ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1553 by ph10, Tue Apr 28 11:36:24 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4    Version 8.37 28-April-2015
5    --------------------------
7    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8        for those parentheses to be closed with whatever has been captured so far.
9        However, it was failing to mark any other groups between the hightest
10        capture so far and the currrent group as "unset". Thus, the ovector for
11        those groups contained whatever was previously there. An example is the
12        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
14    2.  If an assertion condition was quantified with a minimum of zero (an odd
15        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
17    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
18        unrecognized modifier, a crash could occur.
20    4.  An attempt to do global matching in pcretest with a zero-length ovector
21        caused a crash.
23    5.  Fixed a memory leak during matching that could occur for a subpattern
24        subroutine call (recursive or otherwise) if the number of captured groups
25        that had to be saved was greater than ten.
27    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
28        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
31    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
32        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
34    8. When a pattern is compiled, it remembers the highest back reference so that
35       when matching, if the ovector is too small, extra memory can be obtained to
36       use instead. A conditional subpattern whose condition is a check on a
37       capture having happened, such as, for example in the pattern
38       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39       setting the highest backreference number. This mattered only if pcre_exec()
40       was called with an ovector that was too small to hold the capture, and there
41       was no other kind of back reference (a situation which is probably quite
42       rare). The effect of the bug was that the condition was always treated as
43       FALSE when the capture could not be consulted, leading to a incorrect
44       behaviour by pcre_exec(). This bug has been fixed.
46    9. A reference to a duplicated named group (either a back reference or a test
47       for being set in a conditional) that occurred in a part of the pattern where
48       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49       to be incorrectly calculated, leading to overwriting.
51    10. A mutually recursive set of back references such as (\2)(\1) caused a
52        segfault at study time (while trying to find the minimum matching length).
53        The infinite loop is now broken (with the minimum length unset, that is,
54        zero).
56    11. If an assertion that was used as a condition was quantified with a minimum
57        of zero, matching went wrong. In particular, if the whole group had
58        unlimited repetition and could match an empty string, a segfault was
59        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60        assertions to be quantified, but not if they are being used as conditions,
61        so the above pattern is faulted by Perl. PCRE has now been changed so that
62        it also rejects such patterns.
64    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65        failed to allow the zero-repeat case if pcre2_exec() was called with an
66        ovector too small to capture the group.
68    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69        Red Hat Product Security:
71        (a) A crash if /K and /F were both set with the option to save the compiled
72        pattern.
74        (b) Another crash if the option to print captured substrings in a callout
75        was combined with setting a null ovector, for example \O\C+ as a subject
76        string.
78    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79        forward reference repeated a large (but limited) number of times within a
80        repeated outer group that has a zero minimum quantifier, caused incorrect
81        code to be compiled, leading to the error "internal error:
82        previously-checked referenced subpattern not found" when an incorrect
83        memory address was read. This bug was reported as "heap overflow",
84        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
85        CVE-2015-2325.
87    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
88        call within a group that also contained a recursive back reference caused
89        incorrect code to be compiled. This bug was reported as "heap overflow",
90        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
91        number CVE-2015-2326.
93    24. Computing the size of the JIT read-only data in advance has been a source
94        of various issues, and new ones are still appear unfortunately. To fix
95        existing and future issues, size computation is eliminated from the code,
96        and replaced by on-demand memory allocation.
98    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
99        adjacent to the end of the range, and the range contained characters with
100        more than one other case, caused incorrect behaviour when compiled in UTF
101        mode. In that example, the range a-j was left out of the class.
103    26. Fix JIT compilation of conditional blocks, which assertion
104        is converted to (*FAIL). E.g: /(?(?!))/.
106    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
107        discovered by the LLVM fuzzer.
109    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
110        when this assertion was used as a condition, for example (?(?!)a|b). In
111        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
112        error about an unsupported item.
114    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
115        possessification code could take exponential time to complete. A recursion
116        depth limit of 1000 has been imposed to limit the resources used by this
117        optimization.
119    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
120        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
121        because \S ensures they are all in the class. The code for doing this was
122        interacting badly with the code for computing the amount of space needed to
123        compile the pattern, leading to a buffer overflow. This bug was discovered
124        by the LLVM fuzzer.
126    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
127        other kinds of group caused stack overflow at compile time. This bug was
128        discovered by the LLVM fuzzer.
130    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
131        between a subroutine call and its quantifier was incorrectly compiled,
132        leading to buffer overflow or other errors. This bug was discovered by the
133        LLVM fuzzer.
135    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
136        assertion after (?(. The code was failing to check the character after
137        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
138        was discovered by the LLVM fuzzer.
140    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
141        a fixed maximum following a group that contains a subroutine reference was
142        incorrectly compiled and could trigger buffer overflow. This bug was
143        discovered by the LLVM fuzzer.
145    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
146        caused a stack overflow instead of the diagnosis of a non-fixed length
147        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
149    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
150        (e.g. /(?<=\Ka)/) could make pcregrep loop.
152    37. There was a similar problem to 36 in pcretest for global matches.
154    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
155        and a subsequent item in the pattern caused a non-match, backtracking over
156        the repeated \X did not stop, but carried on past the start of the subject,
157        causing reference to random memory and/or a segfault. There were also some
158        other cases where backtracking after \C could crash. This set of bugs was
159        discovered by the LLVM fuzzer.
161    39. The function for finding the minimum length of a matching string could take
162        a very long time if mutual recursion was present many times in a pattern,
163        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
164        method has been implemented. This infelicity was discovered by the LLVM
165        fuzzer.
167    40. Static linking against the PCRE library using the pkg-config module was
168        failing on missing pthread symbols.
171    Version 8.36 26-September-2014
172    ------------------------------
174    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
175        -Wmissing-field-initializers and -Wunused-parameter.
177    2.  The tests for quantifiers being too big (greater than 65535) were being
178        applied after reading the number, and stupidly assuming that integer
179        overflow would give a negative number. The tests are now applied as the
180        numbers are read.
182    3.  Tidy code in pcre_exec.c where two branches that used to be different are
183        now the same.
185    4.  The JIT compiler did not generate match limit checks for certain
186        bracketed expressions with quantifiers. This may lead to exponential
187        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
188        issue should be resolved now.
190    5.  Fixed an issue, which occures when nested alternatives are optimized
191        with table jumps.
193    6.  Inserted two casts and changed some ints to size_t in the light of some
194        reported 64-bit compiler warnings (Bugzilla 1477).
196    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
197        an empty string, which sometimes were behaving incorrectly in the
198        interpreter (though correctly in the JIT matcher). This pcretest input is
199        an example:
201          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
204        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
205        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
206        for an empty string was breaking the inner loop and carrying on at a lower
207        level, when possessive repeated groups should always return to a higher
208        level as they have no backtrack points in them. The empty string test now
209        occurs at the outer level.
211    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
212        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
214    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
215        doesn't).
217    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
218        to hold a pattern to be incorrectly computed (too small) when there were
219        named back references to duplicated names. This could cause "internal
220        error: code overflow" or "double free or corruption" or other memory
221        handling errors.
223    11. When named subpatterns had the same prefixes, back references could be
224        confused. For example, in this pattern:
226          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
228        the reference to 'Name' was incorrectly treated as a reference to a
229        duplicate name.
231    12. A pattern such as /^s?c/mi8 where the optional character has more than
232        one "other case" was incorrectly compiled such that it would only try to
233        match starting at "c".
235    13. When a pattern starting with \s was studied, VT was not included in the
236        list of possible starting characters; this should have been part of the
237        8.34/18 patch.
239    14. If a character class started [\Qx]... where x is any character, the class
240        was incorrectly terminated at the ].
242    15. If a pattern that started with a caseless match for a character with more
243        than one "other case" was studied, PCRE did not set up the starting code
244        unit bit map for the list of possible characters. Now it does. This is an
245        optimization improvement, not a bug fix.
247    16. The Unicode data tables have been updated to Unicode 7.0.0.
249    17. Fixed a number of memory leaks in pcregrep.
251    18. Avoid a compiler warning (from some compilers) for a function call with
252        a cast that removes "const" from an lvalue by using an intermediate
253        variable (to which the compiler does not object).
255    19. Incorrect code was compiled if a group that contained an internal recursive
256        back reference was optional (had quantifier with a minimum of zero). This
257        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
258        caused segmentation faults because of stack overflows at compile time.
260    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
261        group that is quantified with an indefinite repeat, caused a compile-time
262        loop which used up all the system stack and provoked a segmentation fault.
263        This was not the same bug as 19 above.
265    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
266        Patch by Mike Frysinger.
269    Version 8.35 04-April-2014
270    --------------------------
272    1.  A new flag is set, when property checks are present in an XCLASS.
273        When this flag is not set, PCRE can perform certain optimizations
274        such as studying these XCLASS-es.
276    2.  The auto-possessification of character sets were improved: a normal
277        and an extended character set can be compared now. Furthermore
278        the JIT compiler optimizes more character set checks.
280    3.  Got rid of some compiler warnings for potentially uninitialized variables
281        that show up only when compiled with -O2.
283    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
284        of a match later then the end of the match. The pcretest program was not
285        handling the case sensibly - it was outputting from the start to the next
286        binary zero. It now reports this situation in a message, and outputs the
287        text from the end to the start.
289    5.  Fast forward search is improved in JIT. Instead of the first three
290        characters, any three characters with fixed position can be searched.
291        Search order: first, last, middle.
293    6.  Improve character range checks in JIT. Characters are read by an inprecise
294        function now, which returns with an unknown value if the character code is
295        above a certain threshold (e.g: 256). The only limitation is that the value
296        must be bigger than the threshold as well. This function is useful when
297        the characters above the threshold are handled in the same way.
299    7.  The macros whose names start with RAWUCHAR are placeholders for a future
300        mode in which only the bottom 21 bits of 32-bit data items are used. To
301        make this more memorable for those maintaining the code, the names have
302        been changed to start with UCHAR21, and an extensive comment has been added
303        to their definition.
305    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
306        to the export list in Makefile.am (they were accidentally omitted from the
307        8.34 tarball).
309    9.  The informational output from pcretest used the phrase "starting byte set"
310        which is inappropriate for the 16-bit and 32-bit libraries. As the output
311        for "first char" and "need char" really means "non-UTF-char", I've changed
312        "byte" to "char", and slightly reworded the output. The documentation about
313        these values has also been (I hope) clarified.
315    10. Another JIT related optimization: use table jumps for selecting the correct
316        backtracking path, when more than four alternatives are present inside a
317        bracket.
319    11. Empty match is not possible, when the minimum length is greater than zero,
320        and there is no \K in the pattern. JIT should avoid empty match checks in
321        such cases.
323    12. In a caseless character class with UCP support, when a character with more
324        than one alternative case was not the first character of a range, not all
325        the alternative cases were added to the class. For example, s and \x{17f}
326        are both alternative cases for S: the class [RST] was handled correctly,
327        but [R-T] was not.
329    13. The configure.ac file always checked for pthread support when JIT was
330        enabled. This is not used in Windows, so I have put this test inside a
331        check for the presence of windows.h (which was already tested for).
333    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
334        The algorithm provides a way to skip certain starting offsets, and usually
335        faster than linear prefix searches.
337    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
338        as for 'fr_FR' and 'french'. For some reason, however, it then used the
339        Windows-specific input and output files, which have 'french' screwed in.
340        So this could never have worked. One of the problems with locales is that
341        they aren't always the same. I have now updated RunTest so that it checks
342        the output of the locale test (test 3) against three different output
343        files, and it allows the test to pass if any one of them matches. With luck
344        this should make the test pass on some versions of Solaris where it was
345        failing. Because of the uncertainty, the script did not used to stop if
346        test 3 failed; it now does. If further versions of a French locale ever
347        come to light, they can now easily be added.
349    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
350        there was a message during ./configure, but it did not stop. This now
351        provokes an error. The invalid example in README has been corrected.
352        If a value less than the minimum is given, the minimum value has always
353        been used, but now a warning is given.
355    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
356        was a bug in the test system, which is now fixed. Also, the list of various
357        configurations that are tested for each release did not have one with both
358        16/32 bits and --enable-bar-anycrlf. It now does.
360    18. pcretest was missing "-C bsr" for displaying the \R default setting.
362    19. Little endian PowerPC systems are supported now by the JIT compiler.
364    20. The fast forward newline mechanism could enter to an infinite loop on
365        certain invalid UTF-8 input. Although we don't support these cases
366        this issue can be fixed by a performance optimization.
368    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
369        not take account if existing stack usage. There is now a new global
370        variable called pcre_stack_guard that can be set to point to an external
371        function to check stack availability. It is called at the start of
372        processing every parenthesized group.
374    22. A typo in the code meant that in ungreedy mode the max/min qualifier
375        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
376        not match "ab".
378    23. When UTF was disabled, the JIT program reported some incorrect compile
379        errors. These messages are silenced now.
381    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
382        compiler.
384    25. Change all the temporary files used in RunGrepTest to be different to those
385        used by RunTest so that the tests can be run simultaneously, for example by
386        "make -j check".
389    Version 8.34 15-December-2013
390    -----------------------------
392    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
393        executable memory. Patch inspired by Carsten Klein.
395    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
396        this macro is never tested and has no effect, because the work to support
397        coverage involves only compiling and linking options and special targets in
398        the Makefile. The comment in config.h implied that defining the macro would
399        enable coverage support, which is totally false. There was also support for
400        setting this macro in the CMake files (my fault, I just copied it from
401        configure). SUPPORT_GCOV has now been removed.
403    3.  Make a small performance improvement in strlen16() and strlen32() in
404        pcretest.
406    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
407        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
408        these cases). There was also one in pcretest.c.
410    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
412    6.  In UTF mode, the code for checking whether a group could match an empty
413        string (which is used for indefinitely repeated groups to allow for
414        breaking an infinite loop) was broken when the group contained a repeated
415        negated single-character class with a character that occupied more than one
416        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
417        UTF-8 mode). The effect was undefined: the group might or might not be
418        deemed as matching an empty string, or the program might have crashed.
420    7.  The code for checking whether a group could match an empty string was not
421        recognizing that \h, \H, \v, \V, and \R must match a character.
423    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
424        an empty string. If it can, pcretest shows this in its information output.
426    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
427        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
428        matched by pcre_exec() without using JIT:
430        (a) If the rest of the pattern did not match after a maximal run of
431            grapheme clusters, the code for backing up to try with fewer of them
432            did not always back up over a full grapheme when characters that do not
433            have the modifier quality were involved, e.g. Hangul syllables.
435        (b) If the match point in a subject started with modifier character, and
436            there was no match, the code could incorrectly back up beyond the match
437            point, and potentially beyond the first character in the subject,
438            leading to a segfault or an incorrect match result.
440    10. A conditional group with an assertion condition could lead to PCRE
441        recording an incorrect first data item for a match if no other first data
442        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
443        first data item, and therefore matched "ca" after "c" instead of at the
444        start.
446    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
447        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
449    12. The source of pcregrep now includes z/OS-specific code so that it can be
450        compiled for z/OS as part of the special z/OS distribution.
452    13. Added the -T and -TM options to pcretest.
454    14. The code in pcre_compile.c for creating the table of named capturing groups
455        has been refactored. Instead of creating the table dynamically during the
456        actual compiling pass, the information is remembered during the pre-compile
457        pass (on the stack unless there are more than 20 named groups, in which
458        case malloc() is used) and the whole table is created before the actual
459        compile happens. This has simplified the code (it is now nearly 150 lines
460        shorter) and prepared the way for better handling of references to groups
461        with duplicate names.
