ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 538 by ph10, Wed Jun 9 19:30:57 2010 UTC revision 1649 by ph10, Sat May 21 13:34:44 2016 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4  Version 8.10 03-Jun-2010  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
7    Version 8.39 xx-xxxxxx-201x
8    ---------------------------
10    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12        misbehaved. This bug was found by the LLVM fuzzer.
14    2.  Similar to the above, if an isolated \E was present between an item and its
15        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16        bug was found by the LLVM fuzzer.
18    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19        working correctly in UCP mode.
21    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22        was set when the pmatch argument was NULL. It now returns REG_INVARG.
24    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
26    6.  An empty \Q\E sequence between an item and its qualifier caused
27        pcre_compile() to misbehave when auto callouts were enabled. This bug was
28        found by the LLVM fuzzer.
30    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
31        space or a #-type comment that was followed by (?-x), which turns off
32        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
33        pcre_compile() assumed that (?-x) applied to the whole pattern and
34        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
36    8.  An call of pcre_copy_named_substring() for a named substring whose number
37        was greater than the space in the ovector could cause a crash.
39    9.  Yet another buffer overflow bug involved duplicate named groups with a
40        group that reset capture numbers (compare 8.38/7 below). Once again, I have
41        just allowed for more memory, even if not needed. (A proper fix is
42        implemented in PCRE2, but it involves a lot of refactoring.)
44    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
45        start of the match to be earlier than the end.
47    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
49    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
50        assertion, caused pcretest to generate incorrect output, and also to read
51        uninitialized memory (detected by ASAN or valgrind).
53    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
54        nested set of parentheses of sufficient size caused an overflow of the
55        compiling workspace (which was diagnosed, but of course is not desirable).
57    14. And yet another buffer overflow bug involving duplicate named groups, this
58        time nested, with a nested back reference. Yet again, I have just allowed
59        for more memory, because anything more needs all the refactoring that has
60        been done for PCRE2. An example pattern that provoked this bug is:
61        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
62        registered as CVE-2016-1283.
64    15. pcretest went into a loop if global matching was requested with an ovector
65        size less than 2. It now gives an error message. This bug was found by
66        afl-fuzz.
68    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
69        ("assertion expected") when (?(?C) was not followed by an opening
70        parenthesis.
72    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
73        actually affect anything, by sheer luck.
75    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
76        static compilation.
78    19. Modified the RunTest script to incorporate a valgrind suppressions file so
79        that certain errors, provoked by the SSE2 instruction set when JIT is used,
80        are ignored.
82    20. A racing condition is fixed in JIT reported by Mozilla.
84    21. Minor code refactor to avoid "array subscript is below array bounds"
85        compiler warning.
87    22. Minor code refactor to avoid "left shift of negative number" warning.
90    Version 8.38 23-November-2015
91    -----------------------------
93    1.  If a group that contained a recursive back reference also contained a
94        forward reference subroutine call followed by a non-forward-reference
95        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
96        compile correct code, leading to undefined behaviour or an internally
97        detected error. This bug was discovered by the LLVM fuzzer.
99    2.  Quantification of certain items (e.g. atomic back references) could cause
100        incorrect code to be compiled when recursive forward references were
101        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
102        This bug was discovered by the LLVM fuzzer.
104    3.  A repeated conditional group whose condition was a reference by name caused
105        a buffer overflow if there was more than one group with the given name.
106        This bug was discovered by the LLVM fuzzer.
108    4.  A recursive back reference by name within a group that had the same name as
109        another group caused a buffer overflow. For example:
110        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
112    5.  A forward reference by name to a group whose number is the same as the
113        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
114        a buffer overflow at compile time. This bug was discovered by the LLVM
115        fuzzer.
117    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
118        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
120    7.  Another buffer overflow bug involved duplicate named groups with a
121        reference between their definition, with a group that reset capture
122        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
123        fixed by always allowing for more memory, even if not needed. (A proper fix
124        is implemented in PCRE2, but it involves more refactoring.)
126    8.  There was no check for integer overflow in subroutine calls such as (?123).
128    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
129        being treated as a literal 'l' instead of causing an error.
131    10. There was a buffer overflow if pcre_exec() was called with an ovector of
132        size 1. This bug was found by american fuzzy lop.
134    11. If a non-capturing group containing a conditional group that could match
135        an empty string was repeated, it was not identified as matching an empty
136        string itself. For example: /^(?:(?(1)x|)+)+$()/.
138    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
139        \a and \e in test subject lines.
141    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
142        instead of the EBCDIC value.
144    14. The handling of \c in an EBCDIC environment has been revised so that it is
145        now compatible with the specification in Perl's perlebcdic page.
147    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
148        ASCII/Unicode. This has now been added to the list of characters that are
149        recognized as white space in EBCDIC.
151    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
152        error (correctly) when used outside a class, but did not give an error
153        within a class.
155    17. \h within a class was incorrectly compiled in EBCDIC environments.
157    18. A pattern with an unmatched closing parenthesis that contained a backward
158        assertion which itself contained a forward reference caused buffer
159        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
161    19. JIT should return with error when the compiled pattern requires more stack
162        space than the maximum.
164    20. A possessively repeated conditional group that could match an empty string,
165        for example, /(?(R))*+/, was incorrectly compiled.
167    21. Fix infinite recursion in the JIT compiler when certain patterns such as
168        /(?:|a|){100}x/ are analysed.
170    22. Some patterns with character classes involving [: and \\ were incorrectly
171        compiled and could cause reading from uninitialized memory or an incorrect
172        error diagnosis.
174    23. Pathological patterns containing many nested occurrences of [: caused
175        pcre_compile() to run for a very long time.
177    24. A conditional group with only one branch has an implicit empty alternative
178        branch and must therefore be treated as potentially matching an empty
179        string.
181    25. If (?R was followed by - or + incorrect behaviour happened instead of a
182        diagnostic.
184    26. Arrange to give up on finding the minimum matching length for overly
185        complex patterns.
187    27. Similar to (4) above: in a pattern with duplicated named groups and an
188        occurrence of (?| it is possible for an apparently non-recursive back
189        reference to become recursive if a later named group with the relevant
190        number is encountered. This could lead to a buffer overflow. Wen Guanxing
191        from Venustech ADLAB discovered this bug.
193    28. If pcregrep was given the -q option with -c or -l, or when handling a
194        binary file, it incorrectly wrote output to stdout.
196    29. The JIT compiler did not restore the control verb head in case of *THEN
197        control verbs. This issue was found by Karl Skomski with a custom LLVM
198        fuzzer.
200    30. Error messages for syntax errors following \g and \k were giving inaccurate
201        offsets in the pattern.
203    31. Added a check for integer overflow in conditions (?(<digits>) and
204        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
205        fuzzer.
207    32. Handling recursive references such as (?2) when the reference is to a group
208        later in the pattern uses code that is very hacked about and error-prone.
209        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
210        give an internal error if it is obvious that compiling has gone wrong.
212    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
213        This issue was found by Karl Skomski with a custom LLVM fuzzer.
215    34. The JIT compiler should restore the control chain for empty possessive
216        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
218    35. Match limit check added to JIT recursion. This issue was found by Karl
219        Skomski with a custom LLVM fuzzer.
221    36. Yet another case similar to 27 above has been circumvented by an
222        unconditional allocation of extra memory. This issue is fixed "properly" in
223        PCRE2 by refactoring the way references are handled. Wen Guanxing
224        from Venustech ADLAB discovered this bug.
226    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
227        with a custom LLVM fuzzer.
229    38. Fixed a corner case of range optimization in JIT.
231    39. An incorrect error "overran compiling workspace" was given if there were
232        exactly enough group forward references such that the last one extended
233        into the workspace safety margin. The next one would have expanded the
234        workspace. The test for overflow was not including the safety margin.
236    40. A match limit issue is fixed in JIT which was found by Karl Skomski
237        with a custom LLVM fuzzer.
239    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
240        work under Windows. (Why has it taken so long for anyone to notice?)
242    42. In a character class such as [\W\p{Any}] where both a negative-type escape
243        ("not a word character") and a property escape were present, the property
244        escape was being ignored.
246    43. Fix crash caused by very long (*MARK) or (*THEN) names.
248    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
249        by a single ASCII character in a class item, was incorrectly compiled in
250        UCP mode. The POSIX class got lost, but only if the single character
251        followed it.
253    45. [:punct:] in UCP mode was matching some characters in the range 128-255
254        that should not have been matched.
256    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
257        class, all characters with code points greater than 255 are in the class.
258        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
259        such as \w are turned into Unicode properties), wide characters were not
260        correctly handled, and could fail to match.
263    Version 8.37 28-April-2015
264    --------------------------
266    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
267        for those parentheses to be closed with whatever has been captured so far.
