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

Diff of /code/trunk/ChangeLog

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

revision 65 by nigel, Sat Feb 24 21:40:08 2007 UTC revision 1558 by ph10, Fri May 15 17:17:03 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.38 xx-xxx-xxxx
5    ------------------------
6    
7    1.  If a group that contained a recursive back reference also contained a
8        forward reference subroutine call followed by a non-forward-reference
9        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
10        compile correct code, leading to undefined behaviour or an internally
11        detected error. This bug was discovered by the LLVM fuzzer.
12    
13    2.  Quantification of certain items (e.g. atomic back references) could cause
14        incorrect code to be compiled when recursive forward references were
15        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
16        This bug was discovered by the LLVM fuzzer.
17    
18    3.  A repeated conditional group whose condition was a reference by name caused
19        a buffer overflow if there was more than one group with the given name.
20        This bug was discovered by the LLVM fuzzer.
21    
22    4.  A recursive back reference by name within a group that had the same name as
23        another group caused a buffer overflow. For example:
24        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
25    
26    
27    Version 8.37 28-April-2015
28    --------------------------
29    
30    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
31        for those parentheses to be closed with whatever has been captured so far.
32        However, it was failing to mark any other groups between the hightest
33        capture so far and the currrent group as "unset". Thus, the ovector for
34        those groups contained whatever was previously there. An example is the
35        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
36    
37    2.  If an assertion condition was quantified with a minimum of zero (an odd
38        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
39    
40    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
41        unrecognized modifier, a crash could occur.
42    
43    4.  An attempt to do global matching in pcretest with a zero-length ovector
44        caused a crash.
45    
46    5.  Fixed a memory leak during matching that could occur for a subpattern
47        subroutine call (recursive or otherwise) if the number of captured groups
48        that had to be saved was greater than ten.
49    
50    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
51        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
52        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
53    
54    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
55        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
56    
57    8. When a pattern is compiled, it remembers the highest back reference so that
58       when matching, if the ovector is too small, extra memory can be obtained to
59       use instead. A conditional subpattern whose condition is a check on a
60       capture having happened, such as, for example in the pattern
61       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
62       setting the highest backreference number. This mattered only if pcre_exec()
63       was called with an ovector that was too small to hold the capture, and there
64       was no other kind of back reference (a situation which is probably quite
65       rare). The effect of the bug was that the condition was always treated as
66       FALSE when the capture could not be consulted, leading to a incorrect
67       behaviour by pcre_exec(). This bug has been fixed.
68    
69    9. A reference to a duplicated named group (either a back reference or a test
70       for being set in a conditional) that occurred in a part of the pattern where
71       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
72       to be incorrectly calculated, leading to overwriting.
73    
74    10. A mutually recursive set of back references such as (\2)(\1) caused a
75        segfault at study time (while trying to find the minimum matching length).
76        The infinite loop is now broken (with the minimum length unset, that is,
77        zero).
78    
79    11. If an assertion that was used as a condition was quantified with a minimum
80        of zero, matching went wrong. In particular, if the whole group had
81        unlimited repetition and could match an empty string, a segfault was
82        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
83        assertions to be quantified, but not if they are being used as conditions,
84        so the above pattern is faulted by Perl. PCRE has now been changed so that
85        it also rejects such patterns.
86    
87    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
88        failed to allow the zero-repeat case if pcre2_exec() was called with an
89        ovector too small to capture the group.
90    
91    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
92        Red Hat Product Security:
93    
94        (a) A crash if /K and /F were both set with the option to save the compiled
95        pattern.
96    
97        (b) Another crash if the option to print captured substrings in a callout
98        was combined with setting a null ovector, for example \O\C+ as a subject
99        string.
100    
101    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
102        forward reference repeated a large (but limited) number of times within a
103        repeated outer group that has a zero minimum quantifier, caused incorrect
104        code to be compiled, leading to the error "internal error:
105        previously-checked referenced subpattern not found" when an incorrect
106        memory address was read. This bug was reported as "heap overflow",
107        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
108        CVE-2015-2325.
109    
110    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
111        call within a group that also contained a recursive back reference caused
112        incorrect code to be compiled. This bug was reported as "heap overflow",
113        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
114        number CVE-2015-2326.
115    
116    24. Computing the size of the JIT read-only data in advance has been a source
117        of various issues, and new ones are still appear unfortunately. To fix
118        existing and future issues, size computation is eliminated from the code,
119        and replaced by on-demand memory allocation.
120    
121    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
122        adjacent to the end of the range, and the range contained characters with
123        more than one other case, caused incorrect behaviour when compiled in UTF
124        mode. In that example, the range a-j was left out of the class.
125    
126    26. Fix JIT compilation of conditional blocks, which assertion
127        is converted to (*FAIL). E.g: /(?(?!))/.
128    
129    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
130        discovered by the LLVM fuzzer.
131    
132    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
133        when this assertion was used as a condition, for example (?(?!)a|b). In
134        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
135        error about an unsupported item.
136    
137    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
138        possessification code could take exponential time to complete. A recursion
139        depth limit of 1000 has been imposed to limit the resources used by this
140        optimization.
141    
142    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
143        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
144        because \S ensures they are all in the class. The code for doing this was
145        interacting badly with the code for computing the amount of space needed to
146        compile the pattern, leading to a buffer overflow. This bug was discovered
147        by the LLVM fuzzer.
148    
149    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
150        other kinds of group caused stack overflow at compile time. This bug was
151        discovered by the LLVM fuzzer.
152    
153    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
154        between a subroutine call and its quantifier was incorrectly compiled,
155        leading to buffer overflow or other errors. This bug was discovered by the
156        LLVM fuzzer.
157    
158    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
159        assertion after (?(. The code was failing to check the character after
160        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
161        was discovered by the LLVM fuzzer.
162    
163    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
164        a fixed maximum following a group that contains a subroutine reference was
165        incorrectly compiled and could trigger buffer overflow. This bug was
166        discovered by the LLVM fuzzer.
167    
168    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
169        caused a stack overflow instead of the diagnosis of a non-fixed length
170        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
171    
172    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
173        (e.g. /(?<=\Ka)/) could make pcregrep loop.
174    
175    37. There was a similar problem to 36 in pcretest for global matches.
176    
177    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
178        and a subsequent item in the pattern caused a non-match, backtracking over
179        the repeated \X did not stop, but carried on past the start of the subject,
180        causing reference to random memory and/or a segfault. There were also some
181        other cases where backtracking after \C could crash. This set of bugs was
182        discovered by the LLVM fuzzer.
183    
184    39. The function for finding the minimum length of a matching string could take
185        a very long time if mutual recursion was present many times in a pattern,
186        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
187        method has been implemented. This infelicity was discovered by the LLVM
188        fuzzer.
189    
190    40. Static linking against the PCRE library using the pkg-config module was
191        failing on missing pthread symbols.
192    
193    
194    Version 8.36 26-September-2014
195    ------------------------------
196    
197    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
198        -Wmissing-field-initializers and -Wunused-parameter.
199    
200    2.  The tests for quantifiers being too big (greater than 65535) were being
201        applied after reading the number, and stupidly assuming that integer
202        overflow would give a negative number. The tests are now applied as the
203        numbers are read.
204    
205    3.  Tidy code in pcre_exec.c where two branches that used to be different are
206        now the same.
207    
208    4.  The JIT compiler did not generate match limit checks for certain
209        bracketed expressions with quantifiers. This may lead to exponential
210        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
211        issue should be resolved now.
212    
213    5.  Fixed an issue, which occures when nested alternatives are optimized
214        with table jumps.
215    
216    6.  Inserted two casts and changed some ints to size_t in the light of some
217        reported 64-bit compiler warnings (Bugzilla 1477).
218    
219    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
220        an empty string, which sometimes were behaving incorrectly in the
221        interpreter (though correctly in the JIT matcher). This pcretest input is
222        an example:
223    
224          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
225          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
226    
227        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
228        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
229        for an empty string was breaking the inner loop and carrying on at a lower
230        level, when possessive repeated groups should always return to a higher
231        level as they have no backtrack points in them. The empty string test now
232        occurs at the outer level.
233    
234    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
235        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
236    
237    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
238        doesn't).
239    
240    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
241        to hold a pattern to be incorrectly computed (too small) when there were
242        named back references to duplicated names. This could cause "internal
243        error: code overflow" or "double free or corruption" or other memory
244        handling errors.
245    
246    11. When named subpatterns had the same prefixes, back references could be
247        confused. For example, in this pattern:
248    
249          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
250    
251        the reference to 'Name' was incorrectly treated as a reference to a
252        duplicate name.
253    
254    12. A pattern such as /^s?c/mi8 where the optional character has more than
255        one "other case" was incorrectly compiled such that it would only try to
256        match starting at "c".
257    
258    13. When a pattern starting with \s was studied, VT was not included in the
259        list of possible starting characters; this should have been part of the
260        8.34/18 patch.
261    
262    14. If a character class started [\Qx]... where x is any character, the class
263        was incorrectly terminated at the ].
264    
265    15. If a pattern that started with a caseless match for a character with more
266        than one "other case" was studied, PCRE did not set up the starting code
267        unit bit map for the list of possible characters. Now it does. This is an
268        optimization improvement, not a bug fix.
269    
270    16. The Unicode data tables have been updated to Unicode 7.0.0.
271    
272    17. Fixed a number of memory leaks in pcregrep.
273    
274    18. Avoid a compiler warning (from some compilers) for a function call with
275        a cast that removes "const" from an lvalue by using an intermediate
276        variable (to which the compiler does not object).
277    
278    19. Incorrect code was compiled if a group that contained an internal recursive
279        back reference was optional (had quantifier with a minimum of zero). This
280        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
281        caused segmentation faults because of stack overflows at compile time.
282    
283    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
284        group that is quantified with an indefinite repeat, caused a compile-time
285        loop which used up all the system stack and provoked a segmentation fault.
286        This was not the same bug as 19 above.
287    
288    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
289        Patch by Mike Frysinger.
290    
291    
292    Version 8.35 04-April-2014
293    --------------------------
294    
295    1.  A new flag is set, when property checks are present in an XCLASS.
296        When this flag is not set, PCRE can perform certain optimizations
297        such as studying these XCLASS-es.
298    
299    2.  The auto-possessification of character sets were improved: a normal
300        and an extended character set can be compared now. Furthermore
301        the JIT compiler optimizes more character set checks.
302    
303    3.  Got rid of some compiler warnings for potentially uninitialized variables
304        that show up only when compiled with -O2.
305    
306    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
307        of a match later then the end of the match. The pcretest program was not
308        handling the case sensibly - it was outputting from the start to the next
309        binary zero. It now reports this situation in a message, and outputs the
310        text from the end to the start.
311    
312    5.  Fast forward search is improved in JIT. Instead of the first three
313        characters, any three characters with fixed position can be searched.
314        Search order: first, last, middle.
315    
316    6.  Improve character range checks in JIT. Characters are read by an inprecise
317        function now, which returns with an unknown value if the character code is
318        above a certain threshold (e.g: 256). The only limitation is that the value
319        must be bigger than the threshold as well. This function is useful when
320        the characters above the threshold are handled in the same way.
321    
322    7.  The macros whose names start with RAWUCHAR are placeholders for a future
323        mode in which only the bottom 21 bits of 32-bit data items are used. To
324        make this more memorable for those maintaining the code, the names have
325        been changed to start with UCHAR21, and an extensive comment has been added
326        to their definition.
327    
328    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
329        to the export list in Makefile.am (they were accidentally omitted from the
330        8.34 tarball).
331    
332    9.  The informational output from pcretest used the phrase "starting byte set"
333        which is inappropriate for the 16-bit and 32-bit libraries. As the output
334        for "first char" and "need char" really means "non-UTF-char", I've changed
335        "byte" to "char", and slightly reworded the output. The documentation about
336        these values has also been (I hope) clarified.
337    
338    10. Another JIT related optimization: use table jumps for selecting the correct
339        backtracking path, when more than four alternatives are present inside a
340        bracket.
341    
342    11. Empty match is not possible, when the minimum length is greater than zero,
343        and there is no \K in the pattern. JIT should avoid empty match checks in
344        such cases.
345    
346    12. In a caseless character class with UCP support, when a character with more
347        than one alternative case was not the first character of a range, not all
348        the alternative cases were added to the class. For example, s and \x{17f}
349        are both alternative cases for S: the class [RST] was handled correctly,
350        but [R-T] was not.
351    
352    13. The configure.ac file always checked for pthread support when JIT was
353        enabled. This is not used in Windows, so I have put this test inside a
354        check for the presence of windows.h (which was already tested for).
355    
356    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
357        The algorithm provides a way to skip certain starting offsets, and usually
358        faster than linear prefix searches.
359    
360    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
361        as for 'fr_FR' and 'french'. For some reason, however, it then used the
362        Windows-specific input and output files, which have 'french' screwed in.
363        So this could never have worked. One of the problems with locales is that
364        they aren't always the same. I have now updated RunTest so that it checks
365        the output of the locale test (test 3) against three different output
366        files, and it allows the test to pass if any one of them matches. With luck
367        this should make the test pass on some versions of Solaris where it was
368        failing. Because of the uncertainty, the script did not used to stop if
369        test 3 failed; it now does. If further versions of a French locale ever
370        come to light, they can now easily be added.
371    
372    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
373        there was a message during ./configure, but it did not stop. This now
374        provokes an error. The invalid example in README has been corrected.
375        If a value less than the minimum is given, the minimum value has always
376        been used, but now a warning is given.
377    
378    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
379        was a bug in the test system, which is now fixed. Also, the list of various
380        configurations that are tested for each release did not have one with both
381        16/32 bits and --enable-bar-anycrlf. It now does.
382    
383    18. pcretest was missing "-C bsr" for displaying the \R default setting.
384    
385    19. Little endian PowerPC systems are supported now by the JIT compiler.
386    
387    20. The fast forward newline mechanism could enter to an infinite loop on
388        certain invalid UTF-8 input. Although we don't support these cases
389        this issue can be fixed by a performance optimization.
390    
391    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
392        not take account if existing stack usage. There is now a new global
393        variable called pcre_stack_guard that can be set to point to an external
394        function to check stack availability. It is called at the start of
395        processing every parenthesized group.
396    
397    22. A typo in the code meant that in ungreedy mode the max/min qualifier
398        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
399        not match "ab".
400    
401    23. When UTF was disabled, the JIT program reported some incorrect compile
402        errors. These messages are silenced now.
403    
404    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
405        compiler.
406    
407    25. Change all the temporary files used in RunGrepTest to be different to those
408        used by RunTest so that the tests can be run simultaneously, for example by
409        "make -j check".
410    
411    
412    Version 8.34 15-December-2013
413    -----------------------------
414    
415    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
416        executable memory. Patch inspired by Carsten Klein.
417    
418    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
419        this macro is never tested and has no effect, because the work to support
420        coverage involves only compiling and linking options and special targets in
421        the Makefile. The comment in config.h implied that defining the macro would
422        enable coverage support, which is totally false. There was also support for
423        setting this macro in the CMake files (my fault, I just copied it from
424        configure). SUPPORT_GCOV has now been removed.
425    
426    3.  Make a small performance improvement in strlen16() and strlen32() in
427        pcretest.
428    
429    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
430        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
431        these cases). There was also one in pcretest.c.
432    
433    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
434    
435    6.  In UTF mode, the code for checking whether a group could match an empty
436        string (which is used for indefinitely repeated groups to allow for
437        breaking an infinite loop) was broken when the group contained a repeated
438        negated single-character class with a character that occupied more than one
439        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
440        UTF-8 mode). The effect was undefined: the group might or might not be
441        deemed as matching an empty string, or the program might have crashed.
442    
443    7.  The code for checking whether a group could match an empty string was not
444        recognizing that \h, \H, \v, \V, and \R must match a character.
445    
446    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
447        an empty string. If it can, pcretest shows this in its information output.
448    
449    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
450        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
451        matched by pcre_exec() without using JIT:
452    
453        (a) If the rest of the pattern did not match after a maximal run of
454            grapheme clusters, the code for backing up to try with fewer of them
455            did not always back up over a full grapheme when characters that do not
456            have the modifier quality were involved, e.g. Hangul syllables.
457    
458        (b) If the match point in a subject started with modifier character, and
459            there was no match, the code could incorrectly back up beyond the match
460            point, and potentially beyond the first character in the subject,
461            leading to a segfault or an incorrect match result.