463    15. A back reference to a named subpattern when there is more than one of the
464        same name now checks them in the order in which they appear in the pattern.
465        The first one that is set is used for the reference. Previously only the
466        first one was inspected. This change makes PCRE more compatible with Perl.
468    16. Unicode character properties were updated from Unicode 6.3.0.
470    17. The compile-time code for auto-possessification has been refactored, based
471        on a patch by Zoltan Herczeg. It now happens after instead of during
472        compilation. The code is cleaner, and more cases are handled. The option
473        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
474        options in pcretest are provided to set it. It can also be set by
475        (*NO_AUTO_POSSESS) at the start of a pattern.
477    18. The character VT has been added to the default ("C" locale) set of
478        characters that match \s and are generally treated as white space,
479        following this same change in Perl 5.18. There is now no difference between
480        "Perl space" and "POSIX space". Whether VT is treated as white space in
481        other locales depends on the locale.
483    19. The code for checking named groups as conditions, either for being set or
484        for being recursed, has been refactored (this is related to 14 and 15
485        above). Processing unduplicated named groups should now be as fast at
486        numerical groups, and processing duplicated groups should be faster than
487        before.
489    20. Two patches to the CMake build system, by Alexander Barkov:
491          (1) Replace the "source" command by "." in CMakeLists.txt because
492              "source" is a bash-ism.
494          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
495              without these the CMake build does not work on Solaris.
497    21. Perl has changed its handling of \8 and \9. If there is no previously
498        encountered capturing group of those numbers, they are treated as the
499        literal characters 8 and 9 instead of a binary zero followed by the
500        literals. PCRE now does the same.
502    22. Following Perl, added \o{} to specify codepoints in octal, making it
503        possible to specify values greater than 0777 and also making them
504        unambiguous.
506    23. Perl now gives an error for missing closing braces after \x{... instead of
507        treating the string as literal. PCRE now does the same.
509    24. RunTest used to grumble if an inappropriate test was selected explicitly,
510        but just skip it when running all tests. This make it awkward to run ranges
511        of tests when one of them was inappropriate. Now it just skips any
512        inappropriate tests, as it always did when running all tests.
514    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
515        character types such as \d or \w, too many callouts were inserted, and the
516        data that they returned was rubbish.
518    26. In UCP mode, \s was not matching two of the characters that Perl matches,
519        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
520        were matched by \h. The code has now been refactored so that the lists of
521        the horizontal and vertical whitespace characters used for \h and \v (which
522        are defined only in one place) are now also used for \s.
524    27. Add JIT support for the 64 bit TileGX architecture.
525        Patch by Jiong Wang (Tilera Corporation).
527    28. Possessive quantifiers for classes (both explicit and automatically
528        generated) now use special opcodes instead of wrapping in ONCE brackets.
530    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
531        (because it's meaningless), this was not happening when PCRE_CASELESS was
532        set. Not wrong, but inefficient.
534    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
535        properties for \w, \d, etc) is present in a test regex. Otherwise if the
536        test contains no characters greater than 255, Perl doesn't realise it
537        should be using Unicode semantics.
539    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
540        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
541        does in Unicode mode.
543    32. Added the "forbid" facility to pcretest so that putting tests into the
544        wrong test files can sometimes be quickly detected.
546    33. There is now a limit (default 250) on the depth of nesting of parentheses.
547        This limit is imposed to control the amount of system stack used at compile
548        time. It can be changed at build time by --with-parens-nest-limit=xxx or
549        the equivalent in CMake.
551    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
552        errors. Perl warns for these when in warning mode, but PCRE has no facility
553        for giving warnings.
555    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
556        However, this was not working for (?!) because it is optimized to (*FAIL),
557        for which PCRE does not allow quantifiers. The optimization is now disabled
558        when a quantifier follows (?!). I can't see any use for this, but it makes
559        things uniform.
561    36. Perl no longer allows group names to start with digits, so I have made this
562        change also in PCRE. It simplifies the code a bit.
564    37. In extended mode, Perl ignores spaces before a + that indicates a
565        possessive quantifier. PCRE allowed a space before the quantifier, but not
566        before the possessive +. It now does.
568    38. The use of \K (reset reported match start) within a repeated possessive
569        group such as (a\Kb)*+ was not working.
571    40. Document that the same character tables must be used at compile time and
572        run time, and that the facility to pass tables to pcre_exec() and
573        pcre_dfa_exec() is for use only with saved/restored patterns.
575    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
576        features for Builds with MSVC:
578        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
579           builds). The .rc files can be used to set FileDescription and many other
580           attributes.
582        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
583           This allows higher-level build scripts which want .pdb files to avoid
584           hard-coding the exact files needed."
586    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
587        mean "start of word" and "end of word", respectively, as a transition aid.
589    43. A minimizing repeat of a class containing codepoints greater than 255 in
590        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
591        compiled to use the heap for recursion.
593    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
594        is configured.
597    Version 8.33 28-May-2013
598    ------------------------
600    1.  Added 'U' to some constants that are compared to unsigned integers, to
601        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
602        variables that are added to signed variables, to ensure the result is
603        signed and can be negated.
605    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
606        CMake config files.
608    3.  Revise the creation of config.h.generic so that all boolean macros are
609        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
610        overriding via -D on the command line possible.
612    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
613        to unsigned int is reported to make a quite noticeable speed difference in
614        a specific Windows environment. Testing on Linux did also appear to show
615        some benefit (and it is clearly not harmful). Also fixed the definition of
616        Xop which should be unsigned.
618    5.  Related to (4), changing the definition of the intermediate variable cc
619        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
620        improvements.
622    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
623        unnecessary spaces.
625    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
626        and later.
628    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
630    9.  Optimizing fast_forward_start_bits in JIT.
632    10. Adding support for callouts in JIT, and fixing some issues revealed
633        during this work. Namely:
635        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
637        (b) Minimum length was not checked before the matching is started.
639    11. The value of capture_last that is passed to callouts was incorrect in some
640        cases when there was a capture on one path that was subsequently abandoned
641        after a backtrack. Also, the capture_last value is now reset after a
642        recursion, since all captures are also reset in this case.
644    12. The interpreter no longer returns the "too many substrings" error in the
645        case when an overflowing capture is in a branch that is subsequently
646        abandoned after a backtrack.
648    13. In the pathological case when an offset vector of size 2 is used, pcretest
649        now prints out the matched string after a yield of 0 or 1.
651    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
652        Only supported by the JIT compiler at the moment.
654    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
656    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
657        offset of the starting point of the matching process, provided the offsets
658        vector is large enough.
660    17. The \A escape now records a lookbehind value of 1, though its execution
661        does not actually inspect the previous character. This is to ensure that,
662        in partial multi-segment matching, at least one character from the old
663        segment is retained when a new segment is processed. Otherwise, if there
664        are no lookbehinds in the pattern, \A might match incorrectly at the start
665        of a new segment.
667    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
669    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
670        gives some modest performance improvement in 8-bit mode.
672    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
673        be expressed in certain programming languages using Universal Character
674        Names.
676    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
677        which points out that "non characters" are not "characters that may not
678        appear in Unicode strings" but rather "characters that are reserved for
679        internal use and have only local meaning".
681    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
682        there was a conditional group that depended on an assertion, if the
683        assertion was false, the callout that immediately followed the alternation
684        in the condition was skipped when pcre_exec() was used for matching.
686    23. Allow an explicit callout to be inserted before an assertion that is the
687        condition for a conditional group, for compatibility with automatic
688        callouts, which always insert a callout at this point.
690    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
691        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
693    25. (*PRUNE) is now supported by the JIT compiler.
695    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
697    27. Fix the case where there are two or more SKIPs with arguments that may be
698        ignored.
700    28. (*SKIP) is now supported by the JIT compiler.
702    29. (*THEN) is now supported by the JIT compiler.
704    30. Update RunTest with additional test selector options.
706    31. The way PCRE handles backtracking verbs has been changed in two ways.
708        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
709        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
710        backtracking. In some cases this makes it more Perl-compatible, but Perl's
711        rather obscure rules do not always do the same thing.
713        (2) Previously, backtracking verbs were confined within assertions. This is
714        no longer the case for positive assertions, except for (*ACCEPT). Again,
715        this sometimes improves Perl compatibility, and sometimes does not.
717    32. A number of tests that were in test 2 because Perl did things differently
718        have been moved to test 1, because either Perl or PCRE has changed, and
719        these tests are now compatible.
721    32. Backtracking control verbs are now handled in the same way in JIT and
722        interpreter.
724    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
725        contained a forward subroutine reference caused a compile error.
727    34. Auto-detect and optimize limited repetitions in JIT.
729    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
730        blocking (*UTF) etc.
732    36. In the interpreter, maximizing pattern repetitions for characters and
733        character types now use tail recursion, which reduces stack usage.
735    37. The value of the max lookbehind was not correctly preserved if a compiled
736        and saved regex was reloaded on a host of different endianness.
738    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
739        of the compiled pattern block, expand the flags field from 16 to 32 bits
740        because it was almost full.
742    39. Try madvise first before posix_madvise.
744    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
745        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
746        original change locked out all matches of empty strings. This has been
747        changed so that one match of an empty string per line is recognized.
748        Subsequent searches on the same line (for colouring or for --only-matching,
749        for example) do not recognize empty strings.
751    41. Applied a user patch to fix a number of spelling mistakes in comments.
753    42. Data lines longer than 65536 caused pcretest to crash.
755    43. Clarified the data type for length and startoffset arguments for pcre_exec
756        and pcre_dfa_exec in the function-specific man pages, where they were
757        explicitly stated to be in bytes, never having been updated. I also added
758        some clarification to the pcreapi man page.
760    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
761        a segmentation fault.
764    Version 8.32 30-November-2012
765    -----------------------------
767    1.  Improved JIT compiler optimizations for first character search and single
768        character iterators.
770    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
771        Patch by Daniel Richard G.
773    3.  Single character iterator optimizations in the JIT compiler.
775    4.  Improved JIT compiler optimizations for character ranges.
777    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
778        Reported by Giuseppe D'Angelo.
780    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
781        of a line, was being set incorrectly in cases where .* appeared inside
782        atomic brackets at the start of a pattern, or where there was a subsequent
783        *PRUNE or *SKIP.
785    7.  Improved instruction cache flush for POWER/PowerPC.
786        Patch by Daniel Richard G.
788    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
789        grep:
791        (a) There is now no limit to the number of patterns to be matched.
793        (b) An error is given if a pattern is too long.
795        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
796            are now supported.
798        (d) --exclude-from and --include-from (multiple use) have been added.
800        (e) Exclusions and inclusions now apply to all files and directories, not
801            just to those obtained from scanning a directory recursively.
803        (f) Multiple uses of -f and --file-list are now supported.
805        (g) In a Windows environment, the default for -d has been changed from
806            "read" (the GNU grep default) to "skip", because otherwise the presence
807            of a directory in the file list provokes an error.
809        (h) The documentation has been revised and clarified in places.
811    9.  Improve the matching speed of capturing brackets.
813    10. Changed the meaning of \X so that it now matches a Unicode extended
814        grapheme cluster.
816    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
817        out POSIX threads when JIT support is configured.
819    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
821    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
822        structure could go wrong in environments where size_t is not the same size
823        as int.
825    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
826        set.
828    15. The EBCDIC support had decayed; later updates to the code had included
829        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
830        general tidy up of EBCDIC-related issues, and the documentation was also
831        not quite right. There is now a test that can be run on ASCII systems to
832        check some of the EBCDIC-related things (but is it not a full test).
834    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
835        in a small tidy to the code.
837    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
839    18. If the --only-matching (-o) option in pcregrep is specified multiple
840        times, each one causes appropriate output. For example, -o1 -o2 outputs the
841        substrings matched by the 1st and 2nd capturing parentheses. A separating
842        string can be specified by --om-separator (default empty).
844    19. Improving the first n character searches.
846    20. Turn case lists for horizontal and vertical white space into macros so that
847        they are defined only once.
849    21. This set of changes together give more compatible Unicode case-folding
850        behaviour for characters that have more than one other case when UCP
851        support is available.
853        (a) The Unicode property table now has offsets into a new table of sets of
854            three or more characters that are case-equivalent. The MultiStage2.py
855            script that generates these tables (the pcre_ucd.c file) now scans
856            CaseFolding.txt instead of UnicodeData.txt for character case
857            information.
859        (b) The code for adding characters or ranges of characters to a character
860            class has been abstracted into a generalized function that also handles
861            case-independence. In UTF-mode with UCP support, this uses the new data
862            to handle characters with more than one other case.
864        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
865            whose other case is greater than 256 are now correctly matched
866            caselessly. Previously, the high codepoint matched the low one, but not
867            vice versa.
869        (d) The processing of \h, \H, \v, and \ in character classes now makes use
870            of the new class addition function, using character lists defined as
871            macros alongside the case definitions of 20 above.
873        (e) Caseless back references now work with characters that have more than
874            one other case.
876        (f) General caseless matching of characters with more than one other case
877            is supported.
879    22. Unicode character properties were updated from Unicode 6.2.0
881    23. Improved CMake support under Windows. Patch by Daniel Richard G.
883    24. Add support for 32-bit character strings, and UTF-32
885    25. Major JIT compiler update (code refactoring and bugfixing).
886        Experimental Sparc 32 support is added.
888    26. Applied a modified version of Daniel Richard G's patch to create
889        pcre.h.generic and config.h.generic by "make" instead of in the
890        PrepareRelease script.
892    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
893        pcre_compile.c when checking for a zero character.
895    28. Introducing a native interface for JIT. Through this interface, the compiled
896        machine code can be directly executed. The purpose of this interface is to
897        provide fast pattern matching, so several sanity checks are not performed.
898        However, feature tests are still performed. The new interface provides
899        1.4x speedup compared to the old one.
901    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
902        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
903        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
905    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
906        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
907        the "old" RFC 2279). Instead, it ended up passing a negative length to
908        pcre_exec().
910    31. Add support for GCC's visibility feature to hide internal functions.
912    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
913        "unknown -C option" after outputting 0 or 1.
915    33. There is now support for generating a code coverage report for the test
916        suite in environments where gcc is the compiler and lcov is installed. This
917        is mainly for the benefit of the developers.
919    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
920        unaddressable using valgrind annotations, allowing valgrind to detect
921        invalid memory accesses. This is mainly for the benefit of the developers.
923    25. (*UTF) can now be used to start a pattern in any of the three libraries.
925    26. Give configure error if --enable-cpp but no C++ compiler found.
928    Version 8.31 06-July-2012
929    -------------------------
931    1.  Fixing a wrong JIT test case and some compiler warnings.
933    2.  Removed a bashism from the RunTest script.
935    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
936        to unsigned type, result still unsigned" that was given by an MS compiler
937        on encountering the code "-sizeof(xxx)".
939    4.  Partial matching support is added to the JIT compiler.
941    5.  Fixed several bugs concerned with partial matching of items that consist
942        of more than one character:
944        (a) /^(..)\1/ did not partially match "aba" because checking references was
945            done on an "all or nothing" basis. This also applied to repeated
946            references.
948        (b) \R did not give a hard partial match if \r was found at the end of the
949            subject.
