pcregrep specification - which I improved to make it absolutely clear).
484            The action now is always to scan all levels of directory, and just
485            apply the include/exclude patterns to regular files.
486    
487    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
488        --exclude_dir in the tests to avoid scanning .svn directories.
489    
490    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
491        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
492        doesn't support NULs in patterns.
493    
494    7.  Added some missing "const"s to declarations of static tables in
495        pcre_compile.c and pcre_dfa_exec.c.
496    
497    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
498        caused by fix #2  above. (Subsequently also a second patch to fix the
499        first patch. And a third patch - this was a messy problem.)
500    
501    9.  Applied Craig's patch to remove the use of push_back().
502    
503    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
504        matching function regexec().
505    
506    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
507        which, however, unlike Perl's \g{...}, are subroutine calls, not back
508        references. PCRE supports relative numbers with this syntax (I don't think
509        Oniguruma does).
510    
511    12. Previously, a group with a zero repeat such as (...){0} was completely
512        omitted from the compiled regex. However, this means that if the group
513        was called as a subroutine from elsewhere in the pattern, things went wrong
514        (an internal error was given). Such groups are now left in the compiled
515        pattern, with a new opcode that causes them to be skipped at execution
516        time.
517    
518    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
519        to the way PCRE behaves:
520    
521        (a) A lone ] character is dis-allowed (Perl treats it as data).
522    
523        (b) A back reference to an unmatched subpattern matches an empty string
524            (Perl fails the current match path).
525    
526        (c) A data ] in a character class must be notated as \] because if the
527            first data character in a class is ], it defines an empty class. (In
528            Perl it is not possible to have an empty class.) The empty class []
529            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
530            The negative empty class [^] matches any one character, independently
531            of the DOTALL setting.
532    
533    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
534        non-existent subpattern following a character class starting with ']' and
535        containing () gave an internal compiling error instead of "reference to
536        non-existent subpattern". Fortunately, when the pattern did exist, the
537        compiled code was correct. (When scanning forwards to check for the
538        existencd of the subpattern, it was treating the data ']' as terminating
539        the class, so got the count wrong. When actually compiling, the reference
540        was subsequently set up correctly.)
541    
542    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
543        it was being rejected as not supported by pcre_dfa_exec(), even though
544        other assertions are supported. I have made pcre_dfa_exec() support
545        (*FAIL).
546    
547    16. The implementation of 13c above involved the invention of a new opcode,
548        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
549        cannot be changed at match time, I realized I could make a small
550        improvement to matching performance by compiling OP_ALLANY instead of
551        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
552        on the OP_ANY path.
553    
554    17. Compiling pcretest on Windows with readline support failed without the
555        following two fixes: (1) Make the unistd.h include conditional on
556        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
557    
558    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
559        ncurses library to be included for pcretest when ReadLine support is
560        requested, but also to allow for it to be overridden. This patch came from
561        Daniel Bergström.
562    
563    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
564        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
565        any errors with the current Unicode tables. Thanks to Peter Kankowski for
566        spotting this.
567    
568    
569    Version 7.6 28-Jan-08
570    ---------------------
571    
572    1.  A character class containing a very large number of characters with
573        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
574        overflow.
575    
576    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
577        HAVE_LONG_LONG is not defined.
578    
579    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
580        bring them up to date and include new features. This patch includes:
581    
582        - Fixed PH's badly added libz and libbz2 support.
583        - Fixed a problem with static linking.
584        - Added pcredemo. [But later removed - see 7 below.]
585        - Fixed dftables problem and added an option.
586        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
587            HAVE_LONG_LONG.
588        - Added readline support for pcretest.
589        - Added an listing of the option settings after cmake has run.
590    
591    4.  A user submitted a patch to Makefile that makes it easy to create
592        "pcre.dll" under mingw when using Configure/Make. I added stuff to
593        Makefile.am that cause it to include this special target, without
594        affecting anything else. Note that the same mingw target plus all
595        the other distribution libraries and programs are now supported
596        when configuring with CMake (see 6 below) instead of with
597        Configure/Make.
598    
599    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
600        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
601        exported in the Windows port". It has not yet been confirmed that the patch
602        solves the problem, but it does no harm.
603    
604    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
605        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
606        with CMake, and also correct the comment about stack recursion.
607    
608    7.  Remove the automatic building of pcredemo from the ./configure system and
609        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
610        of a program that users should build themselves after PCRE is installed, so
611        building it automatically is not really right. What is more, it gave
612        trouble in some build environments.
613    
614    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
615    
616    
617    Version 7.5 10-Jan-08
618    ---------------------
619    
620    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
621        values in parens when parsing an RE using the C++ wrapper."
622    
623    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
624        Characters greater than 255 were excluded from the class instead of being
625        included.
626    
627    3.  The same bug as (2) above applied to negated POSIX classes such as
628        [:^space:].
629    
630    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
631        defined or documented. It seems to have been a typo for PCRE_STATIC, so
632        I have changed it.
633    
634    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
635        first named subpattern) and a construct such as (?&a) would reference the
636        first named subpattern whose name started with "a" (in other words, the
637        length check was missing). Both these problems are fixed. "Subpattern name
638        expected" is now given for (?&) (a zero-length name), and this patch also
639        makes it give the same error for \k'' (previously it complained that that
640        was a reference to a non-existent subpattern).
641    
642    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
643        this is right because (?- can be followed by option settings as well as by
644        digits. I have, however, made the messages clearer.
645    
646    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
647        than the number used in the conditional) now cause a compile-time error.
648        This is actually not compatible with Perl, which accepts such patterns, but
649        treats the conditional as always being FALSE (as PCRE used to), but it
650        seems to me that giving a diagnostic is better.
651    
652    8.  Change "alphameric" to the more common word "alphanumeric" in comments
653        and messages.
654    
655    9.  Fix two occurrences of "backslash" in comments that should have been
656        "backspace".
657    
658    10. Remove two redundant lines of code that can never be obeyed (their function
659        was moved elsewhere).
660    
661    11. The program that makes PCRE's Unicode character property table had a bug
662        which caused it to generate incorrect table entries for sequences of
663        characters that have the same character type, but are in different scripts.
