/[pcre]/code/trunk/ChangeLog
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 1594 by ph10, Fri Aug 21 16:08:33 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.38 xx-xxx-xxxx
8    ------------------------
9    
10    1.  If a group that contained a recursive back reference also contained a
11        forward reference subroutine call followed by a non-forward-reference
12        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13        compile correct code, leading to undefined behaviour or an internally
14        detected error. This bug was discovered by the LLVM fuzzer.
15    
16    2.  Quantification of certain items (e.g. atomic back references) could cause
17        incorrect code to be compiled when recursive forward references were
18        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19        This bug was discovered by the LLVM fuzzer.
20    
21    3.  A repeated conditional group whose condition was a reference by name caused
22        a buffer overflow if there was more than one group with the given name.
23        This bug was discovered by the LLVM fuzzer.
24    
25    4.  A recursive back reference by name within a group that had the same name as
26        another group caused a buffer overflow. For example:
27        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28    
29    5.  A forward reference by name to a group whose number is the same as the
30        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31        a buffer overflow at compile time. This bug was discovered by the LLVM
32        fuzzer.
33    
34    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
35        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36    
37    7.  Another buffer overflow bug involved duplicate named groups with a
38        reference between their definition, with a group that reset capture
39        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40        fixed by always allowing for more memory, even if not needed. (A proper fix
41        is implemented in PCRE2, but it involves more refactoring.)
42    
43    8.  There was no check for integer overflow in subroutine calls such as (?123).
44    
45    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
46        being treated as a literal 'l' instead of causing an error.
47    
48    10. There was a buffer overflow if pcre_exec() was called with an ovector of
49        size 1. This bug was found by american fuzzy lop.
50    
51    11. If a non-capturing group containing a conditional group that could match
52        an empty string was repeated, it was not identified as matching an empty
53        string itself. For example: /^(?:(?(1)x|)+)+$()/.
54    
55    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
56        \a and \e in test subject lines.
57    
58    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
59        instead of the EBCDIC value.
60    
61    14. The handling of \c in an EBCDIC environment has been revised so that it is
62        now compatible with the specification in Perl's perlebcdic page.
63    
64    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
65        ASCII/Unicode. This has now been added to the list of characters that are
66        recognized as white space in EBCDIC.
67    
68    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
69        error (correctly) when used outside a class, but did not give an error
70        within a class.
71    
72    17. \h within a class was incorrectly compiled in EBCDIC environments.
73    
74    18. A pattern with an unmatched closing parenthesis that contained a backward
75        assertion which itself contained a forward reference caused buffer
76        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
77    
78    19. JIT should return with error when the compiled pattern requires more stack
79        space than the maximum.
80    
81    20. A possessively repeated conditional group that could match an empty string,
82        for example, /(?(R))*+/, was incorrectly compiled.
83    
84    21. Fix infinite recursion in the JIT compiler when certain patterns such as
85        /(?:|a|){100}x/ are analysed.
86    
87    22. Some patterns with character classes involving [: and \\ were incorrectly
88        compiled and could cause reading from uninitialized memory or an incorrect
89        error diagnosis.
90    
91    23. Pathological patterns containing many nested occurrences of [: caused
92        pcre_compile() to run for a very long time.
93    
94    24. A conditional group with only one branch has an implicit empty alternative
95        branch and must therefore be treated as potentially matching an empty
96        string.
97    
98    25. If (?R was followed by - or + incorrect behaviour happened instead of a
99        diagnostic.
100    
101    26. Arrange to give up on finding the minimum matching length for overly
102        complex patterns.
103    
104    27. Similar to (4) above: in a pattern with duplicated named groups and an
105        occurrence of (?| it is possible for an apparently non-recursive back
106        reference to become recursive if a later named group with the relevant
107        number is encountered. This could lead to a buffer overflow. Wen Guanxing
108        from Venustech ADLAB discovered this bug.
109    
110    28. If pcregrep was given the -q option with -c or -l, or when handling a
111        binary file, it incorrectly wrote output to stdout.
112    
113    29. The JIT compiler did not restore the control verb head in case of *THEN
114        control verbs. This issue was found by Karl Skomski with a custom LLVM
115        fuzzer.
116    
117    30. Error messages for syntax errors following \g and \k were giving inaccurate
118        offsets in the pattern.
119    
120    31. Added a check for integer overflow in conditions (?(<digits>) and
121        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
122        fuzzer.
123    
124    32. Handling recursive references such as (?2) when the reference is to a group
125        later in the pattern uses code that is very hacked about and error-prone.
126        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
127        give an internal error if it is obvious that compiling has gone wrong.
128    
129    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
130        This issue was found by Karl Skomski with a custom LLVM fuzzer.
131    
132    34. The JIT compiler should restore the control chain for empty possessive
133        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
134    
135    35. Match limit check added to JIT recursion. This issue was found by Karl
136        Skomski with a custom LLVM fuzzer.
137    
138    36. Yet another case similar to 27 above has been circumvented by an
139        unconditional allocation of extra memory. This issue is fixed "properly" in
140        PCRE2 by refactoring the way references are handled. Wen Guanxing
141        from Venustech ADLAB discovered this bug.
142    
143    
144    Version 8.37 28-April-2015
145    --------------------------
146    
147    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
148        for those parentheses to be closed with whatever has been captured so far.
149        However, it was failing to mark any other groups between the hightest
150        capture so far and the currrent group as "unset". Thus, the ovector for
151        those groups contained whatever was previously there. An example is the
152        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
153    
154    2.  If an assertion condition was quantified with a minimum of zero (an odd
155        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
156    
157    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
158        unrecognized modifier, a crash could occur.
159    
160    4.  An attempt to do global matching in pcretest with a zero-length ovector
161        caused a crash.
162    
163    5.  Fixed a memory leak during matching that could occur for a subpattern
164        subroutine call (recursive or otherwise) if the number of captured groups
165        that had to be saved was greater than ten.
166    
167    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
168        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
169        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
170    
171    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
172        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
173    
174    8. When a pattern is compiled, it remembers the highest back reference so that
175       when matching, if the ovector is too small, extra memory can be obtained to
176       use instead. A conditional subpattern whose condition is a check on a
177       capture having happened, such as, for example in the pattern
178       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
179       setting the highest backreference number. This mattered only if pcre_exec()
180       was called with an ovector that was too small to hold the capture, and there
181       was no other kind of back reference (a situation which is probably quite
182       rare). The effect of the bug was that the condition was always treated as
183       FALSE when the capture could not be consulted, leading to a incorrect
184       behaviour by pcre_exec(). This bug has been fixed.
185    
186    9. A reference to a duplicated named group (either a back reference or a test
187       for being set in a conditional) that occurred in a part of the pattern where
188       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
189       to be incorrectly calculated, leading to overwriting.
190    
191    10. A mutually recursive set of back references such as (\2)(\1) caused a
192        segfault at study time (while trying to find the minimum matching length).
193        The infinite loop is now broken (with the minimum length unset, that is,
194        zero).
195    
196    11. If an assertion that was used as a condition was quantified with a minimum
197        of zero, matching went wrong. In particular, if the whole group had
198        unlimited repetition and could match an empty string, a segfault was
199        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
200        assertions to be quantified, but not if they are being used as conditions,
201        so the above pattern is faulted by Perl. PCRE has now been changed so that
202        it also rejects such patterns.
203    
204    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
205        failed to allow the zero-repeat case if pcre2_exec() was called with an
206        ovector too small to capture the group.
207    
208    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
209        Red Hat Product Security:
210    
211        (a) A crash if /K and /F were both set with the option to save the compiled
212        pattern.
213    
214        (b) Another crash if the option to print captured substrings in a callout
215        was combined with setting a null ovector, for example \O\C+ as a subject
216        string.
217    
218    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
219        forward reference repeated a large (but limited) number of times within a
220        repeated outer group that has a zero minimum quantifier, caused incorrect
221        code to be compiled, leading to the error "internal error:
222        previously-checked referenced subpattern not found" when an incorrect
223        memory address was read. This bug was reported as "heap overflow",
224        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
225        CVE-2015-2325.
226    
227    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
228        call within a group that also contained a recursive back reference caused
229        incorrect code to be compiled. This bug was reported as "heap overflow",
230        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
231        number CVE-2015-2326.
232    
233    24. Computing the size of the JIT read-only data in advance has been a source
234        of various issues, and new ones are still appear unfortunately. To fix
235        existing and future issues, size computation is eliminated from the code,
236        and replaced by on-demand memory allocation.
237    
238    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
239        adjacent to the end of the range, and the range contained characters with
240        more than one other case, caused incorrect behaviour when compiled in UTF
241        mode. In that example, the range a-j was left out of the class.
242    
243    26. Fix JIT compilation of conditional blocks, which assertion
244        is converted to (*FAIL). E.g: /(?(?!))/.
245    
246    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
247        discovered by the LLVM fuzzer.
248    
249    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
250        when this assertion was used as a condition, for example (?(?!)a|b). In
251        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
252        error about an unsupported item.
253    
254    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
255        possessification code could take exponential time to complete. A recursion
256        depth limit of 1000 has been imposed to limit the resources used by this
257        optimization.
258    
259    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
260        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
261        because \S ensures they are all in the class. The code for doing this was
262        interacting badly with the code for computing the amount of space needed to
263        compile the pattern, leading to a buffer overflow. This bug was discovered
264        by the LLVM fuzzer.
265    
266    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
267        other kinds of group caused stack overflow at compile time. This bug was
268        discovered by the LLVM fuzzer.
269    
270    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
271        between a subroutine call and its quantifier was incorrectly compiled,
272        leading to buffer overflow or other errors. This bug was discovered by the
273        LLVM fuzzer.
274    
275    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
276        assertion after (?(. The code was failing to check the character after
277        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
278        was discovered by the LLVM fuzzer.
279    
280    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
281        a fixed maximum following a group that contains a subroutine reference was
282        incorrectly compiled and could trigger buffer overflow. This bug was
283        discovered by the LLVM fuzzer.
284    
285    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
286        caused a stack overflow instead of the diagnosis of a non-fixed length
287        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
288    
289    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
290        (e.g. /(?<=\Ka)/) could make pcregrep loop.
291    
292    37. There was a similar problem to 36 in pcretest for global matches.
293    
294    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
295        and a subsequent item in the pattern caused a non-match, backtracking over
296        the repeated \X did not stop, but carried on past the start of the subject,
297        causing reference to random memory and/or a segfault. There were also some
298        other cases where backtracking after \C could crash. This set of bugs was
299        discovered by the LLVM fuzzer.
300    
301    39. The function for finding the minimum length of a matching string could take
302        a very long time if mutual recursion was present many times in a pattern,
303        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
304        method has been implemented. This infelicity was discovered by the LLVM
305        fuzzer.
306    
307    40. Static linking against the PCRE library using the pkg-config module was
308        failing on missing pthread symbols.
309    
310    
311    Version 8.36 26-September-2014
312    ------------------------------
313    
314    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
315        -Wmissing-field-initializers and -Wunused-parameter.
316    
317    2.  The tests for quantifiers being too big (greater than 65535) were being
318        applied after reading the number, and stupidly assuming that integer
319        overflow would give a negative number. The tests are now applied as the
320        numbers are read.
321    
322    3.  Tidy code in pcre_exec.c where two branches that used to be different are
323        now the same.
324    
325    4.  The JIT compiler did not generate match limit checks for certain
326        bracketed expressions with quantifiers. This may lead to exponential
327        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
328        issue should be resolved now.
329    
330    5.  Fixed an issue, which occures when nested alternatives are optimized
331        with table jumps.
332    
333    6.  Inserted two casts and changed some ints to size_t in the light of some
334        reported 64-bit compiler warnings (Bugzilla 1477).
335    
336    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
337        an empty string, which sometimes were behaving incorrectly in the
338        interpreter (though correctly in the JIT matcher). This pcretest input is
339        an example:
340    
341          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
342          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
343    
344        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
345        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
346        for an empty string was breaking the inner loop and carrying on at a lower
347        level, when possessive repeated groups should always return to a higher
348        level as they have no backtrack points in them. The empty string test now
349        occurs at the outer level.
350    
351    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
352        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
353    
354    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
355        doesn't).
356    
357    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
358        to hold a pattern to be incorrectly computed (too small) when there were
359        named back references to duplicated names. This could cause "internal
360        error: code overflow" or "double free or corruption" or other memory
361        handling errors.
362    
363    11. When named subpatterns had the same prefixes, back references could be
364        confused. For example, in this pattern:
365    
366          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
367    
368        the reference to 'Name' was incorrectly treated as a reference to a
369        duplicate name.
370    
371    12. A pattern such as /^s?c/mi8 where the optional character has more than
372        one "other case" was incorrectly compiled such that it would only try to
373        match starting at "c".
374    
375    13. When a pattern starting with \s was studied, VT was not included in the
376        list of possible starting characters; this should have been part of the
377        8.34/18 patch.
378    
379    14. If a character class started [\Qx]... where x is any character, the class
380        was incorrectly terminated at the ].
381    
382    15. If a pattern that started with a caseless match for a character with more
383        than one "other case" was studied, PCRE did not set up the starting code
384        unit bit map for the list of possible characters. Now it does. This is an
385        optimization improvement, not a bug fix.
386    
387    16. The Unicode data tables have been updated to Unicode 7.0.0.
388    
389    17. Fixed a number of memory leaks in pcregrep.
390    
391    18. Avoid a compiler warning (from some compilers) for a function call with
392        a cast that removes "const" from an lvalue by using an intermediate
393        variable (to which the compiler does not object).
394    
395    19. Incorrect code was compiled if a group that contained an internal recursive
396        back reference was optional (had quantifier with a minimum of zero). This
397        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
398        caused segmentation faults because of stack overflows at compile time.
399    
400    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
401        group that is quantified with an indefinite repeat, caused a compile-time
402        loop which used up all the system stack and provoked a segmentation fault.
403        This was not the same bug as 19 above.
404    
405    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
406        Patch by Mike Frysinger.
407    
408    
409    Version 8.35 04-April-2014
410    --------------------------
411    
412    1.  A new flag is set, when property checks are present in an XCLASS.
413        When this flag is not set, PCRE can perform certain optimizations
414        such as studying these XCLASS-es.
415    
416    2.  The auto-possessification of character sets were improved: a normal
417        and an extended character set can be compared now. Furthermore
418        the JIT compiler optimizes more character set checks.
419    
420    3.  Got rid of some compiler warnings for potentially uninitialized variables
421        that show up only when compiled with -O2.
422    
423    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
424        of a match later then the end of the match. The pcretest program was not
425        handling the case sensibly - it was outputting from the start to the next
426        binary zero. It now reports this situation in a message, and outputs the
427        text from the end to the start.
428    
429    5.  Fast forward search is improved in JIT. Instead of the first three
430        characters, any three characters with fixed position can be searched.
431        Search order: first, last, middle.
432    
433    6.  Improve character range checks in JIT. Characters are read by an inprecise
434        function now, which returns with an unknown value if the character code is
435        above a certain threshold (e.g: 256). The only limitation is that the value
436        must be bigger than the threshold as well. This function is useful when
437        the characters above the threshold are handled in the same way.
438    
439    7.  The macros whose names start with RAWUCHAR are placeholders for a future
440        mode in which only the bottom 21 bits of 32-bit data items are used. To
441        make this more memorable for those maintaining the code, the names have
442        been changed to start with UCHAR21, and an extensive comment has been added
443        to their definition.