951        (c) \X did not give a hard partial match after matching one or more
952            characters at the end of the subject.
954        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
955            a partial match for the string "\r".
957        (e) When newline was set to CRLF, the metacharacter "." did not recognize
958            a partial match for a CR character at the end of the subject string.
960    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
961        running pcretest, the text "(JIT)" added to the output whenever JIT is
962        actually used to run the match.
964    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
965        or /S+[+] with a digit between 1 and 7.
967    8.  OP_NOT now supports any UTF character not just single-byte ones.
969    9.  (*MARK) control verb is now supported by the JIT compiler.
971    10. The command "./RunTest list" lists the available tests without actually
972        running any of them. (Because I keep forgetting what they all are.)
976    12. Applied a (slightly modified) user-supplied patch that improves performance
977        when the heap is used for recursion (compiled with --disable-stack-for-
978        recursion). Instead of malloc and free for each heap frame each time a
979        logical recursion happens, frames are retained on a chain and re-used where
980        possible. This sometimes gives as much as 30% improvement.
982    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
983        call.
985    14. As documented, (*COMMIT) is now confined to within a positive assertion.
987    15. It is now possible to link pcretest with libedit as an alternative to
988        libreadline.
990    16. (*COMMIT) control verb is now supported by the JIT compiler.
992    17. The Unicode data tables have been updated to Unicode 6.1.0.
994    18. Added --file-list option to pcregrep.
996    19. Added binary file support to pcregrep, including the -a, --binary-files,
997        -I, and --text options.
999    20. The madvise function is renamed for posix_madvise for QNX compatibility
1000        reasons. Fixed by Giuseppe D'Angelo.
1002    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1004    22. Changed the option for creating symbolic links for 16-bit man pages from
1005        -s to -sf so that re-installing does not cause issues.
1007    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1009    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1010        matches in certain environments (the workspace was not being correctly
1011        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1012        some of the workspace data at the beginning of a restart.
1014    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1015        was not doing so when it should - probably a typo introduced by SVN 528
1016        (change 8.10/14).
1018    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1019        \w+ when the character tables indicated that \x{c4} was a word character.
1020        There were several related cases, all because the tests for doing a table
1021        lookup were testing for characters less than 127 instead of 255.
1023    27. If a pattern contains capturing parentheses that are not used in a match,
1024        their slots in the ovector are set to -1. For those that are higher than
1025        any matched groups, this happens at the end of processing. In the case when
1026        there were back references that the ovector was too small to contain
1027        (causing temporary malloc'd memory to be used during matching), and the
1028        highest capturing number was not used, memory off the end of the ovector
1029        was incorrectly being set to -1. (It was using the size of the temporary
1030        memory instead of the true size.)
1032    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1033        ovector size, it uses memory at the end of the block that it has got.
1035    29. Check for an overlong MARK name and give an error at compile time. The
1036        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1038    30. JIT compiler update.
1040    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1041        Rill for the patch.
1043    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1045    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1047    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1048        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1049        was enabled.
1051    35. Improve JIT code generation for greedy plus quantifier.
1053    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1054        1 to "aa" instead of to an empty string. The bug affected repeated groups
1055        that could potentially match an empty string.
1057    37. Optimizing single character iterators in JIT.
1059    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1060        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1061        codepoints that are too big for the mode are faulted, and in a UTF mode,
1062        disallowed codepoints are also faulted.
1064    39. If PCRE was compiled with UTF support, in three places in the DFA
1065        matcher there was code that should only have been obeyed in UTF mode, but
1066        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1067        processing when bytes with values greater than 127 were present. In 16-bit
1068        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1069        both cases the values are those that cannot be the first data item in a UTF
1070        character. The three items that might have provoked this were recursions,
1071        possessively repeated groups, and atomic groups.
1073    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1074        and pcregrep, because some OS require shared objects to be explicitly
1075        passed to ld, causing the link step to fail if they are not.
1077    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1078        mode, patterns that started with \h* or \R* might be incorrectly matched.
1081    Version 8.30 04-February-2012
1082    -----------------------------
1084    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1085        name is defined in ctype.h.
1087    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1088        only in quite long subpatterns.
1090    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1091        since it was replaced by pcre_fullinfo() in February 2000.
1093    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1094        match a (*MARK), and the match failed at the start of the subject, a
1095        reference to memory before the start of the subject could occur. This bug
1096        was introduced by fix 17 of release 8.21.
1098    5.  A reference to an unset group with zero minimum repetition was giving
1099        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1100        /(another)?(\1?)test/ matched against "hello world test". This bug was
1101        introduced in release 8.13.
1103    6.  Add support for 16-bit character strings (a large amount of work involving
1104        many changes and refactorings).
1106    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1107        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1108        from a file.
1110    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1111        rounding is not applied in this particular case).
1113    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1114        if they appear, or are escaped, in patterns.
1116    10. Get rid of a number of -Wunused-but-set-variable warnings.
1118    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1119        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1120        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1121        also returns the mark "x". This bug applied to capturing parentheses,
1122        non-capturing parentheses, and atomic parentheses. It also applied to some
1123        assertions.
1125    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1126        information out of configure.ac instead of relying on pcre.h.generic, which
1127        is not stored in the repository.
1129    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1130        -lreadline.
1132    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1134    15. Applied Graycode's patch to put the top-level frame on the stack rather
1135        than the heap when not using the stack for recursion. This gives a
1136        performance improvement in many cases when recursion is not deep.
1138    16. Experimental code added to "pcretest -C" to output the stack frame size.
1141    Version 8.21 12-Dec-2011
1142    ------------------------
1144    1.  Updating the JIT compiler.
1146    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1147        are added as well.
1149    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1150        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1151        calling _pcre_jit_exec. Some extra comments are added.
1153    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1154        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1155        was introduced by change 18 for 8.20.
1157    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1158        ECMA-262 standard.
1160    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1161        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1162        This bug was probably introduced by change 9 of 8.13.
1164    7.  While fixing 6 above, I noticed that a number of other items were being
1165        incorrectly rejected as "not fixed length". This arose partly because newer
1166        opcodes had not been added to the fixed-length checking code. I have (a)
1167        corrected the bug and added tests for these items, and (b) arranged for an
1168        error to occur if an unknown opcode is encountered while checking for fixed
1169        length instead of just assuming "not fixed length". The items that were
1170        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1171        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1172        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1174    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1175        being incorrectly compiled and would have given unpredicatble results.
1177    9.  A possessively repeated subpattern with minimum repeat count greater than
1178        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1179        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1180        the first (A) could occur when it should not.
1182    10. Add a cast and remove a redundant test from the code.
1184    11. JIT should use pcre_malloc/pcre_free for allocation.
1186    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1187        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1188        is anything other than /usr, -L is still shown).
1190    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1192    14. Perl does not support \N without a following name in a [] class; PCRE now
1193        also gives an error.
1195    15. If a forward reference was repeated with an upper limit of around 2000,
1196        it caused the error "internal error: overran compiling workspace". The
1197        maximum number of forward references (including repeats) was limited by the
1198        internal workspace, and dependent on the LINK_SIZE. The code has been
1199        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1200        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1201        of around 200,000 forward references. While doing this, I also speeded up
1202        the filling in of repeated forward references.
1204    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1205        incorrectly expecting the subject to contain another "a" after the start.
1207    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1208        in the match, the SKIP should be ignored. This was not happening; instead
1209        the SKIP was being treated as NOMATCH. For patterns such as
1210        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1211        tested.
1213    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1214        now much more compatible with Perl, in particular in cases where the result
1215        is a non-match for a non-anchored pattern. For example, if
1216        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1217        "m", where previously it did not return a name. A side effect of this
1218        change is that for partial matches, the last encountered mark name is
1219        returned, as for non matches. A number of tests that were previously not
1220        Perl-compatible have been moved into the Perl-compatible test files. The
1221        refactoring has had the pleasing side effect of removing one argument from
1222        the match() function, thus reducing its stack requirements.
1224    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1225        subsequent uses of /S (without +) incorrectly behaved like /S+.
1227    21. Retrieve executable code size support for the JIT compiler and fixing
1228        some warnings.
1230    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1231        not work when the shorter character appeared right at the end of the
1232        subject string.
1234    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1235        systems.
1237    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1238        output it when the /M option is used in pcretest.
1240    25. The CheckMan script was not being included in the distribution. Also, added
1241        an explicit "perl" to run Perl scripts from the PrepareRelease script
1242        because this is reportedly needed in Windows.
1244    26. If study data was being save in a file and studying had not found a set of
1245        "starts with" bytes for the pattern, the data written to the file (though
1246        never used) was taken from uninitialized memory and so caused valgrind to
1247        complain.
1249    27. Updated RunTest.bat as provided by Sheri Pierce.
1251    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1253    29. Computation of memory usage for the table of capturing group names was
1254        giving an unnecessarily large value.
1257    Version 8.20 21-Oct-2011
1258    ------------------------
1260    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1261        a POSIX class. After further experiments with Perl, which convinced me that
1262        Perl has bugs and confusions, a closing square bracket is no longer allowed
1263        in a POSIX name. This bug also affected patterns with classes that started
1264        with full stops.
1266    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1267        captured substring, but while checking the failing first alternative,
1268        substring 1 is temporarily captured. If the output vector supplied to
1269        pcre_exec() was not big enough for this capture, the yield of the function
1270        was still zero ("insufficient space for captured substrings"). This cannot
1271        be totally fixed without adding another stack variable, which seems a lot
1272        of expense for a edge case. However, I have improved the situation in cases
1273        such as /(a)(b)x|abc/ matched against "abc", where the return code
1274        indicates that fewer than the maximum number of slots in the ovector have
1275        been set.
1277    3.  Related to (2) above: when there are more back references in a pattern than
1278        slots in the output vector, pcre_exec() uses temporary memory during
1279        matching, and copies in the captures as far as possible afterwards. It was
1280        using the entire output vector, but this conflicts with the specification
1281        that only 2/3 is used for passing back captured substrings. Now it uses
1282        only the first 2/3, for compatibility. This is, of course, another edge
1283        case.
1285    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1286        main code base, and can be used by building with --enable-jit. When this is
1287        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1288        runtime --no-jit option is given.
1290    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1291        ovector, the return from the function was zero, implying that there were
1292        other matches that did not fit. The correct "exactly full" value is now
1293        returned.
1295    6.  If a subpattern that was called recursively or as a subroutine contained
1296        (*PRUNE) or any other control that caused it to give a non-standard return,
1297        invalid errors such as "Error -26 (nested recursion at the same subject
1298        position)" or even infinite loops could occur.
1300    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1301        computing the minimum length on reaching *ACCEPT, and so ended up with the
1302        wrong value of 1 rather than 0. Further investigation indicates that
1303        computing a minimum subject length in the presence of *ACCEPT is difficult
1304        (think back references, subroutine calls), and so I have changed the code
1305        so that no minimum is registered for a pattern that contains *ACCEPT.
1307    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1308        it was not handled as intended. [But see 16 below.]
1310    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1311        Sheri Pierce.
1313    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1314        the first byte in a match must be "a".
1316    11. Change 17 for 8.13 increased the recursion depth for patterns like
1317        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1318        pattern contains any instances of (*THEN). If it does not, the old
1319        optimizations are restored. It would be nice to do this on a per-group
1320        basis, but at the moment that is not feasible.
1322    12. In some environments, the output of pcretest -C is CRLF terminated. This
1323        broke RunTest's code that checks for the link size. A single white space
1324        character after the value is now allowed for.
1326    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1327        For "fr", it uses the Windows-specific input and output files.
1329    14. If (*THEN) appeared in a group that was called recursively or as a
1330        subroutine, it did not work as intended. [But see next item.]
1332    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1333        pattern fragments (but not containing any | characters). If A and B are
1334        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1335        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1336        D. In other words, Perl considers parentheses that do not contain any |
1337        characters to be part of a surrounding alternative, whereas PCRE was
1338        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1339        differently. PCRE now behaves in the same way as Perl, except in the case
1340        of subroutine/recursion calls such as (?1) which have in any case always
1341        been different (but PCRE had them first :-).
1343    16. Related to 15 above: Perl does not treat the | in a conditional group as
1344        creating alternatives. Such a group is treated in the same way as an
1345        ordinary group without any | characters when processing (*THEN). PCRE has
1346        been changed to match Perl's behaviour.
1348    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1349        RunGrepTest script failed.
1351    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1352        inevitable for groups that contain captures, but it can lead to a lot of
1353        stack use in large patterns. The old behaviour has been restored for atomic
1354        groups that do not contain any capturing parentheses.
1356    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1357        suppress the check for a minimum subject length at run time. (If it was
1358        given to pcre_exec() or pcre_dfa_exec() it did work.)
1360    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1361        fail to work when decoding hex characters in data strings in EBCDIC
1362        environments.
1364    21. It appears that in at least one Mac OS environment, the isxdigit() function
1365        is implemented as a macro that evaluates to its argument more than once,
1366        contravening the C 90 Standard (I haven't checked a later standard). There
1367        was an instance in pcretest which caused it to go wrong when processing
1368        \x{...} escapes in subject strings. The has been rewritten to avoid using
1369        things like p++ in the argument of isxdigit().
1372    Version 8.13 16-Aug-2011
1373    ------------------------
1375    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1377    2.  Two minor typos in pcre_internal.h have been fixed.
1379    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1380        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1381        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1383    4.  There were a number of related bugs in the code for matching backrefences
1384        caselessly in UTF-8 mode when codes for the characters concerned were
1385        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1386        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1387        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1388        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1389        2-byte code at the end of the subject (it thought there wasn't enough data
1390        left).
1392    5.  Comprehensive information about what went wrong is now returned by
1393        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1394        as the output vector has at least 2 elements. The offset of the start of
1395        the failing character and a reason code are placed in the vector.
1397    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1398        now returned is for the first byte of the failing character, instead of the
1399        last byte inspected. This is an incompatible change, but I hope it is small
1400        enough not to be a problem. It makes the returned offset consistent with
1401        pcre_exec() and pcre_dfa_exec().
1403    7.  pcretest now gives a text phrase as well as the error number when
1404        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1405        failure, the offset and reason code are output.
1407    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1408        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1409        back over a single character (\n). This seems wrong (because it treated the
1410        two characters as a single entity when going forwards), conflicts with the
1411        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1412        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1413        has been changed.
1415    9.  Some internal refactoring has changed the processing so that the handling
1416        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1417        time (the PCRE_DOTALL option was changed this way some time ago: version
1418        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1419        which was always a bit of a fudge. It also means that there is one less
1420        argument for the match() function, which reduces its stack requirements
1421        slightly. This change also fixes an incompatibility with Perl: the pattern
1422        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1424    10. More internal refactoring has drastically reduced the number of recursive
1425        calls to match() for possessively repeated groups such as (abc)++ when
1426        using pcre_exec().
1428    11. While implementing 10, a number of bugs in the handling of groups were
1429        discovered and fixed:
1431        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1432        (a|)*(?1) gave a compile-time internal error.
1433        ((a|)+)+  did not notice that the outer group could match an empty string.
1434        (^a|^)+   was not marked as anchored.
1435        (.*a|.*)+ was not marked as matching at start or after a newline.
1437    12. Yet more internal refactoring has removed another argument from the match()
1438        function. Special calls to this function are now indicated by setting a
1439        value in a variable in the "match data" data block.