664        It amalgamated them into a single range, with the script of the first of
665        them. In other words, some characters were in the wrong script. There were
666        thirteen such cases, affecting characters in the following ranges:
667    
668          U+002b0 - U+002c1
669          U+0060c - U+0060d
670          U+0061e - U+00612
671          U+0064b - U+0065e
672          U+0074d - U+0076d
673          U+01800 - U+01805
674          U+01d00 - U+01d77
675          U+01d9b - U+01dbf
676          U+0200b - U+0200f
677          U+030fc - U+030fe
678          U+03260 - U+0327f
679          U+0fb46 - U+0fbb1
680          U+10450 - U+1049d
681    
682    12. The -o option (show only the matching part of a line) for pcregrep was not
683        compatible with GNU grep in that, if there was more than one match in a
684        line, it showed only the first of them. It now behaves in the same way as
685        GNU grep.
686    
687    13. If the -o and -v options were combined for pcregrep, it printed a blank
688        line for every non-matching line. GNU grep prints nothing, and pcregrep now
689        does the same. The return code can be used to tell if there were any
690        non-matching lines.
691    
692    14. Added --file-offsets and --line-offsets to pcregrep.
693    
694    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
695        infinitely looping recursion. The bug was that positive lookaheads were not
696        being skipped when checking for a possible empty match (negative lookaheads
697        and both kinds of lookbehind were skipped).
698    
699    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
700        inclusion of <windows.h> to before rather than after the definition of
701        INVALID_FILE_ATTRIBUTES (patch from David Byron).
702    
703    17. Specifying a possessive quantifier with a specific limit for a Unicode
704        character property caused pcre_compile() to compile bad code, which led at
705        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
706        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
707        caused the error; without that there was no problem.
708    
709    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
710    
711    19. Added --enable-pcretest-libreadline.
712    
713    20. In pcrecpp.cc, the variable 'count' was incremented twice in
714        RE::GlobalReplace(). As a result, the number of replacements returned was
715        double what it should be. I removed one of the increments, but Craig sent a
716        later patch that removed the other one (the right fix) and added unit tests
717        that check the return values (which was not done before).
718    
719    21. Several CMake things:
720    
721        (1) Arranged that, when cmake is used on Unix, the libraries end up with
722            the names libpcre and libpcreposix, not just pcre and pcreposix.
723    
724        (2) The above change means that pcretest and pcregrep are now correctly
725            linked with the newly-built libraries, not previously installed ones.
726    
727        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
728    
729    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
730        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
731        UTF-8 newline character). The key issue is that the pattern starts .*;
732        this means that the match must be either at the beginning, or after a
733        newline. The bug was in the code for advancing after a failed match and
734        checking that the new position followed a newline. It was not taking
735        account of UTF-8 characters correctly.
736    
737    23. PCRE was behaving differently from Perl in the way it recognized POSIX
738        character classes. PCRE was not treating the sequence [:...:] as a
739        character class unless the ... were all letters. Perl, however, seems to
740        allow any characters between [: and :], though of course it rejects as
741        unknown any "names" that contain non-letters, because all the known class
742        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
743        for example, whereas PCRE did not - it did not recognize a POSIX character
744        class. This seemed a bit dangerous, so the code has been changed to be
745        closer to Perl. The behaviour is not identical to Perl, because PCRE will
746        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
747        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
748        Perl does, and where it didn't before.
749    
750    24. Rewrite so as to remove the single use of %n from pcregrep because in some
751        Windows environments %n is disabled by default.
752    
753    
754    Version 7.4 21-Sep-07
755    ---------------------
756    
757    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
758        means that a class such as [\s] counted as "explicit reference to CR or
759        LF". That isn't really right - the whole point of the change was to try to
760        help when there was an actual mention of one of the two characters. So now
761        the change happens only if \r or \n (or a literal CR or LF) character is
762        encountered.
763    
764    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
765        of both had grown to the point where there were only 3 bits left.
766        Fortunately, there was spare space in the data structure, and so I have
767        moved the internal flags into a new 16-bit field to free up more option
768        bits.
769    
770    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
771        but did not set the internal JCHANGED flag - either of these is enough to
772        control the way the "get" function works - but the PCRE_INFO_JCHANGED
773        facility is supposed to tell if (?J) was ever used, so now (?J) at the
774        start sets both bits.
775    
776    4.  Added options (at build time, compile time, exec time) to change \R from
777        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
778    
779    5.  doc/pcresyntax.html was missing from the distribution.
780    
781    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
782        compatibility, even though it is no longer used.
783    
784    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
785        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
786        windows.h file is present (where different names are used). [This was
787        reversed later after testing - see 16 below.]
788    
789    8.  Changed all #include <config.h> to #include "config.h". There were also
790        some further <pcre.h> cases that I changed to "pcre.h".
791    
792    9.  When pcregrep was used with the --colour option, it missed the line ending
793        sequence off the lines that it output.
794    
795    10. It was pointed out to me that arrays of string pointers cause lots of
796        relocations when a shared library is dynamically loaded. A technique of
797        using a single long string with a table of offsets can drastically reduce
798        these. I have refactored PCRE in four places to do this. The result is
799        dramatic:
800    
801          Originally:                          290
802          After changing UCP table:            187
803          After changing error message table:   43
804          After changing table of "verbs"       36
805          After changing table of Posix names   22
806    
807        Thanks to the folks working on Gregex for glib for this insight.
808    
809    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
810        unicode-properties was also set.
811    
812    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
813    
814    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
815        checked only for CRLF.
816    
817    14. Added casts to pcretest.c to avoid compiler warnings.
818    
819    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
820    
821    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
822        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
823        entirely. This removes changes made in 7 above.
824    
825    17. The CMake files have been updated, and there is now more information about
826        building with CMake in the NON-UNIX-USE document.
827    
828    
829    Version 7.3 28-Aug-07
830    ---------------------
831    
832     1. In the rejigging of the build system that eventually resulted in 7.1, the
833        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
834        brackets there is not right, since it causes compilers to look for an
835        installed pcre.h, not the version that is in the source that is being
836        compiled (which of course may be different). I have changed it back to:
837    
838          #include "pcre.h"
839    
840        I have a vague recollection that the change was concerned with compiling in
841        different directories, but in the new build system, that is taken care of
842        by the VPATH setting the Makefile.
843    
844     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
845        when the subject happened to end in the byte 0x85 (e.g. if the last
846        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
847        characters but of course it shouldn't be taken as a newline when it is part
848        of another character. The bug was that, for an unlimited repeat of . in
849        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
850        characters when looking for a newline.
851    
852     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
853    
854     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
855        in debug output.
856    
857     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
858        long printing in the pcrecpp unittest when running under MinGW.
859    
860     6. ESC_K was left out of the EBCDIC table.