444    
445    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
446        to the export list in Makefile.am (they were accidentally omitted from the
447        8.34 tarball).
448    
449    9.  The informational output from pcretest used the phrase "starting byte set"
450        which is inappropriate for the 16-bit and 32-bit libraries. As the output
451        for "first char" and "need char" really means "non-UTF-char", I've changed
452        "byte" to "char", and slightly reworded the output. The documentation about
453        these values has also been (I hope) clarified.
454    
455    10. Another JIT related optimization: use table jumps for selecting the correct
456        backtracking path, when more than four alternatives are present inside a
457        bracket.
458    
459    11. Empty match is not possible, when the minimum length is greater than zero,
460        and there is no \K in the pattern. JIT should avoid empty match checks in
461        such cases.
462    
463    12. In a caseless character class with UCP support, when a character with more
464        than one alternative case was not the first character of a range, not all
465        the alternative cases were added to the class. For example, s and \x{17f}
466        are both alternative cases for S: the class [RST] was handled correctly,
467        but [R-T] was not.
468    
469    13. The configure.ac file always checked for pthread support when JIT was
470        enabled. This is not used in Windows, so I have put this test inside a
471        check for the presence of windows.h (which was already tested for).
472    
473    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
474        The algorithm provides a way to skip certain starting offsets, and usually
475        faster than linear prefix searches.
476    
477    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
478        as for 'fr_FR' and 'french'. For some reason, however, it then used the
479        Windows-specific input and output files, which have 'french' screwed in.
480        So this could never have worked. One of the problems with locales is that
481        they aren't always the same. I have now updated RunTest so that it checks
482        the output of the locale test (test 3) against three different output
483        files, and it allows the test to pass if any one of them matches. With luck
484        this should make the test pass on some versions of Solaris where it was
485        failing. Because of the uncertainty, the script did not used to stop if
486        test 3 failed; it now does. If further versions of a French locale ever
487        come to light, they can now easily be added.
488    
489    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
490        there was a message during ./configure, but it did not stop. This now
491        provokes an error. The invalid example in README has been corrected.
492        If a value less than the minimum is given, the minimum value has always
493        been used, but now a warning is given.
494    
495    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
496        was a bug in the test system, which is now fixed. Also, the list of various
497        configurations that are tested for each release did not have one with both
498        16/32 bits and --enable-bar-anycrlf. It now does.
499    
500    18. pcretest was missing "-C bsr" for displaying the \R default setting.
501    
502    19. Little endian PowerPC systems are supported now by the JIT compiler.
503    
504    20. The fast forward newline mechanism could enter to an infinite loop on
505        certain invalid UTF-8 input. Although we don't support these cases
506        this issue can be fixed by a performance optimization.
507    
508    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
509        not take account if existing stack usage. There is now a new global
510        variable called pcre_stack_guard that can be set to point to an external
511        function to check stack availability. It is called at the start of
512        processing every parenthesized group.
513    
514    22. A typo in the code meant that in ungreedy mode the max/min qualifier
515        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
516        not match "ab".
517    
518    23. When UTF was disabled, the JIT program reported some incorrect compile
519        errors. These messages are silenced now.
520    
521    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
522        compiler.
523    
524    25. Change all the temporary files used in RunGrepTest to be different to those
525        used by RunTest so that the tests can be run simultaneously, for example by
526        "make -j check".
527    
528    
529    Version 8.34 15-December-2013
530    -----------------------------
531    
532    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
533        executable memory. Patch inspired by Carsten Klein.
534    
535    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
536        this macro is never tested and has no effect, because the work to support
537        coverage involves only compiling and linking options and special targets in
538        the Makefile. The comment in config.h implied that defining the macro would
539        enable coverage support, which is totally false. There was also support for
540        setting this macro in the CMake files (my fault, I just copied it from
541        configure). SUPPORT_GCOV has now been removed.
542    
543    3.  Make a small performance improvement in strlen16() and strlen32() in
544        pcretest.
545    
546    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
547        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
548        these cases). There was also one in pcretest.c.
549    
550    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
551    
552    6.  In UTF mode, the code for checking whether a group could match an empty
553        string (which is used for indefinitely repeated groups to allow for
554        breaking an infinite loop) was broken when the group contained a repeated
555        negated single-character class with a character that occupied more than one
556        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
557        UTF-8 mode). The effect was undefined: the group might or might not be
558        deemed as matching an empty string, or the program might have crashed.
559    
560    7.  The code for checking whether a group could match an empty string was not
561        recognizing that \h, \H, \v, \V, and \R must match a character.
562    
563    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
564        an empty string. If it can, pcretest shows this in its information output.
565    
566    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
567        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
568        matched by pcre_exec() without using JIT:
569    
570        (a) If the rest of the pattern did not match after a maximal run of
571            grapheme clusters, the code for backing up to try with fewer of them
572            did not always back up over a full grapheme when characters that do not
573            have the modifier quality were involved, e.g. Hangul syllables.
574    
575        (b) If the match point in a subject started with modifier character, and
576            there was no match, the code could incorrectly back up beyond the match
577            point, and potentially beyond the first character in the subject,
578            leading to a segfault or an incorrect match result.
579    
580    10. A conditional group with an assertion condition could lead to PCRE
581        recording an incorrect first data item for a match if no other first data
582        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
583        first data item, and therefore matched "ca" after "c" instead of at the
584        start.
585    
586    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
587        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
588    
589    12. The source of pcregrep now includes z/OS-specific code so that it can be
590        compiled for z/OS as part of the special z/OS distribution.
591    
592    13. Added the -T and -TM options to pcretest.
593    
594    14. The code in pcre_compile.c for creating the table of named capturing groups
595        has been refactored. Instead of creating the table dynamically during the
596        actual compiling pass, the information is remembered during the pre-compile
597        pass (on the stack unless there are more than 20 named groups, in which
598        case malloc() is used) and the whole table is created before the actual
599        compile happens. This has simplified the code (it is now nearly 150 lines
600        shorter) and prepared the way for better handling of references to groups
601        with duplicate names.
602    
603    15. A back reference to a named subpattern when there is more than one of the
604        same name now checks them in the order in which they appear in the pattern.
605        The first one that is set is used for the reference. Previously only the
606        first one was inspected. This change makes PCRE more compatible with Perl.
607    
608    16. Unicode character properties were updated from Unicode 6.3.0.
609    
610    17. The compile-time code for auto-possessification has been refactored, based
611        on a patch by Zoltan Herczeg. It now happens after instead of during
612        compilation. The code is cleaner, and more cases are handled. The option
613        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
614        options in pcretest are provided to set it. It can also be set by
615        (*NO_AUTO_POSSESS) at the start of a pattern.
616    
617    18. The character VT has been added to the default ("C" locale) set of
618        characters that match \s and are generally treated as white space,
619        following this same change in Perl 5.18. There is now no difference between
620        "Perl space" and "POSIX space". Whether VT is treated as white space in
621        other locales depends on the locale.
622    
623    19. The code for checking named groups as conditions, either for being set or
624        for being recursed, has been refactored (this is related to 14 and 15
625        above). Processing unduplicated named groups should now be as fast at
626        numerical groups, and processing duplicated groups should be faster than
627        before.
628    
629    20. Two patches to the CMake build system, by Alexander Barkov:
630    
631          (1) Replace the "source" command by "." in CMakeLists.txt because
632              "source" is a bash-ism.
633    
634          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
635              without these the CMake build does not work on Solaris.
636    
637    21. Perl has changed its handling of \8 and \9. If there is no previously
638        encountered capturing group of those numbers, they are treated as the
639        literal characters 8 and 9 instead of a binary zero followed by the
640        literals. PCRE now does the same.
641    
642    22. Following Perl, added \o{} to specify codepoints in octal, making it
643        possible to specify values greater than 0777 and also making them
644        unambiguous.
645    
646    23. Perl now gives an error for missing closing braces after \x{... instead of
647        treating the string as literal. PCRE now does the same.
648    
649    24. RunTest used to grumble if an inappropriate test was selected explicitly,
650        but just skip it when running all tests. This make it awkward to run ranges
651        of tests when one of them was inappropriate. Now it just skips any
652        inappropriate tests, as it always did when running all tests.
653    
654    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
655        character types such as \d or \w, too many callouts were inserted, and the
656        data that they returned was rubbish.
657    
658    26. In UCP mode, \s was not matching two of the characters that Perl matches,
659        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
660        were matched by \h. The code has now been refactored so that the lists of
661        the horizontal and vertical whitespace characters used for \h and \v (which
662        are defined only in one place) are now also used for \s.
663    
664    27. Add JIT support for the 64 bit TileGX architecture.
665        Patch by Jiong Wang (Tilera Corporation).
666    
667    28. Possessive quantifiers for classes (both explicit and automatically
668        generated) now use special opcodes instead of wrapping in ONCE brackets.
669    
670    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
671        (because it's meaningless), this was not happening when PCRE_CASELESS was
672        set. Not wrong, but inefficient.
673    
674    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
675        properties for \w, \d, etc) is present in a test regex. Otherwise if the
676        test contains no characters greater than 255, Perl doesn't realise it
677        should be using Unicode semantics.
678    
679    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
680        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
681        does in Unicode mode.
682    
683    32. Added the "forbid" facility to pcretest so that putting tests into the
684        wrong test files can sometimes be quickly detected.
685    
686    33. There is now a limit (default 250) on the depth of nesting of parentheses.
687        This limit is imposed to control the amount of system stack used at compile
688        time. It can be changed at build time by --with-parens-nest-limit=xxx or
689        the equivalent in CMake.
690    
691    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
692        errors. Perl warns for these when in warning mode, but PCRE has no facility
693        for giving warnings.
694    
695    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
696        However, this was not working for (?!) because it is optimized to (*FAIL),
697        for which PCRE does not allow quantifiers. The optimization is now disabled
698        when a quantifier follows (?!). I can't see any use for this, but it makes
699        things uniform.
700    
701    36. Perl no longer allows group names to start with digits, so I have made this
702        change also in PCRE. It simplifies the code a bit.
703    
704    37. In extended mode, Perl ignores spaces before a + that indicates a
705        possessive quantifier. PCRE allowed a space before the quantifier, but not
706        before the possessive +. It now does.
707    
708    38. The use of \K (reset reported match start) within a repeated possessive
709        group such as (a\Kb)*+ was not working.
710    
711    40. Document that the same character tables must be used at compile time and
712        run time, and that the facility to pass tables to pcre_exec() and
713        pcre_dfa_exec() is for use only with saved/restored patterns.
714    
715    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
716        features for Builds with MSVC:
717    
718        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
719           builds). The .rc files can be used to set FileDescription and many other
720           attributes.
721    
722        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
723           This allows higher-level build scripts which want .pdb files to avoid
724           hard-coding the exact files needed."
725    
726    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
727        mean "start of word" and "end of word", respectively, as a transition aid.
728    
729    43. A minimizing repeat of a class containing codepoints greater than 255 in
730        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
731        compiled to use the heap for recursion.
732    
733    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
734        is configured.
735    
736    
737    Version 8.33 28-May-2013
738    ------------------------
739    
740    1.  Added 'U' to some constants that are compared to unsigned integers, to
741        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
742        variables that are added to signed variables, to ensure the result is
743        signed and can be negated.
744    
745    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
746        CMake config files.
747    
748    3.  Revise the creation of config.h.generic so that all boolean macros are
749        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
750        overriding via -D on the command line possible.
751    
752    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
753        to unsigned int is reported to make a quite noticeable speed difference in
754        a specific Windows environment. Testing on Linux did also appear to show
755        some benefit (and it is clearly not harmful). Also fixed the definition of
756        Xop which should be unsigned.
757    
758    5.  Related to (4), changing the definition of the intermediate variable cc
759        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
760        improvements.
761    
762    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
763        unnecessary spaces.
764    
765    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
766        and later.
767    
768    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
769    
770    9.  Optimizing fast_forward_start_bits in JIT.
771    
772    10. Adding support for callouts in JIT, and fixing some issues revealed
773        during this work. Namely:
774    
775        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
776    
777        (b) Minimum length was not checked before the matching is started.
778    
779    11. The value of capture_last that is passed to callouts was incorrect in some
780        cases when there was a capture on one path that was subsequently abandoned
781        after a backtrack. Also, the capture_last value is now reset after a
782        recursion, since all captures are also reset in this case.
783    
784    12. The interpreter no longer returns the "too many substrings" error in the
785        case when an overflowing capture is in a branch that is subsequently
786        abandoned after a backtrack.
787    
788    13. In the pathological case when an offset vector of size 2 is used, pcretest
789        now prints out the matched string after a yield of 0 or 1.
790    
791    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
792        Only supported by the JIT compiler at the moment.
793    
794    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
795    
796    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
797        offset of the starting point of the matching process, provided the offsets
798        vector is large enough.
799    
800    17. The \A escape now records a lookbehind value of 1, though its execution
801        does not actually inspect the previous character. This is to ensure that,
802        in partial multi-segment matching, at least one character from the old
803        segment is retained when a new segment is processed. Otherwise, if there
804        are no lookbehinds in the pattern, \A might match incorrectly at the start
805        of a new segment.
806    
807    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
808    
809    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
810        gives some modest performance improvement in 8-bit mode.
811    
812    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
813        be expressed in certain programming languages using Universal Character
814        Names.
815    
816    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
817        which points out that "non characters" are not "characters that may not
818        appear in Unicode strings" but rather "characters that are reserved for
819        internal use and have only local meaning".
820    
821    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
822        there was a conditional group that depended on an assertion, if the
823        assertion was false, the callout that immediately followed the alternation
824        in the condition was skipped when pcre_exec() was used for matching.
825    
826    23. Allow an explicit callout to be inserted before an assertion that is the
827        condition for a conditional group, for compatibility with automatic
828        callouts, which always insert a callout at this point.
829    
830    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
831        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
832    
833    25. (*PRUNE) is now supported by the JIT compiler.
834    
835    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
836    
837    27. Fix the case where there are two or more SKIPs with arguments that may be
838        ignored.
839    
840    28. (*SKIP) is now supported by the JIT compiler.
841    
842    29. (*THEN) is now supported by the JIT compiler.
843    
844    30. Update RunTest with additional test selector options.
845    
846    31. The way PCRE handles backtracking verbs has been changed in two ways.
847    
848        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
849        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
850        backtracking. In some cases this makes it more Perl-compatible, but Perl's
851        rather obscure rules do not always do the same thing.
852    
853        (2) Previously, backtracking verbs were confined within assertions. This is
854        no longer the case for positive assertions, except for (*ACCEPT). Again,
855        this sometimes improves Perl compatibility, and sometimes does not.
856    
857    32. A number of tests that were in test 2 because Perl did things differently
858        have been moved to test 1, because either Perl or PCRE has changed, and
859        these tests are now compatible.
860    
861    32. Backtracking control verbs are now handled in the same way in JIT and
862        interpreter.
863    
864    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
865        contained a forward subroutine reference caused a compile error.
866    
867    34. Auto-detect and optimize limited repetitions in JIT.
868    
869    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
870        blocking (*UTF) etc.
871    
872    36. In the interpreter, maximizing pattern repetitions for characters and
873        character types now use tail recursion, which reduces stack usage.
874    
875    37. The value of the max lookbehind was not correctly preserved if a compiled
876        and saved regex was reloaded on a host of different endianness.
877    
878    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
879        of the compiled pattern block, expand the flags field from 16 to 32 bits
880        because it was almost full.
881    
882    39. Try madvise first before posix_madvise.