268        However, it was failing to mark any other groups between the hightest
269        capture so far and the currrent group as "unset". Thus, the ovector for
270        those groups contained whatever was previously there. An example is the
271        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
273    2.  If an assertion condition was quantified with a minimum of zero (an odd
274        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
276    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
277        unrecognized modifier, a crash could occur.
279    4.  An attempt to do global matching in pcretest with a zero-length ovector
280        caused a crash.
282    5.  Fixed a memory leak during matching that could occur for a subpattern
283        subroutine call (recursive or otherwise) if the number of captured groups
284        that had to be saved was greater than ten.
286    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
287        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
288        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
290    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
291        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
293    8. When a pattern is compiled, it remembers the highest back reference so that
294       when matching, if the ovector is too small, extra memory can be obtained to
295       use instead. A conditional subpattern whose condition is a check on a
296       capture having happened, such as, for example in the pattern
297       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
298       setting the highest backreference number. This mattered only if pcre_exec()
299       was called with an ovector that was too small to hold the capture, and there
300       was no other kind of back reference (a situation which is probably quite
301       rare). The effect of the bug was that the condition was always treated as
302       FALSE when the capture could not be consulted, leading to a incorrect
303       behaviour by pcre_exec(). This bug has been fixed.
305    9. A reference to a duplicated named group (either a back reference or a test
306       for being set in a conditional) that occurred in a part of the pattern where
307       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
308       to be incorrectly calculated, leading to overwriting.
310    10. A mutually recursive set of back references such as (\2)(\1) caused a
311        segfault at study time (while trying to find the minimum matching length).
312        The infinite loop is now broken (with the minimum length unset, that is,
313        zero).
315    11. If an assertion that was used as a condition was quantified with a minimum
316        of zero, matching went wrong. In particular, if the whole group had
317        unlimited repetition and could match an empty string, a segfault was
318        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
319        assertions to be quantified, but not if they are being used as conditions,
320        so the above pattern is faulted by Perl. PCRE has now been changed so that
321        it also rejects such patterns.
323    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
324        failed to allow the zero-repeat case if pcre2_exec() was called with an
325        ovector too small to capture the group.
327    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
328        Red Hat Product Security:
330        (a) A crash if /K and /F were both set with the option to save the compiled
331        pattern.
333        (b) Another crash if the option to print captured substrings in a callout
334        was combined with setting a null ovector, for example \O\C+ as a subject
335        string.
337    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
338        forward reference repeated a large (but limited) number of times within a
339        repeated outer group that has a zero minimum quantifier, caused incorrect
340        code to be compiled, leading to the error "internal error:
341        previously-checked referenced subpattern not found" when an incorrect
342        memory address was read. This bug was reported as "heap overflow",
343        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
344        CVE-2015-2325.
346    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
347        call within a group that also contained a recursive back reference caused
348        incorrect code to be compiled. This bug was reported as "heap overflow",
349        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
350        number CVE-2015-2326.
352    24. Computing the size of the JIT read-only data in advance has been a source
353        of various issues, and new ones are still appear unfortunately. To fix
354        existing and future issues, size computation is eliminated from the code,
355        and replaced by on-demand memory allocation.
357    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
358        adjacent to the end of the range, and the range contained characters with
359        more than one other case, caused incorrect behaviour when compiled in UTF
360        mode. In that example, the range a-j was left out of the class.
362    26. Fix JIT compilation of conditional blocks, which assertion
363        is converted to (*FAIL). E.g: /(?(?!))/.
365    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
366        discovered by the LLVM fuzzer.
368    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
369        when this assertion was used as a condition, for example (?(?!)a|b). In
370        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
371        error about an unsupported item.
373    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
374        possessification code could take exponential time to complete. A recursion
375        depth limit of 1000 has been imposed to limit the resources used by this
376        optimization.
378    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
379        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
380        because \S ensures they are all in the class. The code for doing this was
381        interacting badly with the code for computing the amount of space needed to
382        compile the pattern, leading to a buffer overflow. This bug was discovered
383        by the LLVM fuzzer.
385    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
386        other kinds of group caused stack overflow at compile time. This bug was
387        discovered by the LLVM fuzzer.
389    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
390        between a subroutine call and its quantifier was incorrectly compiled,
391        leading to buffer overflow or other errors. This bug was discovered by the
392        LLVM fuzzer.
394    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
395        assertion after (?(. The code was failing to check the character after
396        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
397        was discovered by the LLVM fuzzer.
399    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
400        a fixed maximum following a group that contains a subroutine reference was
401        incorrectly compiled and could trigger buffer overflow. This bug was
402        discovered by the LLVM fuzzer.
404    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
405        caused a stack overflow instead of the diagnosis of a non-fixed length
406        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
408    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
409        (e.g. /(?<=\Ka)/) could make pcregrep loop.
411    37. There was a similar problem to 36 in pcretest for global matches.
413    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
414        and a subsequent item in the pattern caused a non-match, backtracking over
415        the repeated \X did not stop, but carried on past the start of the subject,
416        causing reference to random memory and/or a segfault. There were also some
417        other cases where backtracking after \C could crash. This set of bugs was
418        discovered by the LLVM fuzzer.
420    39. The function for finding the minimum length of a matching string could take
421        a very long time if mutual recursion was present many times in a pattern,
422        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
423        method has been implemented. This infelicity was discovered by the LLVM
424        fuzzer.
426    40. Static linking against the PCRE library using the pkg-config module was
427        failing on missing pthread symbols.
430    Version 8.36 26-September-2014
431    ------------------------------
433    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
434        -Wmissing-field-initializers and -Wunused-parameter.
436    2.  The tests for quantifiers being too big (greater than 65535) were being
437        applied after reading the number, and stupidly assuming that integer
438        overflow would give a negative number. The tests are now applied as the
439        numbers are read.
441    3.  Tidy code in pcre_exec.c where two branches that used to be different are
442        now the same.
444    4.  The JIT compiler did not generate match limit checks for certain
445        bracketed expressions with quantifiers. This may lead to exponential
446        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
447        issue should be resolved now.
449    5.  Fixed an issue, which occures when nested alternatives are optimized
450        with table jumps.
452    6.  Inserted two casts and changed some ints to size_t in the light of some
453        reported 64-bit compiler warnings (Bugzilla 1477).
455    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
456        an empty string, which sometimes were behaving incorrectly in the
457        interpreter (though correctly in the JIT matcher). This pcretest input is
458        an example:
460          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
463        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
464        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
465        for an empty string was breaking the inner loop and carrying on at a lower
466        level, when possessive repeated groups should always return to a higher
467        level as they have no backtrack points in them. The empty string test now
468        occurs at the outer level.
470    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
471        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
473    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
474        doesn't).
476    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
477        to hold a pattern to be incorrectly computed (too small) when there were
478        named back references to duplicated names. This could cause "internal
479        error: code overflow" or "double free or corruption" or other memory
480        handling errors.
482    11. When named subpatterns had the same prefixes, back references could be
483        confused. For example, in this pattern:
485          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
487        the reference to 'Name' was incorrectly treated as a reference to a
488        duplicate name.
490    12. A pattern such as /^s?c/mi8 where the optional character has more than
491        one "other case" was incorrectly compiled such that it would only try to
492        match starting at "c".
494    13. When a pattern starting with \s was studied, VT was not included in the
495        list of possible starting characters; this should have been part of the
496        8.34/18 patch.
498    14. If a character class started [\Qx]... where x is any character, the class
499        was incorrectly terminated at the ].
501    15. If a pattern that started with a caseless match for a character with more
502        than one "other case" was studied, PCRE did not set up the starting code
503        unit bit map for the list of possible characters. Now it does. This is an
504        optimization improvement, not a bug fix.
506    16. The Unicode data tables have been updated to Unicode 7.0.0.
508    17. Fixed a number of memory leaks in pcregrep.
510    18. Avoid a compiler warning (from some compilers) for a function call with
511        a cast that removes "const" from an lvalue by using an intermediate
512        variable (to which the compiler does not object).
514    19. Incorrect code was compiled if a group that contained an internal recursive
515        back reference was optional (had quantifier with a minimum of zero). This
516        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
517        caused segmentation faults because of stack overflows at compile time.
519    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
520        group that is quantified with an indefinite repeat, caused a compile-time
521        loop which used up all the system stack and provoked a segmentation fault.
522        This was not the same bug as 19 above.
524    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
525        Patch by Mike Frysinger.
528    Version 8.35 04-April-2014
529    --------------------------
531    1.  A new flag is set, when property checks are present in an XCLASS.
532        When this flag is not set, PCRE can perform certain optimizations
533        such as studying these XCLASS-es.
535    2.  The auto-possessification of character sets were improved: a normal
536        and an extended character set can be compared now. Furthermore
537        the JIT compiler optimizes more character set checks.
539    3.  Got rid of some compiler warnings for potentially uninitialized variables
540        that show up only when compiled with -O2.
