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

Diff of /code/trunk/ChangeLog

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

revision 319 by ph10, Mon Jan 28 15:27:45 2008 UTC revision 1688 by ph10, Fri Feb 24 17:30:30 2017 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
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.
6    
7    Version 8.41
8    ------------
9    
10    1.  Fixed typo in CMakeLists.txt (wrong number of arguments for
11    PCRE_STATIC_RUNTIME (affects MSVC only).
12    
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.
17    
18    3.  Fix a missing else in the JIT compiler reported by 'idaifish'.
19    
20    4.  A (?# style comment is now ignored between a basic quantifier and a
21    following '+' or '?' (example: /X+(?#comment)?Y/.
22    
23    5.  Avoid use of a potentially overflowing buffer in pcregrep (patch by Petr
24    Pisar).
25    
26    6.  Fuzzers have reported issues in pcretest. These are NOT serious (it is,
27    after all, just a test program). However, to stop the reports, some easy ones
28    are fixed:
29    
30        (a) Check for values < 256 when calling isprint() in pcretest.
31        (b) Give an error for too big a number after \O.
32    
33    7.  In the 32-bit library in non-UTF mode, an attempt to find a Unicode
34    property for a character with a code point greater than 0x10ffff (the Unicode
35    maximum) caused a crash.
36    
37    
38    Version 8.40 11-January-2017
39    ----------------------------
40    
41    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
42        the match extended over a line boundary.
43    
44    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
45        MSVC static compilation, putting the first patch under a new option.
46    
47    3.  Fix register overwite in JIT when SSE2 acceleration is enabled.
48    
49    4.  Ignore "show all captures" (/=) for DFA matching.
50    
51    5.  Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
52    
53    6.  In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
54        without PCRE_UCP set, a negative character type such as \D in a positive
55        class should cause all characters greater than 255 to match, whatever else
56        is in the class. There was a bug that caused this not to happen if a
57        Unicode property item was added to such a class, for example [\D\P{Nd}] or
58        [\W\pL].
59    
60    7.  When pcretest was outputing information from a callout, the caret indicator
61        for the current position in the subject line was incorrect if it was after
62        an escape sequence for a character whose code point was greater than
63        \x{ff}.
64    
65    8.  A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
66        the conditional was interpreted as a reference to capturing group 1 instead
67        of a test for recursion. Any group whose name began with R was
68        misinterpreted in this way. (The reference interpretation should only
69        happen if the group's name is precisely "R".)
70    
71    9.  A number of bugs have been mended relating to match start-up optimizations
72        when the first thing in a pattern is a positive lookahead. These all
73        applied only when PCRE_NO_START_OPTIMIZE was *not* set:
74    
75        (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
76            both an initial 'X' and a following 'X'.
77        (b) Some patterns starting with an assertion that started with .* were
78            incorrectly optimized as having to match at the start of the subject or
79            after a newline. There are cases where this is not true, for example,
80            (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
81            start with spaces. Starting .* in an assertion is no longer taken as an
82            indication of matching at the start (or after a newline).
83    
84    
85    Version 8.39 14-June-2016
86    -------------------------
87    
88    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
89        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
90        misbehaved. This bug was found by the LLVM fuzzer.
91    
92    2.  Similar to the above, if an isolated \E was present between an item and its
93        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
94        bug was found by the LLVM fuzzer.
95    
96    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
97        working correctly in UCP mode.
98    
99    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
100        was set when the pmatch argument was NULL. It now returns REG_INVARG.
101    
102    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
103    
104    6.  An empty \Q\E sequence between an item and its qualifier caused
105        pcre_compile() to misbehave when auto callouts were enabled. This bug was
106        found by the LLVM fuzzer.
107    
108    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
109        space or a #-type comment that was followed by (?-x), which turns off
110        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
111        pcre_compile() assumed that (?-x) applied to the whole pattern and
112        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
113    
114    8.  A call of pcre_copy_named_substring() for a named substring whose number
115        was greater than the space in the ovector could cause a crash.
116    
117    9.  Yet another buffer overflow bug involved duplicate named groups with a
118        group that reset capture numbers (compare 8.38/7 below). Once again, I have
119        just allowed for more memory, even if not needed. (A proper fix is
120        implemented in PCRE2, but it involves a lot of refactoring.)
121    
122    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
123        start of the match to be earlier than the end.
124    
125    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
126    
127    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
128        assertion, caused pcretest to generate incorrect output, and also to read
129        uninitialized memory (detected by ASAN or valgrind).
130    
131    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
132        nested set of parentheses of sufficient size caused an overflow of the
133        compiling workspace (which was diagnosed, but of course is not desirable).
134    
135    14. And yet another buffer overflow bug involving duplicate named groups, this
136        time nested, with a nested back reference. Yet again, I have just allowed
137        for more memory, because anything more needs all the refactoring that has
138        been done for PCRE2. An example pattern that provoked this bug is:
139        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
140        registered as CVE-2016-1283.
141    
142    15. pcretest went into a loop if global matching was requested with an ovector
143        size less than 2. It now gives an error message. This bug was found by
144        afl-fuzz.
145    
146    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
147        ("assertion expected") when (?(?C) was not followed by an opening
148        parenthesis.
149    
150    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
151        actually affect anything, by sheer luck.
152    
153    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
154        static compilation.
155    
156    19. Modified the RunTest script to incorporate a valgrind suppressions file so
157        that certain errors, provoked by the SSE2 instruction set when JIT is used,
158        are ignored.
159    
160    20. A racing condition is fixed in JIT reported by Mozilla.
161    
162    21. Minor code refactor to avoid "array subscript is below array bounds"
163        compiler warning.
164    
165    22. Minor code refactor to avoid "left shift of negative number" warning.
166    
167    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
168        UCP support.
169    
170    24. Refactor to avoid compiler warnings in pcrecpp.cc.
171    
172    25. Refactor to fix a typo in pcre_jit_test.c
173    
174    26. Patch to support compiling pcrecpp.cc with Intel compiler.
175    
176    
177    Version 8.38 23-November-2015
178    -----------------------------
179    
180    1.  If a group that contained a recursive back reference also contained a
181        forward reference subroutine call followed by a non-forward-reference
182        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
183        compile correct code, leading to undefined behaviour or an internally
184        detected error. This bug was discovered by the LLVM fuzzer.
185    
186    2.  Quantification of certain items (e.g. atomic back references) could cause
187        incorrect code to be compiled when recursive forward references were
188        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
189        This bug was discovered by the LLVM fuzzer.
190    
191    3.  A repeated conditional group whose condition was a reference by name caused
192        a buffer overflow if there was more than one group with the given name.
193        This bug was discovered by the LLVM fuzzer.
194    
195    4.  A recursive back reference by name within a group that had the same name as
196        another group caused a buffer overflow. For example:
197        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
198    
199    5.  A forward reference by name to a group whose number is the same as the
200        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
201        a buffer overflow at compile time. This bug was discovered by the LLVM
202        fuzzer.
203    
204    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
205        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
206    
207    7.  Another buffer overflow bug involved duplicate named groups with a
208        reference between their definition, with a group that reset capture
209        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
210        fixed by always allowing for more memory, even if not needed. (A proper fix
211        is implemented in PCRE2, but it involves more refactoring.)
212    
213    8.  There was no check for integer overflow in subroutine calls such as (?123).
214    
215    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
216        being treated as a literal 'l' instead of causing an error.
217    
218    10. There was a buffer overflow if pcre_exec() was called with an ovector of
219        size 1. This bug was found by american fuzzy lop.
220    
221    11. If a non-capturing group containing a conditional group that could match
222        an empty string was repeated, it was not identified as matching an empty
223        string itself. For example: /^(?:(?(1)x|)+)+$()/.
224    
225    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
226        \a and \e in test subject lines.
227    
228    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
229        instead of the EBCDIC value.
230    
231    14. The handling of \c in an EBCDIC environment has been revised so that it is
232        now compatible with the specification in Perl's perlebcdic page.
233    
234    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
235        ASCII/Unicode. This has now been added to the list of characters that are
236        recognized as white space in EBCDIC.
237    
238    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
239        error (correctly) when used outside a class, but did not give an error
240        within a class.
241    
242    17. \h within a class was incorrectly compiled in EBCDIC environments.
243    
244    18. A pattern with an unmatched closing parenthesis that contained a backward
245        assertion which itself contained a forward reference caused buffer
246        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
247    
248    19. JIT should return with error when the compiled pattern requires more stack
249        space than the maximum.
250    
251    20. A possessively repeated conditional group that could match an empty string,
252        for example, /(?(R))*+/, was incorrectly compiled.
253    
254    21. Fix infinite recursion in the JIT compiler when certain patterns such as
255        /(?:|a|){100}x/ are analysed.
256    
257    22. Some patterns with character classes involving [: and \\ were incorrectly
258        compiled and could cause reading from uninitialized memory or an incorrect
259        error diagnosis.
260    
261    23. Pathological patterns containing many nested occurrences of [: caused
262        pcre_compile() to run for a very long time.
263    
264    24. A conditional group with only one branch has an implicit empty alternative
265        branch and must therefore be treated as potentially matching an empty
266        string.
267    
268    25. If (?R was followed by - or + incorrect behaviour happened instead of a
269        diagnostic.
270    
271    26. Arrange to give up on finding the minimum matching length for overly
272        complex patterns.
273    
274    27. Similar to (4) above: in a pattern with duplicated named groups and an
275        occurrence of (?| it is possible for an apparently non-recursive back
276        reference to become recursive if a later named group with the relevant
277        number is encountered. This could lead to a buffer overflow. Wen Guanxing
278        from Venustech ADLAB discovered this bug.
279    
280    28. If pcregrep was given the -q option with -c or -l, or when handling a
281        binary file, it incorrectly wrote output to stdout.
282    
283    29. The JIT compiler did not restore the control verb head in case of *THEN
284        control verbs. This issue was found by Karl Skomski with a custom LLVM
285        fuzzer.
286    
287    30. Error messages for syntax errors following \g and \k were giving inaccurate
288        offsets in the pattern.
289    
290    31. Added a check for integer overflow in conditions (?(<digits>) and
291        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
292        fuzzer.
293    
294    32. Handling recursive references such as (?2) when the reference is to a group
295        later in the pattern uses code that is very hacked about and error-prone.
296        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
297        give an internal error if it is obvious that compiling has gone wrong.
298    
299    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
300        This issue was found by Karl Skomski with a custom LLVM fuzzer.
301    
302    34. The JIT compiler should restore the control chain for empty possessive
303        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
304    
305    35. Match limit check added to JIT recursion. This issue was found by Karl
306        Skomski with a custom LLVM fuzzer.
307    
308    36. Yet another case similar to 27 above has been circumvented by an
309        unconditional allocation of extra memory. This issue is fixed "properly" in
310        PCRE2 by refactoring the way references are handled. Wen Guanxing
311        from Venustech ADLAB discovered this bug.
312    
313    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
314        with a custom LLVM fuzzer.
315    
316    38. Fixed a corner case of range optimization in JIT.
317    
318    39. An incorrect error "overran compiling workspace" was given if there were
319        exactly enough group forward references such that the last one extended
320        into the workspace safety margin. The next one would have expanded the
321        workspace. The test for overflow was not including the safety margin.
322    
323    40. A match limit issue is fixed in JIT which was found by Karl Skomski
324        with a custom LLVM fuzzer.
325    
326    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
327        work under Windows. (Why has it taken so long for anyone to notice?)
328    
329    42. In a character class such as [\W\p{Any}] where both a negative-type escape
330        ("not a word character") and a property escape were present, the property
331        escape was being ignored.
332    
333    43. Fix crash caused by very long (*MARK) or (*THEN) names.
334    
335    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
336        by a single ASCII character in a class item, was incorrectly compiled in
337        UCP mode. The POSIX class got lost, but only if the single character
338        followed it.
339    
340    45. [:punct:] in UCP mode was matching some characters in the range 128-255
341        that should not have been matched.
342    
343    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
344        class, all characters with code points greater than 255 are in the class.
345        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
346        such as \w are turned into Unicode properties), wide characters were not
347        correctly handled, and could fail to match.
348    
349    
350    Version 8.37 28-April-2015
351    --------------------------
352    
353    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
354        for those parentheses to be closed with whatever has been captured so far.
355        However, it was failing to mark any other groups between the hightest
356        capture so far and the currrent group as "unset". Thus, the ovector for
357        those groups contained whatever was previously there. An example is the
358        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
359    
360    2.  If an assertion condition was quantified with a minimum of zero (an odd
361        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
362    
363    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
364        unrecognized modifier, a crash could occur.