883    
884    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
885        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
886        original change locked out all matches of empty strings. This has been
887        changed so that one match of an empty string per line is recognized.
888        Subsequent searches on the same line (for colouring or for --only-matching,
889        for example) do not recognize empty strings.
890    
891    41. Applied a user patch to fix a number of spelling mistakes in comments.
892    
893    42. Data lines longer than 65536 caused pcretest to crash.
894    
895    43. Clarified the data type for length and startoffset arguments for pcre_exec
896        and pcre_dfa_exec in the function-specific man pages, where they were
897        explicitly stated to be in bytes, never having been updated. I also added
898        some clarification to the pcreapi man page.
899    
900    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
901        a segmentation fault.
902    
903    
904    Version 8.32 30-November-2012
905    -----------------------------
906    
907    1.  Improved JIT compiler optimizations for first character search and single
908        character iterators.
909    
910    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
911        Patch by Daniel Richard G.
912    
913    3.  Single character iterator optimizations in the JIT compiler.
914    
915    4.  Improved JIT compiler optimizations for character ranges.
916    
917    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
918        Reported by Giuseppe D'Angelo.
919    
920    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
921        of a line, was being set incorrectly in cases where .* appeared inside
922        atomic brackets at the start of a pattern, or where there was a subsequent
923        *PRUNE or *SKIP.
924    
925    7.  Improved instruction cache flush for POWER/PowerPC.
926        Patch by Daniel Richard G.
927    
928    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
929        grep:
930    
931        (a) There is now no limit to the number of patterns to be matched.
932    
933        (b) An error is given if a pattern is too long.
934    
935        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
936            are now supported.
937    
938        (d) --exclude-from and --include-from (multiple use) have been added.
939    
940        (e) Exclusions and inclusions now apply to all files and directories, not
941            just to those obtained from scanning a directory recursively.
942    
943        (f) Multiple uses of -f and --file-list are now supported.
944    
945        (g) In a Windows environment, the default for -d has been changed from
946            "read" (the GNU grep default) to "skip", because otherwise the presence
947            of a directory in the file list provokes an error.
948    
949        (h) The documentation has been revised and clarified in places.
950    
951    9.  Improve the matching speed of capturing brackets.
952    
953    10. Changed the meaning of \X so that it now matches a Unicode extended
954        grapheme cluster.
955    
956    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
957        out POSIX threads when JIT support is configured.
958    
959    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
960    
961    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
962        structure could go wrong in environments where size_t is not the same size
963        as int.
964    
965    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
966        set.
967    
968    15. The EBCDIC support had decayed; later updates to the code had included
969        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
970        general tidy up of EBCDIC-related issues, and the documentation was also
971        not quite right. There is now a test that can be run on ASCII systems to
972        check some of the EBCDIC-related things (but is it not a full test).
973    
974    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
975        in a small tidy to the code.
976    
977    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
978    
979    18. If the --only-matching (-o) option in pcregrep is specified multiple
980        times, each one causes appropriate output. For example, -o1 -o2 outputs the
981        substrings matched by the 1st and 2nd capturing parentheses. A separating
982        string can be specified by --om-separator (default empty).
983    
984    19. Improving the first n character searches.
985    
986    20. Turn case lists for horizontal and vertical white space into macros so that
987        they are defined only once.
988    
989    21. This set of changes together give more compatible Unicode case-folding
990        behaviour for characters that have more than one other case when UCP
991        support is available.
992    
993        (a) The Unicode property table now has offsets into a new table of sets of
994            three or more characters that are case-equivalent. The MultiStage2.py
995            script that generates these tables (the pcre_ucd.c file) now scans
996            CaseFolding.txt instead of UnicodeData.txt for character case
997            information.
998    
999        (b) The code for adding characters or ranges of characters to a character
1000            class has been abstracted into a generalized function that also handles
1001            case-independence. In UTF-mode with UCP support, this uses the new data
1002            to handle characters with more than one other case.
1003    
1004        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1005            whose other case is greater than 256 are now correctly matched
1006            caselessly. Previously, the high codepoint matched the low one, but not
1007            vice versa.
1008    
1009        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1010            of the new class addition function, using character lists defined as
1011            macros alongside the case definitions of 20 above.
1012    
1013        (e) Caseless back references now work with characters that have more than
1014            one other case.
1015    
1016        (f) General caseless matching of characters with more than one other case
1017            is supported.
1018    
1019    22. Unicode character properties were updated from Unicode 6.2.0
1020    
1021    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1022    
1023    24. Add support for 32-bit character strings, and UTF-32
1024    
1025    25. Major JIT compiler update (code refactoring and bugfixing).
1026        Experimental Sparc 32 support is added.
1027    
1028    26. Applied a modified version of Daniel Richard G's patch to create
1029        pcre.h.generic and config.h.generic by "make" instead of in the
1030        PrepareRelease script.
1031    
1032    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1033        pcre_compile.c when checking for a zero character.
1034    
1035    28. Introducing a native interface for JIT. Through this interface, the compiled
1036        machine code can be directly executed. The purpose of this interface is to
1037        provide fast pattern matching, so several sanity checks are not performed.
1038        However, feature tests are still performed. The new interface provides
1039        1.4x speedup compared to the old one.
1040    
1041    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1042        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1043        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1044    
1045    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1046        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1047        the "old" RFC 2279). Instead, it ended up passing a negative length to
1048        pcre_exec().
1049    
1050    31. Add support for GCC's visibility feature to hide internal functions.
1051    
1052    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1053        "unknown -C option" after outputting 0 or 1.
1054    
1055    33. There is now support for generating a code coverage report for the test
1056        suite in environments where gcc is the compiler and lcov is installed. This
1057        is mainly for the benefit of the developers.
1058    
1059    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1060        unaddressable using valgrind annotations, allowing valgrind to detect
1061        invalid memory accesses. This is mainly for the benefit of the developers.
1062    
1063    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1064    
1065    26. Give configure error if --enable-cpp but no C++ compiler found.
1066    
1067    
1068    Version 8.31 06-July-2012
1069    -------------------------
1070    
1071    1.  Fixing a wrong JIT test case and some compiler warnings.
1072    
1073    2.  Removed a bashism from the RunTest script.
1074    
1075    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1076        to unsigned type, result still unsigned" that was given by an MS compiler
1077        on encountering the code "-sizeof(xxx)".
1078    
1079    4.  Partial matching support is added to the JIT compiler.
1080    
1081    5.  Fixed several bugs concerned with partial matching of items that consist
1082        of more than one character:
1083    
1084        (a) /^(..)\1/ did not partially match "aba" because checking references was
1085            done on an "all or nothing" basis. This also applied to repeated
1086            references.
1087    
1088        (b) \R did not give a hard partial match if \r was found at the end of the
1089            subject.
1090    
1091        (c) \X did not give a hard partial match after matching one or more
1092            characters at the end of the subject.
1093    
1094        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1095            a partial match for the string "\r".
1096    
1097        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1098            a partial match for a CR character at the end of the subject string.
1099    
1100    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1101        running pcretest, the text "(JIT)" added to the output whenever JIT is
1102        actually used to run the match.
1103    
1104    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1105        or /S+[+] with a digit between 1 and 7.
1106    
1107    8.  OP_NOT now supports any UTF character not just single-byte ones.
1108    
1109    9.  (*MARK) control verb is now supported by the JIT compiler.
1110    
1111    10. The command "./RunTest list" lists the available tests without actually
1112        running any of them. (Because I keep forgetting what they all are.)
1113    
1114    11. Add PCRE_INFO_MAXLOOKBEHIND.
1115    
1116    12. Applied a (slightly modified) user-supplied patch that improves performance
1117        when the heap is used for recursion (compiled with --disable-stack-for-
1118        recursion). Instead of malloc and free for each heap frame each time a
1119        logical recursion happens, frames are retained on a chain and re-used where
1120        possible. This sometimes gives as much as 30% improvement.
1121    
1122    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1123        call.
1124    
1125    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1126    
1127    15. It is now possible to link pcretest with libedit as an alternative to
1128        libreadline.
1129    
1130    16. (*COMMIT) control verb is now supported by the JIT compiler.
1131    
1132    17. The Unicode data tables have been updated to Unicode 6.1.0.
1133    
1134    18. Added --file-list option to pcregrep.
1135    
1136    19. Added binary file support to pcregrep, including the -a, --binary-files,
1137        -I, and --text options.
1138    
1139    20. The madvise function is renamed for posix_madvise for QNX compatibility
1140        reasons. Fixed by Giuseppe D'Angelo.
1141    
1142    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1143    
1144    22. Changed the option for creating symbolic links for 16-bit man pages from
1145        -s to -sf so that re-installing does not cause issues.
1146    
1147    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1148    
1149    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1150        matches in certain environments (the workspace was not being correctly
1151        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1152        some of the workspace data at the beginning of a restart.
1153    
1154    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1155        was not doing so when it should - probably a typo introduced by SVN 528
1156        (change 8.10/14).
1157    
1158    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1159        \w+ when the character tables indicated that \x{c4} was a word character.
1160        There were several related cases, all because the tests for doing a table
1161        lookup were testing for characters less than 127 instead of 255.
1162    
1163    27. If a pattern contains capturing parentheses that are not used in a match,
1164        their slots in the ovector are set to -1. For those that are higher than
1165        any matched groups, this happens at the end of processing. In the case when
1166        there were back references that the ovector was too small to contain
1167        (causing temporary malloc'd memory to be used during matching), and the
1168        highest capturing number was not used, memory off the end of the ovector
1169        was incorrectly being set to -1. (It was using the size of the temporary
1170        memory instead of the true size.)
1171    
1172    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1173        ovector size, it uses memory at the end of the block that it has got.
1174    
1175    29. Check for an overlong MARK name and give an error at compile time. The
1176        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1177    
1178    30. JIT compiler update.
1179    
1180    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1181        Rill for the patch.
1182    
1183    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1184    
1185    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1186    
1187    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1188        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1189        was enabled.
1190    
1191    35. Improve JIT code generation for greedy plus quantifier.
1192    
1193    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1194        1 to "aa" instead of to an empty string. The bug affected repeated groups
1195        that could potentially match an empty string.
1196    
1197    37. Optimizing single character iterators in JIT.
1198    
1199    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1200        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1201        codepoints that are too big for the mode are faulted, and in a UTF mode,
1202        disallowed codepoints are also faulted.
1203    
1204    39. If PCRE was compiled with UTF support, in three places in the DFA
1205        matcher there was code that should only have been obeyed in UTF mode, but
1206        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1207        processing when bytes with values greater than 127 were present. In 16-bit
1208        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1209        both cases the values are those that cannot be the first data item in a UTF
1210        character. The three items that might have provoked this were recursions,
1211        possessively repeated groups, and atomic groups.
1212    
1213    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1214        and pcregrep, because some OS require shared objects to be explicitly
1215        passed to ld, causing the link step to fail if they are not.
1216    
1217    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1218        mode, patterns that started with \h* or \R* might be incorrectly matched.
1219    
1220    
1221    Version 8.30 04-February-2012
1222    -----------------------------
1223    
1224    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1225        name is defined in ctype.h.
1226    
1227    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1228        only in quite long subpatterns.
1229    
1230    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1231        since it was replaced by pcre_fullinfo() in February 2000.
1232    
1233    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1234        match a (*MARK), and the match failed at the start of the subject, a
1235        reference to memory before the start of the subject could occur. This bug
1236        was introduced by fix 17 of release 8.21.
1237    
1238    5.  A reference to an unset group with zero minimum repetition was giving
1239        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1240        /(another)?(\1?)test/ matched against "hello world test". This bug was
1241        introduced in release 8.13.
1242    
1243    6.  Add support for 16-bit character strings (a large amount of work involving
1244        many changes and refactorings).
1245    
1246    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1247        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1248        from a file.
1249    
1250    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1251        rounding is not applied in this particular case).
1252    
1253    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1254        if they appear, or are escaped, in patterns.
1255    
1256    10. Get rid of a number of -Wunused-but-set-variable warnings.
1257    
1258    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1259        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1260        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1261        also returns the mark "x". This bug applied to capturing parentheses,
1262        non-capturing parentheses, and atomic parentheses. It also applied to some
1263        assertions.
1264    
1265    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1266        information out of configure.ac instead of relying on pcre.h.generic, which
1267        is not stored in the repository.
1268    
1269    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1270        -lreadline.
1271    
1272    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1273    
1274    15. Applied Graycode's patch to put the top-level frame on the stack rather
1275        than the heap when not using the stack for recursion. This gives a
1276        performance improvement in many cases when recursion is not deep.
1277    
1278    16. Experimental code added to "pcretest -C" to output the stack frame size.
1279    
1280    
1281    Version 8.21 12-Dec-2011
1282    ------------------------
1283    
1284    1.  Updating the JIT compiler.
1285    
1286    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1287        are added as well.
1288    
1289    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1290        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1291        calling _pcre_jit_exec. Some extra comments are added.
1292    
1293    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1294        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1295        was introduced by change 18 for 8.20.
1296    
1297    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1298        ECMA-262 standard.
1299    
1300    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1301        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1302        This bug was probably introduced by change 9 of 8.13.
1303    
1304    7.  While fixing 6 above, I noticed that a number of other items were being
1305        incorrectly rejected as "not fixed length". This arose partly because newer
1306        opcodes had not been added to the fixed-length checking code. I have (a)
1307        corrected the bug and added tests for these items, and (b) arranged for an
1308        error to occur if an unknown opcode is encountered while checking for fixed
1309        length instead of just assuming "not fixed length". The items that were
1310        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1311        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1312        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1313    
1314    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1315        being incorrectly compiled and would have given unpredicatble results.
1316    
1317    9.  A possessively repeated subpattern with minimum repeat count greater than
1318        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1319        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1320        the first (A) could occur when it should not.
1321    
1322    10. Add a cast and remove a redundant test from the code.
1323    
1324    11. JIT should use pcre_malloc/pcre_free for allocation.
1325    
1326    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1327        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1328        is anything other than /usr, -L is still shown).
1329    
1330    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1331    
1332    14. Perl does not support \N without a following name in a [] class; PCRE now
1333        also gives an error.
1334    
1335    15. If a forward reference was repeated with an upper limit of around 2000,
1336        it caused the error "internal error: overran compiling workspace". The
1337        maximum number of forward references (including repeats) was limited by the
1338        internal workspace, and dependent on the LINK_SIZE. The code has been
1339        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1340        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1341        of around 200,000 forward references. While doing this, I also speeded up
1342        the filling in of repeated forward references.
1343    
1344    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1345        incorrectly expecting the subject to contain another "a" after the start.
1346    
1347    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1348        in the match, the SKIP should be ignored. This was not happening; instead
1349        the SKIP was being treated as NOMATCH. For patterns such as
1350        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1351        tested.
1352    
1353    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1354        now much more compatible with Perl, in particular in cases where the result
1355        is a non-match for a non-anchored pattern. For example, if
1356        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1357        "m", where previously it did not return a name. A side effect of this
1358        change is that for partial matches, the last encountered mark name is
1359        returned, as for non matches. A number of tests that were previously not
1360        Perl-compatible have been moved into the Perl-compatible test files. The
1361        refactoring has had the pleasing side effect of removing one argument from
1362        the match() function, thus reducing its stack requirements.
1363    
1364    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1365        subsequent uses of /S (without +) incorrectly behaved like /S+.
1366    
1367    21. Retrieve executable code size support for the JIT compiler and fixing
1368        some warnings.
