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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.221  
changed lines
  Added in v.1555

  ViewVC Help
Powered by ViewVC 1.1.5