1441    13. Be more explicit in pcre_study() instead of relying on "default" for
1442        opcodes that mean there is no starting character; this means that when new
1443        ones are added and accidentally left out of pcre_study(), testing should
1444        pick them up.
1446    14. The -s option of pcretest has been documented for ages as being an old
1447        synonym of -m (show memory usage). I have changed it to mean "force study
1448        for every regex", that is, assume /S for every regex. This is similar to -i
1449        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1450        using it. It makes it easier to run collections of tests with and without
1451        study enabled, and thereby test pcre_study() more easily. All the standard
1452        tests are now run with and without -s (but some patterns can be marked as
1453        "never study" - see 20 below).
1455    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1456        restoration of the capturing data to the outer values was not happening
1457        correctly.
1459    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1460        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1461        pattern had matched an empty string, and so incorrectly returned a no
1462        match.
1464    17. There was optimizing code for the last branch of non-capturing parentheses,
1465        and also for the obeyed branch of a conditional subexpression, which used
1466        tail recursion to cut down on stack usage. Unfortunately, now that there is
1467        the possibility of (*THEN) occurring in these branches, tail recursion is
1468        no longer possible because the return has to be checked for (*THEN). These
1469        two optimizations have therefore been removed. [But see 8.20/11 above.]
1471    18. If a pattern containing \R was studied, it was assumed that \R always
1472        matched two bytes, thus causing the minimum subject length to be
1473        incorrectly computed because \R can also match just one byte.
1475    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1476        was incorrectly computed.
1478    20. If /S is present twice on a test pattern in pcretest input, it now
1479        *disables* studying, thereby overriding the use of -s on the command line
1480        (see 14 above). This is necessary for one or two tests to keep the output
1481        identical in both cases.
1483    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1484        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1486    22. When an atomic group that contained a capturing parenthesis was
1487        successfully matched, but the branch in which it appeared failed, the
1488        capturing was not being forgotten if a higher numbered group was later
1489        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1490        group 1 to "a", when in fact it should be unset. This applied to multi-
1491        branched capturing and non-capturing groups, repeated or not, and also to
1492        positive assertions (capturing in negative assertions does not happen
1493        in PCRE) and also to nested atomic groups.
1495    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1496        subject after a captured substring, to make it easier to tell which of a
1497        number of identical substrings has been captured.
1499    24. The way atomic groups are processed by pcre_exec() has been changed so that
1500        if they are repeated, backtracking one repetition now resets captured
1501        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1502        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1503        "aaa". Previously, it would have been "a". As part of this code
1504        refactoring, the way recursive calls are handled has also been changed.
1506    25. If an assertion condition captured any substrings, they were not passed
1507        back unless some other capturing happened later. For example, if
1508        (?(?=(a))a) was matched against "a", no capturing was returned.
1510    26. When studying a pattern that contained subroutine calls or assertions,
1511        the code for finding the minimum length of a possible match was handling
1512        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1513        group 1 called group 2 while simultaneously a separate group 2 called group
1514        1). A stack overflow occurred in this case. I have fixed this by limiting
1515        the recursion depth to 10.
1517    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1518        Fortmann. This supports explicit test numbers on the command line, and has
1519        argument validation and error reporting.
1521    28. An instance of \X with an unlimited repeat could fail if at any point the
1522        first character it looked at was a mark character.
1524    29. Some minor code refactoring concerning Unicode properties and scripts
1525        should reduce the stack requirement of match() slightly.
1527    30. Added the '=' option to pcretest to check the setting of unused capturing
1528        slots at the end of the pattern, which are documented as being -1, but are
1529        not included in the return count.
1531    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1532        compiled something random. Now it gives a compile-time error (as does
1533        Perl).
1535    32. A *MARK encountered during the processing of a positive assertion is now
1536        recorded and passed back (compatible with Perl).
1538    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1539        had alternative anchored branches, the search for a second match in a line
1540        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1541        matched the line "0102" twice. The same bug affected patterns that started
1542        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1543        twice.
1545    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1546        does, and because of capturing effects, quantifying parenthesized
1547        assertions may at times be useful. Quantifiers are now allowed for
1548        parenthesized assertions.
1550    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1552    36. \g was being checked for fancy things in a character class, when it should
1553        just be a literal "g".
1555    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1556        appearance of a nested POSIX class supersedes an apparent external class.
1557        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1558        unescaped square brackets may also appear as part of class names. For
1559        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1560        more like Perl. (But see 8.20/1 above.)
1562    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1563        was because it thought it was \N{name}, which is not supported).
1565    39. Add minix to OS list not supporting the -S option in pcretest.
1567    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1568        cannot analyze patterns in sufficient detail to catch mutual recursions
1569        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1570        subgroup is called recursively as a subpattern for a second time at the
1571        same position in the subject string. In previous releases this might have
1572        been caught by the recursion limit, or it might have run out of stack.
1574    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1575        happen only once. PCRE was, however incorrectly giving a compile time error
1576        "recursive call could loop indefinitely" because it cannot analyze the
1577        pattern in sufficient detail. The compile time test no longer happens when
1578        PCRE is compiling a conditional subpattern, but actual runaway loops are
1579        now caught at runtime (see 40 above).
1581    42. It seems that Perl allows any characters other than a closing parenthesis
1582        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1583        has been changed to be the same.
1585    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1586        as not to get warnings when autogen.sh is called. Also changed
1587        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1589    44. To help people who use pcregrep to scan files containing exceedingly long
1590        lines, the following changes have been made:
1592        (a) The default value of the buffer size parameter has been increased from
1593            8K to 20K. (The actual buffer used is three times this size.)
1595        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1596            PCRE is built.
1598        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1599            to be set at run time.
1601        (d) Numerical values in pcregrep options can be followed by K or M, for
1602            example --buffer-size=50K.
1604        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1605            given and the return code is set to 2.
1607    45. Add a pointer to the latest mark to the callout data block.
1609    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1610        partial match of an empty string instead of no match. This was specific to
1611        the use of ".".
1613    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1614        complete match instead of a partial match. This bug was dependent on both
1615        the PCRE_UTF8 and PCRE_DOTALL options being set.
1617    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1618        starting byte set, because \b was not being ignored.
1621    Version 8.12 15-Jan-2011
1622    ------------------------
1624    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1625        checks for such things as part of the documentation building process.
1627    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1628        --match-limit and --recursion-limit options (added for 8.11). In
1629        particular, this made one of the standard tests fail. (The integer value
1630        went into the wrong half of a long int.)
1632    3.  If the --colour option was given to pcregrep with -v (invert match), it
1633        did strange things, either producing crazy output, or crashing. It should,
1634        of course, ignore a request for colour when reporting lines that do not
1635        match.
1637    4.  Another pcregrep bug caused similar problems if --colour was specified with
1638        -M (multiline) and the pattern match finished with a line ending.
1640    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1641        matched in multiline mode, the following line was shown as part of the
1642        match. This seems wrong, so I have changed it.
1644    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1645        the check for further matches in the same line (so they could be coloured)
1646        to overrun the end of the current line. If another match was found, it was
1647        incorrectly shown (and then shown again when found in the next line).
1649    7.  If pcregrep was compiled under Windows, there was a reference to the
1650        function pcregrep_exit() before it was defined. I am assuming this was
1651        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1652        reported by a user. I've moved the definition above the reference.
1655    Version 8.11 10-Dec-2010
1656    ------------------------
1658    1.  (*THEN) was not working properly if there were untried alternatives prior
1659        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1660        backtracked to try for "b" instead of moving to the next alternative branch
1661        at the same level (in this case, to look for "c"). The Perl documentation
1662        is clear that when (*THEN) is backtracked onto, it goes to the "next
1663        alternative in the innermost enclosing group".
1665    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1666        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1667        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1668        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1669        (*THEN).
1671    3.  If \s appeared in a character class, it removed the VT character from
1672        the class, even if it had been included by some previous item, for example
1673        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1674        of \s, but is part of the POSIX "space" class.)
1676    4.  A partial match never returns an empty string (because you can always
1677        match an empty string at the end of the subject); however the checking for
1678        an empty string was starting at the "start of match" point. This has been
1679        changed to the "earliest inspected character" point, because the returned
1680        data for a partial match starts at this character. This means that, for
1681        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1682        (previously it gave "no match").
1684    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1685        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1686        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1687        has an implication that the given string is incomplete (because a partial
1688        match is preferred over a full match). For this reason, these items now
1689        give a partial match in this situation. [Aside: previously, the one case
1690        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1691        match rather than a full match, which was wrong by the old rules, but is
1692        now correct.]
1694    6.  There was a bug in the handling of #-introduced comments, recognized when
1695        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1696        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1697        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1698        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1699        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1700        places in pcre_compile().
1702    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1703        comments when looking ahead for named forward references to subpatterns,
1704        the only newline sequence it recognized was NL. It now handles newlines
1705        according to the set newline convention.
1707    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1708        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1709        cater for a lack of strerror(). These oversights have been fixed.
1711    9.  Added --match-limit and --recursion-limit to pcregrep.
1713    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1715    11. When the -o option was used, pcregrep was setting a return code of 1, even
1716        when matches were found, and --line-buffered was not being honoured.
1718    12. Added an optional parentheses number to the -o and --only-matching options
1719        of pcregrep.
1721    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1722        can match an empty string. The code to do it in pcretest and pcredemo
1723        needed fixing:
1725        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1726            only one byte after an empty string match just before CRLF (this case
1727            just got forgotten; "any" and "anycrlf" were OK).
1729        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1730            mode when an empty string match preceded an ASCII character followed by
1731            a non-ASCII character. (The code for advancing by one character rather
1732            than one byte was nonsense.)
1734        (c) The pcredemo.c sample program did not have any code at all to handle
1735            the cases when CRLF is a valid newline sequence.
1737    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1738        as a starting offset was within the subject string. There is now a new
1739        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1740        negative or greater than the length of the string. In order to test this,
1741        pcretest is extended to allow the setting of negative starting offsets.
1743    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1744        starting offset points to the beginning of a UTF-8 character was
1745        unnecessarily clumsy. I tidied it up.
1747    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1748        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1750    17. Nobody had reported that the --include_dir option, which was added in
1751        release 7.7 should have been called --include-dir (hyphen, not underscore)
1752        for compatibility with GNU grep. I have changed it to --include-dir, but
1753        left --include_dir as an undocumented synonym, and the same for
1754        --exclude-dir, though that is not available in GNU grep, at least as of
1755        release 2.5.4.
1757    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1758        characters from a string of bytes) have been redefined so as not to use
1759        loops, in order to improve performance in some environments. At the same
1760        time, I abstracted some of the common code into auxiliary macros to save
1761        repetition (this should not affect the compiled code).
1763    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1764        compile-time error is now given if \c is not followed by an ASCII
1765        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1766        different, and any byte value is allowed.)
1768    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1769        START_OPTIMIZE option, which is now allowed at compile time - but just
1770        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1771        to pcregrep and other applications that have no direct access to PCRE
1772        options. The new /Y option in pcretest sets this option when calling
1773        pcre_compile().
1775    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1776        back references. Groups containing recursive back references were forced to
1777        be atomic by that change, but in the case of named groups, the amount of
1778        memory required was incorrectly computed, leading to "Failed: internal
1779        error: code overflow". This has been fixed.
1781    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1782        pcretest.c, to avoid build problems in some Borland environments.
1785    Version 8.10 25-Jun-2010
1786    ------------------------
1788    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1789        THEN.
1791    2.  (*ACCEPT) was not working when inside an atomic group.
1793    3.  Inside a character class, \B is treated as a literal by default, but
1794        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1795        causes the error). The code is unchanged, but I tidied the documentation.
1797    4.  Inside a character class, PCRE always treated \R and \X as literals,
1798        whereas Perl faults them if its -w option is set. I have changed PCRE so
1799        that it faults them when PCRE_EXTRA is set.
1801    5.  Added support for \N, which always matches any character other than
1802        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1804    6.  When compiling pcregrep with newer versions of gcc which may have
1805        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1806        declared with attribute warn_unused_result" were given. Just casting the
1807        result to (void) does not stop the warnings; a more elaborate fudge is
1808        needed. I've used a macro to implement this.
1810    7.  Minor change to pcretest.c to avoid a compiler warning.
1812    8.  Added four artifical Unicode properties to help with an option to make
1813        \s etc use properties (see next item). The new properties are: Xan
1814        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1816    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1817        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1818        this option. Modified pcretest to add /W to test this facility. Added
1819        REG_UCP to make it available via the POSIX interface.
1821    10. Added --line-buffered to pcregrep.
1823    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1824        studied, and the match started with a letter with a code point greater than
1825        127 whose first byte was different to the first byte of the other case of
1826        the letter, the other case of this starting letter was not recognized
1827        (#976).
1829    12. If a pattern that was studied started with a repeated Unicode property
1830        test, for example, \p{Nd}+, there was the theoretical possibility of
1831        setting up an incorrect bitmap of starting bytes, but fortunately it could
1832        not have actually happened in practice until change 8 above was made (it
1833        added property types that matched character-matching opcodes).
1835    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1836        possible starting bytes for non-anchored patterns.
1838    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1839        \R, and also a number of cases that involve Unicode properties, both
1840        explicit and implicit when PCRE_UCP is set.
1842    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1843        input, it could crash or give wrong results if characters with values
1844        greater than 0xc0 were present in the subject string. (Detail: it assumed
1845        UTF-8 input when processing these items.)
1847    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1848        size_t is 64-bit (#991).
1850    17. Added a check for running out of memory when PCRE is compiled with
1851        --disable-stack-for-recursion (#990).
1853    18. If the last data line in a file for pcretest does not have a newline on
1854        the end, a newline was missing in the output.
1856    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1857        less than 128) in its various bitmaps. However, there is a facility for
1858        generating tables according to the current locale when PCRE is compiled. It
1859        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1860        characters, are recognized by isspace() and therefore were getting set in
1861        these tables, and indeed these tables seem to approximate to ISO 8859. This
1862        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1863        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1864        which of course cannot start UTF-8 characters. I have changed the code so
1865        that only real ASCII characters (less than 128) and the correct starting
1866        bytes for UTF-8 encodings are set for characters greater than 127 when in
1867        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1868        altogether.)
1870    20. Added the /T option to pcretest so as to be able to run tests with non-
1871        standard character tables, thus making it possible to include the tests
1872        used for 19 above in the standard set of tests.
1874    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1875        reference to a subpattern the other side of a comment that contains an
1876        opening parenthesis caused either an internal compiling error, or a
1877        reference to the wrong subpattern.
1880    Version 8.02 19-Mar-2010
1881    ------------------------
1883    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1885    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1886        configured.
1888    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1889        original author of that file, following a query about its status.
1891    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1892        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1894    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1895        quantifier applied to a forward-referencing subroutine call, could compile
1896        incorrect code or give the error "internal error: previously-checked
1897        referenced subpattern not found".
1899    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1900        variables to point to external functions. For these systems, therefore,
1901        pcre_malloc etc. are now initialized to local functions that call the
1902        relevant global functions.
1904    7.  There were two entries missing in the vectors called coptable and poptable
1905        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1906        I've fixed the data, and added a kludgy way of testing at compile time that
1907        the lengths are correct (equal to the number of opcodes).
1909    8.  Following on from 7, I added a similar kludge to check the length of the
1910        eint vector in pcreposix.c.