1369    
1370    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1371        not work when the shorter character appeared right at the end of the
1372        subject string.
1373    
1374    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1375        systems.
1376    
1377    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1378        output it when the /M option is used in pcretest.
1379    
1380    25. The CheckMan script was not being included in the distribution. Also, added
1381        an explicit "perl" to run Perl scripts from the PrepareRelease script
1382        because this is reportedly needed in Windows.
1383    
1384    26. If study data was being save in a file and studying had not found a set of
1385        "starts with" bytes for the pattern, the data written to the file (though
1386        never used) was taken from uninitialized memory and so caused valgrind to
1387        complain.
1388    
1389    27. Updated RunTest.bat as provided by Sheri Pierce.
1390    
1391    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1392    
1393    29. Computation of memory usage for the table of capturing group names was
1394        giving an unnecessarily large value.
1395    
1396    
1397    Version 8.20 21-Oct-2011
1398    ------------------------
1399    
1400    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1401        a POSIX class. After further experiments with Perl, which convinced me that
1402        Perl has bugs and confusions, a closing square bracket is no longer allowed
1403        in a POSIX name. This bug also affected patterns with classes that started
1404        with full stops.
1405    
1406    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1407        captured substring, but while checking the failing first alternative,
1408        substring 1 is temporarily captured. If the output vector supplied to
1409        pcre_exec() was not big enough for this capture, the yield of the function
1410        was still zero ("insufficient space for captured substrings"). This cannot
1411        be totally fixed without adding another stack variable, which seems a lot
1412        of expense for a edge case. However, I have improved the situation in cases
1413        such as /(a)(b)x|abc/ matched against "abc", where the return code
1414        indicates that fewer than the maximum number of slots in the ovector have
1415        been set.
1416    
1417    3.  Related to (2) above: when there are more back references in a pattern than
1418        slots in the output vector, pcre_exec() uses temporary memory during
1419        matching, and copies in the captures as far as possible afterwards. It was
1420        using the entire output vector, but this conflicts with the specification
1421        that only 2/3 is used for passing back captured substrings. Now it uses
1422        only the first 2/3, for compatibility. This is, of course, another edge
1423        case.
1424    
1425    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1426        main code base, and can be used by building with --enable-jit. When this is
1427        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1428        runtime --no-jit option is given.
1429    
1430    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1431        ovector, the return from the function was zero, implying that there were
1432        other matches that did not fit. The correct "exactly full" value is now
1433        returned.
1434    
1435    6.  If a subpattern that was called recursively or as a subroutine contained
1436        (*PRUNE) or any other control that caused it to give a non-standard return,
1437        invalid errors such as "Error -26 (nested recursion at the same subject
1438        position)" or even infinite loops could occur.
1439    
1440    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1441        computing the minimum length on reaching *ACCEPT, and so ended up with the
1442        wrong value of 1 rather than 0. Further investigation indicates that
1443        computing a minimum subject length in the presence of *ACCEPT is difficult
1444        (think back references, subroutine calls), and so I have changed the code
1445        so that no minimum is registered for a pattern that contains *ACCEPT.
1446    
1447    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1448        it was not handled as intended. [But see 16 below.]
1449    
1450    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1451        Sheri Pierce.
1452    
1453    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1454        the first byte in a match must be "a".
1455    
1456    11. Change 17 for 8.13 increased the recursion depth for patterns like
1457        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1458        pattern contains any instances of (*THEN). If it does not, the old
1459        optimizations are restored. It would be nice to do this on a per-group
1460        basis, but at the moment that is not feasible.
1461    
1462    12. In some environments, the output of pcretest -C is CRLF terminated. This
1463        broke RunTest's code that checks for the link size. A single white space
1464        character after the value is now allowed for.
1465    
1466    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1467        For "fr", it uses the Windows-specific input and output files.
1468    
1469    14. If (*THEN) appeared in a group that was called recursively or as a
1470        subroutine, it did not work as intended. [But see next item.]
1471    
1472    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1473        pattern fragments (but not containing any | characters). If A and B are
1474        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1475        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1476        D. In other words, Perl considers parentheses that do not contain any |
1477        characters to be part of a surrounding alternative, whereas PCRE was
1478        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1479        differently. PCRE now behaves in the same way as Perl, except in the case
1480        of subroutine/recursion calls such as (?1) which have in any case always
1481        been different (but PCRE had them first :-).
1482    
1483    16. Related to 15 above: Perl does not treat the | in a conditional group as
1484        creating alternatives. Such a group is treated in the same way as an
1485        ordinary group without any | characters when processing (*THEN). PCRE has
1486        been changed to match Perl's behaviour.
1487    
1488    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1489        RunGrepTest script failed.
1490    
1491    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1492        inevitable for groups that contain captures, but it can lead to a lot of
1493        stack use in large patterns. The old behaviour has been restored for atomic
1494        groups that do not contain any capturing parentheses.
1495    
1496    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1497        suppress the check for a minimum subject length at run time. (If it was
1498        given to pcre_exec() or pcre_dfa_exec() it did work.)
1499    
1500    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1501        fail to work when decoding hex characters in data strings in EBCDIC
1502        environments.
1503    
1504    21. It appears that in at least one Mac OS environment, the isxdigit() function
1505        is implemented as a macro that evaluates to its argument more than once,
1506        contravening the C 90 Standard (I haven't checked a later standard). There
1507        was an instance in pcretest which caused it to go wrong when processing
1508        \x{...} escapes in subject strings. The has been rewritten to avoid using
1509        things like p++ in the argument of isxdigit().
1510    
1511    
1512    Version 8.13 16-Aug-2011
1513    ------------------------
1514    
1515    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1516    
1517    2.  Two minor typos in pcre_internal.h have been fixed.
1518    
1519    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1520        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1521        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1522    
1523    4.  There were a number of related bugs in the code for matching backrefences
1524        caselessly in UTF-8 mode when codes for the characters concerned were
1525        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1526        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1527        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1528        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1529        2-byte code at the end of the subject (it thought there wasn't enough data
1530        left).
1531    
1532    5.  Comprehensive information about what went wrong is now returned by
1533        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1534        as the output vector has at least 2 elements. The offset of the start of
1535        the failing character and a reason code are placed in the vector.
1536    
1537    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1538        now returned is for the first byte of the failing character, instead of the
1539        last byte inspected. This is an incompatible change, but I hope it is small
1540        enough not to be a problem. It makes the returned offset consistent with
1541        pcre_exec() and pcre_dfa_exec().
1542    
1543    7.  pcretest now gives a text phrase as well as the error number when
1544        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1545        failure, the offset and reason code are output.
1546    
1547    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1548        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1549        back over a single character (\n). This seems wrong (because it treated the
1550        two characters as a single entity when going forwards), conflicts with the
1551        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1552        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1553        has been changed.
1554    
1555    9.  Some internal refactoring has changed the processing so that the handling
1556        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1557        time (the PCRE_DOTALL option was changed this way some time ago: version
1558        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1559        which was always a bit of a fudge. It also means that there is one less
1560        argument for the match() function, which reduces its stack requirements
1561        slightly. This change also fixes an incompatibility with Perl: the pattern
1562        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1563    
1564    10. More internal refactoring has drastically reduced the number of recursive
1565        calls to match() for possessively repeated groups such as (abc)++ when
1566        using pcre_exec().
1567    
1568    11. While implementing 10, a number of bugs in the handling of groups were
1569        discovered and fixed:
1570    
1571        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1572        (a|)*(?1) gave a compile-time internal error.
1573        ((a|)+)+  did not notice that the outer group could match an empty string.
1574        (^a|^)+   was not marked as anchored.
1575        (.*a|.*)+ was not marked as matching at start or after a newline.
1576    
1577    12. Yet more internal refactoring has removed another argument from the match()
1578        function. Special calls to this function are now indicated by setting a
1579        value in a variable in the "match data" data block.
1580    
1581    13. Be more explicit in pcre_study() instead of relying on "default" for
1582        opcodes that mean there is no starting character; this means that when new
1583        ones are added and accidentally left out of pcre_study(), testing should
1584        pick them up.
1585    
1586    14. The -s option of pcretest has been documented for ages as being an old
1587        synonym of -m (show memory usage). I have changed it to mean "force study
1588        for every regex", that is, assume /S for every regex. This is similar to -i
1589        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1590        using it. It makes it easier to run collections of tests with and without
1591        study enabled, and thereby test pcre_study() more easily. All the standard
1592        tests are now run with and without -s (but some patterns can be marked as
1593        "never study" - see 20 below).
1594    
1595    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1596        restoration of the capturing data to the outer values was not happening
1597        correctly.
1598    
1599    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1600        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1601        pattern had matched an empty string, and so incorrectly returned a no
1602        match.
1603    
1604    17. There was optimizing code for the last branch of non-capturing parentheses,
1605        and also for the obeyed branch of a conditional subexpression, which used
1606        tail recursion to cut down on stack usage. Unfortunately, now that there is
1607        the possibility of (*THEN) occurring in these branches, tail recursion is
1608        no longer possible because the return has to be checked for (*THEN). These
1609        two optimizations have therefore been removed. [But see 8.20/11 above.]
1610    
1611    18. If a pattern containing \R was studied, it was assumed that \R always
1612        matched two bytes, thus causing the minimum subject length to be
1613        incorrectly computed because \R can also match just one byte.
1614    
1615    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1616        was incorrectly computed.
1617    
1618    20. If /S is present twice on a test pattern in pcretest input, it now
1619        *disables* studying, thereby overriding the use of -s on the command line
1620        (see 14 above). This is necessary for one or two tests to keep the output
1621        identical in both cases.
1622    
1623    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1624        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1625    
1626    22. When an atomic group that contained a capturing parenthesis was
1627        successfully matched, but the branch in which it appeared failed, the
1628        capturing was not being forgotten if a higher numbered group was later
1629        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1630        group 1 to "a", when in fact it should be unset. This applied to multi-
1631        branched capturing and non-capturing groups, repeated or not, and also to
1632        positive assertions (capturing in negative assertions does not happen
1633        in PCRE) and also to nested atomic groups.
1634    
1635    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1636        subject after a captured substring, to make it easier to tell which of a
1637        number of identical substrings has been captured.
1638    
1639    24. The way atomic groups are processed by pcre_exec() has been changed so that
1640        if they are repeated, backtracking one repetition now resets captured
1641        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1642        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1643        "aaa". Previously, it would have been "a". As part of this code
1644        refactoring, the way recursive calls are handled has also been changed.
1645    
1646    25. If an assertion condition captured any substrings, they were not passed
1647        back unless some other capturing happened later. For example, if
1648        (?(?=(a))a) was matched against "a", no capturing was returned.
1649    
1650    26. When studying a pattern that contained subroutine calls or assertions,
1651        the code for finding the minimum length of a possible match was handling
1652        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1653        group 1 called group 2 while simultaneously a separate group 2 called group
1654        1). A stack overflow occurred in this case. I have fixed this by limiting
1655        the recursion depth to 10.
1656    
1657    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1658        Fortmann. This supports explicit test numbers on the command line, and has
1659        argument validation and error reporting.
1660    
1661    28. An instance of \X with an unlimited repeat could fail if at any point the
1662        first character it looked at was a mark character.
1663    
1664    29. Some minor code refactoring concerning Unicode properties and scripts
1665        should reduce the stack requirement of match() slightly.
1666    
1667    30. Added the '=' option to pcretest to check the setting of unused capturing
1668        slots at the end of the pattern, which are documented as being -1, but are
1669        not included in the return count.
1670    
1671    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1672        compiled something random. Now it gives a compile-time error (as does
1673        Perl).
1674    
1675    32. A *MARK encountered during the processing of a positive assertion is now
1676        recorded and passed back (compatible with Perl).
1677    
1678    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1679        had alternative anchored branches, the search for a second match in a line
1680        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1681        matched the line "0102" twice. The same bug affected patterns that started
1682        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1683        twice.
1684    
1685    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1686        does, and because of capturing effects, quantifying parenthesized
1687        assertions may at times be useful. Quantifiers are now allowed for
1688        parenthesized assertions.
1689    
1690    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1691    
1692    36. \g was being checked for fancy things in a character class, when it should
1693        just be a literal "g".
1694    
1695    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1696        appearance of a nested POSIX class supersedes an apparent external class.
1697        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1698        unescaped square brackets may also appear as part of class names. For
1699        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1700        more like Perl. (But see 8.20/1 above.)
1701    
1702    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1703        was because it thought it was \N{name}, which is not supported).
1704    
1705    39. Add minix to OS list not supporting the -S option in pcretest.
1706    
1707    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1708        cannot analyze patterns in sufficient detail to catch mutual recursions
1709        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1710        subgroup is called recursively as a subpattern for a second time at the
1711        same position in the subject string. In previous releases this might have
1712        been caught by the recursion limit, or it might have run out of stack.
1713    
1714    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1715        happen only once. PCRE was, however incorrectly giving a compile time error
1716        "recursive call could loop indefinitely" because it cannot analyze the
1717        pattern in sufficient detail. The compile time test no longer happens when
1718        PCRE is compiling a conditional subpattern, but actual runaway loops are
1719        now caught at runtime (see 40 above).
1720    
1721    42. It seems that Perl allows any characters other than a closing parenthesis
1722        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1723        has been changed to be the same.
1724    
1725    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1726        as not to get warnings when autogen.sh is called. Also changed
1727        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1728    
1729    44. To help people who use pcregrep to scan files containing exceedingly long
1730        lines, the following changes have been made:
1731    
1732        (a) The default value of the buffer size parameter has been increased from
1733            8K to 20K. (The actual buffer used is three times this size.)
1734    
1735        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1736            PCRE is built.
1737    
1738        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1739            to be set at run time.
1740    
1741        (d) Numerical values in pcregrep options can be followed by K or M, for
1742            example --buffer-size=50K.
1743    
1744        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1745            given and the return code is set to 2.
1746    
1747    45. Add a pointer to the latest mark to the callout data block.
1748    
1749    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1750        partial match of an empty string instead of no match. This was specific to
1751        the use of ".".
1752    
1753    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1754        complete match instead of a partial match. This bug was dependent on both
1755        the PCRE_UTF8 and PCRE_DOTALL options being set.
1756    
1757    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1758        starting byte set, because \b was not being ignored.
1759    
1760    
1761    Version 8.12 15-Jan-2011
1762    ------------------------
1763    
1764    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1765        checks for such things as part of the documentation building process.
1766    
1767    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1768        --match-limit and --recursion-limit options (added for 8.11). In
1769        particular, this made one of the standard tests fail. (The integer value
1770        went into the wrong half of a long int.)
1771    
1772    3.  If the --colour option was given to pcregrep with -v (invert match), it
1773        did strange things, either producing crazy output, or crashing. It should,
1774        of course, ignore a request for colour when reporting lines that do not
1775        match.
1776    
1777    4.  Another pcregrep bug caused similar problems if --colour was specified with
1778        -M (multiline) and the pattern match finished with a line ending.
1779    
1780    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1781        matched in multiline mode, the following line was shown as part of the
1782        match. This seems wrong, so I have changed it.
1783    
1784    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1785        the check for further matches in the same line (so they could be coloured)
1786        to overrun the end of the current line. If another match was found, it was
1787        incorrectly shown (and then shown again when found in the next line).
1788    
1789    7.  If pcregrep was compiled under Windows, there was a reference to the
1790        function pcregrep_exit() before it was defined. I am assuming this was
1791        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1792        reported by a user. I've moved the definition above the reference.