462    
463    10. A conditional group with an assertion condition could lead to PCRE
464        recording an incorrect first data item for a match if no other first data
465        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
466        first data item, and therefore matched "ca" after "c" instead of at the
467        start.
468    
469    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
470        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
471    
472    12. The source of pcregrep now includes z/OS-specific code so that it can be
473        compiled for z/OS as part of the special z/OS distribution.
474    
475    13. Added the -T and -TM options to pcretest.
476    
477    14. The code in pcre_compile.c for creating the table of named capturing groups
478        has been refactored. Instead of creating the table dynamically during the
479        actual compiling pass, the information is remembered during the pre-compile
480        pass (on the stack unless there are more than 20 named groups, in which
481        case malloc() is used) and the whole table is created before the actual
482        compile happens. This has simplified the code (it is now nearly 150 lines
483        shorter) and prepared the way for better handling of references to groups
484        with duplicate names.
485    
486    15. A back reference to a named subpattern when there is more than one of the
487        same name now checks them in the order in which they appear in the pattern.
488        The first one that is set is used for the reference. Previously only the
489        first one was inspected. This change makes PCRE more compatible with Perl.
490    
491    16. Unicode character properties were updated from Unicode 6.3.0.
492    
493    17. The compile-time code for auto-possessification has been refactored, based
494        on a patch by Zoltan Herczeg. It now happens after instead of during
495        compilation. The code is cleaner, and more cases are handled. The option
496        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
497        options in pcretest are provided to set it. It can also be set by
498        (*NO_AUTO_POSSESS) at the start of a pattern.
499    
500    18. The character VT has been added to the default ("C" locale) set of
501        characters that match \s and are generally treated as white space,
502        following this same change in Perl 5.18. There is now no difference between
503        "Perl space" and "POSIX space". Whether VT is treated as white space in
504        other locales depends on the locale.
505    
506    19. The code for checking named groups as conditions, either for being set or
507        for being recursed, has been refactored (this is related to 14 and 15
508        above). Processing unduplicated named groups should now be as fast at
509        numerical groups, and processing duplicated groups should be faster than
510        before.
511    
512    20. Two patches to the CMake build system, by Alexander Barkov:
513    
514          (1) Replace the "source" command by "." in CMakeLists.txt because
515              "source" is a bash-ism.
516    
517          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
518              without these the CMake build does not work on Solaris.
519    
520    21. Perl has changed its handling of \8 and \9. If there is no previously
521        encountered capturing group of those numbers, they are treated as the
522        literal characters 8 and 9 instead of a binary zero followed by the
523        literals. PCRE now does the same.
524    
525    22. Following Perl, added \o{} to specify codepoints in octal, making it
526        possible to specify values greater than 0777 and also making them
527        unambiguous.
528    
529    23. Perl now gives an error for missing closing braces after \x{... instead of
530        treating the string as literal. PCRE now does the same.
531    
532    24. RunTest used to grumble if an inappropriate test was selected explicitly,
533        but just skip it when running all tests. This make it awkward to run ranges
534        of tests when one of them was inappropriate. Now it just skips any
535        inappropriate tests, as it always did when running all tests.
536    
537    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
538        character types such as \d or \w, too many callouts were inserted, and the
539        data that they returned was rubbish.
540    
541    26. In UCP mode, \s was not matching two of the characters that Perl matches,
542        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
543        were matched by \h. The code has now been refactored so that the lists of
544        the horizontal and vertical whitespace characters used for \h and \v (which
545        are defined only in one place) are now also used for \s.
546    
547    27. Add JIT support for the 64 bit TileGX architecture.
548        Patch by Jiong Wang (Tilera Corporation).
549    
550    28. Possessive quantifiers for classes (both explicit and automatically
551        generated) now use special opcodes instead of wrapping in ONCE brackets.
552    
553    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
554        (because it's meaningless), this was not happening when PCRE_CASELESS was
555        set. Not wrong, but inefficient.
556    
557    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
558        properties for \w, \d, etc) is present in a test regex. Otherwise if the
559        test contains no characters greater than 255, Perl doesn't realise it
560        should be using Unicode semantics.
561    
562    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
563        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
564        does in Unicode mode.
565    
566    32. Added the "forbid" facility to pcretest so that putting tests into the
567        wrong test files can sometimes be quickly detected.
568    
569    33. There is now a limit (default 250) on the depth of nesting of parentheses.
570        This limit is imposed to control the amount of system stack used at compile
571        time. It can be changed at build time by --with-parens-nest-limit=xxx or
572        the equivalent in CMake.
573    
574    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
575        errors. Perl warns for these when in warning mode, but PCRE has no facility
576        for giving warnings.
577    
578    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
579        However, this was not working for (?!) because it is optimized to (*FAIL),
580        for which PCRE does not allow quantifiers. The optimization is now disabled
581        when a quantifier follows (?!). I can't see any use for this, but it makes
582        things uniform.
583    
584    36. Perl no longer allows group names to start with digits, so I have made this
585        change also in PCRE. It simplifies the code a bit.
586    
587    37. In extended mode, Perl ignores spaces before a + that indicates a
588        possessive quantifier. PCRE allowed a space before the quantifier, but not
589        before the possessive +. It now does.
590    
591    38. The use of \K (reset reported match start) within a repeated possessive
592        group such as (a\Kb)*+ was not working.
593    
594    40. Document that the same character tables must be used at compile time and
595        run time, and that the facility to pass tables to pcre_exec() and
596        pcre_dfa_exec() is for use only with saved/restored patterns.
597    
598    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
599        features for Builds with MSVC:
600    
601        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
602           builds). The .rc files can be used to set FileDescription and many other
603           attributes.
604    
605        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
606           This allows higher-level build scripts which want .pdb files to avoid
607           hard-coding the exact files needed."
608    
609    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
610        mean "start of word" and "end of word", respectively, as a transition aid.
611    
612    43. A minimizing repeat of a class containing codepoints greater than 255 in
613        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
614        compiled to use the heap for recursion.
615    
616    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
617        is configured.
618    
619    
620    Version 8.33 28-May-2013
621    ------------------------
622    
623    1.  Added 'U' to some constants that are compared to unsigned integers, to
624        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
625        variables that are added to signed variables, to ensure the result is
626        signed and can be negated.
627    
628    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
629        CMake config files.
630    
631    3.  Revise the creation of config.h.generic so that all boolean macros are
632        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
633        overriding via -D on the command line possible.
634    
635    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
636        to unsigned int is reported to make a quite noticeable speed difference in
637        a specific Windows environment. Testing on Linux did also appear to show
638        some benefit (and it is clearly not harmful). Also fixed the definition of
639        Xop which should be unsigned.
640    
641    5.  Related to (4), changing the definition of the intermediate variable cc
642        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
643        improvements.
644    
645    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
646        unnecessary spaces.
647    
648    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
649        and later.
650    
651    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
652    
653    9.  Optimizing fast_forward_start_bits in JIT.
654    
655    10. Adding support for callouts in JIT, and fixing some issues revealed
656        during this work. Namely:
657    
658        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
659    
660        (b) Minimum length was not checked before the matching is started.
661    
662    11. The value of capture_last that is passed to callouts was incorrect in some
663        cases when there was a capture on one path that was subsequently abandoned
664        after a backtrack. Also, the capture_last value is now reset after a
665        recursion, since all captures are also reset in this case.
666    
667    12. The interpreter no longer returns the "too many substrings" error in the
668        case when an overflowing capture is in a branch that is subsequently
669        abandoned after a backtrack.
670    
671    13. In the pathological case when an offset vector of size 2 is used, pcretest
672        now prints out the matched string after a yield of 0 or 1.
673    
674    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
675        Only supported by the JIT compiler at the moment.
676    
677    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
678    
679    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
680        offset of the starting point of the matching process, provided the offsets
681        vector is large enough.
682    
683    17. The \A escape now records a lookbehind value of 1, though its execution
684        does not actually inspect the previous character. This is to ensure that,
685        in partial multi-segment matching, at least one character from the old
686        segment is retained when a new segment is processed. Otherwise, if there
687        are no lookbehinds in the pattern, \A might match incorrectly at the start
688        of a new segment.
689    
690    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
691    
692    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
693        gives some modest performance improvement in 8-bit mode.
694    
695    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
696        be expressed in certain programming languages using Universal Character
697        Names.
698    
699    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
700        which points out that "non characters" are not "characters that may not
701        appear in Unicode strings" but rather "characters that are reserved for
702        internal use and have only local meaning".
703    
704    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
705        there was a conditional group that depended on an assertion, if the
706        assertion was false, the callout that immediately followed the alternation
707        in the condition was skipped when pcre_exec() was used for matching.
708    
709    23. Allow an explicit callout to be inserted before an assertion that is the
710        condition for a conditional group, for compatibility with automatic
711        callouts, which always insert a callout at this point.
712    
713    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
714        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
715    
716    25. (*PRUNE) is now supported by the JIT compiler.
717    
718    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
719    
720    27. Fix the case where there are two or more SKIPs with arguments that may be
721        ignored.
722    
723    28. (*SKIP) is now supported by the JIT compiler.
724    
725    29. (*THEN) is now supported by the JIT compiler.
726    
727    30. Update RunTest with additional test selector options.
728    
729    31. The way PCRE handles backtracking verbs has been changed in two ways.
730    
731        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
732        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
733        backtracking. In some cases this makes it more Perl-compatible, but Perl's
734        rather obscure rules do not always do the same thing.
735    
736        (2) Previously, backtracking verbs were confined within assertions. This is
737        no longer the case for positive assertions, except for (*ACCEPT). Again,
738        this sometimes improves Perl compatibility, and sometimes does not.
739    
740    32. A number of tests that were in test 2 because Perl did things differently
741        have been moved to test 1, because either Perl or PCRE has changed, and
742        these tests are now compatible.
743    
744    32. Backtracking control verbs are now handled in the same way in JIT and
745        interpreter.
746    
747    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
748        contained a forward subroutine reference caused a compile error.
749    
750    34. Auto-detect and optimize limited repetitions in JIT.
751    
752    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
753        blocking (*UTF) etc.
754    
755    36. In the interpreter, maximizing pattern repetitions for characters and
756        character types now use tail recursion, which reduces stack usage.
757    
758    37. The value of the max lookbehind was not correctly preserved if a compiled
759        and saved regex was reloaded on a host of different endianness.
760    
761    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
762        of the compiled pattern block, expand the flags field from 16 to 32 bits
763        because it was almost full.
764    
765    39. Try madvise first before posix_madvise.
766    
767    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
768        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
769        original change locked out all matches of empty strings. This has been
770        changed so that one match of an empty string per line is recognized.
771        Subsequent searches on the same line (for colouring or for --only-matching,
772        for example) do not recognize empty strings.
773    
774    41. Applied a user patch to fix a number of spelling mistakes in comments.
775    
776    42. Data lines longer than 65536 caused pcretest to crash.
777    
778    43. Clarified the data type for length and startoffset arguments for pcre_exec
779        and pcre_dfa_exec in the function-specific man pages, where they were
780        explicitly stated to be in bytes, never having been updated. I also added
781        some clarification to the pcreapi man page.
782    
783    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
784        a segmentation fault.
785    
786    
787    Version 8.32 30-November-2012
788    -----------------------------
789    
790    1.  Improved JIT compiler optimizations for first character search and single
791        character iterators.
792    
793    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
794        Patch by Daniel Richard G.
795    
796    3.  Single character iterator optimizations in the JIT compiler.
797    
798    4.  Improved JIT compiler optimizations for character ranges.
799    
800    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
801        Reported by Giuseppe D'Angelo.
802    
803    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
804        of a line, was being set incorrectly in cases where .* appeared inside
805        atomic brackets at the start of a pattern, or where there was a subsequent
806        *PRUNE or *SKIP.
807    
808    7.  Improved instruction cache flush for POWER/PowerPC.
809        Patch by Daniel Richard G.
810    
811    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
812        grep:
813    
814        (a) There is now no limit to the number of patterns to be matched.
815    
816        (b) An error is given if a pattern is too long.
817    
818        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
819            are now supported.
820    
821        (d) --exclude-from and --include-from (multiple use) have been added.
822    
823        (e) Exclusions and inclusions now apply to all files and directories, not
824            just to those obtained from scanning a directory recursively.
825    
826        (f) Multiple uses of -f and --file-list are now supported.
827    
828        (g) In a Windows environment, the default for -d has been changed from
829            "read" (the GNU grep default) to "skip", because otherwise the presence
830            of a directory in the file list provokes an error.
831    
832        (h) The documentation has been revised and clarified in places.
833    
834    9.  Improve the matching speed of capturing brackets.
835    
836    10. Changed the meaning of \X so that it now matches a Unicode extended
837        grapheme cluster.
838    
839    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
840        out POSIX threads when JIT support is configured.
841    
842    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
843    
844    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
845        structure could go wrong in environments where size_t is not the same size
846        as int.
847    
848    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
849        set.
850    
851    15. The EBCDIC support had decayed; later updates to the code had included
852        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
853        general tidy up of EBCDIC-related issues, and the documentation was also
854        not quite right. There is now a test that can be run on ASCII systems to
855        check some of the EBCDIC-related things (but is it not a full test).
856    
857    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
858        in a small tidy to the code.
859    
860    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
861    
862    18. If the --only-matching (-o) option in pcregrep is specified multiple
863        times, each one causes appropriate output. For example, -o1 -o2 outputs the
864        substrings matched by the 1st and 2nd capturing parentheses. A separating
865        string can be specified by --om-separator (default empty).
866    
867    19. Improving the first n character searches.
868    
869    20. Turn case lists for horizontal and vertical white space into macros so that
870        they are defined only once.
871    
872    21. This set of changes together give more compatible Unicode case-folding
873        behaviour for characters that have more than one other case when UCP
874        support is available.
875    
876        (a) The Unicode property table now has offsets into a new table of sets of
877            three or more characters that are case-equivalent. The MultiStage2.py
878            script that generates these tables (the pcre_ucd.c file) now scans
879            CaseFolding.txt instead of UnicodeData.txt for character case
880            information.
881    
882        (b) The code for adding characters or ranges of characters to a character
883            class has been abstracted into a generalized function that also handles
884            case-independence. In UTF-mode with UCP support, this uses the new data
885            to handle characters with more than one other case.
886    
887        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
888            whose other case is greater than 256 are now correctly matched
889            caselessly. Previously, the high codepoint matched the low one, but not
890            vice versa.
891    
892        (d) The processing of \h, \H, \v, and \ in character classes now makes use
893            of the new class addition function, using character lists defined as
894            macros alongside the case definitions of 20 above.
895    
896        (e) Caseless back references now work with characters that have more than
897            one other case.
898    
899        (f) General caseless matching of characters with more than one other case
900            is supported.
901    
902    22. Unicode character properties were updated from Unicode 6.2.0
903    
904    23. Improved CMake support under Windows. Patch by Daniel Richard G.
905    
906    24. Add support for 32-bit character strings, and UTF-32
907    
908    25. Major JIT compiler update (code refactoring and bugfixing).
909        Experimental Sparc 32 support is added.
910    
911    26. Applied a modified version of Daniel Richard G's patch to create
912        pcre.h.generic and config.h.generic by "make" instead of in the
913        PrepareRelease script.
914    
915    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
916        pcre_compile.c when checking for a zero character.
917    
918    28. Introducing a native interface for JIT. Through this interface, the compiled
919        machine code can be directly executed. The purpose of this interface is to
920        provide fast pattern matching, so several sanity checks are not performed.
921        However, feature tests are still performed. The new interface provides
922        1.4x speedup compared to the old one.
923    
924    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
925        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
926        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
927    
928    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
929        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
930        the "old" RFC 2279). Instead, it ended up passing a negative length to
931        pcre_exec().
932    
933    31. Add support for GCC's visibility feature to hide internal functions.
934    
935    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
936        "unknown -C option" after outputting 0 or 1.
937    
938    33. There is now support for generating a code coverage report for the test
939        suite in environments where gcc is the compiler and lcov is installed. This
940        is mainly for the benefit of the developers.
941    
942    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
943        unaddressable using valgrind annotations, allowing valgrind to detect
944        invalid memory accesses. This is mainly for the benefit of the developers.
945    
946    25. (*UTF) can now be used to start a pattern in any of the three libraries.
947    
948    26. Give configure error if --enable-cpp but no C++ compiler found.
949    
950    
951    Version 8.31 06-July-2012
952    -------------------------
953    
954    1.  Fixing a wrong JIT test case and some compiler warnings.