861    
862     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
863        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
864        limit also applies to "virtual nesting" when a pattern is recursive, and in
865        this case 1000 isn't so big. I have been able to remove this limit at the
866        expense of backing off one optimization in certain circumstances. Normally,
867        when pcre_exec() would call its internal match() function recursively and
868        immediately return the result unconditionally, it uses a "tail recursion"
869        feature to save stack. However, when a subpattern that can match an empty
870        string has an unlimited repetition quantifier, it no longer makes this
871        optimization. That gives it a stack frame in which to save the data for
872        checking that an empty string has been matched. Previously this was taken
873        from the 1000-entry workspace that had been reserved. So now there is no
874        explicit limit, but more stack is used.
875    
876     8. Applied Daniel's patches to solve problems with the import/export magic
877        syntax that is required for Windows, and which was going wrong for the
878        pcreposix and pcrecpp parts of the library. These were overlooked when this
879        problem was solved for the main library.
880    
881     9. There were some crude static tests to avoid integer overflow when computing
882        the size of patterns that contain repeated groups with explicit upper
883        limits. As the maximum quantifier is 65535, the maximum group length was
884        set at 30,000 so that the product of these two numbers did not overflow a
885        32-bit integer. However, it turns out that people want to use groups that
886        are longer than 30,000 bytes (though not repeat them that many times).
887        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
888        made it possible to implement the integer overflow checks in a much more
889        dynamic way, which I have now done. The artificial limitation on group
890        length has been removed - we now have only the limit on the total length of
891        the compiled pattern, which depends on the LINK_SIZE setting.
892    
893    10. Fixed a bug in the documentation for get/copy named substring when
894        duplicate names are permitted. If none of the named substrings are set, the
895        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
896        empty string.
897    
898    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
899        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
900        because the ] is interpreted as the first data character and the
901        terminating ] is not found. PCRE has been made compatible with Perl in this
902        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
903        cause memory overwriting.
904    
905    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
906        string has been matched (to stop an infinite loop). It was not recognizing
907        a conditional subpattern that could match an empty string if that
908        subpattern was within another subpattern. For example, it looped when
909        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
910        condition was not nested. This bug has been fixed.
911    
912    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
913        past the start of the subject in the presence of bytes with the top bit
914        set, for example "\x8aBCD".
915    
916    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
917        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
918    
919    14. Optimized (?!) to (*FAIL).
920    
921    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
922        This restricts code points to be within the range 0 to 0x10FFFF, excluding
923        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
924        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
925        does: it's just the validity check that is more restrictive.
926    
927    16. Inserted checks for integer overflows during escape sequence (backslash)
928        processing, and also fixed erroneous offset values for syntax errors during
929        backslash processing.
930    
931    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
932        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
933    
934    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
935        caused an overrun.
936    
937    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
938        something other than just ASCII characters) inside a group that had an
939        unlimited repeat caused a loop at compile time (while checking to see
940        whether the group could match an empty string).
941    
942    20. Debugging a pattern containing \p or \P could cause a crash. For example,
943        [\P{Any}] did so. (Error in the code for printing property names.)
944    
945    21. An orphan \E inside a character class could cause a crash.
946    
947    22. A repeated capturing bracket such as (A)? could cause a wild memory
948        reference during compilation.
949    
950    23. There are several functions in pcre_compile() that scan along a compiled
951        expression for various reasons (e.g. to see if it's fixed length for look
952        behind). There were bugs in these functions when a repeated \p or \P was
953        present in the pattern. These operators have additional parameters compared
954        with \d, etc, and these were not being taken into account when moving along
955        the compiled data. Specifically:
956    
957        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
958            length.
959    
960        (b) An item such as \pL+ within a repeated group could cause crashes or
961            loops.
962    
963        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
964            "reference to non-existent subpattern" error.
965    
966        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
967    
968    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
969        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
970    
971    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
972    
973    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
974        character were causing crashes (broken optimization).
975    
976    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
977        \p or \P) caused a compile-time loop.
978    
979    28. More problems have arisen in unanchored patterns when CRLF is a valid line
980        break. For example, the unstudied pattern [\r\n]A does not match the string
981        "\r\nA" because change 7.0/46 below moves the current point on by two
982        characters after failing to match at the start. However, the pattern \nA
983        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
984        the same is true. There doesn't seem any very clean way out of this, but
985        what I have chosen to do makes the common cases work: PCRE now takes note
986        of whether there can be an explicit match for \r or \n anywhere in the
987        pattern, and if so, 7.0/46 no longer applies. As part of this change,
988        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
989        pattern has explicit CR or LF references.
990    
991    29. Added (*CR) etc for changing newline setting at start of pattern.
992    
993    
994    Version 7.2 19-Jun-07
995    ---------------------
996    
997     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
998        which is apparently normally available under Windows.
999    
1000     2. Re-jig the pcregrep tests with different newline settings in an attempt
1001        to make them independent of the local environment's newline setting.
1002    
1003     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1004    
1005     4. Some of the "internals" tests were previously cut out when the link size
1006        was not 2, because the output contained actual offsets. The recent new
1007        "Z" feature of pcretest means that these can be cut out, making the tests
1008        usable with all link sizes.
1009    
1010     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1011        stack recursion. This gives a massive performance boost under BSD, but just
1012        a small improvement under Linux. However, it saves one field in the frame
1013        in all cases.
1014    
1015     6. Added more features from the forthcoming Perl 5.10:
1016    
1017        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1018            recursion call. It refers to the nth most recently opened parentheses.
1019    
1020        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1021            to be opened parentheses.
1022    
1023        (c) Conditions that refer to capturing parentheses can be specified
1024            relatively, for example, (?(-2)... or (?(+3)...
1025    
1026        (d) \K resets the start of the current match so that everything before
1027            is not part of it.
1028    
1029        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1030    
1031        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1032            reference syntax.
1033    
1034        (g) (?| introduces a group in which the numbering of parentheses in each
1035            alternative starts with the same number.
1036    
1037        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1038    
1039     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1040        PCRE_INFO_JCHANGED.
1041    
1042     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1043        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1044        for detecting groups that can match an empty string.
1045    
1046     9. A pattern with a very large number of alternatives (more than several
1047        hundred) was running out of internal workspace during the pre-compile
1048        phase, where pcre_compile() figures out how much memory will be needed. A
1049        bit of new cunning has reduced the workspace needed for groups with
1050        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1051        workspace instead of running out of the 4096 that are available.
1052    
1053    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1054    
1055    11. Applied patch from Google to remove an optimization that didn't quite work.