365    
366    4.  An attempt to do global matching in pcretest with a zero-length ovector
367        caused a crash.
368    
369    5.  Fixed a memory leak during matching that could occur for a subpattern
370        subroutine call (recursive or otherwise) if the number of captured groups
371        that had to be saved was greater than ten.
372    
373    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
374        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
375        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
376    
377    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
378        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
379    
380    8. When a pattern is compiled, it remembers the highest back reference so that
381       when matching, if the ovector is too small, extra memory can be obtained to
382       use instead. A conditional subpattern whose condition is a check on a
383       capture having happened, such as, for example in the pattern
384       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
385       setting the highest backreference number. This mattered only if pcre_exec()
386       was called with an ovector that was too small to hold the capture, and there
387       was no other kind of back reference (a situation which is probably quite
388       rare). The effect of the bug was that the condition was always treated as
389       FALSE when the capture could not be consulted, leading to a incorrect
390       behaviour by pcre_exec(). This bug has been fixed.
391    
392    9. A reference to a duplicated named group (either a back reference or a test
393       for being set in a conditional) that occurred in a part of the pattern where
394       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
395       to be incorrectly calculated, leading to overwriting.
396    
397    10. A mutually recursive set of back references such as (\2)(\1) caused a
398        segfault at study time (while trying to find the minimum matching length).
399        The infinite loop is now broken (with the minimum length unset, that is,
400        zero).
401    
402    11. If an assertion that was used as a condition was quantified with a minimum
403        of zero, matching went wrong. In particular, if the whole group had
404        unlimited repetition and could match an empty string, a segfault was
405        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
406        assertions to be quantified, but not if they are being used as conditions,
407        so the above pattern is faulted by Perl. PCRE has now been changed so that
408        it also rejects such patterns.
409    
410    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
411        failed to allow the zero-repeat case if pcre2_exec() was called with an
412        ovector too small to capture the group.
413    
414    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
415        Red Hat Product Security:
416    
417        (a) A crash if /K and /F were both set with the option to save the compiled
418        pattern.
419    
420        (b) Another crash if the option to print captured substrings in a callout
421        was combined with setting a null ovector, for example \O\C+ as a subject
422        string.
423    
424    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
425        forward reference repeated a large (but limited) number of times within a
426        repeated outer group that has a zero minimum quantifier, caused incorrect
427        code to be compiled, leading to the error "internal error:
428        previously-checked referenced subpattern not found" when an incorrect
429        memory address was read. This bug was reported as "heap overflow",
430        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
431        CVE-2015-2325.
432    
433    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
434        call within a group that also contained a recursive back reference caused
435        incorrect code to be compiled. This bug was reported as "heap overflow",
436        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
437        number CVE-2015-2326.
438    
439    24. Computing the size of the JIT read-only data in advance has been a source
440        of various issues, and new ones are still appear unfortunately. To fix
441        existing and future issues, size computation is eliminated from the code,
442        and replaced by on-demand memory allocation.
443    
444    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
445        adjacent to the end of the range, and the range contained characters with
446        more than one other case, caused incorrect behaviour when compiled in UTF
447        mode. In that example, the range a-j was left out of the class.
448    
449    26. Fix JIT compilation of conditional blocks, which assertion
450        is converted to (*FAIL). E.g: /(?(?!))/.
451    
452    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
453        discovered by the LLVM fuzzer.
454    
455    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
456        when this assertion was used as a condition, for example (?(?!)a|b). In
457        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
458        error about an unsupported item.
459    
460    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
461        possessification code could take exponential time to complete. A recursion
462        depth limit of 1000 has been imposed to limit the resources used by this
463        optimization.
464    
465    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
466        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
467        because \S ensures they are all in the class. The code for doing this was
468        interacting badly with the code for computing the amount of space needed to
469        compile the pattern, leading to a buffer overflow. This bug was discovered
470        by the LLVM fuzzer.
471    
472    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
473        other kinds of group caused stack overflow at compile time. This bug was
474        discovered by the LLVM fuzzer.
475    
476    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
477        between a subroutine call and its quantifier was incorrectly compiled,
478        leading to buffer overflow or other errors. This bug was discovered by the
479        LLVM fuzzer.
480    
481    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
482        assertion after (?(. The code was failing to check the character after
483        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
484        was discovered by the LLVM fuzzer.
485    
486    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
487        a fixed maximum following a group that contains a subroutine reference was
488        incorrectly compiled and could trigger buffer overflow. This bug was
489        discovered by the LLVM fuzzer.
490    
491    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
492        caused a stack overflow instead of the diagnosis of a non-fixed length
493        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
494    
495    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
496        (e.g. /(?<=\Ka)/) could make pcregrep loop.
497    
498    37. There was a similar problem to 36 in pcretest for global matches.
499    
500    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
501        and a subsequent item in the pattern caused a non-match, backtracking over
502        the repeated \X did not stop, but carried on past the start of the subject,
503        causing reference to random memory and/or a segfault. There were also some
504        other cases where backtracking after \C could crash. This set of bugs was
505        discovered by the LLVM fuzzer.
506    
507    39. The function for finding the minimum length of a matching string could take
508        a very long time if mutual recursion was present many times in a pattern,
509        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
510        method has been implemented. This infelicity was discovered by the LLVM
511        fuzzer.
512    
513    40. Static linking against the PCRE library using the pkg-config module was
514        failing on missing pthread symbols.
515    
516    
517    Version 8.36 26-September-2014
518    ------------------------------
519    
520    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
521        -Wmissing-field-initializers and -Wunused-parameter.
522    
523    2.  The tests for quantifiers being too big (greater than 65535) were being
524        applied after reading the number, and stupidly assuming that integer
525        overflow would give a negative number. The tests are now applied as the
526        numbers are read.
527    
528    3.  Tidy code in pcre_exec.c where two branches that used to be different are
529        now the same.
530    
531    4.  The JIT compiler did not generate match limit checks for certain
532        bracketed expressions with quantifiers. This may lead to exponential
533        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
534        issue should be resolved now.
535    
536    5.  Fixed an issue, which occures when nested alternatives are optimized
537        with table jumps.
538    
539    6.  Inserted two casts and changed some ints to size_t in the light of some
540        reported 64-bit compiler warnings (Bugzilla 1477).
541    
542    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
543        an empty string, which sometimes were behaving incorrectly in the
544        interpreter (though correctly in the JIT matcher). This pcretest input is
545        an example:
546    
547          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
548          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
549    
550        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
551        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
552        for an empty string was breaking the inner loop and carrying on at a lower
553        level, when possessive repeated groups should always return to a higher
554        level as they have no backtrack points in them. The empty string test now
555        occurs at the outer level.
556    
557    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
558        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
559    
560    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
561        doesn't).
562    
563    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
564        to hold a pattern to be incorrectly computed (too small) when there were
565        named back references to duplicated names. This could cause "internal
566        error: code overflow" or "double free or corruption" or other memory
567        handling errors.
568    
569    11. When named subpatterns had the same prefixes, back references could be
570        confused. For example, in this pattern:
571    
572          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
573    
574        the reference to 'Name' was incorrectly treated as a reference to a
575        duplicate name.
576    
577    12. A pattern such as /^s?c/mi8 where the optional character has more than
578        one "other case" was incorrectly compiled such that it would only try to
579        match starting at "c".
580    
581    13. When a pattern starting with \s was studied, VT was not included in the
582        list of possible starting characters; this should have been part of the
583        8.34/18 patch.
584    
585    14. If a character class started [\Qx]... where x is any character, the class
586        was incorrectly terminated at the ].
587    
588    15. If a pattern that started with a caseless match for a character with more
589        than one "other case" was studied, PCRE did not set up the starting code
590        unit bit map for the list of possible characters. Now it does. This is an
591        optimization improvement, not a bug fix.
592    
593    16. The Unicode data tables have been updated to Unicode 7.0.0.
594    
595    17. Fixed a number of memory leaks in pcregrep.
596    
597    18. Avoid a compiler warning (from some compilers) for a function call with
598        a cast that removes "const" from an lvalue by using an intermediate
599        variable (to which the compiler does not object).
600    
601    19. Incorrect code was compiled if a group that contained an internal recursive
602        back reference was optional (had quantifier with a minimum of zero). This
603        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
604        caused segmentation faults because of stack overflows at compile time.
605    
606    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
607        group that is quantified with an indefinite repeat, caused a compile-time
608        loop which used up all the system stack and provoked a segmentation fault.
609        This was not the same bug as 19 above.
610    
611    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
612        Patch by Mike Frysinger.
613    
614    
615    Version 8.35 04-April-2014
616    --------------------------
617    
618    1.  A new flag is set, when property checks are present in an XCLASS.
619        When this flag is not set, PCRE can perform certain optimizations
620        such as studying these XCLASS-es.
621    
622    2.  The auto-possessification of character sets were improved: a normal
623        and an extended character set can be compared now. Furthermore
624        the JIT compiler optimizes more character set checks.
625    
626    3.  Got rid of some compiler warnings for potentially uninitialized variables
627        that show up only when compiled with -O2.
628    
629    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
630        of a match later then the end of the match. The pcretest program was not
631        handling the case sensibly - it was outputting from the start to the next
632        binary zero. It now reports this situation in a message, and outputs the
633        text from the end to the start.
634    
635    5.  Fast forward search is improved in JIT. Instead of the first three
636        characters, any three characters with fixed position can be searched.
637        Search order: first, last, middle.
638    
639    6.  Improve character range checks in JIT. Characters are read by an inprecise
640        function now, which returns with an unknown value if the character code is
641        above a certain threshold (e.g: 256). The only limitation is that the value
642        must be bigger than the threshold as well. This function is useful when
643        the characters above the threshold are handled in the same way.
644    
645    7.  The macros whose names start with RAWUCHAR are placeholders for a future
646        mode in which only the bottom 21 bits of 32-bit data items are used. To
647        make this more memorable for those maintaining the code, the names have
648        been changed to start with UCHAR21, and an extensive comment has been added
649        to their definition.
650    
651    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
652        to the export list in Makefile.am (they were accidentally omitted from the
653        8.34 tarball).
654    
655    9.  The informational output from pcretest used the phrase "starting byte set"
656        which is inappropriate for the 16-bit and 32-bit libraries. As the output
657        for "first char" and "need char" really means "non-UTF-char", I've changed
658        "byte" to "char", and slightly reworded the output. The documentation about
659        these values has also been (I hope) clarified.
660    
661    10. Another JIT related optimization: use table jumps for selecting the correct
662        backtracking path, when more than four alternatives are present inside a
663        bracket.
664    
665    11. Empty match is not possible, when the minimum length is greater than zero,
666        and there is no \K in the pattern. JIT should avoid empty match checks in
667        such cases.
668    
669    12. In a caseless character class with UCP support, when a character with more
670        than one alternative case was not the first character of a range, not all
671        the alternative cases were added to the class. For example, s and \x{17f}
672        are both alternative cases for S: the class [RST] was handled correctly,
673        but [R-T] was not.
674    
675    13. The configure.ac file always checked for pthread support when JIT was
676        enabled. This is not used in Windows, so I have put this test inside a
677        check for the presence of windows.h (which was already tested for).
678    
679    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
680        The algorithm provides a way to skip certain starting offsets, and usually
681        faster than linear prefix searches.
682    
683    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
684        as for 'fr_FR' and 'french'. For some reason, however, it then used the
685        Windows-specific input and output files, which have 'french' screwed in.
686        So this could never have worked. One of the problems with locales is that
687        they aren't always the same. I have now updated RunTest so that it checks
688        the output of the locale test (test 3) against three different output
689        files, and it allows the test to pass if any one of them matches. With luck
690        this should make the test pass on some versions of Solaris where it was
691        failing. Because of the uncertainty, the script did not used to stop if
692        test 3 failed; it now does. If further versions of a French locale ever
693        come to light, they can now easily be added.
694    
695    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
696        there was a message during ./configure, but it did not stop. This now
697        provokes an error. The invalid example in README has been corrected.
698        If a value less than the minimum is given, the minimum value has always
699        been used, but now a warning is given.
700    
701    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
702        was a bug in the test system, which is now fixed. Also, the list of various
703        configurations that are tested for each release did not have one with both
704        16/32 bits and --enable-bar-anycrlf. It now does.
705    
706    18. pcretest was missing "-C bsr" for displaying the \R default setting.
707    
708    19. Little endian PowerPC systems are supported now by the JIT compiler.
709    
710    20. The fast forward newline mechanism could enter to an infinite loop on
711        certain invalid UTF-8 input. Although we don't support these cases
712        this issue can be fixed by a performance optimization.
