ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1682 by ph10, Mon Feb 20 17:45:21 2017 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4    Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
7    Version 8.41
8    ------------
10    1.  Fixed typo in CMakeLists.txt (wrong number of arguments for
11    PCRE_STATIC_RUNTIME (affects MSVC only).
13    2. Issue 1 for 8.40 below was not correctly fixed. If pcregrep in multiline
14    mode with --only-matching matched several lines, it restarted scanning at the
15    next line instead of moving on to the end of the matched string, which can be
16    several lines after the start.
18    3.  Fix a missing else in the JIT compiler reported by 'idaifish'.
20    4.  A (?# style comment is now ignored between a basic quantifier and a
21    following '+' or '?' (example: /X+(?#comment)?Y/.
24    Version 8.40 11-January-2017
25    ----------------------------
27    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
28        the match extended over a line boundary.
30    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
31        MSVC static compilation, putting the first patch under a new option.
33    3.  Fix register overwite in JIT when SSE2 acceleration is enabled.
35    4.  Ignore "show all captures" (/=) for DFA matching.
37    5.  Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
39    6.  In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
40        without PCRE_UCP set, a negative character type such as \D in a positive
41        class should cause all characters greater than 255 to match, whatever else
42        is in the class. There was a bug that caused this not to happen if a
43        Unicode property item was added to such a class, for example [\D\P{Nd}] or
44        [\W\pL].
46    7.  When pcretest was outputing information from a callout, the caret indicator
47        for the current position in the subject line was incorrect if it was after
48        an escape sequence for a character whose code point was greater than
49        \x{ff}.
51    8.  A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
52        the conditional was interpreted as a reference to capturing group 1 instead
53        of a test for recursion. Any group whose name began with R was
54        misinterpreted in this way. (The reference interpretation should only
55        happen if the group's name is precisely "R".)
57    9.  A number of bugs have been mended relating to match start-up optimizations
58        when the first thing in a pattern is a positive lookahead. These all
59        applied only when PCRE_NO_START_OPTIMIZE was *not* set:
61        (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
62            both an initial 'X' and a following 'X'.
63        (b) Some patterns starting with an assertion that started with .* were
64            incorrectly optimized as having to match at the start of the subject or
65            after a newline. There are cases where this is not true, for example,
66            (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
67            start with spaces. Starting .* in an assertion is no longer taken as an
68            indication of matching at the start (or after a newline).
71    Version 8.39 14-June-2016
72    -------------------------
74    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
75        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
76        misbehaved. This bug was found by the LLVM fuzzer.
78    2.  Similar to the above, if an isolated \E was present between an item and its
79        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
80        bug was found by the LLVM fuzzer.
82    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
83        working correctly in UCP mode.
85    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
86        was set when the pmatch argument was NULL. It now returns REG_INVARG.
88    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
90    6.  An empty \Q\E sequence between an item and its qualifier caused
91        pcre_compile() to misbehave when auto callouts were enabled. This bug was
92        found by the LLVM fuzzer.
94    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
95        space or a #-type comment that was followed by (?-x), which turns off
96        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
97        pcre_compile() assumed that (?-x) applied to the whole pattern and
98        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
100    8.  A call of pcre_copy_named_substring() for a named substring whose number
101        was greater than the space in the ovector could cause a crash.
103    9.  Yet another buffer overflow bug involved duplicate named groups with a
104        group that reset capture numbers (compare 8.38/7 below). Once again, I have
105        just allowed for more memory, even if not needed. (A proper fix is
106        implemented in PCRE2, but it involves a lot of refactoring.)
108    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
109        start of the match to be earlier than the end.
111    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
113    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
114        assertion, caused pcretest to generate incorrect output, and also to read
115        uninitialized memory (detected by ASAN or valgrind).
117    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
118        nested set of parentheses of sufficient size caused an overflow of the
119        compiling workspace (which was diagnosed, but of course is not desirable).
121    14. And yet another buffer overflow bug involving duplicate named groups, this
122        time nested, with a nested back reference. Yet again, I have just allowed
123        for more memory, because anything more needs all the refactoring that has
124        been done for PCRE2. An example pattern that provoked this bug is:
125        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
126        registered as CVE-2016-1283.
128    15. pcretest went into a loop if global matching was requested with an ovector
129        size less than 2. It now gives an error message. This bug was found by
130        afl-fuzz.
132    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
133        ("assertion expected") when (?(?C) was not followed by an opening
134        parenthesis.
136    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
137        actually affect anything, by sheer luck.
139    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
140        static compilation.
142    19. Modified the RunTest script to incorporate a valgrind suppressions file so
143        that certain errors, provoked by the SSE2 instruction set when JIT is used,
144        are ignored.
146    20. A racing condition is fixed in JIT reported by Mozilla.
148    21. Minor code refactor to avoid "array subscript is below array bounds"
149        compiler warning.
151    22. Minor code refactor to avoid "left shift of negative number" warning.
153    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
154        UCP support.
156    24. Refactor to avoid compiler warnings in pcrecpp.cc.
158    25. Refactor to fix a typo in pcre_jit_test.c
160    26. Patch to support compiling pcrecpp.cc with Intel compiler.
163    Version 8.38 23-November-2015
164    -----------------------------
166    1.  If a group that contained a recursive back reference also contained a
167        forward reference subroutine call followed by a non-forward-reference
168        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
169        compile correct code, leading to undefined behaviour or an internally
170        detected error. This bug was discovered by the LLVM fuzzer.
172    2.  Quantification of certain items (e.g. atomic back references) could cause
173        incorrect code to be compiled when recursive forward references were
174        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
175        This bug was discovered by the LLVM fuzzer.
177    3.  A repeated conditional group whose condition was a reference by name caused
178        a buffer overflow if there was more than one group with the given name.
179        This bug was discovered by the LLVM fuzzer.
181    4.  A recursive back reference by name within a group that had the same name as
182        another group caused a buffer overflow. For example:
183        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
185    5.  A forward reference by name to a group whose number is the same as the
186        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
187        a buffer overflow at compile time. This bug was discovered by the LLVM
188        fuzzer.
190    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
191        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
193    7.  Another buffer overflow bug involved duplicate named groups with a
194        reference between their definition, with a group that reset capture
195        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
196        fixed by always allowing for more memory, even if not needed. (A proper fix
197        is implemented in PCRE2, but it involves more refactoring.)
199    8.  There was no check for integer overflow in subroutine calls such as (?123).
201    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
202        being treated as a literal 'l' instead of causing an error.
204    10. There was a buffer overflow if pcre_exec() was called with an ovector of
205        size 1. This bug was found by american fuzzy lop.
207    11. If a non-capturing group containing a conditional group that could match
208        an empty string was repeated, it was not identified as matching an empty
209        string itself. For example: /^(?:(?(1)x|)+)+$()/.
211    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
212        \a and \e in test subject lines.
214    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
215        instead of the EBCDIC value.
217    14. The handling of \c in an EBCDIC environment has been revised so that it is
218        now compatible with the specification in Perl's perlebcdic page.
220    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
221        ASCII/Unicode. This has now been added to the list of characters that are
222        recognized as white space in EBCDIC.
224    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
225        error (correctly) when used outside a class, but did not give an error
226        within a class.
228    17. \h within a class was incorrectly compiled in EBCDIC environments.
230    18. A pattern with an unmatched closing parenthesis that contained a backward
231        assertion which itself contained a forward reference caused buffer
232        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
234    19. JIT should return with error when the compiled pattern requires more stack
235        space than the maximum.
237    20. A possessively repeated conditional group that could match an empty string,
238        for example, /(?(R))*+/, was incorrectly compiled.
240    21. Fix infinite recursion in the JIT compiler when certain patterns such as
241        /(?:|a|){100}x/ are analysed.
243    22. Some patterns with character classes involving [: and \\ were incorrectly
244        compiled and could cause reading from uninitialized memory or an incorrect
245        error diagnosis.
247    23. Pathological patterns containing many nested occurrences of [: caused
248        pcre_compile() to run for a very long time.
250    24. A conditional group with only one branch has an implicit empty alternative
251        branch and must therefore be treated as potentially matching an empty
252        string.
254    25. If (?R was followed by - or + incorrect behaviour happened instead of a
255        diagnostic.
257    26. Arrange to give up on finding the minimum matching length for overly
258        complex patterns.
260    27. Similar to (4) above: in a pattern with duplicated named groups and an
261        occurrence of (?| it is possible for an apparently non-recursive back
262        reference to become recursive if a later named group with the relevant
263        number is encountered. This could lead to a buffer overflow. Wen Guanxing
264        from Venustech ADLAB discovered this bug.
266    28. If pcregrep was given the -q option with -c or -l, or when handling a
267        binary file, it incorrectly wrote output to stdout.
269    29. The JIT compiler did not restore the control verb head in case of *THEN
270        control verbs. This issue was found by Karl Skomski with a custom LLVM
271        fuzzer.
273    30. Error messages for syntax errors following \g and \k were giving inaccurate
274        offsets in the pattern.
276    31. Added a check for integer overflow in conditions (?(<digits>) and
277        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
278        fuzzer.
280    32. Handling recursive references such as (?2) when the reference is to a group
281        later in the pattern uses code that is very hacked about and error-prone.
282        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
283        give an internal error if it is obvious that compiling has gone wrong.
285    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
286        This issue was found by Karl Skomski with a custom LLVM fuzzer.
288    34. The JIT compiler should restore the control chain for empty possessive
289        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
291    35. Match limit check added to JIT recursion. This issue was found by Karl
292        Skomski with a custom LLVM fuzzer.
294    36. Yet another case similar to 27 above has been circumvented by an
295        unconditional allocation of extra memory. This issue is fixed "properly" in
296        PCRE2 by refactoring the way references are handled. Wen Guanxing
297        from Venustech ADLAB discovered this bug.
299    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
300        with a custom LLVM fuzzer.
302    38. Fixed a corner case of range optimization in JIT.
304    39. An incorrect error "overran compiling workspace" was given if there were
305        exactly enough group forward references such that the last one extended
306        into the workspace safety margin. The next one would have expanded the
307        workspace. The test for overflow was not including the safety margin.
309    40. A match limit issue is fixed in JIT which was found by Karl Skomski
310        with a custom LLVM fuzzer.
312    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
313        work under Windows. (Why has it taken so long for anyone to notice?)
315    42. In a character class such as [\W\p{Any}] where both a negative-type escape
316        ("not a word character") and a property escape were present, the property
317        escape was being ignored.
319    43. Fix crash caused by very long (*MARK) or (*THEN) names.
321    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
322        by a single ASCII character in a class item, was incorrectly compiled in
323        UCP mode. The POSIX class got lost, but only if the single character
324        followed it.
326    45. [:punct:] in UCP mode was matching some characters in the range 128-255
327        that should not have been matched.
329    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
330        class, all characters with code points greater than 255 are in the class.
331        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
332        such as \w are turned into Unicode properties), wide characters were not
333        correctly handled, and could fail to match.
336    Version 8.37 28-April-2015
337    --------------------------
339    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
340        for those parentheses to be closed with whatever has been captured so far.
341        However, it was failing to mark any other groups between the hightest
342        capture so far and the currrent group as "unset". Thus, the ovector for
343        those groups contained whatever was previously there. An example is the
344        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
346    2.  If an assertion condition was quantified with a minimum of zero (an odd
347        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
349    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
350        unrecognized modifier, a crash could occur.
352    4.  An attempt to do global matching in pcretest with a zero-length ovector
353        caused a crash.
355    5.  Fixed a memory leak during matching that could occur for a subpattern
356        subroutine call (recursive or otherwise) if the number of captured groups
357        that had to be saved was greater than ten.
359    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
360        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
361        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
363    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
364        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
366    8. When a pattern is compiled, it remembers the highest back reference so that
367       when matching, if the ovector is too small, extra memory can be obtained to
368       use instead. A conditional subpattern whose condition is a check on a
369       capture having happened, such as, for example in the pattern
370       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
371       setting the highest backreference number. This mattered only if pcre_exec()
372       was called with an ovector that was too small to hold the capture, and there
373       was no other kind of back reference (a situation which is probably quite
374       rare). The effect of the bug was that the condition was always treated as
375       FALSE when the capture could not be consulted, leading to a incorrect
376       behaviour by pcre_exec(). This bug has been fixed.
378    9. A reference to a duplicated named group (either a back reference or a test
379       for being set in a conditional) that occurred in a part of the pattern where
380       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
381       to be incorrectly calculated, leading to overwriting.
383    10. A mutually recursive set of back references such as (\2)(\1) caused a
384        segfault at study time (while trying to find the minimum matching length).
385        The infinite loop is now broken (with the minimum length unset, that is,
386        zero).
388    11. If an assertion that was used as a condition was quantified with a minimum
389        of zero, matching went wrong. In particular, if the whole group had
390        unlimited repetition and could match an empty string, a segfault was
391        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
392        assertions to be quantified, but not if they are being used as conditions,
393        so the above pattern is faulted by Perl. PCRE has now been changed so that
394        it also rejects such patterns.
396    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
397        failed to allow the zero-repeat case if pcre2_exec() was called with an
398        ovector too small to capture the group.
400    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
401        Red Hat Product Security:
403        (a) A crash if /K and /F were both set with the option to save the compiled
404        pattern.
406        (b) Another crash if the option to print captured substrings in a callout
407        was combined with setting a null ovector, for example \O\C+ as a subject
408        string.
410    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
411        forward reference repeated a large (but limited) number of times within a
412        repeated outer group that has a zero minimum quantifier, caused incorrect
413        code to be compiled, leading to the error "internal error:
414        previously-checked referenced subpattern not found" when an incorrect
415        memory address was read. This bug was reported as "heap overflow",
416        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
417        CVE-2015-2325.
419    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
420        call within a group that also contained a recursive back reference caused
421        incorrect code to be compiled. This bug was reported as "heap overflow",
422        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
423        number CVE-2015-2326.
425    24. Computing the size of the JIT read-only data in advance has been a source
426        of various issues, and new ones are still appear unfortunately. To fix
427        existing and future issues, size computation is eliminated from the code,
428        and replaced by on-demand memory allocation.
430    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
431        adjacent to the end of the range, and the range contained characters with
432        more than one other case, caused incorrect behaviour when compiled in UTF
433        mode. In that example, the range a-j was left out of the class.
435    26. Fix JIT compilation of conditional blocks, which assertion
436        is converted to (*FAIL). E.g: /(?(?!))/.
438    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
439        discovered by the LLVM fuzzer.
441    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
442        when this assertion was used as a condition, for example (?(?!)a|b). In
443        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
444        error about an unsupported item.