1056        The report of the bug said:
1057    
1058          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1059          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1060          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1061    
1062    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1063        it matched the wrong number of bytes.
1064    
1065    
1066    Version 7.1 24-Apr-07
1067    ---------------------
1068    
1069     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1070        that is more "standard", making use of automake and other Autotools. There
1071        is some re-arrangement of the files and adjustment of comments consequent
1072        on this.
1073    
1074     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1075        for recursive directory scanning broke on some systems because the files
1076        are not scanned in any specific order and on different systems the order
1077        was different. A call to "sort" has been inserted into RunGrepTest for the
1078        approprate test as a short-term fix. In the longer term there may be an
1079        alternative.
1080    
1081     3. I had an email from Eric Raymond about problems translating some of PCRE's
1082        man pages to HTML (despite the fact that I distribute HTML pages, some
1083        people do their own conversions for various reasons). The problems
1084        concerned the use of low-level troff macros .br and .in. I have therefore
1085        removed all such uses from the man pages (some were redundant, some could
1086        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1087        HTML has been updated to handle .nf/.fi and to complain if it encounters
1088        .br or .in.
1089    
1090     4. Updated comments in configure.ac that get placed in config.h.in and also
1091        arranged for config.h to be included in the distribution, with the name
1092        config.h.generic, for the benefit of those who have to compile without
1093        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1094    
1095     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1096        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1097        makevp.bat for the current PCRE, using the additional files
1098        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1099    
1100     6. A Windows user reported a minor discrepancy with test 2, which turned out
1101        to be caused by a trailing space on an input line that had got lost in his
1102        copy. The trailing space was an accident, so I've just removed it.
1103    
1104     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1105        that is needed.
1106    
1107     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1108        as "const" (a) because they are and (b) because it helps the PHP
1109        maintainers who have recently made a script to detect big data structures
1110        in the php code that should be moved to the .rodata section. I remembered
1111        to update Builducptable as well, so it won't revert if ucptable.h is ever
1112        re-created.
1113    
1114     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1115        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1116        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1117        support is not required. This saves 1.5-2K of code, which is important in
1118        some applications.
1119    
1120        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1121        so as not to refer to the tables, even though these functions will never be
1122        called when UTF-8 support is disabled. Otherwise there are problems with a
1123        shared library.
1124    
1125    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1126    
1127        (a) It was defining its arguments as char * instead of void *.
1128    
1129        (b) It was assuming that all moves were upwards in memory; this was true
1130            a long time ago when I wrote it, but is no longer the case.
1131    
1132        The emulated memove() is provided for those environments that have neither
1133        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1134        is clearly not the case, as these two bugs were recently reported.
1135    
1136    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1137        and Detrail to create the HTML documentation, the .txt form of the man
1138        pages, and it removes trailing spaces from listed files. It also creates
1139        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1140        case, it wraps all the #defines with #ifndefs. This script should be run
1141        before "make dist".
1142    
1143    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1144        with Unicode property support.
1145    
1146        (a) For a maximizing quantifier, if the two different cases of the
1147            character were of different lengths in their UTF-8 codings (there are
1148            some cases like this - I found 11), and the matching function had to
1149            back up over a mixture of the two cases, it incorrectly assumed they
1150            were both the same length.
1151    
1152        (b) When PCRE was configured to use the heap rather than the stack for
1153            recursion during matching, it was not correctly preserving the data for
1154            the other case of a UTF-8 character when checking ahead for a match
1155            while processing a minimizing repeat. If the check also involved
1156            matching a wide character, but failed, corruption could cause an
1157            erroneous result when trying to check for a repeat of the original
1158            character.
1159    
1160    13. Some tidying changes to the testing mechanism:
1161    
1162        (a) The RunTest script now detects the internal link size and whether there
1163            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1164            values substituted by "configure". (The RunGrepTest script already did
1165            this for UTF-8.) The configure.ac script no longer substitutes the
1166            relevant variables.
1167    
1168        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1169            with length and offset values. This means that the output is different
1170            for different internal link sizes. Test 2 is skipped for link sizes
1171            other than 2 because of this, bypassing the problem. Unfortunately,
1172            there was also a test in test 3 (the locale tests) that used /B and
1173            failed for link sizes other than 2. Rather than cut the whole test out,
1174            I have added a new /Z option to pcretest that replaces the length and
1175            offset values with spaces. This is now used to make test 3 independent
1176            of link size. (Test 2 will be tidied up later.)
1177    
1178    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1179        segmentation fault instead of returning the appropriate error message.
1180    
1181    15. In multiline mode when the newline sequence was set to "any", the pattern
1182        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1183        This doesn't seem right; it now treats the CRLF combination as the line
1184        ending, and so does not match in that case. It's only a pattern such as ^$
1185        that would hit this one: something like ^ABC$ would have failed after \r
1186        and then tried again after \r\n.
1187    
1188    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1189        in an attempt to make files that differ only in their line terminators
1190        compare equal. This works on Linux.
1191    
1192    17. Under certain error circumstances pcregrep might try to free random memory
1193        as it exited. This is now fixed, thanks to valgrind.
1194    
1195    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1196        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1197        was because its rules for how to advance for /g after matching an empty
1198        string at the end of a line did not allow for this case. They now check for
1199        it specially.
1200    
1201    20. pcretest is supposed to handle patterns and data of any length, by
1202        extending its buffers when necessary. It was getting this wrong when the
1203        buffer for a data line had to be extended.
1204    
1205    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1206        CRLF as a newline sequence.
1207    
1208    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1209        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1210        I have nevertheless tidied it up.
1211    
1212    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1213    
1214    24. Added a man page for pcre-config.
1215    
1216    
1217    Version 7.0 19-Dec-06
1218    ---------------------
1219    
1220     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1221        moving to gcc 4.1.1.
1222    
1223     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1224        sys/time.h, which is documented as needed for this function. It doesn't
1225        seem to matter on Linux, but it showed up on some releases of OS X.
1226    
1227     3. It seems that there are systems where bytes whose values are greater than
1228        127 match isprint() in the "C" locale. The "C" locale should be the
1229        default when a C program starts up. In most systems, only ASCII printing
1230        characters match isprint(). This difference caused the output from pcretest
1231        to vary, making some of the tests fail. I have changed pcretest so that:
1232    
1233        (a) When it is outputting text in the compiled version of a pattern, bytes
1234            other than 32-126 are always shown as hex escapes.