1793    
1794    
1795    Version 8.11 10-Dec-2010
1796    ------------------------
1797    
1798    1.  (*THEN) was not working properly if there were untried alternatives prior
1799        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1800        backtracked to try for "b" instead of moving to the next alternative branch
1801        at the same level (in this case, to look for "c"). The Perl documentation
1802        is clear that when (*THEN) is backtracked onto, it goes to the "next
1803        alternative in the innermost enclosing group".
1804    
1805    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1806        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1807        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1808        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1809        (*THEN).
1810    
1811    3.  If \s appeared in a character class, it removed the VT character from
1812        the class, even if it had been included by some previous item, for example
1813        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1814        of \s, but is part of the POSIX "space" class.)
1815    
1816    4.  A partial match never returns an empty string (because you can always
1817        match an empty string at the end of the subject); however the checking for
1818        an empty string was starting at the "start of match" point. This has been
1819        changed to the "earliest inspected character" point, because the returned
1820        data for a partial match starts at this character. This means that, for
1821        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1822        (previously it gave "no match").
1823    
1824    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1825        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1826        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1827        has an implication that the given string is incomplete (because a partial
1828        match is preferred over a full match). For this reason, these items now
1829        give a partial match in this situation. [Aside: previously, the one case
1830        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1831        match rather than a full match, which was wrong by the old rules, but is
1832        now correct.]
1833    
1834    6.  There was a bug in the handling of #-introduced comments, recognized when
1835        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1836        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1837        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1838        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1839        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1840        places in pcre_compile().
1841    
1842    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1843        comments when looking ahead for named forward references to subpatterns,
1844        the only newline sequence it recognized was NL. It now handles newlines
1845        according to the set newline convention.
1846    
1847    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1848        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1849        cater for a lack of strerror(). These oversights have been fixed.
1850    
1851    9.  Added --match-limit and --recursion-limit to pcregrep.
1852    
1853    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1854    
1855    11. When the -o option was used, pcregrep was setting a return code of 1, even
1856        when matches were found, and --line-buffered was not being honoured.
1857    
1858    12. Added an optional parentheses number to the -o and --only-matching options
1859        of pcregrep.
1860    
1861    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1862        can match an empty string. The code to do it in pcretest and pcredemo
1863        needed fixing:
1864    
1865        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1866            only one byte after an empty string match just before CRLF (this case
1867            just got forgotten; "any" and "anycrlf" were OK).
1868    
1869        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1870            mode when an empty string match preceded an ASCII character followed by
1871            a non-ASCII character. (The code for advancing by one character rather
1872            than one byte was nonsense.)
1873    
1874        (c) The pcredemo.c sample program did not have any code at all to handle
1875            the cases when CRLF is a valid newline sequence.
1876    
1877    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1878        as a starting offset was within the subject string. There is now a new
1879        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1880        negative or greater than the length of the string. In order to test this,
1881        pcretest is extended to allow the setting of negative starting offsets.
1882    
1883    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1884        starting offset points to the beginning of a UTF-8 character was
1885        unnecessarily clumsy. I tidied it up.
1886    
1887    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1888        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1889    
1890    17. Nobody had reported that the --include_dir option, which was added in
1891        release 7.7 should have been called --include-dir (hyphen, not underscore)
1892        for compatibility with GNU grep. I have changed it to --include-dir, but
1893        left --include_dir as an undocumented synonym, and the same for
1894        --exclude-dir, though that is not available in GNU grep, at least as of
1895        release 2.5.4.
1896    
1897    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1898        characters from a string of bytes) have been redefined so as not to use
1899        loops, in order to improve performance in some environments. At the same
1900        time, I abstracted some of the common code into auxiliary macros to save
1901        repetition (this should not affect the compiled code).
1902    
1903    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1904        compile-time error is now given if \c is not followed by an ASCII
1905        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1906        different, and any byte value is allowed.)
1907    
1908    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1909        START_OPTIMIZE option, which is now allowed at compile time - but just
1910        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1911        to pcregrep and other applications that have no direct access to PCRE
1912        options. The new /Y option in pcretest sets this option when calling
1913        pcre_compile().
1914    
1915    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1916        back references. Groups containing recursive back references were forced to
1917        be atomic by that change, but in the case of named groups, the amount of
1918        memory required was incorrectly computed, leading to "Failed: internal
1919        error: code overflow". This has been fixed.
1920    
1921    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1922        pcretest.c, to avoid build problems in some Borland environments.
1923    
1924    
1925    Version 8.10 25-Jun-2010
1926    ------------------------
1927    
1928    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1929        THEN.
1930    
1931    2.  (*ACCEPT) was not working when inside an atomic group.
1932    
1933    3.  Inside a character class, \B is treated as a literal by default, but
1934        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1935        causes the error). The code is unchanged, but I tidied the documentation.
1936    
1937    4.  Inside a character class, PCRE always treated \R and \X as literals,
1938        whereas Perl faults them if its -w option is set. I have changed PCRE so
1939        that it faults them when PCRE_EXTRA is set.
1940    
1941    5.  Added support for \N, which always matches any character other than
1942        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1943    
1944    6.  When compiling pcregrep with newer versions of gcc which may have
1945        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1946        declared with attribute warn_unused_result" were given. Just casting the
1947        result to (void) does not stop the warnings; a more elaborate fudge is
1948        needed. I've used a macro to implement this.
1949    
1950    7.  Minor change to pcretest.c to avoid a compiler warning.
1951    
1952    8.  Added four artifical Unicode properties to help with an option to make
1953        \s etc use properties (see next item). The new properties are: Xan
1954        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1955    
1956    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1957        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1958        this option. Modified pcretest to add /W to test this facility. Added
1959        REG_UCP to make it available via the POSIX interface.
1960    
1961    10. Added --line-buffered to pcregrep.
1962    
1963    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1964        studied, and the match started with a letter with a code point greater than
1965        127 whose first byte was different to the first byte of the other case of
1966        the letter, the other case of this starting letter was not recognized
1967        (#976).
1968    
1969    12. If a pattern that was studied started with a repeated Unicode property
1970        test, for example, \p{Nd}+, there was the theoretical possibility of
1971        setting up an incorrect bitmap of starting bytes, but fortunately it could
1972        not have actually happened in practice until change 8 above was made (it
1973        added property types that matched character-matching opcodes).
1974    
1975    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1976        possible starting bytes for non-anchored patterns.
1977    
1978    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1979        \R, and also a number of cases that involve Unicode properties, both
1980        explicit and implicit when PCRE_UCP is set.
1981    
1982    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1983        input, it could crash or give wrong results if characters with values
1984        greater than 0xc0 were present in the subject string. (Detail: it assumed
1985        UTF-8 input when processing these items.)
1986    
1987    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1988        size_t is 64-bit (#991).
1989    
1990    17. Added a check for running out of memory when PCRE is compiled with
1991        --disable-stack-for-recursion (#990).
1992    
1993    18. If the last data line in a file for pcretest does not have a newline on
1994        the end, a newline was missing in the output.
1995    
1996    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1997        less than 128) in its various bitmaps. However, there is a facility for
1998        generating tables according to the current locale when PCRE is compiled. It
1999        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2000        characters, are recognized by isspace() and therefore were getting set in
2001        these tables, and indeed these tables seem to approximate to ISO 8859. This
2002        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2003        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2004        which of course cannot start UTF-8 characters. I have changed the code so
2005        that only real ASCII characters (less than 128) and the correct starting
2006        bytes for UTF-8 encodings are set for characters greater than 127 when in
2007        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2008        altogether.)
2009    
2010    20. Added the /T option to pcretest so as to be able to run tests with non-
2011        standard character tables, thus making it possible to include the tests
2012        used for 19 above in the standard set of tests.
2013    
2014    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2015        reference to a subpattern the other side of a comment that contains an
2016        opening parenthesis caused either an internal compiling error, or a
2017        reference to the wrong subpattern.
2018    
2019    
2020    Version 8.02 19-Mar-2010
2021    ------------------------
2022    
2023    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2024    
2025    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2026        configured.
2027    
2028    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2029        original author of that file, following a query about its status.
2030    
2031    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2032        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2033    
2034    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2035        quantifier applied to a forward-referencing subroutine call, could compile
2036        incorrect code or give the error "internal error: previously-checked
2037        referenced subpattern not found".
2038    
2039    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2040        variables to point to external functions. For these systems, therefore,
2041        pcre_malloc etc. are now initialized to local functions that call the
2042        relevant global functions.
2043    
2044    7.  There were two entries missing in the vectors called coptable and poptable
2045        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2046        I've fixed the data, and added a kludgy way of testing at compile time that
2047        the lengths are correct (equal to the number of opcodes).
2048    
2049    8.  Following on from 7, I added a similar kludge to check the length of the
2050        eint vector in pcreposix.c.
2051    
2052    9.  Error texts for pcre_compile() are held as one long string to avoid too
2053        much relocation at load time. To find a text, the string is searched,
2054        counting zeros. There was no check for running off the end of the string,
2055        which could happen if a new error number was added without updating the
2056        string.
2057    
2058    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2059    
2060    11. \K was not working if it appeared in an atomic group or in a group that
2061        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2062        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2063        the assertion is positive, but not if it is negative.
2064    
2065    12. Change 11 fortuitously reduced the size of the stack frame used in the
2066        "match()" function of pcre_exec.c by one pointer. Forthcoming
2067        implementation of support for (*MARK) will need an extra pointer on the
2068        stack; I have reserved it now, so that the stack frame size does not
2069        decrease.
2070    
2071    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2072        item in branch that calls a recursion is a subroutine call - as in the
2073        second branch in the above example - was incorrectly given the compile-
2074        time error "recursive call could loop indefinitely" because pcre_compile()
2075        was not correctly checking the subroutine for matching a non-empty string.
2076    
2077    14. The checks for overrunning compiling workspace could trigger after an
2078        overrun had occurred. This is a "should never occur" error, but it can be
2079        triggered by pathological patterns such as hundreds of nested parentheses.
2080        The checks now trigger 100 bytes before the end of the workspace.
2081    
2082    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2083    
2084    
2085    Version 8.01 19-Jan-2010
2086    ------------------------
2087    
2088    1.  If a pattern contained a conditional subpattern with only one branch (in
2089        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2090        computed the wrong minimum data length (which is of course zero for such
2091        subpatterns). This could cause incorrect "no match" results.
2092    
2093    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2094        the pattern is reset in the first branch, pcre_compile() failed with
2095        "internal error: code overflow at offset...". This happened only when
2096        the reset was to the original external option setting. (An optimization
2097        abstracts leading options settings into an external setting, which was the
2098        cause of this.)
2099    
2100    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2101        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2102        assertion pattern did not match (meaning that the assertion was true), it
2103        was incorrectly treated as false if the SKIP had been reached during the
2104        matching. This also applied to assertions used as conditions.
2105    
2106    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2107        assertion subpattern, including such a pattern used as a condition,
2108        unpredictable results occurred, instead of the error return
2109        PCRE_ERROR_DFA_UITEM.
2110    
2111    5.  The C++ GlobalReplace function was not working like Perl for the special
2112        situation when an empty string is matched. It now does the fancy magic
2113        stuff that is necessary.
2114    
2115    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2116        removed. (These were left over from very, very early versions of PCRE.)
2117    
2118    7.  Some cosmetic changes to the code to make life easier when compiling it
2119        as part of something else:
2120    
2121        (a) Change DEBUG to PCRE_DEBUG.
2122    
2123        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2124            called "current" as "current_branch", to prevent a collision with the
2125            Linux macro when compiled as a kernel module.
2126    
2127        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2128            prevent a collision with the Linux macro when compiled as a kernel
2129            module.
2130    
2131    8.  In pcre_compile() there are some checks for integer overflows that used to
2132        cast potentially large values to (double). This has been changed to that
2133        when building, a check for int64_t is made, and if it is found, it is used
2134        instead, thus avoiding the use of floating point arithmetic. (There is no
2135        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2136        double.
2137    
2138    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2139        2005 (difference between two addresses compared to an unsigned value).
2140    
2141    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2142        custom one, because of the following reported problem in Windows:
2143    
2144          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2145              under Win32.
2146          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2147              therefore missing the function definition.
2148          - The compiler thus generates a "C" signature for the test function.
2149          - The linker fails to find the "C" function.
2150          - PCRE fails to configure if asked to do so against libbz2.
2151    
2152    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2153        messages were output:
2154    
2155          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2156          rerunning libtoolize, to keep the correct libtool macros in-tree.
2157          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2158    
2159        I have done both of these things.
2160    
2161    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2162        most of the time, it *can* run out if it is given a pattern that contains a
2163        runaway infinite recursion. I updated the discussion in the pcrestack man
2164        page.
2165    
2166    13. Now that we have gone to the x.xx style of version numbers, the minor
2167        version may start with zero. Using 08 or 09 is a bad idea because users
2168        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2169        interpreted as invalid octal numbers. I've updated the previous comment in
2170        configure.ac, and also added a check that gives an error if 08 or 09 are
2171        used.
2172    
2173    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2174        causing partial matching to fail when the end of the subject matched \W
2175        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2176    
2177    15. There were some discrepancies between the declarations in pcre_internal.h
2178        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2179        their definitions. The declarations used "const uschar *" and the
2180        definitions used USPTR. Even though USPTR is normally defined as "const
2181        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2182        reported that: "This difference in casting confuses some C++ compilers, for
2183        example, SunCC recognizes above declarations as different functions and
2184        generates broken code for hbpcre." I have changed the declarations to use
2185        USPTR.
2186    
2187    16. GNU libtool is named differently on some systems. The autogen.sh script now
2188        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2189        (FreeBSD).
2190    
2191    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2192        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2193        comment: "Figure out how to create a longlong from a string: strtoll and
2194        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2195        instance, but it only takes 2 args instead of 3!"
2196    
2197    18. A subtle bug concerned with back references has been fixed by a change of
2198        specification, with a corresponding code fix. A pattern such as
2199        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2200        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2201        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2202        same bug. Such groups have to be quantified to be useful, or contained
2203        inside another quantified group. (If there's no repetition, the reference
2204        can never match.) The problem arises because, having left the group and
2205        moved on to the rest of the pattern, a later failure that backtracks into
2206        the group uses the captured value from the final iteration of the group
2207        rather than the correct earlier one. I have fixed this in PCRE by forcing
2208        any group that contains a reference to itself to be an atomic group; that
2209        is, there cannot be any backtracking into it once it has completed. This is
2210        similar to recursive and subroutine calls.
2211    
2212    
2213    Version 8.00 19-Oct-09
2214    ----------------------
2215    
2216    1.  The table for translating pcre_compile() error codes into POSIX error codes
2217        was out-of-date, and there was no check on the pcre_compile() error code
2218        being within the table. This could lead to an OK return being given in
2219        error.
2220    
2221    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2222        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2223        in a Windows environment.
2224    
2225    3.  The pcregrep --count option prints the count for each file even when it is
2226        zero, as does GNU grep. However, pcregrep was also printing all files when
2227        --files-with-matches was added. Now, when both options are given, it prints
2228        counts only for those files that have at least one match. (GNU grep just
2229        prints the file name in this circumstance, but including the count seems
2230        more useful - otherwise, why use --count?) Also ensured that the
2231        combination -clh just lists non-zero counts, with no names.
2232    
2233    4.  The long form of the pcregrep -F option was incorrectly implemented as
2234        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2235        but it seems right to fix it, and I didn't think it was worth preserving
2236        the old behaviour.