713    
714    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
715        not take account if existing stack usage. There is now a new global
716        variable called pcre_stack_guard that can be set to point to an external
717        function to check stack availability. It is called at the start of
718        processing every parenthesized group.
719    
720    22. A typo in the code meant that in ungreedy mode the max/min qualifier
721        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
722        not match "ab".
723    
724    23. When UTF was disabled, the JIT program reported some incorrect compile
725        errors. These messages are silenced now.
726    
727    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
728        compiler.
729    
730    25. Change all the temporary files used in RunGrepTest to be different to those
731        used by RunTest so that the tests can be run simultaneously, for example by
732        "make -j check".
733    
734    
735    Version 8.34 15-December-2013
736    -----------------------------
737    
738    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
739        executable memory. Patch inspired by Carsten Klein.
740    
741    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
742        this macro is never tested and has no effect, because the work to support
743        coverage involves only compiling and linking options and special targets in
744        the Makefile. The comment in config.h implied that defining the macro would
745        enable coverage support, which is totally false. There was also support for
746        setting this macro in the CMake files (my fault, I just copied it from
747        configure). SUPPORT_GCOV has now been removed.
748    
749    3.  Make a small performance improvement in strlen16() and strlen32() in
750        pcretest.
751    
752    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
753        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
754        these cases). There was also one in pcretest.c.
755    
756    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
757    
758    6.  In UTF mode, the code for checking whether a group could match an empty
759        string (which is used for indefinitely repeated groups to allow for
760        breaking an infinite loop) was broken when the group contained a repeated
761        negated single-character class with a character that occupied more than one
762        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
763        UTF-8 mode). The effect was undefined: the group might or might not be
764        deemed as matching an empty string, or the program might have crashed.
765    
766    7.  The code for checking whether a group could match an empty string was not
767        recognizing that \h, \H, \v, \V, and \R must match a character.
768    
769    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
770        an empty string. If it can, pcretest shows this in its information output.
771    
772    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
773        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
774        matched by pcre_exec() without using JIT:
775    
776        (a) If the rest of the pattern did not match after a maximal run of
777            grapheme clusters, the code for backing up to try with fewer of them
778            did not always back up over a full grapheme when characters that do not
779            have the modifier quality were involved, e.g. Hangul syllables.
780    
781        (b) If the match point in a subject started with modifier character, and
782            there was no match, the code could incorrectly back up beyond the match
783            point, and potentially beyond the first character in the subject,
784            leading to a segfault or an incorrect match result.
785    
786    10. A conditional group with an assertion condition could lead to PCRE
787        recording an incorrect first data item for a match if no other first data
788        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
789        first data item, and therefore matched "ca" after "c" instead of at the
790        start.
791    
792    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
793        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
794    
795    12. The source of pcregrep now includes z/OS-specific code so that it can be
796        compiled for z/OS as part of the special z/OS distribution.
797    
798    13. Added the -T and -TM options to pcretest.
799    
800    14. The code in pcre_compile.c for creating the table of named capturing groups
801        has been refactored. Instead of creating the table dynamically during the
802        actual compiling pass, the information is remembered during the pre-compile
803        pass (on the stack unless there are more than 20 named groups, in which
804        case malloc() is used) and the whole table is created before the actual
805        compile happens. This has simplified the code (it is now nearly 150 lines
806        shorter) and prepared the way for better handling of references to groups
807        with duplicate names.
808    
809    15. A back reference to a named subpattern when there is more than one of the
810        same name now checks them in the order in which they appear in the pattern.
811        The first one that is set is used for the reference. Previously only the
812        first one was inspected. This change makes PCRE more compatible with Perl.
813    
814    16. Unicode character properties were updated from Unicode 6.3.0.
815    
816    17. The compile-time code for auto-possessification has been refactored, based
817        on a patch by Zoltan Herczeg. It now happens after instead of during
818        compilation. The code is cleaner, and more cases are handled. The option
819        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
820        options in pcretest are provided to set it. It can also be set by
821        (*NO_AUTO_POSSESS) at the start of a pattern.
822    
823    18. The character VT has been added to the default ("C" locale) set of
824        characters that match \s and are generally treated as white space,
825        following this same change in Perl 5.18. There is now no difference between
826        "Perl space" and "POSIX space". Whether VT is treated as white space in
827        other locales depends on the locale.
828    
829    19. The code for checking named groups as conditions, either for being set or
830        for being recursed, has been refactored (this is related to 14 and 15
831        above). Processing unduplicated named groups should now be as fast at
832        numerical groups, and processing duplicated groups should be faster than
833        before.
834    
835    20. Two patches to the CMake build system, by Alexander Barkov:
836    
837          (1) Replace the "source" command by "." in CMakeLists.txt because
838              "source" is a bash-ism.
839    
840          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
841              without these the CMake build does not work on Solaris.
842    
843    21. Perl has changed its handling of \8 and \9. If there is no previously
844        encountered capturing group of those numbers, they are treated as the
845        literal characters 8 and 9 instead of a binary zero followed by the
846        literals. PCRE now does the same.
847    
848    22. Following Perl, added \o{} to specify codepoints in octal, making it
849        possible to specify values greater than 0777 and also making them
850        unambiguous.
851    
852    23. Perl now gives an error for missing closing braces after \x{... instead of
853        treating the string as literal. PCRE now does the same.
854    
855    24. RunTest used to grumble if an inappropriate test was selected explicitly,
856        but just skip it when running all tests. This make it awkward to run ranges
857        of tests when one of them was inappropriate. Now it just skips any
858        inappropriate tests, as it always did when running all tests.
859    
860    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
861        character types such as \d or \w, too many callouts were inserted, and the
862        data that they returned was rubbish.
863    
864    26. In UCP mode, \s was not matching two of the characters that Perl matches,
865        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
866        were matched by \h. The code has now been refactored so that the lists of
867        the horizontal and vertical whitespace characters used for \h and \v (which
868        are defined only in one place) are now also used for \s.
869    
870    27. Add JIT support for the 64 bit TileGX architecture.
871        Patch by Jiong Wang (Tilera Corporation).
872    
873    28. Possessive quantifiers for classes (both explicit and automatically
874        generated) now use special opcodes instead of wrapping in ONCE brackets.
875    
876    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
877        (because it's meaningless), this was not happening when PCRE_CASELESS was
878        set. Not wrong, but inefficient.
879    
880    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
881        properties for \w, \d, etc) is present in a test regex. Otherwise if the
882        test contains no characters greater than 255, Perl doesn't realise it
883        should be using Unicode semantics.
884    
885    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
886        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
887        does in Unicode mode.
888    
889    32. Added the "forbid" facility to pcretest so that putting tests into the
890        wrong test files can sometimes be quickly detected.
891    
892    33. There is now a limit (default 250) on the depth of nesting of parentheses.
893        This limit is imposed to control the amount of system stack used at compile
894        time. It can be changed at build time by --with-parens-nest-limit=xxx or
895        the equivalent in CMake.
896    
897    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
898        errors. Perl warns for these when in warning mode, but PCRE has no facility
899        for giving warnings.
900    
901    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
902        However, this was not working for (?!) because it is optimized to (*FAIL),
903        for which PCRE does not allow quantifiers. The optimization is now disabled
904        when a quantifier follows (?!). I can't see any use for this, but it makes
905        things uniform.
906    
907    36. Perl no longer allows group names to start with digits, so I have made this
908        change also in PCRE. It simplifies the code a bit.
909    
910    37. In extended mode, Perl ignores spaces before a + that indicates a
911        possessive quantifier. PCRE allowed a space before the quantifier, but not
912        before the possessive +. It now does.
913    
914    38. The use of \K (reset reported match start) within a repeated possessive
915        group such as (a\Kb)*+ was not working.
916    
917    40. Document that the same character tables must be used at compile time and
918        run time, and that the facility to pass tables to pcre_exec() and
919        pcre_dfa_exec() is for use only with saved/restored patterns.
920    
921    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
922        features for Builds with MSVC:
923    
924        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
925           builds). The .rc files can be used to set FileDescription and many other
926           attributes.
927    
928        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
929           This allows higher-level build scripts which want .pdb files to avoid
930           hard-coding the exact files needed."
931    
932    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
933        mean "start of word" and "end of word", respectively, as a transition aid.
934    
935    43. A minimizing repeat of a class containing codepoints greater than 255 in
936        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
937        compiled to use the heap for recursion.
938    
939    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
940        is configured.
941    
942    
943    Version 8.33 28-May-2013
944    ------------------------
945    
946    1.  Added 'U' to some constants that are compared to unsigned integers, to
947        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
948        variables that are added to signed variables, to ensure the result is
949        signed and can be negated.
950    
951    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
952        CMake config files.
953    
954    3.  Revise the creation of config.h.generic so that all boolean macros are
955        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
956        overriding via -D on the command line possible.
957    
958    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
959        to unsigned int is reported to make a quite noticeable speed difference in
960        a specific Windows environment. Testing on Linux did also appear to show
961        some benefit (and it is clearly not harmful). Also fixed the definition of
962        Xop which should be unsigned.
963    
964    5.  Related to (4), changing the definition of the intermediate variable cc
965        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
966        improvements.
967    
968    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
969        unnecessary spaces.
970    
971    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
972        and later.
973    
974    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
975    
976    9.  Optimizing fast_forward_start_bits in JIT.
977    
978    10. Adding support for callouts in JIT, and fixing some issues revealed
979        during this work. Namely:
980    
981        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
982    
983        (b) Minimum length was not checked before the matching is started.
984    
985    11. The value of capture_last that is passed to callouts was incorrect in some
986        cases when there was a capture on one path that was subsequently abandoned
987        after a backtrack. Also, the capture_last value is now reset after a
988        recursion, since all captures are also reset in this case.
989    
990    12. The interpreter no longer returns the "too many substrings" error in the
991        case when an overflowing capture is in a branch that is subsequently
992        abandoned after a backtrack.
993    
994    13. In the pathological case when an offset vector of size 2 is used, pcretest
995        now prints out the matched string after a yield of 0 or 1.
996    
997    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
998        Only supported by the JIT compiler at the moment.
999    
1000    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
1001    
1002    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
1003        offset of the starting point of the matching process, provided the offsets
1004        vector is large enough.
1005    
1006    17. The \A escape now records a lookbehind value of 1, though its execution
1007        does not actually inspect the previous character. This is to ensure that,
1008        in partial multi-segment matching, at least one character from the old
1009        segment is retained when a new segment is processed. Otherwise, if there
1010        are no lookbehinds in the pattern, \A might match incorrectly at the start
1011        of a new segment.
1012    
1013    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
1014    
1015    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
1016        gives some modest performance improvement in 8-bit mode.
1017    
1018    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1019        be expressed in certain programming languages using Universal Character
1020        Names.
1021    
1022    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1023        which points out that "non characters" are not "characters that may not
1024        appear in Unicode strings" but rather "characters that are reserved for
1025        internal use and have only local meaning".
1026    
1027    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1028        there was a conditional group that depended on an assertion, if the
1029        assertion was false, the callout that immediately followed the alternation
1030        in the condition was skipped when pcre_exec() was used for matching.
1031    
1032    23. Allow an explicit callout to be inserted before an assertion that is the
1033        condition for a conditional group, for compatibility with automatic
1034        callouts, which always insert a callout at this point.
1035    
1036    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1037        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1038    
1039    25. (*PRUNE) is now supported by the JIT compiler.
1040    
1041    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1042    
1043    27. Fix the case where there are two or more SKIPs with arguments that may be
1044        ignored.
1045    
1046    28. (*SKIP) is now supported by the JIT compiler.
1047    
1048    29. (*THEN) is now supported by the JIT compiler.
1049    
1050    30. Update RunTest with additional test selector options.
1051    
1052    31. The way PCRE handles backtracking verbs has been changed in two ways.
1053    
1054        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1055        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1056        backtracking. In some cases this makes it more Perl-compatible, but Perl's
1057        rather obscure rules do not always do the same thing.
1058    
1059        (2) Previously, backtracking verbs were confined within assertions. This is
1060        no longer the case for positive assertions, except for (*ACCEPT). Again,
1061        this sometimes improves Perl compatibility, and sometimes does not.
1062    
1063    32. A number of tests that were in test 2 because Perl did things differently
1064        have been moved to test 1, because either Perl or PCRE has changed, and
1065        these tests are now compatible.
1066    
1067    32. Backtracking control verbs are now handled in the same way in JIT and
1068        interpreter.
1069    
1070    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1071        contained a forward subroutine reference caused a compile error.