1912    9.  Error texts for pcre_compile() are held as one long string to avoid too
1913        much relocation at load time. To find a text, the string is searched,
1914        counting zeros. There was no check for running off the end of the string,
1915        which could happen if a new error number was added without updating the
1916        string.
1918    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1920    11. \K was not working if it appeared in an atomic group or in a group that
1921        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1922        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1923        the assertion is positive, but not if it is negative.
1925    12. Change 11 fortuitously reduced the size of the stack frame used in the
1926        "match()" function of pcre_exec.c by one pointer. Forthcoming
1927        implementation of support for (*MARK) will need an extra pointer on the
1928        stack; I have reserved it now, so that the stack frame size does not
1929        decrease.
1931    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1932        item in branch that calls a recursion is a subroutine call - as in the
1933        second branch in the above example - was incorrectly given the compile-
1934        time error "recursive call could loop indefinitely" because pcre_compile()
1935        was not correctly checking the subroutine for matching a non-empty string.
1937    14. The checks for overrunning compiling workspace could trigger after an
1938        overrun had occurred. This is a "should never occur" error, but it can be
1939        triggered by pathological patterns such as hundreds of nested parentheses.
1940        The checks now trigger 100 bytes before the end of the workspace.
1942    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1945    Version 8.01 19-Jan-2010
1946    ------------------------
1948    1.  If a pattern contained a conditional subpattern with only one branch (in
1949        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1950        computed the wrong minimum data length (which is of course zero for such
1951        subpatterns). This could cause incorrect "no match" results.
1953    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1954        the pattern is reset in the first branch, pcre_compile() failed with
1955        "internal error: code overflow at offset...". This happened only when
1956        the reset was to the original external option setting. (An optimization
1957        abstracts leading options settings into an external setting, which was the
1958        cause of this.)
1960    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1961        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1962        assertion pattern did not match (meaning that the assertion was true), it
1963        was incorrectly treated as false if the SKIP had been reached during the
1964        matching. This also applied to assertions used as conditions.
1966    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1967        assertion subpattern, including such a pattern used as a condition,
1968        unpredictable results occurred, instead of the error return
1971    5.  The C++ GlobalReplace function was not working like Perl for the special
1972        situation when an empty string is matched. It now does the fancy magic
1973        stuff that is necessary.
1975    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1976        removed. (These were left over from very, very early versions of PCRE.)
1978    7.  Some cosmetic changes to the code to make life easier when compiling it
1979        as part of something else:
1981        (a) Change DEBUG to PCRE_DEBUG.
1983        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1984            called "current" as "current_branch", to prevent a collision with the
1985            Linux macro when compiled as a kernel module.
1987        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1988            prevent a collision with the Linux macro when compiled as a kernel
1989            module.
1991    8.  In pcre_compile() there are some checks for integer overflows that used to
1992        cast potentially large values to (double). This has been changed to that
1993        when building, a check for int64_t is made, and if it is found, it is used
1994        instead, thus avoiding the use of floating point arithmetic. (There is no
1995        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1996        double.
1998    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1999        2005 (difference between two addresses compared to an unsigned value).
2001    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2002        custom one, because of the following reported problem in Windows:
2004          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2005              under Win32.
2006          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2007              therefore missing the function definition.
2008          - The compiler thus generates a "C" signature for the test function.
2009          - The linker fails to find the "C" function.
2010          - PCRE fails to configure if asked to do so against libbz2.
2012    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2013        messages were output:
2015          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2016          rerunning libtoolize, to keep the correct libtool macros in-tree.
2017          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2019        I have done both of these things.
2021    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2022        most of the time, it *can* run out if it is given a pattern that contains a
2023        runaway infinite recursion. I updated the discussion in the pcrestack man
2024        page.
2026    13. Now that we have gone to the x.xx style of version numbers, the minor
2027        version may start with zero. Using 08 or 09 is a bad idea because users
2028        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2029        interpreted as invalid octal numbers. I've updated the previous comment in
2030        configure.ac, and also added a check that gives an error if 08 or 09 are
2031        used.
2033    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2034        causing partial matching to fail when the end of the subject matched \W
2035        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2037    15. There were some discrepancies between the declarations in pcre_internal.h
2038        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2039        their definitions. The declarations used "const uschar *" and the
2040        definitions used USPTR. Even though USPTR is normally defined as "const
2041        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2042        reported that: "This difference in casting confuses some C++ compilers, for
2043        example, SunCC recognizes above declarations as different functions and
2044        generates broken code for hbpcre." I have changed the declarations to use
2045        USPTR.
2047    16. GNU libtool is named differently on some systems. The autogen.sh script now
2048        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2049        (FreeBSD).
2051    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2052        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2053        comment: "Figure out how to create a longlong from a string: strtoll and
2054        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2055        instance, but it only takes 2 args instead of 3!"
2057    18. A subtle bug concerned with back references has been fixed by a change of
2058        specification, with a corresponding code fix. A pattern such as
2059        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2060        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2061        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2062        same bug. Such groups have to be quantified to be useful, or contained
2063        inside another quantified group. (If there's no repetition, the reference
2064        can never match.) The problem arises because, having left the group and
2065        moved on to the rest of the pattern, a later failure that backtracks into
2066        the group uses the captured value from the final iteration of the group
2067        rather than the correct earlier one. I have fixed this in PCRE by forcing
2068        any group that contains a reference to itself to be an atomic group; that
2069        is, there cannot be any backtracking into it once it has completed. This is
2070        similar to recursive and subroutine calls.
2073    Version 8.00 19-Oct-09
2074    ----------------------
2076    1.  The table for translating pcre_compile() error codes into POSIX error codes
2077        was out-of-date, and there was no check on the pcre_compile() error code
2078        being within the table. This could lead to an OK return being given in
2079        error.
2081    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2082        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2083        in a Windows environment.
2085    3.  The pcregrep --count option prints the count for each file even when it is
2086        zero, as does GNU grep. However, pcregrep was also printing all files when
2087        --files-with-matches was added. Now, when both options are given, it prints
2088        counts only for those files that have at least one match. (GNU grep just
2089        prints the file name in this circumstance, but including the count seems
2090        more useful - otherwise, why use --count?) Also ensured that the
2091        combination -clh just lists non-zero counts, with no names.
2093    4.  The long form of the pcregrep -F option was incorrectly implemented as
2094        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2095        but it seems right to fix it, and I didn't think it was worth preserving
2096        the old behaviour.
2098    5.  The command line items --regex=pattern and --regexp=pattern were not
2099        recognized by pcregrep, which required --regex pattern or --regexp pattern
2100        (with a space rather than an '='). The man page documented the '=' forms,
2101        which are compatible with GNU grep; these now work.
2103    6.  No libpcreposix.pc file was created for pkg-config; there was just
2104        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2106    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2107        when UCP support is not needed, by modifying the Python script that
2108        generates it from Unicode data files. This should not matter if the module
2109        is correctly used as a library, but I received one complaint about 50K of
2110        unwanted data. My guess is that the person linked everything into his
2111        program rather than using a library. Anyway, it does no harm.
2113    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2114        was a minimum greater than 1 for a wide character in a possessive
2115        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2116        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2117        character. Chaos in the form of incorrect output or a compiling loop could
2118        result.
2120    9.  The restrictions on what a pattern can contain when partial matching is
2121        requested for pcre_exec() have been removed. All patterns can now be
2122        partially matched by this function. In addition, if there are at least two
2123        slots in the offset vector, the offset of the earliest inspected character
2124        for the match and the offset of the end of the subject are set in them when
2125        PCRE_ERROR_PARTIAL is returned.
2127    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2128        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2129        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2130        and may be more useful for multi-segment matching.
2132    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2133        used to be given if ever the end of the subject was reached; now it is
2134        given only if matching could not proceed because another character was
2135        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2136        string "Z", which now yields "no match" instead of "partial match". In the
2137        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2138        final character ended with (*FAIL).
2140    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2141        if the pattern had a "must contain" character that was already found in the
2142        earlier partial match, unless partial matching was again requested. For
2143        example, with the pattern /dog.(body)?/, the "must contain" character is
2144        "g". If the first part-match was for the string "dog", restarting with
2145        "sbody" failed. This bug has been fixed.
2147    13. The string returned by pcre_dfa_exec() after a partial match has been
2148        changed so that it starts at the first inspected character rather than the
2149        first character of the match. This makes a difference only if the pattern
2150        starts with a lookbehind assertion or \b or \B (\K is not supported by
2151        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2152        matching functions compatible, and I think it's the right thing to do.
2154    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2155        so that the demonstration program is easily available in environments where
2156        PCRE has not been installed from source.
2158    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2159        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2160        library.
2162    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2163        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2164        is not the first non-POSIX option to be added. Clearly some people find
2165        these options useful.
2167    17. If a caller to the POSIX matching function regexec() passes a non-zero
2168        value for nmatch with a NULL value for pmatch, the value of
2169        nmatch is forced to zero.
2171    18. RunGrepTest did not have a test for the availability of the -u option of
2172        the diff command, as RunTest does. It now checks in the same way as
2173        RunTest, and also checks for the -b option.
2175    19. If an odd number of negated classes containing just a single character
2176        interposed, within parentheses, between a forward reference to a named
2177        subpattern and the definition of the subpattern, compilation crashed with
2178        an internal error, complaining that it could not find the referenced
2179        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2180        [The bug was that it was starting one character too far in when skipping
2181        over the character class, thus treating the ] as data rather than
2182        terminating the class. This meant it could skip too much.]
2184    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2185        /g option in pcretest when the pattern contains \K, which makes it possible
2186        to have an empty string match not at the start, even when the pattern is
2187        anchored. Updated pcretest and pcredemo to use this option.
2189    21. If the maximum number of capturing subpatterns in a recursion was greater
2190        than the maximum at the outer level, the higher number was returned, but
2191        with unset values at the outer level. The correct (outer level) value is
2192        now given.
2194    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2195        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2196        make it do so. The string so far is captured, making this feature
2197        compatible with Perl.
2199    23. The tests have been re-organized, adding tests 11 and 12, to make it
2200        possible to check the Perl 5.10 features against Perl 5.10.
2202    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2203        pattern matches a fixed length string. PCRE did not allow this; now it
2204        does. Neither allows recursion.
2206    25. I finally figured out how to implement a request to provide the minimum
2207        length of subject string that was needed in order to match a given pattern.
2208        (It was back references and recursion that I had previously got hung up
2209        on.) This code has now been added to pcre_study(); it finds a lower bound
2210        to the length of subject needed. It is not necessarily the greatest lower
2211        bound, but using it to avoid searching strings that are too short does give
2212        some useful speed-ups. The value is available to calling programs via
2213        pcre_fullinfo().
2215    26. While implementing 25, I discovered to my embarrassment that pcretest had
2216        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2217        study optimizations had never been tested with that matching function.
2218        Oops. What is worse, even when it was passed study data, there was a bug in
2219        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2220        were also very few tests of studied patterns with pcre_dfa_exec().
2222    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2223        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2224        on the other side of the coin, they are no longer allowed to have different
2225        names, because these cannot be distinguished in PCRE, and this has caused
2226        confusion. (This is a difference from Perl.)
2228    28. When duplicate subpattern names are present (necessarily with different
2229        numbers, as required by 27 above), and a test is made by name in a
2230        conditional pattern, either for a subpattern having been matched, or for
2231        recursion in such a pattern, all the associated numbered subpatterns are
2232        tested, and the overall condition is true if the condition is true for any
2233        one of them. This is the way Perl works, and is also more like the way
2234        testing by number works.
2237    Version 7.9 11-Apr-09
2238    ---------------------
2240    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2241        (pcretest), all targets were linked against these libraries. This included
2242        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2243        libraries. This caused unwanted dependencies to be created. This problem
2244        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2245        pcretest is linked with readline.
2247    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2248        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2249        moved outside it again, because FALSE and TRUE are already defined in AIX,
2250        but BOOL is not.
2252    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2253        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2255    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2256        hyphens) following file names and line numbers when outputting matching
2257        lines. This is not true; no spaces are inserted. I have also clarified the
2258        wording for the --colour (or --color) option.
2260    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2261        was not coloured; this is different to GNU grep, so I have changed it to be
2262        the same.
2264    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2265        each matching line was coloured. Now it goes on to look for further matches
2266        of any of the test patterns, which is the same behaviour as GNU grep.
2268    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2269        doesn't make sense to accept an empty string match in pcregrep, so I have
2270        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2271        seems to be how GNU grep behaves. [But see later change 40 for release
2272        8.33.]
2274    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2275        start or after a newline", because the conditional assertion was not being
2276        correctly handled. The rule now is that both the assertion and what follows
2277        in the first alternative must satisfy the test.
2279    9.  If auto-callout was enabled in a pattern with a conditional group whose
2280        condition was an assertion, PCRE could crash during matching, both with
2281        pcre_exec() and pcre_dfa_exec().
2283    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2284        used for matching.
2286    11. Unicode property support in character classes was not working for
2287        characters (bytes) greater than 127 when not in UTF-8 mode.
2289    12. Added the -M command line option to pcretest.
2291    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2293    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2295    16. Added comments and documentation about mis-use of no_arg in the C++
2296        wrapper.
2298    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2299        from Martin Jerabek that uses macro names for all relevant character and
2300        string constants.
2302    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2303        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2304        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2305        these, but not everybody uses configure.
2307    19. A conditional group that had only one branch was not being correctly
2308        recognized as an item that could match an empty string. This meant that an
2309        enclosing group might also not be so recognized, causing infinite looping
2310        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2311        with the subject "ab", where knowledge that the repeated group can match
2312        nothing is needed in order to break the loop.
2314    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2315        exec(), but without supplying a callout function, matching went wrong.
2317    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2318        leak if the size of the offset vector was greater than 30. When the vector
2319        is smaller, the saved offsets during recursion go onto a local stack
2320        vector, but for larger vectors malloc() is used. It was failing to free
2321        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2322        error, in fact).
2324    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2325        heapframe that is used only when UTF-8 support is enabled. This caused no
2326        problem, but was untidy.
2328    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2329        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2330        included within another project.
2332    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2333        slightly modified by me:
2335          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2336              not building pcregrep.
2338          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2339              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2341    25. Forward references, both numeric and by name, in patterns that made use of
2342        duplicate group numbers, could behave incorrectly or give incorrect errors,
2343        because when scanning forward to find the reference group, PCRE was not
2344        taking into account the duplicate group numbers. A pattern such as
2345        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2347    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2348        the feature of a custom pointer more persuasive (as requested by a user).
2350    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2351        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2352        pre-defined.
2354    28. Added support for (*UTF8) at the start of a pattern.
2356    29. Arrange for flags added by the "release type" setting in CMake to be shown
2357        in the configuration summary.
2360    Version 7.8 05-Sep-08
2361    ---------------------
2363    1.  Replaced UCP searching code with optimized version as implemented for Ad
2364        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2365        stage table and inline lookup instead of a function, giving speed ups of 2
2366        to 5 times on some simple patterns that I tested. Permission was given to
2367        distribute the MultiStage2.py script that generates the tables (it's not in
2368        the tarball, but is in the Subversion repository).
2370    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2371        scripts.
2373    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2374        a group with a zero qualifier. The result of the study could be incorrect,
2375        or the function might crash, depending on the pattern.
2377    4.  Caseless matching was not working for non-ASCII characters in back
2378        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2379        It now works when Unicode Property Support is available.