2237    
2238    5.  The command line items --regex=pattern and --regexp=pattern were not
2239        recognized by pcregrep, which required --regex pattern or --regexp pattern
2240        (with a space rather than an '='). The man page documented the '=' forms,
2241        which are compatible with GNU grep; these now work.
2242    
2243    6.  No libpcreposix.pc file was created for pkg-config; there was just
2244        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2245    
2246    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2247        when UCP support is not needed, by modifying the Python script that
2248        generates it from Unicode data files. This should not matter if the module
2249        is correctly used as a library, but I received one complaint about 50K of
2250        unwanted data. My guess is that the person linked everything into his
2251        program rather than using a library. Anyway, it does no harm.
2252    
2253    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2254        was a minimum greater than 1 for a wide character in a possessive
2255        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2256        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2257        character. Chaos in the form of incorrect output or a compiling loop could
2258        result.
2259    
2260    9.  The restrictions on what a pattern can contain when partial matching is
2261        requested for pcre_exec() have been removed. All patterns can now be
2262        partially matched by this function. In addition, if there are at least two
2263        slots in the offset vector, the offset of the earliest inspected character
2264        for the match and the offset of the end of the subject are set in them when
2265        PCRE_ERROR_PARTIAL is returned.
2266    
2267    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2268        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2269        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2270        and may be more useful for multi-segment matching.
2271    
2272    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2273        used to be given if ever the end of the subject was reached; now it is
2274        given only if matching could not proceed because another character was
2275        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2276        string "Z", which now yields "no match" instead of "partial match". In the
2277        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2278        final character ended with (*FAIL).
2279    
2280    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2281        if the pattern had a "must contain" character that was already found in the
2282        earlier partial match, unless partial matching was again requested. For
2283        example, with the pattern /dog.(body)?/, the "must contain" character is
2284        "g". If the first part-match was for the string "dog", restarting with
2285        "sbody" failed. This bug has been fixed.
2286    
2287    13. The string returned by pcre_dfa_exec() after a partial match has been
2288        changed so that it starts at the first inspected character rather than the
2289        first character of the match. This makes a difference only if the pattern
2290        starts with a lookbehind assertion or \b or \B (\K is not supported by
2291        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2292        matching functions compatible, and I think it's the right thing to do.
2293    
2294    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2295        so that the demonstration program is easily available in environments where
2296        PCRE has not been installed from source.
2297    
2298    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2299        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2300        library.
2301    
2302    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2303        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2304        is not the first non-POSIX option to be added. Clearly some people find
2305        these options useful.
2306    
2307    17. If a caller to the POSIX matching function regexec() passes a non-zero
2308        value for nmatch with a NULL value for pmatch, the value of
2309        nmatch is forced to zero.
2310    
2311    18. RunGrepTest did not have a test for the availability of the -u option of
2312        the diff command, as RunTest does. It now checks in the same way as
2313        RunTest, and also checks for the -b option.
2314    
2315    19. If an odd number of negated classes containing just a single character
2316        interposed, within parentheses, between a forward reference to a named
2317        subpattern and the definition of the subpattern, compilation crashed with
2318        an internal error, complaining that it could not find the referenced
2319        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2320        [The bug was that it was starting one character too far in when skipping
2321        over the character class, thus treating the ] as data rather than
2322        terminating the class. This meant it could skip too much.]
2323    
2324    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2325        /g option in pcretest when the pattern contains \K, which makes it possible
2326        to have an empty string match not at the start, even when the pattern is
2327        anchored. Updated pcretest and pcredemo to use this option.
2328    
2329    21. If the maximum number of capturing subpatterns in a recursion was greater
2330        than the maximum at the outer level, the higher number was returned, but
2331        with unset values at the outer level. The correct (outer level) value is
2332        now given.
2333    
2334    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2335        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2336        make it do so. The string so far is captured, making this feature
2337        compatible with Perl.
2338    
2339    23. The tests have been re-organized, adding tests 11 and 12, to make it
2340        possible to check the Perl 5.10 features against Perl 5.10.
2341    
2342    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2343        pattern matches a fixed length string. PCRE did not allow this; now it
2344        does. Neither allows recursion.
2345    
2346    25. I finally figured out how to implement a request to provide the minimum
2347        length of subject string that was needed in order to match a given pattern.
2348        (It was back references and recursion that I had previously got hung up
2349        on.) This code has now been added to pcre_study(); it finds a lower bound
2350        to the length of subject needed. It is not necessarily the greatest lower
2351        bound, but using it to avoid searching strings that are too short does give
2352        some useful speed-ups. The value is available to calling programs via
2353        pcre_fullinfo().
2354    
2355    26. While implementing 25, I discovered to my embarrassment that pcretest had
2356        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2357        study optimizations had never been tested with that matching function.
2358        Oops. What is worse, even when it was passed study data, there was a bug in
2359        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2360        were also very few tests of studied patterns with pcre_dfa_exec().
2361    
2362    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2363        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2364        on the other side of the coin, they are no longer allowed to have different
2365        names, because these cannot be distinguished in PCRE, and this has caused
2366        confusion. (This is a difference from Perl.)
2367    
2368    28. When duplicate subpattern names are present (necessarily with different
2369        numbers, as required by 27 above), and a test is made by name in a
2370        conditional pattern, either for a subpattern having been matched, or for
2371        recursion in such a pattern, all the associated numbered subpatterns are
2372        tested, and the overall condition is true if the condition is true for any
2373        one of them. This is the way Perl works, and is also more like the way
2374        testing by number works.
2375    
2376    
2377    Version 7.9 11-Apr-09
2378    ---------------------
2379    
2380    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2381        (pcretest), all targets were linked against these libraries. This included
2382        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2383        libraries. This caused unwanted dependencies to be created. This problem
2384        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2385        pcretest is linked with readline.
2386    
2387    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2388        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2389        moved outside it again, because FALSE and TRUE are already defined in AIX,
2390        but BOOL is not.
2391    
2392    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2393        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2394    
2395    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2396        hyphens) following file names and line numbers when outputting matching
2397        lines. This is not true; no spaces are inserted. I have also clarified the
2398        wording for the --colour (or --color) option.
2399    
2400    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2401        was not coloured; this is different to GNU grep, so I have changed it to be
2402        the same.
2403    
2404    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2405        each matching line was coloured. Now it goes on to look for further matches
2406        of any of the test patterns, which is the same behaviour as GNU grep.
2407    
2408    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2409        doesn't make sense to accept an empty string match in pcregrep, so I have
2410        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2411        seems to be how GNU grep behaves. [But see later change 40 for release
2412        8.33.]
2413    
2414    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2415        start or after a newline", because the conditional assertion was not being
2416        correctly handled. The rule now is that both the assertion and what follows
2417        in the first alternative must satisfy the test.
2418    
2419    9.  If auto-callout was enabled in a pattern with a conditional group whose
2420        condition was an assertion, PCRE could crash during matching, both with
2421        pcre_exec() and pcre_dfa_exec().
2422    
2423    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2424        used for matching.
2425    
2426    11. Unicode property support in character classes was not working for
2427        characters (bytes) greater than 127 when not in UTF-8 mode.
2428    
2429    12. Added the -M command line option to pcretest.
2430    
2431    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2432    
2433    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2434    
2435    16. Added comments and documentation about mis-use of no_arg in the C++
2436        wrapper.
2437    
2438    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2439        from Martin Jerabek that uses macro names for all relevant character and
2440        string constants.
2441    
2442    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2443        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2444        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2445        these, but not everybody uses configure.
2446    
2447    19. A conditional group that had only one branch was not being correctly
2448        recognized as an item that could match an empty string. This meant that an
2449        enclosing group might also not be so recognized, causing infinite looping
2450        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2451        with the subject "ab", where knowledge that the repeated group can match
2452        nothing is needed in order to break the loop.
2453    
2454    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2455        exec(), but without supplying a callout function, matching went wrong.
2456    
2457    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2458        leak if the size of the offset vector was greater than 30. When the vector
2459        is smaller, the saved offsets during recursion go onto a local stack
2460        vector, but for larger vectors malloc() is used. It was failing to free
2461        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2462        error, in fact).
2463    
2464    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2465        heapframe that is used only when UTF-8 support is enabled. This caused no
2466        problem, but was untidy.
2467    
2468    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2469        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2470        included within another project.
2471    
2472    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2473        slightly modified by me:
2474    
2475          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2476              not building pcregrep.
2477    
2478          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2479              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2480    
2481    25. Forward references, both numeric and by name, in patterns that made use of
2482        duplicate group numbers, could behave incorrectly or give incorrect errors,
2483        because when scanning forward to find the reference group, PCRE was not
2484        taking into account the duplicate group numbers. A pattern such as
2485        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2486    
2487    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2488        the feature of a custom pointer more persuasive (as requested by a user).
2489    
2490    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2491        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2492        pre-defined.
2493    
2494    28. Added support for (*UTF8) at the start of a pattern.
2495    
2496    29. Arrange for flags added by the "release type" setting in CMake to be shown
2497        in the configuration summary.
2498    
2499    
2500    Version 7.8 05-Sep-08
2501    ---------------------
2502    
2503    1.  Replaced UCP searching code with optimized version as implemented for Ad
2504        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2505        stage table and inline lookup instead of a function, giving speed ups of 2
2506        to 5 times on some simple patterns that I tested. Permission was given to
2507        distribute the MultiStage2.py script that generates the tables (it's not in
2508        the tarball, but is in the Subversion repository).
2509    
2510    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2511        scripts.
2512    
2513    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2514        a group with a zero qualifier. The result of the study could be incorrect,
2515        or the function might crash, depending on the pattern.
2516    
2517    4.  Caseless matching was not working for non-ASCII characters in back
2518        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2519        It now works when Unicode Property Support is available.
2520    
2521    5.  In pcretest, an escape such as \x{de} in the data was always generating
2522        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2523        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2524        truncation.
2525    
2526    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2527    
2528    7.  Added two (int) casts to pcregrep when printing the difference of two
2529        pointers, in case they are 64-bit values.
2530    
2531    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2532        test 2 if it fails.
2533    
2534    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2535        and a #define of that name to empty if it is not externally set. This is to
2536        allow users of MSVC to set it if necessary.
2537    
2538    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2539        the convenience functions in the pcre_get.c source file.
2540    
2541    11. An option change at the start of a pattern that had top-level alternatives
2542        could cause overwriting and/or a crash. This command provoked a crash in
2543        some environments:
2544    
2545          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2546    
2547        This potential security problem was recorded as CVE-2008-2371.
2548    
2549    12. For a pattern where the match had to start at the beginning or immediately
2550        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2551        pcre_dfa_exec() could read past the end of the passed subject if there was
2552        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2553        pcretest so that it places the subject at the end of its malloc-ed buffer.
2554    
2555    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2556        exec() might read past the end of the data buffer in UTF-8 mode.
2557    
2558    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2559        the data contained the byte 0x85 as part of a UTF-8 character within its
2560        first line. This applied both to normal and DFA matching.
2561    
2562    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2563        /^[^d]*?$/8 failed to match "abc".
2564    
2565    16. Added a missing copyright notice to pcrecpp_internal.h.
2566    
2567    17. Make it more clear in the documentation that values returned from
2568        pcre_exec() in ovector are byte offsets, not character counts.
2569    
2570    18. Tidied a few places to stop certain compilers from issuing warnings.
2571    
2572    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2573        supplied by Stefan Weber. I made a further small update for 7.8 because
2574        there is a change of source arrangements: the pcre_searchfuncs.c module is
2575        replaced by pcre_ucd.c.
2576    
2577    
2578    Version 7.7 07-May-08
2579    ---------------------
2580    
2581    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2582        a string to a long long, pretend we don't even have a long long." This is
2583        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2584    
2585    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2586        pre-7.6 versions, which defined a global no_arg variable instead of putting
2587        it in the RE class. (See also #8 below.)
2588    
2589    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2590        Lopes.
2591    
2592    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2593    
2594        (1) The include/exclude patterns were being applied to the whole pathnames
2595            of files, instead of just to the final components.
2596    
2597        (2) If there was more than one level of directory, the subdirectories were
2598            skipped unless they satisfied the include/exclude conditions. This is
2599            inconsistent with GNU grep (and could even be seen as contrary to the
2600            pcregrep specification - which I improved to make it absolutely clear).
2601            The action now is always to scan all levels of directory, and just
2602            apply the include/exclude patterns to regular files.
2603    
2604    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2605        --exclude_dir in the tests to avoid scanning .svn directories.
2606    
2607    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2608        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2609        doesn't support NULs in patterns.
2610    
2611    7.  Added some missing "const"s to declarations of static tables in
2612        pcre_compile.c and pcre_dfa_exec.c.
2613    
2614    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2615        caused by fix #2  above. (Subsequently also a second patch to fix the
2616        first patch. And a third patch - this was a messy problem.)
2617    
2618    9.  Applied Craig's patch to remove the use of push_back().
2619    
2620    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2621        matching function regexec().
2622    
2623    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2624        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2625        references. PCRE supports relative numbers with this syntax (I don't think
2626        Oniguruma does).
2627    
2628    12. Previously, a group with a zero repeat such as (...){0} was completely
2629        omitted from the compiled regex. However, this means that if the group
2630        was called as a subroutine from elsewhere in the pattern, things went wrong
2631        (an internal error was given). Such groups are now left in the compiled
2632        pattern, with a new opcode that causes them to be skipped at execution
2633        time.
2634    
2635    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2636        to the way PCRE behaves:
2637    
2638        (a) A lone ] character is dis-allowed (Perl treats it as data).
2639    
2640        (b) A back reference to an unmatched subpattern matches an empty string
2641            (Perl fails the current match path).
2642    
2643        (c) A data ] in a character class must be notated as \] because if the
2644            first data character in a class is ], it defines an empty class. (In
2645            Perl it is not possible to have an empty class.) The empty class []
2646            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2647            The negative empty class [^] matches any one character, independently
2648            of the DOTALL setting.
2649    
2650    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2651        non-existent subpattern following a character class starting with ']' and
2652        containing () gave an internal compiling error instead of "reference to
2653        non-existent subpattern". Fortunately, when the pattern did exist, the
2654        compiled code was correct. (When scanning forwards to check for the
2655        existence of the subpattern, it was treating the data ']' as terminating
2656        the class, so got the count wrong. When actually compiling, the reference
2657        was subsequently set up correctly.)
2658    
2659    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2660        it was being rejected as not supported by pcre_dfa_exec(), even though
2661        other assertions are supported. I have made pcre_dfa_exec() support
2662        (*FAIL).
2663    
2664    16. The implementation of 13c above involved the invention of a new opcode,
2665        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2666        cannot be changed at match time, I realized I could make a small
2667        improvement to matching performance by compiling OP_ALLANY instead of
2668        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2669        on the OP_ANY path.
2670    
2671    17. Compiling pcretest on Windows with readline support failed without the
2672        following two fixes: (1) Make the unistd.h include conditional on
2673        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2674    
2675    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2676        ncurses library to be included for pcretest when ReadLine support is
2677        requested, but also to allow for it to be overridden. This patch came from
2678        Daniel Bergström.
2679    
2680    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2681        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2682        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2683        spotting this.
2684    
2685    
2686    Version 7.6 28-Jan-08
2687    ---------------------
2688    
2689    1.  A character class containing a very large number of characters with
2690        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2691        overflow.
2692    
2693    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2694        HAVE_LONG_LONG is not defined.