1072    
1073    34. Auto-detect and optimize limited repetitions in JIT.
1074    
1075    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1076        blocking (*UTF) etc.
1077    
1078    36. In the interpreter, maximizing pattern repetitions for characters and
1079        character types now use tail recursion, which reduces stack usage.
1080    
1081    37. The value of the max lookbehind was not correctly preserved if a compiled
1082        and saved regex was reloaded on a host of different endianness.
1083    
1084    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1085        of the compiled pattern block, expand the flags field from 16 to 32 bits
1086        because it was almost full.
1087    
1088    39. Try madvise first before posix_madvise.
1089    
1090    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1091        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1092        original change locked out all matches of empty strings. This has been
1093        changed so that one match of an empty string per line is recognized.
1094        Subsequent searches on the same line (for colouring or for --only-matching,
1095        for example) do not recognize empty strings.
1096    
1097    41. Applied a user patch to fix a number of spelling mistakes in comments.
1098    
1099    42. Data lines longer than 65536 caused pcretest to crash.
1100    
1101    43. Clarified the data type for length and startoffset arguments for pcre_exec
1102        and pcre_dfa_exec in the function-specific man pages, where they were
1103        explicitly stated to be in bytes, never having been updated. I also added
1104        some clarification to the pcreapi man page.
1105    
1106    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1107        a segmentation fault.
1108    
1109    
1110    Version 8.32 30-November-2012
1111    -----------------------------
1112    
1113    1.  Improved JIT compiler optimizations for first character search and single
1114        character iterators.
1115    
1116    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1117        Patch by Daniel Richard G.
1118    
1119    3.  Single character iterator optimizations in the JIT compiler.
1120    
1121    4.  Improved JIT compiler optimizations for character ranges.
1122    
1123    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1124        Reported by Giuseppe D'Angelo.
1125    
1126    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1127        of a line, was being set incorrectly in cases where .* appeared inside
1128        atomic brackets at the start of a pattern, or where there was a subsequent
1129        *PRUNE or *SKIP.
1130    
1131    7.  Improved instruction cache flush for POWER/PowerPC.
1132        Patch by Daniel Richard G.
1133    
1134    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1135        grep:
1136    
1137        (a) There is now no limit to the number of patterns to be matched.
1138    
1139        (b) An error is given if a pattern is too long.
1140    
1141        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1142            are now supported.
1143    
1144        (d) --exclude-from and --include-from (multiple use) have been added.
1145    
1146        (e) Exclusions and inclusions now apply to all files and directories, not
1147            just to those obtained from scanning a directory recursively.
1148    
1149        (f) Multiple uses of -f and --file-list are now supported.
1150    
1151        (g) In a Windows environment, the default for -d has been changed from
1152            "read" (the GNU grep default) to "skip", because otherwise the presence
1153            of a directory in the file list provokes an error.
1154    
1155        (h) The documentation has been revised and clarified in places.
1156    
1157    9.  Improve the matching speed of capturing brackets.
1158    
1159    10. Changed the meaning of \X so that it now matches a Unicode extended
1160        grapheme cluster.
1161    
1162    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1163        out POSIX threads when JIT support is configured.
1164    
1165    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1166    
1167    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1168        structure could go wrong in environments where size_t is not the same size
1169        as int.
1170    
1171    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1172        set.
1173    
1174    15. The EBCDIC support had decayed; later updates to the code had included
1175        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1176        general tidy up of EBCDIC-related issues, and the documentation was also
1177        not quite right. There is now a test that can be run on ASCII systems to
1178        check some of the EBCDIC-related things (but is it not a full test).
1179    
1180    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1181        in a small tidy to the code.
1182    
1183    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1184    
1185    18. If the --only-matching (-o) option in pcregrep is specified multiple
1186        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1187        substrings matched by the 1st and 2nd capturing parentheses. A separating
1188        string can be specified by --om-separator (default empty).
1189    
1190    19. Improving the first n character searches.
1191    
1192    20. Turn case lists for horizontal and vertical white space into macros so that
1193        they are defined only once.
1194    
1195    21. This set of changes together give more compatible Unicode case-folding
1196        behaviour for characters that have more than one other case when UCP
1197        support is available.
1198    
1199        (a) The Unicode property table now has offsets into a new table of sets of
1200            three or more characters that are case-equivalent. The MultiStage2.py
1201            script that generates these tables (the pcre_ucd.c file) now scans
1202            CaseFolding.txt instead of UnicodeData.txt for character case
1203            information.
1204    
1205        (b) The code for adding characters or ranges of characters to a character
1206            class has been abstracted into a generalized function that also handles
1207            case-independence. In UTF-mode with UCP support, this uses the new data
1208            to handle characters with more than one other case.
1209    
1210        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1211            whose other case is greater than 256 are now correctly matched
1212            caselessly. Previously, the high codepoint matched the low one, but not
1213            vice versa.
1214    
1215        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1216            of the new class addition function, using character lists defined as
1217            macros alongside the case definitions of 20 above.
1218    
1219        (e) Caseless back references now work with characters that have more than
1220            one other case.
1221    
1222        (f) General caseless matching of characters with more than one other case
1223            is supported.
1224    
1225    22. Unicode character properties were updated from Unicode 6.2.0
1226    
1227    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1228    
1229    24. Add support for 32-bit character strings, and UTF-32
1230    
1231    25. Major JIT compiler update (code refactoring and bugfixing).
1232        Experimental Sparc 32 support is added.
1233    
1234    26. Applied a modified version of Daniel Richard G's patch to create
1235        pcre.h.generic and config.h.generic by "make" instead of in the
1236        PrepareRelease script.
1237    
1238    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1239        pcre_compile.c when checking for a zero character.
1240    
1241    28. Introducing a native interface for JIT. Through this interface, the compiled
1242        machine code can be directly executed. The purpose of this interface is to
1243        provide fast pattern matching, so several sanity checks are not performed.
1244        However, feature tests are still performed. The new interface provides
1245        1.4x speedup compared to the old one.
1246    
1247    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1248        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1249        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1250    
1251    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1252        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1253        the "old" RFC 2279). Instead, it ended up passing a negative length to
1254        pcre_exec().
1255    
1256    31. Add support for GCC's visibility feature to hide internal functions.
1257    
1258    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1259        "unknown -C option" after outputting 0 or 1.
1260    
1261    33. There is now support for generating a code coverage report for the test
1262        suite in environments where gcc is the compiler and lcov is installed. This
1263        is mainly for the benefit of the developers.
1264    
1265    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1266        unaddressable using valgrind annotations, allowing valgrind to detect
1267        invalid memory accesses. This is mainly for the benefit of the developers.
1268    
1269    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1270    
1271    26. Give configure error if --enable-cpp but no C++ compiler found.
1272    
1273    
1274    Version 8.31 06-July-2012
1275    -------------------------
1276    
1277    1.  Fixing a wrong JIT test case and some compiler warnings.
1278    
1279    2.  Removed a bashism from the RunTest script.
1280    
1281    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1282        to unsigned type, result still unsigned" that was given by an MS compiler
1283        on encountering the code "-sizeof(xxx)".
1284    
1285    4.  Partial matching support is added to the JIT compiler.
1286    
1287    5.  Fixed several bugs concerned with partial matching of items that consist
1288        of more than one character:
1289    
1290        (a) /^(..)\1/ did not partially match "aba" because checking references was
1291            done on an "all or nothing" basis. This also applied to repeated
1292            references.
1293    
1294        (b) \R did not give a hard partial match if \r was found at the end of the
1295            subject.
1296    
1297        (c) \X did not give a hard partial match after matching one or more
1298            characters at the end of the subject.
1299    
1300        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1301            a partial match for the string "\r".
1302    
1303        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1304            a partial match for a CR character at the end of the subject string.
1305    
1306    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1307        running pcretest, the text "(JIT)" added to the output whenever JIT is
1308        actually used to run the match.
1309    
1310    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1311        or /S+[+] with a digit between 1 and 7.
1312    
1313    8.  OP_NOT now supports any UTF character not just single-byte ones.
1314    
1315    9.  (*MARK) control verb is now supported by the JIT compiler.
1316    
1317    10. The command "./RunTest list" lists the available tests without actually
1318        running any of them. (Because I keep forgetting what they all are.)
1319    
1320    11. Add PCRE_INFO_MAXLOOKBEHIND.
1321    
1322    12. Applied a (slightly modified) user-supplied patch that improves performance
1323        when the heap is used for recursion (compiled with --disable-stack-for-
1324        recursion). Instead of malloc and free for each heap frame each time a
1325        logical recursion happens, frames are retained on a chain and re-used where
1326        possible. This sometimes gives as much as 30% improvement.
1327    
1328    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1329        call.
1330    
1331    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1332    
1333    15. It is now possible to link pcretest with libedit as an alternative to
1334        libreadline.
1335    
1336    16. (*COMMIT) control verb is now supported by the JIT compiler.
1337    
1338    17. The Unicode data tables have been updated to Unicode 6.1.0.
1339    
1340    18. Added --file-list option to pcregrep.
1341    
1342    19. Added binary file support to pcregrep, including the -a, --binary-files,
1343        -I, and --text options.
1344    
1345    20. The madvise function is renamed for posix_madvise for QNX compatibility
1346        reasons. Fixed by Giuseppe D'Angelo.
1347    
1348    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1349    
1350    22. Changed the option for creating symbolic links for 16-bit man pages from
1351        -s to -sf so that re-installing does not cause issues.
1352    
1353    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1354    
1355    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1356        matches in certain environments (the workspace was not being correctly
1357        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1358        some of the workspace data at the beginning of a restart.
1359    
1360    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1361        was not doing so when it should - probably a typo introduced by SVN 528
1362        (change 8.10/14).
1363    
1364    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1365        \w+ when the character tables indicated that \x{c4} was a word character.
1366        There were several related cases, all because the tests for doing a table
1367        lookup were testing for characters less than 127 instead of 255.
1368    
1369    27. If a pattern contains capturing parentheses that are not used in a match,
1370        their slots in the ovector are set to -1. For those that are higher than
1371        any matched groups, this happens at the end of processing. In the case when
1372        there were back references that the ovector was too small to contain
1373        (causing temporary malloc'd memory to be used during matching), and the
1374        highest capturing number was not used, memory off the end of the ovector
1375        was incorrectly being set to -1. (It was using the size of the temporary
1376        memory instead of the true size.)
1377    
1378    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1379        ovector size, it uses memory at the end of the block that it has got.
1380    
1381    29. Check for an overlong MARK name and give an error at compile time. The
1382        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1383    
1384    30. JIT compiler update.
1385    
1386    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1387        Rill for the patch.
1388    
1389    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1390    
1391    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1392    
1393    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1394        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1395        was enabled.
1396    
1397    35. Improve JIT code generation for greedy plus quantifier.
1398    
1399    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1400        1 to "aa" instead of to an empty string. The bug affected repeated groups
1401        that could potentially match an empty string.
1402    
1403    37. Optimizing single character iterators in JIT.
1404    
1405    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1406        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1407        codepoints that are too big for the mode are faulted, and in a UTF mode,
1408        disallowed codepoints are also faulted.
1409    
1410    39. If PCRE was compiled with UTF support, in three places in the DFA
1411        matcher there was code that should only have been obeyed in UTF mode, but
1412        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1413        processing when bytes with values greater than 127 were present. In 16-bit
1414        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1415        both cases the values are those that cannot be the first data item in a UTF
1416        character. The three items that might have provoked this were recursions,
1417        possessively repeated groups, and atomic groups.
1418    
1419    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1420        and pcregrep, because some OS require shared objects to be explicitly
1421        passed to ld, causing the link step to fail if they are not.
1422    
1423    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1424        mode, patterns that started with \h* or \R* might be incorrectly matched.
1425    
1426    
1427    Version 8.30 04-February-2012
1428    -----------------------------
1429    
1430    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1431        name is defined in ctype.h.
1432    
1433    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1434        only in quite long subpatterns.
1435    
1436    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1437        since it was replaced by pcre_fullinfo() in February 2000.
1438    
1439    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1440        match a (*MARK), and the match failed at the start of the subject, a
1441        reference to memory before the start of the subject could occur. This bug
1442        was introduced by fix 17 of release 8.21.
1443    
1444    5.  A reference to an unset group with zero minimum repetition was giving
1445        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1446        /(another)?(\1?)test/ matched against "hello world test". This bug was
1447        introduced in release 8.13.
1448    
1449    6.  Add support for 16-bit character strings (a large amount of work involving
1450        many changes and refactorings).
1451    
1452    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1453        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1454        from a file.
1455    
1456    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1457        rounding is not applied in this particular case).