446    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
447        possessification code could take exponential time to complete. A recursion
448        depth limit of 1000 has been imposed to limit the resources used by this
449        optimization.
451    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
452        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
453        because \S ensures they are all in the class. The code for doing this was
454        interacting badly with the code for computing the amount of space needed to
455        compile the pattern, leading to a buffer overflow. This bug was discovered
456        by the LLVM fuzzer.
458    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
459        other kinds of group caused stack overflow at compile time. This bug was
460        discovered by the LLVM fuzzer.
462    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
463        between a subroutine call and its quantifier was incorrectly compiled,
464        leading to buffer overflow or other errors. This bug was discovered by the
465        LLVM fuzzer.
467    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
468        assertion after (?(. The code was failing to check the character after
469        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
470        was discovered by the LLVM fuzzer.
472    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
473        a fixed maximum following a group that contains a subroutine reference was
474        incorrectly compiled and could trigger buffer overflow. This bug was
475        discovered by the LLVM fuzzer.
477    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
478        caused a stack overflow instead of the diagnosis of a non-fixed length
479        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
481    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
482        (e.g. /(?<=\Ka)/) could make pcregrep loop.
484    37. There was a similar problem to 36 in pcretest for global matches.
486    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
487        and a subsequent item in the pattern caused a non-match, backtracking over
488        the repeated \X did not stop, but carried on past the start of the subject,
489        causing reference to random memory and/or a segfault. There were also some
490        other cases where backtracking after \C could crash. This set of bugs was
491        discovered by the LLVM fuzzer.
493    39. The function for finding the minimum length of a matching string could take
494        a very long time if mutual recursion was present many times in a pattern,
495        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
496        method has been implemented. This infelicity was discovered by the LLVM
497        fuzzer.
499    40. Static linking against the PCRE library using the pkg-config module was
500        failing on missing pthread symbols.
503    Version 8.36 26-September-2014
504    ------------------------------
506    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
507        -Wmissing-field-initializers and -Wunused-parameter.
509    2.  The tests for quantifiers being too big (greater than 65535) were being
510        applied after reading the number, and stupidly assuming that integer
511        overflow would give a negative number. The tests are now applied as the
512        numbers are read.
514    3.  Tidy code in pcre_exec.c where two branches that used to be different are
515        now the same.
517    4.  The JIT compiler did not generate match limit checks for certain
518        bracketed expressions with quantifiers. This may lead to exponential
519        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
520        issue should be resolved now.
522    5.  Fixed an issue, which occures when nested alternatives are optimized
523        with table jumps.
525    6.  Inserted two casts and changed some ints to size_t in the light of some
526        reported 64-bit compiler warnings (Bugzilla 1477).
528    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
529        an empty string, which sometimes were behaving incorrectly in the
530        interpreter (though correctly in the JIT matcher). This pcretest input is
531        an example:
533          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
536        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
537        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
538        for an empty string was breaking the inner loop and carrying on at a lower
539        level, when possessive repeated groups should always return to a higher
540        level as they have no backtrack points in them. The empty string test now
541        occurs at the outer level.
543    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
544        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
546    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
547        doesn't).
549    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
550        to hold a pattern to be incorrectly computed (too small) when there were
551        named back references to duplicated names. This could cause "internal
552        error: code overflow" or "double free or corruption" or other memory
553        handling errors.
555    11. When named subpatterns had the same prefixes, back references could be
556        confused. For example, in this pattern:
558          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
560        the reference to 'Name' was incorrectly treated as a reference to a
561        duplicate name.
563    12. A pattern such as /^s?c/mi8 where the optional character has more than
564        one "other case" was incorrectly compiled such that it would only try to
565        match starting at "c".
567    13. When a pattern starting with \s was studied, VT was not included in the
568        list of possible starting characters; this should have been part of the
569        8.34/18 patch.
571    14. If a character class started [\Qx]... where x is any character, the class
572        was incorrectly terminated at the ].
574    15. If a pattern that started with a caseless match for a character with more
575        than one "other case" was studied, PCRE did not set up the starting code
576        unit bit map for the list of possible characters. Now it does. This is an
577        optimization improvement, not a bug fix.
579    16. The Unicode data tables have been updated to Unicode 7.0.0.
581    17. Fixed a number of memory leaks in pcregrep.
583    18. Avoid a compiler warning (from some compilers) for a function call with
584        a cast that removes "const" from an lvalue by using an intermediate
585        variable (to which the compiler does not object).
587    19. Incorrect code was compiled if a group that contained an internal recursive
588        back reference was optional (had quantifier with a minimum of zero). This
589        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
590        caused segmentation faults because of stack overflows at compile time.
592    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
593        group that is quantified with an indefinite repeat, caused a compile-time
594        loop which used up all the system stack and provoked a segmentation fault.
595        This was not the same bug as 19 above.
597    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
598        Patch by Mike Frysinger.
601    Version 8.35 04-April-2014
602    --------------------------
604    1.  A new flag is set, when property checks are present in an XCLASS.
605        When this flag is not set, PCRE can perform certain optimizations
606        such as studying these XCLASS-es.
608    2.  The auto-possessification of character sets were improved: a normal
609        and an extended character set can be compared now. Furthermore
610        the JIT compiler optimizes more character set checks.
612    3.  Got rid of some compiler warnings for potentially uninitialized variables
613        that show up only when compiled with -O2.
615    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
616        of a match later then the end of the match. The pcretest program was not
617        handling the case sensibly - it was outputting from the start to the next
618        binary zero. It now reports this situation in a message, and outputs the
619        text from the end to the start.
621    5.  Fast forward search is improved in JIT. Instead of the first three
622        characters, any three characters with fixed position can be searched.
623        Search order: first, last, middle.
625    6.  Improve character range checks in JIT. Characters are read by an inprecise
626        function now, which returns with an unknown value if the character code is
627        above a certain threshold (e.g: 256). The only limitation is that the value
628        must be bigger than the threshold as well. This function is useful when
629        the characters above the threshold are handled in the same way.
631    7.  The macros whose names start with RAWUCHAR are placeholders for a future
632        mode in which only the bottom 21 bits of 32-bit data items are used. To
633        make this more memorable for those maintaining the code, the names have
634        been changed to start with UCHAR21, and an extensive comment has been added
635        to their definition.
637    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
638        to the export list in Makefile.am (they were accidentally omitted from the
639        8.34 tarball).
641    9.  The informational output from pcretest used the phrase "starting byte set"
642        which is inappropriate for the 16-bit and 32-bit libraries. As the output
643        for "first char" and "need char" really means "non-UTF-char", I've changed
644        "byte" to "char", and slightly reworded the output. The documentation about
645        these values has also been (I hope) clarified.
647    10. Another JIT related optimization: use table jumps for selecting the correct
648        backtracking path, when more than four alternatives are present inside a
649        bracket.
651    11. Empty match is not possible, when the minimum length is greater than zero,
652        and there is no \K in the pattern. JIT should avoid empty match checks in
653        such cases.
655    12. In a caseless character class with UCP support, when a character with more
656        than one alternative case was not the first character of a range, not all
657        the alternative cases were added to the class. For example, s and \x{17f}
658        are both alternative cases for S: the class [RST] was handled correctly,
659        but [R-T] was not.
661    13. The configure.ac file always checked for pthread support when JIT was
662        enabled. This is not used in Windows, so I have put this test inside a
663        check for the presence of windows.h (which was already tested for).
665    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
666        The algorithm provides a way to skip certain starting offsets, and usually
667        faster than linear prefix searches.
669    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
670        as for 'fr_FR' and 'french'. For some reason, however, it then used the
671        Windows-specific input and output files, which have 'french' screwed in.
672        So this could never have worked. One of the problems with locales is that
673        they aren't always the same. I have now updated RunTest so that it checks
674        the output of the locale test (test 3) against three different output
675        files, and it allows the test to pass if any one of them matches. With luck
676        this should make the test pass on some versions of Solaris where it was
677        failing. Because of the uncertainty, the script did not used to stop if
678        test 3 failed; it now does. If further versions of a French locale ever
679        come to light, they can now easily be added.
681    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
682        there was a message during ./configure, but it did not stop. This now
683        provokes an error. The invalid example in README has been corrected.
684        If a value less than the minimum is given, the minimum value has always
685        been used, but now a warning is given.
687    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
688        was a bug in the test system, which is now fixed. Also, the list of various
689        configurations that are tested for each release did not have one with both
690        16/32 bits and --enable-bar-anycrlf. It now does.
692    18. pcretest was missing "-C bsr" for displaying the \R default setting.
694    19. Little endian PowerPC systems are supported now by the JIT compiler.
696    20. The fast forward newline mechanism could enter to an infinite loop on
697        certain invalid UTF-8 input. Although we don't support these cases
698        this issue can be fixed by a performance optimization.
700    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
701        not take account if existing stack usage. There is now a new global
702        variable called pcre_stack_guard that can be set to point to an external
703        function to check stack availability. It is called at the start of
704        processing every parenthesized group.
706    22. A typo in the code meant that in ungreedy mode the max/min qualifier
707        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
708        not match "ab".
710    23. When UTF was disabled, the JIT program reported some incorrect compile
711        errors. These messages are silenced now.
713    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
714        compiler.
716    25. Change all the temporary files used in RunGrepTest to be different to those
717        used by RunTest so that the tests can be run simultaneously, for example by
718        "make -j check".
721    Version 8.34 15-December-2013
722    -----------------------------
724    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
725        executable memory. Patch inspired by Carsten Klein.
727    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
728        this macro is never tested and has no effect, because the work to support
729        coverage involves only compiling and linking options and special targets in
730        the Makefile. The comment in config.h implied that defining the macro would
731        enable coverage support, which is totally false. There was also support for
732        setting this macro in the CMake files (my fault, I just copied it from
733        configure). SUPPORT_GCOV has now been removed.
735    3.  Make a small performance improvement in strlen16() and strlen32() in
736        pcretest.
738    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
739        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
740        these cases). There was also one in pcretest.c.
742    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
744    6.  In UTF mode, the code for checking whether a group could match an empty
745        string (which is used for indefinitely repeated groups to allow for
746        breaking an infinite loop) was broken when the group contained a repeated
747        negated single-character class with a character that occupied more than one
748        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
749        UTF-8 mode). The effect was undefined: the group might or might not be
750        deemed as matching an empty string, or the program might have crashed.
752    7.  The code for checking whether a group could match an empty string was not
753        recognizing that \h, \H, \v, \V, and \R must match a character.
755    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
756        an empty string. If it can, pcretest shows this in its information output.
758    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
759        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
760        matched by pcre_exec() without using JIT:
762        (a) If the rest of the pattern did not match after a maximal run of
763            grapheme clusters, the code for backing up to try with fewer of them
764            did not always back up over a full grapheme when characters that do not
765            have the modifier quality were involved, e.g. Hangul syllables.
767        (b) If the match point in a subject started with modifier character, and
768            there was no match, the code could incorrectly back up beyond the match
769            point, and potentially beyond the first character in the subject,
770            leading to a segfault or an incorrect match result.
772    10. A conditional group with an assertion condition could lead to PCRE
773        recording an incorrect first data item for a match if no other first data
774        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
775        first data item, and therefore matched "ca" after "c" instead of at the
776        start.
778    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
779        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
781    12. The source of pcregrep now includes z/OS-specific code so that it can be
782        compiled for z/OS as part of the special z/OS distribution.
784    13. Added the -T and -TM options to pcretest.
786    14. The code in pcre_compile.c for creating the table of named capturing groups
787        has been refactored. Instead of creating the table dynamically during the
788        actual compiling pass, the information is remembered during the pre-compile
789        pass (on the stack unless there are more than 20 named groups, in which
790        case malloc() is used) and the whole table is created before the actual
791        compile happens. This has simplified the code (it is now nearly 150 lines
792        shorter) and prepared the way for better handling of references to groups
793        with duplicate names.
795    15. A back reference to a named subpattern when there is more than one of the
796        same name now checks them in the order in which they appear in the pattern.
797        The first one that is set is used for the reference. Previously only the
798        first one was inspected. This change makes PCRE more compatible with Perl.
800    16. Unicode character properties were updated from Unicode 6.3.0.
802    17. The compile-time code for auto-possessification has been refactored, based
803        on a patch by Zoltan Herczeg. It now happens after instead of during
804        compilation. The code is cleaner, and more cases are handled. The option
805        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
806        options in pcretest are provided to set it. It can also be set by
807        (*NO_AUTO_POSSESS) at the start of a pattern.
809    18. The character VT has been added to the default ("C" locale) set of
810        characters that match \s and are generally treated as white space,
811        following this same change in Perl 5.18. There is now no difference between
812        "Perl space" and "POSIX space". Whether VT is treated as white space in
813        other locales depends on the locale.
815    19. The code for checking named groups as conditions, either for being set or
816        for being recursed, has been refactored (this is related to 14 and 15
817        above). Processing unduplicated named groups should now be as fast at
818        numerical groups, and processing duplicated groups should be faster than
819        before.
821    20. Two patches to the CMake build system, by Alexander Barkov:
823          (1) Replace the "source" command by "." in CMakeLists.txt because
824              "source" is a bash-ism.
826          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
827              without these the CMake build does not work on Solaris.
829    21. Perl has changed its handling of \8 and \9. If there is no previously
830        encountered capturing group of those numbers, they are treated as the
831        literal characters 8 and 9 instead of a binary zero followed by the
832        literals. PCRE now does the same.
834    22. Following Perl, added \o{} to specify codepoints in octal, making it
835        possible to specify values greater than 0777 and also making them
836        unambiguous.
838    23. Perl now gives an error for missing closing braces after \x{... instead of
839        treating the string as literal. PCRE now does the same.
841    24. RunTest used to grumble if an inappropriate test was selected explicitly,
842        but just skip it when running all tests. This make it awkward to run ranges
843        of tests when one of them was inappropriate. Now it just skips any
844        inappropriate tests, as it always did when running all tests.
846    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
847        character types such as \d or \w, too many callouts were inserted, and the
848        data that they returned was rubbish.
850    26. In UCP mode, \s was not matching two of the characters that Perl matches,
851        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
852        were matched by \h. The code has now been refactored so that the lists of
853        the horizontal and vertical whitespace characters used for \h and \v (which
854        are defined only in one place) are now also used for \s.
856    27. Add JIT support for the 64 bit TileGX architecture.
857        Patch by Jiong Wang (Tilera Corporation).
859    28. Possessive quantifiers for classes (both explicit and automatically
860        generated) now use special opcodes instead of wrapping in ONCE brackets.