2695    
2696    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2697        bring them up to date and include new features. This patch includes:
2698    
2699        - Fixed PH's badly added libz and libbz2 support.
2700        - Fixed a problem with static linking.
2701        - Added pcredemo. [But later removed - see 7 below.]
2702        - Fixed dftables problem and added an option.
2703        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2704            HAVE_LONG_LONG.
2705        - Added readline support for pcretest.
2706        - Added an listing of the option settings after cmake has run.
2707    
2708    4.  A user submitted a patch to Makefile that makes it easy to create
2709        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2710        Makefile.am that cause it to include this special target, without
2711        affecting anything else. Note that the same mingw target plus all
2712        the other distribution libraries and programs are now supported
2713        when configuring with CMake (see 6 below) instead of with
2714        Configure/Make.
2715    
2716    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2717        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2718        exported in the Windows port". It has not yet been confirmed that the patch
2719        solves the problem, but it does no harm.
2720    
2721    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2722        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2723        with CMake, and also correct the comment about stack recursion.
2724    
2725    7.  Remove the automatic building of pcredemo from the ./configure system and
2726        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2727        of a program that users should build themselves after PCRE is installed, so
2728        building it automatically is not really right. What is more, it gave
2729        trouble in some build environments.
2730    
2731    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2732    
2733    
2734    Version 7.5 10-Jan-08
2735    ---------------------
2736    
2737    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2738        values in parens when parsing an RE using the C++ wrapper."
2739    
2740    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2741        Characters greater than 255 were excluded from the class instead of being
2742        included.
2743    
2744    3.  The same bug as (2) above applied to negated POSIX classes such as
2745        [:^space:].
2746    
2747    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2748        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2749        I have changed it.
2750    
2751    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2752        first named subpattern) and a construct such as (?&a) would reference the
2753        first named subpattern whose name started with "a" (in other words, the
2754        length check was missing). Both these problems are fixed. "Subpattern name
2755        expected" is now given for (?&) (a zero-length name), and this patch also
2756        makes it give the same error for \k'' (previously it complained that that
2757        was a reference to a non-existent subpattern).
2758    
2759    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2760        this is right because (?- can be followed by option settings as well as by
2761        digits. I have, however, made the messages clearer.
2762    
2763    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2764        than the number used in the conditional) now cause a compile-time error.
2765        This is actually not compatible with Perl, which accepts such patterns, but
2766        treats the conditional as always being FALSE (as PCRE used to), but it
2767        seems to me that giving a diagnostic is better.
2768    
2769    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2770        and messages.
2771    
2772    9.  Fix two occurrences of "backslash" in comments that should have been
2773        "backspace".
2774    
2775    10. Remove two redundant lines of code that can never be obeyed (their function
2776        was moved elsewhere).
2777    
2778    11. The program that makes PCRE's Unicode character property table had a bug
2779        which caused it to generate incorrect table entries for sequences of
2780        characters that have the same character type, but are in different scripts.
2781        It amalgamated them into a single range, with the script of the first of
2782        them. In other words, some characters were in the wrong script. There were
2783        thirteen such cases, affecting characters in the following ranges:
2784    
2785          U+002b0 - U+002c1
2786          U+0060c - U+0060d
2787          U+0061e - U+00612
2788          U+0064b - U+0065e
2789          U+0074d - U+0076d
2790          U+01800 - U+01805
2791          U+01d00 - U+01d77
2792          U+01d9b - U+01dbf
2793          U+0200b - U+0200f
2794          U+030fc - U+030fe
2795          U+03260 - U+0327f
2796          U+0fb46 - U+0fbb1
2797          U+10450 - U+1049d
2798    
2799    12. The -o option (show only the matching part of a line) for pcregrep was not
2800        compatible with GNU grep in that, if there was more than one match in a
2801        line, it showed only the first of them. It now behaves in the same way as
2802        GNU grep.
2803    
2804    13. If the -o and -v options were combined for pcregrep, it printed a blank
2805        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2806        does the same. The return code can be used to tell if there were any
2807        non-matching lines.
2808    
2809    14. Added --file-offsets and --line-offsets to pcregrep.
2810    
2811    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2812        infinitely looping recursion. The bug was that positive lookaheads were not
2813        being skipped when checking for a possible empty match (negative lookaheads
2814        and both kinds of lookbehind were skipped).
2815    
2816    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2817        inclusion of <windows.h> to before rather than after the definition of
2818        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2819    
2820    17. Specifying a possessive quantifier with a specific limit for a Unicode
2821        character property caused pcre_compile() to compile bad code, which led at
2822        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2823        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2824        caused the error; without that there was no problem.
2825    
2826    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2827    
2828    19. Added --enable-pcretest-libreadline.
2829    
2830    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2831        RE::GlobalReplace(). As a result, the number of replacements returned was
2832        double what it should be. I removed one of the increments, but Craig sent a
2833        later patch that removed the other one (the right fix) and added unit tests
2834        that check the return values (which was not done before).
2835    
2836    21. Several CMake things:
2837    
2838        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2839            the names libpcre and libpcreposix, not just pcre and pcreposix.
2840    
2841        (2) The above change means that pcretest and pcregrep are now correctly
2842            linked with the newly-built libraries, not previously installed ones.
2843    
2844        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2845    
2846    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2847        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2848        UTF-8 newline character). The key issue is that the pattern starts .*;
2849        this means that the match must be either at the beginning, or after a
2850        newline. The bug was in the code for advancing after a failed match and
2851        checking that the new position followed a newline. It was not taking
2852        account of UTF-8 characters correctly.
2853    
2854    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2855        character classes. PCRE was not treating the sequence [:...:] as a
2856        character class unless the ... were all letters. Perl, however, seems to
2857        allow any characters between [: and :], though of course it rejects as
2858        unknown any "names" that contain non-letters, because all the known class
2859        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2860        for example, whereas PCRE did not - it did not recognize a POSIX character
2861        class. This seemed a bit dangerous, so the code has been changed to be
2862        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2863        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2864        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2865        Perl does, and where it didn't before.
2866    
2867    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2868        Windows environments %n is disabled by default.
2869    
2870    
2871    Version 7.4 21-Sep-07
2872    ---------------------
2873    
2874    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2875        means that a class such as [\s] counted as "explicit reference to CR or
2876        LF". That isn't really right - the whole point of the change was to try to
2877        help when there was an actual mention of one of the two characters. So now
2878        the change happens only if \r or \n (or a literal CR or LF) character is
2879        encountered.
2880    
2881    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2882        of both had grown to the point where there were only 3 bits left.
2883        Fortunately, there was spare space in the data structure, and so I have
2884        moved the internal flags into a new 16-bit field to free up more option
2885        bits.
2886    
2887    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2888        but did not set the internal JCHANGED flag - either of these is enough to
2889        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2890        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2891        start sets both bits.
2892    
2893    4.  Added options (at build time, compile time, exec time) to change \R from
2894        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2895    
2896    5.  doc/pcresyntax.html was missing from the distribution.
2897    
2898    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2899        compatibility, even though it is no longer used.
2900    
2901    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2902        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2903        windows.h file is present (where different names are used). [This was
2904        reversed later after testing - see 16 below.]
2905    
2906    8.  Changed all #include <config.h> to #include "config.h". There were also
2907        some further <pcre.h> cases that I changed to "pcre.h".
2908    
2909    9.  When pcregrep was used with the --colour option, it missed the line ending
2910        sequence off the lines that it output.
2911    
2912    10. It was pointed out to me that arrays of string pointers cause lots of
2913        relocations when a shared library is dynamically loaded. A technique of
2914        using a single long string with a table of offsets can drastically reduce
2915        these. I have refactored PCRE in four places to do this. The result is
2916        dramatic:
2917    
2918          Originally:                          290
2919          After changing UCP table:            187
2920          After changing error message table:   43
2921          After changing table of "verbs"       36
2922          After changing table of Posix names   22
2923    
2924        Thanks to the folks working on Gregex for glib for this insight.
2925    
2926    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2927        unicode-properties was also set.
2928    
2929    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2930    
2931    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2932        checked only for CRLF.
2933    
2934    14. Added casts to pcretest.c to avoid compiler warnings.
2935    
2936    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2937    
2938    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2939        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2940        entirely. This removes changes made in 7 above.
2941    
2942    17. The CMake files have been updated, and there is now more information about
2943        building with CMake in the NON-UNIX-USE document.
2944    
2945    
2946    Version 7.3 28-Aug-07
2947    ---------------------
2948    
2949     1. In the rejigging of the build system that eventually resulted in 7.1, the
2950        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2951        brackets there is not right, since it causes compilers to look for an
2952        installed pcre.h, not the version that is in the source that is being
2953        compiled (which of course may be different). I have changed it back to:
2954    
2955          #include "pcre.h"
2956    
2957        I have a vague recollection that the change was concerned with compiling in
2958        different directories, but in the new build system, that is taken care of
2959        by the VPATH setting the Makefile.
2960    
2961     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2962        when the subject happened to end in the byte 0x85 (e.g. if the last
2963        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2964        characters but of course it shouldn't be taken as a newline when it is part
2965        of another character. The bug was that, for an unlimited repeat of . in
2966        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2967        characters when looking for a newline.
2968    
2969     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2970    
2971     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2972        in debug output.
2973    
2974     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2975        long printing in the pcrecpp unittest when running under MinGW.
2976    
2977     6. ESC_K was left out of the EBCDIC table.
2978    
2979     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2980        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2981        limit also applies to "virtual nesting" when a pattern is recursive, and in
2982        this case 1000 isn't so big. I have been able to remove this limit at the
2983        expense of backing off one optimization in certain circumstances. Normally,
2984        when pcre_exec() would call its internal match() function recursively and
2985        immediately return the result unconditionally, it uses a "tail recursion"
2986        feature to save stack. However, when a subpattern that can match an empty
2987        string has an unlimited repetition quantifier, it no longer makes this
2988        optimization. That gives it a stack frame in which to save the data for
2989        checking that an empty string has been matched. Previously this was taken
2990        from the 1000-entry workspace that had been reserved. So now there is no
2991        explicit limit, but more stack is used.
2992    
2993     8. Applied Daniel's patches to solve problems with the import/export magic
2994        syntax that is required for Windows, and which was going wrong for the
2995        pcreposix and pcrecpp parts of the library. These were overlooked when this
2996        problem was solved for the main library.
2997    
2998     9. There were some crude static tests to avoid integer overflow when computing
2999        the size of patterns that contain repeated groups with explicit upper
3000        limits. As the maximum quantifier is 65535, the maximum group length was
3001        set at 30,000 so that the product of these two numbers did not overflow a
3002        32-bit integer. However, it turns out that people want to use groups that
3003        are longer than 30,000 bytes (though not repeat them that many times).
3004        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3005        made it possible to implement the integer overflow checks in a much more
3006        dynamic way, which I have now done. The artificial limitation on group
3007        length has been removed - we now have only the limit on the total length of
3008        the compiled pattern, which depends on the LINK_SIZE setting.
3009    
3010    10. Fixed a bug in the documentation for get/copy named substring when
3011        duplicate names are permitted. If none of the named substrings are set, the
3012        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3013        empty string.
3014    
3015    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3016        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3017        because the ] is interpreted as the first data character and the
3018        terminating ] is not found. PCRE has been made compatible with Perl in this
3019        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3020        cause memory overwriting.
3021    
3022    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3023        string has been matched (to stop an infinite loop). It was not recognizing
3024        a conditional subpattern that could match an empty string if that
3025        subpattern was within another subpattern. For example, it looped when
3026        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3027        condition was not nested. This bug has been fixed.
3028    
3029    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3030        past the start of the subject in the presence of bytes with the top bit
3031        set, for example "\x8aBCD".
3032    
3033    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3034        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3035    
3036    14. Optimized (?!) to (*FAIL).
3037    
3038    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3039        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3040        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3041        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3042        does: it's just the validity check that is more restrictive.
3043    
3044    16. Inserted checks for integer overflows during escape sequence (backslash)
3045        processing, and also fixed erroneous offset values for syntax errors during
3046        backslash processing.
3047    
3048    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3049        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3050    
3051    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3052        caused an overrun.
3053    
3054    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3055        something other than just ASCII characters) inside a group that had an
3056        unlimited repeat caused a loop at compile time (while checking to see
3057        whether the group could match an empty string).
3058    
3059    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3060        [\P{Any}] did so. (Error in the code for printing property names.)
3061    
3062    21. An orphan \E inside a character class could cause a crash.
3063    
3064    22. A repeated capturing bracket such as (A)? could cause a wild memory
3065        reference during compilation.
3066    
3067    23. There are several functions in pcre_compile() that scan along a compiled
3068        expression for various reasons (e.g. to see if it's fixed length for look
3069        behind). There were bugs in these functions when a repeated \p or \P was
3070        present in the pattern. These operators have additional parameters compared
3071        with \d, etc, and these were not being taken into account when moving along
3072        the compiled data. Specifically:
3073    
3074        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3075            length.
3076    
3077        (b) An item such as \pL+ within a repeated group could cause crashes or
3078            loops.
3079    
3080        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3081            "reference to non-existent subpattern" error.
3082    
3083        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3084    
3085    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3086        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3087    
3088    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3089    
3090    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3091        character were causing crashes (broken optimization).
3092    
3093    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3094        \p or \P) caused a compile-time loop.
3095    
3096    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3097        break. For example, the unstudied pattern [\r\n]A does not match the string
3098        "\r\nA" because change 7.0/46 below moves the current point on by two
3099        characters after failing to match at the start. However, the pattern \nA
3100        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3101        the same is true. There doesn't seem any very clean way out of this, but
3102        what I have chosen to do makes the common cases work: PCRE now takes note
3103        of whether there can be an explicit match for \r or \n anywhere in the
3104        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3105        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3106        pattern has explicit CR or LF references.
3107    
3108    29. Added (*CR) etc for changing newline setting at start of pattern.
3109    
3110    
3111    Version 7.2 19-Jun-07
3112    ---------------------
3113    
3114     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3115        which is apparently normally available under Windows.
3116    
3117     2. Re-jig the pcregrep tests with different newline settings in an attempt
3118        to make them independent of the local environment's newline setting.
3119    
3120     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3121    
3122     4. Some of the "internals" tests were previously cut out when the link size
3123        was not 2, because the output contained actual offsets. The recent new
3124        "Z" feature of pcretest means that these can be cut out, making the tests
3125        usable with all link sizes.
3126    
3127     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3128        stack recursion. This gives a massive performance boost under BSD, but just
3129        a small improvement under Linux. However, it saves one field in the frame
3130        in all cases.
3131    
3132     6. Added more features from the forthcoming Perl 5.10:
3133    
3134        (a) (?-n) (where n is a string of digits) is a relative subroutine or
3135            recursion call. It refers to the nth most recently opened parentheses.
3136    
3137        (b) (?+n) is also a relative subroutine call; it refers to the nth next
3138            to be opened parentheses.
3139    
3140        (c) Conditions that refer to capturing parentheses can be specified
3141            relatively, for example, (?(-2)... or (?(+3)...
3142    
3143        (d) \K resets the start of the current match so that everything before
3144            is not part of it.
3145    
3146        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3147    
3148        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3149            reference syntax.
3150    
3151        (g) (?| introduces a group in which the numbering of parentheses in each
3152            alternative starts with the same number.
3153    
3154        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3155    
3156     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3157        PCRE_INFO_JCHANGED.
3158    
3159     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3160        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3161        for detecting groups that can match an empty string.