1458    
1459    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1460        if they appear, or are escaped, in patterns.
1461    
1462    10. Get rid of a number of -Wunused-but-set-variable warnings.
1463    
1464    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1465        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1466        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1467        also returns the mark "x". This bug applied to capturing parentheses,
1468        non-capturing parentheses, and atomic parentheses. It also applied to some
1469        assertions.
1470    
1471    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1472        information out of configure.ac instead of relying on pcre.h.generic, which
1473        is not stored in the repository.
1474    
1475    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1476        -lreadline.
1477    
1478    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1479    
1480    15. Applied Graycode's patch to put the top-level frame on the stack rather
1481        than the heap when not using the stack for recursion. This gives a
1482        performance improvement in many cases when recursion is not deep.
1483    
1484    16. Experimental code added to "pcretest -C" to output the stack frame size.
1485    
1486    
1487    Version 8.21 12-Dec-2011
1488    ------------------------
1489    
1490    1.  Updating the JIT compiler.
1491    
1492    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1493        are added as well.
1494    
1495    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1496        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1497        calling _pcre_jit_exec. Some extra comments are added.
1498    
1499    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1500        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1501        was introduced by change 18 for 8.20.
1502    
1503    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1504        ECMA-262 standard.
1505    
1506    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1507        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1508        This bug was probably introduced by change 9 of 8.13.
1509    
1510    7.  While fixing 6 above, I noticed that a number of other items were being
1511        incorrectly rejected as "not fixed length". This arose partly because newer
1512        opcodes had not been added to the fixed-length checking code. I have (a)
1513        corrected the bug and added tests for these items, and (b) arranged for an
1514        error to occur if an unknown opcode is encountered while checking for fixed
1515        length instead of just assuming "not fixed length". The items that were
1516        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1517        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1518        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1519    
1520    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1521        being incorrectly compiled and would have given unpredicatble results.
1522    
1523    9.  A possessively repeated subpattern with minimum repeat count greater than
1524        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1525        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1526        the first (A) could occur when it should not.
1527    
1528    10. Add a cast and remove a redundant test from the code.
1529    
1530    11. JIT should use pcre_malloc/pcre_free for allocation.
1531    
1532    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1533        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1534        is anything other than /usr, -L is still shown).
1535    
1536    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1537    
1538    14. Perl does not support \N without a following name in a [] class; PCRE now
1539        also gives an error.
1540    
1541    15. If a forward reference was repeated with an upper limit of around 2000,
1542        it caused the error "internal error: overran compiling workspace". The
1543        maximum number of forward references (including repeats) was limited by the
1544        internal workspace, and dependent on the LINK_SIZE. The code has been
1545        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1546        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1547        of around 200,000 forward references. While doing this, I also speeded up
1548        the filling in of repeated forward references.
1549    
1550    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1551        incorrectly expecting the subject to contain another "a" after the start.
1552    
1553    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1554        in the match, the SKIP should be ignored. This was not happening; instead
1555        the SKIP was being treated as NOMATCH. For patterns such as
1556        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1557        tested.
1558    
1559    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1560        now much more compatible with Perl, in particular in cases where the result
1561        is a non-match for a non-anchored pattern. For example, if
1562        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1563        "m", where previously it did not return a name. A side effect of this
1564        change is that for partial matches, the last encountered mark name is
1565        returned, as for non matches. A number of tests that were previously not
1566        Perl-compatible have been moved into the Perl-compatible test files. The
1567        refactoring has had the pleasing side effect of removing one argument from
1568        the match() function, thus reducing its stack requirements.
1569    
1570    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1571        subsequent uses of /S (without +) incorrectly behaved like /S+.
1572    
1573    21. Retrieve executable code size support for the JIT compiler and fixing
1574        some warnings.
1575    
1576    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1577        not work when the shorter character appeared right at the end of the
1578        subject string.
1579    
1580    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1581        systems.
1582    
1583    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1584        output it when the /M option is used in pcretest.
1585    
1586    25. The CheckMan script was not being included in the distribution. Also, added
1587        an explicit "perl" to run Perl scripts from the PrepareRelease script
1588        because this is reportedly needed in Windows.
1589    
1590    26. If study data was being save in a file and studying had not found a set of
1591        "starts with" bytes for the pattern, the data written to the file (though
1592        never used) was taken from uninitialized memory and so caused valgrind to
1593        complain.
1594    
1595    27. Updated RunTest.bat as provided by Sheri Pierce.
1596    
1597    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1598    
1599    29. Computation of memory usage for the table of capturing group names was
1600        giving an unnecessarily large value.
1601    
1602    
1603    Version 8.20 21-Oct-2011
1604    ------------------------
1605    
1606    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1607        a POSIX class. After further experiments with Perl, which convinced me that
1608        Perl has bugs and confusions, a closing square bracket is no longer allowed
1609        in a POSIX name. This bug also affected patterns with classes that started
1610        with full stops.
1611    
1612    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1613        captured substring, but while checking the failing first alternative,
1614        substring 1 is temporarily captured. If the output vector supplied to
1615        pcre_exec() was not big enough for this capture, the yield of the function
1616        was still zero ("insufficient space for captured substrings"). This cannot
1617        be totally fixed without adding another stack variable, which seems a lot
1618        of expense for a edge case. However, I have improved the situation in cases
1619        such as /(a)(b)x|abc/ matched against "abc", where the return code
1620        indicates that fewer than the maximum number of slots in the ovector have
1621        been set.
1622    
1623    3.  Related to (2) above: when there are more back references in a pattern than
1624        slots in the output vector, pcre_exec() uses temporary memory during
1625        matching, and copies in the captures as far as possible afterwards. It was
1626        using the entire output vector, but this conflicts with the specification
1627        that only 2/3 is used for passing back captured substrings. Now it uses
1628        only the first 2/3, for compatibility. This is, of course, another edge
1629        case.
1630    
1631    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1632        main code base, and can be used by building with --enable-jit. When this is
1633        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1634        runtime --no-jit option is given.
1635    
1636    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1637        ovector, the return from the function was zero, implying that there were
1638        other matches that did not fit. The correct "exactly full" value is now
1639        returned.
1640    
1641    6.  If a subpattern that was called recursively or as a subroutine contained
1642        (*PRUNE) or any other control that caused it to give a non-standard return,
1643        invalid errors such as "Error -26 (nested recursion at the same subject
1644        position)" or even infinite loops could occur.
1645    
1646    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1647        computing the minimum length on reaching *ACCEPT, and so ended up with the
1648        wrong value of 1 rather than 0. Further investigation indicates that
1649        computing a minimum subject length in the presence of *ACCEPT is difficult
1650        (think back references, subroutine calls), and so I have changed the code
1651        so that no minimum is registered for a pattern that contains *ACCEPT.
1652    
1653    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1654        it was not handled as intended. [But see 16 below.]
1655    
1656    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1657        Sheri Pierce.
1658    
1659    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1660        the first byte in a match must be "a".
1661    
1662    11. Change 17 for 8.13 increased the recursion depth for patterns like
1663        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1664        pattern contains any instances of (*THEN). If it does not, the old
1665        optimizations are restored. It would be nice to do this on a per-group
1666        basis, but at the moment that is not feasible.
1667    
1668    12. In some environments, the output of pcretest -C is CRLF terminated. This
1669        broke RunTest's code that checks for the link size. A single white space
1670        character after the value is now allowed for.
1671    
1672    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1673        For "fr", it uses the Windows-specific input and output files.
1674    
1675    14. If (*THEN) appeared in a group that was called recursively or as a
1676        subroutine, it did not work as intended. [But see next item.]
1677    
1678    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1679        pattern fragments (but not containing any | characters). If A and B are
1680        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1681        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1682        D. In other words, Perl considers parentheses that do not contain any |
1683        characters to be part of a surrounding alternative, whereas PCRE was
1684        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1685        differently. PCRE now behaves in the same way as Perl, except in the case
1686        of subroutine/recursion calls such as (?1) which have in any case always
1687        been different (but PCRE had them first :-).
1688    
1689    16. Related to 15 above: Perl does not treat the | in a conditional group as
1690        creating alternatives. Such a group is treated in the same way as an
1691        ordinary group without any | characters when processing (*THEN). PCRE has
1692        been changed to match Perl's behaviour.
1693    
1694    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1695        RunGrepTest script failed.
1696    
1697    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1698        inevitable for groups that contain captures, but it can lead to a lot of
1699        stack use in large patterns. The old behaviour has been restored for atomic
1700        groups that do not contain any capturing parentheses.
1701    
1702    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1703        suppress the check for a minimum subject length at run time. (If it was
1704        given to pcre_exec() or pcre_dfa_exec() it did work.)
1705    
1706    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1707        fail to work when decoding hex characters in data strings in EBCDIC
1708        environments.
1709    
1710    21. It appears that in at least one Mac OS environment, the isxdigit() function
1711        is implemented as a macro that evaluates to its argument more than once,
1712        contravening the C 90 Standard (I haven't checked a later standard). There
1713        was an instance in pcretest which caused it to go wrong when processing
1714        \x{...} escapes in subject strings. The has been rewritten to avoid using
1715        things like p++ in the argument of isxdigit().
1716    
1717    
1718    Version 8.13 16-Aug-2011
1719    ------------------------
1720    
1721    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1722    
1723    2.  Two minor typos in pcre_internal.h have been fixed.
1724    
1725    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1726        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1727        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1728    
1729    4.  There were a number of related bugs in the code for matching backrefences
1730        caselessly in UTF-8 mode when codes for the characters concerned were
1731        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1732        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1733        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1734        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1735        2-byte code at the end of the subject (it thought there wasn't enough data
1736        left).
1737    
1738    5.  Comprehensive information about what went wrong is now returned by
1739        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1740        as the output vector has at least 2 elements. The offset of the start of
1741        the failing character and a reason code are placed in the vector.
1742    
1743    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1744        now returned is for the first byte of the failing character, instead of the
1745        last byte inspected. This is an incompatible change, but I hope it is small
1746        enough not to be a problem. It makes the returned offset consistent with
1747        pcre_exec() and pcre_dfa_exec().
1748    
1749    7.  pcretest now gives a text phrase as well as the error number when
1750        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1751        failure, the offset and reason code are output.
1752    
1753    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1754        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1755        back over a single character (\n). This seems wrong (because it treated the
1756        two characters as a single entity when going forwards), conflicts with the
1757        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1758        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1759        has been changed.
1760    
1761    9.  Some internal refactoring has changed the processing so that the handling
1762        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1763        time (the PCRE_DOTALL option was changed this way some time ago: version
1764        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1765        which was always a bit of a fudge. It also means that there is one less
1766        argument for the match() function, which reduces its stack requirements
1767        slightly. This change also fixes an incompatibility with Perl: the pattern
1768        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1769    
1770    10. More internal refactoring has drastically reduced the number of recursive
1771        calls to match() for possessively repeated groups such as (abc)++ when
1772        using pcre_exec().
1773    
1774    11. While implementing 10, a number of bugs in the handling of groups were
1775        discovered and fixed:
1776    
1777        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1778        (a|)*(?1) gave a compile-time internal error.
1779        ((a|)+)+  did not notice that the outer group could match an empty string.
1780        (^a|^)+   was not marked as anchored.
1781        (.*a|.*)+ was not marked as matching at start or after a newline.
1782    
1783    12. Yet more internal refactoring has removed another argument from the match()
1784        function. Special calls to this function are now indicated by setting a
1785        value in a variable in the "match data" data block.
1786    
1787    13. Be more explicit in pcre_study() instead of relying on "default" for
1788        opcodes that mean there is no starting character; this means that when new
1789        ones are added and accidentally left out of pcre_study(), testing should
1790        pick them up.
1791    
1792    14. The -s option of pcretest has been documented for ages as being an old
1793        synonym of -m (show memory usage). I have changed it to mean "force study
1794        for every regex", that is, assume /S for every regex. This is similar to -i
1795        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1796        using it. It makes it easier to run collections of tests with and without
1797        study enabled, and thereby test pcre_study() more easily. All the standard
1798        tests are now run with and without -s (but some patterns can be marked as
1799        "never study" - see 20 below).
1800    
1801    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1802        restoration of the capturing data to the outer values was not happening
1803        correctly.
1804    
1805    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1806        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1807        pattern had matched an empty string, and so incorrectly returned a no
1808        match.
1809    
1810    17. There was optimizing code for the last branch of non-capturing parentheses,
1811        and also for the obeyed branch of a conditional subexpression, which used
1812        tail recursion to cut down on stack usage. Unfortunately, now that there is
1813        the possibility of (*THEN) occurring in these branches, tail recursion is
1814        no longer possible because the return has to be checked for (*THEN). These
1815        two optimizations have therefore been removed. [But see 8.20/11 above.]