955    
956    2.  Removed a bashism from the RunTest script.
957    
958    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
959        to unsigned type, result still unsigned" that was given by an MS compiler
960        on encountering the code "-sizeof(xxx)".
961    
962    4.  Partial matching support is added to the JIT compiler.
963    
964    5.  Fixed several bugs concerned with partial matching of items that consist
965        of more than one character:
966    
967        (a) /^(..)\1/ did not partially match "aba" because checking references was
968            done on an "all or nothing" basis. This also applied to repeated
969            references.
970    
971        (b) \R did not give a hard partial match if \r was found at the end of the
972            subject.
973    
974        (c) \X did not give a hard partial match after matching one or more
975            characters at the end of the subject.
976    
977        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
978            a partial match for the string "\r".
979    
980        (e) When newline was set to CRLF, the metacharacter "." did not recognize
981            a partial match for a CR character at the end of the subject string.
982    
983    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
984        running pcretest, the text "(JIT)" added to the output whenever JIT is
985        actually used to run the match.
986    
987    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
988        or /S+[+] with a digit between 1 and 7.
989    
990    8.  OP_NOT now supports any UTF character not just single-byte ones.
991    
992    9.  (*MARK) control verb is now supported by the JIT compiler.
993    
994    10. The command "./RunTest list" lists the available tests without actually
995        running any of them. (Because I keep forgetting what they all are.)
996    
997    11. Add PCRE_INFO_MAXLOOKBEHIND.
998    
999    12. Applied a (slightly modified) user-supplied patch that improves performance
1000        when the heap is used for recursion (compiled with --disable-stack-for-
1001        recursion). Instead of malloc and free for each heap frame each time a
1002        logical recursion happens, frames are retained on a chain and re-used where
1003        possible. This sometimes gives as much as 30% improvement.
1004    
1005    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1006        call.
1007    
1008    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1009    
1010    15. It is now possible to link pcretest with libedit as an alternative to
1011        libreadline.
1012    
1013    16. (*COMMIT) control verb is now supported by the JIT compiler.
1014    
1015    17. The Unicode data tables have been updated to Unicode 6.1.0.
1016    
1017    18. Added --file-list option to pcregrep.
1018    
1019    19. Added binary file support to pcregrep, including the -a, --binary-files,
1020        -I, and --text options.
1021    
1022    20. The madvise function is renamed for posix_madvise for QNX compatibility
1023        reasons. Fixed by Giuseppe D'Angelo.
1024    
1025    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1026    
1027    22. Changed the option for creating symbolic links for 16-bit man pages from
1028        -s to -sf so that re-installing does not cause issues.
1029    
1030    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1031    
1032    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1033        matches in certain environments (the workspace was not being correctly
1034        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1035        some of the workspace data at the beginning of a restart.
1036    
1037    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1038        was not doing so when it should - probably a typo introduced by SVN 528
1039        (change 8.10/14).
1040    
1041    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1042        \w+ when the character tables indicated that \x{c4} was a word character.
1043        There were several related cases, all because the tests for doing a table
1044        lookup were testing for characters less than 127 instead of 255.
1045    
1046    27. If a pattern contains capturing parentheses that are not used in a match,
1047        their slots in the ovector are set to -1. For those that are higher than
1048        any matched groups, this happens at the end of processing. In the case when
1049        there were back references that the ovector was too small to contain
1050        (causing temporary malloc'd memory to be used during matching), and the
1051        highest capturing number was not used, memory off the end of the ovector
1052        was incorrectly being set to -1. (It was using the size of the temporary
1053        memory instead of the true size.)
1054    
1055    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1056        ovector size, it uses memory at the end of the block that it has got.
1057    
1058    29. Check for an overlong MARK name and give an error at compile time. The
1059        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1060    
1061    30. JIT compiler update.
1062    
1063    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1064        Rill for the patch.
1065    
1066    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1067    
1068    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1069    
1070    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1071        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1072        was enabled.
1073    
1074    35. Improve JIT code generation for greedy plus quantifier.
1075    
1076    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1077        1 to "aa" instead of to an empty string. The bug affected repeated groups
1078        that could potentially match an empty string.
1079    
1080    37. Optimizing single character iterators in JIT.
1081    
1082    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1083        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1084        codepoints that are too big for the mode are faulted, and in a UTF mode,
1085        disallowed codepoints are also faulted.
1086    
1087    39. If PCRE was compiled with UTF support, in three places in the DFA
1088        matcher there was code that should only have been obeyed in UTF mode, but
1089        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1090        processing when bytes with values greater than 127 were present. In 16-bit
1091        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1092        both cases the values are those that cannot be the first data item in a UTF
1093        character. The three items that might have provoked this were recursions,
1094        possessively repeated groups, and atomic groups.
1095    
1096    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1097        and pcregrep, because some OS require shared objects to be explicitly
1098        passed to ld, causing the link step to fail if they are not.
1099    
1100    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1101        mode, patterns that started with \h* or \R* might be incorrectly matched.
1102    
1103    
1104    Version 8.30 04-February-2012
1105    -----------------------------
1106    
1107    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1108        name is defined in ctype.h.
1109    
1110    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1111        only in quite long subpatterns.
1112    
1113    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1114        since it was replaced by pcre_fullinfo() in February 2000.
1115    
1116    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1117        match a (*MARK), and the match failed at the start of the subject, a
1118        reference to memory before the start of the subject could occur. This bug
1119        was introduced by fix 17 of release 8.21.
1120    
1121    5.  A reference to an unset group with zero minimum repetition was giving
1122        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1123        /(another)?(\1?)test/ matched against "hello world test". This bug was
1124        introduced in release 8.13.
1125    
1126    6.  Add support for 16-bit character strings (a large amount of work involving
1127        many changes and refactorings).
1128    
1129    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1130        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1131        from a file.
1132    
1133    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1134        rounding is not applied in this particular case).
1135    
1136    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1137        if they appear, or are escaped, in patterns.
1138    
1139    10. Get rid of a number of -Wunused-but-set-variable warnings.
1140    
1141    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1142        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1143        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1144        also returns the mark "x". This bug applied to capturing parentheses,
1145        non-capturing parentheses, and atomic parentheses. It also applied to some
1146        assertions.
1147    
1148    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1149        information out of configure.ac instead of relying on pcre.h.generic, which
1150        is not stored in the repository.
1151    
1152    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1153        -lreadline.
1154    
1155    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1156    
1157    15. Applied Graycode's patch to put the top-level frame on the stack rather
1158        than the heap when not using the stack for recursion. This gives a
1159        performance improvement in many cases when recursion is not deep.
1160    
1161    16. Experimental code added to "pcretest -C" to output the stack frame size.
1162    
1163    
1164    Version 8.21 12-Dec-2011
1165    ------------------------
1166    
1167    1.  Updating the JIT compiler.
1168    
1169    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1170        are added as well.
1171    
1172    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1173        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1174        calling _pcre_jit_exec. Some extra comments are added.
1175    
1176    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1177        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1178        was introduced by change 18 for 8.20.
1179    
1180    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1181        ECMA-262 standard.
1182    
1183    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1184        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1185        This bug was probably introduced by change 9 of 8.13.
1186    
1187    7.  While fixing 6 above, I noticed that a number of other items were being
1188        incorrectly rejected as "not fixed length". This arose partly because newer
1189        opcodes had not been added to the fixed-length checking code. I have (a)
1190        corrected the bug and added tests for these items, and (b) arranged for an
1191        error to occur if an unknown opcode is encountered while checking for fixed
1192        length instead of just assuming "not fixed length". The items that were
1193        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1194        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1195        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1196    
1197    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1198        being incorrectly compiled and would have given unpredicatble results.
1199    
1200    9.  A possessively repeated subpattern with minimum repeat count greater than
1201        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1202        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1203        the first (A) could occur when it should not.
1204    
1205    10. Add a cast and remove a redundant test from the code.
1206    
1207    11. JIT should use pcre_malloc/pcre_free for allocation.
1208    
1209    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1210        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1211        is anything other than /usr, -L is still shown).
1212    
1213    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1214    
1215    14. Perl does not support \N without a following name in a [] class; PCRE now
1216        also gives an error.
1217    
1218    15. If a forward reference was repeated with an upper limit of around 2000,
1219        it caused the error "internal error: overran compiling workspace". The
1220        maximum number of forward references (including repeats) was limited by the
1221        internal workspace, and dependent on the LINK_SIZE. The code has been
1222        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1223        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1224        of around 200,000 forward references. While doing this, I also speeded up
1225        the filling in of repeated forward references.
1226    
1227    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1228        incorrectly expecting the subject to contain another "a" after the start.
1229    
1230    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1231        in the match, the SKIP should be ignored. This was not happening; instead
1232        the SKIP was being treated as NOMATCH. For patterns such as
1233        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1234        tested.
1235    
1236    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1237        now much more compatible with Perl, in particular in cases where the result
1238        is a non-match for a non-anchored pattern. For example, if
1239        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1240        "m", where previously it did not return a name. A side effect of this
1241        change is that for partial matches, the last encountered mark name is
1242        returned, as for non matches. A number of tests that were previously not
1243        Perl-compatible have been moved into the Perl-compatible test files. The
1244        refactoring has had the pleasing side effect of removing one argument from
1245        the match() function, thus reducing its stack requirements.
1246    
1247    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1248        subsequent uses of /S (without +) incorrectly behaved like /S+.
1249    
1250    21. Retrieve executable code size support for the JIT compiler and fixing
1251        some warnings.
1252    
1253    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1254        not work when the shorter character appeared right at the end of the
1255        subject string.
1256    
1257    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1258        systems.
1259    
1260    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1261        output it when the /M option is used in pcretest.
1262    
1263    25. The CheckMan script was not being included in the distribution. Also, added
1264        an explicit "perl" to run Perl scripts from the PrepareRelease script
1265        because this is reportedly needed in Windows.
1266    
1267    26. If study data was being save in a file and studying had not found a set of
1268        "starts with" bytes for the pattern, the data written to the file (though
1269        never used) was taken from uninitialized memory and so caused valgrind to
1270        complain.
1271    
1272    27. Updated RunTest.bat as provided by Sheri Pierce.
1273    
1274    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1275    
1276    29. Computation of memory usage for the table of capturing group names was
1277        giving an unnecessarily large value.
1278    
1279    
1280    Version 8.20 21-Oct-2011
1281    ------------------------
1282    
1283    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1284        a POSIX class. After further experiments with Perl, which convinced me that
1285        Perl has bugs and confusions, a closing square bracket is no longer allowed
1286        in a POSIX name. This bug also affected patterns with classes that started
1287        with full stops.
1288    
1289    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1290        captured substring, but while checking the failing first alternative,
1291        substring 1 is temporarily captured. If the output vector supplied to
1292        pcre_exec() was not big enough for this capture, the yield of the function
1293        was still zero ("insufficient space for captured substrings"). This cannot
1294        be totally fixed without adding another stack variable, which seems a lot
1295        of expense for a edge case. However, I have improved the situation in cases
1296        such as /(a)(b)x|abc/ matched against "abc", where the return code
1297        indicates that fewer than the maximum number of slots in the ovector have
1298        been set.
1299    
1300    3.  Related to (2) above: when there are more back references in a pattern than
1301        slots in the output vector, pcre_exec() uses temporary memory during
1302        matching, and copies in the captures as far as possible afterwards. It was
1303        using the entire output vector, but this conflicts with the specification
1304        that only 2/3 is used for passing back captured substrings. Now it uses
1305        only the first 2/3, for compatibility. This is, of course, another edge
1306        case.
1307    
1308    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1309        main code base, and can be used by building with --enable-jit. When this is
1310        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1311        runtime --no-jit option is given.
1312    
1313    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1314        ovector, the return from the function was zero, implying that there were
1315        other matches that did not fit. The correct "exactly full" value is now
1316        returned.
1317    
1318    6.  If a subpattern that was called recursively or as a subroutine contained
1319        (*PRUNE) or any other control that caused it to give a non-standard return,
1320        invalid errors such as "Error -26 (nested recursion at the same subject
1321        position)" or even infinite loops could occur.
1322    
1323    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1324        computing the minimum length on reaching *ACCEPT, and so ended up with the
1325        wrong value of 1 rather than 0. Further investigation indicates that
1326        computing a minimum subject length in the presence of *ACCEPT is difficult
1327        (think back references, subroutine calls), and so I have changed the code
1328        so that no minimum is registered for a pattern that contains *ACCEPT.
1329    
1330    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1331        it was not handled as intended. [But see 16 below.]
1332    
1333    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1334        Sheri Pierce.
1335    
1336    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1337        the first byte in a match must be "a".
1338    
1339    11. Change 17 for 8.13 increased the recursion depth for patterns like
1340        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1341        pattern contains any instances of (*THEN). If it does not, the old
1342        optimizations are restored. It would be nice to do this on a per-group
1343        basis, but at the moment that is not feasible.
1344    
1345    12. In some environments, the output of pcretest -C is CRLF terminated. This
1346        broke RunTest's code that checks for the link size. A single white space
1347        character after the value is now allowed for.
1348    
1349    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1350        For "fr", it uses the Windows-specific input and output files.
1351    
1352    14. If (*THEN) appeared in a group that was called recursively or as a
1353        subroutine, it did not work as intended. [But see next item.]
1354    
1355    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1356        pattern fragments (but not containing any | characters). If A and B are
1357        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1358        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1359        D. In other words, Perl considers parentheses that do not contain any |
1360        characters to be part of a surrounding alternative, whereas PCRE was
1361        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1362        differently. PCRE now behaves in the same way as Perl, except in the case
1363        of subroutine/recursion calls such as (?1) which have in any case always
1364        been different (but PCRE had them first :-).
1365    
1366    16. Related to 15 above: Perl does not treat the | in a conditional group as
1367        creating alternatives. Such a group is treated in the same way as an
1368        ordinary group without any | characters when processing (*THEN). PCRE has
1369        been changed to match Perl's behaviour.
1370    
1371    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1372        RunGrepTest script failed.
1373    
1374    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1375        inevitable for groups that contain captures, but it can lead to a lot of
1376        stack use in large patterns. The old behaviour has been restored for atomic
1377        groups that do not contain any capturing parentheses.
1378    
1379    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1380        suppress the check for a minimum subject length at run time. (If it was
1381        given to pcre_exec() or pcre_dfa_exec() it did work.)
1382    
1383    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1384        fail to work when decoding hex characters in data strings in EBCDIC
1385        environments.
1386    
1387    21. It appears that in at least one Mac OS environment, the isxdigit() function
1388        is implemented as a macro that evaluates to its argument more than once,
1389        contravening the C 90 Standard (I haven't checked a later standard). There
1390        was an instance in pcretest which caused it to go wrong when processing
1391        \x{...} escapes in subject strings. The has been rewritten to avoid using
1392        things like p++ in the argument of isxdigit().
1393    
1394    
1395    Version 8.13 16-Aug-2011
1396    ------------------------
1397    
1398    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1399    
1400    2.  Two minor typos in pcre_internal.h have been fixed.
1401    
1402    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1403        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1404        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1405    
1406    4.  There were a number of related bugs in the code for matching backrefences
1407        caselessly in UTF-8 mode when codes for the characters concerned were
1408        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1409        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1410        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1411        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1412        2-byte code at the end of the subject (it thought there wasn't enough data
1413        left).
1414    
1415    5.  Comprehensive information about what went wrong is now returned by
1416        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1417        as the output vector has at least 2 elements. The offset of the start of
1418        the failing character and a reason code are placed in the vector.
1419    
1420    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1421        now returned is for the first byte of the failing character, instead of the
1422        last byte inspected. This is an incompatible change, but I hope it is small
1423        enough not to be a problem. It makes the returned offset consistent with
1424        pcre_exec() and pcre_dfa_exec().
1425    
1426    7.  pcretest now gives a text phrase as well as the error number when
1427        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1428        failure, the offset and reason code are output.
1429    
1430    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1431        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1432        back over a single character (\n). This seems wrong (because it treated the
1433        two characters as a single entity when going forwards), conflicts with the
1434        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1435        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1436        has been changed.
1437    
1438    9.  Some internal refactoring has changed the processing so that the handling
1439        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1440        time (the PCRE_DOTALL option was changed this way some time ago: version
1441        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1442        which was always a bit of a fudge. It also means that there is one less
1443        argument for the match() function, which reduces its stack requirements
1444        slightly. This change also fixes an incompatibility with Perl: the pattern
1445        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1446    
1447    10. More internal refactoring has drastically reduced the number of recursive
1448        calls to match() for possessively repeated groups such as (abc)++ when
1449        using pcre_exec().