862    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
863        (because it's meaningless), this was not happening when PCRE_CASELESS was
864        set. Not wrong, but inefficient.
866    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
867        properties for \w, \d, etc) is present in a test regex. Otherwise if the
868        test contains no characters greater than 255, Perl doesn't realise it
869        should be using Unicode semantics.
871    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
872        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
873        does in Unicode mode.
875    32. Added the "forbid" facility to pcretest so that putting tests into the
876        wrong test files can sometimes be quickly detected.
878    33. There is now a limit (default 250) on the depth of nesting of parentheses.
879        This limit is imposed to control the amount of system stack used at compile
880        time. It can be changed at build time by --with-parens-nest-limit=xxx or
881        the equivalent in CMake.
883    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
884        errors. Perl warns for these when in warning mode, but PCRE has no facility
885        for giving warnings.
887    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
888        However, this was not working for (?!) because it is optimized to (*FAIL),
889        for which PCRE does not allow quantifiers. The optimization is now disabled
890        when a quantifier follows (?!). I can't see any use for this, but it makes
891        things uniform.
893    36. Perl no longer allows group names to start with digits, so I have made this
894        change also in PCRE. It simplifies the code a bit.
896    37. In extended mode, Perl ignores spaces before a + that indicates a
897        possessive quantifier. PCRE allowed a space before the quantifier, but not
898        before the possessive +. It now does.
900    38. The use of \K (reset reported match start) within a repeated possessive
901        group such as (a\Kb)*+ was not working.
903    40. Document that the same character tables must be used at compile time and
904        run time, and that the facility to pass tables to pcre_exec() and
905        pcre_dfa_exec() is for use only with saved/restored patterns.
907    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
908        features for Builds with MSVC:
910        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
911           builds). The .rc files can be used to set FileDescription and many other
912           attributes.
914        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
915           This allows higher-level build scripts which want .pdb files to avoid
916           hard-coding the exact files needed."
918    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
919        mean "start of word" and "end of word", respectively, as a transition aid.
921    43. A minimizing repeat of a class containing codepoints greater than 255 in
922        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
923        compiled to use the heap for recursion.
925    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
926        is configured.
929    Version 8.33 28-May-2013
930    ------------------------
932    1.  Added 'U' to some constants that are compared to unsigned integers, to
933        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
934        variables that are added to signed variables, to ensure the result is
935        signed and can be negated.
937    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
938        CMake config files.
940    3.  Revise the creation of config.h.generic so that all boolean macros are
941        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
942        overriding via -D on the command line possible.
944    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
945        to unsigned int is reported to make a quite noticeable speed difference in
946        a specific Windows environment. Testing on Linux did also appear to show
947        some benefit (and it is clearly not harmful). Also fixed the definition of
948        Xop which should be unsigned.
950    5.  Related to (4), changing the definition of the intermediate variable cc
951        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
952        improvements.
954    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
955        unnecessary spaces.
957    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
958        and later.
960    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
962    9.  Optimizing fast_forward_start_bits in JIT.
964    10. Adding support for callouts in JIT, and fixing some issues revealed
965        during this work. Namely:
967        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
969        (b) Minimum length was not checked before the matching is started.
971    11. The value of capture_last that is passed to callouts was incorrect in some
972        cases when there was a capture on one path that was subsequently abandoned
973        after a backtrack. Also, the capture_last value is now reset after a
974        recursion, since all captures are also reset in this case.
976    12. The interpreter no longer returns the "too many substrings" error in the
977        case when an overflowing capture is in a branch that is subsequently
978        abandoned after a backtrack.
980    13. In the pathological case when an offset vector of size 2 is used, pcretest
981        now prints out the matched string after a yield of 0 or 1.
983    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
984        Only supported by the JIT compiler at the moment.
986    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
988    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
989        offset of the starting point of the matching process, provided the offsets
990        vector is large enough.
992    17. The \A escape now records a lookbehind value of 1, though its execution
993        does not actually inspect the previous character. This is to ensure that,
994        in partial multi-segment matching, at least one character from the old
995        segment is retained when a new segment is processed. Otherwise, if there
996        are no lookbehinds in the pattern, \A might match incorrectly at the start
997        of a new segment.
999    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
1001    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
1002        gives some modest performance improvement in 8-bit mode.
1004    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1005        be expressed in certain programming languages using Universal Character
1006        Names.
1008    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1009        which points out that "non characters" are not "characters that may not
1010        appear in Unicode strings" but rather "characters that are reserved for
1011        internal use and have only local meaning".
1013    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1014        there was a conditional group that depended on an assertion, if the
1015        assertion was false, the callout that immediately followed the alternation
1016        in the condition was skipped when pcre_exec() was used for matching.
1018    23. Allow an explicit callout to be inserted before an assertion that is the
1019        condition for a conditional group, for compatibility with automatic
1020        callouts, which always insert a callout at this point.
1022    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1023        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1025    25. (*PRUNE) is now supported by the JIT compiler.
1027    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1029    27. Fix the case where there are two or more SKIPs with arguments that may be
1030        ignored.
1032    28. (*SKIP) is now supported by the JIT compiler.
1034    29. (*THEN) is now supported by the JIT compiler.
1036    30. Update RunTest with additional test selector options.
1038    31. The way PCRE handles backtracking verbs has been changed in two ways.
1040        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1041        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1042        backtracking. In some cases this makes it more Perl-compatible, but Perl's
1043        rather obscure rules do not always do the same thing.
1045        (2) Previously, backtracking verbs were confined within assertions. This is
1046        no longer the case for positive assertions, except for (*ACCEPT). Again,
1047        this sometimes improves Perl compatibility, and sometimes does not.
1049    32. A number of tests that were in test 2 because Perl did things differently
1050        have been moved to test 1, because either Perl or PCRE has changed, and
1051        these tests are now compatible.
1053    32. Backtracking control verbs are now handled in the same way in JIT and
1054        interpreter.
1056    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1057        contained a forward subroutine reference caused a compile error.
1059    34. Auto-detect and optimize limited repetitions in JIT.
1061    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1062        blocking (*UTF) etc.
1064    36. In the interpreter, maximizing pattern repetitions for characters and
1065        character types now use tail recursion, which reduces stack usage.
1067    37. The value of the max lookbehind was not correctly preserved if a compiled
1068        and saved regex was reloaded on a host of different endianness.
1070    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1071        of the compiled pattern block, expand the flags field from 16 to 32 bits
1072        because it was almost full.
1074    39. Try madvise first before posix_madvise.
1076    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1077        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1078        original change locked out all matches of empty strings. This has been
1079        changed so that one match of an empty string per line is recognized.
1080        Subsequent searches on the same line (for colouring or for --only-matching,
1081        for example) do not recognize empty strings.
1083    41. Applied a user patch to fix a number of spelling mistakes in comments.
1085    42. Data lines longer than 65536 caused pcretest to crash.
1087    43. Clarified the data type for length and startoffset arguments for pcre_exec
1088        and pcre_dfa_exec in the function-specific man pages, where they were
1089        explicitly stated to be in bytes, never having been updated. I also added
1090        some clarification to the pcreapi man page.
1092    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1093        a segmentation fault.
1096    Version 8.32 30-November-2012
1097    -----------------------------
1099    1.  Improved JIT compiler optimizations for first character search and single
1100        character iterators.
1102    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1103        Patch by Daniel Richard G.
1105    3.  Single character iterator optimizations in the JIT compiler.
1107    4.  Improved JIT compiler optimizations for character ranges.
1109    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1110        Reported by Giuseppe D'Angelo.
1112    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1113        of a line, was being set incorrectly in cases where .* appeared inside
1114        atomic brackets at the start of a pattern, or where there was a subsequent
1115        *PRUNE or *SKIP.
1117    7.  Improved instruction cache flush for POWER/PowerPC.
1118        Patch by Daniel Richard G.
1120    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1121        grep:
1123        (a) There is now no limit to the number of patterns to be matched.
1125        (b) An error is given if a pattern is too long.
1127        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1128            are now supported.
1130        (d) --exclude-from and --include-from (multiple use) have been added.
1132        (e) Exclusions and inclusions now apply to all files and directories, not
1133            just to those obtained from scanning a directory recursively.
1135        (f) Multiple uses of -f and --file-list are now supported.
1137        (g) In a Windows environment, the default for -d has been changed from
1138            "read" (the GNU grep default) to "skip", because otherwise the presence
1139            of a directory in the file list provokes an error.
1141        (h) The documentation has been revised and clarified in places.
1143    9.  Improve the matching speed of capturing brackets.
1145    10. Changed the meaning of \X so that it now matches a Unicode extended
1146        grapheme cluster.
1148    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1149        out POSIX threads when JIT support is configured.
1151    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1153    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1154        structure could go wrong in environments where size_t is not the same size
1155        as int.
1157    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1158        set.
1160    15. The EBCDIC support had decayed; later updates to the code had included
1161        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1162        general tidy up of EBCDIC-related issues, and the documentation was also
1163        not quite right. There is now a test that can be run on ASCII systems to
1164        check some of the EBCDIC-related things (but is it not a full test).
1166    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1167        in a small tidy to the code.
1169    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1171    18. If the --only-matching (-o) option in pcregrep is specified multiple
1172        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1173        substrings matched by the 1st and 2nd capturing parentheses. A separating
1174        string can be specified by --om-separator (default empty).
1176    19. Improving the first n character searches.
1178    20. Turn case lists for horizontal and vertical white space into macros so that
1179        they are defined only once.
1181    21. This set of changes together give more compatible Unicode case-folding
1182        behaviour for characters that have more than one other case when UCP
1183        support is available.
1185        (a) The Unicode property table now has offsets into a new table of sets of
1186            three or more characters that are case-equivalent. The MultiStage2.py
1187            script that generates these tables (the pcre_ucd.c file) now scans
1188            CaseFolding.txt instead of UnicodeData.txt for character case
1189            information.
1191        (b) The code for adding characters or ranges of characters to a character
1192            class has been abstracted into a generalized function that also handles
1193            case-independence. In UTF-mode with UCP support, this uses the new data
1194            to handle characters with more than one other case.
1196        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1197            whose other case is greater than 256 are now correctly matched
1198            caselessly. Previously, the high codepoint matched the low one, but not
1199            vice versa.
1201        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1202            of the new class addition function, using character lists defined as
1203            macros alongside the case definitions of 20 above.
1205        (e) Caseless back references now work with characters that have more than
1206            one other case.
1208        (f) General caseless matching of characters with more than one other case
1209            is supported.
1211    22. Unicode character properties were updated from Unicode 6.2.0
1213    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1215    24. Add support for 32-bit character strings, and UTF-32
1217    25. Major JIT compiler update (code refactoring and bugfixing).
1218        Experimental Sparc 32 support is added.
1220    26. Applied a modified version of Daniel Richard G's patch to create
1221        pcre.h.generic and config.h.generic by "make" instead of in the
1222        PrepareRelease script.
1224    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1225        pcre_compile.c when checking for a zero character.
1227    28. Introducing a native interface for JIT. Through this interface, the compiled
1228        machine code can be directly executed. The purpose of this interface is to
1229        provide fast pattern matching, so several sanity checks are not performed.
1230        However, feature tests are still performed. The new interface provides
1231        1.4x speedup compared to the old one.
1233    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1234        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1235        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1237    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1238        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1239        the "old" RFC 2279). Instead, it ended up passing a negative length to
1240        pcre_exec().
1242    31. Add support for GCC's visibility feature to hide internal functions.
1244    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1245        "unknown -C option" after outputting 0 or 1.
1247    33. There is now support for generating a code coverage report for the test
1248        suite in environments where gcc is the compiler and lcov is installed. This
1249        is mainly for the benefit of the developers.
1251    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1252        unaddressable using valgrind annotations, allowing valgrind to detect
1253        invalid memory accesses. This is mainly for the benefit of the developers.
1255    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1257    26. Give configure error if --enable-cpp but no C++ compiler found.
1260    Version 8.31 06-July-2012
1261    -------------------------
1263    1.  Fixing a wrong JIT test case and some compiler warnings.
1265    2.  Removed a bashism from the RunTest script.
1267    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1268        to unsigned type, result still unsigned" that was given by an MS compiler
1269        on encountering the code "-sizeof(xxx)".
1271    4.  Partial matching support is added to the JIT compiler.
1273    5.  Fixed several bugs concerned with partial matching of items that consist
1274        of more than one character:
1276        (a) /^(..)\1/ did not partially match "aba" because checking references was
1277            done on an "all or nothing" basis. This also applied to repeated
1278            references.
1280        (b) \R did not give a hard partial match if \r was found at the end of the
1281            subject.
1283        (c) \X did not give a hard partial match after matching one or more
1284            characters at the end of the subject.
1286        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1287            a partial match for the string "\r".
1289        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1290            a partial match for a CR character at the end of the subject string.
1292    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1293        running pcretest, the text "(JIT)" added to the output whenever JIT is
1294        actually used to run the match.
1296    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1297        or /S+[+] with a digit between 1 and 7.
1299    8.  OP_NOT now supports any UTF character not just single-byte ones.
1301    9.  (*MARK) control verb is now supported by the JIT compiler.
1303    10. The command "./RunTest list" lists the available tests without actually
1304        running any of them. (Because I keep forgetting what they all are.)
1308    12. Applied a (slightly modified) user-supplied patch that improves performance
1309        when the heap is used for recursion (compiled with --disable-stack-for-
1310        recursion). Instead of malloc and free for each heap frame each time a
1311        logical recursion happens, frames are retained on a chain and re-used where
1312        possible. This sometimes gives as much as 30% improvement.
1314    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1315        call.
1317    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1319    15. It is now possible to link pcretest with libedit as an alternative to
1320        libreadline.
1322    16. (*COMMIT) control verb is now supported by the JIT compiler.
1324    17. The Unicode data tables have been updated to Unicode 6.1.0.
1326    18. Added --file-list option to pcregrep.
1328    19. Added binary file support to pcregrep, including the -a, --binary-files,
1329        -I, and --text options.
1331    20. The madvise function is renamed for posix_madvise for QNX compatibility
1332        reasons. Fixed by Giuseppe D'Angelo.
1334    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1336    22. Changed the option for creating symbolic links for 16-bit man pages from
1337        -s to -sf so that re-installing does not cause issues.
1339    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1341    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1342        matches in certain environments (the workspace was not being correctly
1343        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1344        some of the workspace data at the beginning of a restart.
1346    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1347        was not doing so when it should - probably a typo introduced by SVN 528
1348        (change 8.10/14).