1816    
1817    18. If a pattern containing \R was studied, it was assumed that \R always
1818        matched two bytes, thus causing the minimum subject length to be
1819        incorrectly computed because \R can also match just one byte.
1820    
1821    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1822        was incorrectly computed.
1823    
1824    20. If /S is present twice on a test pattern in pcretest input, it now
1825        *disables* studying, thereby overriding the use of -s on the command line
1826        (see 14 above). This is necessary for one or two tests to keep the output
1827        identical in both cases.
1828    
1829    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1830        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1831    
1832    22. When an atomic group that contained a capturing parenthesis was
1833        successfully matched, but the branch in which it appeared failed, the
1834        capturing was not being forgotten if a higher numbered group was later
1835        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1836        group 1 to "a", when in fact it should be unset. This applied to multi-
1837        branched capturing and non-capturing groups, repeated or not, and also to
1838        positive assertions (capturing in negative assertions does not happen
1839        in PCRE) and also to nested atomic groups.
1840    
1841    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1842        subject after a captured substring, to make it easier to tell which of a
1843        number of identical substrings has been captured.
1844    
1845    24. The way atomic groups are processed by pcre_exec() has been changed so that
1846        if they are repeated, backtracking one repetition now resets captured
1847        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1848        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1849        "aaa". Previously, it would have been "a". As part of this code
1850        refactoring, the way recursive calls are handled has also been changed.
1851    
1852    25. If an assertion condition captured any substrings, they were not passed
1853        back unless some other capturing happened later. For example, if
1854        (?(?=(a))a) was matched against "a", no capturing was returned.
1855    
1856    26. When studying a pattern that contained subroutine calls or assertions,
1857        the code for finding the minimum length of a possible match was handling
1858        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1859        group 1 called group 2 while simultaneously a separate group 2 called group
1860        1). A stack overflow occurred in this case. I have fixed this by limiting
1861        the recursion depth to 10.
1862    
1863    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1864        Fortmann. This supports explicit test numbers on the command line, and has
1865        argument validation and error reporting.
1866    
1867    28. An instance of \X with an unlimited repeat could fail if at any point the
1868        first character it looked at was a mark character.
1869    
1870    29. Some minor code refactoring concerning Unicode properties and scripts
1871        should reduce the stack requirement of match() slightly.
1872    
1873    30. Added the '=' option to pcretest to check the setting of unused capturing
1874        slots at the end of the pattern, which are documented as being -1, but are
1875        not included in the return count.
1876    
1877    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1878        compiled something random. Now it gives a compile-time error (as does
1879        Perl).
1880    
1881    32. A *MARK encountered during the processing of a positive assertion is now
1882        recorded and passed back (compatible with Perl).
1883    
1884    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1885        had alternative anchored branches, the search for a second match in a line
1886        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1887        matched the line "0102" twice. The same bug affected patterns that started
1888        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1889        twice.
1890    
1891    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1892        does, and because of capturing effects, quantifying parenthesized
1893        assertions may at times be useful. Quantifiers are now allowed for
1894        parenthesized assertions.
1895    
1896    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1897    
1898    36. \g was being checked for fancy things in a character class, when it should
1899        just be a literal "g".
1900    
1901    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1902        appearance of a nested POSIX class supersedes an apparent external class.
1903        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1904        unescaped square brackets may also appear as part of class names. For
1905        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1906        more like Perl. (But see 8.20/1 above.)
1907    
1908    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1909        was because it thought it was \N{name}, which is not supported).
1910    
1911    39. Add minix to OS list not supporting the -S option in pcretest.
1912    
1913    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1914        cannot analyze patterns in sufficient detail to catch mutual recursions
1915        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1916        subgroup is called recursively as a subpattern for a second time at the
1917        same position in the subject string. In previous releases this might have
1918        been caught by the recursion limit, or it might have run out of stack.
1919    
1920    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1921        happen only once. PCRE was, however incorrectly giving a compile time error
1922        "recursive call could loop indefinitely" because it cannot analyze the
1923        pattern in sufficient detail. The compile time test no longer happens when
1924        PCRE is compiling a conditional subpattern, but actual runaway loops are
1925        now caught at runtime (see 40 above).
1926    
1927    42. It seems that Perl allows any characters other than a closing parenthesis
1928        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1929        has been changed to be the same.
1930    
1931    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1932        as not to get warnings when autogen.sh is called. Also changed
1933        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1934    
1935    44. To help people who use pcregrep to scan files containing exceedingly long
1936        lines, the following changes have been made:
1937    
1938        (a) The default value of the buffer size parameter has been increased from
1939            8K to 20K. (The actual buffer used is three times this size.)
1940    
1941        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1942            PCRE is built.
1943    
1944        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1945            to be set at run time.
1946    
1947        (d) Numerical values in pcregrep options can be followed by K or M, for
1948            example --buffer-size=50K.
1949    
1950        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1951            given and the return code is set to 2.
1952    
1953    45. Add a pointer to the latest mark to the callout data block.
1954    
1955    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1956        partial match of an empty string instead of no match. This was specific to
1957        the use of ".".
1958    
1959    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1960        complete match instead of a partial match. This bug was dependent on both
1961        the PCRE_UTF8 and PCRE_DOTALL options being set.
1962    
1963    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1964        starting byte set, because \b was not being ignored.
1965    
1966    
1967    Version 8.12 15-Jan-2011
1968    ------------------------
1969    
1970    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1971        checks for such things as part of the documentation building process.
1972    
1973    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1974        --match-limit and --recursion-limit options (added for 8.11). In
1975        particular, this made one of the standard tests fail. (The integer value
1976        went into the wrong half of a long int.)
1977    
1978    3.  If the --colour option was given to pcregrep with -v (invert match), it
1979        did strange things, either producing crazy output, or crashing. It should,
1980        of course, ignore a request for colour when reporting lines that do not
1981        match.
1982    
1983    4.  Another pcregrep bug caused similar problems if --colour was specified with
1984        -M (multiline) and the pattern match finished with a line ending.
1985    
1986    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1987        matched in multiline mode, the following line was shown as part of the
1988        match. This seems wrong, so I have changed it.
1989    
1990    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1991        the check for further matches in the same line (so they could be coloured)
1992        to overrun the end of the current line. If another match was found, it was
1993        incorrectly shown (and then shown again when found in the next line).
1994    
1995    7.  If pcregrep was compiled under Windows, there was a reference to the
1996        function pcregrep_exit() before it was defined. I am assuming this was
1997        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1998        reported by a user. I've moved the definition above the reference.
1999    
2000    
2001    Version 8.11 10-Dec-2010
2002    ------------------------
2003    
2004    1.  (*THEN) was not working properly if there were untried alternatives prior
2005        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
2006        backtracked to try for "b" instead of moving to the next alternative branch
2007        at the same level (in this case, to look for "c"). The Perl documentation
2008        is clear that when (*THEN) is backtracked onto, it goes to the "next
2009        alternative in the innermost enclosing group".
2010    
2011    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
2012        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
2013        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
2014        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
2015        (*THEN).
2016    
2017    3.  If \s appeared in a character class, it removed the VT character from
2018        the class, even if it had been included by some previous item, for example
2019        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
2020        of \s, but is part of the POSIX "space" class.)
2021    
2022    4.  A partial match never returns an empty string (because you can always
2023        match an empty string at the end of the subject); however the checking for
2024        an empty string was starting at the "start of match" point. This has been
2025        changed to the "earliest inspected character" point, because the returned
2026        data for a partial match starts at this character. This means that, for
2027        example, /(?<=abc)def/ gives a partial match for the subject "abc"
2028        (previously it gave "no match").
2029    
2030    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
2031        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
2032        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
2033        has an implication that the given string is incomplete (because a partial
2034        match is preferred over a full match). For this reason, these items now
2035        give a partial match in this situation. [Aside: previously, the one case
2036        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
2037        match rather than a full match, which was wrong by the old rules, but is
2038        now correct.]
2039    
2040    6.  There was a bug in the handling of #-introduced comments, recognized when
2041        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
2042        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
2043        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
2044        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
2045        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
2046        places in pcre_compile().
2047    
2048    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
2049        comments when looking ahead for named forward references to subpatterns,
2050        the only newline sequence it recognized was NL. It now handles newlines
2051        according to the set newline convention.
2052    
2053    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
2054        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
2055        cater for a lack of strerror(). These oversights have been fixed.
2056    
2057    9.  Added --match-limit and --recursion-limit to pcregrep.
2058    
2059    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
2060    
2061    11. When the -o option was used, pcregrep was setting a return code of 1, even
2062        when matches were found, and --line-buffered was not being honoured.
2063    
2064    12. Added an optional parentheses number to the -o and --only-matching options
2065        of pcregrep.
2066    
2067    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2068        can match an empty string. The code to do it in pcretest and pcredemo
2069        needed fixing:
2070    
2071        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2072            only one byte after an empty string match just before CRLF (this case
2073            just got forgotten; "any" and "anycrlf" were OK).
2074    
2075        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2076            mode when an empty string match preceded an ASCII character followed by
2077            a non-ASCII character. (The code for advancing by one character rather
2078            than one byte was nonsense.)
2079    
2080        (c) The pcredemo.c sample program did not have any code at all to handle
2081            the cases when CRLF is a valid newline sequence.
2082    
2083    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2084        as a starting offset was within the subject string. There is now a new
2085        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2086        negative or greater than the length of the string. In order to test this,
2087        pcretest is extended to allow the setting of negative starting offsets.
2088    
2089    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2090        starting offset points to the beginning of a UTF-8 character was
2091        unnecessarily clumsy. I tidied it up.
2092    
2093    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2094        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2095    
2096    17. Nobody had reported that the --include_dir option, which was added in
2097        release 7.7 should have been called --include-dir (hyphen, not underscore)
2098        for compatibility with GNU grep. I have changed it to --include-dir, but
2099        left --include_dir as an undocumented synonym, and the same for
2100        --exclude-dir, though that is not available in GNU grep, at least as of
2101        release 2.5.4.
2102    
2103    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2104        characters from a string of bytes) have been redefined so as not to use
2105        loops, in order to improve performance in some environments. At the same
2106        time, I abstracted some of the common code into auxiliary macros to save
2107        repetition (this should not affect the compiled code).
2108    
2109    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2110        compile-time error is now given if \c is not followed by an ASCII
2111        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2112        different, and any byte value is allowed.)
2113    
2114    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2115        START_OPTIMIZE option, which is now allowed at compile time - but just
2116        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2117        to pcregrep and other applications that have no direct access to PCRE
2118        options. The new /Y option in pcretest sets this option when calling
2119        pcre_compile().
2120    
2121    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2122        back references. Groups containing recursive back references were forced to
2123        be atomic by that change, but in the case of named groups, the amount of
2124        memory required was incorrectly computed, leading to "Failed: internal
2125        error: code overflow". This has been fixed.
2126    
2127    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2128        pcretest.c, to avoid build problems in some Borland environments.
2129    
2130    
2131    Version 8.10 25-Jun-2010
2132    ------------------------
2133    
2134    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2135        THEN.
2136    
2137    2.  (*ACCEPT) was not working when inside an atomic group.
2138    
2139    3.  Inside a character class, \B is treated as a literal by default, but
2140        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2141        causes the error). The code is unchanged, but I tidied the documentation.
2142    
2143    4.  Inside a character class, PCRE always treated \R and \X as literals,
2144        whereas Perl faults them if its -w option is set. I have changed PCRE so
2145        that it faults them when PCRE_EXTRA is set.
2146    
2147    5.  Added support for \N, which always matches any character other than
2148        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2149    
2150    6.  When compiling pcregrep with newer versions of gcc which may have
2151        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2152        declared with attribute warn_unused_result" were given. Just casting the
2153        result to (void) does not stop the warnings; a more elaborate fudge is
2154        needed. I've used a macro to implement this.
2155    
2156    7.  Minor change to pcretest.c to avoid a compiler warning.
2157    
2158    8.  Added four artifical Unicode properties to help with an option to make
2159        \s etc use properties (see next item). The new properties are: Xan
2160        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2161    
2162    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2163        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2164        this option. Modified pcretest to add /W to test this facility. Added
2165        REG_UCP to make it available via the POSIX interface.
2166    
2167    10. Added --line-buffered to pcregrep.
2168    
2169    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2170        studied, and the match started with a letter with a code point greater than
2171        127 whose first byte was different to the first byte of the other case of
2172        the letter, the other case of this starting letter was not recognized
2173        (#976).