1450    
1451    11. While implementing 10, a number of bugs in the handling of groups were
1452        discovered and fixed:
1453    
1454        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1455        (a|)*(?1) gave a compile-time internal error.
1456        ((a|)+)+  did not notice that the outer group could match an empty string.
1457        (^a|^)+   was not marked as anchored.
1458        (.*a|.*)+ was not marked as matching at start or after a newline.
1459    
1460    12. Yet more internal refactoring has removed another argument from the match()
1461        function. Special calls to this function are now indicated by setting a
1462        value in a variable in the "match data" data block.
1463    
1464    13. Be more explicit in pcre_study() instead of relying on "default" for
1465        opcodes that mean there is no starting character; this means that when new
1466        ones are added and accidentally left out of pcre_study(), testing should
1467        pick them up.
1468    
1469    14. The -s option of pcretest has been documented for ages as being an old
1470        synonym of -m (show memory usage). I have changed it to mean "force study
1471        for every regex", that is, assume /S for every regex. This is similar to -i
1472        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1473        using it. It makes it easier to run collections of tests with and without
1474        study enabled, and thereby test pcre_study() more easily. All the standard
1475        tests are now run with and without -s (but some patterns can be marked as
1476        "never study" - see 20 below).
1477    
1478    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1479        restoration of the capturing data to the outer values was not happening
1480        correctly.
1481    
1482    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1483        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1484        pattern had matched an empty string, and so incorrectly returned a no
1485        match.
1486    
1487    17. There was optimizing code for the last branch of non-capturing parentheses,
1488        and also for the obeyed branch of a conditional subexpression, which used
1489        tail recursion to cut down on stack usage. Unfortunately, now that there is
1490        the possibility of (*THEN) occurring in these branches, tail recursion is
1491        no longer possible because the return has to be checked for (*THEN). These
1492        two optimizations have therefore been removed. [But see 8.20/11 above.]
1493    
1494    18. If a pattern containing \R was studied, it was assumed that \R always
1495        matched two bytes, thus causing the minimum subject length to be
1496        incorrectly computed because \R can also match just one byte.
1497    
1498    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1499        was incorrectly computed.
1500    
1501    20. If /S is present twice on a test pattern in pcretest input, it now
1502        *disables* studying, thereby overriding the use of -s on the command line
1503        (see 14 above). This is necessary for one or two tests to keep the output
1504        identical in both cases.
1505    
1506    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1507        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1508    
1509    22. When an atomic group that contained a capturing parenthesis was
1510        successfully matched, but the branch in which it appeared failed, the
1511        capturing was not being forgotten if a higher numbered group was later
1512        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1513        group 1 to "a", when in fact it should be unset. This applied to multi-
1514        branched capturing and non-capturing groups, repeated or not, and also to
1515        positive assertions (capturing in negative assertions does not happen
1516        in PCRE) and also to nested atomic groups.
1517    
1518    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1519        subject after a captured substring, to make it easier to tell which of a
1520        number of identical substrings has been captured.
1521    
1522    24. The way atomic groups are processed by pcre_exec() has been changed so that
1523        if they are repeated, backtracking one repetition now resets captured
1524        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1525        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1526        "aaa". Previously, it would have been "a". As part of this code
1527        refactoring, the way recursive calls are handled has also been changed.
1528    
1529    25. If an assertion condition captured any substrings, they were not passed
1530        back unless some other capturing happened later. For example, if
1531        (?(?=(a))a) was matched against "a", no capturing was returned.
1532    
1533    26. When studying a pattern that contained subroutine calls or assertions,
1534        the code for finding the minimum length of a possible match was handling
1535        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1536        group 1 called group 2 while simultaneously a separate group 2 called group
1537        1). A stack overflow occurred in this case. I have fixed this by limiting
1538        the recursion depth to 10.
1539    
1540    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1541        Fortmann. This supports explicit test numbers on the command line, and has
1542        argument validation and error reporting.
1543    
1544    28. An instance of \X with an unlimited repeat could fail if at any point the
1545        first character it looked at was a mark character.
1546    
1547    29. Some minor code refactoring concerning Unicode properties and scripts
1548        should reduce the stack requirement of match() slightly.
1549    
1550    30. Added the '=' option to pcretest to check the setting of unused capturing
1551        slots at the end of the pattern, which are documented as being -1, but are
1552        not included in the return count.
1553    
1554    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1555        compiled something random. Now it gives a compile-time error (as does
1556        Perl).
1557    
1558    32. A *MARK encountered during the processing of a positive assertion is now
1559        recorded and passed back (compatible with Perl).
1560    
1561    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1562        had alternative anchored branches, the search for a second match in a line
1563        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1564        matched the line "0102" twice. The same bug affected patterns that started
1565        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1566        twice.
1567    
1568    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1569        does, and because of capturing effects, quantifying parenthesized
1570        assertions may at times be useful. Quantifiers are now allowed for
1571        parenthesized assertions.
1572    
1573    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1574    
1575    36. \g was being checked for fancy things in a character class, when it should
1576        just be a literal "g".
1577    
1578    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1579        appearance of a nested POSIX class supersedes an apparent external class.
1580        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1581        unescaped square brackets may also appear as part of class names. For
1582        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1583        more like Perl. (But see 8.20/1 above.)
1584    
1585    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1586        was because it thought it was \N{name}, which is not supported).
1587    
1588    39. Add minix to OS list not supporting the -S option in pcretest.
1589    
1590    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1591        cannot analyze patterns in sufficient detail to catch mutual recursions
1592        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1593        subgroup is called recursively as a subpattern for a second time at the
1594        same position in the subject string. In previous releases this might have
1595        been caught by the recursion limit, or it might have run out of stack.
1596    
1597    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1598        happen only once. PCRE was, however incorrectly giving a compile time error
1599        "recursive call could loop indefinitely" because it cannot analyze the
1600        pattern in sufficient detail. The compile time test no longer happens when
1601        PCRE is compiling a conditional subpattern, but actual runaway loops are
1602        now caught at runtime (see 40 above).
1603    
1604    42. It seems that Perl allows any characters other than a closing parenthesis
1605        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1606        has been changed to be the same.
1607    
1608    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1609        as not to get warnings when autogen.sh is called. Also changed
1610        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1611    
1612    44. To help people who use pcregrep to scan files containing exceedingly long
1613        lines, the following changes have been made:
1614    
1615        (a) The default value of the buffer size parameter has been increased from
1616            8K to 20K. (The actual buffer used is three times this size.)
1617    
1618        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1619            PCRE is built.
1620    
1621        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1622            to be set at run time.
1623    
1624        (d) Numerical values in pcregrep options can be followed by K or M, for
1625            example --buffer-size=50K.
1626    
1627        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1628            given and the return code is set to 2.
1629    
1630    45. Add a pointer to the latest mark to the callout data block.
1631    
1632    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1633        partial match of an empty string instead of no match. This was specific to
1634        the use of ".".
1635    
1636    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1637        complete match instead of a partial match. This bug was dependent on both
1638        the PCRE_UTF8 and PCRE_DOTALL options being set.
1639    
1640    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1641        starting byte set, because \b was not being ignored.
1642    
1643    
1644    Version 8.12 15-Jan-2011
1645    ------------------------
1646    
1647    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1648        checks for such things as part of the documentation building process.
1649    
1650    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1651        --match-limit and --recursion-limit options (added for 8.11). In
1652        particular, this made one of the standard tests fail. (The integer value
1653        went into the wrong half of a long int.)
1654    
1655    3.  If the --colour option was given to pcregrep with -v (invert match), it
1656        did strange things, either producing crazy output, or crashing. It should,
1657        of course, ignore a request for colour when reporting lines that do not
1658        match.
1659    
1660    4.  Another pcregrep bug caused similar problems if --colour was specified with
1661        -M (multiline) and the pattern match finished with a line ending.
1662    
1663    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1664        matched in multiline mode, the following line was shown as part of the
1665        match. This seems wrong, so I have changed it.
1666    
1667    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1668        the check for further matches in the same line (so they could be coloured)
1669        to overrun the end of the current line. If another match was found, it was
1670        incorrectly shown (and then shown again when found in the next line).
1671    
1672    7.  If pcregrep was compiled under Windows, there was a reference to the
1673        function pcregrep_exit() before it was defined. I am assuming this was
1674        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1675        reported by a user. I've moved the definition above the reference.
1676    
1677    
1678    Version 8.11 10-Dec-2010
1679    ------------------------
1680    
1681    1.  (*THEN) was not working properly if there were untried alternatives prior
1682        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1683        backtracked to try for "b" instead of moving to the next alternative branch
1684        at the same level (in this case, to look for "c"). The Perl documentation
1685        is clear that when (*THEN) is backtracked onto, it goes to the "next
1686        alternative in the innermost enclosing group".
1687    
1688    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1689        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1690        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1691        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1692        (*THEN).
1693    
1694    3.  If \s appeared in a character class, it removed the VT character from
1695        the class, even if it had been included by some previous item, for example
1696        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1697        of \s, but is part of the POSIX "space" class.)
1698    
1699    4.  A partial match never returns an empty string (because you can always
1700        match an empty string at the end of the subject); however the checking for
1701        an empty string was starting at the "start of match" point. This has been
1702        changed to the "earliest inspected character" point, because the returned
1703        data for a partial match starts at this character. This means that, for
1704        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1705        (previously it gave "no match").
1706    
1707    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1708        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1709        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1710        has an implication that the given string is incomplete (because a partial
1711        match is preferred over a full match). For this reason, these items now
1712        give a partial match in this situation. [Aside: previously, the one case
1713        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1714        match rather than a full match, which was wrong by the old rules, but is
1715        now correct.]
1716    
1717    6.  There was a bug in the handling of #-introduced comments, recognized when
1718        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1719        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1720        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1721        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1722        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1723        places in pcre_compile().
1724    
1725    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1726        comments when looking ahead for named forward references to subpatterns,
1727        the only newline sequence it recognized was NL. It now handles newlines
1728        according to the set newline convention.
1729    
1730    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1731        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1732        cater for a lack of strerror(). These oversights have been fixed.
1733    
1734    9.  Added --match-limit and --recursion-limit to pcregrep.
1735    
1736    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1737    
1738    11. When the -o option was used, pcregrep was setting a return code of 1, even
1739        when matches were found, and --line-buffered was not being honoured.
1740    
1741    12. Added an optional parentheses number to the -o and --only-matching options
1742        of pcregrep.
1743    
1744    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1745        can match an empty string. The code to do it in pcretest and pcredemo
1746        needed fixing:
1747    
1748        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1749            only one byte after an empty string match just before CRLF (this case
1750            just got forgotten; "any" and "anycrlf" were OK).
1751    
1752        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1753            mode when an empty string match preceded an ASCII character followed by
1754            a non-ASCII character. (The code for advancing by one character rather
1755            than one byte was nonsense.)
1756    
1757        (c) The pcredemo.c sample program did not have any code at all to handle
1758            the cases when CRLF is a valid newline sequence.
1759    
1760    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1761        as a starting offset was within the subject string. There is now a new
1762        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1763        negative or greater than the length of the string. In order to test this,
1764        pcretest is extended to allow the setting of negative starting offsets.
1765    
1766    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1767        starting offset points to the beginning of a UTF-8 character was
1768        unnecessarily clumsy. I tidied it up.
1769    
1770    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1771        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1772    
1773    17. Nobody had reported that the --include_dir option, which was added in
1774        release 7.7 should have been called --include-dir (hyphen, not underscore)
1775        for compatibility with GNU grep. I have changed it to --include-dir, but
1776        left --include_dir as an undocumented synonym, and the same for
1777        --exclude-dir, though that is not available in GNU grep, at least as of
1778        release 2.5.4.
1779    
1780    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1781        characters from a string of bytes) have been redefined so as not to use
1782        loops, in order to improve performance in some environments. At the same
1783        time, I abstracted some of the common code into auxiliary macros to save
1784        repetition (this should not affect the compiled code).
1785    
1786    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1787        compile-time error is now given if \c is not followed by an ASCII
1788        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1789        different, and any byte value is allowed.)
1790    
1791    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1792        START_OPTIMIZE option, which is now allowed at compile time - but just
1793        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1794        to pcregrep and other applications that have no direct access to PCRE
1795        options. The new /Y option in pcretest sets this option when calling
1796        pcre_compile().
1797    
1798    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1799        back references. Groups containing recursive back references were forced to
1800        be atomic by that change, but in the case of named groups, the amount of
1801        memory required was incorrectly computed, leading to "Failed: internal
1802        error: code overflow". This has been fixed.
1803    
1804    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1805        pcretest.c, to avoid build problems in some Borland environments.
1806    
1807    
1808    Version 8.10 25-Jun-2010
1809    ------------------------
1810    
1811    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1812        THEN.
1813    
1814    2.  (*ACCEPT) was not working when inside an atomic group.
1815    
1816    3.  Inside a character class, \B is treated as a literal by default, but
1817        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1818        causes the error). The code is unchanged, but I tidied the documentation.
1819    
1820    4.  Inside a character class, PCRE always treated \R and \X as literals,
1821        whereas Perl faults them if its -w option is set. I have changed PCRE so
1822        that it faults them when PCRE_EXTRA is set.
1823    
1824    5.  Added support for \N, which always matches any character other than
1825        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1826    
1827    6.  When compiling pcregrep with newer versions of gcc which may have
1828        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1829        declared with attribute warn_unused_result" were given. Just casting the
1830        result to (void) does not stop the warnings; a more elaborate fudge is
1831        needed. I've used a macro to implement this.
1832    
1833    7.  Minor change to pcretest.c to avoid a compiler warning.
1834    
1835    8.  Added four artifical Unicode properties to help with an option to make
1836        \s etc use properties (see next item). The new properties are: Xan
1837        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1838    
1839    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1840        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1841        this option. Modified pcretest to add /W to test this facility. Added
1842        REG_UCP to make it available via the POSIX interface.
1843    
1844    10. Added --line-buffered to pcregrep.
1845    
1846    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1847        studied, and the match started with a letter with a code point greater than
1848        127 whose first byte was different to the first byte of the other case of
1849        the letter, the other case of this starting letter was not recognized
1850        (#976).
1851    
1852    12. If a pattern that was studied started with a repeated Unicode property
1853        test, for example, \p{Nd}+, there was the theoretical possibility of
1854        setting up an incorrect bitmap of starting bytes, but fortunately it could
1855        not have actually happened in practice until change 8 above was made (it
1856        added property types that matched character-matching opcodes).
1857    
1858    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1859        possible starting bytes for non-anchored patterns.
1860    
1861    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1862        \R, and also a number of cases that involve Unicode properties, both
1863        explicit and implicit when PCRE_UCP is set.
1864    
1865    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1866        input, it could crash or give wrong results if characters with values
1867        greater than 0xc0 were present in the subject string. (Detail: it assumed
1868        UTF-8 input when processing these items.)
1869    
1870    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1871        size_t is 64-bit (#991).
1872    
1873    17. Added a check for running out of memory when PCRE is compiled with
1874        --disable-stack-for-recursion (#990).
1875    
1876    18. If the last data line in a file for pcretest does not have a newline on
1877        the end, a newline was missing in the output.
1878    
1879    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1880        less than 128) in its various bitmaps. However, there is a facility for
1881        generating tables according to the current locale when PCRE is compiled. It
1882        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1883        characters, are recognized by isspace() and therefore were getting set in
1884        these tables, and indeed these tables seem to approximate to ISO 8859. This
1885        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1886        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1887        which of course cannot start UTF-8 characters. I have changed the code so
1888        that only real ASCII characters (less than 128) and the correct starting
1889        bytes for UTF-8 encodings are set for characters greater than 127 when in
1890        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1891        altogether.)
1892    
1893    20. Added the /T option to pcretest so as to be able to run tests with non-
1894        standard character tables, thus making it possible to include the tests
1895        used for 19 above in the standard set of tests.
1896    
1897    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1898        reference to a subpattern the other side of a comment that contains an
1899        opening parenthesis caused either an internal compiling error, or a
1900        reference to the wrong subpattern.
1901    
1902    
1903    Version 8.02 19-Mar-2010
1904    ------------------------
1905    
1906    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1907    
1908    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1909        configured.
1910    
1911    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1912        original author of that file, following a query about its status.