2381    5.  In pcretest, an escape such as \x{de} in the data was always generating
2382        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2383        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2384        truncation.
2386    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2388    7.  Added two (int) casts to pcregrep when printing the difference of two
2389        pointers, in case they are 64-bit values.
2391    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2392        test 2 if it fails.
2394    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2395        and a #define of that name to empty if it is not externally set. This is to
2396        allow users of MSVC to set it if necessary.
2398    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2399        the convenience functions in the pcre_get.c source file.
2401    11. An option change at the start of a pattern that had top-level alternatives
2402        could cause overwriting and/or a crash. This command provoked a crash in
2403        some environments:
2405          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2407        This potential security problem was recorded as CVE-2008-2371.
2409    12. For a pattern where the match had to start at the beginning or immediately
2410        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2411        pcre_dfa_exec() could read past the end of the passed subject if there was
2412        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2413        pcretest so that it places the subject at the end of its malloc-ed buffer.
2415    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2416        exec() might read past the end of the data buffer in UTF-8 mode.
2418    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2419        the data contained the byte 0x85 as part of a UTF-8 character within its
2420        first line. This applied both to normal and DFA matching.
2422    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2423        /^[^d]*?$/8 failed to match "abc".
2425    16. Added a missing copyright notice to pcrecpp_internal.h.
2427    17. Make it more clear in the documentation that values returned from
2428        pcre_exec() in ovector are byte offsets, not character counts.
2430    18. Tidied a few places to stop certain compilers from issuing warnings.
2432    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2433        supplied by Stefan Weber. I made a further small update for 7.8 because
2434        there is a change of source arrangements: the pcre_searchfuncs.c module is
2435        replaced by pcre_ucd.c.
2438    Version 7.7 07-May-08
2439    ---------------------
2441    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2442        a string to a long long, pretend we don't even have a long long." This is
2443        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2445    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2446        pre-7.6 versions, which defined a global no_arg variable instead of putting
2447        it in the RE class. (See also #8 below.)
2449    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2450        Lopes.
2452    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2454        (1) The include/exclude patterns were being applied to the whole pathnames
2455            of files, instead of just to the final components.
2457        (2) If there was more than one level of directory, the subdirectories were
2458            skipped unless they satisfied the include/exclude conditions. This is
2459            inconsistent with GNU grep (and could even be seen as contrary to the
2460            pcregrep specification - which I improved to make it absolutely clear).
2461            The action now is always to scan all levels of directory, and just
2462            apply the include/exclude patterns to regular files.
2464    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2465        --exclude_dir in the tests to avoid scanning .svn directories.
2467    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2468        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2469        doesn't support NULs in patterns.
2471    7.  Added some missing "const"s to declarations of static tables in
2472        pcre_compile.c and pcre_dfa_exec.c.
2474    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2475        caused by fix #2  above. (Subsequently also a second patch to fix the
2476        first patch. And a third patch - this was a messy problem.)
2478    9.  Applied Craig's patch to remove the use of push_back().
2480    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2481        matching function regexec().
2483    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2484        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2485        references. PCRE supports relative numbers with this syntax (I don't think
2486        Oniguruma does).
2488    12. Previously, a group with a zero repeat such as (...){0} was completely
2489        omitted from the compiled regex. However, this means that if the group
2490        was called as a subroutine from elsewhere in the pattern, things went wrong
2491        (an internal error was given). Such groups are now left in the compiled
2492        pattern, with a new opcode that causes them to be skipped at execution
2493        time.
2495    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2496        to the way PCRE behaves:
2498        (a) A lone ] character is dis-allowed (Perl treats it as data).
2500        (b) A back reference to an unmatched subpattern matches an empty string
2501            (Perl fails the current match path).
2503        (c) A data ] in a character class must be notated as \] because if the
2504            first data character in a class is ], it defines an empty class. (In
2505            Perl it is not possible to have an empty class.) The empty class []
2506            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2507            The negative empty class [^] matches any one character, independently
2508            of the DOTALL setting.
2510    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2511        non-existent subpattern following a character class starting with ']' and
2512        containing () gave an internal compiling error instead of "reference to
2513        non-existent subpattern". Fortunately, when the pattern did exist, the
2514        compiled code was correct. (When scanning forwards to check for the
2515        existence of the subpattern, it was treating the data ']' as terminating
2516        the class, so got the count wrong. When actually compiling, the reference
2517        was subsequently set up correctly.)
2519    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2520        it was being rejected as not supported by pcre_dfa_exec(), even though
2521        other assertions are supported. I have made pcre_dfa_exec() support
2522        (*FAIL).
2524    16. The implementation of 13c above involved the invention of a new opcode,
2525        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2526        cannot be changed at match time, I realized I could make a small
2527        improvement to matching performance by compiling OP_ALLANY instead of
2528        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2529        on the OP_ANY path.
2531    17. Compiling pcretest on Windows with readline support failed without the
2532        following two fixes: (1) Make the unistd.h include conditional on
2533        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2535    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2536        ncurses library to be included for pcretest when ReadLine support is
2537        requested, but also to allow for it to be overridden. This patch came from
2538        Daniel Bergström.
2540    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2541        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2542        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2543        spotting this.
2546    Version 7.6 28-Jan-08
2547    ---------------------
2549    1.  A character class containing a very large number of characters with
2550        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2551        overflow.
2553    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2554        HAVE_LONG_LONG is not defined.
2556    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2557        bring them up to date and include new features. This patch includes:
2559        - Fixed PH's badly added libz and libbz2 support.
2560        - Fixed a problem with static linking.
2561        - Added pcredemo. [But later removed - see 7 below.]
2562        - Fixed dftables problem and added an option.
2563        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2564            HAVE_LONG_LONG.
2565        - Added readline support for pcretest.
2566        - Added an listing of the option settings after cmake has run.
2568    4.  A user submitted a patch to Makefile that makes it easy to create
2569        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2570        Makefile.am that cause it to include this special target, without
2571        affecting anything else. Note that the same mingw target plus all
2572        the other distribution libraries and programs are now supported
2573        when configuring with CMake (see 6 below) instead of with
2574        Configure/Make.
2576    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2577        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2578        exported in the Windows port". It has not yet been confirmed that the patch
2579        solves the problem, but it does no harm.
2581    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2582        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2583        with CMake, and also correct the comment about stack recursion.
2585    7.  Remove the automatic building of pcredemo from the ./configure system and
2586        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2587        of a program that users should build themselves after PCRE is installed, so
2588        building it automatically is not really right. What is more, it gave
2589        trouble in some build environments.
2591    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2594    Version 7.5 10-Jan-08
2595    ---------------------
2597    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2598        values in parens when parsing an RE using the C++ wrapper."
2600    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2601        Characters greater than 255 were excluded from the class instead of being
2602        included.
2604    3.  The same bug as (2) above applied to negated POSIX classes such as
2605        [:^space:].
2607    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2608        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2609        I have changed it.
2611    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2612        first named subpattern) and a construct such as (?&a) would reference the
2613        first named subpattern whose name started with "a" (in other words, the
2614        length check was missing). Both these problems are fixed. "Subpattern name
2615        expected" is now given for (?&) (a zero-length name), and this patch also
2616        makes it give the same error for \k'' (previously it complained that that
2617        was a reference to a non-existent subpattern).
2619    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2620        this is right because (?- can be followed by option settings as well as by
2621        digits. I have, however, made the messages clearer.
2623    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2624        than the number used in the conditional) now cause a compile-time error.
2625        This is actually not compatible with Perl, which accepts such patterns, but
2626        treats the conditional as always being FALSE (as PCRE used to), but it
2627        seems to me that giving a diagnostic is better.
2629    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2630        and messages.
2632    9.  Fix two occurrences of "backslash" in comments that should have been
2633        "backspace".
2635    10. Remove two redundant lines of code that can never be obeyed (their function
2636        was moved elsewhere).
2638    11. The program that makes PCRE's Unicode character property table had a bug
2639        which caused it to generate incorrect table entries for sequences of
2640        characters that have the same character type, but are in different scripts.
2641        It amalgamated them into a single range, with the script of the first of
2642        them. In other words, some characters were in the wrong script. There were
2643        thirteen such cases, affecting characters in the following ranges:
2645          U+002b0 - U+002c1
2646          U+0060c - U+0060d
2647          U+0061e - U+00612
2648          U+0064b - U+0065e
2649          U+0074d - U+0076d
2650          U+01800 - U+01805
2651          U+01d00 - U+01d77
2652          U+01d9b - U+01dbf
2653          U+0200b - U+0200f
2654          U+030fc - U+030fe
2655          U+03260 - U+0327f
2656          U+0fb46 - U+0fbb1
2657          U+10450 - U+1049d
2659    12. The -o option (show only the matching part of a line) for pcregrep was not
2660        compatible with GNU grep in that, if there was more than one match in a
2661        line, it showed only the first of them. It now behaves in the same way as
2662        GNU grep.
2664    13. If the -o and -v options were combined for pcregrep, it printed a blank
2665        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2666        does the same. The return code can be used to tell if there were any
2667        non-matching lines.
2669    14. Added --file-offsets and --line-offsets to pcregrep.
2671    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2672        infinitely looping recursion. The bug was that positive lookaheads were not
2673        being skipped when checking for a possible empty match (negative lookaheads
2674        and both kinds of lookbehind were skipped).
2676    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2677        inclusion of <windows.h> to before rather than after the definition of
2678        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2680    17. Specifying a possessive quantifier with a specific limit for a Unicode
2681        character property caused pcre_compile() to compile bad code, which led at
2682        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2683        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2684        caused the error; without that there was no problem.
2686    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2688    19. Added --enable-pcretest-libreadline.
2690    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2691        RE::GlobalReplace(). As a result, the number of replacements returned was
2692        double what it should be. I removed one of the increments, but Craig sent a
2693        later patch that removed the other one (the right fix) and added unit tests
2694        that check the return values (which was not done before).
2696    21. Several CMake things:
2698        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2699            the names libpcre and libpcreposix, not just pcre and pcreposix.
2701        (2) The above change means that pcretest and pcregrep are now correctly
2702            linked with the newly-built libraries, not previously installed ones.
2706    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2707        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2708        UTF-8 newline character). The key issue is that the pattern starts .*;
2709        this means that the match must be either at the beginning, or after a
2710        newline. The bug was in the code for advancing after a failed match and
2711        checking that the new position followed a newline. It was not taking
2712        account of UTF-8 characters correctly.
2714    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2715        character classes. PCRE was not treating the sequence [:...:] as a
2716        character class unless the ... were all letters. Perl, however, seems to
2717        allow any characters between [: and :], though of course it rejects as
2718        unknown any "names" that contain non-letters, because all the known class
2719        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2720        for example, whereas PCRE did not - it did not recognize a POSIX character
2721        class. This seemed a bit dangerous, so the code has been changed to be
2722        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2723        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2724        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2725        Perl does, and where it didn't before.
2727    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2728        Windows environments %n is disabled by default.
2731    Version 7.4 21-Sep-07
2732    ---------------------
2734    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2735        means that a class such as [\s] counted as "explicit reference to CR or
2736        LF". That isn't really right - the whole point of the change was to try to
2737        help when there was an actual mention of one of the two characters. So now
2738        the change happens only if \r or \n (or a literal CR or LF) character is
2739        encountered.
2741    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2742        of both had grown to the point where there were only 3 bits left.
2743        Fortunately, there was spare space in the data structure, and so I have
2744        moved the internal flags into a new 16-bit field to free up more option
2745        bits.
2747    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2748        but did not set the internal JCHANGED flag - either of these is enough to
2749        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2750        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2751        start sets both bits.
2753    4.  Added options (at build time, compile time, exec time) to change \R from
2754        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2756    5.  doc/pcresyntax.html was missing from the distribution.
2758    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2759        compatibility, even though it is no longer used.
2761    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2762        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2763        windows.h file is present (where different names are used). [This was
2764        reversed later after testing - see 16 below.]
2766    8.  Changed all #include <config.h> to #include "config.h". There were also
2767        some further <pcre.h> cases that I changed to "pcre.h".
2769    9.  When pcregrep was used with the --colour option, it missed the line ending
2770        sequence off the lines that it output.
2772    10. It was pointed out to me that arrays of string pointers cause lots of
2773        relocations when a shared library is dynamically loaded. A technique of
2774        using a single long string with a table of offsets can drastically reduce
2775        these. I have refactored PCRE in four places to do this. The result is
2776        dramatic:
2778          Originally:                          290
2779          After changing UCP table:            187
2780          After changing error message table:   43
2781          After changing table of "verbs"       36
2782          After changing table of Posix names   22
2784        Thanks to the folks working on Gregex for glib for this insight.
2786    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2787        unicode-properties was also set.
2789    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2791    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2792        checked only for CRLF.
2794    14. Added casts to pcretest.c to avoid compiler warnings.
2796    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2798    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2799        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2800        entirely. This removes changes made in 7 above.
2802    17. The CMake files have been updated, and there is now more information about
2803        building with CMake in the NON-UNIX-USE document.
2806    Version 7.3 28-Aug-07
2807    ---------------------
2809     1. In the rejigging of the build system that eventually resulted in 7.1, the
2810        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2811        brackets there is not right, since it causes compilers to look for an
2812        installed pcre.h, not the version that is in the source that is being
2813        compiled (which of course may be different). I have changed it back to:
2815          #include "pcre.h"
2817        I have a vague recollection that the change was concerned with compiling in
2818        different directories, but in the new build system, that is taken care of
2819        by the VPATH setting the Makefile.
2821     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2822        when the subject happened to end in the byte 0x85 (e.g. if the last
2823        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2824        characters but of course it shouldn't be taken as a newline when it is part
2825        of another character. The bug was that, for an unlimited repeat of . in
2826        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2827        characters when looking for a newline.
2829     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2831     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2832        in debug output.
2834     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2835        long printing in the pcrecpp unittest when running under MinGW.
2837     6. ESC_K was left out of the EBCDIC table.
2839     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2840        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2841        limit also applies to "virtual nesting" when a pattern is recursive, and in
2842        this case 1000 isn't so big. I have been able to remove this limit at the
2843        expense of backing off one optimization in certain circumstances. Normally,
2844        when pcre_exec() would call its internal match() function recursively and
2845        immediately return the result unconditionally, it uses a "tail recursion"
2846        feature to save stack. However, when a subpattern that can match an empty
2847        string has an unlimited repetition quantifier, it no longer makes this
2848        optimization. That gives it a stack frame in which to save the data for
2849        checking that an empty string has been matched. Previously this was taken
2850        from the 1000-entry workspace that had been reserved. So now there is no
2851        explicit limit, but more stack is used.
2853     8. Applied Daniel's patches to solve problems with the import/export magic
2854        syntax that is required for Windows, and which was going wrong for the
2855        pcreposix and pcrecpp parts of the library. These were overlooked when this
2856        problem was solved for the main library.
2858     9. There were some crude static tests to avoid integer overflow when computing
2859        the size of patterns that contain repeated groups with explicit upper
2860        limits. As the maximum quantifier is 65535, the maximum group length was
2861        set at 30,000 so that the product of these two numbers did not overflow a
2862        32-bit integer. However, it turns out that people want to use groups that
2863        are longer than 30,000 bytes (though not repeat them that many times).
2864        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2865        made it possible to implement the integer overflow checks in a much more
2866        dynamic way, which I have now done. The artificial limitation on group
2867        length has been removed - we now have only the limit on the total length of
2868        the compiled pattern, which depends on the LINK_SIZE setting.