3162    
3163     9. A pattern with a very large number of alternatives (more than several
3164        hundred) was running out of internal workspace during the pre-compile
3165        phase, where pcre_compile() figures out how much memory will be needed. A
3166        bit of new cunning has reduced the workspace needed for groups with
3167        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3168        workspace instead of running out of the 4096 that are available.
3169    
3170    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3171    
3172    11. Applied patch from Google to remove an optimization that didn't quite work.
3173        The report of the bug said:
3174    
3175          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3176          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3177          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3178    
3179    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3180        it matched the wrong number of bytes.
3181    
3182    
3183    Version 7.1 24-Apr-07
3184    ---------------------
3185    
3186     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3187        that is more "standard", making use of automake and other Autotools. There
3188        is some re-arrangement of the files and adjustment of comments consequent
3189        on this.
3190    
3191     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3192        for recursive directory scanning broke on some systems because the files
3193        are not scanned in any specific order and on different systems the order
3194        was different. A call to "sort" has been inserted into RunGrepTest for the
3195        approprate test as a short-term fix. In the longer term there may be an
3196        alternative.
3197    
3198     3. I had an email from Eric Raymond about problems translating some of PCRE's
3199        man pages to HTML (despite the fact that I distribute HTML pages, some
3200        people do their own conversions for various reasons). The problems
3201        concerned the use of low-level troff macros .br and .in. I have therefore
3202        removed all such uses from the man pages (some were redundant, some could
3203        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3204        HTML has been updated to handle .nf/.fi and to complain if it encounters
3205        .br or .in.
3206    
3207     4. Updated comments in configure.ac that get placed in config.h.in and also
3208        arranged for config.h to be included in the distribution, with the name
3209        config.h.generic, for the benefit of those who have to compile without
3210        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3211    
3212     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3213        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3214        makevp.bat for the current PCRE, using the additional files
3215        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3216    
3217     6. A Windows user reported a minor discrepancy with test 2, which turned out
3218        to be caused by a trailing space on an input line that had got lost in his
3219        copy. The trailing space was an accident, so I've just removed it.
3220    
3221     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3222        that is needed.
3223    
3224     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3225        as "const" (a) because they are and (b) because it helps the PHP
3226        maintainers who have recently made a script to detect big data structures
3227        in the php code that should be moved to the .rodata section. I remembered
3228        to update Builducptable as well, so it won't revert if ucptable.h is ever
3229        re-created.
3230    
3231     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3232        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3233        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3234        support is not required. This saves 1.5-2K of code, which is important in
3235        some applications.
3236    
3237        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3238        so as not to refer to the tables, even though these functions will never be
3239        called when UTF-8 support is disabled. Otherwise there are problems with a
3240        shared library.
3241    
3242    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3243    
3244        (a) It was defining its arguments as char * instead of void *.
3245    
3246        (b) It was assuming that all moves were upwards in memory; this was true
3247            a long time ago when I wrote it, but is no longer the case.
3248    
3249        The emulated memove() is provided for those environments that have neither
3250        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3251        is clearly not the case, as these two bugs were recently reported.
3252    
3253    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3254        and Detrail to create the HTML documentation, the .txt form of the man
3255        pages, and it removes trailing spaces from listed files. It also creates
3256        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3257        case, it wraps all the #defines with #ifndefs. This script should be run
3258        before "make dist".
3259    
3260    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3261        with Unicode property support.
3262    
3263        (a) For a maximizing quantifier, if the two different cases of the
3264            character were of different lengths in their UTF-8 codings (there are
3265            some cases like this - I found 11), and the matching function had to
3266            back up over a mixture of the two cases, it incorrectly assumed they
3267            were both the same length.
3268    
3269        (b) When PCRE was configured to use the heap rather than the stack for
3270            recursion during matching, it was not correctly preserving the data for
3271            the other case of a UTF-8 character when checking ahead for a match
3272            while processing a minimizing repeat. If the check also involved
3273            matching a wide character, but failed, corruption could cause an
3274            erroneous result when trying to check for a repeat of the original
3275            character.
3276    
3277    13. Some tidying changes to the testing mechanism:
3278    
3279        (a) The RunTest script now detects the internal link size and whether there
3280            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3281            values substituted by "configure". (The RunGrepTest script already did
3282            this for UTF-8.) The configure.ac script no longer substitutes the
3283            relevant variables.
3284    
3285        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3286            with length and offset values. This means that the output is different
3287            for different internal link sizes. Test 2 is skipped for link sizes
3288            other than 2 because of this, bypassing the problem. Unfortunately,
3289            there was also a test in test 3 (the locale tests) that used /B and
3290            failed for link sizes other than 2. Rather than cut the whole test out,
3291            I have added a new /Z option to pcretest that replaces the length and
3292            offset values with spaces. This is now used to make test 3 independent
3293            of link size. (Test 2 will be tidied up later.)
3294    
3295    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3296        segmentation fault instead of returning the appropriate error message.
3297    
3298    15. In multiline mode when the newline sequence was set to "any", the pattern
3299        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3300        This doesn't seem right; it now treats the CRLF combination as the line
3301        ending, and so does not match in that case. It's only a pattern such as ^$
3302        that would hit this one: something like ^ABC$ would have failed after \r
3303        and then tried again after \r\n.
3304    
3305    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3306        in an attempt to make files that differ only in their line terminators
3307        compare equal. This works on Linux.
3308    
3309    17. Under certain error circumstances pcregrep might try to free random memory
3310        as it exited. This is now fixed, thanks to valgrind.
3311    
3312    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3313        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3314        was because its rules for how to advance for /g after matching an empty
3315        string at the end of a line did not allow for this case. They now check for
3316        it specially.
3317    
3318    20. pcretest is supposed to handle patterns and data of any length, by
3319        extending its buffers when necessary. It was getting this wrong when the
3320        buffer for a data line had to be extended.
3321    
3322    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3323        CRLF as a newline sequence.
3324    
3325    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3326        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3327        I have nevertheless tidied it up.
3328    
3329    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3330    
3331    24. Added a man page for pcre-config.
3332    
3333    
3334    Version 7.0 19-Dec-06
3335    ---------------------
3336    
3337     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3338        moving to gcc 4.1.1.
3339    
3340     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3341        sys/time.h, which is documented as needed for this function. It doesn't
3342        seem to matter on Linux, but it showed up on some releases of OS X.
3343    
3344     3. It seems that there are systems where bytes whose values are greater than
3345        127 match isprint() in the "C" locale. The "C" locale should be the
3346        default when a C program starts up. In most systems, only ASCII printing
3347        characters match isprint(). This difference caused the output from pcretest
3348        to vary, making some of the tests fail. I have changed pcretest so that:
3349    
3350        (a) When it is outputting text in the compiled version of a pattern, bytes
3351            other than 32-126 are always shown as hex escapes.
3352    
3353        (b) When it is outputting text that is a matched part of a subject string,
3354            it does the same, unless a different locale has been set for the match
3355            (using the /L modifier). In this case, it uses isprint() to decide.
3356    
3357     4. Fixed a major bug that caused incorrect computation of the amount of memory
3358        required for a compiled pattern when options that changed within the
3359        pattern affected the logic of the preliminary scan that determines the
3360        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3361        that the computed length was too small. The symptoms of this bug were
3362        either the PCRE error "internal error: code overflow" from pcre_compile(),
3363        or a glibc crash with a message such as "pcretest: free(): invalid next
3364        size (fast)". Examples of patterns that provoked this bug (shown in
3365        pcretest format) are:
3366    
3367          /(?-x: )/x
3368          /(?x)(?-x: \s*#\s*)/
3369          /((?i)[\x{c0}])/8
3370          /(?i:[\x{c0}])/8
3371    
3372        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3373        is now done differently.
3374    
3375     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3376        wrapper classes; (b) implement a new function in the C++ scanner that is
3377        more efficient than the old way of doing things because it avoids levels of
3378        recursion in the regex matching; (c) add a paragraph to the documentation
3379        for the FullMatch() function.
3380    
3381     6. The escape sequence \n was being treated as whatever was defined as
3382        "newline". Not only was this contrary to the documentation, which states
3383        that \n is character 10 (hex 0A), but it also went horribly wrong when
3384        "newline" was defined as CRLF. This has been fixed.
3385    
3386     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3387        was being set to -1 for the "end of line" case (supposedly a value that no
3388        character can have). Though this value is never used (the check for end of
3389        line is "zero bytes in current character"), it caused compiler complaints.
3390        I've changed it to 0xffffffff.
3391    
3392     8. In pcre_version.c, the version string was being built by a sequence of
3393        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3394        string (as it is for production releases) called a macro with an empty
3395        argument. The C standard says the result of this is undefined. The gcc
3396        compiler treats it as an empty string (which was what was wanted) but it is
3397        reported that Visual C gives an error. The source has been hacked around to
3398        avoid this problem.
3399    
3400     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3401        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3402        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3403        of them did).
3404    
3405    10. Originally, pcretest opened its input and output without "b"; then I was
3406        told that "b" was needed in some environments, so it was added for release
3407        5.0 to both the input and output. (It makes no difference on Unix-like
3408        systems.) Later I was told that it is wrong for the input on Windows. I've
3409        now abstracted the modes into two macros, to make it easier to fiddle with
3410        them, and removed "b" from the input mode under Windows.
3411    
3412    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3413    
3414    12. Added -help and --help to pcretest as an official way of being reminded
3415        of the options.
3416    
3417    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3418        and pcrecpp.cc because they annoy compilers at high warning levels.
3419    
3420    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3421    
3422    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3423        scripts are not C programs :-) and which was not noticed because it works
3424        on Linux.
3425    
3426    16. pcretest is supposed to handle any length of pattern and data line (as one
3427        line or as a continued sequence of lines) by extending its input buffer if
3428        necessary. This feature was broken for very long pattern lines, leading to
3429        a string of junk being passed to pcre_compile() if the pattern was longer
3430        than about 50K.
3431    
3432    17. I have done a major re-factoring of the way pcre_compile() computes the
3433        amount of memory needed for a compiled pattern. Previously, there was code
3434        that made a preliminary scan of the pattern in order to do this. That was
3435        OK when PCRE was new, but as the facilities have expanded, it has become
3436        harder and harder to keep it in step with the real compile phase, and there
3437        have been a number of bugs (see for example, 4 above). I have now found a
3438        cunning way of running the real compile function in a "fake" mode that
3439        enables it to compute how much memory it would need, while actually only
3440        ever using a few hundred bytes of working memory and without too many
3441        tests of the mode. This should make future maintenance and development
3442        easier. A side effect of this work is that the limit of 200 on the nesting
3443        depth of parentheses has been removed (though this was never a serious
3444        limitation, I suspect). However, there is a downside: pcre_compile() now
3445        runs more slowly than before (30% or more, depending on the pattern). I
3446        hope this isn't a big issue. There is no effect on runtime performance.
3447    
3448    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3449        newline (only possible for the last line of a file) and it was a
3450        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3451    
3452    19. Added additional timing features to pcretest. (1) The -tm option now times
3453        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3454        separate command line item, by a number that specifies the number of
3455        repeats to use when timing. The default is 50000; this gives better
3456        precision, but takes uncomfortably long for very large patterns.
3457    
3458    20. Extended pcre_study() to be more clever in cases where a branch of a
3459        subpattern has no definite first character. For example, (a*|b*)[cd] would
3460        previously give no result from pcre_study(). Now it recognizes that the
3461        first character must be a, b, c, or d.
3462    
3463    21. There was an incorrect error "recursive call could loop indefinitely" if
3464        a subpattern (or the entire pattern) that was being tested for matching an
3465        empty string contained only one non-empty item after a nested subpattern.
3466        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3467        incorrectly, because the \d was being skipped in the check.
3468    
3469    22. The pcretest program now has a new pattern option /B and a command line
3470        option -b, which is equivalent to adding /B to every pattern. This causes
3471        it to show the compiled bytecode, without the additional information that
3472        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3473        is the same as /B/I).
3474    
3475    23. A new optimization is now able automatically to treat some sequences such
3476        as a*b as a*+b. More specifically, if something simple (such as a character
3477        or a simple class like \d) has an unlimited quantifier, and is followed by
3478        something that cannot possibly match the quantified thing, the quantifier
3479        is automatically "possessified".
3480    
3481    24. A recursive reference to a subpattern whose number was greater than 39
3482        went wrong under certain circumstances in UTF-8 mode. This bug could also
3483        have affected the operation of pcre_study().
3484    
3485    25. Realized that a little bit of performance could be had by replacing
3486        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3487    
3488    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3489    
3490    27. Possessive quantifiers such as a++ were previously implemented by turning
3491        them into atomic groups such as ($>a+). Now they have their own opcodes,
3492        which improves performance. This includes the automatically created ones
3493        from 23 above.
3494    
3495    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3496        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3497        the first matched character to be a colon. This applied both to named and
3498        numbered groups.
3499    
3500    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3501    
3502    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3503        building PCRE on Windows easier, so I have included it in the distribution.
3504    
3505    31. There is now a check in pcretest against a ridiculously large number being
3506        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3507        loop, the loop is abandoned.
3508    
3509    32. Forward references to subpatterns in conditions such as (?(2)...) where
3510        subpattern 2 is defined later cause pcre_compile() to search forwards in
3511        the pattern for the relevant set of parentheses. This search went wrong
3512        when there were unescaped parentheses in a character class, parentheses
3513        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3514    
3515    33. "Subroutine" calls and backreferences were previously restricted to
3516        referencing subpatterns earlier in the regex. This restriction has now
3517        been removed.
3518    
3519    34. Added a number of extra features that are going to be in Perl 5.10. On the
3520        whole, these are just syntactic alternatives for features that PCRE had
3521        previously implemented using the Python syntax or my own invention. The
3522        other formats are all retained for compatibility.
3523    
3524        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3525            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3526            also .NET compatible.
3527    
3528        (b) A recursion or subroutine call to a named group can now be defined as
3529            (?&name) as well as (?P>name).
3530    
3531        (c) A backreference to a named group can now be defined as \k<name> or
3532            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3533            5.10, are also .NET compatible.
3534    
3535        (d) A conditional reference to a named group can now use the syntax
3536            (?(<name>) or (?('name') as well as (?(name).
3537    
3538        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3539            groups (named and numbered) that are never evaluated inline, but can be
3540            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3541            is always false. There may be only one alternative in such a group.
3542    
3543        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3544            as the simple (?(R). The condition is true only if the most recent
3545            recursion is that of the given number or name. It does not search out
3546            through the entire recursion stack.
3547    
3548        (g) The escape \gN or \g{N} has been added, where N is a positive or
3549            negative number, specifying an absolute or relative reference.
3550    
3551    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3552        some "unreachable code" warnings.
3553    
3554    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3555        things, this adds five new scripts.
3556    
3557    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3558        There were also incompatibilities regarding the handling of \Q..\E inside
3559        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3560        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3561    
3562    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3563        matches an empty string, and forcibly breaks the loop. There were bugs in
3564        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3565        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3566        separate and independent bugs (that affected different cases) have been
3567        fixed.
3568    
3569    39. Refactored the code to abolish the use of different opcodes for small
3570        capturing bracket numbers. This is a tidy that I avoided doing when I
3571        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3572        The new approach is not only tidier, it makes it possible to reduce the
3573        memory needed to fix the previous bug (38).
3574    
3575    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3576        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3577        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3578        mode.
3579    
3580    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3581        report.
3582    
3583    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3584        copy construction and assignment in the C++ wrapper.
3585