1235    
1236        (b) When it is outputting text that is a matched part of a subject string,
1237            it does the same, unless a different locale has been set for the match
1238            (using the /L modifier). In this case, it uses isprint() to decide.
1239    
1240     4. Fixed a major bug that caused incorrect computation of the amount of memory
1241        required for a compiled pattern when options that changed within the
1242        pattern affected the logic of the preliminary scan that determines the
1243        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1244        that the computed length was too small. The symptoms of this bug were
1245        either the PCRE error "internal error: code overflow" from pcre_compile(),
1246        or a glibc crash with a message such as "pcretest: free(): invalid next
1247        size (fast)". Examples of patterns that provoked this bug (shown in
1248        pcretest format) are:
1249    
1250          /(?-x: )/x
1251          /(?x)(?-x: \s*#\s*)/
1252          /((?i)[\x{c0}])/8
1253          /(?i:[\x{c0}])/8
1254    
1255        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1256        is now done differently.
1257    
1258     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1259        wrapper classes; (b) implement a new function in the C++ scanner that is
1260        more efficient than the old way of doing things because it avoids levels of
1261        recursion in the regex matching; (c) add a paragraph to the documentation
1262        for the FullMatch() function.
1263    
1264     6. The escape sequence \n was being treated as whatever was defined as
1265        "newline". Not only was this contrary to the documentation, which states
1266        that \n is character 10 (hex 0A), but it also went horribly wrong when
1267        "newline" was defined as CRLF. This has been fixed.
1268    
1269     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1270        was being set to -1 for the "end of line" case (supposedly a value that no
1271        character can have). Though this value is never used (the check for end of
1272        line is "zero bytes in current character"), it caused compiler complaints.
1273        I've changed it to 0xffffffff.
1274    
1275     8. In pcre_version.c, the version string was being built by a sequence of
1276        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1277        string (as it is for production releases) called a macro with an empty
1278        argument. The C standard says the result of this is undefined. The gcc
1279        compiler treats it as an empty string (which was what was wanted) but it is
1280        reported that Visual C gives an error. The source has been hacked around to
1281        avoid this problem.
1282    
1283     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1284        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1285        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1286        of them did).
1287    
1288    10. Originally, pcretest opened its input and output without "b"; then I was
1289        told that "b" was needed in some environments, so it was added for release
1290        5.0 to both the input and output. (It makes no difference on Unix-like
1291        systems.) Later I was told that it is wrong for the input on Windows. I've
1292        now abstracted the modes into two macros, to make it easier to fiddle with
1293        them, and removed "b" from the input mode under Windows.
1294    
1295    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1296    
1297    12. Added -help and --help to pcretest as an official way of being reminded
1298        of the options.
1299    
1300    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1301        and pcrecpp.cc because they annoy compilers at high warning levels.
1302    
1303    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1304    
1305    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1306        scripts are not C programs :-) and which was not noticed because it works
1307        on Linux.
1308    
1309    16. pcretest is supposed to handle any length of pattern and data line (as one
1310        line or as a continued sequence of lines) by extending its input buffer if
1311        necessary. This feature was broken for very long pattern lines, leading to
1312        a string of junk being passed to pcre_compile() if the pattern was longer
1313        than about 50K.
1314    
1315    17. I have done a major re-factoring of the way pcre_compile() computes the
1316        amount of memory needed for a compiled pattern. Previously, there was code
1317        that made a preliminary scan of the pattern in order to do this. That was
1318        OK when PCRE was new, but as the facilities have expanded, it has become
1319        harder and harder to keep it in step with the real compile phase, and there
1320        have been a number of bugs (see for example, 4 above). I have now found a
1321        cunning way of running the real compile function in a "fake" mode that
1322        enables it to compute how much memory it would need, while actually only
1323        ever using a few hundred bytes of working memory and without too many
1324        tests of the mode. This should make future maintenance and development
1325        easier. A side effect of this work is that the limit of 200 on the nesting
1326        depth of parentheses has been removed (though this was never a serious
1327        limitation, I suspect). However, there is a downside: pcre_compile() now
1328        runs more slowly than before (30% or more, depending on the pattern). I
1329        hope this isn't a big issue. There is no effect on runtime performance.
1330    
1331    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1332        newline (only possible for the last line of a file) and it was a
1333        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1334    
1335    19. Added additional timing features to pcretest. (1) The -tm option now times
1336        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1337        separate command line item, by a number that specifies the number of
1338        repeats to use when timing. The default is 50000; this gives better
1339        precision, but takes uncomfortably long for very large patterns.
1340    
1341    20. Extended pcre_study() to be more clever in cases where a branch of a
1342        subpattern has no definite first character. For example, (a*|b*)[cd] would
1343        previously give no result from pcre_study(). Now it recognizes that the
1344        first character must be a, b, c, or d.
1345    
1346    21. There was an incorrect error "recursive call could loop indefinitely" if
1347        a subpattern (or the entire pattern) that was being tested for matching an
1348        empty string contained only one non-empty item after a nested subpattern.
1349        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1350        incorrectly, because the \d was being skipped in the check.
1351    
1352    22. The pcretest program now has a new pattern option /B and a command line
1353        option -b, which is equivalent to adding /B to every pattern. This causes
1354        it to show the compiled bytecode, without the additional information that
1355        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1356        is the same as /B/I).
1357    
1358    23. A new optimization is now able automatically to treat some sequences such
1359        as a*b as a*+b. More specifically, if something simple (such as a character
1360        or a simple class like \d) has an unlimited quantifier, and is followed by
1361        something that cannot possibly match the quantified thing, the quantifier
1362        is automatically "possessified".
1363    
1364    24. A recursive reference to a subpattern whose number was greater than 39
1365        went wrong under certain circumstances in UTF-8 mode. This bug could also
1366        have affected the operation of pcre_study().
1367    
1368    25. Realized that a little bit of performance could be had by replacing
1369        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1370    
1371    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1372    
1373    27. Possessive quantifiers such as a++ were previously implemented by turning
1374        them into atomic groups such as ($>a+). Now they have their own opcodes,
1375        which improves performance. This includes the automatically created ones
1376        from 23 above.
1377    
1378    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1379        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1380        the first matched character to be a colon. This applied both to named and
1381        numbered groups.
1382    
1383    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1384    
1385    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1386        building PCRE on Windows easier, so I have included it in the distribution.
1387    
1388    31. There is now a check in pcretest against a ridiculously large number being
1389        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1390        loop, the loop is abandoned.