2870    10. Fixed a bug in the documentation for get/copy named substring when
2871        duplicate names are permitted. If none of the named substrings are set, the
2872        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2873        empty string.
2875    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2876        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2877        because the ] is interpreted as the first data character and the
2878        terminating ] is not found. PCRE has been made compatible with Perl in this
2879        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2880        cause memory overwriting.
2882    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2883        string has been matched (to stop an infinite loop). It was not recognizing
2884        a conditional subpattern that could match an empty string if that
2885        subpattern was within another subpattern. For example, it looped when
2886        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2887        condition was not nested. This bug has been fixed.
2889    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2890        past the start of the subject in the presence of bytes with the top bit
2891        set, for example "\x8aBCD".
2893    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2894        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2896    14. Optimized (?!) to (*FAIL).
2898    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2899        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2900        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2901        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2902        does: it's just the validity check that is more restrictive.
2904    16. Inserted checks for integer overflows during escape sequence (backslash)
2905        processing, and also fixed erroneous offset values for syntax errors during
2906        backslash processing.
2908    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2909        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2911    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2912        caused an overrun.
2914    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2915        something other than just ASCII characters) inside a group that had an
2916        unlimited repeat caused a loop at compile time (while checking to see
2917        whether the group could match an empty string).
2919    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2920        [\P{Any}] did so. (Error in the code for printing property names.)
2922    21. An orphan \E inside a character class could cause a crash.
2924    22. A repeated capturing bracket such as (A)? could cause a wild memory
2925        reference during compilation.
2927    23. There are several functions in pcre_compile() that scan along a compiled
2928        expression for various reasons (e.g. to see if it's fixed length for look
2929        behind). There were bugs in these functions when a repeated \p or \P was
2930        present in the pattern. These operators have additional parameters compared
2931        with \d, etc, and these were not being taken into account when moving along
2932        the compiled data. Specifically:
2934        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2935            length.
2937        (b) An item such as \pL+ within a repeated group could cause crashes or
2938            loops.
2940        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2941            "reference to non-existent subpattern" error.
2943        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2945    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2946        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2948    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2950    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2951        character were causing crashes (broken optimization).
2953    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2954        \p or \P) caused a compile-time loop.
2956    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2957        break. For example, the unstudied pattern [\r\n]A does not match the string
2958        "\r\nA" because change 7.0/46 below moves the current point on by two
2959        characters after failing to match at the start. However, the pattern \nA
2960        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2961        the same is true. There doesn't seem any very clean way out of this, but
2962        what I have chosen to do makes the common cases work: PCRE now takes note
2963        of whether there can be an explicit match for \r or \n anywhere in the
2964        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2965        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2966        pattern has explicit CR or LF references.
2968    29. Added (*CR) etc for changing newline setting at start of pattern.
2971    Version 7.2 19-Jun-07
2972    ---------------------
2974     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2975        which is apparently normally available under Windows.
2977     2. Re-jig the pcregrep tests with different newline settings in an attempt
2978        to make them independent of the local environment's newline setting.
2980     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2982     4. Some of the "internals" tests were previously cut out when the link size
2983        was not 2, because the output contained actual offsets. The recent new
2984        "Z" feature of pcretest means that these can be cut out, making the tests
2985        usable with all link sizes.
2987     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2988        stack recursion. This gives a massive performance boost under BSD, but just
2989        a small improvement under Linux. However, it saves one field in the frame
2990        in all cases.
2992     6. Added more features from the forthcoming Perl 5.10:
2994        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2995            recursion call. It refers to the nth most recently opened parentheses.
2997        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2998            to be opened parentheses.
3000        (c) Conditions that refer to capturing parentheses can be specified
3001            relatively, for example, (?(-2)... or (?(+3)...
3003        (d) \K resets the start of the current match so that everything before
3004            is not part of it.
3006        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3008        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3009            reference syntax.
3011        (g) (?| introduces a group in which the numbering of parentheses in each
3012            alternative starts with the same number.
3014        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3016     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3019     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3020        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3021        for detecting groups that can match an empty string.
3023     9. A pattern with a very large number of alternatives (more than several
3024        hundred) was running out of internal workspace during the pre-compile
3025        phase, where pcre_compile() figures out how much memory will be needed. A
3026        bit of new cunning has reduced the workspace needed for groups with
3027        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3028        workspace instead of running out of the 4096 that are available.
3030    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3032    11. Applied patch from Google to remove an optimization that didn't quite work.
3033        The report of the bug said:
3035          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3036          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3037          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3039    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3040        it matched the wrong number of bytes.
3043    Version 7.1 24-Apr-07
3044    ---------------------
3046     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3047        that is more "standard", making use of automake and other Autotools. There
3048        is some re-arrangement of the files and adjustment of comments consequent
3049        on this.
3051     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3052        for recursive directory scanning broke on some systems because the files
3053        are not scanned in any specific order and on different systems the order
3054        was different. A call to "sort" has been inserted into RunGrepTest for the
3055        approprate test as a short-term fix. In the longer term there may be an
3056        alternative.
3058     3. I had an email from Eric Raymond about problems translating some of PCRE's
3059        man pages to HTML (despite the fact that I distribute HTML pages, some
3060        people do their own conversions for various reasons). The problems
3061        concerned the use of low-level troff macros .br and .in. I have therefore
3062        removed all such uses from the man pages (some were redundant, some could
3063        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3064        HTML has been updated to handle .nf/.fi and to complain if it encounters
3065        .br or .in.
3067     4. Updated comments in configure.ac that get placed in config.h.in and also
3068        arranged for config.h to be included in the distribution, with the name
3069        config.h.generic, for the benefit of those who have to compile without
3070        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3072     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3073        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3074        makevp.bat for the current PCRE, using the additional files
3075        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3077     6. A Windows user reported a minor discrepancy with test 2, which turned out
3078        to be caused by a trailing space on an input line that had got lost in his
3079        copy. The trailing space was an accident, so I've just removed it.
3081     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3082        that is needed.
3084     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3085        as "const" (a) because they are and (b) because it helps the PHP
3086        maintainers who have recently made a script to detect big data structures
3087        in the php code that should be moved to the .rodata section. I remembered
3088        to update Builducptable as well, so it won't revert if ucptable.h is ever
3089        re-created.
3091     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3092        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3093        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3094        support is not required. This saves 1.5-2K of code, which is important in
3095        some applications.
3097        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3098        so as not to refer to the tables, even though these functions will never be
3099        called when UTF-8 support is disabled. Otherwise there are problems with a
3100        shared library.
3102    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3104        (a) It was defining its arguments as char * instead of void *.
3106        (b) It was assuming that all moves were upwards in memory; this was true
3107            a long time ago when I wrote it, but is no longer the case.
3109        The emulated memove() is provided for those environments that have neither
3110        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3111        is clearly not the case, as these two bugs were recently reported.
3113    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3114        and Detrail to create the HTML documentation, the .txt form of the man
3115        pages, and it removes trailing spaces from listed files. It also creates
3116        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3117        case, it wraps all the #defines with #ifndefs. This script should be run
3118        before "make dist".
3120    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3121        with Unicode property support.
3123        (a) For a maximizing quantifier, if the two different cases of the
3124            character were of different lengths in their UTF-8 codings (there are
3125            some cases like this - I found 11), and the matching function had to
3126            back up over a mixture of the two cases, it incorrectly assumed they
3127            were both the same length.
3129        (b) When PCRE was configured to use the heap rather than the stack for
3130            recursion during matching, it was not correctly preserving the data for
3131            the other case of a UTF-8 character when checking ahead for a match
3132            while processing a minimizing repeat. If the check also involved
3133            matching a wide character, but failed, corruption could cause an
3134            erroneous result when trying to check for a repeat of the original
3135            character.
3137    13. Some tidying changes to the testing mechanism:
3139        (a) The RunTest script now detects the internal link size and whether there
3140            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3141            values substituted by "configure". (The RunGrepTest script already did
3142            this for UTF-8.) The configure.ac script no longer substitutes the
3143            relevant variables.
3145        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3146            with length and offset values. This means that the output is different
3147            for different internal link sizes. Test 2 is skipped for link sizes
3148            other than 2 because of this, bypassing the problem. Unfortunately,
3149            there was also a test in test 3 (the locale tests) that used /B and
3150            failed for link sizes other than 2. Rather than cut the whole test out,
3151            I have added a new /Z option to pcretest that replaces the length and
3152            offset values with spaces. This is now used to make test 3 independent
3153            of link size. (Test 2 will be tidied up later.)
3155    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3156        segmentation fault instead of returning the appropriate error message.
3158    15. In multiline mode when the newline sequence was set to "any", the pattern
3159        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3160        This doesn't seem right; it now treats the CRLF combination as the line
3161        ending, and so does not match in that case. It's only a pattern such as ^$
3162        that would hit this one: something like ^ABC$ would have failed after \r
3163        and then tried again after \r\n.
3165    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3166        in an attempt to make files that differ only in their line terminators
3167        compare equal. This works on Linux.
3169    17. Under certain error circumstances pcregrep might try to free random memory
3170        as it exited. This is now fixed, thanks to valgrind.
3172    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3173        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3174        was because its rules for how to advance for /g after matching an empty
3175        string at the end of a line did not allow for this case. They now check for
3176        it specially.
3178    20. pcretest is supposed to handle patterns and data of any length, by
3179        extending its buffers when necessary. It was getting this wrong when the
3180        buffer for a data line had to be extended.
3182    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3183        CRLF as a newline sequence.
3185    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3186        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3187        I have nevertheless tidied it up.
3189    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3191    24. Added a man page for pcre-config.
3194    Version 7.0 19-Dec-06
3195    ---------------------
3197     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3198        moving to gcc 4.1.1.
3200     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3201        sys/time.h, which is documented as needed for this function. It doesn't
3202        seem to matter on Linux, but it showed up on some releases of OS X.
3204     3. It seems that there are systems where bytes whose values are greater than
3205        127 match isprint() in the "C" locale. The "C" locale should be the
3206        default when a C program starts up. In most systems, only ASCII printing
3207        characters match isprint(). This difference caused the output from pcretest
3208        to vary, making some of the tests fail. I have changed pcretest so that:
3210        (a) When it is outputting text in the compiled version of a pattern, bytes
3211            other than 32-126 are always shown as hex escapes.
3213        (b) When it is outputting text that is a matched part of a subject string,
3214            it does the same, unless a different locale has been set for the match
3215            (using the /L modifier). In this case, it uses isprint() to decide.
3217     4. Fixed a major bug that caused incorrect computation of the amount of memory
3218        required for a compiled pattern when options that changed within the
3219        pattern affected the logic of the preliminary scan that determines the
3220        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3221        that the computed length was too small. The symptoms of this bug were
3222        either the PCRE error "internal error: code overflow" from pcre_compile(),
3223        or a glibc crash with a message such as "pcretest: free(): invalid next
3224        size (fast)". Examples of patterns that provoked this bug (shown in
3225        pcretest format) are:
3227          /(?-x: )/x
3228          /(?x)(?-x: \s*#\s*)/
3229          /((?i)[\x{c0}])/8
3230          /(?i:[\x{c0}])/8
3232        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3233        is now done differently.
3235     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3236        wrapper classes; (b) implement a new function in the C++ scanner that is
3237        more efficient than the old way of doing things because it avoids levels of
3238        recursion in the regex matching; (c) add a paragraph to the documentation
3239        for the FullMatch() function.
3241     6. The escape sequence \n was being treated as whatever was defined as
3242        "newline". Not only was this contrary to the documentation, which states
3243        that \n is character 10 (hex 0A), but it also went horribly wrong when
3244        "newline" was defined as CRLF. This has been fixed.
3246     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3247        was being set to -1 for the "end of line" case (supposedly a value that no
3248        character can have). Though this value is never used (the check for end of
3249        line is "zero bytes in current character"), it caused compiler complaints.
3250        I've changed it to 0xffffffff.
3252     8. In pcre_version.c, the version string was being built by a sequence of
3253        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3254        string (as it is for production releases) called a macro with an empty
3255        argument. The C standard says the result of this is undefined. The gcc
3256        compiler treats it as an empty string (which was what was wanted) but it is
3257        reported that Visual C gives an error. The source has been hacked around to
3258        avoid this problem.
3260     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3261        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3262        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3263        of them did).
3265    10. Originally, pcretest opened its input and output without "b"; then I was
3266        told that "b" was needed in some environments, so it was added for release
3267        5.0 to both the input and output. (It makes no difference on Unix-like
3268        systems.) Later I was told that it is wrong for the input on Windows. I've
3269        now abstracted the modes into two macros, to make it easier to fiddle with
3270        them, and removed "b" from the input mode under Windows.
3272    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3274    12. Added -help and --help to pcretest as an official way of being reminded
3275        of the options.
3277    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3278        and pcrecpp.cc because they annoy compilers at high warning levels.
3280    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3282    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3283        scripts are not C programs :-) and which was not noticed because it works
3284        on Linux.
3286    16. pcretest is supposed to handle any length of pattern and data line (as one
3287        line or as a continued sequence of lines) by extending its input buffer if
3288        necessary. This feature was broken for very long pattern lines, leading to
3289        a string of junk being passed to pcre_compile() if the pattern was longer
3290        than about 50K.
3292    17. I have done a major re-factoring of the way pcre_compile() computes the
3293        amount of memory needed for a compiled pattern. Previously, there was code
3294        that made a preliminary scan of the pattern in order to do this. That was
3295        OK when PCRE was new, but as the facilities have expanded, it has become
3296        harder and harder to keep it in step with the real compile phase, and there
3297        have been a number of bugs (see for example, 4 above). I have now found a
3298        cunning way of running the real compile function in a "fake" mode that
3299        enables it to compute how much memory it would need, while actually only
3300        ever using a few hundred bytes of working memory and without too many
3301        tests of the mode. This should make future maintenance and development
3302        easier. A side effect of this work is that the limit of 200 on the nesting
3303        depth of parentheses has been removed (though this was never a serious
3304        limitation, I suspect). However, there is a downside: pcre_compile() now
3305        runs more slowly than before (30% or more, depending on the pattern). I
3306        hope this isn't a big issue. There is no effect on runtime performance.
3308    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3309        newline (only possible for the last line of a file) and it was a
3310        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3312    19. Added additional timing features to pcretest. (1) The -tm option now times
3313        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3314        separate command line item, by a number that specifies the number of
3315        repeats to use when timing. The default is 50000; this gives better
3316        precision, but takes uncomfortably long for very large patterns.
3318    20. Extended pcre_study() to be more clever in cases where a branch of a
3319        subpattern has no definite first character. For example, (a*|b*)[cd] would
3320        previously give no result from pcre_study(). Now it recognizes that the
3321        first character must be a, b, c, or d.
3323    21. There was an incorrect error "recursive call could loop indefinitely" if
3324        a subpattern (or the entire pattern) that was being tested for matching an
3325        empty string contained only one non-empty item after a nested subpattern.
3326        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3327        incorrectly, because the \d was being skipped in the check.
3329    22. The pcretest program now has a new pattern option /B and a command line
3330        option -b, which is equivalent to adding /B to every pattern. This causes
3331        it to show the compiled bytecode, without the additional information that
3332        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3333        is the same as /B/I).
