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

Legend:
Removed from v.319  
changed lines
  Added in v.1647

  ViewVC Help
Powered by ViewVC 1.1.5