542    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
543        of a match later then the end of the match. The pcretest program was not
544        handling the case sensibly - it was outputting from the start to the next
545        binary zero. It now reports this situation in a message, and outputs the
546        text from the end to the start.
548    5.  Fast forward search is improved in JIT. Instead of the first three
549        characters, any three characters with fixed position can be searched.
550        Search order: first, last, middle.
552    6.  Improve character range checks in JIT. Characters are read by an inprecise
553        function now, which returns with an unknown value if the character code is
554        above a certain threshold (e.g: 256). The only limitation is that the value
555        must be bigger than the threshold as well. This function is useful when
556        the characters above the threshold are handled in the same way.
558    7.  The macros whose names start with RAWUCHAR are placeholders for a future
559        mode in which only the bottom 21 bits of 32-bit data items are used. To
560        make this more memorable for those maintaining the code, the names have
561        been changed to start with UCHAR21, and an extensive comment has been added
562        to their definition.
564    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
565        to the export list in Makefile.am (they were accidentally omitted from the
566        8.34 tarball).
568    9.  The informational output from pcretest used the phrase "starting byte set"
569        which is inappropriate for the 16-bit and 32-bit libraries. As the output
570        for "first char" and "need char" really means "non-UTF-char", I've changed
571        "byte" to "char", and slightly reworded the output. The documentation about
572        these values has also been (I hope) clarified.
574    10. Another JIT related optimization: use table jumps for selecting the correct
575        backtracking path, when more than four alternatives are present inside a
576        bracket.
578    11. Empty match is not possible, when the minimum length is greater than zero,
579        and there is no \K in the pattern. JIT should avoid empty match checks in
580        such cases.
582    12. In a caseless character class with UCP support, when a character with more
583        than one alternative case was not the first character of a range, not all
584        the alternative cases were added to the class. For example, s and \x{17f}
585        are both alternative cases for S: the class [RST] was handled correctly,
586        but [R-T] was not.
588    13. The configure.ac file always checked for pthread support when JIT was
589        enabled. This is not used in Windows, so I have put this test inside a
590        check for the presence of windows.h (which was already tested for).
592    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
593        The algorithm provides a way to skip certain starting offsets, and usually
594        faster than linear prefix searches.
596    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
597        as for 'fr_FR' and 'french'. For some reason, however, it then used the
598        Windows-specific input and output files, which have 'french' screwed in.
599        So this could never have worked. One of the problems with locales is that
600        they aren't always the same. I have now updated RunTest so that it checks
601        the output of the locale test (test 3) against three different output
602        files, and it allows the test to pass if any one of them matches. With luck
603        this should make the test pass on some versions of Solaris where it was
604        failing. Because of the uncertainty, the script did not used to stop if
605        test 3 failed; it now does. If further versions of a French locale ever
606        come to light, they can now easily be added.
608    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
609        there was a message during ./configure, but it did not stop. This now
610        provokes an error. The invalid example in README has been corrected.
611        If a value less than the minimum is given, the minimum value has always
612        been used, but now a warning is given.
614    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
615        was a bug in the test system, which is now fixed. Also, the list of various
616        configurations that are tested for each release did not have one with both
617        16/32 bits and --enable-bar-anycrlf. It now does.
619    18. pcretest was missing "-C bsr" for displaying the \R default setting.
621    19. Little endian PowerPC systems are supported now by the JIT compiler.
623    20. The fast forward newline mechanism could enter to an infinite loop on
624        certain invalid UTF-8 input. Although we don't support these cases
625        this issue can be fixed by a performance optimization.
627    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
628        not take account if existing stack usage. There is now a new global
629        variable called pcre_stack_guard that can be set to point to an external
630        function to check stack availability. It is called at the start of
631        processing every parenthesized group.
633    22. A typo in the code meant that in ungreedy mode the max/min qualifier
634        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
635        not match "ab".
637    23. When UTF was disabled, the JIT program reported some incorrect compile
638        errors. These messages are silenced now.
640    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
641        compiler.
643    25. Change all the temporary files used in RunGrepTest to be different to those
644        used by RunTest so that the tests can be run simultaneously, for example by
645        "make -j check".
648    Version 8.34 15-December-2013
649    -----------------------------
651    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
652        executable memory. Patch inspired by Carsten Klein.
654    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
655        this macro is never tested and has no effect, because the work to support
656        coverage involves only compiling and linking options and special targets in
657        the Makefile. The comment in config.h implied that defining the macro would
658        enable coverage support, which is totally false. There was also support for
659        setting this macro in the CMake files (my fault, I just copied it from
660        configure). SUPPORT_GCOV has now been removed.
662    3.  Make a small performance improvement in strlen16() and strlen32() in
663        pcretest.
665    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
666        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
667        these cases). There was also one in pcretest.c.
669    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
671    6.  In UTF mode, the code for checking whether a group could match an empty
672        string (which is used for indefinitely repeated groups to allow for
673        breaking an infinite loop) was broken when the group contained a repeated
674        negated single-character class with a character that occupied more than one
675        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
676        UTF-8 mode). The effect was undefined: the group might or might not be
677        deemed as matching an empty string, or the program might have crashed.
679    7.  The code for checking whether a group could match an empty string was not
680        recognizing that \h, \H, \v, \V, and \R must match a character.
682    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
683        an empty string. If it can, pcretest shows this in its information output.
685    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
686        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
687        matched by pcre_exec() without using JIT:
689        (a) If the rest of the pattern did not match after a maximal run of
690            grapheme clusters, the code for backing up to try with fewer of them
691            did not always back up over a full grapheme when characters that do not
692            have the modifier quality were involved, e.g. Hangul syllables.
694        (b) If the match point in a subject started with modifier character, and
695            there was no match, the code could incorrectly back up beyond the match
696            point, and potentially beyond the first character in the subject,
697            leading to a segfault or an incorrect match result.
699    10. A conditional group with an assertion condition could lead to PCRE
700        recording an incorrect first data item for a match if no other first data
701        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
702        first data item, and therefore matched "ca" after "c" instead of at the
703        start.
705    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
706        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
708    12. The source of pcregrep now includes z/OS-specific code so that it can be
709        compiled for z/OS as part of the special z/OS distribution.
711    13. Added the -T and -TM options to pcretest.
713    14. The code in pcre_compile.c for creating the table of named capturing groups
714        has been refactored. Instead of creating the table dynamically during the
715        actual compiling pass, the information is remembered during the pre-compile
716        pass (on the stack unless there are more than 20 named groups, in which
717        case malloc() is used) and the whole table is created before the actual
718        compile happens. This has simplified the code (it is now nearly 150 lines
719        shorter) and prepared the way for better handling of references to groups
720        with duplicate names.
722    15. A back reference to a named subpattern when there is more than one of the
723        same name now checks them in the order in which they appear in the pattern.
724        The first one that is set is used for the reference. Previously only the
725        first one was inspected. This change makes PCRE more compatible with Perl.
727    16. Unicode character properties were updated from Unicode 6.3.0.
729    17. The compile-time code for auto-possessification has been refactored, based
730        on a patch by Zoltan Herczeg. It now happens after instead of during
731        compilation. The code is cleaner, and more cases are handled. The option
732        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
733        options in pcretest are provided to set it. It can also be set by
734        (*NO_AUTO_POSSESS) at the start of a pattern.
736    18. The character VT has been added to the default ("C" locale) set of
737        characters that match \s and are generally treated as white space,
738        following this same change in Perl 5.18. There is now no difference between
739        "Perl space" and "POSIX space". Whether VT is treated as white space in
740        other locales depends on the locale.
742    19. The code for checking named groups as conditions, either for being set or
743        for being recursed, has been refactored (this is related to 14 and 15
744        above). Processing unduplicated named groups should now be as fast at
745        numerical groups, and processing duplicated groups should be faster than
746        before.
748    20. Two patches to the CMake build system, by Alexander Barkov:
750          (1) Replace the "source" command by "." in CMakeLists.txt because
751              "source" is a bash-ism.
753          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
754              without these the CMake build does not work on Solaris.
756    21. Perl has changed its handling of \8 and \9. If there is no previously
757        encountered capturing group of those numbers, they are treated as the
758        literal characters 8 and 9 instead of a binary zero followed by the
759        literals. PCRE now does the same.
761    22. Following Perl, added \o{} to specify codepoints in octal, making it
762        possible to specify values greater than 0777 and also making them
763        unambiguous.
765    23. Perl now gives an error for missing closing braces after \x{... instead of
766        treating the string as literal. PCRE now does the same.
768    24. RunTest used to grumble if an inappropriate test was selected explicitly,
769        but just skip it when running all tests. This make it awkward to run ranges
770        of tests when one of them was inappropriate. Now it just skips any
771        inappropriate tests, as it always did when running all tests.
773    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
774        character types such as \d or \w, too many callouts were inserted, and the
775        data that they returned was rubbish.
777    26. In UCP mode, \s was not matching two of the characters that Perl matches,
778        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
779        were matched by \h. The code has now been refactored so that the lists of
780        the horizontal and vertical whitespace characters used for \h and \v (which
781        are defined only in one place) are now also used for \s.
783    27. Add JIT support for the 64 bit TileGX architecture.
784        Patch by Jiong Wang (Tilera Corporation).
786    28. Possessive quantifiers for classes (both explicit and automatically
787        generated) now use special opcodes instead of wrapping in ONCE brackets.
789    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
790        (because it's meaningless), this was not happening when PCRE_CASELESS was
791        set. Not wrong, but inefficient.
793    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
794        properties for \w, \d, etc) is present in a test regex. Otherwise if the
795        test contains no characters greater than 255, Perl doesn't realise it
796        should be using Unicode semantics.
798    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
799        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
800        does in Unicode mode.
802    32. Added the "forbid" facility to pcretest so that putting tests into the
803        wrong test files can sometimes be quickly detected.
805    33. There is now a limit (default 250) on the depth of nesting of parentheses.
806        This limit is imposed to control the amount of system stack used at compile
807        time. It can be changed at build time by --with-parens-nest-limit=xxx or
808        the equivalent in CMake.
810    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
811        errors. Perl warns for these when in warning mode, but PCRE has no facility
812        for giving warnings.
814    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
815        However, this was not working for (?!) because it is optimized to (*FAIL),
816        for which PCRE does not allow quantifiers. The optimization is now disabled
817        when a quantifier follows (?!). I can't see any use for this, but it makes
818        things uniform.
820    36. Perl no longer allows group names to start with digits, so I have made this
821        change also in PCRE. It simplifies the code a bit.
823    37. In extended mode, Perl ignores spaces before a + that indicates a
824        possessive quantifier. PCRE allowed a space before the quantifier, but not
825        before the possessive +. It now does.
827    38. The use of \K (reset reported match start) within a repeated possessive
828        group such as (a\Kb)*+ was not working.
830    40. Document that the same character tables must be used at compile time and
831        run time, and that the facility to pass tables to pcre_exec() and
832        pcre_dfa_exec() is for use only with saved/restored patterns.
834    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
835        features for Builds with MSVC:
837        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
838           builds). The .rc files can be used to set FileDescription and many other
839           attributes.
841        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
842           This allows higher-level build scripts which want .pdb files to avoid
843           hard-coding the exact files needed."
845    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
846        mean "start of word" and "end of word", respectively, as a transition aid.
848    43. A minimizing repeat of a class containing codepoints greater than 255 in
849        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
850        compiled to use the heap for recursion.
852    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
853        is configured.
856    Version 8.33 28-May-2013
857    ------------------------
859    1.  Added 'U' to some constants that are compared to unsigned integers, to
860        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
861        variables that are added to signed variables, to ensure the result is
862        signed and can be negated.
864    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
865        CMake config files.
867    3.  Revise the creation of config.h.generic so that all boolean macros are
868        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
869        overriding via -D on the command line possible.
871    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
872        to unsigned int is reported to make a quite noticeable speed difference in
873        a specific Windows environment. Testing on Linux did also appear to show
874        some benefit (and it is clearly not harmful). Also fixed the definition of
875        Xop which should be unsigned.
877    5.  Related to (4), changing the definition of the intermediate variable cc
878        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
879        improvements.
881    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
882        unnecessary spaces.
884    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
885        and later.
887    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
889    9.  Optimizing fast_forward_start_bits in JIT.
891    10. Adding support for callouts in JIT, and fixing some issues revealed
892        during this work. Namely:
894        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
896        (b) Minimum length was not checked before the matching is started.
898    11. The value of capture_last that is passed to callouts was incorrect in some
899        cases when there was a capture on one path that was subsequently abandoned
900        after a backtrack. Also, the capture_last value is now reset after a
901        recursion, since all captures are also reset in this case.
903    12. The interpreter no longer returns the "too many substrings" error in the
904        case when an overflowing capture is in a branch that is subsequently
905        abandoned after a backtrack.
907    13. In the pathological case when an offset vector of size 2 is used, pcretest
908        now prints out the matched string after a yield of 0 or 1.
910    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
911        Only supported by the JIT compiler at the moment.
913    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
915    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
916        offset of the starting point of the matching process, provided the offsets
917        vector is large enough.
919    17. The \A escape now records a lookbehind value of 1, though its execution
920        does not actually inspect the previous character. This is to ensure that,
921        in partial multi-segment matching, at least one character from the old
922        segment is retained when a new segment is processed. Otherwise, if there
923        are no lookbehinds in the pattern, \A might match incorrectly at the start
924        of a new segment.
926    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
928    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
929        gives some modest performance improvement in 8-bit mode.
931    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
932        be expressed in certain programming languages using Universal Character
933        Names.
935    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
936        which points out that "non characters" are not "characters that may not
937        appear in Unicode strings" but rather "characters that are reserved for
938        internal use and have only local meaning".
940    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
941        there was a conditional group that depended on an assertion, if the
942        assertion was false, the callout that immediately followed the alternation
943        in the condition was skipped when pcre_exec() was used for matching.
945    23. Allow an explicit callout to be inserted before an assertion that is the
946        condition for a conditional group, for compatibility with automatic
947        callouts, which always insert a callout at this point.
949    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
950        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
952    25. (*PRUNE) is now supported by the JIT compiler.
954    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
956    27. Fix the case where there are two or more SKIPs with arguments that may be
957        ignored.
959    28. (*SKIP) is now supported by the JIT compiler.
961    29. (*THEN) is now supported by the JIT compiler.
963    30. Update RunTest with additional test selector options.
965    31. The way PCRE handles backtracking verbs has been changed in two ways.
967        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
968        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
969        backtracking. In some cases this makes it more Perl-compatible, but Perl's
970        rather obscure rules do not always do the same thing.
972        (2) Previously, backtracking verbs were confined within assertions. This is
973        no longer the case for positive assertions, except for (*ACCEPT). Again,
974        this sometimes improves Perl compatibility, and sometimes does not.
976    32. A number of tests that were in test 2 because Perl did things differently
977        have been moved to test 1, because either Perl or PCRE has changed, and
978        these tests are now compatible.
980    32. Backtracking control verbs are now handled in the same way in JIT and
981        interpreter.
983    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
984        contained a forward subroutine reference caused a compile error.
986    34. Auto-detect and optimize limited repetitions in JIT.
988    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
989        blocking (*UTF) etc.
991    36. In the interpreter, maximizing pattern repetitions for characters and
992        character types now use tail recursion, which reduces stack usage.
994    37. The value of the max lookbehind was not correctly preserved if a compiled
995        and saved regex was reloaded on a host of different endianness.
997    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
998        of the compiled pattern block, expand the flags field from 16 to 32 bits
999        because it was almost full.
1001    39. Try madvise first before posix_madvise.
1003    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1004        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1005        original change locked out all matches of empty strings. This has been
1006        changed so that one match of an empty string per line is recognized.
1007        Subsequent searches on the same line (for colouring or for --only-matching,
1008        for example) do not recognize empty strings.
1010    41. Applied a user patch to fix a number of spelling mistakes in comments.
1012    42. Data lines longer than 65536 caused pcretest to crash.
1014    43. Clarified the data type for length and startoffset arguments for pcre_exec
1015        and pcre_dfa_exec in the function-specific man pages, where they were
1016        explicitly stated to be in bytes, never having been updated. I also added
1017        some clarification to the pcreapi man page.
1019    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1020        a segmentation fault.
1023    Version 8.32 30-November-2012
1024    -----------------------------
1026    1.  Improved JIT compiler optimizations for first character search and single
1027        character iterators.
1029    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1030        Patch by Daniel Richard G.
1032    3.  Single character iterator optimizations in the JIT compiler.
1034    4.  Improved JIT compiler optimizations for character ranges.
1036    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1037        Reported by Giuseppe D'Angelo.
1039    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1040        of a line, was being set incorrectly in cases where .* appeared inside
1041        atomic brackets at the start of a pattern, or where there was a subsequent
1042        *PRUNE or *SKIP.
1044    7.  Improved instruction cache flush for POWER/PowerPC.
1045        Patch by Daniel Richard G.
1047    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1048        grep:
1050        (a) There is now no limit to the number of patterns to be matched.
1052        (b) An error is given if a pattern is too long.
1054        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1055            are now supported.
1057        (d) --exclude-from and --include-from (multiple use) have been added.
1059        (e) Exclusions and inclusions now apply to all files and directories, not
1060            just to those obtained from scanning a directory recursively.
1062        (f) Multiple uses of -f and --file-list are now supported.
1064        (g) In a Windows environment, the default for -d has been changed from
1065            "read" (the GNU grep default) to "skip", because otherwise the presence
1066            of a directory in the file list provokes an error.
1068        (h) The documentation has been revised and clarified in places.
1070    9.  Improve the matching speed of capturing brackets.
1072    10. Changed the meaning of \X so that it now matches a Unicode extended
1073        grapheme cluster.
1075    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1076        out POSIX threads when JIT support is configured.
1078    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1080    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1081        structure could go wrong in environments where size_t is not the same size
1082        as int.
1084    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1085        set.
1087    15. The EBCDIC support had decayed; later updates to the code had included
1088        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1089        general tidy up of EBCDIC-related issues, and the documentation was also
1090        not quite right. There is now a test that can be run on ASCII systems to
1091        check some of the EBCDIC-related things (but is it not a full test).
1093    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1094        in a small tidy to the code.
1096    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1098    18. If the --only-matching (-o) option in pcregrep is specified multiple
1099        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1100        substrings matched by the 1st and 2nd capturing parentheses. A separating
1101        string can be specified by --om-separator (default empty).
1103    19. Improving the first n character searches.
1105    20. Turn case lists for horizontal and vertical white space into macros so that
1106        they are defined only once.
1108    21. This set of changes together give more compatible Unicode case-folding
1109        behaviour for characters that have more than one other case when UCP
1110        support is available.
1112        (a) The Unicode property table now has offsets into a new table of sets of
1113            three or more characters that are case-equivalent. The MultiStage2.py
1114            script that generates these tables (the pcre_ucd.c file) now scans
1115            CaseFolding.txt instead of UnicodeData.txt for character case
1116            information.
1118        (b) The code for adding characters or ranges of characters to a character
1119            class has been abstracted into a generalized function that also handles
1120            case-independence. In UTF-mode with UCP support, this uses the new data
1121            to handle characters with more than one other case.
1123        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1124            whose other case is greater than 256 are now correctly matched
1125            caselessly. Previously, the high codepoint matched the low one, but not
1126            vice versa.
1128        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1129            of the new class addition function, using character lists defined as
1130            macros alongside the case definitions of 20 above.
1132        (e) Caseless back references now work with characters that have more than
1133            one other case.
1135        (f) General caseless matching of characters with more than one other case
1136            is supported.
1138    22. Unicode character properties were updated from Unicode 6.2.0
1140    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1142    24. Add support for 32-bit character strings, and UTF-32
1144    25. Major JIT compiler update (code refactoring and bugfixing).
1145        Experimental Sparc 32 support is added.
1147    26. Applied a modified version of Daniel Richard G's patch to create
1148        pcre.h.generic and config.h.generic by "make" instead of in the
1149        PrepareRelease script.
1151    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1152        pcre_compile.c when checking for a zero character.
1154    28. Introducing a native interface for JIT. Through this interface, the compiled
1155        machine code can be directly executed. The purpose of this interface is to
1156        provide fast pattern matching, so several sanity checks are not performed.
1157        However, feature tests are still performed. The new interface provides
1158        1.4x speedup compared to the old one.
1160    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1161        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1162        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1164    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1165        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1166        the "old" RFC 2279). Instead, it ended up passing a negative length to
1167        pcre_exec().
1169    31. Add support for GCC's visibility feature to hide internal functions.
1171    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1172        "unknown -C option" after outputting 0 or 1.
1174    33. There is now support for generating a code coverage report for the test
1175        suite in environments where gcc is the compiler and lcov is installed. This
1176        is mainly for the benefit of the developers.
1178    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1179        unaddressable using valgrind annotations, allowing valgrind to detect
1180        invalid memory accesses. This is mainly for the benefit of the developers.
1182    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1184    26. Give configure error if --enable-cpp but no C++ compiler found.
1187    Version 8.31 06-July-2012
1188    -------------------------
1190    1.  Fixing a wrong JIT test case and some compiler warnings.
1192    2.  Removed a bashism from the RunTest script.
1194    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1195        to unsigned type, result still unsigned" that was given by an MS compiler
1196        on encountering the code "-sizeof(xxx)".
1198    4.  Partial matching support is added to the JIT compiler.
1200    5.  Fixed several bugs concerned with partial matching of items that consist
1201        of more than one character:
1203        (a) /^(..)\1/ did not partially match "aba" because checking references was
1204            done on an "all or nothing" basis. This also applied to repeated
1205            references.
1207        (b) \R did not give a hard partial match if \r was found at the end of the
1208            subject.
1210        (c) \X did not give a hard partial match after matching one or more
1211            characters at the end of the subject.
1213        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1214            a partial match for the string "\r".
1216        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1217            a partial match for a CR character at the end of the subject string.
1219    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1220        running pcretest, the text "(JIT)" added to the output whenever JIT is
1221        actually used to run the match.
1223    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1224        or /S+[+] with a digit between 1 and 7.
1226    8.  OP_NOT now supports any UTF character not just single-byte ones.
1228    9.  (*MARK) control verb is now supported by the JIT compiler.
1230    10. The command "./RunTest list" lists the available tests without actually
1231        running any of them. (Because I keep forgetting what they all are.)
1235    12. Applied a (slightly modified) user-supplied patch that improves performance
1236        when the heap is used for recursion (compiled with --disable-stack-for-
1237        recursion). Instead of malloc and free for each heap frame each time a
1238        logical recursion happens, frames are retained on a chain and re-used where
1239        possible. This sometimes gives as much as 30% improvement.
1241    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1242        call.
1244    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1246    15. It is now possible to link pcretest with libedit as an alternative to
1247        libreadline.
1249    16. (*COMMIT) control verb is now supported by the JIT compiler.
1251    17. The Unicode data tables have been updated to Unicode 6.1.0.
1253    18. Added --file-list option to pcregrep.
1255    19. Added binary file support to pcregrep, including the -a, --binary-files,
1256        -I, and --text options.
1258    20. The madvise function is renamed for posix_madvise for QNX compatibility
1259        reasons. Fixed by Giuseppe D'Angelo.
1261    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1263    22. Changed the option for creating symbolic links for 16-bit man pages from
1264        -s to -sf so that re-installing does not cause issues.
1266    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1268    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1269        matches in certain environments (the workspace was not being correctly
1270        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1271        some of the workspace data at the beginning of a restart.
1273    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1274        was not doing so when it should - probably a typo introduced by SVN 528
1275        (change 8.10/14).
1277    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1278        \w+ when the character tables indicated that \x{c4} was a word character.
1279        There were several related cases, all because the tests for doing a table
1280        lookup were testing for characters less than 127 instead of 255.
1282    27. If a pattern contains capturing parentheses that are not used in a match,
1283        their slots in the ovector are set to -1. For those that are higher than
1284        any matched groups, this happens at the end of processing. In the case when
1285        there were back references that the ovector was too small to contain
1286        (causing temporary malloc'd memory to be used during matching), and the
1287        highest capturing number was not used, memory off the end of the ovector
1288        was incorrectly being set to -1. (It was using the size of the temporary
1289        memory instead of the true size.)
1291    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1292        ovector size, it uses memory at the end of the block that it has got.
1294    29. Check for an overlong MARK name and give an error at compile time. The
1295        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1297    30. JIT compiler update.
1299    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1300        Rill for the patch.
1302    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1304    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1306    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1307        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1308        was enabled.
1310    35. Improve JIT code generation for greedy plus quantifier.
1312    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1313        1 to "aa" instead of to an empty string. The bug affected repeated groups
1314        that could potentially match an empty string.
1316    37. Optimizing single character iterators in JIT.
1318    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1319        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1320        codepoints that are too big for the mode are faulted, and in a UTF mode,
1321        disallowed codepoints are also faulted.
1323    39. If PCRE was compiled with UTF support, in three places in the DFA
1324        matcher there was code that should only have been obeyed in UTF mode, but
1325        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1326        processing when bytes with values greater than 127 were present. In 16-bit
1327        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1328        both cases the values are those that cannot be the first data item in a UTF
1329        character. The three items that might have provoked this were recursions,
1330        possessively repeated groups, and atomic groups.
1332    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1333        and pcregrep, because some OS require shared objects to be explicitly
1334        passed to ld, causing the link step to fail if they are not.
1336    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1337        mode, patterns that started with \h* or \R* might be incorrectly matched.
1340    Version 8.30 04-February-2012
1341    -----------------------------
1343    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1344        name is defined in ctype.h.
1346    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1347        only in quite long subpatterns.
1349    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1350        since it was replaced by pcre_fullinfo() in February 2000.
1352    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1353        match a (*MARK), and the match failed at the start of the subject, a
1354        reference to memory before the start of the subject could occur. This bug
1355        was introduced by fix 17 of release 8.21.
1357    5.  A reference to an unset group with zero minimum repetition was giving
1358        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1359        /(another)?(\1?)test/ matched against "hello world test". This bug was
1360        introduced in release 8.13.
1362    6.  Add support for 16-bit character strings (a large amount of work involving
1363        many changes and refactorings).
1365    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1366        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1367        from a file.
1369    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1370        rounding is not applied in this particular case).
1372    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1373        if they appear, or are escaped, in patterns.
1375    10. Get rid of a number of -Wunused-but-set-variable warnings.
1377    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1378        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1379        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1380        also returns the mark "x". This bug applied to capturing parentheses,
1381        non-capturing parentheses, and atomic parentheses. It also applied to some
1382        assertions.
1384    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1385        information out of configure.ac instead of relying on pcre.h.generic, which
1386        is not stored in the repository.
1388    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1389        -lreadline.
1391    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1393    15. Applied Graycode's patch to put the top-level frame on the stack rather
1394        than the heap when not using the stack for recursion. This gives a
1395        performance improvement in many cases when recursion is not deep.
1397    16. Experimental code added to "pcretest -C" to output the stack frame size.
1400    Version 8.21 12-Dec-2011
1401    ------------------------
1403    1.  Updating the JIT compiler.
1405    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1406        are added as well.
1408    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1409        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1410        calling _pcre_jit_exec. Some extra comments are added.
1412    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1413        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1414        was introduced by change 18 for 8.20.
1416    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1417        ECMA-262 standard.
1419    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1420        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1421        This bug was probably introduced by change 9 of 8.13.
1423    7.  While fixing 6 above, I noticed that a number of other items were being
1424        incorrectly rejected as "not fixed length". This arose partly because newer
1425        opcodes had not been added to the fixed-length checking code. I have (a)
1426        corrected the bug and added tests for these items, and (b) arranged for an
1427        error to occur if an unknown opcode is encountered while checking for fixed
1428        length instead of just assuming "not fixed length". The items that were
1429        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1430        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1431        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1433    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1434        being incorrectly compiled and would have given unpredicatble results.
1436    9.  A possessively repeated subpattern with minimum repeat count greater than
1437        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1438        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1439        the first (A) could occur when it should not.
1441    10. Add a cast and remove a redundant test from the code.
1443    11. JIT should use pcre_malloc/pcre_free for allocation.
1445    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1446        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1447        is anything other than /usr, -L is still shown).
1449    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1451    14. Perl does not support \N without a following name in a [] class; PCRE now
1452        also gives an error.
1454    15. If a forward reference was repeated with an upper limit of around 2000,
1455        it caused the error "internal error: overran compiling workspace". The
1456        maximum number of forward references (including repeats) was limited by the
1457        internal workspace, and dependent on the LINK_SIZE. The code has been
1458        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1459        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1460        of around 200,000 forward references. While doing this, I also speeded up
1461        the filling in of repeated forward references.
1463    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1464        incorrectly expecting the subject to contain another "a" after the start.
1466    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1467        in the match, the SKIP should be ignored. This was not happening; instead
1468        the SKIP was being treated as NOMATCH. For patterns such as
1469        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1470        tested.
1472    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1473        now much more compatible with Perl, in particular in cases where the result
1474        is a non-match for a non-anchored pattern. For example, if
1475        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1476        "m", where previously it did not return a name. A side effect of this
1477        change is that for partial matches, the last encountered mark name is
1478        returned, as for non matches. A number of tests that were previously not
1479        Perl-compatible have been moved into the Perl-compatible test files. The
1480        refactoring has had the pleasing side effect of removing one argument from
1481        the match() function, thus reducing its stack requirements.
1483    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1484        subsequent uses of /S (without +) incorrectly behaved like /S+.
1486    21. Retrieve executable code size support for the JIT compiler and fixing
1487        some warnings.
1489    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1490        not work when the shorter character appeared right at the end of the
1491        subject string.
1493    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1494        systems.
1496    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1497        output it when the /M option is used in pcretest.
1499    25. The CheckMan script was not being included in the distribution. Also, added
1500        an explicit "perl" to run Perl scripts from the PrepareRelease script
1501        because this is reportedly needed in Windows.
1503    26. If study data was being save in a file and studying had not found a set of
1504        "starts with" bytes for the pattern, the data written to the file (though
1505        never used) was taken from uninitialized memory and so caused valgrind to
1506        complain.
1508    27. Updated RunTest.bat as provided by Sheri Pierce.
1510    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1512    29. Computation of memory usage for the table of capturing group names was
1513        giving an unnecessarily large value.
1516    Version 8.20 21-Oct-2011
1517    ------------------------
1519    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1520        a POSIX class. After further experiments with Perl, which convinced me that
1521        Perl has bugs and confusions, a closing square bracket is no longer allowed
1522        in a POSIX name. This bug also affected patterns with classes that started
1523        with full stops.
1525    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1526        captured substring, but while checking the failing first alternative,
1527        substring 1 is temporarily captured. If the output vector supplied to
1528        pcre_exec() was not big enough for this capture, the yield of the function
1529        was still zero ("insufficient space for captured substrings"). This cannot
1530        be totally fixed without adding another stack variable, which seems a lot
1531        of expense for a edge case. However, I have improved the situation in cases
1532        such as /(a)(b)x|abc/ matched against "abc", where the return code
1533        indicates that fewer than the maximum number of slots in the ovector have
1534        been set.
1536    3.  Related to (2) above: when there are more back references in a pattern than
1537        slots in the output vector, pcre_exec() uses temporary memory during
1538        matching, and copies in the captures as far as possible afterwards. It was
1539        using the entire output vector, but this conflicts with the specification
1540        that only 2/3 is used for passing back captured substrings. Now it uses
1541        only the first 2/3, for compatibility. This is, of course, another edge
1542        case.
1544    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1545        main code base, and can be used by building with --enable-jit. When this is
1546        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1547        runtime --no-jit option is given.
1549    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1550        ovector, the return from the function was zero, implying that there were
1551        other matches that did not fit. The correct "exactly full" value is now
1552        returned.
1554    6.  If a subpattern that was called recursively or as a subroutine contained
1555        (*PRUNE) or any other control that caused it to give a non-standard return,
1556        invalid errors such as "Error -26 (nested recursion at the same subject
1557        position)" or even infinite loops could occur.
1559    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1560        computing the minimum length on reaching *ACCEPT, and so ended up with the
1561        wrong value of 1 rather than 0. Further investigation indicates that
1562        computing a minimum subject length in the presence of *ACCEPT is difficult
1563        (think back references, subroutine calls), and so I have changed the code
1564        so that no minimum is registered for a pattern that contains *ACCEPT.
1566    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1567        it was not handled as intended. [But see 16 below.]
1569    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1570        Sheri Pierce.
1572    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1573        the first byte in a match must be "a".
1575    11. Change 17 for 8.13 increased the recursion depth for patterns like
1576        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1577        pattern contains any instances of (*THEN). If it does not, the old
1578        optimizations are restored. It would be nice to do this on a per-group
1579        basis, but at the moment that is not feasible.
1581    12. In some environments, the output of pcretest -C is CRLF terminated. This
1582        broke RunTest's code that checks for the link size. A single white space
1583        character after the value is now allowed for.
1585    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1586        For "fr", it uses the Windows-specific input and output files.
1588    14. If (*THEN) appeared in a group that was called recursively or as a
1589        subroutine, it did not work as intended. [But see next item.]
1591    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1592        pattern fragments (but not containing any | characters). If A and B are
1593        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1594        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1595        D. In other words, Perl considers parentheses that do not contain any |
1596        characters to be part of a surrounding alternative, whereas PCRE was
1597        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1598        differently. PCRE now behaves in the same way as Perl, except in the case
1599        of subroutine/recursion calls such as (?1) which have in any case always
1600        been different (but PCRE had them first :-).
1602    16. Related to 15 above: Perl does not treat the | in a conditional group as
1603        creating alternatives. Such a group is treated in the same way as an
1604        ordinary group without any | characters when processing (*THEN). PCRE has
1605        been changed to match Perl's behaviour.
1607    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1608        RunGrepTest script failed.
1610    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1611        inevitable for groups that contain captures, but it can lead to a lot of
1612        stack use in large patterns. The old behaviour has been restored for atomic
1613        groups that do not contain any capturing parentheses.
1615    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1616        suppress the check for a minimum subject length at run time. (If it was
1617        given to pcre_exec() or pcre_dfa_exec() it did work.)
1619    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1620        fail to work when decoding hex characters in data strings in EBCDIC
1621        environments.
1623    21. It appears that in at least one Mac OS environment, the isxdigit() function
1624        is implemented as a macro that evaluates to its argument more than once,
1625        contravening the C 90 Standard (I haven't checked a later standard). There
1626        was an instance in pcretest which caused it to go wrong when processing
1627        \x{...} escapes in subject strings. The has been rewritten to avoid using
1628        things like p++ in the argument of isxdigit().
1631    Version 8.13 16-Aug-2011
1632    ------------------------
1634    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1636    2.  Two minor typos in pcre_internal.h have been fixed.
1638    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1639        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1640        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1642    4.  There were a number of related bugs in the code for matching backrefences
1643        caselessly in UTF-8 mode when codes for the characters concerned were
1644        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1645        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1646        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1647        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1648        2-byte code at the end of the subject (it thought there wasn't enough data
1649        left).
1651    5.  Comprehensive information about what went wrong is now returned by
1652        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1653        as the output vector has at least 2 elements. The offset of the start of
1654        the failing character and a reason code are placed in the vector.
1656    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1657        now returned is for the first byte of the failing character, instead of the
1658        last byte inspected. This is an incompatible change, but I hope it is small
1659        enough not to be a problem. It makes the returned offset consistent with
1660        pcre_exec() and pcre_dfa_exec().
1662    7.  pcretest now gives a text phrase as well as the error number when
1663        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1664        failure, the offset and reason code are output.
1666    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1667        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1668        back over a single character (\n). This seems wrong (because it treated the
1669        two characters as a single entity when going forwards), conflicts with the
1670        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1671        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1672        has been changed.
1674    9.  Some internal refactoring has changed the processing so that the handling
1675        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1676        time (the PCRE_DOTALL option was changed this way some time ago: version
1677        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1678        which was always a bit of a fudge. It also means that there is one less
1679        argument for the match() function, which reduces its stack requirements
1680        slightly. This change also fixes an incompatibility with Perl: the pattern
1681        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1683    10. More internal refactoring has drastically reduced the number of recursive
1684        calls to match() for possessively repeated groups such as (abc)++ when
1685        using pcre_exec().
1687    11. While implementing 10, a number of bugs in the handling of groups were
1688        discovered and fixed:
1690        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1691        (a|)*(?1) gave a compile-time internal error.
1692        ((a|)+)+  did not notice that the outer group could match an empty string.
1693        (^a|^)+   was not marked as anchored.
1694        (.*a|.*)+ was not marked as matching at start or after a newline.
1696    12. Yet more internal refactoring has removed another argument from the match()
1697        function. Special calls to this function are now indicated by setting a
1698        value in a variable in the "match data" data block.
1700    13. Be more explicit in pcre_study() instead of relying on "default" for
1701        opcodes that mean there is no starting character; this means that when new
1702        ones are added and accidentally left out of pcre_study(), testing should
1703        pick them up.
1705    14. The -s option of pcretest has been documented for ages as being an old
1706        synonym of -m (show memory usage). I have changed it to mean "force study
1707        for every regex", that is, assume /S for every regex. This is similar to -i
1708        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1709        using it. It makes it easier to run collections of tests with and without
1710        study enabled, and thereby test pcre_study() more easily. All the standard
1711        tests are now run with and without -s (but some patterns can be marked as
1712        "never study" - see 20 below).
1714    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1715        restoration of the capturing data to the outer values was not happening
1716        correctly.
1718    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1719        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1720        pattern had matched an empty string, and so incorrectly returned a no
1721        match.
1723    17. There was optimizing code for the last branch of non-capturing parentheses,
1724        and also for the obeyed branch of a conditional subexpression, which used
1725        tail recursion to cut down on stack usage. Unfortunately, now that there is
1726        the possibility of (*THEN) occurring in these branches, tail recursion is
1727        no longer possible because the return has to be checked for (*THEN). These
1728        two optimizations have therefore been removed. [But see 8.20/11 above.]
1730    18. If a pattern containing \R was studied, it was assumed that \R always
1731        matched two bytes, thus causing the minimum subject length to be
1732        incorrectly computed because \R can also match just one byte.
1734    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1735        was incorrectly computed.
1737    20. If /S is present twice on a test pattern in pcretest input, it now
1738        *disables* studying, thereby overriding the use of -s on the command line
1739        (see 14 above). This is necessary for one or two tests to keep the output
1740        identical in both cases.
1742    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1743        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1745    22. When an atomic group that contained a capturing parenthesis was
1746        successfully matched, but the branch in which it appeared failed, the
1747        capturing was not being forgotten if a higher numbered group was later
1748        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1749        group 1 to "a", when in fact it should be unset. This applied to multi-
1750        branched capturing and non-capturing groups, repeated or not, and also to
1751        positive assertions (capturing in negative assertions does not happen
1752        in PCRE) and also to nested atomic groups.
1754    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1755        subject after a captured substring, to make it easier to tell which of a
1756        number of identical substrings has been captured.
1758    24. The way atomic groups are processed by pcre_exec() has been changed so that
1759        if they are repeated, backtracking one repetition now resets captured
1760        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1761        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1762        "aaa". Previously, it would have been "a". As part of this code
1763        refactoring, the way recursive calls are handled has also been changed.
1765    25. If an assertion condition captured any substrings, they were not passed
1766        back unless some other capturing happened later. For example, if
1767        (?(?=(a))a) was matched against "a", no capturing was returned.
1769    26. When studying a pattern that contained subroutine calls or assertions,
1770        the code for finding the minimum length of a possible match was handling
1771        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1772        group 1 called group 2 while simultaneously a separate group 2 called group
1773        1). A stack overflow occurred in this case. I have fixed this by limiting
1774        the recursion depth to 10.
1776    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1777        Fortmann. This supports explicit test numbers on the command line, and has
1778        argument validation and error reporting.
1780    28. An instance of \X with an unlimited repeat could fail if at any point the
1781        first character it looked at was a mark character.
1783    29. Some minor code refactoring concerning Unicode properties and scripts
1784        should reduce the stack requirement of match() slightly.
1786    30. Added the '=' option to pcretest to check the setting of unused capturing
1787        slots at the end of the pattern, which are documented as being -1, but are
1788        not included in the return count.
1790    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1791        compiled something random. Now it gives a compile-time error (as does
1792        Perl).
1794    32. A *MARK encountered during the processing of a positive assertion is now
1795        recorded and passed back (compatible with Perl).
1797    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1798        had alternative anchored branches, the search for a second match in a line
1799        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1800        matched the line "0102" twice. The same bug affected patterns that started
1801        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1802        twice.
1804    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1805        does, and because of capturing effects, quantifying parenthesized
1806        assertions may at times be useful. Quantifiers are now allowed for
1807        parenthesized assertions.
1809    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1811    36. \g was being checked for fancy things in a character class, when it should
1812        just be a literal "g".
1814    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1815        appearance of a nested POSIX class supersedes an apparent external class.
1816        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1817        unescaped square brackets may also appear as part of class names. For
1818        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1819        more like Perl. (But see 8.20/1 above.)
1821    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1822        was because it thought it was \N{name}, which is not supported).
1824    39. Add minix to OS list not supporting the -S option in pcretest.
1826    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1827        cannot analyze patterns in sufficient detail to catch mutual recursions
1828        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1829        subgroup is called recursively as a subpattern for a second time at the
1830        same position in the subject string. In previous releases this might have
1831        been caught by the recursion limit, or it might have run out of stack.
1833    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1834        happen only once. PCRE was, however incorrectly giving a compile time error
1835        "recursive call could loop indefinitely" because it cannot analyze the
1836        pattern in sufficient detail. The compile time test no longer happens when
1837        PCRE is compiling a conditional subpattern, but actual runaway loops are
1838        now caught at runtime (see 40 above).
1840    42. It seems that Perl allows any characters other than a closing parenthesis
1841        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1842        has been changed to be the same.
1844    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1845        as not to get warnings when autogen.sh is called. Also changed
1846        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1848    44. To help people who use pcregrep to scan files containing exceedingly long
1849        lines, the following changes have been made:
1851        (a) The default value of the buffer size parameter has been increased from
1852            8K to 20K. (The actual buffer used is three times this size.)
1854        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1855            PCRE is built.
1857        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1858            to be set at run time.
1860        (d) Numerical values in pcregrep options can be followed by K or M, for
1861            example --buffer-size=50K.
1863        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1864            given and the return code is set to 2.
1866    45. Add a pointer to the latest mark to the callout data block.
1868    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1869        partial match of an empty string instead of no match. This was specific to
1870        the use of ".".
1872    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1873        complete match instead of a partial match. This bug was dependent on both
1874        the PCRE_UTF8 and PCRE_DOTALL options being set.
1876    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1877        starting byte set, because \b was not being ignored.
1880    Version 8.12 15-Jan-2011
1881    ------------------------
1883    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1884        checks for such things as part of the documentation building process.
1886    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1887        --match-limit and --recursion-limit options (added for 8.11). In
1888        particular, this made one of the standard tests fail. (The integer value
1889        went into the wrong half of a long int.)
1891    3.  If the --colour option was given to pcregrep with -v (invert match), it
1892        did strange things, either producing crazy output, or crashing. It should,
1893        of course, ignore a request for colour when reporting lines that do not
1894        match.
1896    4.  Another pcregrep bug caused similar problems if --colour was specified with
1897        -M (multiline) and the pattern match finished with a line ending.
1899    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1900        matched in multiline mode, the following line was shown as part of the
1901        match. This seems wrong, so I have changed it.
1903    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1904        the check for further matches in the same line (so they could be coloured)
1905        to overrun the end of the current line. If another match was found, it was
1906        incorrectly shown (and then shown again when found in the next line).
1908    7.  If pcregrep was compiled under Windows, there was a reference to the
1909        function pcregrep_exit() before it was defined. I am assuming this was
1910        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1911        reported by a user. I've moved the definition above the reference.
1914    Version 8.11 10-Dec-2010
1915    ------------------------
1917    1.  (*THEN) was not working properly if there were untried alternatives prior
1918        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1919        backtracked to try for "b" instead of moving to the next alternative branch
1920        at the same level (in this case, to look for "c"). The Perl documentation
1921        is clear that when (*THEN) is backtracked onto, it goes to the "next
1922        alternative in the innermost enclosing group".
1924    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1925        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1926        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1927        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1928        (*THEN).
1930    3.  If \s appeared in a character class, it removed the VT character from
1931        the class, even if it had been included by some previous item, for example
1932        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1933        of \s, but is part of the POSIX "space" class.)
1935    4.  A partial match never returns an empty string (because you can always
1936        match an empty string at the end of the subject); however the checking for
1937        an empty string was starting at the "start of match" point. This has been
1938        changed to the "earliest inspected character" point, because the returned
1939        data for a partial match starts at this character. This means that, for
1940        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1941        (previously it gave "no match").
1943    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1944        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1945        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1946        has an implication that the given string is incomplete (because a partial
1947        match is preferred over a full match). For this reason, these items now
1948        give a partial match in this situation. [Aside: previously, the one case
1949        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1950        match rather than a full match, which was wrong by the old rules, but is
1951        now correct.]
1953    6.  There was a bug in the handling of #-introduced comments, recognized when
1954        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1955        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1956        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1957        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1958        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1959        places in pcre_compile().
1961    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1962        comments when looking ahead for named forward references to subpatterns,
1963        the only newline sequence it recognized was NL. It now handles newlines
1964        according to the set newline convention.
1966    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1967        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1968        cater for a lack of strerror(). These oversights have been fixed.
1970    9.  Added --match-limit and --recursion-limit to pcregrep.
1972    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1974    11. When the -o option was used, pcregrep was setting a return code of 1, even
1975        when matches were found, and --line-buffered was not being honoured.
1977    12. Added an optional parentheses number to the -o and --only-matching options
1978        of pcregrep.
1980    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1981        can match an empty string. The code to do it in pcretest and pcredemo
1982        needed fixing:
1984        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1985            only one byte after an empty string match just before CRLF (this case
1986            just got forgotten; "any" and "anycrlf" were OK).
1988        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1989            mode when an empty string match preceded an ASCII character followed by
1990            a non-ASCII character. (The code for advancing by one character rather
1991            than one byte was nonsense.)
1993        (c) The pcredemo.c sample program did not have any code at all to handle
1994            the cases when CRLF is a valid newline sequence.
1996    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1997        as a starting offset was within the subject string. There is now a new
1998        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1999        negative or greater than the length of the string. In order to test this,
2000        pcretest is extended to allow the setting of negative starting offsets.
2002    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2003        starting offset points to the beginning of a UTF-8 character was
2004        unnecessarily clumsy. I tidied it up.
2006    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2007        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2009    17. Nobody had reported that the --include_dir option, which was added in
2010        release 7.7 should have been called --include-dir (hyphen, not underscore)
2011        for compatibility with GNU grep. I have changed it to --include-dir, but
2012        left --include_dir as an undocumented synonym, and the same for
2013        --exclude-dir, though that is not available in GNU grep, at least as of
2014        release 2.5.4.
2016    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2017        characters from a string of bytes) have been redefined so as not to use
2018        loops, in order to improve performance in some environments. At the same
2019        time, I abstracted some of the common code into auxiliary macros to save
2020        repetition (this should not affect the compiled code).
2022    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2023        compile-time error is now given if \c is not followed by an ASCII
2024        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2025        different, and any byte value is allowed.)
2027    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2028        START_OPTIMIZE option, which is now allowed at compile time - but just
2029        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2030        to pcregrep and other applications that have no direct access to PCRE
2031        options. The new /Y option in pcretest sets this option when calling
2032        pcre_compile().
2034    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2035        back references. Groups containing recursive back references were forced to
2036        be atomic by that change, but in the case of named groups, the amount of
2037        memory required was incorrectly computed, leading to "Failed: internal
2038        error: code overflow". This has been fixed.
2040    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2041        pcretest.c, to avoid build problems in some Borland environments.
2044    Version 8.10 25-Jun-2010
2045  ------------------------  ------------------------
2047  1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and  1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
# Line 42  Version 8.10 03-Jun-2010 Line 2082  Version 8.10 03-Jun-2010
2082  11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was  11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2083      studied, and the match started with a letter with a code point greater than      studied, and the match started with a letter with a code point greater than
2084      127 whose first byte was different to the first byte of the other case of      127 whose first byte was different to the first byte of the other case of
2085      the letter, the other case of this starting letter was not recognized      the letter, the other case of this starting letter was not recognized
2086      (#976).      (#976).
2088  12. If a pattern that was studied started with a repeated Unicode property  12. If a pattern that was studied started with a repeated Unicode property
# Line 68  Version 8.10 03-Jun-2010 Line 2108  Version 8.10 03-Jun-2010
2109  17. Added a check for running out of memory when PCRE is compiled with  17. Added a check for running out of memory when PCRE is compiled with
2110      --disable-stack-for-recursion (#990).      --disable-stack-for-recursion (#990).
2112  18. If the last data line in a file for pcretest does not have a newline on  18. If the last data line in a file for pcretest does not have a newline on
2113      the end, a newline was missing in the output.      the end, a newline was missing in the output.
2115  19. The default pcre_chartables.c file recognizes only ASCII characters (values  19. The default pcre_chartables.c file recognizes only ASCII characters (values
2116      less than 128) in its various bitmaps. However, there is a facility for      less than 128) in its various bitmaps. However, there is a facility for
2117      generating tables according to the current locale when PCRE is compiled. It      generating tables according to the current locale when PCRE is compiled. It
2118      turns out that in some environments, 0x85 and 0xa0, which are Unicode space      turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2119      characters, are recognized by isspace() and therefore were getting set in      characters, are recognized by isspace() and therefore were getting set in
2120      these tables. This caused a problem in UTF-8 mode when pcre_study() was      these tables, and indeed these tables seem to approximate to ISO 8859. This
2121      used to create a list of bytes that can start a match. For \s, it was      caused a problem in UTF-8 mode when pcre_study() was used to create a list
2122      including 0x85 and 0xa0, which of course cannot start UTF-8 characters. I      of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2123      have changed the code so that only real ASCII characters (less than 128)      which of course cannot start UTF-8 characters. I have changed the code so
2124      are set in this case because the \s etc escapes are documented as      that only real ASCII characters (less than 128) and the correct starting
2125      recognizing only ASCII characters. (When PCRE_UCP is set - see 9 above -      bytes for UTF-8 encodings are set for characters greater than 127 when in
2126      the code is different altogether.)      UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2127        altogether.)
2129    20. Added the /T option to pcretest so as to be able to run tests with non-
2130        standard character tables, thus making it possible to include the tests
2131        used for 19 above in the standard set of tests.
2133    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2134        reference to a subpattern the other side of a comment that contains an
2135        opening parenthesis caused either an internal compiling error, or a
2136        reference to the wrong subpattern.
2139  Version 8.02 19-Mar-2010  Version 8.02 19-Mar-2010
# Line 477  Version 7.9 11-Apr-09 Line 2527  Version 7.9 11-Apr-09
2527  7.  A pattern that could match an empty string could cause pcregrep to loop; it  7.  A pattern that could match an empty string could cause pcregrep to loop; it
2528      doesn't make sense to accept an empty string match in pcregrep, so I have      doesn't make sense to accept an empty string match in pcregrep, so I have
2529      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2530      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
2531        8.33.]
2533  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2534      start or after a newline", because the conditional assertion was not being      start or after a newline", because the conditional assertion was not being
# Line 720  Version 7.7 07-May-08 Line 2771  Version 7.7 07-May-08
2771      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2772      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2773      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2774      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2775      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2776      was subsequently set up correctly.)      was subsequently set up correctly.)

Removed from v.538  
changed lines
  Added in v.1649

  ViewVC Help
Powered by ViewVC 1.1.5