1350    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1351        \w+ when the character tables indicated that \x{c4} was a word character.
1352        There were several related cases, all because the tests for doing a table
1353        lookup were testing for characters less than 127 instead of 255.
1355    27. If a pattern contains capturing parentheses that are not used in a match,
1356        their slots in the ovector are set to -1. For those that are higher than
1357        any matched groups, this happens at the end of processing. In the case when
1358        there were back references that the ovector was too small to contain
1359        (causing temporary malloc'd memory to be used during matching), and the
1360        highest capturing number was not used, memory off the end of the ovector
1361        was incorrectly being set to -1. (It was using the size of the temporary
1362        memory instead of the true size.)
1364    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1365        ovector size, it uses memory at the end of the block that it has got.
1367    29. Check for an overlong MARK name and give an error at compile time. The
1368        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1370    30. JIT compiler update.
1372    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1373        Rill for the patch.
1375    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1377    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1379    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1380        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1381        was enabled.
1383    35. Improve JIT code generation for greedy plus quantifier.
1385    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1386        1 to "aa" instead of to an empty string. The bug affected repeated groups
1387        that could potentially match an empty string.
1389    37. Optimizing single character iterators in JIT.
1391    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1392        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1393        codepoints that are too big for the mode are faulted, and in a UTF mode,
1394        disallowed codepoints are also faulted.
1396    39. If PCRE was compiled with UTF support, in three places in the DFA
1397        matcher there was code that should only have been obeyed in UTF mode, but
1398        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1399        processing when bytes with values greater than 127 were present. In 16-bit
1400        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1401        both cases the values are those that cannot be the first data item in a UTF
1402        character. The three items that might have provoked this were recursions,
1403        possessively repeated groups, and atomic groups.
1405    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1406        and pcregrep, because some OS require shared objects to be explicitly
1407        passed to ld, causing the link step to fail if they are not.
1409    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1410        mode, patterns that started with \h* or \R* might be incorrectly matched.
1413    Version 8.30 04-February-2012
1414    -----------------------------
1416    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1417        name is defined in ctype.h.
1419    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1420        only in quite long subpatterns.
1422    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1423        since it was replaced by pcre_fullinfo() in February 2000.
1425    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1426        match a (*MARK), and the match failed at the start of the subject, a
1427        reference to memory before the start of the subject could occur. This bug
1428        was introduced by fix 17 of release 8.21.
1430    5.  A reference to an unset group with zero minimum repetition was giving
1431        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1432        /(another)?(\1?)test/ matched against "hello world test". This bug was
1433        introduced in release 8.13.
1435    6.  Add support for 16-bit character strings (a large amount of work involving
1436        many changes and refactorings).
1438    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1439        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1440        from a file.
1442    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1443        rounding is not applied in this particular case).
1445    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1446        if they appear, or are escaped, in patterns.
1448    10. Get rid of a number of -Wunused-but-set-variable warnings.
1450    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1451        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1452        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1453        also returns the mark "x". This bug applied to capturing parentheses,
1454        non-capturing parentheses, and atomic parentheses. It also applied to some
1455        assertions.
1457    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1458        information out of configure.ac instead of relying on pcre.h.generic, which
1459        is not stored in the repository.
1461    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1462        -lreadline.
1464    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1466    15. Applied Graycode's patch to put the top-level frame on the stack rather
1467        than the heap when not using the stack for recursion. This gives a
1468        performance improvement in many cases when recursion is not deep.
1470    16. Experimental code added to "pcretest -C" to output the stack frame size.
1473    Version 8.21 12-Dec-2011
1474    ------------------------
1476    1.  Updating the JIT compiler.
1478    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1479        are added as well.
1481    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1482        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1483        calling _pcre_jit_exec. Some extra comments are added.
1485    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1486        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1487        was introduced by change 18 for 8.20.
1489    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1490        ECMA-262 standard.
1492    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1493        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1494        This bug was probably introduced by change 9 of 8.13.
1496    7.  While fixing 6 above, I noticed that a number of other items were being
1497        incorrectly rejected as "not fixed length". This arose partly because newer
1498        opcodes had not been added to the fixed-length checking code. I have (a)
1499        corrected the bug and added tests for these items, and (b) arranged for an
1500        error to occur if an unknown opcode is encountered while checking for fixed
1501        length instead of just assuming "not fixed length". The items that were
1502        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1503        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1504        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1506    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1507        being incorrectly compiled and would have given unpredicatble results.
1509    9.  A possessively repeated subpattern with minimum repeat count greater than
1510        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1511        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1512        the first (A) could occur when it should not.
1514    10. Add a cast and remove a redundant test from the code.
1516    11. JIT should use pcre_malloc/pcre_free for allocation.
1518    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1519        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1520        is anything other than /usr, -L is still shown).
1522    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1524    14. Perl does not support \N without a following name in a [] class; PCRE now
1525        also gives an error.
1527    15. If a forward reference was repeated with an upper limit of around 2000,
1528        it caused the error "internal error: overran compiling workspace". The
1529        maximum number of forward references (including repeats) was limited by the
1530        internal workspace, and dependent on the LINK_SIZE. The code has been
1531        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1532        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1533        of around 200,000 forward references. While doing this, I also speeded up
1534        the filling in of repeated forward references.
1536    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1537        incorrectly expecting the subject to contain another "a" after the start.
1539    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1540        in the match, the SKIP should be ignored. This was not happening; instead
1541        the SKIP was being treated as NOMATCH. For patterns such as
1542        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1543        tested.
1545    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1546        now much more compatible with Perl, in particular in cases where the result
1547        is a non-match for a non-anchored pattern. For example, if
1548        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1549        "m", where previously it did not return a name. A side effect of this
1550        change is that for partial matches, the last encountered mark name is
1551        returned, as for non matches. A number of tests that were previously not
1552        Perl-compatible have been moved into the Perl-compatible test files. The
1553        refactoring has had the pleasing side effect of removing one argument from
1554        the match() function, thus reducing its stack requirements.
1556    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1557        subsequent uses of /S (without +) incorrectly behaved like /S+.
1559    21. Retrieve executable code size support for the JIT compiler and fixing
1560        some warnings.
1562    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1563        not work when the shorter character appeared right at the end of the
1564        subject string.
1566    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1567        systems.
1569    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1570        output it when the /M option is used in pcretest.
1572    25. The CheckMan script was not being included in the distribution. Also, added
1573        an explicit "perl" to run Perl scripts from the PrepareRelease script
1574        because this is reportedly needed in Windows.
1576    26. If study data was being save in a file and studying had not found a set of
1577        "starts with" bytes for the pattern, the data written to the file (though
1578        never used) was taken from uninitialized memory and so caused valgrind to
1579        complain.
1581    27. Updated RunTest.bat as provided by Sheri Pierce.
1583    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1585    29. Computation of memory usage for the table of capturing group names was
1586        giving an unnecessarily large value.
1589    Version 8.20 21-Oct-2011
1590    ------------------------
1592    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1593        a POSIX class. After further experiments with Perl, which convinced me that
1594        Perl has bugs and confusions, a closing square bracket is no longer allowed
1595        in a POSIX name. This bug also affected patterns with classes that started
1596        with full stops.
1598    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1599        captured substring, but while checking the failing first alternative,
1600        substring 1 is temporarily captured. If the output vector supplied to
1601        pcre_exec() was not big enough for this capture, the yield of the function
1602        was still zero ("insufficient space for captured substrings"). This cannot
1603        be totally fixed without adding another stack variable, which seems a lot
1604        of expense for a edge case. However, I have improved the situation in cases
1605        such as /(a)(b)x|abc/ matched against "abc", where the return code
1606        indicates that fewer than the maximum number of slots in the ovector have
1607        been set.
1609    3.  Related to (2) above: when there are more back references in a pattern than
1610        slots in the output vector, pcre_exec() uses temporary memory during
1611        matching, and copies in the captures as far as possible afterwards. It was
1612        using the entire output vector, but this conflicts with the specification
1613        that only 2/3 is used for passing back captured substrings. Now it uses
1614        only the first 2/3, for compatibility. This is, of course, another edge
1615        case.
1617    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1618        main code base, and can be used by building with --enable-jit. When this is
1619        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1620        runtime --no-jit option is given.
1622    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1623        ovector, the return from the function was zero, implying that there were
1624        other matches that did not fit. The correct "exactly full" value is now
1625        returned.
1627    6.  If a subpattern that was called recursively or as a subroutine contained
1628        (*PRUNE) or any other control that caused it to give a non-standard return,
1629        invalid errors such as "Error -26 (nested recursion at the same subject
1630        position)" or even infinite loops could occur.
1632    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1633        computing the minimum length on reaching *ACCEPT, and so ended up with the
1634        wrong value of 1 rather than 0. Further investigation indicates that
1635        computing a minimum subject length in the presence of *ACCEPT is difficult
1636        (think back references, subroutine calls), and so I have changed the code
1637        so that no minimum is registered for a pattern that contains *ACCEPT.
1639    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1640        it was not handled as intended. [But see 16 below.]
1642    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1643        Sheri Pierce.
1645    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1646        the first byte in a match must be "a".
1648    11. Change 17 for 8.13 increased the recursion depth for patterns like
1649        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1650        pattern contains any instances of (*THEN). If it does not, the old
1651        optimizations are restored. It would be nice to do this on a per-group
1652        basis, but at the moment that is not feasible.
1654    12. In some environments, the output of pcretest -C is CRLF terminated. This
1655        broke RunTest's code that checks for the link size. A single white space
1656        character after the value is now allowed for.
1658    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1659        For "fr", it uses the Windows-specific input and output files.
1661    14. If (*THEN) appeared in a group that was called recursively or as a
1662        subroutine, it did not work as intended. [But see next item.]
1664    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1665        pattern fragments (but not containing any | characters). If A and B are
1666        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1667        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1668        D. In other words, Perl considers parentheses that do not contain any |
1669        characters to be part of a surrounding alternative, whereas PCRE was
1670        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1671        differently. PCRE now behaves in the same way as Perl, except in the case
1672        of subroutine/recursion calls such as (?1) which have in any case always
1673        been different (but PCRE had them first :-).
1675    16. Related to 15 above: Perl does not treat the | in a conditional group as
1676        creating alternatives. Such a group is treated in the same way as an
1677        ordinary group without any | characters when processing (*THEN). PCRE has
1678        been changed to match Perl's behaviour.
1680    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1681        RunGrepTest script failed.
1683    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1684        inevitable for groups that contain captures, but it can lead to a lot of
1685        stack use in large patterns. The old behaviour has been restored for atomic
1686        groups that do not contain any capturing parentheses.
1688    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1689        suppress the check for a minimum subject length at run time. (If it was
1690        given to pcre_exec() or pcre_dfa_exec() it did work.)
1692    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1693        fail to work when decoding hex characters in data strings in EBCDIC
1694        environments.
1696    21. It appears that in at least one Mac OS environment, the isxdigit() function
1697        is implemented as a macro that evaluates to its argument more than once,
1698        contravening the C 90 Standard (I haven't checked a later standard). There
1699        was an instance in pcretest which caused it to go wrong when processing
1700        \x{...} escapes in subject strings. The has been rewritten to avoid using
1701        things like p++ in the argument of isxdigit().
1704    Version 8.13 16-Aug-2011
1705    ------------------------
1707    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1709    2.  Two minor typos in pcre_internal.h have been fixed.
1711    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1712        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1713        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1715    4.  There were a number of related bugs in the code for matching backrefences
1716        caselessly in UTF-8 mode when codes for the characters concerned were
1717        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1718        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1719        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1720        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1721        2-byte code at the end of the subject (it thought there wasn't enough data
1722        left).
1724    5.  Comprehensive information about what went wrong is now returned by
1725        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1726        as the output vector has at least 2 elements. The offset of the start of
1727        the failing character and a reason code are placed in the vector.
1729    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1730        now returned is for the first byte of the failing character, instead of the
1731        last byte inspected. This is an incompatible change, but I hope it is small
1732        enough not to be a problem. It makes the returned offset consistent with
1733        pcre_exec() and pcre_dfa_exec().
1735    7.  pcretest now gives a text phrase as well as the error number when
1736        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1737        failure, the offset and reason code are output.
1739    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1740        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1741        back over a single character (\n). This seems wrong (because it treated the
1742        two characters as a single entity when going forwards), conflicts with the
1743        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1744        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1745        has been changed.
1747    9.  Some internal refactoring has changed the processing so that the handling
1748        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1749        time (the PCRE_DOTALL option was changed this way some time ago: version
1750        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1751        which was always a bit of a fudge. It also means that there is one less
1752        argument for the match() function, which reduces its stack requirements
1753        slightly. This change also fixes an incompatibility with Perl: the pattern
1754        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1756    10. More internal refactoring has drastically reduced the number of recursive
1757        calls to match() for possessively repeated groups such as (abc)++ when
1758        using pcre_exec().
1760    11. While implementing 10, a number of bugs in the handling of groups were
1761        discovered and fixed:
1763        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1764        (a|)*(?1) gave a compile-time internal error.
1765        ((a|)+)+  did not notice that the outer group could match an empty string.
1766        (^a|^)+   was not marked as anchored.
1767        (.*a|.*)+ was not marked as matching at start or after a newline.
1769    12. Yet more internal refactoring has removed another argument from the match()
1770        function. Special calls to this function are now indicated by setting a
1771        value in a variable in the "match data" data block.
1773    13. Be more explicit in pcre_study() instead of relying on "default" for
1774        opcodes that mean there is no starting character; this means that when new
1775        ones are added and accidentally left out of pcre_study(), testing should
1776        pick them up.
1778    14. The -s option of pcretest has been documented for ages as being an old
1779        synonym of -m (show memory usage). I have changed it to mean "force study
1780        for every regex", that is, assume /S for every regex. This is similar to -i
1781        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1782        using it. It makes it easier to run collections of tests with and without
1783        study enabled, and thereby test pcre_study() more easily. All the standard
1784        tests are now run with and without -s (but some patterns can be marked as
1785        "never study" - see 20 below).
1787    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1788        restoration of the capturing data to the outer values was not happening
1789        correctly.
1791    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1792        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1793        pattern had matched an empty string, and so incorrectly returned a no
1794        match.
1796    17. There was optimizing code for the last branch of non-capturing parentheses,
1797        and also for the obeyed branch of a conditional subexpression, which used
1798        tail recursion to cut down on stack usage. Unfortunately, now that there is
1799        the possibility of (*THEN) occurring in these branches, tail recursion is
1800        no longer possible because the return has to be checked for (*THEN). These
1801        two optimizations have therefore been removed. [But see 8.20/11 above.]