1913    
1914    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1915        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1916    
1917    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1918        quantifier applied to a forward-referencing subroutine call, could compile
1919        incorrect code or give the error "internal error: previously-checked
1920        referenced subpattern not found".
1921    
1922    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1923        variables to point to external functions. For these systems, therefore,
1924        pcre_malloc etc. are now initialized to local functions that call the
1925        relevant global functions.
1926    
1927    7.  There were two entries missing in the vectors called coptable and poptable
1928        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1929        I've fixed the data, and added a kludgy way of testing at compile time that
1930        the lengths are correct (equal to the number of opcodes).
1931    
1932    8.  Following on from 7, I added a similar kludge to check the length of the
1933        eint vector in pcreposix.c.
1934    
1935    9.  Error texts for pcre_compile() are held as one long string to avoid too
1936        much relocation at load time. To find a text, the string is searched,
1937        counting zeros. There was no check for running off the end of the string,
1938        which could happen if a new error number was added without updating the
1939        string.
1940    
1941    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1942    
1943    11. \K was not working if it appeared in an atomic group or in a group that
1944        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1945        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1946        the assertion is positive, but not if it is negative.
1947    
1948    12. Change 11 fortuitously reduced the size of the stack frame used in the
1949        "match()" function of pcre_exec.c by one pointer. Forthcoming
1950        implementation of support for (*MARK) will need an extra pointer on the
1951        stack; I have reserved it now, so that the stack frame size does not
1952        decrease.
1953    
1954    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1955        item in branch that calls a recursion is a subroutine call - as in the
1956        second branch in the above example - was incorrectly given the compile-
1957        time error "recursive call could loop indefinitely" because pcre_compile()
1958        was not correctly checking the subroutine for matching a non-empty string.
1959    
1960    14. The checks for overrunning compiling workspace could trigger after an
1961        overrun had occurred. This is a "should never occur" error, but it can be
1962        triggered by pathological patterns such as hundreds of nested parentheses.
1963        The checks now trigger 100 bytes before the end of the workspace.
1964    
1965    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1966    
1967    
1968    Version 8.01 19-Jan-2010
1969    ------------------------
1970    
1971    1.  If a pattern contained a conditional subpattern with only one branch (in
1972        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1973        computed the wrong minimum data length (which is of course zero for such
1974        subpatterns). This could cause incorrect "no match" results.
1975    
1976    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1977        the pattern is reset in the first branch, pcre_compile() failed with
1978        "internal error: code overflow at offset...". This happened only when
1979        the reset was to the original external option setting. (An optimization
1980        abstracts leading options settings into an external setting, which was the
1981        cause of this.)
1982    
1983    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1984        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1985        assertion pattern did not match (meaning that the assertion was true), it
1986        was incorrectly treated as false if the SKIP had been reached during the
1987        matching. This also applied to assertions used as conditions.
1988    
1989    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1990        assertion subpattern, including such a pattern used as a condition,
1991        unpredictable results occurred, instead of the error return
1992        PCRE_ERROR_DFA_UITEM.
1993    
1994    5.  The C++ GlobalReplace function was not working like Perl for the special
1995        situation when an empty string is matched. It now does the fancy magic
1996        stuff that is necessary.
1997    
1998    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1999        removed. (These were left over from very, very early versions of PCRE.)
2000    
2001    7.  Some cosmetic changes to the code to make life easier when compiling it
2002        as part of something else:
2003    
2004        (a) Change DEBUG to PCRE_DEBUG.
2005    
2006        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2007            called "current" as "current_branch", to prevent a collision with the
2008            Linux macro when compiled as a kernel module.
2009    
2010        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2011            prevent a collision with the Linux macro when compiled as a kernel
2012            module.
2013    
2014    8.  In pcre_compile() there are some checks for integer overflows that used to
2015        cast potentially large values to (double). This has been changed to that
2016        when building, a check for int64_t is made, and if it is found, it is used
2017        instead, thus avoiding the use of floating point arithmetic. (There is no
2018        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2019        double.
2020    
2021    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2022        2005 (difference between two addresses compared to an unsigned value).
2023    
2024    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2025        custom one, because of the following reported problem in Windows:
2026    
2027          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2028              under Win32.
2029          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2030              therefore missing the function definition.
2031          - The compiler thus generates a "C" signature for the test function.
2032          - The linker fails to find the "C" function.
2033          - PCRE fails to configure if asked to do so against libbz2.
2034    
2035    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2036        messages were output:
2037    
2038          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2039          rerunning libtoolize, to keep the correct libtool macros in-tree.
2040          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2041    
2042        I have done both of these things.
2043    
2044    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2045        most of the time, it *can* run out if it is given a pattern that contains a
2046        runaway infinite recursion. I updated the discussion in the pcrestack man
2047        page.
2048    
2049    13. Now that we have gone to the x.xx style of version numbers, the minor
2050        version may start with zero. Using 08 or 09 is a bad idea because users
2051        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2052        interpreted as invalid octal numbers. I've updated the previous comment in
2053        configure.ac, and also added a check that gives an error if 08 or 09 are
2054        used.
2055    
2056    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2057        causing partial matching to fail when the end of the subject matched \W
2058        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2059    
2060    15. There were some discrepancies between the declarations in pcre_internal.h
2061        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2062        their definitions. The declarations used "const uschar *" and the
2063        definitions used USPTR. Even though USPTR is normally defined as "const
2064        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2065        reported that: "This difference in casting confuses some C++ compilers, for
2066        example, SunCC recognizes above declarations as different functions and
2067        generates broken code for hbpcre." I have changed the declarations to use
2068        USPTR.
2069    
2070    16. GNU libtool is named differently on some systems. The autogen.sh script now
2071        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2072        (FreeBSD).
2073    
2074    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2075        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2076        comment: "Figure out how to create a longlong from a string: strtoll and
2077        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2078        instance, but it only takes 2 args instead of 3!"
2079    
2080    18. A subtle bug concerned with back references has been fixed by a change of
2081        specification, with a corresponding code fix. A pattern such as
2082        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2083        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2084        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2085        same bug. Such groups have to be quantified to be useful, or contained
2086        inside another quantified group. (If there's no repetition, the reference
2087        can never match.) The problem arises because, having left the group and
2088        moved on to the rest of the pattern, a later failure that backtracks into
2089        the group uses the captured value from the final iteration of the group
2090        rather than the correct earlier one. I have fixed this in PCRE by forcing
2091        any group that contains a reference to itself to be an atomic group; that
2092        is, there cannot be any backtracking into it once it has completed. This is
2093        similar to recursive and subroutine calls.
2094    
2095    
2096    Version 8.00 19-Oct-09
2097    ----------------------
2098    
2099    1.  The table for translating pcre_compile() error codes into POSIX error codes
2100        was out-of-date, and there was no check on the pcre_compile() error code
2101        being within the table. This could lead to an OK return being given in
2102        error.
2103    
2104    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2105        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2106        in a Windows environment.
2107    
2108    3.  The pcregrep --count option prints the count for each file even when it is
2109        zero, as does GNU grep. However, pcregrep was also printing all files when
2110        --files-with-matches was added. Now, when both options are given, it prints
2111        counts only for those files that have at least one match. (GNU grep just
2112        prints the file name in this circumstance, but including the count seems
2113        more useful - otherwise, why use --count?) Also ensured that the
2114        combination -clh just lists non-zero counts, with no names.
2115    
2116    4.  The long form of the pcregrep -F option was incorrectly implemented as
2117        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2118        but it seems right to fix it, and I didn't think it was worth preserving
2119        the old behaviour.
2120    
2121    5.  The command line items --regex=pattern and --regexp=pattern were not
2122        recognized by pcregrep, which required --regex pattern or --regexp pattern
2123        (with a space rather than an '='). The man page documented the '=' forms,
2124        which are compatible with GNU grep; these now work.
2125    
2126    6.  No libpcreposix.pc file was created for pkg-config; there was just
2127        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2128    
2129    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2130        when UCP support is not needed, by modifying the Python script that
2131        generates it from Unicode data files. This should not matter if the module
2132        is correctly used as a library, but I received one complaint about 50K of
2133        unwanted data. My guess is that the person linked everything into his
2134        program rather than using a library. Anyway, it does no harm.
2135    
2136    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2137        was a minimum greater than 1 for a wide character in a possessive
2138        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2139        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2140        character. Chaos in the form of incorrect output or a compiling loop could
2141        result.
2142    
2143    9.  The restrictions on what a pattern can contain when partial matching is
2144        requested for pcre_exec() have been removed. All patterns can now be
2145        partially matched by this function. In addition, if there are at least two
2146        slots in the offset vector, the offset of the earliest inspected character
2147        for the match and the offset of the end of the subject are set in them when
2148        PCRE_ERROR_PARTIAL is returned.
2149    
2150    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2151        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2152        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2153        and may be more useful for multi-segment matching.
2154    
2155    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2156        used to be given if ever the end of the subject was reached; now it is
2157        given only if matching could not proceed because another character was
2158        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2159        string "Z", which now yields "no match" instead of "partial match". In the
2160        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2161        final character ended with (*FAIL).
2162    
2163    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2164        if the pattern had a "must contain" character that was already found in the
2165        earlier partial match, unless partial matching was again requested. For
2166        example, with the pattern /dog.(body)?/, the "must contain" character is
2167        "g". If the first part-match was for the string "dog", restarting with
2168        "sbody" failed. This bug has been fixed.
2169    
2170    13. The string returned by pcre_dfa_exec() after a partial match has been
2171        changed so that it starts at the first inspected character rather than the
2172        first character of the match. This makes a difference only if the pattern
2173        starts with a lookbehind assertion or \b or \B (\K is not supported by
2174        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2175        matching functions compatible, and I think it's the right thing to do.
2176    
2177    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2178        so that the demonstration program is easily available in environments where
2179        PCRE has not been installed from source.
2180    
2181    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2182        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2183        library.
2184    
2185    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2186        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2187        is not the first non-POSIX option to be added. Clearly some people find
2188        these options useful.
2189    
2190    17. If a caller to the POSIX matching function regexec() passes a non-zero
2191        value for nmatch with a NULL value for pmatch, the value of
2192        nmatch is forced to zero.
2193    
2194    18. RunGrepTest did not have a test for the availability of the -u option of
2195        the diff command, as RunTest does. It now checks in the same way as
2196        RunTest, and also checks for the -b option.
2197    
2198    19. If an odd number of negated classes containing just a single character
2199        interposed, within parentheses, between a forward reference to a named
2200        subpattern and the definition of the subpattern, compilation crashed with
2201        an internal error, complaining that it could not find the referenced
2202        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2203        [The bug was that it was starting one character too far in when skipping
2204        over the character class, thus treating the ] as data rather than
2205        terminating the class. This meant it could skip too much.]
2206    
2207    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2208        /g option in pcretest when the pattern contains \K, which makes it possible
2209        to have an empty string match not at the start, even when the pattern is
2210        anchored. Updated pcretest and pcredemo to use this option.
2211    
2212    21. If the maximum number of capturing subpatterns in a recursion was greater
2213        than the maximum at the outer level, the higher number was returned, but
2214        with unset values at the outer level. The correct (outer level) value is
2215        now given.
2216    
2217    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2218        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2219        make it do so. The string so far is captured, making this feature
2220        compatible with Perl.
2221    
2222    23. The tests have been re-organized, adding tests 11 and 12, to make it
2223        possible to check the Perl 5.10 features against Perl 5.10.
2224    
2225    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2226        pattern matches a fixed length string. PCRE did not allow this; now it
2227        does. Neither allows recursion.
2228    
2229    25. I finally figured out how to implement a request to provide the minimum
2230        length of subject string that was needed in order to match a given pattern.
2231        (It was back references and recursion that I had previously got hung up
2232        on.) This code has now been added to pcre_study(); it finds a lower bound
2233        to the length of subject needed. It is not necessarily the greatest lower
2234        bound, but using it to avoid searching strings that are too short does give
2235        some useful speed-ups. The value is available to calling programs via
2236        pcre_fullinfo().
2237    
2238    26. While implementing 25, I discovered to my embarrassment that pcretest had
2239        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2240        study optimizations had never been tested with that matching function.
2241        Oops. What is worse, even when it was passed study data, there was a bug in
2242        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2243        were also very few tests of studied patterns with pcre_dfa_exec().
2244    
2245    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2246        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2247        on the other side of the coin, they are no longer allowed to have different
2248        names, because these cannot be distinguished in PCRE, and this has caused
2249        confusion. (This is a difference from Perl.)
2250    
2251    28. When duplicate subpattern names are present (necessarily with different
2252        numbers, as required by 27 above), and a test is made by name in a
2253        conditional pattern, either for a subpattern having been matched, or for
2254        recursion in such a pattern, all the associated numbered subpatterns are
2255        tested, and the overall condition is true if the condition is true for any
2256        one of them. This is the way Perl works, and is also more like the way
2257        testing by number works.
2258    
2259    
2260    Version 7.9 11-Apr-09
2261    ---------------------
2262    
2263    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2264        (pcretest), all targets were linked against these libraries. This included
2265        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2266        libraries. This caused unwanted dependencies to be created. This problem
2267        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2268        pcretest is linked with readline.
2269    
2270    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2271        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2272        moved outside it again, because FALSE and TRUE are already defined in AIX,
2273        but BOOL is not.
2274    
2275    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2276        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2277    
2278    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2279        hyphens) following file names and line numbers when outputting matching
2280        lines. This is not true; no spaces are inserted. I have also clarified the
2281        wording for the --colour (or --color) option.
2282    
2283    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2284        was not coloured; this is different to GNU grep, so I have changed it to be
2285        the same.
2286    
2287    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2288        each matching line was coloured. Now it goes on to look for further matches
2289        of any of the test patterns, which is the same behaviour as GNU grep.
2290    
2291    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2292        doesn't make sense to accept an empty string match in pcregrep, so I have
2293        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2294        seems to be how GNU grep behaves. [But see later change 40 for release
2295        8.33.]
2296    
2297    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2298        start or after a newline", because the conditional assertion was not being
2299        correctly handled. The rule now is that both the assertion and what follows
2300        in the first alternative must satisfy the test.
2301    
2302    9.  If auto-callout was enabled in a pattern with a conditional group whose
2303        condition was an assertion, PCRE could crash during matching, both with
2304        pcre_exec() and pcre_dfa_exec().
2305    
2306    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2307        used for matching.
2308    
2309    11. Unicode property support in character classes was not working for
2310        characters (bytes) greater than 127 when not in UTF-8 mode.
2311    
2312    12. Added the -M command line option to pcretest.
2313    
2314    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2315    
2316    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2317    
2318    16. Added comments and documentation about mis-use of no_arg in the C++
2319        wrapper.
2320    
2321    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2322        from Martin Jerabek that uses macro names for all relevant character and
2323        string constants.
2324    
2325    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2326        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2327        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2328        these, but not everybody uses configure.
2329    
2330    19. A conditional group that had only one branch was not being correctly
2331        recognized as an item that could match an empty string. This meant that an
2332        enclosing group might also not be so recognized, causing infinite looping
2333        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2334        with the subject "ab", where knowledge that the repeated group can match
2335        nothing is needed in order to break the loop.
2336    
2337    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2338        exec(), but without supplying a callout function, matching went wrong.
2339    
2340    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2341        leak if the size of the offset vector was greater than 30. When the vector
2342        is smaller, the saved offsets during recursion go onto a local stack
2343        vector, but for larger vectors malloc() is used. It was failing to free
2344        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2345        error, in fact).
2346    
2347    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2348        heapframe that is used only when UTF-8 support is enabled. This caused no
2349        problem, but was untidy.
2350    
2351    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2352        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2353        included within another project.
2354    
2355    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2356        slightly modified by me:
2357    
2358          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2359              not building pcregrep.
2360    
2361          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2362              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2363    
2364    25. Forward references, both numeric and by name, in patterns that made use of
2365        duplicate group numbers, could behave incorrectly or give incorrect errors,
2366        because when scanning forward to find the reference group, PCRE was not
2367        taking into account the duplicate group numbers. A pattern such as
2368        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2369    
2370    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2371        the feature of a custom pointer more persuasive (as requested by a user).
2372    
2373    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2374        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2375        pre-defined.
2376    
2377    28. Added support for (*UTF8) at the start of a pattern.
2378    
2379    29. Arrange for flags added by the "release type" setting in CMake to be shown
2380        in the configuration summary.
2381    
2382    
2383    Version 7.8 05-Sep-08
2384    ---------------------
2385    
2386    1.  Replaced UCP searching code with optimized version as implemented for Ad
2387        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2388        stage table and inline lookup instead of a function, giving speed ups of 2
2389        to 5 times on some simple patterns that I tested. Permission was given to
2390        distribute the MultiStage2.py script that generates the tables (it's not in
2391        the tarball, but is in the Subversion repository).
2392    
2393    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2394        scripts.
2395    
2396    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2397        a group with a zero qualifier. The result of the study could be incorrect,
2398        or the function might crash, depending on the pattern.
2399    
2400    4.  Caseless matching was not working for non-ASCII characters in back
2401        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2402        It now works when Unicode Property Support is available.
2403    
2404    5.  In pcretest, an escape such as \x{de} in the data was always generating
2405        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2406        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2407        truncation.
2408    
2409    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2410    
2411    7.  Added two (int) casts to pcregrep when printing the difference of two
2412        pointers, in case they are 64-bit values.
2413    
2414    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2415        test 2 if it fails.
2416    
2417    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2418        and a #define of that name to empty if it is not externally set. This is to
2419        allow users of MSVC to set it if necessary.
2420    
2421    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2422        the convenience functions in the pcre_get.c source file.
2423    
2424    11. An option change at the start of a pattern that had top-level alternatives
2425        could cause overwriting and/or a crash. This command provoked a crash in
2426        some environments:
2427    
2428          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2429    
2430        This potential security problem was recorded as CVE-2008-2371.
2431    
2432    12. For a pattern where the match had to start at the beginning or immediately
2433        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2434        pcre_dfa_exec() could read past the end of the passed subject if there was
2435        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2436        pcretest so that it places the subject at the end of its malloc-ed buffer.
2437    
2438    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2439        exec() might read past the end of the data buffer in UTF-8 mode.
2440    
2441    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2442        the data contained the byte 0x85 as part of a UTF-8 character within its
2443        first line. This applied both to normal and DFA matching.
2444    
2445    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2446        /^[^d]*?$/8 failed to match "abc".
2447    
2448    16. Added a missing copyright notice to pcrecpp_internal.h.
2449    
2450    17. Make it more clear in the documentation that values returned from
2451        pcre_exec() in ovector are byte offsets, not character counts.
2452    
2453    18. Tidied a few places to stop certain compilers from issuing warnings.
2454    
2455    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2456        supplied by Stefan Weber. I made a further small update for 7.8 because
2457        there is a change of source arrangements: the pcre_searchfuncs.c module is
2458        replaced by pcre_ucd.c.
2459    
2460    
2461    Version 7.7 07-May-08
2462    ---------------------
2463    
2464    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2465        a string to a long long, pretend we don't even have a long long." This is
2466        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2467    
2468    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2469        pre-7.6 versions, which defined a global no_arg variable instead of putting
2470        it in the RE class. (See also #8 below.)
2471    
2472    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2473        Lopes.
2474    
2475    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2476    
2477        (1) The include/exclude patterns were being applied to the whole pathnames
2478            of files, instead of just to the final components.
2479    
2480        (2) If there was more than one level of directory, the subdirectories were
2481            skipped unless they satisfied the include/exclude conditions. This is
2482            inconsistent with GNU grep (and could even be seen as contrary to the
2483            pcregrep specification - which I improved to make it absolutely clear).
2484            The action now is always to scan all levels of directory, and just
2485            apply the include/exclude patterns to regular files.
2486    
2487    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2488        --exclude_dir in the tests to avoid scanning .svn directories.
2489    
2490    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2491        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2492        doesn't support NULs in patterns.
2493    
2494    7.  Added some missing "const"s to declarations of static tables in
2495        pcre_compile.c and pcre_dfa_exec.c.
2496    
2497    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2498        caused by fix #2  above. (Subsequently also a second patch to fix the
2499        first patch. And a third patch - this was a messy problem.)
2500    
2501    9.  Applied Craig's patch to remove the use of push_back().
2502    
2503    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2504        matching function regexec().
2505    
2506    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2507        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2508        references. PCRE supports relative numbers with this syntax (I don't think
2509        Oniguruma does).
2510    
2511    12. Previously, a group with a zero repeat such as (...){0} was completely
2512        omitted from the compiled regex. However, this means that if the group
2513        was called as a subroutine from elsewhere in the pattern, things went wrong
2514        (an internal error was given). Such groups are now left in the compiled
2515        pattern, with a new opcode that causes them to be skipped at execution
2516        time.
2517    
2518    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2519        to the way PCRE behaves:
2520    
2521        (a) A lone ] character is dis-allowed (Perl treats it as data).
2522    
2523        (b) A back reference to an unmatched subpattern matches an empty string
2524            (Perl fails the current match path).
2525    
2526        (c) A data ] in a character class must be notated as \] because if the
2527            first data character in a class is ], it defines an empty class. (In
2528            Perl it is not possible to have an empty class.) The empty class []
2529            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2530            The negative empty class [^] matches any one character, independently
2531            of the DOTALL setting.
2532    
2533    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2534        non-existent subpattern following a character class starting with ']' and
2535        containing () gave an internal compiling error instead of "reference to
2536        non-existent subpattern". Fortunately, when the pattern did exist, the
2537        compiled code was correct. (When scanning forwards to check for the
2538        existence of the subpattern, it was treating the data ']' as terminating
2539        the class, so got the count wrong. When actually compiling, the reference
2540        was subsequently set up correctly.)
2541    
2542    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2543        it was being rejected as not supported by pcre_dfa_exec(), even though
2544        other assertions are supported. I have made pcre_dfa_exec() support
2545        (*FAIL).
2546    
2547    16. The implementation of 13c above involved the invention of a new opcode,
2548        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2549        cannot be changed at match time, I realized I could make a small
2550        improvement to matching performance by compiling OP_ALLANY instead of
2551        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2552        on the OP_ANY path.
2553    
2554    17. Compiling pcretest on Windows with readline support failed without the
2555        following two fixes: (1) Make the unistd.h include conditional on
2556        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2557    
2558    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2559        ncurses library to be included for pcretest when ReadLine support is
2560        requested, but also to allow for it to be overridden. This patch came from
2561        Daniel Bergström.
2562    
2563    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2564        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2565        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2566        spotting this.
2567    
2568    
2569    Version 7.6 28-Jan-08
2570    ---------------------
2571    
2572    1.  A character class containing a very large number of characters with
2573        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2574        overflow.
2575    
2576    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2577        HAVE_LONG_LONG is not defined.
2578    
2579    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2580        bring them up to date and include new features. This patch includes:
2581    
2582        - Fixed PH's badly added libz and libbz2 support.
2583        - Fixed a problem with static linking.
2584        - Added pcredemo. [But later removed - see 7 below.]
2585        - Fixed dftables problem and added an option.
2586        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2587            HAVE_LONG_LONG.
2588        - Added readline support for pcretest.
2589        - Added an listing of the option settings after cmake has run.
2590    
2591    4.  A user submitted a patch to Makefile that makes it easy to create
2592        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2593        Makefile.am that cause it to include this special target, without
2594        affecting anything else. Note that the same mingw target plus all
2595        the other distribution libraries and programs are now supported
2596        when configuring with CMake (see 6 below) instead of with
2597        Configure/Make.
2598    
2599    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2600        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2601        exported in the Windows port". It has not yet been confirmed that the patch
2602        solves the problem, but it does no harm.
2603    
2604    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2605        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2606        with CMake, and also correct the comment about stack recursion.
2607    
2608    7.  Remove the automatic building of pcredemo from the ./configure system and
2609        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2610        of a program that users should build themselves after PCRE is installed, so
2611        building it automatically is not really right. What is more, it gave
2612        trouble in some build environments.
2613    
2614    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2615    
2616    
2617    Version 7.5 10-Jan-08
2618    ---------------------
2619    
2620    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2621        values in parens when parsing an RE using the C++ wrapper."
2622    
2623    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2624        Characters greater than 255 were excluded from the class instead of being
2625        included.
2626    
2627    3.  The same bug as (2) above applied to negated POSIX classes such as
2628        [:^space:].
2629    
2630    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2631        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2632        I have changed it.
2633    
2634    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2635        first named subpattern) and a construct such as (?&a) would reference the
2636        first named subpattern whose name started with "a" (in other words, the
2637        length check was missing). Both these problems are fixed. "Subpattern name
2638        expected" is now given for (?&) (a zero-length name), and this patch also
2639        makes it give the same error for \k'' (previously it complained that that
2640        was a reference to a non-existent subpattern).
2641    
2642    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2643        this is right because (?- can be followed by option settings as well as by
2644        digits. I have, however, made the messages clearer.
2645    
2646    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2647        than the number used in the conditional) now cause a compile-time error.
2648        This is actually not compatible with Perl, which accepts such patterns, but
2649        treats the conditional as always being FALSE (as PCRE used to), but it
2650        seems to me that giving a diagnostic is better.
2651    
2652    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2653        and messages.
2654    
2655    9.  Fix two occurrences of "backslash" in comments that should have been
2656        "backspace".
2657    
2658    10. Remove two redundant lines of code that can never be obeyed (their function
2659        was moved elsewhere).
2660    
2661    11. The program that makes PCRE's Unicode character property table had a bug
2662        which caused it to generate incorrect table entries for sequences of
2663        characters that have the same character type, but are in different scripts.
2664        It amalgamated them into a single range, with the script of the first of
2665        them. In other words, some characters were in the wrong script. There were
2666        thirteen such cases, affecting characters in the following ranges:
2667    
2668          U+002b0 - U+002c1
2669          U+0060c - U+0060d
2670          U+0061e - U+00612
2671          U+0064b - U+0065e
2672          U+0074d - U+0076d
2673          U+01800 - U+01805
2674          U+01d00 - U+01d77
2675          U+01d9b - U+01dbf
2676          U+0200b - U+0200f
2677          U+030fc - U+030fe
2678          U+03260 - U+0327f
2679          U+0fb46 - U+0fbb1
2680          U+10450 - U+1049d
2681    
2682    12. The -o option (show only the matching part of a line) for pcregrep was not
2683        compatible with GNU grep in that, if there was more than one match in a
2684        line, it showed only the first of them. It now behaves in the same way as
2685        GNU grep.
2686    
2687    13. If the -o and -v options were combined for pcregrep, it printed a blank
2688        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2689        does the same. The return code can be used to tell if there were any
2690        non-matching lines.
2691    
2692    14. Added --file-offsets and --line-offsets to pcregrep.
2693    
2694    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2695        infinitely looping recursion. The bug was that positive lookaheads were not
2696        being skipped when checking for a possible empty match (negative lookaheads
2697        and both kinds of lookbehind were skipped).
2698    
2699    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2700        inclusion of <windows.h> to before rather than after the definition of
2701        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2702    
2703    17. Specifying a possessive quantifier with a specific limit for a Unicode
2704        character property caused pcre_compile() to compile bad code, which led at
2705        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2706        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2707        caused the error; without that there was no problem.
2708    
2709    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2710    
2711    19. Added --enable-pcretest-libreadline.
2712    
2713    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2714        RE::GlobalReplace(). As a result, the number of replacements returned was
2715        double what it should be. I removed one of the increments, but Craig sent a
2716        later patch that removed the other one (the right fix) and added unit tests
2717        that check the return values (which was not done before).
2718    
2719    21. Several CMake things:
2720    
2721        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2722            the names libpcre and libpcreposix, not just pcre and pcreposix.
2723    
2724        (2) The above change means that pcretest and pcregrep are now correctly
2725            linked with the newly-built libraries, not previously installed ones.
2726    
2727        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2728    
2729    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2730        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2731        UTF-8 newline character). The key issue is that the pattern starts .*;
2732        this means that the match must be either at the beginning, or after a
2733        newline. The bug was in the code for advancing after a failed match and
2734        checking that the new position followed a newline. It was not taking
2735        account of UTF-8 characters correctly.
2736    
2737    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2738        character classes. PCRE was not treating the sequence [:...:] as a
2739        character class unless the ... were all letters. Perl, however, seems to
2740        allow any characters between [: and :], though of course it rejects as
2741        unknown any "names" that contain non-letters, because all the known class
2742        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2743        for example, whereas PCRE did not - it did not recognize a POSIX character
2744        class. This seemed a bit dangerous, so the code has been changed to be
2745        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2746        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2747        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2748        Perl does, and where it didn't before.
2749    
2750    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2751        Windows environments %n is disabled by default.
2752    
2753    
2754    Version 7.4 21-Sep-07
2755    ---------------------
2756    
2757    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2758        means that a class such as [\s] counted as "explicit reference to CR or
2759        LF". That isn't really right - the whole point of the change was to try to
2760        help when there was an actual mention of one of the two characters. So now
2761        the change happens only if \r or \n (or a literal CR or LF) character is
2762        encountered.
2763    
2764    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2765        of both had grown to the point where there were only 3 bits left.
2766        Fortunately, there was spare space in the data structure, and so I have
2767        moved the internal flags into a new 16-bit field to free up more option
2768        bits.
2769    
2770    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2771        but did not set the internal JCHANGED flag - either of these is enough to
2772        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2773        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2774        start sets both bits.
2775    
2776    4.  Added options (at build time, compile time, exec time) to change \R from
2777        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2778    
2779    5.  doc/pcresyntax.html was missing from the distribution.
2780    
2781    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2782        compatibility, even though it is no longer used.
2783    
2784    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2785        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2786        windows.h file is present (where different names are used). [This was
2787        reversed later after testing - see 16 below.]
2788    
2789    8.  Changed all #include <config.h> to #include "config.h". There were also
2790        some further <pcre.h> cases that I changed to "pcre.h".
2791    
2792    9.  When pcregrep was used with the --colour option, it missed the line ending
2793        sequence off the lines that it output.
2794    
2795    10. It was pointed out to me that arrays of string pointers cause lots of
2796        relocations when a shared library is dynamically loaded. A technique of
2797        using a single long string with a table of offsets can drastically reduce
2798        these. I have refactored PCRE in four places to do this. The result is
2799        dramatic:
2800    
2801          Originally:                          290
2802          After changing UCP table:            187
2803          After changing error message table:   43
2804          After changing table of "verbs"       36
2805          After changing table of Posix names   22
2806    
2807        Thanks to the folks working on Gregex for glib for this insight.
2808    
2809    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2810        unicode-properties was also set.
2811    
2812    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2813    
2814    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2815        checked only for CRLF.
2816    
2817    14. Added casts to pcretest.c to avoid compiler warnings.
2818    
2819    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2820    
2821    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2822        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2823        entirely. This removes changes made in 7 above.
2824    
2825    17. The CMake files have been updated, and there is now more information about
2826        building with CMake in the NON-UNIX-USE document.
2827    
2828    
2829    Version 7.3 28-Aug-07
2830    ---------------------
2831    
2832     1. In the rejigging of the build system that eventually resulted in 7.1, the
2833        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2834        brackets there is not right, since it causes compilers to look for an
2835        installed pcre.h, not the version that is in the source that is being
2836        compiled (which of course may be different). I have changed it back to:
2837    
2838          #include "pcre.h"
2839    
2840        I have a vague recollection that the change was concerned with compiling in
2841        different directories, but in the new build system, that is taken care of
2842        by the VPATH setting the Makefile.
2843    
2844     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2845        when the subject happened to end in the byte 0x85 (e.g. if the last
2846        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2847        characters but of course it shouldn't be taken as a newline when it is part
2848        of another character. The bug was that, for an unlimited repeat of . in
2849        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2850        characters when looking for a newline.
2851    
2852     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2853    
2854     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2855        in debug output.
2856    
2857     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2858        long printing in the pcrecpp unittest when running under MinGW.
2859    
2860     6. ESC_K was left out of the EBCDIC table.
2861    
2862     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2863        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2864        limit also applies to "virtual nesting" when a pattern is recursive, and in
2865        this case 1000 isn't so big. I have been able to remove this limit at the
2866        expense of backing off one optimization in certain circumstances. Normally,
2867        when pcre_exec() would call its internal match() function recursively and
2868        immediately return the result unconditionally, it uses a "tail recursion"
2869        feature to save stack. However, when a subpattern that can match an empty
2870        string has an unlimited repetition quantifier, it no longer makes this
2871        optimization. That gives it a stack frame in which to save the data for
2872        checking that an empty string has been matched. Previously this was taken
2873        from the 1000-entry workspace that had been reserved. So now there is no
2874        explicit limit, but more stack is used.
2875    
2876     8. Applied Daniel's patches to solve problems with the import/export magic
2877        syntax that is required for Windows, and which was going wrong for the
2878        pcreposix and pcrecpp parts of the library. These were overlooked when this
2879        problem was solved for the main library.
2880    
2881     9. There were some crude static tests to avoid integer overflow when computing
2882        the size of patterns that contain repeated groups with explicit upper
2883        limits. As the maximum quantifier is 65535, the maximum group length was
2884        set at 30,000 so that the product of these two numbers did not overflow a
2885        32-bit integer. However, it turns out that people want to use groups that
2886        are longer than 30,000 bytes (though not repeat them that many times).
2887        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2888        made it possible to implement the integer overflow checks in a much more
2889        dynamic way, which I have now done. The artificial limitation on group
2890        length has been removed - we now have only the limit on the total length of
2891        the compiled pattern, which depends on the LINK_SIZE setting.
2892    
2893    10. Fixed a bug in the documentation for get/copy named substring when
2894        duplicate names are permitted. If none of the named substrings are set, the
2895        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2896        empty string.
2897    
2898    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2899        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2900        because the ] is interpreted as the first data character and the
2901        terminating ] is not found. PCRE has been made compatible with Perl in this
2902        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2903        cause memory overwriting.
2904    
2905    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2906        string has been matched (to stop an infinite loop). It was not recognizing
2907        a conditional subpattern that could match an empty string if that
2908        subpattern was within another subpattern. For example, it looped when
2909        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2910        condition was not nested. This bug has been fixed.
2911    
2912    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2913        past the start of the subject in the presence of bytes with the top bit
2914        set, for example "\x8aBCD".
2915    
2916    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2917        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2918    
2919    14. Optimized (?!) to (*FAIL).
2920    
2921    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2922        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2923        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2924        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2925        does: it's just the validity check that is more restrictive.
2926    
2927    16. Inserted checks for integer overflows during escape sequence (backslash)
2928        processing, and also fixed erroneous offset values for syntax errors during
2929        backslash processing.
2930    
2931    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2932        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2933    
2934    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2935        caused an overrun.
2936    
2937    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2938        something other than just ASCII characters) inside a group that had an
2939        unlimited repeat caused a loop at compile time (while checking to see
2940        whether the group could match an empty string).
2941    
2942    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2943        [\P{Any}] did so. (Error in the code for printing property names.)
2944    
2945    21. An orphan \E inside a character class could cause a crash.
2946    
2947    22. A repeated capturing bracket such as (A)? could cause a wild memory
2948        reference during compilation.
2949    
2950    23. There are several functions in pcre_compile() that scan along a compiled
2951        expression for various reasons (e.g. to see if it's fixed length for look
2952        behind). There were bugs in these functions when a repeated \p or \P was
2953        present in the pattern. These operators have additional parameters compared
2954        with \d, etc, and these were not being taken into account when moving along
2955        the compiled data. Specifically:
2956    
2957        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2958            length.
2959    
2960        (b) An item such as \pL+ within a repeated group could cause crashes or
2961            loops.
2962    
2963        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2964            "reference to non-existent subpattern" error.
2965    
2966        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2967    
2968    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2969        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2970    
2971    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2972    
2973    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2974        character were causing crashes (broken optimization).
2975    
2976    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2977        \p or \P) caused a compile-time loop.
2978    
2979    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2980        break. For example, the unstudied pattern [\r\n]A does not match the string
2981        "\r\nA" because change 7.0/46 below moves the current point on by two
2982        characters after failing to match at the start. However, the pattern \nA
2983        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2984        the same is true. There doesn't seem any very clean way out of this, but
2985        what I have chosen to do makes the common cases work: PCRE now takes note
2986        of whether there can be an explicit match for \r or \n anywhere in the
2987        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2988        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2989        pattern has explicit CR or LF references.
2990    
2991    29. Added (*CR) etc for changing newline setting at start of pattern.
2992    
2993    
2994    Version 7.2 19-Jun-07
2995    ---------------------
2996    
2997     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2998        which is apparently normally available under Windows.
2999    
3000     2. Re-jig the pcregrep tests with different newline settings in an attempt
3001        to make them independent of the local environment's newline setting.
3002    
3003     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3004    
3005     4. Some of the "internals" tests were previously cut out when the link size
3006        was not 2, because the output contained actual offsets. The recent new
3007        "Z" feature of pcretest means that these can be cut out, making the tests
3008        usable with all link sizes.
3009    
3010     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3011        stack recursion. This gives a massive performance boost under BSD, but just
3012        a small improvement under Linux. However, it saves one field in the frame
3013        in all cases.
3014    
3015     6. Added more features from the forthcoming Perl 5.10:
3016    
3017        (a) (?-n) (where n is a string of digits) is a relative subroutine or
3018            recursion call. It refers to the nth most recently opened parentheses.
3019    
3020        (b) (?+n) is also a relative subroutine call; it refers to the nth next
3021            to be opened parentheses.
3022    
3023        (c) Conditions that refer to capturing parentheses can be specified
3024            relatively, for example, (?(-2)... or (?(+3)...
3025    
3026        (d) \K resets the start of the current match so that everything before
3027            is not part of it.
3028    
3029        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3030    
3031        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3032            reference syntax.
3033    
3034        (g) (?| introduces a group in which the numbering of parentheses in each
3035            alternative starts with the same number.
3036    
3037        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3038    
3039     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3040        PCRE_INFO_JCHANGED.
3041    
3042     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3043        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3044        for detecting groups that can match an empty string.
3045    
3046     9. A pattern with a very large number of alternatives (more than several
3047        hundred) was running out of internal workspace during the pre-compile
3048        phase, where pcre_compile() figures out how much memory will be needed. A
3049        bit of new cunning has reduced the workspace needed for groups with
3050        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3051        workspace instead of running out of the 4096 that are available.
3052    
3053    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3054    
3055    11. Applied patch from Google to remove an optimization that didn't quite work.
3056        The report of the bug said:
3057    
3058          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3059          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3060          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3061    
3062    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3063        it matched the wrong number of bytes.
3064    
3065    
3066    Version 7.1 24-Apr-07
3067    ---------------------
3068    
3069     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3070        that is more "standard", making use of automake and other Autotools. There
3071        is some re-arrangement of the files and adjustment of comments consequent
3072        on this.
3073    
3074     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3075        for recursive directory scanning broke on some systems because the files
3076        are not scanned in any specific order and on different systems the order
3077        was different. A call to "sort" has been inserted into RunGrepTest for the
3078        approprate test as a short-term fix. In the longer term there may be an
3079        alternative.
3080    
3081     3. I had an email from Eric Raymond about problems translating some of PCRE's
3082        man pages to HTML (despite the fact that I distribute HTML pages, some
3083        people do their own conversions for various reasons). The problems
3084        concerned the use of low-level troff macros .br and .in. I have therefore
3085        removed all such uses from the man pages (some were redundant, some could
3086        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3087        HTML has been updated to handle .nf/.fi and to complain if it encounters
3088        .br or .in.
3089    
3090     4. Updated comments in configure.ac that get placed in config.h.in and also
3091        arranged for config.h to be included in the distribution, with the name
3092        config.h.generic, for the benefit of those who have to compile without
3093        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3094    
3095     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3096        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3097        makevp.bat for the current PCRE, using the additional files
3098        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3099    
3100     6. A Windows user reported a minor discrepancy with test 2, which turned out
3101        to be caused by a trailing space on an input line that had got lost in his
3102        copy. The trailing space was an accident, so I've just removed it.
3103    
3104     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3105        that is needed.
3106    
3107     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3108        as "const" (a) because they are and (b) because it helps the PHP
3109        maintainers who have recently made a script to detect big data structures
3110        in the php code that should be moved to the .rodata section. I remembered
3111        to update Builducptable as well, so it won't revert if ucptable.h is ever
3112        re-created.
3113    
3114     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3115        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3116        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3117        support is not required. This saves 1.5-2K of code, which is important in
3118        some applications.
3119    
3120        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3121        so as not to refer to the tables, even though these functions will never be
3122        called when UTF-8 support is disabled. Otherwise there are problems with a
3123        shared library.
3124    
3125    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3126    
3127        (a) It was defining its arguments as char * instead of void *.
3128    
3129        (b) It was assuming that all moves were upwards in memory; this was true
3130            a long time ago when I wrote it, but is no longer the case.
3131    
3132        The emulated memove() is provided for those environments that have neither
3133        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3134        is clearly not the case, as these two bugs were recently reported.
3135    
3136    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3137        and Detrail to create the HTML documentation, the .txt form of the man
3138        pages, and it removes trailing spaces from listed files. It also creates
3139        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3140        case, it wraps all the #defines with #ifndefs. This script should be run
3141        before "make dist".
3142    
3143    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3144        with Unicode property support.
3145    
3146        (a) For a maximizing quantifier, if the two different cases of the
3147            character were of different lengths in their UTF-8 codings (there are
3148            some cases like this - I found 11), and the matching function had to
3149            back up over a mixture of the two cases, it incorrectly assumed they
3150            were both the same length.
3151    
3152        (b) When PCRE was configured to use the heap rather than the stack for
3153            recursion during matching, it was not correctly preserving the data for
3154            the other case of a UTF-8 character when checking ahead for a match
3155            while processing a minimizing repeat. If the check also involved
3156            matching a wide character, but failed, corruption could cause an
3157            erroneous result when trying to check for a repeat of the original
3158            character.
3159    
3160    13. Some tidying changes to the testing mechanism:
3161    
3162        (a) The RunTest script now detects the internal link size and whether there
3163            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3164            values substituted by "configure". (The RunGrepTest script already did
3165            this for UTF-8.) The configure.ac script no longer substitutes the
3166            relevant variables.
3167    
3168        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3169            with length and offset values. This means that the output is different
3170            for different internal link sizes. Test 2 is skipped for link sizes
3171            other than 2 because of this, bypassing the problem. Unfortunately,
3172            there was also a test in test 3 (the locale tests) that used /B and
3173            failed for link sizes other than 2. Rather than cut the whole test out,
3174            I have added a new /Z option to pcretest that replaces the length and
3175            offset values with spaces. This is now used to make test 3 independent
3176            of link size. (Test 2 will be tidied up later.)
3177    
3178    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3179        segmentation fault instead of returning the appropriate error message.
3180    
3181    15. In multiline mode when the newline sequence was set to "any", the pattern
3182        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3183        This doesn't seem right; it now treats the CRLF combination as the line
3184        ending, and so does not match in that case. It's only a pattern such as ^$
3185        that would hit this one: something like ^ABC$ would have failed after \r
3186        and then tried again after \r\n.
3187    
3188    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3189        in an attempt to make files that differ only in their line terminators
3190        compare equal. This works on Linux.
3191    
3192    17. Under certain error circumstances pcregrep might try to free random memory
3193        as it exited. This is now fixed, thanks to valgrind.
3194    
3195    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3196        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3197        was because its rules for how to advance for /g after matching an empty
3198        string at the end of a line did not allow for this case. They now check for
3199        it specially.
3200    
3201    20. pcretest is supposed to handle patterns and data of any length, by
3202        extending its buffers when necessary. It was getting this wrong when the
3203        buffer for a data line had to be extended.
3204    
3205    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3206        CRLF as a newline sequence.
3207    
3208    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3209        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3210        I have nevertheless tidied it up.
3211    
3212    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3213    
3214    24. Added a man page for pcre-config.
3215    
3216    
3217    Version 7.0 19-Dec-06
3218    ---------------------
3219    
3220     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3221        moving to gcc 4.1.1.
3222    
3223     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3224        sys/time.h, which is documented as needed for this function. It doesn't
3225        seem to matter on Linux, but it showed up on some releases of OS X.
3226    
3227     3. It seems that there are systems where bytes whose values are greater than
3228        127 match isprint() in the "C" locale. The "C" locale should be the
3229        default when a C program starts up. In most systems, only ASCII printing
3230        characters match isprint(). This difference caused the output from pcretest
3231        to vary, making some of the tests fail. I have changed pcretest so that:
3232    
3233        (a) When it is outputting text in the compiled version of a pattern, bytes
3234            other than 32-126 are always shown as hex escapes.
3235    
3236        (b) When it is outputting text that is a matched part of a subject string,
3237            it does the same, unless a different locale has been set for the match
3238            (using the /L modifier). In this case, it uses isprint() to decide.
3239    
3240     4. Fixed a major bug that caused incorrect computation of the amount of memory
3241        required for a compiled pattern when options that changed within the
3242        pattern affected the logic of the preliminary scan that determines the
3243        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3244        that the computed length was too small. The symptoms of this bug were
3245        either the PCRE error "internal error: code overflow" from pcre_compile(),
3246        or a glibc crash with a message such as "pcretest: free(): invalid next
3247        size (fast)". Examples of patterns that provoked this bug (shown in
3248        pcretest format) are:
3249    
3250          /(?-x: )/x
3251          /(?x)(?-x: \s*#\s*)/
3252          /((?i)[\x{c0}])/8
3253          /(?i:[\x{c0}])/8
3254    
3255        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3256        is now done differently.
3257    
3258     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3259        wrapper classes; (b) implement a new function in the C++ scanner that is
3260        more efficient than the old way of doing things because it avoids levels of
3261        recursion in the regex matching; (c) add a paragraph to the documentation
3262        for the FullMatch() function.
3263    
3264     6. The escape sequence \n was being treated as whatever was defined as
3265        "newline". Not only was this contrary to the documentation, which states
3266        that \n is character 10 (hex 0A), but it also went horribly wrong when
3267        "newline" was defined as CRLF. This has been fixed.
3268    
3269     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3270        was being set to -1 for the "end of line" case (supposedly a value that no
3271        character can have). Though this value is never used (the check for end of
3272        line is "zero bytes in current character"), it caused compiler complaints.
3273        I've changed it to 0xffffffff.
3274    
3275     8. In pcre_version.c, the version string was being built by a sequence of
3276        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3277        string (as it is for production releases) called a macro with an empty
3278        argument. The C standard says the result of this is undefined. The gcc
3279        compiler treats it as an empty string (which was what was wanted) but it is
3280        reported that Visual C gives an error. The source has been hacked around to
3281        avoid this problem.
3282    
3283     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3284        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3285        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3286        of them did).
3287    
3288    10. Originally, pcretest opened its input and output without "b"; then I was
3289        told that "b" was needed in some environments, so it was added for release
3290        5.0 to both the input and output. (It makes no difference on Unix-like
3291        systems.) Later I was told that it is wrong for the input on Windows. I've
3292        now abstracted the modes into two macros, to make it easier to fiddle with
3293        them, and removed "b" from the input mode under Windows.
3294    
3295    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3296    
3297    12. Added -help and --help to pcretest as an official way of being reminded
3298        of the options.
3299    
3300    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3301        and pcrecpp.cc because they annoy compilers at high warning levels.
3302    
3303    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3304    
3305    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3306        scripts are not C programs :-) and which was not noticed because it works
3307        on Linux.
3308    
3309    16. pcretest is supposed to handle any length of pattern and data line (as one
3310        line or as a continued sequence of lines) by extending its input buffer if
3311        necessary. This feature was broken for very long pattern lines, leading to
3312        a string of junk being passed to pcre_compile() if the pattern was longer
3313        than about 50K.
3314    
3315    17. I have done a major re-factoring of the way pcre_compile() computes the
3316        amount of memory needed for a compiled pattern. Previously, there was code
3317        that made a preliminary scan of the pattern in order to do this. That was
3318        OK when PCRE was new, but as the facilities have expanded, it has become
3319        harder and harder to keep it in step with the real compile phase, and there
3320        have been a number of bugs (see for example, 4 above). I have now found a
3321        cunning way of running the real compile function in a "fake" mode that
3322        enables it to compute how much memory it would need, while actually only
3323        ever using a few hundred bytes of working memory and without too many
3324        tests of the mode. This should make future maintenance and development
3325        easier. A side effect of this work is that the limit of 200 on the nesting
3326        depth of parentheses has been removed (though this was never a serious
3327        limitation, I suspect). However, there is a downside: pcre_compile() now
3328        runs more slowly than before (30% or more, depending on the pattern). I
3329        hope this isn't a big issue. There is no effect on runtime performance.
3330    
3331    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3332        newline (only possible for the last line of a file) and it was a
3333        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3334    
3335    19. Added additional timing features to pcretest. (1) The -tm option now times
3336        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3337        separate command line item, by a number that specifies the number of
3338        repeats to use when timing. The default is 50000; this gives better
3339        precision, but takes uncomfortably long for very large patterns.