1391    
1392    32. Forward references to subpatterns in conditions such as (?(2)...) where
1393        subpattern 2 is defined later cause pcre_compile() to search forwards in
1394        the pattern for the relevant set of parentheses. This search went wrong
1395        when there were unescaped parentheses in a character class, parentheses
1396        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1397    
1398    33. "Subroutine" calls and backreferences were previously restricted to
1399        referencing subpatterns earlier in the regex. This restriction has now
1400        been removed.
1401    
1402    34. Added a number of extra features that are going to be in Perl 5.10. On the
1403        whole, these are just syntactic alternatives for features that PCRE had
1404        previously implemented using the Python syntax or my own invention. The
1405        other formats are all retained for compatibility.
1406    
1407        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1408            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1409            also .NET compatible.
1410    
1411        (b) A recursion or subroutine call to a named group can now be defined as
1412            (?&name) as well as (?P>name).
1413    
1414        (c) A backreference to a named group can now be defined as \k<name> or
1415            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1416            5.10, are also .NET compatible.
1417    
1418        (d) A conditional reference to a named group can now use the syntax
1419            (?(<name>) or (?('name') as well as (?(name).
1420    
1421        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1422            groups (named and numbered) that are never evaluated inline, but can be
1423            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1424            is always false. There may be only one alternative in such a group.
1425    
1426        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1427            as the simple (?(R). The condition is true only if the most recent
1428            recursion is that of the given number or name. It does not search out
1429            through the entire recursion stack.
1430    
1431        (g) The escape \gN or \g{N} has been added, where N is a positive or
1432            negative number, specifying an absolute or relative reference.
1433    
1434    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1435        some "unreachable code" warnings.
1436    
1437    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1438        things, this adds five new scripts.
1439    
1440    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1441        There were also incompatibilities regarding the handling of \Q..\E inside
1442        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1443        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1444    
1445    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1446        matches an empty string, and forcibly breaks the loop. There were bugs in
1447        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1448        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1449        separate and independent bugs (that affected different cases) have been
1450        fixed.
1451    
1452    39. Refactored the code to abolish the use of different opcodes for small
1453        capturing bracket numbers. This is a tidy that I avoided doing when I
1454        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1455        The new approach is not only tidier, it makes it possible to reduce the
1456        memory needed to fix the previous bug (38).
1457    
1458    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1459        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1460        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1461        mode.
1462    
1463    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1464        report.
1465    
1466    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1467        copy construction and assignment in the C++ wrapper.
1468    
1469    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1470        couple of bugs that could have given wrong results in the "--newline=crlf"
1471        case.
1472    
1473    44. Added a number of casts and did some reorganization of signed/unsigned int
1474        variables following suggestions from Dair Grant. Also renamed the variable
1475        "this" as "item" because it is a C++ keyword.
1476    
1477    45. Arranged for dftables to add
1478    
1479          #include "pcre_internal.h"
1480    
1481        to pcre_chartables.c because without it, gcc 4.x may remove the array
1482        definition from the final binary if PCRE is built into a static library and
1483        dead code stripping is activated.
1484    
1485    46. For an unanchored pattern, if a match attempt fails at the start of a
1486        newline sequence, and the newline setting is CRLF or ANY, and the next two
1487        characters are CRLF, advance by two characters instead of one.
1488    
1489    
1490    Version 6.7 04-Jul-06
1491    ---------------------
1492    
1493     1. In order to handle tests when input lines are enormously long, pcretest has
1494        been re-factored so that it automatically extends its buffers when
1495        necessary. The code is crude, but this _is_ just a test program. The
1496        default size has been increased from 32K to 50K.
1497    
1498     2. The code in pcre_study() was using the value of the re argument before
1499        testing it for NULL. (Of course, in any sensible call of the function, it
1500        won't be NULL.)
1501    
1502     3. The memmove() emulation function in pcre_internal.h, which is used on
1503        systems that lack both memmove() and bcopy() - that is, hardly ever -
1504        was missing a "static" storage class specifier.
1505    
1506     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1507        containing an extended class (one that cannot be represented by a bitmap
1508        because it contains high-valued characters or Unicode property items, e.g.
1509        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1510        pattern, but PCRE should not loop if you do not (it no longer does).
1511        [Detail: two cases were found: (a) a repeated subpattern containing an
1512        extended class; (b) a recursive reference to a subpattern that followed a
1513        previous extended class. It wasn't skipping over the extended class
1514        correctly when UTF-8 mode was not set.]
1515    
1516     5. A negated single-character class was not being recognized as fixed-length
1517        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1518        compile error "lookbehind assertion is not fixed length".
1519    
1520     6. The RunPerlTest auxiliary script was showing an unexpected difference
1521        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1522        write a Perl script that can interpret lines of an input file either as
1523        byte characters or as UTF-8, which is what "perltest" was being required to
1524        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1525        can't do is switch easily at run time between having the "use utf8;" pragma
1526        or not. In the end, I fudged it by using the RunPerlTest script to insert
1527        "use utf8;" explicitly for the UTF-8 tests.
1528    
1529     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1530        the end of the subject string, contrary to the documentation and to what
1531        Perl does. This was true of both matching functions. Now it matches only at
1532        the start of the subject and immediately after *internal* newlines.
1533    
1534     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1535        a pointer to an int instead of a pointer to an unsigned long int. This
1536        caused problems on 64-bit systems.
1537    
1538     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1539        instance of the 'standard' template library not being so standard".
1540    
1541    10. There was no check on the number of named subpatterns nor the maximum
1542        length of a subpattern name. The product of these values is used to compute
1543        the size of the memory block for a compiled pattern. By supplying a very
1544        long subpattern name and a large number of named subpatterns, the size
1545        computation could be caused to overflow. This is now prevented by limiting
1546        the length of names to 32 characters, and the number of named subpatterns
1547        to 10,000.
1548    
1549    11. Subpatterns that are repeated with specific counts have to be replicated in
1550        the compiled pattern. The size of memory for this was computed from the
1551        length of the subpattern and the repeat count. The latter is limited to
1552        65535, but there was no limit on the former, meaning that integer overflow
1553        could in principle occur. The compiled length of a repeated subpattern is
1554        now limited to 30,000 bytes in order to prevent this.
1555    
1556    12. Added the optional facility to have named substrings with the same name.
1557    
1558    13. Added the ability to use a named substring as a condition, using the
1559        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1560        are numbers (not recommended). Forward references are permitted.
1561    
1562    14. Added forward references in named backreferences (if you see what I mean).