3335    23. A new optimization is now able automatically to treat some sequences such
3336        as a*b as a*+b. More specifically, if something simple (such as a character
3337        or a simple class like \d) has an unlimited quantifier, and is followed by
3338        something that cannot possibly match the quantified thing, the quantifier
3339        is automatically "possessified".
3341    24. A recursive reference to a subpattern whose number was greater than 39
3342        went wrong under certain circumstances in UTF-8 mode. This bug could also
3343        have affected the operation of pcre_study().
3345    25. Realized that a little bit of performance could be had by replacing
3346        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3348    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3350    27. Possessive quantifiers such as a++ were previously implemented by turning
3351        them into atomic groups such as ($>a+). Now they have their own opcodes,
3352        which improves performance. This includes the automatically created ones
3353        from 23 above.
3355    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3356        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3357        the first matched character to be a colon. This applied both to named and
3358        numbered groups.
3360    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3362    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3363        building PCRE on Windows easier, so I have included it in the distribution.
3365    31. There is now a check in pcretest against a ridiculously large number being
3366        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3367        loop, the loop is abandoned.
3369    32. Forward references to subpatterns in conditions such as (?(2)...) where
3370        subpattern 2 is defined later cause pcre_compile() to search forwards in
3371        the pattern for the relevant set of parentheses. This search went wrong
3372        when there were unescaped parentheses in a character class, parentheses
3373        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3375    33. "Subroutine" calls and backreferences were previously restricted to
3376        referencing subpatterns earlier in the regex. This restriction has now
3377        been removed.
3379    34. Added a number of extra features that are going to be in Perl 5.10. On the
3380        whole, these are just syntactic alternatives for features that PCRE had
3381        previously implemented using the Python syntax or my own invention. The
3382        other formats are all retained for compatibility.
3384        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3385            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3386            also .NET compatible.
3388        (b) A recursion or subroutine call to a named group can now be defined as
3389            (?&name) as well as (?P>name).
3391        (c) A backreference to a named group can now be defined as \k<name> or
3392            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3393            5.10, are also .NET compatible.
3395        (d) A conditional reference to a named group can now use the syntax
3396            (?(<name>) or (?('name') as well as (?(name).
3398        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3399            groups (named and numbered) that are never evaluated inline, but can be
3400            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3401            is always false. There may be only one alternative in such a group.
3403        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3404            as the simple (?(R). The condition is true only if the most recent
3405            recursion is that of the given number or name. It does not search out
3406            through the entire recursion stack.
3408        (g) The escape \gN or \g{N} has been added, where N is a positive or
3409            negative number, specifying an absolute or relative reference.
3411    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3412        some "unreachable code" warnings.
3414    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3415        things, this adds five new scripts.
3417    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3418        There were also incompatibilities regarding the handling of \Q..\E inside
3419        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3420        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3422    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3423        matches an empty string, and forcibly breaks the loop. There were bugs in
3424        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3425        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3426        separate and independent bugs (that affected different cases) have been
3427        fixed.
3429    39. Refactored the code to abolish the use of different opcodes for small
3430        capturing bracket numbers. This is a tidy that I avoided doing when I
3431        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3432        The new approach is not only tidier, it makes it possible to reduce the
3433        memory needed to fix the previous bug (38).
3435    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3436        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3437        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3438        mode.
3440    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3441        report.
3443    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3444        copy construction and assignment in the C++ wrapper.
3446    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3447        couple of bugs that could have given wrong results in the "--newline=crlf"
3448        case.
3450    44. Added a number of casts and did some reorganization of signed/unsigned int
3451        variables following suggestions from Dair Grant. Also renamed the variable
3452        "this" as "item" because it is a C++ keyword.
3454    45. Arranged for dftables to add
3456          #include "pcre_internal.h"
3458        to pcre_chartables.c because without it, gcc 4.x may remove the array
3459        definition from the final binary if PCRE is built into a static library and
3460        dead code stripping is activated.
3462    46. For an unanchored pattern, if a match attempt fails at the start of a
3463        newline sequence, and the newline setting is CRLF or ANY, and the next two
3464        characters are CRLF, advance by two characters instead of one.
3467    Version 6.7 04-Jul-06
3468    ---------------------
3470     1. In order to handle tests when input lines are enormously long, pcretest has
3471        been re-factored so that it automatically extends its buffers when
3472        necessary. The code is crude, but this _is_ just a test program. The
3473        default size has been increased from 32K to 50K.
3475     2. The code in pcre_study() was using the value of the re argument before
3476        testing it for NULL. (Of course, in any sensible call of the function, it
3477        won't be NULL.)
3479     3. The memmove() emulation function in pcre_internal.h, which is used on
3480        systems that lack both memmove() and bcopy() - that is, hardly ever -
3481        was missing a "static" storage class specifier.
3483     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3484        containing an extended class (one that cannot be represented by a bitmap
3485        because it contains high-valued characters or Unicode property items, e.g.
3486        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3487        pattern, but PCRE should not loop if you do not (it no longer does).
3488        [Detail: two cases were found: (a) a repeated subpattern containing an
3489        extended class; (b) a recursive reference to a subpattern that followed a
3490        previous extended class. It wasn't skipping over the extended class
3491        correctly when UTF-8 mode was not set.]
3493     5. A negated single-character class was not being recognized as fixed-length
3494        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3495        compile error "lookbehind assertion is not fixed length".
3497     6. The RunPerlTest auxiliary script was showing an unexpected difference
3498        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3499        write a Perl script that can interpret lines of an input file either as
3500        byte characters or as UTF-8, which is what "perltest" was being required to
3501        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3502        can't do is switch easily at run time between having the "use utf8;" pragma
3503        or not. In the end, I fudged it by using the RunPerlTest script to insert
3504        "use utf8;" explicitly for the UTF-8 tests.
3506     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3507        the end of the subject string, contrary to the documentation and to what
3508        Perl does. This was true of both matching functions. Now it matches only at
3509        the start of the subject and immediately after *internal* newlines.
3511     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3512        a pointer to an int instead of a pointer to an unsigned long int. This
3513        caused problems on 64-bit systems.
3515     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3516        instance of the 'standard' template library not being so standard".
3518    10. There was no check on the number of named subpatterns nor the maximum
3519        length of a subpattern name. The product of these values is used to compute
3520        the size of the memory block for a compiled pattern. By supplying a very
3521        long subpattern name and a large number of named subpatterns, the size
3522        computation could be caused to overflow. This is now prevented by limiting
3523        the length of names to 32 characters, and the number of named subpatterns
3524        to 10,000.
3526    11. Subpatterns that are repeated with specific counts have to be replicated in
3527        the compiled pattern. The size of memory for this was computed from the
3528        length of the subpattern and the repeat count. The latter is limited to
3529        65535, but there was no limit on the former, meaning that integer overflow
3530        could in principle occur. The compiled length of a repeated subpattern is
3531        now limited to 30,000 bytes in order to prevent this.
3533    12. Added the optional facility to have named substrings with the same name.
3535    13. Added the ability to use a named substring as a condition, using the
3536        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3537        are numbers (not recommended). Forward references are permitted.
3539    14. Added forward references in named backreferences (if you see what I mean).
3541    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3542        pattern could run off the end of the subject. For example, the pattern
3543        "(?s)(.{1,5})"8 did this with the subject "ab".
3545    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3546        PCRE_CASELESS was set when matching characters that were quantified with ?
3547        or *.
3549    17. A character class other than a single negated character that had a minimum
3550        but no maximum quantifier - for example [ab]{6,} - was not handled
3551        correctly by pce_dfa_exec(). It would match only one character.
3553    18. A valid (though odd) pattern that looked like a POSIX character
3554        class but used an invalid character after [ (for example [[,abc,]]) caused
3555        pcre_compile() to give the error "Failed: internal error: code overflow" or
3556        in some cases to crash with a glibc free() error. This could even happen if
3557        the pattern terminated after [[ but there just happened to be a sequence of
3558        letters, a binary zero, and a closing ] in the memory that followed.
3560    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3561        over the years. Originally (before any Unicode support), just the bottom 8
3562        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3563        output from "man perlunicode" includes this:
3565          The regular expression compiler produces polymorphic opcodes.  That
3566          is, the pattern adapts to the data and automatically switches to
3567          the Unicode character scheme when presented with Unicode data--or
3568          instead uses a traditional byte scheme when presented with byte
3569          data.
3571        Sadly, a wide octal escape does not cause a switch, and in a string with
3572        no other multibyte characters, these octal escapes are treated as before.
3573        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3574        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3575        Unicode string.
3577        I have not perpetrated such confusion in PCRE. Up till now, it took just
3578        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3579        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3580        translate to the appropriate multibyte character.
3582    29. Applied some refactoring to reduce the number of warnings from Microsoft
3583        and Borland compilers. This has included removing the fudge introduced
3584        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3585        a warning about an unused variable.
3587    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3588        characters since release 4.0, because Perl (from release 5.004) does not.
3589        [Or at least, is documented not to: some releases seem to be in conflict
3590        with the documentation.] However, when a pattern was studied with
3591        pcre_study() and all its branches started with \s, PCRE still included VT
3592        as a possible starting character. Of course, this did no harm; it just
3593        caused an unnecessary match attempt.
3595    22. Removed a now-redundant internal flag bit that recorded the fact that case
3596        dependency changed within the pattern. This was once needed for "required
3597        byte" processing, but is no longer used. This recovers a now-scarce options
3598        bit. Also moved the least significant internal flag bit to the most-
3599        significant bit of the word, which was not previously used (hangover from
3600        the days when it was an int rather than a uint) to free up another bit for
3601        the future.
3603    23. Added support for CRLF line endings as well as CR and LF. As well as the
3604        default being selectable at build time, it can now be changed at runtime
3605        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3606        specify that it is scanning data with non-default line endings.
3608    24. Changed the definition of CXXLINK to make it agree with the definition of
3609        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3611    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3612        recursions. This makes a big different to stack usage for some patterns.
3614    26. If a subpattern containing a named recursion or subroutine reference such
3615        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3616        the space required for the compiled pattern went wrong and gave too small a
3617        value. Depending on the environment, this could lead to "Failed: internal
3618        error: code overflow at offset 49" or "glibc detected double free or
3619        corruption" errors.
3621    27. Applied patches from Google (a) to support the new newline modes and (b) to
3622        advance over multibyte UTF-8 characters in GlobalReplace.
3624    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3625        difference for some implementation of PCRE in some Windows version.
3627    29. Added some extra testing facilities to pcretest:
3629        \q<number>   in a data line sets the "match limit" value
3630        \Q<number>   in a data line sets the "match recursion limt" value
3631        -S <number>  sets the stack size, where <number> is in megabytes
3633        The -S option isn't available for Windows.
3636    Version 6.6 06-Feb-06
3637    ---------------------
3639     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3640        in pcreposix.h. I have copied the definition from pcre.h.
3642     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3643        because pcre.h is no longer a built file.
3645     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3646        not normally included in the compiled code.
3649    Version 6.5 01-Feb-06
3650    ---------------------
3652     1. When using the partial match feature with pcre_dfa_exec(), it was not
3653        anchoring the second and subsequent partial matches at the new starting
3654        point. This could lead to incorrect results. For example, with the pattern
3655        /1234/, partially matching against "123" and then "a4" gave a match.
3657     2. Changes to pcregrep:
3659        (a) All non-match returns from pcre_exec() were being treated as failures
3660            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3661            error message is output. Some extra information is given for the
3662            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3663            probably the only errors that are likely to be caused by users (by
3664            specifying a regex that has nested indefinite repeats, for instance).
3665            If there are more than 20 of these errors, pcregrep is abandoned.
3667        (b) A binary zero was treated as data while matching, but terminated the
3668            output line if it was written out. This has been fixed: binary zeroes
3669            are now no different to any other data bytes.
3671        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3672            used to set a locale for matching. The --locale=xxxx long option has
3673            been added (no short equivalent) to specify a locale explicitly on the
3674            pcregrep command, overriding the environment variables.
3676        (d) When -B was used with -n, some line numbers in the output were one less
3677            than they should have been.
3679        (e) Added the -o (--only-matching) option.
3681        (f) If -A or -C was used with -c (count only), some lines of context were
3682            accidentally printed for the final match.
3684        (g) Added the -H (--with-filename) option.
3686        (h) The combination of options -rh failed to suppress file names for files
3687            that were found from directory arguments.
3689        (i) Added the -D (--devices) and -d (--directories) options.
3691        (j) Added the -F (--fixed-strings) option.
3693        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3695        (l) Added the --colo(u)r option.
3697        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3698            is not present by default.
3700     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3701        items such as (?R) or (?1), when the recursion could match a number of
3702        alternatives. If it matched one of the alternatives, but subsequently,
3703        outside the recursion, there was a failure, the code tried to back up into
3704        the recursion. However, because of the way PCRE is implemented, this is not
3705        possible, and the result was an incorrect result from the match.
3707        In order to prevent this happening, the specification of recursion has
3708        been changed so that all such subpatterns are automatically treated as
3709        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3711     4. I had overlooked the fact that, in some locales, there are characters for
3712        which isalpha() is true but neither isupper() nor islower() are true. In
3713        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3714        and ordfeminine) are like this. This affected the treatment of \w and \W
3715        when they appeared in character classes, but not when they appeared outside
3716        a character class. The bit map for "word" characters is now created
3717        separately from the results of isalnum() instead of just taking it from the
3718        upper, lower, and digit maps. (Plus the underscore character, of course.)
3720     5. The above bug also affected the handling of POSIX character classes such as
3721        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3722        permanent tables. Instead, the bit maps for such a class were previously
3723        created as the appropriate unions of the upper, lower, and digit bitmaps.
3724        Now they are created by subtraction from the [[:word:]] class, which has
3725        its own bitmap.
3727     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3728        It is created by subtracting the vertical space characters (\x09, \x0a,
3729        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3730        subtraction was done in the overall bitmap for a character class, meaning
3731        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3732        be recognized. This bug has been fixed.
3734     7. Patches from the folks at Google:
3736          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3737          real life, but is still worth protecting against".
3739          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3740          regular expressions".
3742          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3743          have it.
3745          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3746          "configure" and the latter not, in order to fix a problem somebody had
3747          with compiling the Arg class on HP-UX.
3749          (e) Improve the error-handling of the C++ wrapper a little bit.
3751          (f) New tests for checking recursion limiting.
3753     8. The pcre_memmove() function, which is used only if the environment does not
3754        have a standard memmove() function (and is therefore rarely compiled),
3755        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3756        returning a result (though PCRE never actually uses the result).
3758     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3759        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3760        returned instead of calling malloc() with an overflowing number that would
3761        most likely cause subsequent chaos.
3763    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3765    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3766        with this option is matched, the nmatch and pmatch options of regexec() are
3767        ignored.
3769    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3770        provided in case anyone wants to the the POSIX interface with UTF-8
3771        strings.
3773    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3774        C++ linking (needed for some HP-UX environments).
3776    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3777        (unused parameter) and in the pcre_printint() function (omitted "default"
3778        switch label when the default is to do nothing).
3780    15. Added some code to make it possible, when PCRE is compiled as a C++
3781        library, to replace subject pointers for pcre_exec() with a smart pointer
3782        class, thus making it possible to process discontinuous strings.
3784    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3785        much the same function. They were added by different people who were trying
3786        to make PCRE easy to compile on non-Unix systems. It has been suggested
3787        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3788        for compiling on Windows systems. I have therefore rep