1803    18. If a pattern containing \R was studied, it was assumed that \R always
1804        matched two bytes, thus causing the minimum subject length to be
1805        incorrectly computed because \R can also match just one byte.
1807    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1808        was incorrectly computed.
1810    20. If /S is present twice on a test pattern in pcretest input, it now
1811        *disables* studying, thereby overriding the use of -s on the command line
1812        (see 14 above). This is necessary for one or two tests to keep the output
1813        identical in both cases.
1815    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1816        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1818    22. When an atomic group that contained a capturing parenthesis was
1819        successfully matched, but the branch in which it appeared failed, the
1820        capturing was not being forgotten if a higher numbered group was later
1821        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1822        group 1 to "a", when in fact it should be unset. This applied to multi-
1823        branched capturing and non-capturing groups, repeated or not, and also to
1824        positive assertions (capturing in negative assertions does not happen
1825        in PCRE) and also to nested atomic groups.
1827    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1828        subject after a captured substring, to make it easier to tell which of a
1829        number of identical substrings has been captured.
1831    24. The way atomic groups are processed by pcre_exec() has been changed so that
1832        if they are repeated, backtracking one repetition now resets captured
1833        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1834        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1835        "aaa". Previously, it would have been "a". As part of this code
1836        refactoring, the way recursive calls are handled has also been changed.
1838    25. If an assertion condition captured any substrings, they were not passed
1839        back unless some other capturing happened later. For example, if
1840        (?(?=(a))a) was matched against "a", no capturing was returned.
1842    26. When studying a pattern that contained subroutine calls or assertions,
1843        the code for finding the minimum length of a possible match was handling
1844        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1845        group 1 called group 2 while simultaneously a separate group 2 called group
1846        1). A stack overflow occurred in this case. I have fixed this by limiting
1847        the recursion depth to 10.
1849    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1850        Fortmann. This supports explicit test numbers on the command line, and has
1851        argument validation and error reporting.
1853    28. An instance of \X with an unlimited repeat could fail if at any point the
1854        first character it looked at was a mark character.
1856    29. Some minor code refactoring concerning Unicode properties and scripts
1857        should reduce the stack requirement of match() slightly.
1859    30. Added the '=' option to pcretest to check the setting of unused capturing
1860        slots at the end of the pattern, which are documented as being -1, but are
1861        not included in the return count.
1863    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1864        compiled something random. Now it gives a compile-time error (as does
1865        Perl).
1867    32. A *MARK encountered during the processing of a positive assertion is now
1868        recorded and passed back (compatible with Perl).
1870    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1871        had alternative anchored branches, the search for a second match in a line
1872        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1873        matched the line "0102" twice. The same bug affected patterns that started
1874        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1875        twice.
1877    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1878        does, and because of capturing effects, quantifying parenthesized
1879        assertions may at times be useful. Quantifiers are now allowed for
1880        parenthesized assertions.
1882    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1884    36. \g was being checked for fancy things in a character class, when it should
1885        just be a literal "g".
1887    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1888        appearance of a nested POSIX class supersedes an apparent external class.
1889        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1890        unescaped square brackets may also appear as part of class names. For
1891        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1892        more like Perl. (But see 8.20/1 above.)
1894    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1895        was because it thought it was \N{name}, which is not supported).
1897    39. Add minix to OS list not supporting the -S option in pcretest.
1899    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1900        cannot analyze patterns in sufficient detail to catch mutual recursions
1901        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1902        subgroup is called recursively as a subpattern for a second time at the
1903        same position in the subject string. In previous releases this might have
1904        been caught by the recursion limit, or it might have run out of stack.
1906    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1907        happen only once. PCRE was, however incorrectly giving a compile time error
1908        "recursive call could loop indefinitely" because it cannot analyze the
1909        pattern in sufficient detail. The compile time test no longer happens when
1910        PCRE is compiling a conditional subpattern, but actual runaway loops are
1911        now caught at runtime (see 40 above).
1913    42. It seems that Perl allows any characters other than a closing parenthesis
1914        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1915        has been changed to be the same.
1917    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1918        as not to get warnings when autogen.sh is called. Also changed
1919        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1921    44. To help people who use pcregrep to scan files containing exceedingly long
1922        lines, the following changes have been made:
1924        (a) The default value of the buffer size parameter has been increased from
1925            8K to 20K. (The actual buffer used is three times this size.)
1927        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1928            PCRE is built.
1930        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1931            to be set at run time.
1933        (d) Numerical values in pcregrep options can be followed by K or M, for
1934            example --buffer-size=50K.
1936        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1937            given and the return code is set to 2.
1939    45. Add a pointer to the latest mark to the callout data block.
1941    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1942        partial match of an empty string instead of no match. This was specific to
1943        the use of ".".
1945    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1946        complete match instead of a partial match. This bug was dependent on both
1947        the PCRE_UTF8 and PCRE_DOTALL options being set.
1949    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1950        starting byte set, because \b was not being ignored.
1953    Version 8.12 15-Jan-2011
1954    ------------------------
1956    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1957        checks for such things as part of the documentation building process.
1959    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1960        --match-limit and --recursion-limit options (added for 8.11). In
1961        particular, this made one of the standard tests fail. (The integer value
1962        went into the wrong half of a long int.)
1964    3.  If the --colour option was given to pcregrep with -v (invert match), it
1965        did strange things, either producing crazy output, or crashing. It should,
1966        of course, ignore a request for colour when reporting lines that do not
1967        match.
1969    4.  Another pcregrep bug caused similar problems if --colour was specified with
1970        -M (multiline) and the pattern match finished with a line ending.
1972    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1973        matched in multiline mode, the following line was shown as part of the
1974        match. This seems wrong, so I have changed it.
1976    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1977        the check for further matches in the same line (so they could be coloured)
1978        to overrun the end of the current line. If another match was found, it was
1979        incorrectly shown (and then shown again when found in the next line).
1981    7.  If pcregrep was compiled under Windows, there was a reference to the
1982        function pcregrep_exit() before it was defined. I am assuming this was
1983        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1984        reported by a user. I've moved the definition above the reference.
1987    Version 8.11 10-Dec-2010
1988    ------------------------
1990    1.  (*THEN) was not working properly if there were untried alternatives prior
1991        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1992        backtracked to try for "b" instead of moving to the next alternative branch
1993        at the same level (in this case, to look for "c"). The Perl documentation
1994        is clear that when (*THEN) is backtracked onto, it goes to the "next
1995        alternative in the innermost enclosing group".
1997    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1998        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1999        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
2000        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
2001        (*THEN).
2003    3.  If \s appeared in a character class, it removed the VT character from
2004        the class, even if it had been included by some previous item, for example
2005        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
2006        of \s, but is part of the POSIX "space" class.)
2008    4.  A partial match never returns an empty string (because you can always
2009        match an empty string at the end of the subject); however the checking for
2010        an empty string was starting at the "start of match" point. This has been
2011        changed to the "earliest inspected character" point, because the returned
2012        data for a partial match starts at this character. This means that, for
2013        example, /(?<=abc)def/ gives a partial match for the subject "abc"
2014        (previously it gave "no match").
2016    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
2017        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
2018        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
2019        has an implication that the given string is incomplete (because a partial
2020        match is preferred over a full match). For this reason, these items now
2021        give a partial match in this situation. [Aside: previously, the one case
2022        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
2023        match rather than a full match, which was wrong by the old rules, but is
2024        now correct.]
2026    6.  There was a bug in the handling of #-introduced comments, recognized when
2027        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
2028        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
2029        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
2030        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
2031        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
2032        places in pcre_compile().
2034    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
2035        comments when looking ahead for named forward references to subpatterns,
2036        the only newline sequence it recognized was NL. It now handles newlines
2037        according to the set newline convention.
2039    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
2040        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
2041        cater for a lack of strerror(). These oversights have been fixed.
2043    9.  Added --match-limit and --recursion-limit to pcregrep.
2045    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
2047    11. When the -o option was used, pcregrep was setting a return code of 1, even
2048        when matches were found, and --line-buffered was not being honoured.
2050    12. Added an optional parentheses number to the -o and --only-matching options
2051        of pcregrep.
2053    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2054        can match an empty string. The code to do it in pcretest and pcredemo
2055        needed fixing:
2057        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2058            only one byte after an empty string match just before CRLF (this case
2059            just got forgotten; "any" and "anycrlf" were OK).
2061        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2062            mode when an empty string match preceded an ASCII character followed by
2063            a non-ASCII character. (The code for advancing by one character rather
2064            than one byte was nonsense.)
2066        (c) The pcredemo.c sample program did not have any code at all to handle
2067            the cases when CRLF is a valid newline sequence.
2069    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2070        as a starting offset was within the subject string. There is now a new
2071        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2072        negative or greater than the length of the string. In order to test this,
2073        pcretest is extended to allow the setting of negative starting offsets.
2075    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2076        starting offset points to the beginning of a UTF-8 character was
2077        unnecessarily clumsy. I tidied it up.
2079    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2080        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2082    17. Nobody had reported that the --include_dir option, which was added in
2083        release 7.7 should have been called --include-dir (hyphen, not underscore)
2084        for compatibility with GNU grep. I have changed it to --include-dir, but
2085        left --include_dir as an undocumented synonym, and the same for
2086        --exclude-dir, though that is not available in GNU grep, at least as of
2087        release 2.5.4.
2089    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2090        characters from a string of bytes) have been redefined so as not to use
2091        loops, in order to improve performance in some environments. At the same
2092        time, I abstracted some of the common code into auxiliary macros to save
2093        repetition (this should not affect the compiled code).
2095    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2096        compile-time error is now given if \c is not followed by an ASCII
2097        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2098        different, and any byte value is allowed.)
2100    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2101        START_OPTIMIZE option, which is now allowed at compile time - but just
2102        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2103        to pcregrep and other applications that have no direct access to PCRE
2104        options. The new /Y option in pcretest sets this option when calling
2105        pcre_compile().
2107    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2108        back references. Groups containing recursive back references were forced to
2109        be atomic by that change, but in the case of named groups, the amount of
2110        memory required was incorrectly computed, leading to "Failed: internal
2111        error: code overflow". This has been fixed.
2113    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2114        pcretest.c, to avoid build problems in some Borland environments.
2117    Version 8.10 25-Jun-2010
2118    ------------------------
2120    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2121        THEN.
2123    2.  (*ACCEPT) was not working when inside an atomic group.
2125    3.  Inside a character class, \B is treated as a literal by default, but
2126        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2127        causes the error). The code is unchanged, but I tidied the documentation.
2129    4.  Inside a character class, PCRE always treated \R and \X as literals,
2130        whereas Perl faults them if its -w option is set. I have changed PCRE so
2131        that it faults them when PCRE_EXTRA is set.
2133    5.  Added support for \N, which always matches any character other than
2134        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2136    6.  When compiling pcregrep with newer versions of gcc which may have
2137        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2138        declared with attribute warn_unused_result" were given. Just casting the
2139        result to (void) does not stop the warnings; a more elaborate fudge is
2140        needed. I've used a macro to implement this.
2142    7.  Minor change to pcretest.c to avoid a compiler warning.
2144    8.  Added four artifical Unicode properties to help with an option to make
2145        \s etc use properties (see next item). The new properties are: Xan
2146        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2148    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2149        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2150        this option. Modified pcretest to add /W to test this facility. Added
2151        REG_UCP to make it available via the POSIX interface.
2153    10. Added --line-buffered to pcregrep.
2155    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2156        studied, and the match started with a letter with a code point greater than
2157        127 whose first byte was different to the first byte of the other case of
2158        the letter, the other case of this starting letter was not recognized
2159        (#976).
2161    12. If a pattern that was studied started with a repeated Unicode property
2162        test, for example, \p{Nd}+, there was the theoretical possibility of
2163        setting up an incorrect bitmap of starting bytes, but fortunately it could
2164        not have actually happened in practice until change 8 above was made (it
2165        added property types that matched character-matching opcodes).
2167    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2168        possible starting bytes for non-anchored patterns.
2170    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2171        \R, and also a number of cases that involve Unicode properties, both
2172        explicit and implicit when PCRE_UCP is set.
2174    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2175        input, it could crash or give wrong results if characters with values
2176        greater than 0xc0 were present in the subject string. (Detail: it assumed
2177        UTF-8 input when processing these items.)
2179    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2180        size_t is 64-bit (#991).
2182    17. Added a check for running out of memory when PCRE is compiled with
2183        --disable-stack-for-recursion (#990).
2185    18. If the last data line in a file for pcretest does not have a newline on
2186        the end, a newline was missing in the output.
2188    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2189        less than 128) in its various bitmaps. However, there is a facility for
2190        generating tables according to the current locale when PCRE is compiled. It
2191        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2192        characters, are recognized by isspace() and therefore were getting set in
2193        these tables, and indeed these tables seem to approximate to ISO 8859. This
2194        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2195        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2196        which of course cannot start UTF-8 characters. I have changed the code so
2197        that only real ASCII characters (less than 128) and the correct starting
2198        bytes for UTF-8 encodings are set for characters greater than 127 when in
2199        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2200        altogether.)
2202    20. Added the /T option to pcretest so as to be able to run tests with non-
2203        standard character tables, thus making it possible to include the tests
2204        used for 19 above in the standard set of tests.
2206    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2207        reference to a subpattern the other side of a comment that contains an
2208        opening parenthesis caused either an internal compiling error, or a
2209        reference to the wrong subpattern.
2212    Version 8.02 19-Mar-2010
2213    ------------------------
2215    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2217    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2218        configured.
2220    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2221        original author of that file, following a query about its status.
2223    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2224        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2226    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2227        quantifier applied to a forward-referencing subroutine call, could compile
2228        incorrect code or give the error "internal error: previously-checked
2229        referenced subpattern not found".
2231    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2232        variables to point to external functions. For these systems, therefore,
2233        pcre_malloc etc. are now initialized to local functions that call the
2234        relevant global functions.
2236    7.  There were two entries missing in the vectors called coptable and poptable
2237        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2238        I've fixed the data, and added a kludgy way of testing at compile time that
2239        the lengths are correct (equal to the number of opcodes).
2241    8.  Following on from 7, I added a similar kludge to check the length of the
2242        eint vector in pcreposix.c.
2244    9.  Error texts for pcre_compile() are held as one long string to avoid too
2245        much relocation at load time. To find a text, the string is searched,
2246        counting zeros. There was no check for running off the end of the string,
2247        which could happen if a new error number was added without updating the
2248        string.