3340    
3341    20. Extended pcre_study() to be more clever in cases where a branch of a
3342        subpattern has no definite first character. For example, (a*|b*)[cd] would
3343        previously give no result from pcre_study(). Now it recognizes that the
3344        first character must be a, b, c, or d.
3345    
3346    21. There was an incorrect error "recursive call could loop indefinitely" if
3347        a subpattern (or the entire pattern) that was being tested for matching an
3348        empty string contained only one non-empty item after a nested subpattern.
3349        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3350        incorrectly, because the \d was being skipped in the check.
3351    
3352    22. The pcretest program now has a new pattern option /B and a command line
3353        option -b, which is equivalent to adding /B to every pattern. This causes
3354        it to show the compiled bytecode, without the additional information that
3355        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3356        is the same as /B/I).
3357    
3358    23. A new optimization is now able automatically to treat some sequences such
3359        as a*b as a*+b. More specifically, if something simple (such as a character
3360        or a simple class like \d) has an unlimited quantifier, and is followed by
3361        something that cannot possibly match the quantified thing, the quantifier
3362        is automatically "possessified".
3363    
3364    24. A recursive reference to a subpattern whose number was greater than 39
3365        went wrong under certain circumstances in UTF-8 mode. This bug could also
3366        have affected the operation of pcre_study().
3367    
3368    25. Realized that a little bit of performance could be had by replacing
3369        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3370    
3371    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3372    
3373    27. Possessive quantifiers such as a++ were previously implemented by turning
3374        them into atomic groups such as ($>a+). Now they have their own opcodes,
3375        which improves performance. This includes the automatically created ones
3376        from 23 above.
3377    
3378    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3379        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3380        the first matched character to be a colon. This applied both to named and
3381        numbered groups.
3382    
3383    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3384    
3385    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3386        building PCRE on Windows easier, so I have included it in the distribution.
3387    
3388    31. There is now a check in pcretest against a ridiculously large number being
3389        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3390        loop, the loop is abandoned.
3391    
3392    32. Forward references to subpatterns in conditions such as (?(2)...) where
3393        subpattern 2 is defined later cause pcre_compile() to search forwards in
3394        the pattern for the relevant set of parentheses. This search went wrong
3395        when there were unescaped parentheses in a character class, parentheses
3396        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3397    
3398    33. "Subroutine" calls and backreferences were previously restricted to
3399        referencing subpatterns earlier in the regex. This restriction has now
3400        been removed.
3401    
3402    34. Added a number of extra features that are going to be in Perl 5.10. On the
3403        whole, these are just syntactic alternatives for features that PCRE had
3404        previously implemented using the Python syntax or my own invention. The
3405        other formats are all retained for compatibility.
3406    
3407        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3408            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3409            also .NET compatible.
3410    
3411        (b) A recursion or subroutine call to a named group can now be defined as
3412            (?&name) as well as (?P>name).
3413    
3414        (c) A backreference to a named group can now be defined as \k<name> or
3415            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3416            5.10, are also .NET compatible.
3417    
3418        (d) A conditional reference to a named group can now use the syntax
3419            (?(<name>) or (?('name') as well as (?(name).
3420    
3421        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3422            groups (named and numbered) that are never evaluated inline, but can be
3423            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3424            is always false. There may be only one alternative in such a group.
3425    
3426        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3427            as the simple (?(R). The condition is true only if the most recent
3428            recursion is that of the given number or name. It does not search out
3429            through the entire recursion stack.
3430    
3431        (g) The escape \gN or \g{N} has been added, where N is a positive or
3432            negative number, specifying an absolute or relative reference.
3433    
3434    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3435        some "unreachable code" warnings.
3436    
3437    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3438        things, this adds five new scripts.
3439    
3440    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3441        There were also incompatibilities regarding the handling of \Q..\E inside
3442        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3443        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3444    
3445    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3446        matches an empty string, and forcibly breaks the loop. There were bugs in
3447        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3448        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3449        separate and independent bugs (that affected different cases) have been
3450        fixed.
3451    
3452    39. Refactored the code to abolish the use of different opcodes for small
3453        capturing bracket numbers. This is a tidy that I avoided doing when I
3454        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3455        The new approach is not only tidier, it makes it possible to reduce the
3456        memory needed to fix the previous bug (38).
3457    
3458    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3459        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3460        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3461        mode.
3462    
3463    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3464        report.
3465    
3466    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3467        copy construction and assignment in the C++ wrapper.
3468    
3469    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3470        couple of bugs that could have given wrong results in the "--newline=crlf"
3471        case.
3472    
3473    44. Added a number of casts and did some reorganization of signed/unsigned int
3474        variables following suggestions from Dair Grant. Also renamed the variable
3475        "this" as "item" because it is a C++ keyword.
3476    
3477    45. Arranged for dftables to add
3478    
3479          #include "pcre_internal.h"
3480    
3481        to pcre_chartables.c because without it, gcc 4.x may remove the array
3482        definition from the final binary if PCRE is built into a static library and
3483        dead code stripping is activated.
3484    
3485    46. For an unanchored pattern, if a match attempt fails at the start of a
3486        newline sequence, and the newline setting is CRLF or ANY, and the next two
3487        characters are CRLF, advance by two characters instead of one.
3488    
3489    
3490    Version 6.7 04-Jul-06
3491    ---------------------
3492    
3493     1. In order to handle tests when input lines are enormously long, pcretest has
3494        been re-factored so that it automatically extends its buffers when
3495        necessary. The code is crude, but this _is_ just a test program. The
3496        default size has been increased from 32K to 50K.
3497    
3498     2. The code in pcre_study() was using the value of the re argument before
3499        testing it for NULL. (Of course, in any sensible call of the function, it
3500        won't be NULL.)
3501    
3502     3. The memmove() emulation function in pcre_internal.h, which is used on
3503        systems that lack both memmove() and bcopy() - that is, hardly ever -
3504        was missing a "static" storage class specifier.
3505    
3506     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3507        containing an extended class (one that cannot be represented by a bitmap
3508        because it contains high-valued characters or Unicode property items, e.g.
3509        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3510        pattern, but PCRE should not loop if you do not (it no longer does).
3511        [Detail: two cases were found: (a) a repeated subpattern containing an
3512        extended class; (b) a recursive reference to a subpattern that followed a
3513        previous extended class. It wasn't skipping over the extended class
3514        correctly when UTF-8 mode was not set.]
3515    
3516     5. A negated single-character class was not being recognized as fixed-length
3517        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3518        compile error "lookbehind assertion is not fixed length".
3519    
3520     6. The RunPerlTest auxiliary script was showing an unexpected difference
3521        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3522        write a Perl script that can interpret lines of an input file either as
3523        byte characters or as UTF-8, which is what "perltest" was being required to
3524        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3525        can't do is switch easily at run time between having the "use utf8;" pragma
3526        or not. In the end, I fudged it by using the RunPerlTest script to insert
3527        "use utf8;" explicitly for the UTF-8 tests.
3528    
3529     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3530        the end of the subject string, contrary to the documentation and to what
3531        Perl does. This was true of both matching functions. Now it matches only at
3532        the start of the subject and immediately after *internal* newlines.
3533    
3534     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3535        a pointer to an int instead of a pointer to an unsigned long int. This
3536        caused problems on 64-bit systems.
3537    
3538     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3539        instance of the 'standard' template library not being so standard".
3540    
3541    10. There was no check on the number of named subpatterns nor the maximum
3542        length of a subpattern name. The product of these values is used to compute
3543        the size of the memory block for a compiled pattern. By supplying a very
3544        long subpattern name and a large number of named subpatterns, the size
3545        computation could be caused to overflow. This is now prevented by limiting
3546        the length of names to 32 characters, and the number of named subpatterns
3547        to 10,000.
3548    
3549    11. Subpatterns that are repeated with specific counts have to be replicated in
3550        the compiled pattern. The size of memory for this was computed from the
3551        length of the subpattern and the repeat count. The latter is limited to
3552        65535, but there was no limit on the former, meaning that integer overflow
3553        could in principle occur. The compiled length of a repeated subpattern is
3554        now limited to 30,000 bytes in order to prevent this.
3555    
3556    12. Added the optional facility to have named substrings with the same name.
3557    
3558    13. Added the ability to use a named substring as a condition, using the
3559        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3560        are numbers (not recommended). Forward references are permitted.
3561    
3562    14. Added forward references in named backreferences (if you see what I mean).
3563    
3564    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3565        pattern could run off the end of the subject. For example, the pattern
3566        "(?s)(.{1,5})"8 did this with the subject "ab".
3567    
3568    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3569        PCRE_CASELESS was set when matching characters that were quantified with ?
3570        or *.
3571    
3572    17. A character class other than a single negated character that had a minimum
3573        but no maximum quantifier - for example [ab]{6,} - was not handled
3574        correctly by pce_dfa_exec(). It would match only one character.
3575    
3576    18. A valid (though odd) pattern that looked like a POSIX character
3577        class but used an invalid character after [ (for example [[,abc,]]) caused
3578        pcre_compile() to give the error "Failed: internal error: code overflow" or
3579        in some cases to crash with a glibc free() error. This could even happen if
3580        the pattern terminated after [[ but there just happened to be a sequence of
3581        letters, a binary zero, and a closing ] in the memory that followed.
3582    
3583    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3584        over the years. Originally (before any Unicode support), just the bottom 8
3585        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3586        output from "man perlunicode" includes this:
3587    
3588          The regular expression compiler produces polymorphic opcodes.  That
3589          is, the pattern adapts to the data and automatically switches to
3590          the Unicode character scheme when presented with Unicode data--or
3591          instead uses a traditional byte scheme when presented with byte
3592          data.
3593    
3594        Sadly, a wide octal escape does not cause a switch, and in a string with
3595        no other multibyte characters, these octal escapes are treated as before.
3596        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3597        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3598        Unicode string.
3599    
3600        I have not perpetrated such confusion in PCRE. Up till now, it took just
3601        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3602        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3603        translate to the appropriate multibyte character.
3604    
3605    29. Applied some refactoring to reduce the number of warnings from Microsoft
3606        and Borland compilers. This has included removing the fudge introduced
3607        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3608        a warning about an unused variable.
3609    
3610    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3611        characters since release 4.0, because Perl (from release 5.004) does not.
3612        [Or at least, is documented not to: some releases seem to be in conflict
3613        with the documentation.] However, when a pattern was studied with
3614        pcre_study() and all its branches started with \s, PCRE still included VT
3615        as a possible starting character. Of course, this did no harm; it just
3616        caused an unnecessary match attempt.
3617    
3618    22. Removed a now-redundant internal flag bit that recorded the fact that case
3619        dependency changed within the pattern. This was once needed for "required
3620        byte" processing, but is no longer used. This recovers a now-scarce options
3621        bit. Also moved the least significant internal flag bit to the most-
3622        significant bit of the word, which was not previously used (hangover from
3623        the days when it was an int rather than a uint) to free up another bit for
3624        the future.
3625    
3626    23. Added support for CRLF line endings as well as CR and LF. As well as the
3627        default being selectable at build time, it can now be changed at runtime
3628        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3629        specify that it is scanning data with non-default line endings.
3630    
3631    24. Changed the definition of CXXLINK to make it agree with the definition of
3632        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3633    
3634    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3635        recursions. This makes a big different to stack usage for some patterns.
3636    
3637    26. If a subpattern containing a named recursion or subroutine reference such
3638        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3639        the space required for the compiled pattern went wrong and gave too small a
3640        value. Depending on the environment, this could lead to "Failed: internal
3641        error: code overflow at offset 49" or "glibc detected double free or
3642        corruption" errors.
3643    
3644    27. Applied patches from Google (a) to support the new newline modes and (b) to
3645        advance over multibyte UTF-8 characters in GlobalReplace.
3646    
3647    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3648        difference for some implementation of PCRE in some Windows version.
3649    
3650    29. Added some extra testing facilities to pcretest:
3651    
3652        \q<number>   in a data line sets the "match limit" value
3653        \Q<number>   in a data line sets the "match recursion limt" value
3654        -S <number>  sets the stack size, where <number> is in megabytes
3655    
3656        The -S option isn't available for Windows.
3657    
3658    
3659    Version 6.6 06-Feb-06
3660    ---------------------
3661    
3662     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3663        in pcreposix.h. I have copied the definition from pcre.h.
3664    
3665     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3666        because pcre.h is no longer a built file.
3667    
3668     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3669        not normally included in the compiled code.
3670    
3671    
3672    Version 6.5 01-Feb-06
3673    ---------------------
3674    
3675     1. When using the partial match feature with pcre_dfa_exec(), it was not
3676        anchoring the second and subsequent partial matches at the new starting
3677        point. This could lead to incorrect results. For example, with the pattern
3678        /1234/, partially matching against "123" and then "a4" gave a match.
3679    
3680     2. Changes to pcregrep:
3681    
3682        (a) All non-match returns from pcre_exec() were being treated as failures
3683            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3684            error message is output. Some extra information is given for the
3685            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3686            probably the only errors that are likely to be caused by users (by
3687            specifying a regex that has nested indefinite repeats, for instance).
3688            If there are more than 20 of these errors, pcregrep is abandoned.
3689    
3690        (b) A binary zero was treated as data while matching, but terminated the
3691            output line if it was written out. This has been fixed: binary zeroes
3692            are now no different to any other data bytes.
3693    
3694        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3695            used to set a locale for matching. The --locale=xxxx long option has
3696            been added (no short equivalent) to specify a locale explicitly on the
3697            pcregrep command, overriding the environment variables.
3698    
3699        (d) When -B was used with -n, some line numbers in the output were one less
3700            than they should have been.
3701    
3702        (e) Added the -o (--only-matching) option.
3703    
3704        (f) If -A or -C was used with -c (count only), some lines of context were
3705            accidentally printed for the final match.
3706    
3707        (g) Added the -H (--with-filename) option.
3708    
3709        (h) The combination of options -rh failed to suppress file names for files
3710            that were found from directory arguments.
3711    
3712        (i) Added the -D (--devices) and -d (--directories) options.
3713    
3714        (j) Added the -F (--fixed-strings) option.
3715    
3716        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3717    
3718        (l) Added the --colo(u)r option.
3719    
3720        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3721            is not present by default.
3722    
3723     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3724        items such as (?R) or (?1), when the recursion could match a number of
3725        alternatives. If it matched one of the alternatives, but subsequently,
3726        outside the recursion, there was a failure, the code tried to back up into
3727        the recursion. However, because of the way PCRE is implemented, this is not
3728        possible, and the result was an incorrect result from the match.
3729    
3730        In order to prevent this happening, the specification of recursion has
3731        been changed so that all such subpatterns are automatically treated as
3732        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3733    
3734     4. I had overlooked the fact that, in some locales, there are characters for
3735        which isalpha() is true but neither isupper() nor islower() are true. In
3736        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3737        and ordfeminine) are like this. This affected the treatment of \w and \W
3738        when they appeared in character classes, but not when they appeared outside
3739        a character class. The bit map for "word" characters is now created
3740        separately from the results of isalnum() instead of just taking it from the
3741        upper, lower, and digit maps. (Plus the underscore character, of course.)
3742    
3743     5. The above bug also affected the handling of POSIX character classes such as
3744        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3745        permanent tables. Instead, the bit maps for such a class were previously
3746        created as the appropriate unions of the upper, lower, and digit bitmaps.
3747        Now they are created by subtraction from the [[:word:]] class, which has
3748        its own bitmap.
3749    
3750     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3751        It is created by subtracting the vertical space characters (\x09, \x0a,
3752        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3753        subtraction was done in the overall bitmap for a character class, meaning
3754        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3755        be recognized. This bug has been fixed.
3756    
3757     7. Patches from the folks at Google:
3758    
3759          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3760          real life, but is still worth protecting against".
3761    
3762          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3763          regular expressions".
3764    
3765          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3766          have it.
3767    
3768          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3769          "configure" and the latter not, in order to fix a problem somebody had
3770          with compiling the Arg class on HP-UX.
3771    
3772          (e) Improve the error-handling of the C++ wrapper a little bit.
3773    
3774          (f) New tests for checking recursion limiting.
3775    
3776     8. The pcre_memmove() function, which is used only if the environment does not
3777        have a standard memmove() function (and is therefore rarely compiled),
3778        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3779        returning a result (though PCRE never actually uses the result).
3780    
3781     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3782        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3783        returned instead of calling malloc() with an overflowing number that would
3784        most likely cause subsequent chaos.
3785    
3786    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3787    
3788    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled