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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.200  
changed lines
  Added in v.1562

  ViewVC Help
Powered by ViewVC 1.1.5