2250    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2252    11. \K was not working if it appeared in an atomic group or in a group that
2253        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2254        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2255        the assertion is positive, but not if it is negative.
2257    12. Change 11 fortuitously reduced the size of the stack frame used in the
2258        "match()" function of pcre_exec.c by one pointer. Forthcoming
2259        implementation of support for (*MARK) will need an extra pointer on the
2260        stack; I have reserved it now, so that the stack frame size does not
2261        decrease.
2263    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2264        item in branch that calls a recursion is a subroutine call - as in the
2265        second branch in the above example - was incorrectly given the compile-
2266        time error "recursive call could loop indefinitely" because pcre_compile()
2267        was not correctly checking the subroutine for matching a non-empty string.
2269    14. The checks for overrunning compiling workspace could trigger after an
2270        overrun had occurred. This is a "should never occur" error, but it can be
2271        triggered by pathological patterns such as hundreds of nested parentheses.
2272        The checks now trigger 100 bytes before the end of the workspace.
2274    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2277    Version 8.01 19-Jan-2010
2278    ------------------------
2280    1.  If a pattern contained a conditional subpattern with only one branch (in
2281        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2282        computed the wrong minimum data length (which is of course zero for such
2283        subpatterns). This could cause incorrect "no match" results.
2285    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2286        the pattern is reset in the first branch, pcre_compile() failed with
2287        "internal error: code overflow at offset...". This happened only when
2288        the reset was to the original external option setting. (An optimization
2289        abstracts leading options settings into an external setting, which was the
2290        cause of this.)
2292    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2293        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2294        assertion pattern did not match (meaning that the assertion was true), it
2295        was incorrectly treated as false if the SKIP had been reached during the
2296        matching. This also applied to assertions used as conditions.
2298    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2299        assertion subpattern, including such a pattern used as a condition,
2300        unpredictable results occurred, instead of the error return
2303    5.  The C++ GlobalReplace function was not working like Perl for the special
2304        situation when an empty string is matched. It now does the fancy magic
2305        stuff that is necessary.
2307    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2308        removed. (These were left over from very, very early versions of PCRE.)
2310    7.  Some cosmetic changes to the code to make life easier when compiling it
2311        as part of something else:
2313        (a) Change DEBUG to PCRE_DEBUG.
2315        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2316            called "current" as "current_branch", to prevent a collision with the
2317            Linux macro when compiled as a kernel module.
2319        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2320            prevent a collision with the Linux macro when compiled as a kernel
2321            module.
2323    8.  In pcre_compile() there are some checks for integer overflows that used to
2324        cast potentially large values to (double). This has been changed to that
2325        when building, a check for int64_t is made, and if it is found, it is used
2326        instead, thus avoiding the use of floating point arithmetic. (There is no
2327        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2328        double.
2330    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2331        2005 (difference between two addresses compared to an unsigned value).
2333    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2334        custom one, because of the following reported problem in Windows:
2336          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2337              under Win32.
2338          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2339              therefore missing the function definition.
2340          - The compiler thus generates a "C" signature for the test function.
2341          - The linker fails to find the "C" function.
2342          - PCRE fails to configure if asked to do so against libbz2.
2344    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2345        messages were output:
2347          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2348          rerunning libtoolize, to keep the correct libtool macros in-tree.
2349          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2351        I have done both of these things.
2353    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2354        most of the time, it *can* run out if it is given a pattern that contains a
2355        runaway infinite recursion. I updated the discussion in the pcrestack man
2356        page.
2358    13. Now that we have gone to the x.xx style of version numbers, the minor
2359        version may start with zero. Using 08 or 09 is a bad idea because users
2360        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2361        interpreted as invalid octal numbers. I've updated the previous comment in
2362        configure.ac, and also added a check that gives an error if 08 or 09 are
2363        used.
2365    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2366        causing partial matching to fail when the end of the subject matched \W
2367        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2369    15. There were some discrepancies between the declarations in pcre_internal.h
2370        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2371        their definitions. The declarations used "const uschar *" and the
2372        definitions used USPTR. Even though USPTR is normally defined as "const
2373        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2374        reported that: "This difference in casting confuses some C++ compilers, for
2375        example, SunCC recognizes above declarations as different functions and
2376        generates broken code for hbpcre." I have changed the declarations to use
2377        USPTR.
2379    16. GNU libtool is named differently on some systems. The autogen.sh script now
2380        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2381        (FreeBSD).
2383    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2384        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2385        comment: "Figure out how to create a longlong from a string: strtoll and
2386        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2387        instance, but it only takes 2 args instead of 3!"
2389    18. A subtle bug concerned with back references has been fixed by a change of
2390        specification, with a corresponding code fix. A pattern such as
2391        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2392        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2393        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2394        same bug. Such groups have to be quantified to be useful, or contained
2395        inside another quantified group. (If there's no repetition, the reference
2396        can never match.) The problem arises because, having left the group and
2397        moved on to the rest of the pattern, a later failure that backtracks into
2398        the group uses the captured value from the final iteration of the group
2399        rather than the correct earlier one. I have fixed this in PCRE by forcing
2400        any group that contains a reference to itself to be an atomic group; that
2401        is, there cannot be any backtracking into it once it has completed. This is
2402        similar to recursive and subroutine calls.
2405    Version 8.00 19-Oct-09
2406    ----------------------
2408    1.  The table for translating pcre_compile() error codes into POSIX error codes
2409        was out-of-date, and there was no check on the pcre_compile() error code
2410        being within the table. This could lead to an OK return being given in
2411        error.
2413    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2414        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2415        in a Windows environment.
2417    3.  The pcregrep --count option prints the count for each file even when it is
2418        zero, as does GNU grep. However, pcregrep was also printing all files when
2419        --files-with-matches was added. Now, when both options are given, it prints
2420        counts only for those files that have at least one match. (GNU grep just
2421        prints the file name in this circumstance, but including the count seems
2422        more useful - otherwise, why use --count?) Also ensured that the
2423        combination -clh just lists non-zero counts, with no names.
2425    4.  The long form of the pcregrep -F option was incorrectly implemented as
2426        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2427        but it seems right to fix it, and I didn't think it was worth preserving
2428        the old behaviour.
2430    5.  The command line items --regex=pattern and --regexp=pattern were not
2431        recognized by pcregrep, which required --regex pattern or --regexp pattern
2432        (with a space rather than an '='). The man page documented the '=' forms,
2433        which are compatible with GNU grep; these now work.
2435    6.  No libpcreposix.pc file was created for pkg-config; there was just
2436        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2438    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2439        when UCP support is not needed, by modifying the Python script that
2440        generates it from Unicode data files. This should not matter if the module
2441        is correctly used as a library, but I received one complaint about 50K of
2442        unwanted data. My guess is that the person linked everything into his
2443        program rather than using a library. Anyway, it does no harm.
2445    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2446        was a minimum greater than 1 for a wide character in a possessive
2447        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2448        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2449        character. Chaos in the form of incorrect output or a compiling loop could
2450        result.
2452    9.  The restrictions on what a pattern can contain when partial matching is
2453        requested for pcre_exec() have been removed. All patterns can now be
2454        partially matched by this function. In addition, if there are at least two
2455        slots in the offset vector, the offset of the earliest inspected character
2456        for the match and the offset of the end of the subject are set in them when
2457        PCRE_ERROR_PARTIAL is returned.
2459    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2460        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2461        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2462        and may be more useful for multi-segment matching.
2464    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2465        used to be given if ever the end of the subject was reached; now it is
2466        given only if matching could not proceed because another character was
2467        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2468        string "Z", which now yields "no match" instead of "partial match". In the
2469        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2470        final character ended with (*FAIL).
2472    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2473        if the pattern had a "must contain" character that was already found in the
2474        earlier partial match, unless partial matching was again requested. For
2475        example, with the pattern /dog.(body)?/, the "must contain" character is
2476        "g". If the first part-match was for the string "dog", restarting with
2477        "sbody" failed. This bug has been fixed.
2479    13. The string returned by pcre_dfa_exec() after a partial match has been
2480        changed so that it starts at the first inspected character rather than the
2481        first character of the match. This makes a difference only if the pattern
2482        starts with a lookbehind assertion or \b or \B (\K is not supported by
2483        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2484        matching functions compatible, and I think it's the right thing to do.
2486    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2487        so that the demonstration program is easily available in environments where
2488        PCRE has not been installed from source.
2490    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2491        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2492        library.
2494    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2495        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2496        is not the first non-POSIX option to be added. Clearly some people find
2497        these options useful.
2499    17. If a caller to the POSIX matching function regexec() passes a non-zero
2500        value for nmatch with a NULL value for pmatch, the value of
2501        nmatch is forced to zero.
2503    18. RunGrepTest did not have a test for the availability of the -u option of
2504        the diff command, as RunTest does. It now checks in the same way as
2505        RunTest, and also checks for the -b option.
2507    19. If an odd number of negated classes containing just a single character
2508        interposed, within parentheses, between a forward reference to a named
2509        subpattern and the definition of the subpattern, compilation crashed with
2510        an internal error, complaining that it could not find the referenced
2511        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2512        [The bug was that it was starting one character too far in when skipping
2513        over the character class, thus treating the ] as data rather than
2514        terminating the class. This meant it could skip too much.]
2516    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2517        /g option in pcretest when the pattern contains \K, which makes it possible
2518        to have an empty string match not at the start, even when the pattern is
2519        anchored. Updated pcretest and pcredemo to use this option.
2521    21. If the maximum number of capturing subpatterns in a recursion was greater
2522        than the maximum at the outer level, the higher number was returned, but
2523        with unset values at the outer level. The correct (outer level) value is
2524        now given.
2526    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2527        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2528        make it do so. The string so far is captured, making this feature
2529        compatible with Perl.
2531    23. The tests have been re-organized, adding tests 11 and 12, to make it
2532        possible to check the Perl 5.10 features against Perl 5.10.
2534    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2535        pattern matches a fixed length string. PCRE did not allow this; now it
2536        does. Neither allows recursion.
2538    25. I finally figured out how to implement a request to provide the minimum
2539        length of subject string that was needed in order to match a given pattern.
2540        (It was back references and recursion that I had previously got hung up
2541        on.) This code has now been added to pcre_study(); it finds a lower bound
2542        to the length of subject needed. It is not necessarily the greatest lower
2543        bound, but using it to avoid searching strings that are too short does give
2544        some useful speed-ups. The value is available to calling programs via
2545        pcre_fullinfo().
2547    26. While implementing 25, I discovered to my embarrassment that pcretest had
2548        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2549        study optimizations had never been tested with that matching function.
2550        Oops. What is worse, even when it was passed study data, there was a bug in
2551        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2552        were also very few tests of studied patterns with pcre_dfa_exec().
2554    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2555        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2556        on the other side of the coin, they are no longer allowed to have different
2557        names, because these cannot be distinguished in PCRE, and this has caused
2558        confusion. (This is a difference from Perl.)
2560    28. When duplicate subpattern names are present (necessarily with different
2561        numbers, as required by 27 above), and a test is made by name in a
2562        conditional pattern, either for a subpattern having been matched, or for
2563        recursion in such a pattern, all the associated numbered subpatterns are
2564        tested, and the overall condition is true if the condition is true for any
2565        one of them. This is the way Perl works, and is also more like the way
2566        testing by number works.
2569    Version 7.9 11-Apr-09
2570    ---------------------
2572    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2573        (pcretest), all targets were linked against these libraries. This included
2574        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2575        libraries. This caused unwanted dependencies to be created. This problem
2576        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2577        pcretest is linked with readline.
2579    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2580        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2581        moved outside it again, because FALSE and TRUE are already defined in AIX,
2582        but BOOL is not.
2584    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2585        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2587    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2588        hyphens) following file names and line numbers when outputting matching
2589        lines. This is not true; no spaces are inserted. I have also clarified the
2590        wording for the --colour (or --color) option.
2592    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2593        was not coloured; this is different to GNU grep, so I have changed it to be
2594        the same.
2596    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2597        each matching line was coloured. Now it goes on to look for further matches
2598        of any of the test patterns, which is the same behaviour as GNU grep.
2600    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2601        doesn't make sense to accept an empty string match in pcregrep, so I have
2602        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2603        seems to be how GNU grep behaves. [But see later change 40 for release
2604        8.33.]
2606    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2607        start or after a newline", because the conditional assertion was not being
2608        correctly handled. The rule now is that both the assertion and what follows
2609        in the first alternative must satisfy the test.
2611    9.  If auto-callout was enabled in a pattern with a conditional group whose
2612        condition was an assertion, PCRE could crash during matching, both with
2613        pcre_exec() and pcre_dfa_exec().
2615    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2616        used for matching.
2618    11. Unicode property support in character classes was not working for
2619        characters (bytes) greater than 127 when not in UTF-8 mode.
2621    12. Added the -M command line option to pcretest.
2623    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2625    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2627    16. Added comments and documentation about mis-use of no_arg in the C++
2628        wrapper.
2630    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2631        from Martin Jerabek that uses macro names for all relevant character and
2632        string constants.
2634    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2635        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2636        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2637        these, but not everybody uses configure.
2639    19. A conditional group that had only one branch was not being correctly
2640        recognized as an item that could match an empty string. This meant that an
2641        enclosing group might also not be so recognized, causing infinite looping
2642        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2643        with the subject "ab", where knowledge that the repeated group can match
2644        nothing is needed in order to break the loop.
2646    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2647        exec(), but without supplying a callout function, matching went wrong.
2649    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2650        leak if the size of the offset vector was greater than 30. When the vector
2651        is smaller, the saved offsets during recursion go onto a local stack
2652        vector, but for larger vectors malloc() is used. It was failing to free
2653        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2654        error, in fact).
2656    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2657        heapframe that is used only when UTF-8 support is enabled. This caused no
2658        problem, but was untidy.
2660    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2661        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2662        included within another project.
2664    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2665        slightly modified by me:
2667          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2668              not building pcregrep.
2670          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2671              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2673    25. Forward references, both numeric and by name, in patterns that made use of
2674        duplicate group numbers, could behave incorrectly or give incorrect errors,
2675        because when scanning forward to find the reference group, PCRE was not
2676        taking into account the duplicate group numbers. A pattern such as
2677        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2679    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2680        the feature of a custom pointer more persuasive (as requested by a user).
2682    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2683        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2684        pre-defined.
2686    28. Added support for (*UTF8) at the start of a pattern.
2688    29. Arrange for flags added by the "release type" setting in CMake to be shown
2689        in the configuration summary.