2174    
2175    12. If a pattern that was studied started with a repeated Unicode property
2176        test, for example, \p{Nd}+, there was the theoretical possibility of
2177        setting up an incorrect bitmap of starting bytes, but fortunately it could
2178        not have actually happened in practice until change 8 above was made (it
2179        added property types that matched character-matching opcodes).
2180    
2181    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2182        possible starting bytes for non-anchored patterns.
2183    
2184    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2185        \R, and also a number of cases that involve Unicode properties, both
2186        explicit and implicit when PCRE_UCP is set.
2187    
2188    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2189        input, it could crash or give wrong results if characters with values
2190        greater than 0xc0 were present in the subject string. (Detail: it assumed
2191        UTF-8 input when processing these items.)
2192    
2193    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2194        size_t is 64-bit (#991).
2195    
2196    17. Added a check for running out of memory when PCRE is compiled with
2197        --disable-stack-for-recursion (#990).
2198    
2199    18. If the last data line in a file for pcretest does not have a newline on
2200        the end, a newline was missing in the output.
2201    
2202    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2203        less than 128) in its various bitmaps. However, there is a facility for
2204        generating tables according to the current locale when PCRE is compiled. It
2205        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2206        characters, are recognized by isspace() and therefore were getting set in
2207        these tables, and indeed these tables seem to approximate to ISO 8859. This
2208        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2209        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2210        which of course cannot start UTF-8 characters. I have changed the code so
2211        that only real ASCII characters (less than 128) and the correct starting
2212        bytes for UTF-8 encodings are set for characters greater than 127 when in
2213        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2214        altogether.)
2215    
2216    20. Added the /T option to pcretest so as to be able to run tests with non-
2217        standard character tables, thus making it possible to include the tests
2218        used for 19 above in the standard set of tests.
2219    
2220    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2221        reference to a subpattern the other side of a comment that contains an
2222        opening parenthesis caused either an internal compiling error, or a
2223        reference to the wrong subpattern.
2224    
2225    
2226    Version 8.02 19-Mar-2010
2227    ------------------------
2228    
2229    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2230    
2231    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2232        configured.
2233    
2234    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2235        original author of that file, following a query about its status.
2236    
2237    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2238        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2239    
2240    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2241        quantifier applied to a forward-referencing subroutine call, could compile
2242        incorrect code or give the error "internal error: previously-checked
2243        referenced subpattern not found".
2244    
2245    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2246        variables to point to external functions. For these systems, therefore,
2247        pcre_malloc etc. are now initialized to local functions that call the
2248        relevant global functions.
2249    
2250    7.  There were two entries missing in the vectors called coptable and poptable
2251        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2252        I've fixed the data, and added a kludgy way of testing at compile time that
2253        the lengths are correct (equal to the number of opcodes).
2254    
2255    8.  Following on from 7, I added a similar kludge to check the length of the
2256        eint vector in pcreposix.c.
2257    
2258    9.  Error texts for pcre_compile() are held as one long string to avoid too
2259        much relocation at load time. To find a text, the string is searched,
2260        counting zeros. There was no check for running off the end of the string,
2261        which could happen if a new error number was added without updating the
2262        string.
2263    
2264    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2265    
2266    11. \K was not working if it appeared in an atomic group or in a group that
2267        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2268        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2269        the assertion is positive, but not if it is negative.
2270    
2271    12. Change 11 fortuitously reduced the size of the stack frame used in the
2272        "match()" function of pcre_exec.c by one pointer. Forthcoming
2273        implementation of support for (*MARK) will need an extra pointer on the
2274        stack; I have reserved it now, so that the stack frame size does not
2275        decrease.
2276    
2277    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2278        item in branch that calls a recursion is a subroutine call - as in the
2279        second branch in the above example - was incorrectly given the compile-
2280        time error "recursive call could loop indefinitely" because pcre_compile()
2281        was not correctly checking the subroutine for matching a non-empty string.
2282    
2283    14. The checks for overrunning compiling workspace could trigger after an
2284        overrun had occurred. This is a "should never occur" error, but it can be
2285        triggered by pathological patterns such as hundreds of nested parentheses.
2286        The checks now trigger 100 bytes before the end of the workspace.
2287    
2288    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2289    
2290    
2291    Version 8.01 19-Jan-2010
2292    ------------------------
2293    
2294    1.  If a pattern contained a conditional subpattern with only one branch (in
2295        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2296        computed the wrong minimum data length (which is of course zero for such
2297        subpatterns). This could cause incorrect "no match" results.
2298    
2299    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2300        the pattern is reset in the first branch, pcre_compile() failed with
2301        "internal error: code overflow at offset...". This happened only when
2302        the reset was to the original external option setting. (An optimization
2303        abstracts leading options settings into an external setting, which was the
2304        cause of this.)
2305    
2306    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2307        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2308        assertion pattern did not match (meaning that the assertion was true), it
2309        was incorrectly treated as false if the SKIP had been reached during the
2310        matching. This also applied to assertions used as conditions.
2311    
2312    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2313        assertion subpattern, including such a pattern used as a condition,
2314        unpredictable results occurred, instead of the error return
2315        PCRE_ERROR_DFA_UITEM.
2316    
2317    5.  The C++ GlobalReplace function was not working like Perl for the special
2318        situation when an empty string is matched. It now does the fancy magic
2319        stuff that is necessary.
2320    
2321    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2322        removed. (These were left over from very, very early versions of PCRE.)
2323    
2324    7.  Some cosmetic changes to the code to make life easier when compiling it
2325        as part of something else:
2326    
2327        (a) Change DEBUG to PCRE_DEBUG.
2328    
2329        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2330            called "current" as "current_branch", to prevent a collision with the
2331            Linux macro when compiled as a kernel module.
2332    
2333        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2334            prevent a collision with the Linux macro when compiled as a kernel
2335            module.
2336    
2337    8.  In pcre_compile() there are some checks for integer overflows that used to
2338        cast potentially large values to (double). This has been changed to that
2339        when building, a check for int64_t is made, and if it is found, it is used
2340        instead, thus avoiding the use of floating point arithmetic. (There is no
2341        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2342        double.
2343    
2344    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2345        2005 (difference between two addresses compared to an unsigned value).
2346    
2347    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2348        custom one, because of the following reported problem in Windows:
2349    
2350          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2351              under Win32.
2352          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2353              therefore missing the function definition.
2354          - The compiler thus generates a "C" signature for the test function.
2355          - The linker fails to find the "C" function.
2356          - PCRE fails to configure if asked to do so against libbz2.
2357    
2358    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2359        messages were output:
2360    
2361          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2362          rerunning libtoolize, to keep the correct libtool macros in-tree.
2363          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2364    
2365        I have done both of these things.
2366    
2367    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2368        most of the time, it *can* run out if it is given a pattern that contains a
2369        runaway infinite recursion. I updated the discussion in the pcrestack man
2370        page.
2371    
2372    13. Now that we have gone to the x.xx style of version numbers, the minor
2373        version may start with zero. Using 08 or 09 is a bad idea because users
2374        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2375        interpreted as invalid octal numbers. I've updated the previous comment in
2376        configure.ac, and also added a check that gives an error if 08 or 09 are
2377        used.
2378    
2379    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2380        causing partial matching to fail when the end of the subject matched \W
2381        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2382    
2383    15. There were some discrepancies between the declarations in pcre_internal.h
2384        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2385        their definitions. The declarations used "const uschar *" and the
2386        definitions used USPTR. Even though USPTR is normally defined as "const
2387        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2388        reported that: "This difference in casting confuses some C++ compilers, for
2389        example, SunCC recognizes above declarations as different functions and
2390        generates broken code for hbpcre." I have changed the declarations to use
2391        USPTR.
2392    
2393    16. GNU libtool is named differently on some systems. The autogen.sh script now
2394        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2395        (FreeBSD).
2396    
2397    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2398        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2399        comment: "Figure out how to create a longlong from a string: strtoll and
2400        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2401        instance, but it only takes 2 args instead of 3!"
2402    
2403    18. A subtle bug concerned with back references has been fixed by a change of
2404        specification, with a corresponding code fix. A pattern such as
2405        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2406        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2407        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2408        same bug. Such groups have to be quantified to be useful, or contained
2409        inside another quantified group. (If there's no repetition, the reference
2410        can never match.) The problem arises because, having left the group and
2411        moved on to the rest of the pattern, a later failure that backtracks into
2412        the group uses the captured value from the final iteration of the group
2413        rather than the correct earlier one. I have fixed this in PCRE by forcing
2414        any group that contains a reference to itself to be an atomic group; that
2415        is, there cannot be any backtracking into it once it has completed. This is
2416        similar to recursive and subroutine calls.
2417    
2418    
2419    Version 8.00 19-Oct-09
2420    ----------------------
2421    
2422    1.  The table for translating pcre_compile() error codes into POSIX error codes
2423        was out-of-date, and there was no check on the pcre_compile() error code
2424        being within the table. This could lead to an OK return being given in
2425        error.
2426    
2427    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2428        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2429        in a Windows environment.
2430    
2431    3.  The pcregrep --count option prints the count for each file even when it is
2432        zero, as does GNU grep. However, pcregrep was also printing all files when
2433        --files-with-matches was added. Now, when both options are given, it prints
2434        counts only for those files that have at least one match. (GNU grep just
2435        prints the file name in this circumstance, but including the count seems
2436        more useful - otherwise, why use --count?) Also ensured that the
2437        combination -clh just lists non-zero counts, with no names.
2438    
2439    4.  The long form of the pcregrep -F option was incorrectly implemented as
2440        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2441        but it seems right to fix it, and I didn't think it was worth preserving
2442        the old behaviour.
2443    
2444    5.  The command line items --regex=pattern and --regexp=pattern were not
2445        recognized by pcregrep, which required --regex pattern or --regexp pattern
2446        (with a space rather than an '='). The man page documented the '=' forms,
2447        which are compatible with GNU grep; these now work.
2448    
2449    6.  No libpcreposix.pc file was created for pkg-config; there was just
2450        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2451    
2452    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2453        when UCP support is not needed, by modifying the Python script that
2454        generates it from Unicode data files. This should not matter if the module
2455        is correctly used as a library, but I received one complaint about 50K of
2456        unwanted data. My guess is that the person linked everything into his
2457        program rather than using a library. Anyway, it does no harm.
2458    
2459    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2460        was a minimum greater than 1 for a wide character in a possessive
2461        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2462        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2463        character. Chaos in the form of incorrect output or a compiling loop could
2464        result.
2465    
2466    9.  The restrictions on what a pattern can contain when partial matching is
2467        requested for pcre_exec() have been removed. All patterns can now be
2468        partially matched by this function. In addition, if there are at least two
2469        slots in the offset vector, the offset of the earliest inspected character
2470        for the match and the offset of the end of the subject are set in them when
2471        PCRE_ERROR_PARTIAL is returned.
2472    
2473    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2474        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2475        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2476        and may be more useful for multi-segment matching.
2477    
2478    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2479        used to be given if ever the end of the subject was reached; now it is
2480        given only if matching could not proceed because another character was
2481        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2482        string "Z", which now yields "no match" instead of "partial match". In the
2483        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2484        final character ended with (*FAIL).
2485    
2486    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2487        if the pattern had a "must contain" character that was already found in the
2488        earlier partial match, unless partial matching was again requested. For
2489        example, with the pattern /dog.(body)?/, the "must contain" character is
2490        "g". If the first part-match was for the string "dog", restarting with
2491        "sbody" failed. This bug has been fixed.
2492    
2493    13. The string returned by pcre_dfa_exec() after a partial match has been
2494        changed so that it starts at the first inspected character rather than the
2495        first character of the match. This makes a difference only if the pattern
2496        starts with a lookbehind assertion or \b or \B (\K is not supported by
2497        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2498        matching functions compatible, and I think it's the right thing to do.
2499    
2500    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2501        so that the demonstration program is easily available in environments where
2502        PCRE has not been installed from source.
2503    
2504    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2505        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2506        library.
2507    
2508    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2509        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2510        is not the first non-POSIX option to be added. Clearly some people find
2511        these options useful.
2512    
2513    17. If a caller to the POSIX matching function regexec() passes a non-zero
2514        value for nmatch with a NULL value for pmatch, the value of
2515        nmatch is forced to zero.
2516    
2517    18. RunGrepTest did not have a test for the availability of the -u option of
2518        the diff command, as RunTest does. It now checks in the same way as
2519        RunTest, and also checks for the -b option.
2520    
2521    19. If an odd number of negated classes containing just a single character
2522        interposed, within parentheses, between a forward reference to a named
2523        subpattern and the definition of the subpattern, compilation crashed with
2524        an internal error, complaining that it could not find the referenced
2525        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2526        [The bug was that it was starting one character too far in when skipping
2527        over the character class, thus treating the ] as data rather than
2528        terminating the class. This meant it could skip too much.]
2529    
2530    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2531        /g option in pcretest when the pattern contains \K, which makes it possible
2532        to have an empty string match not at the start, even when the pattern is
2533        anchored. Updated pcretest and pcredemo to use this option.
2534    
2535    21. If the maximum number of capturing subpatterns in a recursion was greater
2536        than the maximum at the outer level, the higher number was returned, but
2537        with unset values at the outer level. The correct (outer level) value is
2538        now given.
2539    
2540    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2541        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2542        make it do so. The string so far is captured, making this feature
2543        compatible with Perl.
2544    
2545    23. The tests have been re-organized, adding tests 11 and 12, to make it
2546        possible to check the Perl 5.10 features against Perl 5.10.
2547    
2548    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2549        pattern matches a fixed length string. PCRE did not allow this; now it
2550        does. Neither allows recursion.
2551    
2552    25. I finally figured out how to implement a request to provide the minimum
2553        length of subject string that was needed in order to match a given pattern.
2554        (It was back references and recursion that I had previously got hung up
2555        on.) This code has now been added to pcre_study(); it finds a lower bound
2556        to the length of subject needed. It is not necessarily the greatest lower
2557        bound, but using it to avoid searching strings that are too short does give
2558        some useful speed-ups. The value is available to calling programs via
2559        pcre_fullinfo().
2560    
2561    26. While implementing 25, I discovered to my embarrassment that pcretest had
2562        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2563        study optimizations had never been tested with that matching function.
2564        Oops. What is worse, even when it was passed study data, there was a bug in
2565        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2566        were also very few tests of studied patterns with pcre_dfa_exec().
2567    
2568    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2569        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2570        on the other side of the coin, they are no longer allowed to have different
2571        names, because these cannot be distinguished in PCRE, and this has caused
2572        confusion. (This is a difference from Perl.)
2573    
2574    28. When duplicate subpattern names are present (necessarily with different
2575        numbers, as required by 27 above), and a test is made by name in a
2576        conditional pattern, either for a subpattern having been matched, or for
2577        recursion in such a pattern, all the associated numbered subpatterns are
2578        tested, and the overall condition is true if the condition is true for any
2579        one of them. This is the way Perl works, and is also more like the way
2580        testing by number works.
2581    
2582    
2583    Version 7.9 11-Apr-09
2584    ---------------------
2585    
2586    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2587        (pcretest), all targets were linked against these libraries. This included
2588        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2589        libraries. This caused unwanted dependencies to be created. This problem
2590        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2591        pcretest is linked with readline.
2592    
2593    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2594        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2595        moved outside it again, because FALSE and TRUE are already defined in AIX,
2596        but BOOL is not.
2597    
2598    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2599        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2600    
2601    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2602        hyphens) following file names and line numbers when outputting matching
2603        lines. This is not true; no spaces are inserted. I have also clarified the
2604        wording for the --colour (or --color) option.
2605    
2606    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2607        was not coloured; this is different to GNU grep, so I have changed it to be
2608        the same.
2609    
2610    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2611        each matching line was coloured. Now it goes on to look for further matches
2612        of any of the test patterns, which is the same behaviour as GNU grep.
2613    
2614    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2615        doesn't make sense to accept an empty string match in pcregrep, so I have
2616        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2617        seems to be how GNU grep behaves. [But see later change 40 for release
2618        8.33.]
2619    
2620    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2621        start or after a newline", because the conditional assertion was not being
2622        correctly handled. The rule now is that both the assertion and what follows
2623        in the first alternative must satisfy the test.
2624    
2625    9.  If auto-callout was enabled in a pattern with a conditional group whose
2626        condition was an assertion, PCRE could crash during matching, both with
2627        pcre_exec() and pcre_dfa_exec().
2628    
2629    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2630        used for matching.
2631    
2632    11. Unicode property support in character classes was not working for
2633        characters (bytes) greater than 127 when not in UTF-8 mode.
2634    
2635    12. Added the -M command line option to pcretest.
2636    
2637    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2638    
2639    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2640    
2641    16. Added comments and documentation about mis-use of no_arg in the C++
2642        wrapper.
2643    
2644    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2645        from Martin Jerabek that uses macro names for all relevant character and
2646        string constants.
2647    
2648    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2649        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2650        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2651        these, but not everybody uses configure.
2652    
2653    19. A conditional group that had only one branch was not being correctly
2654        recognized as an item that could match an empty string. This meant that an
2655        enclosing group might also not be so recognized, causing infinite looping
2656        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2657        with the subject "ab", where knowledge that the repeated group can match
2658        nothing is needed in order to break the loop.
2659    
2660    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2661        exec(), but without supplying a callout function, matching went wrong.
2662    
2663    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2664        leak if the size of the offset vector was greater than 30. When the vector
2665        is smaller, the saved offsets during recursion go onto a local stack
2666        vector, but for larger vectors malloc() is used. It was failing to free
2667        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2668        error, in fact).
2669    
2670    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2671        heapframe that is used only when UTF-8 support is enabled. This caused no
2672        problem, but was untidy.
2673    
2674    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2675        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2676        included within another project.
2677    
2678    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2679        slightly modified by me:
2680    
2681          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2682              not building pcregrep.
2683    
2684          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2685              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2686    
2687    25. Forward references, both numeric and by name, in patterns that made use of
2688        duplicate group numbers, could behave incorrectly or give incorrect errors,
2689        because when scanning forward to find the reference group, PCRE was not
2690        taking into account the duplicate group numbers. A pattern such as
2691        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2692    
2693    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2694        the feature of a custom pointer more persuasive (as requested by a user).
2695    
2696    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2697        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2698        pre-defined.
2699    
2700    28. Added support for (*UTF8) at the start of a pattern.
2701    
2702    29. Arrange for flags added by the "release type" setting in CMake to be shown
2703        in the configuration summary.
2704    
2705    
2706    Version 7.8 05-Sep-08
2707    ---------------------
2708    
2709    1.  Replaced UCP searching code with optimized version as implemented for Ad
2710        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2711        stage table and inline lookup instead of a function, giving speed ups of 2
2712        to 5 times on some simple patterns that I tested. Permission was given to
2713        distribute the MultiStage2.py script that generates the tables (it's not in
2714        the tarball, but is in the Subversion repository).
2715    
2716    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2717        scripts.
2718    
2719    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2720        a group with a zero qualifier. The result of the study could be incorrect,
2721        or the function might crash, depending on the pattern.
2722    
2723    4.  Caseless matching was not working for non-ASCII characters in back
2724        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2725        It now works when Unicode Property Support is available.
2726    
2727    5.  In pcretest, an escape such as \x{de} in the data was always generating
2728        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2729        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2730        truncation.
2731    
2732    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2733    
2734    7.  Added two (int) casts to pcregrep when printing the difference of two
2735        pointers, in case they are 64-bit values.
2736    
2737    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2738        test 2 if it fails.
2739    
2740    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2741        and a #define of that name to empty if it is not externally set. This is to
2742        allow users of MSVC to set it if necessary.
2743    
2744    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2745        the convenience functions in the pcre_get.c source file.
2746    
2747    11. An option change at the start of a pattern that had top-level alternatives
2748        could cause overwriting and/or a crash. This command provoked a crash in
2749        some environments:
2750    
2751          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2752    
2753        This potential security problem was recorded as CVE-2008-2371.
2754    
2755    12. For a pattern where the match had to start at the beginning or immediately
2756        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2757        pcre_dfa_exec() could read past the end of the passed subject if there was
2758        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2759        pcretest so that it places the subject at the end of its malloc-ed buffer.
2760    
2761    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2762        exec() might read past the end of the data buffer in UTF-8 mode.
2763    
2764    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2765        the data contained the byte 0x85 as part of a UTF-8 character within its
2766        first line. This applied both to normal and DFA matching.
2767    
2768    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2769        /^[^d]*?$/8 failed to match "abc".
2770    
2771    16. Added a missing copyright notice to pcrecpp_internal.h.
2772    
2773    17. Make it more clear in the documentation that values returned from
2774        pcre_exec() in ovector are byte offsets, not character counts.
2775    
2776    18. Tidied a few places to stop certain compilers from issuing warnings.
2777    
2778    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2779        supplied by Stefan Weber. I made a further small update for 7.8 because
2780        there is a change of source arrangements: the pcre_searchfuncs.c module is
2781        replaced by pcre_ucd.c.
2782    
2783    
2784    Version 7.7 07-May-08
2785    ---------------------
2786    
2787    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2788        a string to a long long, pretend we don't even have a long long." This is
2789        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2790    
2791    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2792        pre-7.6 versions, which defined a global no_arg variable instead of putting
2793        it in the RE class. (See also #8 below.)
2794    
2795    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2796        Lopes.
2797    
2798    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2799    
2800        (1) The include/exclude patterns were being applied to the whole pathnames
2801            of files, instead of just to the final components.
2802    
2803        (2) If there was more than one level of directory, the subdirectories were
2804            skipped unless they satisfied the include/exclude conditions. This is
2805            inconsistent with GNU grep (and could even be seen as contrary to the
2806            pcregrep specification - which I improved to make it absolutely clear).
2807            The action now is always to scan all levels of directory, and just
2808            apply the include/exclude patterns to regular files.
2809    
2810    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2811        --exclude_dir in the tests to avoid scanning .svn directories.
2812    
2813    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2814        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2815        doesn't support NULs in patterns.
2816    
2817    7.  Added some missing "const"s to declarations of static tables in
2818        pcre_compile.c and pcre_dfa_exec.c.
2819    
2820    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2821        caused by fix #2  above. (Subsequently also a second patch to fix the
2822        first patch. And a third patch - this was a messy problem.)
2823    
2824    9.  Applied Craig's patch to remove the use of push_back().
2825    
2826    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2827        matching function regexec().
2828    
2829    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2830        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2831        references. PCRE supports relative numbers with this syntax (I don't think
2832        Oniguruma does).
2833    
2834    12. Previously, a group with a zero repeat such as (...){0} was completely
2835        omitted from the compiled regex. However, this means that if the group
2836        was called as a subroutine from elsewhere in the pattern, things went wrong
2837        (an internal error was given). Such groups are now left in the compiled
2838        pattern, with a new opcode that causes them to be skipped at execution
2839        time.
2840    
2841    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2842        to the way PCRE behaves:
2843    
2844        (a) A lone ] character is dis-allowed (Perl treats it as data).
2845    
2846        (b) A back reference to an unmatched subpattern matches an empty string
2847            (Perl fails the current match path).
2848    
2849        (c) A data ] in a character class must be notated as \] because if the
2850            first data character in a class is ], it defines an empty class. (In
2851            Perl it is not possible to have an empty class.) The empty class []
2852            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2853            The negative empty class [^] matches any one character, independently
2854            of the DOTALL setting.
2855    
2856    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2857        non-existent subpattern following a character class starting with ']' and
2858        containing () gave an internal compiling error instead of "reference to
2859        non-existent subpattern". Fortunately, when the pattern did exist, the
2860        compiled code was correct. (When scanning forwards to check for the
2861        existence of the subpattern, it was treating the data ']' as terminating
2862        the class, so got the count wrong. When actually compiling, the reference
2863        was subsequently set up correctly.)
2864    
2865    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2866        it was being rejected as not supported by pcre_dfa_exec(), even though
2867        other assertions are supported. I have made pcre_dfa_exec() support
2868        (*FAIL).
2869    
2870    16. The implementation of 13c above involved the invention of a new opcode,
2871        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2872        cannot be changed at match time, I realized I could make a small
2873        improvement to matching performance by compiling OP_ALLANY instead of
2874        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2875        on the OP_ANY path.
2876    
2877    17. Compiling pcretest on Windows with readline support failed without the
2878        following two fixes: (1) Make the unistd.h include conditional on
2879        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2880    
2881    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2882        ncurses library to be included for pcretest when ReadLine support is
2883        requested, but also to allow for it to be overridden. This patch came from
2884        Daniel Bergström.
2885    
2886    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2887        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2888        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2889        spotting this.
2890    
2891    
2892  Version 7.6 28-Jan-08  Version 7.6 28-Jan-08
2893  ---------------------  ---------------------
2894    

Legend:
Removed from v.319  
changed lines
  Added in v.1688

  ViewVC Help
Powered by ViewVC 1.1.5