1563    
1564    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1565        pattern could run off the end of the subject. For example, the pattern
1566        "(?s)(.{1,5})"8 did this with the subject "ab".
1567    
1568    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1569        PCRE_CASELESS was set when matching characters that were quantified with ?
1570        or *.
1571    
1572    17. A character class other than a single negated character that had a minimum
1573        but no maximum quantifier - for example [ab]{6,} - was not handled
1574        correctly by pce_dfa_exec(). It would match only one character.
1575    
1576    18. A valid (though odd) pattern that looked like a POSIX character
1577        class but used an invalid character after [ (for example [[,abc,]]) caused
1578        pcre_compile() to give the error "Failed: internal error: code overflow" or
1579        in some cases to crash with a glibc free() error. This could even happen if
1580        the pattern terminated after [[ but there just happened to be a sequence of
1581        letters, a binary zero, and a closing ] in the memory that followed.
1582    
1583    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1584        over the years. Originally (before any Unicode support), just the bottom 8
1585        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1586        output from "man perlunicode" includes this:
1587    
1588          The regular expression compiler produces polymorphic opcodes.  That
1589          is, the pattern adapts to the data and automatically switches to
1590          the Unicode character scheme when presented with Unicode data--or
1591          instead uses a traditional byte scheme when presented with byte
1592          data.
1593    
1594        Sadly, a wide octal escape does not cause a switch, and in a string with
1595        no other multibyte characters, these octal escapes are treated as before.
1596        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1597        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1598        Unicode string.
1599    
1600        I have not perpetrated such confusion in PCRE. Up till now, it took just
1601        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1602        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1603        translate to the appropriate multibyte character.
1604    
1605    29. Applied some refactoring to reduce the number of warnings from Microsoft
1606        and Borland compilers. This has included removing the fudge introduced
1607        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1608        a warning about an unused variable.
1609    
1610    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1611        characters since release 4.0, because Perl (from release 5.004) does not.
1612        [Or at least, is documented not to: some releases seem to be in conflict
1613        with the documentation.] However, when a pattern was studied with
1614        pcre_study() and all its branches started with \s, PCRE still included VT
1615        as a possible starting character. Of course, this did no harm; it just
1616        caused an unnecessary match attempt.
1617    
1618    22. Removed a now-redundant internal flag bit that recorded the fact that case
1619        dependency changed within the pattern. This was once needed for "required
1620        byte" processing, but is no longer used. This recovers a now-scarce options
1621        bit. Also moved the least significant internal flag bit to the most-
1622        significant bit of the word, which was not previously used (hangover from
1623        the days when it was an int rather than a uint) to free up another bit for
1624        the future.
1625    
1626    23. Added support for CRLF line endings as well as CR and LF. As well as the
1627        default being selectable at build time, it can now be changed at runtime
1628        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1629        specify that it is scanning data with non-default line endings.
1630    
1631    24. Changed the definition of CXXLINK to make it agree with the definition of
1632        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1633    
1634    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1635        recursions. This makes a big different to stack usage for some patterns.
1636    
1637    26. If a subpattern containing a named recursion or subroutine reference such
1638        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1639        the space required for the compiled pattern went wrong and gave too small a
1640        value. Depending on the environment, this could lead to "Failed: internal
1641        error: code overflow at offset 49" or "glibc detected double free or
1642        corruption" errors.
1643    
1644    27. Applied patches from Google (a) to support the new newline modes and (b) to
1645        advance over multibyte UTF-8 characters in GlobalReplace.
1646    
1647    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1648        difference for some implementation of PCRE in some Windows version.
1649    
1650    29. Added some extra testing facilities to pcretest:
1651    
1652        \q<number>   in a data line sets the "match limit" value
1653        \Q<number>   in a data line sets the "match recursion limt" value
1654        -S <number>  sets the stack size, where <number> is in megabytes
1655    
1656        The -S option isn't available for Windows.
1657    
1658    
1659    Version 6.6 06-Feb-06
1660    ---------------------
1661    
1662     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1663        in pcreposix.h. I have copied the definition from pcre.h.
1664    
1665     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1666        because pcre.h is no longer a built file.
1667    
1668     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1669        not normally included in the compiled code.
1670    
1671    
1672    Version 6.5 01-Feb-06
1673    ---------------------
1674    
1675     1. When using the partial match feature with pcre_dfa_exec(), it was not
1676        anchoring the second and subsequent partial matches at the new starting
1677        point. This could lead to incorrect results. For example, with the pattern
1678        /1234/, partially matching against "123" and then "a4" gave a match.
1679    
1680     2. Changes to pcregrep:
1681    
1682        (a) All non-match returns from pcre_exec() were being treated as failures
1683            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1684            error message is output. Some extra information is given for the
1685            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1686            probably the only errors that are likely to be caused by users (by
1687            specifying a regex that has nested indefinite repeats, for instance).
1688            If there are more than 20 of these errors, pcregrep is abandoned.
1689    
1690        (b) A binary zero was treated as data while matching, but terminated the
1691            output line if it was written out. This has been fixed: binary zeroes
1692            are now no different to any other data bytes.
1693    
1694        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1695            used to set a locale for matching. The --locale=xxxx long option has
1696            been added (no short equivalent) to specify a locale explicitly on the
1697            pcregrep command, overriding the environment variables.
1698    
1699        (d) When -B was used with -n, some line numbers in the output were one less
1700            than they should have been.
1701    
1702        (e) Added the -o (--only-matching) option.
1703    
1704        (f) If -A or -C was used with -c (count only), some lines of context were
1705            accidentally printed for the final match.
1706    
1707        (g) Added the -H (--with-filename) option.
1708    
1709        (h) The combination of options -rh failed to suppress file names for files
1710            that were found from directory arguments.
1711    
1712        (i) Added the -D (--devices) and -d (--directories) options.
1713    
1714        (j) Added the -F (--fixed-strings) option.
1715    
1716        (k) Allow "-" to be used as a file name for -f as well as for a data file.
1717    
1718        (l) Added the --colo(u)r option.
1719    
1720        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1721            is not present by default.
1722    
1723     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1724        items such as (?R) or (?1), when the recursion could match a number of
1725        alternatives. If it matched one of the alternatives, but subsequently,
1726        outside the recursion, there was a failure, the code tried to back up into
1727        the recursion. However, because of the way PCRE is implemented, this is not
1728        possible, and the result was an incorrect result from the match.
1729    
1730        In order to prevent this happening, the specification of recursion has
1731        been changed so that all such subpatterns are automatically treated as
1732        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1733    
1734     4. I had overlooked the fact that, in some locales, there are characters for
1735        which isalpha() is true but neither isupper() nor islower() are true. In
1736        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1737        and ordfeminine) are like this. This affected the treatment of \w and \W
1738        when they appeared in character classes, but not when they appeared outside
1739        a character class. The bit map for "word" characters is now created
1740        separately from the results of isalnum() instead of just taking it from the
1741        upper, lower, and digit maps. (Plus the underscore character, of course.)
1742    
1743     5. The above bug also affected the handling of POSIX character classes such as
1744        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1745        permanent tables. Instead, the bit maps for such a class were previously
1746        created as the appropriate unions of the upper, lower, and digit bitmaps.
1747        Now they are created by subtraction from the [[:word:]] class, which has
1748        its own bitmap.
1749    
1750     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1751        It is created by subtracting the vertical space characters (\x09, \x0a,
1752        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1753        subtraction was done in the overall bitmap for a character class, meaning
1754        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1755        be recognized. This bug has been fixed.
1756    
1757     7. Patches from the folks at Google:
1758    
1759          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1760          real life, but is still worth protecting against".
1761    
1762          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1763          regular expressions".
1764    
1765          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1766          have it.
1767    
1768          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1769          "configure" and the latter not, in order to fix a problem somebody had
1770          with compiling the Arg class on HP-UX.
1771    
1772          (e) Improve the error-handling of the C++ wrapper a little bit.
1773    
1774          (f) New tests for checking recursion limiting.
1775    
1776     8. The pcre_memmove() function, which is used only if the environment does not
1777        have a standard memmove() function (and is therefore rarely compiled),
1778        contained two bugs: (a) use of int instead of size_t, and (b) it was not
1779        returning a result (though PCRE never actually uses the result).
1780    
1781     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1782        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1783        returned instead of calling malloc() with an overflowing number that would
1784        most likely cause subsequent chaos.
1785    
1786    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1787    
1788    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1789        with this option is matched, the nmatch and pmatch options of regexec() are
1790        ignored.
1791    
1792    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1793        provided in case anyone wants to the the POSIX interface with UTF-8
1794        strings.
1795    
1796    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1797        C++ linking (needed for some HP-UX environments).
1798    
1799    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1800        (unused parameter) and in the pcre_printint() function (omitted "default"
1801        switch label when the default is to do nothing).
1802    
1803    15. Added some code to make it possible, when PCRE is compiled as a C++
1804        library, to replace subject pointers for pcre_exec() with a smart pointer
1805        class, thus making it possible to process discontinuous strings.
1806    
1807    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1808        much the same function. They were added by different people who were trying
1809        to make PCRE easy to compile on non-Unix systems. It has been suggested
1810        that PCRE_EXPORT be abolished now that there is more automatic apparatus
1811        for compiling on Windows systems. I have therefore replaced it with
1812        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1813        defaults to "extern" for C or "extern C" for C++, which works fine on
1814        Unix-like systems. It is now possible to override the value of PCRE_DATA_
1815        SCOPE with something explicit in config.h. In addition:
1816    
1817        (a) pcreposix.h still had just "extern" instead of either of these macros;
1818            I have replaced it with PCRE_DATA_SCOPE.
1819    
1820        (b) Functions such as _pcre_xclass(), which are internal to the library,
1821            but external in the C sense, all had PCRE_EXPORT in their definitions.
1822            This is apparently wrong for the Windows case, so I have removed it.
1823            (It makes no difference on Unix-like systems.)
1824    
1825    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1826        of recursive calls to match(). This is different to MATCH_LIMIT because
1827        that limits the total number of calls to match(), not all of which increase
1828        the depth of recursion. Limiting the recursion depth limits the amount of
1829        stack (or heap if NO_RECURSE is set) that is used. The default can be set
1830        when PCRE is compiled, and changed at run time. A patch from Google adds
1831        this functionality to the C++ interface.
1832    
1833    18. Changes to the handling of Unicode character properties:
1834    
1835        (a) Updated the table to Unicode 4.1.0.
1836    
1837        (b) Recognize characters that are not in the table as "Cn" (undefined).
1838    
1839        (c) I revised the way the table is implemented to a much improved format
1840            which includes recognition of ranges. It now supports the ranges that
1841            are defined in UnicodeData.txt, and it also amalgamates other
1842            characters into ranges. This has reduced the number of entries in the
1843            table from around 16,000 to around 3,000, thus reducing its size
1844            considerably. I realized I did not need to use a tree structure after
1845            all - a binary chop search is just as efficient. Having reduced the
1846            number of entries, I extended their size from 6 bytes to 8 bytes to
1847            allow for more data.
1848    
1849        (d) Added support for Unicode script names via properties such as \p{Han}.
1850    
1851    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1852        matching that character.
1853    
1854    20. When matching a repeated Unicode property with a minimum greater than zero,
1855        (for example \pL{2,}), PCRE could look past the end of the subject if it
1856        reached it while seeking the minimum number of characters. This could
1857        happen only if some of the characters were more than one byte long, because
1858        there is a check for at least the minimum number of bytes.
1859    
1860    21. Refactored the implementation of \p and \P so as to be more general, to
1861        allow for more different types of property in future. This has changed the
1862        compiled form incompatibly. Anybody with saved compiled patterns that use
1863        \p or \P will have to recompile them.
1864    
1865    22. Added "Any" and "L&" to the supported property types.
1866    
1867    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1868        but give a compile time error if the value is greater than 0xff.
1869    
1870    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1871        accidentally not being installed or uninstalled.
1872    
1873    25. The pcre.h file was built from pcre.h.in, but the only changes that were
1874        made were to insert the current release number. This seemed silly, because
1875        it made things harder for people building PCRE on systems that don't run
1876        "configure". I have turned pcre.h into a distributed file, no longer built
1877        by "configure", with the version identification directly included. There is
1878        no longer a pcre.h.in file.
1879    
1880        However, this change necessitated a change to the pcre-config script as
1881        well. It is built from pcre-config.in, and one of the substitutions was the
1882        release number. I have updated configure.ac so that ./configure now finds
1883        the release number by grepping pcre.h.
1884    
1885    26. Added the ability to run the tests under valgrind.
1886    
1887    
1888  Version 6.4 05-Sep-05  Version 6.4 05-Sep-05
1889  ---------------------  ---------------------
1890    

Legend:
Removed from v.85  
changed lines
  Added in v.482

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12