2692    Version 7.8 05-Sep-08
2693    ---------------------
2695    1.  Replaced UCP searching code with optimized version as implemented for Ad
2696        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2697        stage table and inline lookup instead of a function, giving speed ups of 2
2698        to 5 times on some simple patterns that I tested. Permission was given to
2699        distribute the MultiStage2.py script that generates the tables (it's not in
2700        the tarball, but is in the Subversion repository).
2702    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2703        scripts.
2705    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2706        a group with a zero qualifier. The result of the study could be incorrect,
2707        or the function might crash, depending on the pattern.
2709    4.  Caseless matching was not working for non-ASCII characters in back
2710        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2711        It now works when Unicode Property Support is available.
2713    5.  In pcretest, an escape such as \x{de} in the data was always generating
2714        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2715        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2716        truncation.
2718    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2720    7.  Added two (int) casts to pcregrep when printing the difference of two
2721        pointers, in case they are 64-bit values.
2723    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2724        test 2 if it fails.
2726    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2727        and a #define of that name to empty if it is not externally set. This is to
2728        allow users of MSVC to set it if necessary.
2730    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2731        the convenience functions in the pcre_get.c source file.
2733    11. An option change at the start of a pattern that had top-level alternatives
2734        could cause overwriting and/or a crash. This command provoked a crash in
2735        some environments:
2737          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2739        This potential security problem was recorded as CVE-2008-2371.
2741    12. For a pattern where the match had to start at the beginning or immediately
2742        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2743        pcre_dfa_exec() could read past the end of the passed subject if there was
2744        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2745        pcretest so that it places the subject at the end of its malloc-ed buffer.
2747    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2748        exec() might read past the end of the data buffer in UTF-8 mode.
2750    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2751        the data contained the byte 0x85 as part of a UTF-8 character within its
2752        first line. This applied both to normal and DFA matching.
2754    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2755        /^[^d]*?$/8 failed to match "abc".
2757    16. Added a missing copyright notice to pcrecpp_internal.h.
2759    17. Make it more clear in the documentation that values returned from
2760        pcre_exec() in ovector are byte offsets, not character counts.
2762    18. Tidied a few places to stop certain compilers from issuing warnings.
2764    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2765        supplied by Stefan Weber. I made a further small update for 7.8 because
2766        there is a change of source arrangements: the pcre_searchfuncs.c module is
2767        replaced by pcre_ucd.c.
2770    Version 7.7 07-May-08
2771    ---------------------
2773    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2774        a string to a long long, pretend we don't even have a long long." This is
2775        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2777    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2778        pre-7.6 versions, which defined a global no_arg variable instead of putting
2779        it in the RE class. (See also #8 below.)
2781    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2782        Lopes.
2784    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2786        (1) The include/exclude patterns were being applied to the whole pathnames
2787            of files, instead of just to the final components.
2789        (2) If there was more than one level of directory, the subdirectories were
2790            skipped unless they satisfied the include/exclude conditions. This is
2791            inconsistent with GNU grep (and could even be seen as contrary to the
2792            pcregrep specification - which I improved to make it absolutely clear).
2793            The action now is always to scan all levels of directory, and just
2794            apply the include/exclude patterns to regular files.
2796    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2797        --exclude_dir in the tests to avoid scanning .svn directories.
2799    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2800        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2801        doesn't support NULs in patterns.
2803    7.  Added some missing "const"s to declarations of static tables in
2804        pcre_compile.c and pcre_dfa_exec.c.
2806    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2807        caused by fix #2  above. (Subsequently also a second patch to fix the
2808        first patch. And a third patch - this was a messy problem.)
2810    9.  Applied Craig's patch to remove the use of push_back().
2812    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2813        matching function regexec().
2815    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2816        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2817        references. PCRE supports relative numbers with this syntax (I don't think
2818        Oniguruma does).
2820    12. Previously, a group with a zero repeat such as (...){0} was completely
2821        omitted from the compiled regex. However, this means that if the group
2822        was called as a subroutine from elsewhere in the pattern, things went wrong
2823        (an internal error was given). Such groups are now left in the compiled
2824        pattern, with a new opcode that causes them to be skipped at execution
2825        time.
2827    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2828        to the way PCRE behaves:
2830        (a) A lone ] character is dis-allowed (Perl treats it as data).
2832        (b) A back reference to an unmatched subpattern matches an empty string
2833            (Perl fails the current match path).
2835        (c) A data ] in a character class must be notated as \] because if the
2836            first data character in a class is ], it defines an empty class. (In
2837            Perl it is not possible to have an empty class.) The empty class []
2838            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2839            The negative empty class [^] matches any one character, independently
2840            of the DOTALL setting.
2842    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2843        non-existent subpattern following a character class starting with ']' and
2844        containing () gave an internal compiling error instead of "reference to
2845        non-existent subpattern". Fortunately, when the pattern did exist, the
2846        compiled code was correct. (When scanning forwards to check for the
2847        existence of the subpattern, it was treating the data ']' as terminating
2848        the class, so got the count wrong. When actually compiling, the reference
2849        was subsequently set up correctly.)
2851    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2852        it was being rejected as not supported by pcre_dfa_exec(), even though
2853        other assertions are supported. I have made pcre_dfa_exec() support
2854        (*FAIL).
2856    16. The implementation of 13c above involved the invention of a new opcode,
2857        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2858        cannot be changed at match time, I realized I could make a small
2859        improvement to matching performance by compiling OP_ALLANY instead of
2860        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2861        on the OP_ANY path.
2863    17. Compiling pcretest on Windows with readline support failed without the
2864        following two fixes: (1) Make the unistd.h include conditional on
2865        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2867    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2868        ncurses library to be included for pcretest when ReadLine support is
2869        requested, but also to allow for it to be overridden. This patch came from
2870        Daniel Bergström.
2872    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2873        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2874        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2875        spotting this.
2878    Version 7.6 28-Jan-08
2879    ---------------------
2881    1.  A character class containing a very large number of characters with
2882        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2883        overflow.
2885    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2886        HAVE_LONG_LONG is not defined.
2888    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2889        bring them up to date and include new features. This patch includes:
2891        - Fixed PH's badly added libz and libbz2 support.
2892        - Fixed a problem with static linking.
2893        - Added pcredemo. [But later removed - see 7 below.]
2894        - Fixed dftables problem and added an option.
2895        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2896            HAVE_LONG_LONG.
2897        - Added readline support for pcretest.
2898        - Added an listing of the option settings after cmake has run.
2900    4.  A user submitted a patch to Makefile that makes it easy to create
2901        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2902        Makefile.am that cause it to include this special target, without
2903        affecting anything else. Note that the same mingw target plus all
2904        the other distribution libraries and programs are now supported
2905        when configuring with CMake (see 6 below) instead of with
2906        Configure/Make.
2908    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2909        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2910        exported in the Windows port". It has not yet been confirmed that the patch
2911        solves the problem, but it does no harm.
2913    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2914        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2915        with CMake, and also correct the comment about stack recursion.
2917    7.  Remove the automatic building of pcredemo from the ./configure system and
2918        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2919        of a program that users should build themselves after PCRE is installed, so
2920        building it automatically is not really right. What is more, it gave
2921        trouble in some build environments.
2923    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2926    Version 7.5 10-Jan-08
2927    ---------------------
2929    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2930        values in parens when parsing an RE using the C++ wrapper."
2932    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2933        Characters greater than 255 were excluded from the class instead of being
2934        included.
2936    3.  The same bug as (2) above applied to negated POSIX classes such as
2937        [:^space:].
2939    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2940        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2941        I have changed it.
2943    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2944        first named subpattern) and a construct such as (?&a) would reference the
2945        first named subpattern whose name started with "a" (in other words, the
2946        length check was missing). Both these problems are fixed. "Subpattern name
2947        expected" is now given for (?&) (a zero-length name), and this patch also
2948        makes it give the same error for \k'' (previously it complained that that
2949        was a reference to a non-existent subpattern).
2951    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2952        this is right because (?- can be followed by option settings as well as by
2953        digits. I have, however, made the messages clearer.
2955    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2956        than the number used in the conditional) now cause a compile-time error.
2957        This is actually not compatible with Perl, which accepts such patterns, but
2958        treats the conditional as always being FALSE (as PCRE used to), but it
2959        seems to me that giving a diagnostic is better.
2961    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2962        and messages.
2964    9.  Fix two occurrences of "backslash" in comments that should have been
2965        "backspace".
2967    10. Remove two redundant lines of code that can never be obeyed (their function
2968        was moved elsewhere).
2970    11. The program that makes PCRE's Unicode character property table had a bug
2971        which caused it to generate incorrect table entries for sequences of
2972        characters that have the same character type, but are in different scripts.
2973        It amalgamated them into a single range, with the script of the first of
2974        them. In other words, some characters were in the wrong script. There were
2975        thirteen such cases, affecting characters in the following ranges:
2977          U+002b0 - U+002c1
2978          U+0060c - U+0060d
2979          U+0061e - U+00612
2980          U+0064b - U+0065e
2981          U+0074d - U+0076d
2982          U+01800 - U+01805
2983          U+01d00 - U+01d77
2984          U+01d9b - U+01dbf
2985          U+0200b - U+0200f
2986          U+030fc - U+030fe
2987          U+03260 - U+0327f
2988          U+0fb46 - U+0fbb1
2989          U+10450 - U+1049d
2991    12. The -o option (show only the matching part of a line) for pcregrep was not
2992        compatible with GNU grep in that, if there was more than one match in a
2993        line, it showed only the first of them. It now behaves in the same way as
2994        GNU grep.
2996    13. If the -o and -v options were combined for pcregrep, it printed a blank
2997        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2998        does the same. The return code can be used to tell if there were any
2999        non-matching lines.
3001    14. Added --file-offsets and --line-offsets to pcregrep.
3003    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
3004        infinitely looping recursion. The bug was that positive lookaheads were not
3005        being skipped when checking for a possible empty match (negative lookaheads
3006        and both kinds of lookbehind were skipped).
3008    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
3009        inclusion of <windows.h> to before rather than after the definition of
3010        INVALID_FILE_ATTRIBUTES (patch from David Byron).
3012    17. Specifying a possessive quantifier with a specific limit for a Unicode
3013        character property caused pcre_compile() to compile bad code, which led at
3014        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
3015        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
3016        caused the error; without that there was no problem.
3018    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
3020    19. Added --enable-pcretest-libreadline.
3022    20. In pcrecpp.cc, the variable 'count' was incremented twice in
3023        RE::GlobalReplace(). As a result, the number of replacements returned was
3024        double what it should be. I removed one of the increments, but Craig sent a
3025        later patch that removed the other one (the right fix) and added unit tests
3026        that check the return values (which was not done before).
3028    21. Several CMake things:
3030        (1) Arranged that, when cmake is used on Unix, the libraries end up with
3031            the names libpcre and libpcreposix, not just pcre and pcreposix.
3033        (2) The above change means that pcretest and pcregrep are now correctly
3034            linked with the newly-built libraries, not previously installed ones.
3038    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
3039        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
3040        UTF-8 newline character). The key issue is that the pattern starts .*;
3041        this means that the match must be either at the beginning, or after a
3042        newline. The bug was in the code for advancing after a failed match and
3043        checking that the new position followed a newline. It was not taking
3044        account of UTF-8 characters correctly.
3046    23. PCRE was behaving differently from Perl in the way it recognized POSIX
3047        character classes. PCRE was not treating the sequence [:...:] as a
3048        character class unless the ... were all letters. Perl, however, seems to
3049        allow any characters between [: and :], though of course it rejects as
3050        unknown any "names" that contain non-letters, because all the known class
3051        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
3052        for example, whereas PCRE did not - it did not recognize a POSIX character
3053        class. This seemed a bit dangerous, so the code has been changed to be
3054        closer to Perl. The behaviour is not identical to Perl, because PCRE will
3055        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
3056        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
3057        Perl does, and where it didn't before.
3059    24. Rewrite so as to remove the single use of %n from pcregrep because in some
3060        Windows environments %n is disabled by default.
3063    Version 7.4 21-Sep-07
3064    ---------------------
3066    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
3067        means that a class such as [\s] counted as "explicit reference to CR or
3068        LF". That isn't really right - the whole point of the change was to try to
3069        help when there was an actual mention of one of the two characters. So now
3070        the change happens only if \r or \n (or a literal CR or LF) character is
3071        encountered.
3073    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
3074        of both had grown to the point where there were only 3 bits left.
3075        Fortunately, there was spare space in the data structure, and so I have
3076        moved the internal flags into a new 16-bit field to free up more option
3077        bits.
3079    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3080        but did not set the internal JCHANGED flag - either of these is enough to
3081        control the way the "get" function works - but the PCRE_INFO_JCHANGED
3082        facility is supposed to tell if (?J) was ever used, so now (?J) at the
3083        start sets both bits.
3085    4.  Added options (at build time, compile time, exec time) to change \R from
3086        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3088    5.  doc/pcresyntax.html was missing from the distribution.
3090    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3091        compatibility, even though it is no longer used.
3093    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3094        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3095        windows.h file is present (where different names are used). [This was
3096        reversed later after testing - see 16 below.]
3098    8.  Changed all #include <config.h> to #include "config.h". There were also
3099        some further <pcre.h> cases that I changed to "pcre.h".
3101    9.  When pcregrep was used with the --colour option, it missed the line ending
3102        sequence off the lines that it output.
3104    10. It was pointed out to me that arrays of string pointers cause lots of
3105        relocations when a shared library is dynamically loaded. A technique of
3106        using a single long string with a table of offsets can drastically reduce
3107        these. I have refactored PCRE in four places to do this. The result is
3108        dramatic:
3110          Originally:                          290
3111          After changing UCP table:            187
3112          After changing error message table:   43
3113          After changing table of "verbs"       36
3114          After changing table of Posix names   22
3116        Thanks to the folks working on Gregex for glib for this insight.
3118    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3119        unicode-properties was also set.
3121    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3123    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3124        checked only for CRLF.
3126    14. Added casts to pcretest.c to avoid compiler warnings.
3128    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3130    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3131        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3132        entirely. This removes changes made in 7 above.
3134    17. The CMake files have been updated, and there is now more information about
3135        building with CMake in the NON-UNIX-USE document.
3138    Version 7.3 28-Aug-07
3139    ---------------------
3141     1. In the rejigging of the build system that eventually resulted in 7.1, the
3142        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3143        brackets there is not right, since it causes compilers to look for an
3144        installed pcre.h, not the version that is in the source that is being
3145        compiled (which of course may be different). I have changed it back to:
3147          #include "pcre.h"
3149        I have a vague recollection that the change was concerned with compiling in
3150        different directories, but in the new build system, that is taken care of
3151        by the VPATH setting the Makefile.
3153     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3154        when the subject happened to end in the byte 0x85 (e.g. if the last
3155        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3156        characters but of course it shouldn't be taken as a newline when it is part
3157        of another character. The bug was that, for an unlimited repeat of . in
3158        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3159        characters when looking for a newline.
3161     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3163     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3164        in debug output.
3166     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3167        long printing in the pcrecpp unittest when running under MinGW.
3169     6. ESC_K was left out of the EBCDIC table.
3171     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3172        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3173        limit also applies to "virtual nesting" when a pattern is recursive, and in
3174        this case 1000 isn't so big. I have been able to remove this limit at the
3175        expense of backing off one optimization in certain circumstances. Normally,
3176        when pcre_exec() would call its internal match() function recursively and
3177        immediately return the result unconditionally, it uses a "tail recursion"
3178        feature to save stack. However, when a subpattern that can match an empty
3179        string has an unlimited repetition quantifier, it no longer makes this
3180        optimization. That gives it a stack frame in which to save the data for
3181        checking that an empty string has been matched. Previously this was taken
3182        from the 1000-entry workspace that had been reserved. So now there is no
3183        explicit limit, but more stack is used.
3185     8. Applied Daniel's patches to solve problems with the import/export magic
3186        syntax that is required for Windows, and which was going wrong for the
3187        pcreposix and pcrecpp parts of the library. These were overlooked when this
3188        problem was solved for the main library.
3190     9. There were some crude static tests to avoid integer overflow when computing
3191        the size of patterns that contain repeated groups with explicit upper
3192        limits. As the maximum quantifier is 65535, the maximum group length was
3193        set at 30,000 so that the product of these two numbers did not overflow a
3194        32-bit integer. However, it turns out that people want to use groups that
3195        are longer than 30,000 bytes (though not repeat them that many times).
3196        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3197        made it possible to implement the integer overflow checks in a much more
3198        dynamic way, which I have now done. The artificial limitation on group
3199        length has been removed - we now have only the limit on the total length of
3200        the compiled pattern, which depends on the LINK_SIZE setting.
3202    10. Fixed a bug in the documentation for get/copy named substring when
3203        duplicate names are permitted. If none of the named substrings are set, the
3204        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3205        empty string.
3207    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3208        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3209        because the ] is interpreted as the first data character and the
3210        terminating ] is not found. PCRE has been made compatible with Perl in this
3211        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3212        cause memory overwriting.
3214    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3215        string has been matched (to stop an infinite loop). It was not recognizing
3216        a conditional subpattern that could match an empty string if that
3217        subpattern was within another subpattern. For example, it looped when
3218        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3219        condition was not nested. This bug has been fixed.
3221    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3222        past the start of the subject in the presence of bytes with the top bit
3223        set, for example "\x8aBCD".
3225    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3226        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3228    14. Optimized (?!) to (*FAIL).
3230    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3231        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3232        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3233        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3234        does: it's just the validity check that is more restrictive.
3236    16. Inserted checks for integer overflows during escape sequence (backslash)
3237        processing, and also fixed erroneous offset values for syntax errors during
3238        backslash processing.
3240    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3241        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3243    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3244        caused an overrun.
3246    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3247        something other than just ASCII characters) inside a group that had an
3248        unlimited repeat caused a loop at compile time (while checking to see
3249        whether the group could match an empty string).
3251    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3252        [\P{Any}] did so. (Error in the code for printing property names.)
3254    21. An orphan \E inside a character class could cause a crash.
3256    22. A repeated capturing bracket such as (A)? could cause a wild memory
3257        reference during compilation.
3259    23. There are several functions in pcre_compile() that scan along a compiled
3260        expression for various reasons (e.g. to see if it's fixed length for look
3261        behind). There were bugs in these functions when a repeated \p or \P was
3262        present in the pattern. These operators have additional parameters compared
3263        with \d, etc, and these were not being taken into account when moving along
3264        the compiled data. Specifically:
3266        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3267            length.
3269        (b) An item such as \pL+ within a repeated group could cause crashes or
3270            loops.
3272        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3273            "reference to non-existent subpattern" error.
3275        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3277    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3278        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3280    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3282    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3283        character were causing crashes (broken optimization).
3285    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3286        \p or \P) caused a compile-time loop.
3288    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3289        break. For example, the unstudied pattern [\r\n]A does not match the string
3290        "\r\nA" because change 7.0/46 below moves the current point on by two
3291        characters after failing to match at the start. However, the pattern \nA
3292        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3293        the same is true. There doesn't seem any very clean way out of this, but
3294        what I have chosen to do makes the common cases work: PCRE now takes note
3295        of whether there can be an explicit match for \r or \n anywhere in the
3296        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3297        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3298        pattern has explicit CR or LF references.
3300    29. Added (*CR) etc for changing newline setting at start of pattern.
3303    Version 7.2 19-Jun-07
3304    ---------------------
3306     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3307        which is apparently normally available under Windows.
3309     2. Re-jig the pcregrep tests with different newline settings in an attempt
3310        to make them independent of the local environment's newline setting.
3312     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3314     4. Some of the "internals" tests were previously cut out when the link size
3315        was not 2, because the output contained actual offsets. The recent new
3316        "Z" feature of pcretest means that these can be cut out, making the tests
3317        usable with all link sizes.
3319     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3320        stack recursion. This gives a massive performance boost under BSD, but just
3321        a small improvement under Linux. However, it saves one field in the frame
3322        in all cases.
3324     6. Added more features from the forthcoming Perl 5.10:
3326        (a) (?-n) (where n is a string of digits) is a relative subroutine or
3327            recursion call. It refers to the nth most recently opened parentheses.
3329        (b) (?+n) is also a relative subroutine call; it refers to the nth next
3330            to be opened parentheses.
3332        (c) Conditions that refer to capturing parentheses can be specified
3333            relatively, for example, (?(-2)... or (?(+3)...
3335        (d) \K resets the start of the current match so that everything before
3336            is not part of it.
3338        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3340        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3341            reference syntax.
3343        (g) (?| introduces a group in which the numbering of parentheses in each
3344            alternative starts with the same number.
3346        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3348     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3351     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3352        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3353        for detecting groups that can match an empty string.
3355     9. A pattern with a very large number of alternatives (more than several
3356        hundred) was running out of internal workspace during the pre-compile
3357        phase, where pcre_compile() figures out how much memory will be needed. A
3358        bit of new cunning has reduced the workspace needed for groups with
3359        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3360        workspace instead of running out of the 4096 that are available.
3362    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3364    11. Applied patch from Google to remove an optimization that didn't quite work.
3365        The report of the bug said:
3367          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3368          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3369          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3371    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3372        it matched the wrong number of bytes.
3375    Version 7.1 24-Apr-07
3376    ---------------------
3378     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3379        that is more "standard", making use of automake and other Autotools. There
3380        is some re-arrangement of the files and adjustment of comments consequent
3381        on this.
3383     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3384        for recursive directory scanning broke on some systems because the files
3385        are not scanned in any specific order and on different systems the order
3386        was different. A call to "sort" has been inserted into RunGrepTest for the
3387        approprate test as a short-term fix. In the longer term there may be an
3388        alternative.
3390     3. I had an email from Eric Raymond about problems translating some of PCRE's
3391        man pages to HTML (despite the fact that I distribute HTML pages, some
3392        people do their own conversions for various reasons). The problems
3393        concerned the use of low-level troff macros .br and .in. I have therefore
3394        removed all such uses from the man pages (some were redundant, some could
3395        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3396        HTML has been updated to handle .nf/.fi and to complain if it encounters
3397        .br or .in.
3399     4. Updated comments in configure.ac that get placed in config.h.in and also
3400        arranged for config.h to be included in the distribution, with the name
3401        config.h.generic, for the benefit of those who have to compile without
3402        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3404     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3405        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3406        makevp.bat for the current PCRE, using the additional files
3407        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3409     6. A Windows user reported a minor discrepancy with test 2, which turned out
3410        to be caused by a trailing space on an input line that had got lost in his
3411        copy. The trailing space was an accident, so I've just removed it.
3413     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3414        that is needed.
3416     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3417        as "const" (a) because they are and (b) because it helps the PHP
3418        maintainers who have recently made a script to detect big data structures
3419        in the php code that should be moved to the .rodata section. I remembered
3420        to update Builducptable as well, so it won't revert if ucptable.h is ever
3421        re-created.
3423     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3424        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3425        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3426        support is not required. This saves 1.5-2K of code, which is important in
3427        some applications.
3429        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3430        so as not to refer to the tables, even though these functions will never be
3431        called when UTF-8 support is disabled. Otherwise there are problems with a
3432        shared library.
3434    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3436        (a) It was defining its arguments as char * instead of void *.
3438        (b) It was assuming that all moves were upwards in memory; this was true
3439            a long time ago when I wrote it, but is no longer the case.
3441        The emulated memove() is provided for those environments that have neither
3442        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3443        is clearly not the case, as these two bugs were recently reported.
3445    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3446        and Detrail to create the HTML documentation, the .txt form of the man
3447        pages, and it removes trailing spaces from listed files. It also creates
3448        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3449        case, it wraps all the #defines with #ifndefs. This script should be run
3450        before "make dist".
3452    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3453        with Unicode property support.
3455        (a) For a maximizing quantifier, if the two different cases of the
3456            character were of different lengths in their UTF-8 codings (there are
3457            some cases like this - I found 11), and the matching function had to
3458            back up over a mixture of the two cases, it incorrectly assumed they
3459            were both the same length.
3461        (b) When PCRE was configured to use the heap rather than the stack for
3462            recursion during matching, it was not correctly preserving the data for
3463            the other case of a UTF-8 character when checking ahead for a match
3464            while processing a minimizing repeat. If the check also involved
3465            matching a wide character, but failed, corruption could cause an
3466            erroneous result when trying to check for a repeat of the original
3467            character.
3469    13. Some tidying changes to the testing mechanism:
3471        (a) The RunTest script now detects the internal link size and whether there
3472            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3473            values substituted by "configure". (The RunGrepTest script already did
3474            this for UTF-8.) The configure.ac script no longer substitutes the
3475            relevant variables.
3477        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3478            with length and offset values. This means that the output is different
3479            for different internal link sizes. Test 2 is skipped for link sizes
3480            other than 2 because of this, bypassing the problem. Unfortunately,
3481            there was also a test in test 3 (the locale tests) that used /B and
3482            failed for link sizes other than 2. Rather than cut the whole test out,
3483            I have added a new /Z option to pcretest that replaces the length and
3484            offset values with spaces. This is now used to make test 3 independent
3485            of link size. (Test 2 will be tidied up later.)
3487    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3488        segmentation fault instead of returning the appropriate error message.
3490    15. In multiline mode when the newline sequence was set to "any", the pattern
3491        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3492        This doesn't seem right; it now treats the CRLF combination as the line
3493        ending, and so does not match in that case. It's only a pattern such as ^$
3494        that would hit this one: something like ^ABC$ would have failed after \r
3495        and then tried again after \r\n.
3497    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3498        in an attempt to make files that differ only in their line terminators
3499        compare equal. This works on Linux.
3501    17. Under certain error circumstances pcregrep might try to free random memory
3502        as it exited. This is now fixed, thanks to valgrind.
3504    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3505        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3506        was because its rules for how to advance for /g after matching an empty
3507        string at the end of a line did not allow for this case. They now check for
3508        it specially.
3510    20. pcretest is supposed to handle patterns and data of any length, by
3511        extending its buffers when necessary. It was getting this wrong when the
3512        buffer for a data line had to be extended.
3514    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3515        CRLF as a newline sequence.
3517    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3518        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3519        I have nevertheless tidied it up.
3521    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3523    24. Added a man page for pcre-config.
3526    Version 7.0 19-Dec-06
3527    ---------------------
3529     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3530        moving to gcc 4.1.1.
3532     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3533        sys/time.h, which is documented as needed for this function. It doesn't
3534        seem to matter on Linux, but it showed up on some releases of OS X.
3536     3. It seems that there are systems where bytes whose values are greater than
3537        127 match isprint() in the "C" locale. The "C" locale should be the
3538        default when a C program starts up. In most systems, only ASCII printing
3539        characters match isprint(). This difference caused the output from pcretest
3540        to vary, making some of the tests fail. I have changed pcretest so that:
3542        (a) When it is outputting text in the compiled version of a pattern, bytes
3543            other than 32-126 are always shown as hex escapes.
3545        (b) When it is outputting text that is a matched part of a subject string,
3546            it does the same, unless a different locale has been set for the match
3547            (using the /L modifier). In this case, it uses isprint() to decide.
3549     4. Fixed a major bug that caused incorrect computation of the amount of memory
3550        required for a compiled pattern when options that changed within the
3551        pattern affected the logic of the preliminary scan that determines the
3552        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3553        that the computed length was too small. The symptoms of this bug were
3554        either the PCRE error "internal error: code overflow" from pcre_compile(),
3555        or a glibc crash with a message such as "pcretest: free(): invalid next
3556        size (fast)". Examples of patterns that provoked this bug (shown in
3557        pcretest format) are:
3559          /(?-x: )/x
3560          /(?x)(?-x: \s*#\s*)/
3561          /((?i)[\x{c0}])/8
3562          /(?i:[\x{c0}])/8
3564        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3565        is now done differently.
3567     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3568        wrapper classes; (b) implement a new function in the C++ scanner that is
3569        more efficient than the old way of doing things because it avoids levels of
3570        recursion in the regex matching; (c) add a paragraph to the documentation
3571        for the FullMatch() function.
3573     6. The escape sequence \n was being treated as whatever was defined as
3574        "newline". Not only was this contrary to the documentation, which states
3575        that \n is character 10 (hex 0A), but it also went horribly wrong when
3576        "newline" was defined as CRLF. This has been fixed.
3578     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3579        was being set to -1 for the "end of line" case (supposedly a value that no
3580        character can have). Though this value is never used (the check for end of
3581        line is "zero bytes in current character"), it caused compiler complaints.
3582        I've changed it to 0xffffffff.
3584     8. In pcre_version.c, the version string was being built by a sequence of
3585        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3586        string (as it is for production releases) called a macro with an empty
3587        argument. The C standard says the result of this is undefined. The gcc
3588        compiler treats it as an empty string (which was what was wanted) but it is
3589        reported that Visual C gives an error. The source has been hacked around to
3590        avoid this problem.
3592     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows